roadmap2-08 next generation hydrocarbon bio refineries

BASED ON THE JUNE 25-26,2007 WORKSHOPWASHINGTON, D.C.

A RESEARCH ROADMAP FOR MAKING LIGNOCELLULOSIC BIOFUELS

A PRACTICAL REALITY

UNIVERSITY OF

MASSACHUSETTSAMHERST

SPONSORED BY:

Breaking the Chemical and Engineering Barriers to

Lignocellulosic Biofuels:

Next Generation Hydrocarbon Biorefineries

THE NATIONAL SCIENCE FOUNDATION

AMERICAN CHEMICALSOCIETY

THE DEPARTMENT OF ENERGY

Publication Date: March 2008 Suggested citation for this document:NSF. 2008. Breaking the Chemical and Engineering Barriers to Lignocellulosic Biofuels: Next Generation Hydrocarbon Biorefineries.

Ed. George W. Huber, University of Massachusetts Amherst. National Science Foundation. Chemical, Bioengineering,Environmental, and Transport Systems Division. Washington D.C. 180 p.

Breaking the Chemical and Engineering Barriers to

Lignocellulosic Biofuels:Next Generation

Hydrocarbon Biorefineries

BASED ON THEJUNE 25-26, 2007

WORKSHOPWASHINGTON, D.C.

WORKSHOP CHAIR:GEORGE W. HUBER

UNIVERSITY OFMASSACHUSETTS

AMHERST

Roadmap 2007 • Breaking the Chemical and Engineering Barriers to Lignocellulosic Biofuels2

Liquid biofuels produced from lignocellulosic biomass can significantly reduce our dependenceon foreign oil, create new jobs, improve ruraleconomies, reduce greenhouse gas emissions,

and improvenational security. Therehas been deepbipartisan support formeasures suchas the Vehicleand FuelChoices forAmericanSecurity Act. Inhis 2006 Stateof the Unionaddress, thePresident notedthat “WithAmerica on theverge ofbreakthroughs

in advanced energy technologies the best way tobreak the addiction to foreign oil is through new technologies.”

Advances in agriculture and biotechnology havemade it possible to inexpensively produce lignocellulosic biomass at costs that are significantlylower (about $15 per barrel of oil energyequivalent) than crude oil. Significant amounts oflignocellulosic biomass can be sustainably producedon US agriculture and forestry land with the energy content of 60 % of the current US petroleum consumption. The key bottleneck forlignocellulosic-derived biofuels is the lack of technology for the efficient conversion of biomassinto liquid fuels.

While the U.S. has made a significant investmentin technologies focusing on breaking the biologicalbarriers to biofuels, principally ethanol, there hasnot been a commensurate investment in theresearch needed to break the chemical andengineering barriers to hydrocarbon fuels such asgasoline, diesel, and jet fuel.

The production of hydrocarbon fuels frombiomass has many important advantages. First,“green” hydrocarbon fuels are essentially the sameas those currently derived from petroleum, exceptthat they are made from biomass.Therefore, it willnot be necessary to modify existing infrastructure(e.g. pipelines, engines) and hydrocarbon biorefiningprocesses can be tied into the fuel productionsystems of existing petroleum refineries. Second,biomass-based hydrocarbon fuels are energyequivalent to fuels derived from petroleum. Incontrast to the lower energy density of E85 flexfuel, there will be no penalty in gas mileage withbiomass-based hydrocarbon fuels.Third,hydrocarbons produced from lignocellulosicbiomass are immiscible in water; they self-separate,which eliminates the need for an expensive,energy-consuming distillation step. Fourth, biomass-based hydrocarbon fuels are produced at hightemperatures, which allows for faster conversionreactions in smaller reactors. Thus, processingunits can be placed close to the biomass source oreven transported on truck trailers. Fifth, theamount of water needed for processinghydrocarbon fuels from biomass can be greatlyreduced, compared with the dilute sugar solutionsto which enzymes are constrained. This is becauseorganic or heterogeneous catalysts work well inconcentrated water solutions or even in theabsence of water if ionic liquids are used. Finally,heterogeneous catalysts are inherently recyclable.So they can be used over the course of monthsand even years, which significantly reduces costscompared to biological catalysts.The elimination ofenergy-intensive distillation, the higher reaction

E xe c u t i ve Summary

“With America on the

verge of breakthroughs

in advanced energy

technologies – the best

way to break the

addiction to foreign

oil is through new

technologies.”

- President George W. Bush

State of the Union

Address/January 31, 2006

Next Generation Hydrocarbon Biorefineries3

rates, and the much smaller process footprints canalso lead to lower biofuel costs than are possibleusing currently available biological pathways forproducing cellulosic ethanol.

To articulate the essential role of chemistry,chemical catalysis, thermal processing, andengineering in the conversion of lignocellulosicbiomass into green gasoline, green diesel and greenjet fuel, the National Science Foundation and theDepartment of Energy convened a workshopentitled “Breaking the Chemical and EngineeringBarriers to Lignocellulosic Biofuels” on June 25-26,2007 in Washington D.C., ancillary to the 2007 ACSGreen Chemistry and Engineering Conference.Over 70 participants from 24 academic institutions,20 petroleum, chemical and biofuel companies, and7 national laboratories contributed expertise inchemical catalysis, chemistry, petroleum refining,nanotechnology, quantum chemistry, andengineering. This document is the result of thatworkshop.

This roadmap to “Next Generation HydrocarbonBiorefineries” outlines a number of novel process

pathways for biofuels productionbased on sound scientific and

engineering proofs of concept demonstrated in laboratories

around the world.Roadmap highlights are as follows:

Selective thermalprocessing of lignocellulosicbiomass to produce liquidfuels (bio-oils) indistributed biorefineries(Chapter 1).

Utilization of petroleumrefining technology for

conversion of biomass-derivedoxygenates within existing

petroleum refineries (Chapter 2).

Hydrocarbon production by liquid phaseprocessing of sugars to a heretofore“sleeping giant” intermediate,hydroxymethylfurfural (HMF), followed byHMF conversion to “green” diesel and jetfuel (Chapter 3).

Process intensification for diesel and gasoline production from synthesis gas (CO and H2) by Fisher-Tropsch synthesis (FTS), which dramatically decreases the economically viable size compared to traditional FTS processes with petroleumderived feedstocks (Chapter 4).

Conceptual design of biorefining processesin conjunction with experimental studies atthe beginning of research projects to allowrapid development of commercial biofueltechnologies (Chapter 5).

Design of recyclable, highly active and selective heterogeneous catalysts for biofuelproduction using advanced nanotechnology,synthesis methods and quantum chemicalcalculations (Chapter 6).

Years of engineering research and design wentinto the development of the modern petroleumindustry. A similarly expansive and sustained effortmust also be applied to develop hydrocarbonbiorefineries.Advances in nanoscience over the last several decades have given us anunprecedented ability to understand and controlchemistry at the molecular scale, which promisesto accelerate the development of biomass-to-fuelsproduction technologies.As the tremendousexpertise of the chemistry, catalysis andengineering communities—which was soinstrumental in the development of petroleumrefining technologies—is brought fully to bear onthis field as well, there is every reason to believethat we can rapidly develop cost-effectivehydrocarbon biorefineries.


Ta ble of Contents

EXECUTIVE SUMMARY 2

I N T RODUCTION 6

0.1 Workshop Challenges the Interdisciplinary Community of Biofuels Scientists and Engineers 7

0.2 America’s Energy Challenge 80.3 The Promise of Lignocellulosic Biofuels 90.4 A National Mandate for Biofuels Production 100.5 The Feasibility of Producing Fuels from Lignocellulosic Biomass 130.6 Biofuels from the Catalytic Processing of Biomass 150.7 Designing Biofuels Production Processes 200.8 Next Generation Biorefineries for Production of Liquid Fuels 230.9 Overview of Roadmap and Workshop 25

THRUST 1. SELECTIVE THERMAL PROCESSING OF CELLULOSIC BIOMASS AND LIGNIN 3 0

1.1 Introduction 3 01.2 Overall Process Description 3 11.3 Resulting Fuels 3 51.4 Summary of Previous Research 3 71.5 Economics and Potential of Technology 3 91.6 Current Technology Limitations and Research/Development Needs 4 11.7 Recommendations 4 21.8 References 4 7

THRUST 2. U T I L I Z ATION OF PETROLEUM REFINING TECHNOLOGIES FOR BIOFUEL PRO D U C T I O N 5 0

2.1 Introduction 5 02.2 Overall Process Description 5 12.3 Resulting Fuels 5 12.4 Summary of Previous Research 5 32.5 Economics and Potential of Technology 5 72.6 Current Technology Limitations and Research/Development Needs 5 82.7 Recommendations 6 22.8 References 6 4

THRUST 3. LIQUID-PHASE CATA LYTIC PROCESSING OF BIOMASS-DERIVED COMPOUNDS 66

3.1 Introduction 6 63.2 Process Description and Resulting Biofuels 6 63.3 Feedstocks 6 73.4 Review of Catalytic Reactions for Liquid Phase Processing 6 83.5 Advantages of Liquid-Phase Processing 8 03.6 Current Technology Limitations and Research/Development Needs 8 43.7 Recommendations 8 83.8 References 8 9


THRUST 4. CATA LYTIC CONVERSION OF SYNGAS 9 2

4.1 Introduction 9 24.2 Overall Process Description 9 34.3 Resulting Fuels 9 54.4 Summary of Previous Research 9 64.5 Economics and Potential of Technology 1 0 84.6 Current Technology Limitations and Research/Development Needs 1 0 94.7 Recommendations 1 1 94.8 References 1 2 0

THRUST 5. P ROCESS ENGINEERING & DESIGN 1 2 4

5.1 Introduction 1 2 45.2 Process Analysis 1 2 45.3 Priorities for Research & Development 1 2 65.4 Summary 1 3 25.5 References 1 3 3

THRUST 6. C ROSSCUTTING SCIENTIFIC ISSUES 1 3 4

6.1 Analytical Database for Biomolecules 1 3 56.2 Thermodynamics 1 3 76.3 Chemical Reaction Engineering 1 3 86.4 Catalyst Engineering 1 4 16.5 Catalyst Characterization 1 4 46.6 Computational Catalysis 1 4 96.7 Recommendations 1 5 56.8 References 1 5 6

APPENDICES 1 5 2

Participants 160Organizational Committee 164Participant Biographies 166


A CONCERTED EFFORT TO ACCELERATE THE DEVELOPMENT OF DOMESTICALLYPRODUCED ALTERNATIVE TRANSPORTATIONFUELS PROMISES TO REDUCE OUR NATIONALDEPENDENCE ON FOREIGN OIL, SPURECONOMIC DEVELOPMENT,AND IMPROVEENVIRONMENTAL QUALITY IN THE UNITEDSTATES. In his 2007 State of the Union address,President Bush proposed to increase the domesticsupply of alternative fuels to 35 million gallonsannually over the next 10 years [White House2007].The U.S. Department of Energy hasdetermined that more than 1 billion dry tons ofbiomass could be sustainably harvested from U.S.fields and forests, enough to displace 30 percent ofthe nation's annual petroleum consumption fortransportation fuels [Perlack, et al. 2005]. TheNatural Resources Defense Council has furtherprojected that an aggressive plan to producelignocellulosic biofuels in the US could produce theequivalent of nearly 7.9 million barrels of oil perday by 2050, or more than 50 percent of currenttotal oil use in the transportation sector, and morethan three times total Persian Gulf oil importsalone [Greene, et al. 2004].

The promise of these domestic fuel productiongoals, strategies, and projections will only berealized through technological advances derivedfrom concerted and long-term programs ininterdisciplinary science and engineering. AsPresident Bush himself said, “With America on theverge of breakthroughs in advanced energytechnologies the best way to break the addictionto foreign oil is through new technologies.”

In order for biofuels to emerge as a viablealternative to petroleum-based fuels we mustdevelop cost-effective production technologies.The obvious national imperative of the problemhas produced a groundswell of research activity in

laboratories across the country. The range of high-tech analytical tools, sophisticated software, andpowerful computers now available to the scientificcommunity allow researchers to “see” chemicalreactions as they unfold. This unprecedentedanalytical capacity presents exciting opportunitiesto fine tune biomass conversion reactions andengineer efficient and economical processes forbiofuels production. Today, we have both theincentive and the technical means to solve thisproblem. Through concerted effort and a sustainedinvestment in research we will solve it.

I n t ro d u c t i o n

“In challenge, there is great opportunity. We must seize theinitiative – as past generations ofAmericans have done – and harness the unique American ingenuity that has made us theworld leader in innovation, invention and technology. If wedo, we can achieve real energyindependence that strengthens ournational security, boosts our economy, creates more jobs andprotects our environment for future generations.”

-Steny Hoyer, Majority LeaderU.S. House of Representatives.

March 1, 2007

0.1 JOINT WORKSHOP CHALLENGESTHE INTERDISCIPLINARY COMMUNITY OF BIOFUEL SCIENTISTS AND ENGINEERS

On June 25th and 26th 2007 more than 70 leadingbiofuels scientists and engineers representingindustry, academia, and government agenciesgathered in Washington D.C. to develop a roadmapfor making lignocellulosic biofuels a practical reality.The objective of the workshop was to articulatethe critical role that chemistry, chemical catalysis,thermal processing, and engineering play in theconversion of lignocellulosic biomass into liquidtransportation fuels including “green” gasoline,diesel, and jet fuel. Workshop participants includeda range of academic, government, and industryscientists and engineers with backgrounds inpetroleum and petrochemical production, catalysis,chemistry, process design and engineering, andbiorefining. An array of companies participated inthe workshop, including representatives from thefollowing industries: oil (Conoco-Phillips, Exxon-Mobil, British Petroleum, Chevron, UOP), chemical(Dow, Dupont, GE Plastics,ABB Lumus, Symyx),venture capital (Khosla Ventures), agriculture(Cargill), engine (United Technology) and smallbusinesses (Virent Energy Systems, Renewable OilInternational). The following five nationallaboratories were also represented: NationalRenewable Energy Laboratory (NREL), PacificNational Northwest Laboratory (PNNL), OakRidge National Lab, Brookhaven National Lab, andSavannah River National Lab. Attendees alsoincluded professors from twenty-four U.S.universities. Each workshop thrust area was co-led by a professor and a national laboratoryrepresentative . The substantial interest in thisworkshop, evidenced by the diversity ofparticipants, demonstrates the vital importance ofthis topic to our nation.

The following roadmap report paints a picture ofconcerted, transformational science. It outlines afar-reaching research program focused on thedevelopment of fundamental knowledge aboutbiomass feedstock properties and reactionproducts, which will lay the groundwork for thebiorefineries of the future. At this stage, there aremany more questions than answers, but thetremendous potential for domestic production ofessential fuels and products compels us to workdiligently to develop the technologies necessary torealize this potential.


SECTIONS OF INTRO D U C T I O N :

0.1 Workshop Challenges theInterdisciplinary Community of Biofuels Scientists and Engineers

0.2 America’s Energy Challenge

0.3 The Promise of LignocellulosicBiofuels

0.4 A National Mandate for BiofuelsProduction

0.5 The Feasibility of Producing Fuelsfrom Lignocellulosic Biomass

0.6 Biofuels from the CatalyticProcessing of Biomass

0.7 Designing Biofuels ProductionProcesses

0.8 Next Generation Biorefineries forProduction of Liquid Fuels

0.9 Overview of Roadmap and Workshop


Roadmap 2007 • Breaking the Chemical and Engineering Barriers to Lignocellulosic Biofuels

0.2 A M E R I CA’S ENERGY C H A L L E N G E

It is now widely acknowledged that the currentU.S. energy supply-demand problem and the myriadenvironmental, economic, and social challengesassociated with global climate change are urgent.Our increasing demand for energy, especiallytransportation fuels, far outstrips our domesticproduction capacity, leaving the nation vulnerableto the political, economic, and national securityconsequences of importing foreign oil. Thegreenhouse gas emissions produced in large partby fossil fuel combustion may cause unprecedentedclimactic upheavals worldwide. To head off thepotentially severe consequences of these trends anew national energy research program must beinitiated with the intensity and commitment of theManhattan Project or President Kennedy’s Apollolunar landing challenge. Then, this research mustbe sustained until the problems are solved.

World demand for oil now stands at 85.9 millionbarrels a day [International Energy Agency 2007].The U.S. Energy Information Administrationexpects it to reach 117.6 million barrels a day by2030 (Figure 0.1) [Energy InformationAdministration 2007].That extra 31.7 millionbarrels of daily production is the equivalent to theoil production of three Saudi Arabias.Transportation accounts for 94 percent of theprojected increase in liquid fuel consumption.

With crude oil prices currently approaching $100per barrel, and projections for continued heavy useof petroleum for transportation fuels, it isimperative to develop alternatives fuels that can beused in existing cars, trucks, boats, and planes.Through chemical processing and upgrading, the fullrange of transportation fuels including gasoline,diesel, and jet fuel can be produced from biomassfeedstocks.

Delivered Energy Consumption for Transporation

1980-2030 (quadrillion Btu)

0

5

10

15

20

25

30

35

40

45

50

1980 1995 2005 2015 2030

Reference

Low growth

High growth

F i g u re 0.1: P rojections of Pe t roleum Consumption forTr a n s p o rt ation through 2030

[ E n e rgy Information A d m i n i s t r ation 2007]



9

Plant biomass is the only sustainable source oforganic carbon currently available on earth, andbiofuels, are the only sustainable source of liquidfuels [Wyman et al 2005, Klass 2004]. However,the complex nature of biomass and the currentlyill-defined issues related to biomass-to-biofuelsconversion pose a substantial, but notinsurmountable, challenge to the large-scaleproduction and widespread use of biofuels.

0.3 THE PROMISE OF LIGNOCELLULOSIC BIOFUELS

Abundant and inexpensive, lignocellulosic biomassdoes not compete with the production of foodcrops.1 As such, it can serve as a source of fungiblefuels to fill the gap between demand and availabilityof petroleum. The current delivered cost oflignocellulosic biomass is significantly cheaper thancrude oil. However, the cost of biomass in the U.S.varies according to type and region, ranging from$5 to $15 per barrel of oil energy equivalent[Huber, et al. 2006]. Furthermore, the pricedifference between biomass and petroleum wouldbe even greater if the expense of confrontingnegative geopolitical challenges was added to thetotal cost of crude oil.

Biomass is abundant, inexpensive, and does notcompete with the production of food crops.Economically, lignocellulosic biomass has anadvantage over other agriculturally importantbiofuels feedstocks such as corn starch, soybeans,and sugar cane, because it can be produced quicklyand at significantly lower cost than food crops

[Klass 2004]. Lignocellulosic biomass is also animportant component of the major food cropslisted above; it is the non-edible portion of theplant, which is currently underutilized, but could beused for biofuel production. In short, lignocellulosicbiomass holds the key to supplying society’s basicneeds for sustainable production of liquidtransportation fuels without impacting the nation'sfood supply.

Availability of domestic lignocellulosic biomass isnot a limitation. In fact, the U.S. has a large amountof underutilized biomass (Table 0.1). In fact,non-food biomass, including trees, grasses andagricultural residues, constitutes more than 80% ofthe total biomass in the U.S. A 2005 studydetermined that 1.3 billion dry tons of this non-food biomass could be available for large-scalebioenergy and biorefining industries by the middle


1 Lignocellulosic biomass is the fibrous, woody, and generally inedibleportions of plants that are composed of hemicellulose, and lignin—keystructural components of plant cell walls. Cellulose and hemicellulose arecomplex carbohydrates (i.e. long chains of sugar molecules). Lignin is acomplex, noncarbohydrate polymer that binds cellulose and hemicelluloseand gives plant cell walls their rigidity. Dry cellulosic biomass consists ofabout 75% carbohydrates (including cellulose, the most abundant form ofcarbon in the biosphere and hemicellulose, polymers of 5- and 6-carbonsugars) and 25% lignin (a complex ringed aromatic structure).

Ta ble 0.1: Potential U. S . Biomass Resources (adapted from Pe rl a c k , et al. 2 0 0 5 )

Biomass Resources Million Dry Tons per Year

FOREST BIOMASS

Forest Products industry residues 145

Logging and site-clearing residues 64

Forest thinning 60

Fuelwood 52

Urban wood residues 47

Subtotal for Forest Residues 368

AGRICULTURAL BIOMASS

Animal crop residues 428

Perennial crops 377

Misc. process residues, manure 106

Grains 87

Subtotal for Agricultural Resources 998

TOTAL BIOMASS

RESOURCE POTENTIAL 1,366


of the 21st Century [Perlack, et al. 2005]. Thismuch biomass has the energy content of 3.8 billionbarrels of oil; an amount equivalent toapproximately half the oil consumed in the U.S. in2006. Lignocellulosic biomass feedstocks forbiofuels production can be derived from bothforest and agricultural resources. Forest resourcesinclude residues such as tree bark and scrap wood,the logging and site clearing debris that is usuallyleft on site or burned, and urban wood residuesconsisting mainly of municipal solid waste.Agricultural resources consist mainly of cropresidues, which are mostly leaves and stems (e.g.corn stover), from crops grown for food and fibersuch as soybeans, corn, and wheat. Additionally, thestudy included grasses (e.g. switchgrass) and fast-growing trees (e.g. poplars) grown specifically forbioenergy. The results of the so-called “billion-tonstudy” illustrate that ample lignocellulosic biomassresources are readily available for sustainableextraction from U.S. fields and forests.

The limiting factor to biofuels production is simplythat low-cost processing technologies to efficientlyconvert a large fraction of the lignocellulosicbiomass energy into liquid fuels do not yet exist.

0.4 A NATIONAL MANDATE FORDOMESTIC BIOFUELS P RO D U C T I O N

Today, the American public overwhelminglysupports the development of technologies toproduce alternative fuels and reduce greenhousegas emissions [Saad 2007]. The environmental andgeopolitical challenges resulting from the nation’sdependence on fossil fuels, combined withdiminishing petroleum resources, have spurred oursociety to search for renewable sources of liquidfuels and chemicals.

In 2005, the United States consumed about 874million gallons of petroleum per day of which 385million were consumed as gasoline, more than anyother country. The 3.9 million gallons of ethanolproduced domestically in 2005 represents just 1percent of our country’s annual gasolineconsumption [Renewable Fuels Association 2007].Clearly, achieving the President’s goal of reducingthe nation’s gasoline usage 20 percent by 2017 willrequire substantial fuel efficiency and conservationefforts in combination with increased domesticbiofuels production.

"When it comes to energy, wemust think big and lead thefuture. We must be bold; wemust declare independencefrom yesterday’s thinking andinvest in energy solutions of tomorrow"

- Nancy Pelosi, Speaker U.S. House of Representatives

June 28, 2007



11

In the last six years, the average annual price ofregular gasoline has risen from $1.46 to $2.70 pergallon.The price of a crude oil per barrel has morethan doubled from $21-$32 in 2000 to a recordhigh of nearly $100 in the fall of 2007 [EnergyInformation Administration 2007b] (Figure 0.2).From 2003 to 2025, U.S. energy demand isprojected to rise by 35%, an increase much greaterthan the projected growth in domestic production[Energy Information Administration 2007]. Thecurrent level of petroleum consumption in theUnited States, and projections for ever-increasingdemand, has wide-ranging environmentalconsequences and leaves the country vulnerable tosevere economic and social impacts resulting fromdisruptions in oil supplies.

In the challenges associated with our nationaldependence on oil there is also great opportunity.Today, the United States is fundamentally apetroleum-based economy. The daily consumptionof 874 million gallons of petroleum that is used toproduce a vast range of fuels, chemicals, andproducts represents about 3 gallons of petroleum

World Crude Oil PricesHighest Price per year: 1989-2007

(U.S. dollars per barrel)

$0.00

$10.00

$20.00

$30.00

$40.00

$50.00

$60.00

$70.00

$80.00

$90.00

$100.00

1989 1992 1995 1998 2001 2004 2007

F i g u re 0.2: Wo rld Crude Oil Prices, Annual Highs [ E n e rgy Information A d m i n i s t r ation 2007b]

each day for every man, woman, and child in thecountry. Given the enormous demand forpetroleum-based transportation fuels andconsumer goods in the U.S. and around the world,the opportunity to develop alternatives cannot beoverstated. Ultimately, the solution will be manysolutions and will require creative use of existinginfrastructure in parallel with development of anew biorefinery infrastructure.

The benefits of moving toward domesticproduction of fuel from lignocellulosic biomasssources will also yield valuable benefits for nationalsecurity, the economy, and the environment.

N ational Security Benefits of Biofuels

Achieving independence from foreign oil, andthereby making the country less vulnerable topolitical instability in the oil producing regions ofthe Middle East, is perhaps our foremost energyissue. America’s oil consumption accounts forapproximately 25 percent of the global total, yetAmerica holds only 3 percent of the world’s


plants are likely to be prevalent in the biofuelsproduction industry. In contrast to fossil fuels, thequantity of biomass is not confined to certainlocalities and the resource cost of biomass is muchlower than that of crude oil. Therefore, conversionto fuels or fuel precursors must be done on a locallevel to eliminate the expense of transporting low-cost biomass, which would likely limit biomassharvest to a radius of 50-75 miles around theconversion plant; such a plant would produce theliquid fuel equivalent of 10,000-20,000 barrels ofoil/day. Distributed production of fuels fromdomestically grown biomass would reduceinfrastructure vulnerability by not consolidating allextraction and production activities in singlegeographic area. This paradigm of distributedproduction will create good jobs in biorefineriesand may offer the opportunity to found newbio-based industries that will benefit ruraleconomies across the nation.

Biofuels and Fo o dBiofuels can and should be produced sustainablywith food and animal feed as co-products. Ethicaland moral questions arise when edible biomassproducts are converted into biofuels. Therefore,conversion of non-edible biomass is the preferredstrategy for long-term, large-scale biofuelsproduction in the U.S. However, the economicsare currently more favorable for conversion ofedible biomass (e.g. corn starch, soybeans) intofuels due to their chemical structure, which can bemore efficiently processed.Therefore, it isimportant to continue developing technologies forthe cost-effective conversion of non-ediblelignocellulosic biomass into fuels.

Agricultural practices in the U.S. and otherindustrialized countries are very advanced, andmost industrialized regions produce more thanenough food for domestic consumption. Farmersdo not pick the crops based on how efficiently

known oil reserves. Roughly 60% of the 319 billiongallons of petroleum consumed in the U.S. annuallyis imported, with about 13% (~42 billion gallons)coming from Persian Gulf countries [EnergyInformation Administration 2007c].The UnitedStates primarily imports crude oil but also importspetroleum products including gasoline, aviation fuel,and fuel oil. A concerted effort to develop thechemical and engineering methods for cost-effective production of biomass-derived alternativesto conventional transportation fuels couldsignificantly reduce our dependence on foreign oil.

Economic Benefits of BiofuelsP roduction in the United Stat e s

According to the U.S. Council on Competitiveness,energy security and sustainability are key U.S.competitiveness issues because of the direct impactthey have on the productivity of U.S. companiesand the standard of living of all Americans [Councilon Competitiveness 2007]. A vibrantlignocellulosic biofuels industry would create alarge amount of high paying domestic jobs in theagricultural, forest management, and oil/chemicalindustries.

The U.S. Bureau of Labor Statistics (BLS) hasstated that production workers in chemicalmanufacturing, which would be similar to those in abiorefinery, earn more money ($820 weekly) thanin all other manufacturing sectors ($659/week) [USDepartment of Labor 2005].The BLS has furtherstated that the median hourly wages in 2004 weremore than $19/hour for plant operators,maintenance and repair workers, chemicaltechnicians, and chemical equipment operators andtenders. Thus, jobs created in the biofuel industryare likely to be good-paying jobs for skilledworkers.

Because of the variable nature of biomassresources, small-scale, geographically localized



13

they produce edible food products. Insteadfarmers’ goals are to grow crops that maximizetheir income, even though more efficient crops canbe grown.Therefore, to the extent that bioenergycrops can be produced economically anddependably command a good price, they canprovide farmers another market for their products,which could improve the economic situation ofagricultural communities.

E nv i ronmental Benefits of BiofuelsOne of the biggest benefits of biofuels is theassociated reduction in net emissions of CO2.The release of CO2 into the atmosphere is directlyrelated to an increase in temperatures worldwide.According to a landmark report released by theUnited Nation’s Intergovernmental Panel onClimate Change (IPCC), an international bodycomprised of representatives from 113 worldgovernments,“Most of the observed increase inglobally averaged temperatures since the mid-20thcentury is very likely due to the observed increasein anthropogenic greenhouse gas concentrations.”[Climate Change 2007]. The IPCC report furtherstates that burning fossil fuels is one of globalwarming's main drivers.

Projections of rising sea levels and more frequentepisodes of severe weather have prompted policymakers and the general public to demand action tostabilize atmospheric CO2 concentrations. Overthe long term, stabilizing CO2 concentrations inthe atmosphere means reducing emissions close tozero. Fossil fuel combustion releases CO2 into theatmosphere. While the biofuels combustion alsoreleases CO2, it is consumed during subsequentbiomass re-growth. Thus, biofuels made fromlignocellulosic biomass can be carbon-neutraltransportation fuels if efficient processes forbiomass conversion are developed (i.e. the amountof CO2 produced during fuel production andcombustion is equal to the amount of CO2consumed by the biomass during its growth).

0.5 THE FEASIBILITY OF P RODUCING FUELS FROM LIGNOCELLULOSIC BIOMASS

The objective of the research program outlined inthe following sections of this roadmap report is to develop biomass-to-biofuels conversiontechnologies and processes that are cost- andperformance-competitive with petroleum-basedfuels. The nation’s non-food agricultural and forestbiomass provides an ample and underutilizedresource base for development of thesetechnologies. A gradual shift from our petroleum-based economy to a biomass-based economyis possible.

As this workshop report illustrates, it istechnologically feasible to convert lignocellulosicmaterials and organic wastes into biofuels.However, costs have to be lowered and newtechnologies must be demonstrated at acommercial scale (i.e. greater than 150,000 metrictons per year). If this is achieved, second-generation biofuels (i.e. fuels made from non-foodbiomass feedstocks) will secure a higher marketshare by allowing the use of a wider range of rawmaterials. Moreover, the cultivation process forbioenergy crops could be environmentally lessintensive than for ordinary agricultural crops, withcorresponding cost decreases and lowergreenhouse gas emissions.

A concerted program of scientific research basedon the recommendations described herein, will setbiofuels on a trajectory toward becoming a majorU.S. industry sector during the first half of the 21stCentury. The projections for development of amature biofuels industry resembles an acceleratedversion of the petroleum industry’s developmentover the 20th Century. As with the petroleumindustry, the biofuels production sector will bedriven by intensive research and development, but


growth of this new bio-based industry will beaccelerated by the innovative and powerful toolsnow available.

In the past two decades, many new techniques have become available to probe the catalyst andcatalyst-reaction medium interface, includingvarious high-resolution electron microscopies, laserspectroscopies, and high field nuclear magneticresonance spectroscopy. Such techniques haveelevated our understanding of the structure-property relationship of biomass-conversioncatalytic reactions to a higher level. Together, theseand other related tools are helping researchers tounravel the secrets of chemical reactions at themolecular scale. These analytical tools, combinedwith powerful computer simulations of catalyticreaction conditions, can be used to elucidatereaction mechanisms and design new catalysts withthe necessary properties to create high-performance biofuels from a range of biomassfeedstocks. However, the complexity of the

reaction systems for biomass conversion to fuelshas posed new challenges that can be met only bystretching the limits of these tools and techniquesand developing new ones.

The projections for accelerated growth andmaturation of the emerging biofuels industry mustbegin with a prolonged period of investigation intothe fundamental properties of biofuels feedstocksand their chemical intermediates. This foundationalwork must then be followed by construction of thenecessary infrastructure and/or adaptation ofexisting petroleum refining and transportationinfrastructure for biofuels production.

Carbon dioxide, water, light, air, and nutrients arethe inputs for biofuel production. Energy to powertransportation vehicles and food are the outputs.The three main technology areas that need to beimproved in order to realize this vision of a maturebiorefining industry are: (1) growth of the biomassfeedstock; (2) biomass conversion into a fuel; and

F i g u re 0.3 The three primary needs (biomass grow t h , fuel p roduction and fuel utilization) for the biofuel economy.

FUELUTILIZATION Energy

FUELPRODUCTION Energy

BIOMASSGROWTH

CO2

H2O

Light

Air

SEPARATION

NutrientsRecycle

BiomassTransport

NutrientsEnergy

EdibleBiomassfor Feedand Food

Non-EdibleBiomass

Recycle

RecycleCO2

H2O

CO2 H2O

Nutrients



15

0.6 BIOFUELS FROM CATA LY T I CP ROCESSING OF BIOMASS

Biomass can be converted into different types ofhydrocarbon fuels, such as “green” gasoline, diesel,and jet fuel. Via the conversion technologies

(3) fuel utilization. (Figure 0.3) [Huber, et al. 2006].Although the other areas are clearly important, thisreport focuses exclusively on technologies for theconversion of biomass into fuels.

A biofuel is a liquid transportation fuel made frombiomass. Feedstocks suitable for conversion intobiofuels include starches such as corn, sugar, animalor vegetable oils, lignocellulosic materials such astrees, grasses or corn stover, waste paper, etc.Unlike fuels derived from fossil sources such as crudeoil, biofuels are renewable and have a smaller carbonfootprint. A wide variety of biofuels are possible andinclude single chemical fuels or additives, as well astraditional, complex mixtures of chemicals:

Single molecule fuels or add i t i ve s :ETHANOL or ethyl alcohol (C2H5OH) can be madefrom cellulosic biomass via fermentation routes.Polysaccharides are depolymerized to yieldmonomeric sugars, which are then enzymaticallyfermented into ethanol. In addition, newtechnologies are emerging for the synthesis ofethanol and other higher alcohols from biomass-derived syngas via non-biological catalysis.

BUTANOL or butyl alcohol is a four-carbonalcohol. This biofuel can be made from cellulosicbiomass via fermentation routes or synthesized from syngas.

HYDROXYMETHYLFURFURAL (HMF) orFurfural is derived from biomass and does not haveto be biocatalytically processed via fermentation tomake fuels. Sugars can also be dehydrated viachemical catalysis to yield HMF (from 6-carbonsugars like glucose) and furfural (from 5-carbonsugars like xylose). These molecules are buildingblocks for transformation into potentially viabletransportation fuels such as ethyl levulinate (ELV),dimethylfuran (DMF), and -valerolactone (GVL).Routes to prepare DIMETHYFURAN (DMF), a 6-

THE WIDE RANGE OF BIOFUELS FROM LIGNOCELLULOSIC BIOMASS

carbon cyclic ether, from sugars in high yields haverecently been reported. DMF is a molecule that hasmany appealing properties for potential use as atransportation fuel [Roman-Leshkov et al. 2007]

-VALEROLACTONE (GVL) is similar to DMF andcan be synthesized from decomposition products ofsugars. Processes exist for the production oflevulinic acid in high yield from monomeric sugars,as well as for the catalytic transformation oflevulinic acid into GVL [Manzer 2005]. GVL hasrecently been suggested to have numerous propertiesthat make it suitable for use as a transportation fuel[Horvath, 2008].

ETHYL LEVULINATE (ELV) is made by thereaction of ethanol and levulinic acid to make anester. EVL has been suggested as a fuel additive bynumerous parties [Manzer 2005].

M i x t u re of compounds – classical fuels:“GREEN” GASOLINE OR DIESEL can besynthesized from lignocellulosic biomass by catalyticdeoxygenation. Technologies such as biomassreforming can be used to provide hydrogen forreduction of components of cellulosic biomass, suchas sugars, as well as lignin fractions of biomass, intogasoline or diesel range hydrocarbons. In somecases, the small biomass fragments need to becoupled to allow for proper molecular weights[Huber, et al. 2005]. Green diesel can also beprepared via the catalytic deoxygenation of fattyacids derived from virgin or waste vegetable oranimal oils. These same oils can be transformed into biodiesel by a transesterification reaction with methanol.


Biomass FeedstocksCellulosic Biomass(wood, wood wastes, corn stover, switch grass, agricultural waste, straw, etc.)Chemical Structure:cellulose, hemicellulose, lignin

Gasification

Fast Pyrolysis

Liquefaction

Pretreatment &Hydrolysis

Syn-gasCO + H2

Bio-oils(Sugars, Acids, Aldehydes, Aromatics)

Lignin(coumaryl, coniferyl and sinapyl alcohols)

C5 Sugars(Xylose)

C6 Sugars(Glucose, Fructose)

CornHydrolysis

Sugarcane

Triglycerides(VegetableOils, Algae)

Water-gas shift

MeOH Synthesis

Fischer-Tropsch Synthesis

Hydrogen

Methanol

Alkanes (Diesel Fuel)

Hydrodeoxygenation

Zeolite Upgrading

Emulsions

Gasoline

Olefins

Steam-Reforming

Hydrodeoxygenation

Zeolite upgrading

FermentationEthanol, Butanol

Corn Stover

Sucrose (90%)Glucose (10 %)

CornGrain

Bagasse

Aromatics, hydrocarbons (Gasoline)

Aromatics, light alkanes, (Gasoline)

Direct Use (Blend with Diesel)

All Sugars

FurfuralDehydration

MTHF (Methyltetrahydrofuran)

C6 Sugars

Alkyl benzenes, paraffins (Gasoline)

Aromatics, coke (Gasoline)

Transesterification

Zeolite/Pyrolysis

Aqu. Phase Proc.C8-C13 n-Alkanes, Alcohols

Hydrogenation

LevulinicAcid MTHF

EsterificationLevulinic Esters

Hydrogenation

Dehydration

APD/HC1-C6 n-Alkanes

ZeoliteAromatics, alkanes, coke

Aqueous or S.C. Reforming Hydrogen

Hydrodeoxygenation

Alkyl esters (Bio-diesel)

C1-C14 Alkanes/Alkenes

C12-C18 n-Alkanes

Blending/Direct UseDirect Use

Key: Black - Chemical ConversionGreen - Biological ConversionBlue - Both Chemical & Biological Conv.

F i g u re 0.4: Routes to Make a Biofuel, adapted from Huber et al. [ 2 0 0 6 ] .

described in this report, hydrocarbon fuels derivedfrom biomass are virtually indistinguishable frompetroleum-based hydrocarbons with respect totheir energy density. Additional advantages ofbiomass-derived hydrocarbons include theircapacity to self separate, which removes the greatexpense of distillation, and their compatibility withthe current fuel utilization infrastructure – no needfor engine modifications or new distributionsystems.

Catalytic processing is the technique by whichliquid fuels and chemicals are made frompetroleum-based feedstocks. Catalytic processingcan also be applied to the production of liquidhydrocarbon fuels derived from lignocellulosicbiomass. However, just as it is impossible toconvert all the energy in crude oil into gasolineand diesel fuels, it is equally impossible to convertall the energy of biomass into a fuel. Different

conversion technology methods have a wide rangeof efficiencies. Further advances in conversiontechnologies and process integration will ultimatelyimprove overall energy and economic efficiency of biofuels.

Liquid biofuels can be produced through a widerange of processes (Figure 0.4).The two main typesof catalysts used in these processes are eitherbiological or chemical (Table 0.2). Biologicalcatalysts, such as the yeast used to produceethanol, are homogeneous catalysts, meaning theyare in the same liquid phase as the biomass feed.Chemical catalysts range from homogeneous acidsto solid heterogeneous catalysts. As shown inFigure 0.4, the majority of the pathways to biofuelsproduction use chemical catalysts. While mankindhas been using biological catalysts in fermentation(i.e. ethanol production) for thousands of years,heterogeneous catalysts have only recently been



17

Ta ble 0.2: Comparison of Biological and Chemical Catalysts for Making Fuels from Lignocellulosic Biomass

Biological Catalysts Chemical Catalysts

Products Alcohols A Wide Range of Hydrocarbon Fuels

Reaction Conditions Less than 70oC, 1 atm 100-1200oC, 1-250 atm

Residence Time 2-5 days 0.01 second to 1 hour

Selectivity Can be tuned to be very Depends on reaction. New catalysts

selective (greater than 95 %) need to be developed that are greater

than 95 % selective.

Catalyst Cost $0.50/gallon ethanol $0.01/gallon gasoline

(cost for cellulase enzymes, (cost in mature petroleum industry)

and they require sugars to grow)

$0.04/gallon of corn ethanol

Sterilization Sterilize all Feeds No sterilization needed

(enzymes are being developed

that do not require sterilization

of feed)

Recyclability Not possible Yes with Solid Catalysts

Size of Cellulosic Plant 2000-5000 tons/day 100-2000 tons/day

applied to many of these biofuels productionpathways.

Chemical catalysts differ from biological catalysts ina number of respects (Table 0.2). Chemicalcatalysts can operate at significantly highertemperatures and over a broader set of conditionsthan biological catalysts. Thus, the residence timefor a reaction using biological catalysts is measuredin days compared with seconds or minutes forchemical catalysts. Biological catalysts are veryselective for certain classes of reactions such ashydrolysis and fermentation. Chemical catalysts canalso be selective for certain classes of reactions,and new classes of chemical catalysts will bedeveloped expressly for use with specific biomass-derived feedstocks. Biological catalysts are alsomore expensive than most chemical catalysts. Forexample Department of Energy projectionsindicate that the cost of cellulase enzymes forethanol production is between $0.30-.50 per gallon

of ethanol [EERE 2007]. In contrast, the cost ofchemical catalysts in the petroleum industry rangesaround $0.01 per gallon of gasoline. The majorityof biological catalyst-based processes requirefeedstocks to be sterilized prior to enzymaticconversion. No sterilization step is required forchemical conversion. Solid chemical catalysts canbe recycled, lasting for weeks and even years.In contrast, it is difficult to recycle biologicalcatalysts because they cannot easily be separatedfrom aqueous media once fuel is produced.Furthermore, chemical catalysts present anopportunity for small-scale distributedbiorefineries, which may not be possible forprocesses that utilize biological catalysts exclusivelybecause of the need to scale up the process inorder to make it economical. While most researchin biofuels to date has focused on development ofbiological catalysts it should be emphasized thatfuture biorefineries will likely use a combination ofbiological and chemical catalysts to make biofuels.


If you ask most drivers what they want when itcomes to fueling their cars or trucks you will hearthat they want low cost, good mileage, and wideavailability. Some, more thoughtful, drivers will addlow impact on the environment and improvedengine maintenance. While this sounds easy enough,the particulars of providing a fuel that can meetthese requirements are quite intricate and involvedetailed specifications derived from environmentalregulations and engine requirements.

New biomass-derived fuels that can meet theserequirements while minimizing the impact oncurrent vehicle engines, fueling infrastructure, andthe environment, will have a significant advantageover fuels that require new engines, new tanks, ornew types of fueling stations. For example, ethanoland bio-diesel fit into the current transportationinfrastructure, which explains why these fuels havehad greater commercial success than previousattempts at introducing less compatible fuels, suchas compressed natural gas or liquid propane gas.

Engine, fuel, and environmental technologies changewith time. Fuels that were excellent for engines ofthe past would destroy catalytic converters, polluteour air, and give poor driving performance in today’sengines. While it is impossible to know the exactdetails of future fuel specifications since bothengines and environmental regulations will continueto evolve, a number of key fuel properties arecertain to be important to future fuels.

I M P O RTANT GASOLINE PRO P E RT I E SSome of the most important properties of a “green”gasoline, made entirely of chemicals derived fromprocessed biomass or blended with petroleumproducts, are: cost; fuel economy; volatility; watertolerance; material compatibility; and environmentalimpact.

Fuel EconomyFuel economy is commonly calculated as the numberof miles traveled on a gallon of gasoline — miles pergallon (mpg). Some drivers also think of it in termsof fill-up frequency. Although Brazil’s experience withsubstituting fuel ethanol for petroleum-derivedgasoline has shown that consumers can cope withmore frequent fill-ups when fuels have reducedenergy content. However, a new fuel componentwith high volumetric energy density would be moredesirable.

Vo l at i l i t yThe rate at which a liquid vaporizes into the air isits volatility. Fuels only burn once they havevaporized into the air. When a liquid appears to beburning, actually it is the invisible vapor above the surface that is burning.

It is important to note that there is no single bestvolatility for gasoline. In cold weather, gasoline isblended to vaporize easily. This allows an engine tostart quickly and run smoothly until it is warm. Inwarm weather, gasoline is blended to vaporize lesseasily in the vehicle to prevent vapor lock or otherhot-fuel handling problems and minimizeevaporation, which contributes to air pollution.

Three properties are used to measure gasolinevolatility in the U.S.: vapor pressure, distillationprofile, and vapor-liquid ratio. A fourth property,driveability index, is calculated from the distillationprofile. Instead of a vapor-liquid ratio, a Vapor LockIndex (VLI) is used outside the U.S. to control hot-fuel-handling problems.

O c t a n eFuels with poor octane rating can damage engineswhen the fuel ignites prematurely. The measure of afuel’s propensity to pre-ignite is called the octanenumber. The minimum acceptable octane number fora fuel varies with engine design. Work is being doneto redefine the minimum octane requirements for

WHAT MAKES A GOOD FUEL?



19

modern vehicles based on a measure of accelerationperformance.

H e ating Va l u eHeating value is usually thought of as the amount of energy in a gallon of fuel. Conventional fuels varyin heating value. One cause is the formulationdifferences among batches and among refiners. A1990–1991 survey of conventional gasolines foundthat the heating value of summer gasolines variedover an 8 percent range. The heating value alsovaries by grade and by season. On average, theheating value of premium-grade gasoline is about 0.7percent higher than regular-grade because premium-grade, in general, contains more aromatichydrocarbons — the class of hydrocarbons with thehighest densities. The heating value of wintergasoline is about 0.5 percent lower than summergasoline because winter gasoline contains morevolatile, less dense hydrocarbons. Since ethanol hasless energy per gallon than gasoline, the heatingvalue of ethanol/gasoline blends decreases as theamount of ethanol increases.

Water Tolerance Fuels always come into contact with water. New fuelblending components that do not absorb water orsplit into separate phases when mixed with waterare highly desirable. Conventional gasoline candissolve up to 150 parts per million (ppm) water atnormal conditions. Oxygenating gasoline by addingethers to it can increase water solubility to 600ppm. Bringin either conventional gasoline or ether-oxygenated gasoline into contact with additionalwater will not affect the properties of the gasolinebut can make it hazy in appearance. In contrast,ethanol/gasoline blends can phase separate whenthey come into contact with too much water.

M aterial Compat i b i l i t yNew biomass-derived fuel blending components mustnot corrode fuel system metal components ordissolve polymers used in fuel tanks and lines.Oxygenates can swell and soften natural and

synthetic rubbers (elastomers). Oxygenated gasolinesaffect elastomers less, the extent of which alsodepends on the hydrocarbon chemistry of thegasoline, particularly the aromatics content. Theeffect is of potential concern because fuel systemscontain elastomers in hoses, connectors ("O" rings),valves and diaphragms. The elastomeric materialsused in today's vehicles in the U.S. have beenselected to be compatible with oxygenated gasolines.Owner's manuals approve the use of gasolineoxygenated with 10 vol % ethanol or 15 vol % MTBE(the compatibility of the other ethers is the same asthat of MTBE). New gasoline blending componentsthat do not require additional automobile upgradeswould be highly desirable.

E nv i ronmental and Safety ImpactNew fuel components must be safe for the peoplethat use them and for the environment. Leadadditives for gasoline were banned decades ago inthe U.S. Limits on the amounts of benzene andaromatics in gasoline have been in effect for severalyears. Recently safety concerns over methyl t-butylether (MTBE) caused the use of this material (as wellas higher molecular weight ethers and alcohols) as afuel additive to be banned in some states. No fuelproducer will be willing to consider adding a newcomponent to its gasoline without extensivescientific and regulatory review.

Biomass feedstocks have lower levels of sulfur-containing compounds than either crude oil or coal,resulting in lower levels of SOx emissions. Sulfurpoisons catalytic converters and on-board diagnosticequipment in vehicles, leading to increased levels ofpollutants, which adversely affect human health andthe environment. In 2005, the U.S. governmentimposed a maximum limit on sulfur in gasoline of 30ppm; the European Union will require even stricterlimits of 10 ppm by 2009. Biomass-derived fuels willbe required to meet these sulfur limits as well.These lower sulfur concentrations also eliminate theneed for a costly sulfur removal step prior tocombustion.


continued next page. . .


I M P O RTANT DIESEL FUEL PRO P E RT I E SThe properties of fuel economy, low engine wear(lubricity), low-temperature operability, long filterlife (stability), ease of startup, and low emissions areimportant to diesel fuels. Their relative importancedepends on engine type and duty cycle (truck,passenger car, stationary generator, marine vessel,etc.). “Green” diesel, comprised at least in partfrom blend stocks produced from biomass, can beformulated to have a number of these desirableproperties.

Cetane Rat i n gWhen a cold diesel engine is started, the heat ofcompression is the only energy source available toheat gas in the combustion chamber to atemperature that will initiate spontaneous fuelcombustion (about 750°F [400°C]). A fuel thatcombusts easily will require less cranking to start anengine. The measure of combustibility is referred toas the cetane number. Thus, all other conditionsbeing equal, a higher cetane number fuel makesstarting easier. Modern diesel engines require aminimum cetane number of 40 for adequate coldstarting unless temperatures are very low. Futureengine design advances may reduce this requirement.

Fuel EconomyAs for gasoline, fuel economy is related to heatingvalue. The heating value (also referred to as energycontent) of diesel fuel is its heat of combustion (i.e.the heat released when a known quantity of fuel isburned under specific conditions). In the U.S., theheating value is usually expressed as British thermalunits (Btu) per pound or per gallon at 60°F.(International metric [SI] units are kilojoules perkilogram or per cubic meter at 15°C.)

Lubricity Some moving parts of diesel fuel pumps andinjectors are lubricated by the fuel. To avoidexcessive wear, the fuel must have some minimumamount of lubricity. (Lubricity is the ability toreduce friction between solid surfaces in relativemotion.) Many diesel fuel components are goodlubricants. This is not due to the hydrocarbons thatconstitute the bulk of the fuel. Instead it isattributed to trace amounts of oxygen- andnitrogen-containing compounds and certain classesof aromatic compounds. Evidence for the role oftrace quantities is the fact that the lubricity of afuel can be restored with the addition of as little as 10 ppm of an additive.


0.7 DESIGNING BIOFUELS P RODUCTION PRO C E S S E S

A comprehensive understanding of the fundamentalchemistry, science and engineering underpinningthe chemical catalytic production of lignocellulosicbiofuels is necessary to build on the many advancesthat have already been made in the development ofbiofuels production processes. Empirical “know-how” must be translated into sound theory inorder to advance the science and engineering ofthe biorefinery.

Process optimization is critical to the success ofscaled-up biofuels production operations. Thisrequires the combination of process modelsderived from known feedstock and reactantproperties and application of property predictionmethods in tandem with laboratory-basedresearch. Through careful process design andengineering using reliable data, 80 percent ofbiomass-to-fuel production facility can be locked inat the end of the process design phase.

An economic process model is also critical to thedevelopment of cost-effective, large scale biofuelsproduction processes. Process economic



21

Acidity Organic acids in diesel fuel can also cause corrosivewear of the fuel system. While this may be asignificant wear mechanism for high sulfur diesel, itis less significant for low sulfur diesel becausehydrotreating to reduce sulfur also destroys organicacids.

L ow Te m p e r at u re OperabilityLow temperature operability is an issue with middledistillate fuels because they contain straight andbranched chain hydrocarbons (paraffin waxes) thatbecome solid at ambient wintertime temperatures incolder geographic areas. When this happens, the waxmay plug the fuel filter or it may completely gel thefuel, making it impossible for the fuel system todeliver fuel to the engine. This has also been aconcern with the current generation of bio-dieselscomprised of fatty acid methyl esters. The ability toremain liquid at low temperature is a highlydesirable characteristic of diesel blend stocks.

E m i s s i o n sRegulated emissions are hydrocarbons (HC), carbonmonoxide (CO), nitrogen oxides (NOx), andparticulate matter (PM). Unlike gasoline poweredengines, diesel engines don't have much trouble

meeting the HC and CO standards. The standards forNOx and PM are much more challenging and areaffected by fuel components. Gums in the fuel aredeposited on the injectors, causing sticking, whichinterferes with fuel metering.

Petroleum residue or inorganic salts in the fuelresult in injector tip deposits that prevent theinjector from creating the desired fuel spray pattern.Excessive abrasive contaminants, organic acids in thefuel, or inadequate fuel lubricity cause abrasive orcorrosive injector wear.

I M P O RTANT JET FUEL PRO P E RT I E SJet fuels are comprised of materials that distill inthe temperature range between gasoline and diesel.The specifications are more wide acceptedinternationally since airplanes cross internationalborders more frequently than cars and trucks. Theimportant properties of jet fuels are energy content,lubricity, combustion characteristics, stability,distillation range, corrosivity, density, cleanliness,fluidity, and electrical conductivity.

The ASTM specifications for several jet fuels areshown in Table S-0.1 on page 22.


bottlenecks include the needs to: reduce hydrogenrequirements; demonstrate catalyst stability; andincrease product value. Experiments in chemicalengineering and theoretical chemistry should becarried out in tandem with process engineers inorder to develop methodologies and processesthat are both thermodynamically efficient andeconomical.

However, a lack of fundamental information aboutthe chemical properties of various biomassfeedstocks hinders the development of essentialproduction process models. Detailed models ofthe chemical and thermodynamic properties of

these biostreams would enable the development ofadvanced bioprocessing systems to create fuels andvaluable organic chemicals. The lack of basicunderstanding about the reaction system limits theability to draw general relationships betweenbiomass composition, pyrolysis conditions,condensation conditions, and bio-oil composition.The creation of a feedstock and reactantproperties database, which will enable thedevelopment of properties prediction methods, isessential to the design of cost-effective biofuelsproduction processes.


Property Jet A or ASTMJet A-1 Jet B Test Method

COMPOSITIONAcidity, total, mg KOH/g, max 0.10 – D 3242 Aromatics, % vol, max 25 25 D 1319 Sulfur, mercaptan, % mass, max 0.003 0.003 D 3227 Sulfur, total, % mass, max 0.30 0.3 D 1266, D 1552, D 2622

D 4294, or D 5453

VOLATILITYDistillation, ºC D 86

Volume percent recovered10, max 205 –20, max – 14550, max Report 19090, max Report 245

Final boiling point, max 300 –Distillation yields, % vol:

Residue, max 1.5 1.5Loss, max 1.5 1.5

Flash point, ºC, min 38 – D 56 or D 3828 Density, 15ºC, kg/m3 775 to 840 751 to 802 D 1298 or D 4052 Vapor pressure at 38ºC, kPa, max – 21 D 323 or D 5191

FLUIDITYFreezing point, ºC, max – 40 (Jet A) – 50 D 2386, D 4305, D 5501, or

– 47 (Jet A-1) D 5972 Viscosity at –20ºC, mm2/sec max 8.0 – D 445

COMBUSTIONNet heat of combustion, MJ/kg, min 42.8 42.8 D 4529, D 3338, or D 4809 One of the following requirements:

1. Luminometer number, min 45 45 D 1740 2. Smoke point, mm, min 25 25 D 1322 3. Smoke point, mm, min and 18 18 D 1322 naphthalenes, % vol, max 3.0 3.0 D 1840

CORROSIONCopper strip, 2 hr. at 100ºC, max No. 1 No. 1 D 130

STABILITYThermal stability, 2.5 hr. at 260ºC:

Filter pressure drop, mm Hg, max 25 25 D 3241 Tube deposit, less than Code 3 Code 3

CONTAMINANTSExistent gum, mg/100 mL, max 7 7 D 381 Water reaction, interface rating, max 1b 1b D 1094

* See the current version of D 1655 for complete requirements.

Table S-0.1 Summary of ASTM D 1655 Table 1 Requirements*




23

0.8 NEXT GENERAT I O NBIOREFINERIES FORP RODUCTION OF LIQUID FUELS

A biorefinery is a facility that integrates biomassconversion processes and equipment to producefuels, power, and chemicals from biomass.Thebiorefinery concept is analogous to today'spetroleum refineries, which produce multiple fuelsand products from petroleum. Industrialbiorefineries have been identified as the mostpromising route to the creation of a new domesticbio-based industry. By producing multipleproducts, a biorefinery can take advantage of the

differences in biomass components andintermediates and maximize the value derivedfrom the biomass feedstock.

A biorefinery might, for example, produce one orseveral low-volume, but high-value, chemicalproducts and a low-value, but high-volume liquidtransportation fuel, while generating electricity andprocess heat for its own use, and perhaps enoughfor sale of electricity.The high-value productsenhance profitability, the high-volume fuel helpsmeet national energy needs, and the powerproduction reduces costs and avoids greenhouse-gas emissions [NREL 2007].

T h e re are many different types of biofuels to choose fro m , including fuels currently in

use and those that are still on the drawing board . In terms of “picking a winner”, we

s t rongly advo c ate that science should determine the best possible solution. W h e t h e r

bio-based fuels are best fit for gasoline (spark ignited engines), diesel (compre s s i o n

ignited engines) or jet fuel (turbines) is not critical. While we are not advo c ating that

the end product application be completely ignore d , re s e a rch should be encouraged to

f a c i l i t ate the best possible opportunity for a fuel technology bre a k t h rough that could

fit into any of these end-product are a s .

As the gove rn m e n t a l , a c a d e m i c, and industrial communities debate technology options to deve l o p

p re f e r red bio-based fuel pat h way s , a frequently raised question is, “ W h at fuel or fuel pro p e rties do we

want to design for?” While this is a key long term question, the immediate issues that deserve

p r i o r i t i zed attention include: reducing the oxygen content (lower oxygen content increases energ y

density in the context of liquid-phase catalytic processing of biomass-derived stre a m s ) , ensuring that

the molecular weight is within the range of fuels compat i ble with internal combustion engines, and that

the end product is designed to optimize compatibility with conventional hy d rocarbon fuels.

In the global fuels marke t , a dvances in biofuel technology would be welcomed in any of these are a s .

F u rt h e r m o re , the use of conventional hy d rocarbon processing facilities is an option that would ex p l o i t

the availability of existing fuel processing and distribution infrastru c t u re for the production of biofuels.

To summarize , maximum liberty to re s e a rch and develop a desirabl e , c o s t - e f f e c t i ve bio-based fuel

p roduct should be encouraged and we should not limit ourselves to a “designer fuel” s p e c i f i c at i o n

mentality that hinders cre ativity or opport u n i t y.

PICKING A WINNER!


RAPID BIOFUELS PRODUCTION USING HIGH-EFFICIENCY,SMALL-SCALE REACTORS

The heart of any biorefinery is the reactor (or

reactors) where the biomass is converted into a fuel

or fuel precursor. The residence time is the time a

reactant or product is in the chemical reactor.

Ideally the residence time should be as low as

possible since this leads to a smaller reactor and

lower capital and operating costs. Chemical

reactions exhibit a clear, well-understood relationship

between temperature and residence time as shown

in Figure S-0.1. This figure shows a range of

industrial processes, which use both chemical and

biological catalysts. Chemical catalysts can be used

over a wide range of temperature from 50-1100°C

(black dots in Figure S-0.1). In contrast, biological

catalysts are only stable at a very limited

temperature range (20-50°C), which is required to

preserve organic enzymes and microorganisms

[Schmidt and Dauenhauer 2007]. Figure S-0.1 also

contains two chemical routes to make biofuels. A

technique called “biomass fast pyrolysis” (blue dots

in Figure S-0.1) rapidly heats wood chips to about

500°C to produce a brown liquid called “bio-oil” in

about one second. This transformation of biomass to

fuel occurs four to five orders of magnitude faster

than biological routes, permitting reactor systems

that are several thousand times smaller.

Figure S-0.1: The residence time of a reaction versus the reaction temperature. Processesthat operate at higher temperature require smaller reactor residence times. By selecting ahigher temperature process, or adding catalyst, the residence time of chemicals in a reactor

can decrease by orders of magnitude.

(continued next page)



25

The biorefinery of the future will be analogous tothe petroleum refinery of the present: a highlyintegrated system of processes and chemicalmechanisms at a fundamental level. Detailedprocess analysis made the unprecedentedefficiencies of the petrochemical industry possible.Similarly, there is a great need for extensive datacollection and the development of computermodels representing the reactants, products andintermediates involved in the conversion ofbiomass feedstocks into valuable fuels.

0.9 OV E RVIEW OF ROADMAP AND WO R K S H O P

This workshop builds on the success of fourprevious programs:

Catalysis for Biorenewables Conversion(2004; Chair: Dennis Miller) http://www.egr.msu.edu/apps/nsfworkshop/

Design of Catalyst Systems forBiorenewables (report)(2005; Chair: Brent Shanks)http://www.nsf.gov/eng/cbet/workshops/catalysis_for_biofuels_workshop_2005.pdf

Breaking the Biological Barriers toCellulosic Ethanol (2005; Chair: Sharlene Weatherwax) http://genomicsgtl.energy.gov/biofuels/b2bworkshop.shtml

Thermochemical Conversion ofBiomass(2007; Chair: Paul Grabowski, David Dayton)http://www.thermochem.biomass.govtools.us/

A wide range of experts from industry,governmental agencies and academic researchinstitutions were invited to participate in theworkshop:

- 71 invited participants- 27 academics from 24 universities- 19 companies represented- 13 representatives from 5 national laboratores- 10 program managers (NSF, DOE, USDA)

The workshop was divided into six thrust areas(below). The first four thrusts cover differenttechnical strategies to convert lignocellulosicbiomass into biofuels. The last two strategiesaddress important cross-cutting areas which affectall of biomass conversion. Figure 0.5 shows a

Chemical routes to biofuels occur even faster by

the addition of chemical catalysts combined with

higher operating temperatures. It has recently

been shown that catalysts can be added to

biofuel reaction at high temperature in a process

called “catalytic partial oxidation of biomass”

(red in Figure S-0.1) permitting particles of

biomass to be converted to a liquid fuel

precursor called “synthesis gas” as fast as 0.05

seconds residence time [Salge, et al. 2006]. The

catalyst also reduces the amount of waste

byproduct significantly improving the overall

efficiency of the biomass-to-liquids process. As

shown in Figure 0.5, white particles of biomass

directly impact the bright orange metal catalyst

at about 800 °C where they rapidly convert to a

gaseous mixture [Dauenhauer, et al. 2006]. The

gaseous mixture flows into the catalyst where it

undergoes reactions producing predominately

liquid fuel precursors. The reaction sustains the

necessary high temperatures without a heat

source permitting the reaction to operate

independently of extra equipment. Small,

independent chemical reactions of this nature

have the potential to permit production of

liquid fuels from biomass at locations as small as

individual farms and homes.


BiomassConversion

BiomassConversion

THRUST 4:SYN-GAS

CONVERSIONAND TAR REMOVAL

THRUST 2:UTILIZATION OF

PETROLEUMREFINING

THRUST 3LIQUID-PHASE

CATALYTICPROCESSING

LignocellulosicBiomassFeedstocks(DOE-GTLWorkshop)

Biofuels

GreenGasoline

GreenDiesel

Green JetFuel

GreenChemicals

Syn-gas& Tars

Carbohydrates

Lignin

Bio-oils

Gasification(DOE Thermochemical

Workshop)

THRUST 1:SELECTIVETHERMAL

PROCESSING

Pretreatment w/Acid/Enzymatic

Hydrolysis(DOE-GTL Workshop)

F i g u re 0.5: O ve rv i ew of Wo r k s h o p

The Key Findings of the workshop are outlined inthe following sections. We have identified the basicresearch needs and opportunities in catalyticchemistry and materials science that underpinbiomass conversion and fuel utilization, with a focuson new, emerging and scientifically challenging areasthat have the potential for significant impact.Thereport illuminates the principal technologicalbarriers and the underlying scientific limitationsassociated with efficient processing of biomassresources into finished fuels. Major developmentsin chemistry, biochemistry, materials and otherdisciplines associated with biomass-to-fuelsprocessing are underscored and strategies toovercome the long-term grand challenges areoutlined.

2Biomass gasification was not discussed because it was the primary focusof a previous DOE workshop held in January of this year.(www.thermochem.biomass.govtools.us/).

general scheme that illustrates the overallapproaches to the conversion of lignocellulosicbiomass to fuels. This scheme also provided theorganization structure of the workshop.

1. Selective Thermal Processing of Lignocellulosic Biomass

2. Utilization of Petroleum Refining Technologies for Biofuel Production

3. Liquid-phase Catalytic Processing of Sugars and Bio-oils

4. Catalytic Conversion of Syn-gas2

5. Process Engineering and Design

6. Cross Cutting 21st CenturyScience,Technology, and Infrastructure for a NewGeneration of Biofuel Research



27

REFERENCES CITED:

Climate Change 2007: The Physical Science Basis.Contribution of Working Group I to the FourthAssessment Report of the Intergovernmental Panelon Climate Change. Summary for Policy Makers.Intergovernmental Panel on Climate Change. 2007.Accessed on 20 September 2007:http://www.ipcc.ch/SPM2feb07.pdf

Council on Competitiveness. Future of U.S. Energyand Competitiveness Tightly Linked.1 August 2007.Washington D.C.Accessed on 17September 2007:http://www.compete.org/newsroom/temp_news/readnews_08-01-07_energy_initiative.asp

Dauenhauer, P.J., Dreyer, B.J., Degenstein, N.J.,Schmidt, L.D. “Millisecond reforming of solidbiomass for sustainable fuels,” Angewandte Chemie46 (2006).

EERE. Cellulase Enzyme Research. U.S.Department of Energy, Division of Energy Efficiencyand Renewable Energy, Biomass Program.Accessed 24 September 2007:http://www1.eere.energy.gov/biomass/cellulase_enzyme.html

Energy Information Administration. 2007. AnnualEnergy Outlook 2007 with Projections to 2030.Accessed 15 September 2007:http://www.eia.doe.gov/oiaf/aeo/index.html

Energy Information Administration. 2007b.WorldCrude Oil Prices 1978-2007.Accessed 15 September 2007:http://tonto.eia.doe.gov/dnav/pet/pet_pri_wco_k_w.htm

Energy Information Administration 2007c. U.S.Imports by Country of Origin.Accessed 15 September 2007.http://tonto.eia.doe.gov/dnav/pet/pet_move_impcus_a2_nus_ep00_im0_mbbl_m.htm

Greene, N., et al. “Growing Energy: How BiofuelsCan Help End America’s Oil Dependence.” NaturalResources Defense Council, New York. 2004.Accessed 17 September 2007:http://www.nrdc.org/air/energy/biofuels/biofuels.pdf

Horváth, I.T., Mehdi, H., Fábos,V., Boda, L., Mika,L.T. Green Chemistry, 2008

Huber, G.W., J. N. Chheda, et al. (2005).“Production of Liquid alkanes by Aqueous-PhaseProcessing of Biomass-Derived Carbohydrates.”Science 308: 1446-1450.

Huber, G.W., et al. “Synthesis of TransportationFuels from Biomass: Chemistry, Catalysis andEngineering.” Chemical Reviews. 2006. 106:4044-4098.

International Energy Agency. Highlights ofSeptember Oil Market Report. 2007. Accessed 20September 2007: http://omrpublic.iea.org/

Klass, D. L. In Encyclopedia of Energy; Cleveland, C.J., Ed.; Elsevier: London, 2004;Vol. 1.

Manzer L.E. , Preparation of Levulinic Acid Estersfrom Alpha-Angelica Lactone and AlcoholsAssigned to DuPont US Patent Application2005/0210738 A1 2005.


28

REFERENCES CITED:

NREL. What is a biorefinery? National RenewableEnergy Laboratory, Biomass Program. Accessed on20 September 2007:http://www.nrel.gov/biomass/biorefinery.html

Perlack, R.D., et al. Biomass as Feedstock for aBioenergy and Bioproducts Industry: The TechnicalFeasibility of a Billion-Ton Annual Supply.DOE/GO-102005-2135, Oak Ridge NationalLaboratory, Oak Ridge,Tennessee. 2005.Accessed 12 September 2007:http://feedstockreview.ornl.gov/pdf/billion_ton_vision.pdf.

Renewable Fuels Association. Historic U.S. EthanolProduction. Accessed 15 September 2007:http://www.ethanolrfa.org/industry/statistics/

Roman-Leshkov,Y., Barrett, C. J., Liu, Z.Y., Dumesic,J.A. (2007). Production of dimethylfuran for liquidfuels from biomass-derived carbohydrates, 447,Nature 447 982 - 985

Saad, L. 2007. Most Americans Back Curbs onAuto Emissions, Other Environmental Proposals.Gallup News Service. 5 April 2007.Accessed 15 September 2007:http://www.galluppoll.com/content/?ci=27100

Salge, J.R, Dreyer, B.J., P.J. Dauenhauer, “Renewablehydrogen from nonvolatile fuels by reactive flashvolatilization,” Science 314 (2006) 801-804.

Schmidt, L.D., P. J. Dauenhauer,“Hybrid routes tobiofuels,” Nature 447 (2007) 914-915.

US Department of Labor, "Chemical Manufacturing,Except Pharmaceutical and MedicineManufacturing", U.S. Department of Labor, 2005.

White House. President Bush Delivers State ofthe Union Address. 23 January 2007.Accessed: 17 September 2007:http://www.whitehouse.gov/news/releases/2007/01/20070123-2.html

Wyman, C. E.; Decker, S. R.; Himmel, M. E.; Brady, J.W.; Skopec, C. E.;Viikari, L. In Polysaccharides, 2nded.; Dumitriu, S., Ed.; Marcel Dekker: New York,2005.

An IntegratedChemical and

Engineering Platfo r mfor Ove rcoming

the Barrier toCellulosic Biofuels...


1 . S e l e c t i ve Thermal Processing of Cellulosic Biomass and Lignin

30

F i g u re 1.1 A schematic showing the chemical products resulting from the thermochemical conversion of biomass in the absence of any

c atalytic mat e r i a l s .

PrimaryProcesses

CO, CO2,H2O

VaporPhase

LiquidPhase

SolidPhase

SecondaryProcesses

Tertiary Processes

PrimaryVapors

Light HCs,Aromatics,

& Oxygenates

Olefins, AromaticsCO, H2, CO2, H2O

PNAs, CO, H2,CO2, H2O, CH4

CO, H2,CO2, H2O

Charcoal Coke SootBiomass

PrimaryLiquids

Low P

High P

Low P

High P

Pyrolysis Severity

Condensed Oils(phenols, aromatics)

OV E RV I E W:

The production of bio-oils, through thermal

processing of biomass, is a versatile and economical

first step in the fabrication of liquid transportation

fuels. Bio-oils are multi-component mixtures

comprised of different sized molecules derived

primarily from depolymerization and fragmentation

reactions of the three key biomass building blocks:

cellulose, hemicellulose, and lignin. Bio-oils

production technologies are relatively low cost and

suitable for small-scale applications. Several types

of thermal processes for producing bio-oils are in

the commercial pilot plant stage, demonstrating

that the economics of these processes are

competitive. However, current bio-oils production

technology is not very selective, resulting in a

bio-oil composed of more than 300 chemical

species. New techniques for increasing the

control of bio-oil composition are needed to make

this technology more attractive. Among the

technological advances needed are better

characterizations of the thermal reactions and an

examination of how catalysts may be incorporated

into the reaction environment to produce the

preferred bio-oil compositions.

1.1 INTRODUCTION

Lignocellulosic biomass is difficult to deconstructinto easily utilized sub-fractions because of itsheterogeneous composition. The approaches usedfor lignocellulose deconstruction can be broadlylumped into three classes: gasification, hydrolysis,and selective thermal processing. Of these threeapproaches, gasification creates a primary productwith the simplest chemical distribution, but thissynthesis gas (“syngas”) product must besubsequently reconstructed into fuels andchemicals, as discussed in Chapter 4. Hydrolyticprocesses, particularly if performed with


biocatalysts (as in fermentation), provide a mixedsugar and a lignin product with slightly higherchemical species complexity than with gasification.Of these three approaches, selective thermalprocessing results in a product with the highestlevel of chemical complexity. In spite of thechemical complexity of the resulting bio-oil, theadvantage of this approach is the simplicity of the process.

The complexity of the chemical products resultingfrom thermal fractionation is illustrated in Figure 1.1. Depending on the severity of thepyrolysis conditions the product slate can access arange of chemical processes; increased severity isdefined as higher temperature and/or longerreaction times. These chemical processes, whichare a function of pyrolysis severity, can be lumpedinto primary, secondary and tertiary reactions.Primary processes represent the first productsproduced when lignocellulosic biomass isthermochemically treated. However, these primaryproducts can go through subsequent conversion tosecondary products. Finally, as the process severityincreases, the secondary products can react again.Products from the tertiary processes representthose obtained via gasification. It is important tonote that the chemical products shown in Figure1.1 represent only the products produced bythermal reactions. Incorporating catalytic materialsinto this process provides an opportunity tofurther alter this chemical landscape.

The liquid bio-oil produced by thermalfractionation processes requires requiresdownstream catalytic upgrading, discussed inChapter 2 and 3. The current chapter and thefollowing chapters on upgrading are inherentlylinked because the composition of chemicalproducts obtained by a given thermal fractionationprocess will drive the type of processing needed toupgrade the resulting bio-oil. However, the inverseproblem of determining what thermal fractionation

products are most amenable to upgrading is alsotremendously important. Substantial processoptimization can only achieved by concerted work in both the areas of thermal fractionationand upgrading.

1.2 OVERALL PROCESS DESCRIPTION

1.2.1 Fast Pyro l y s i sBiomass pyrolysis is the thermal depolymerizationof biomass at modest temperatures in the absenceof added oxygen. The slate of products from

SECTIONS OF T H RUST 1:

1.1 Introduction

1.2 Overall Process Description

1.3 Resulting Fuels

1.4 Summary of Previous Research

1.5 Economics and Potential ofTechnology

1.6 Current Technology Limitationsand Research/Development Needs

1.7 Recommendations

1.8 References

PA RT I C I PA N T S :Phillip Badger, Paul Blommel,A.A. (Kwesi) Boateng, Robert Brown,Steven Czernik, Anthony Dean, MarkA. Dietenberger, Douglas Elliott, CalvinFeik, Kristi Fjare,Anne Gaffney, FrankGerry, George W. Huber, Mark Jones,Mike Malone, Leo Manzer, RichardMarinangeli, Raul Miranada,ValeriaReed, John Regalbuto, Bob Saxton.Lanny Schmidt, James F Stevens,Chunshan Song, Phillip Steele, GalenSuppes, Scot Turn, Phil Westmoreland,Charlie Wyman.


biomass pyrolysis depends on the processtemperature, pressure, and residence time of theliberated pyrolysis vapors (Bridgwater, 1999;Bridgwater and Peacocke, 2000; Czernik andBridgwater, 2004;Yaman, 2004). Production of thebio-oil products is maximized by fast pyrolysis,which is typically performed at temperatures of450 to 500ºC at atmospheric pressure (Czernikand Bridgwater, 2004). High heating rates (i.e.,500ºC/s) and short residence times (1-2 s)

A fundamental challenge to cost-effective biomass-conversion is the expense of transporting low energy density

biomass feedstocks. An advantage of fast pyrolysis is that this technology is economical for use on a small

scale (i.e. 50-100 tons-biomass/day) where fast pyrolysis systems are built on portable units and distributed

close to the biomass source. This can significantly reduce costs. For example the cost of biomass has been

reported to be $22/dry-ton for a 24 tons/day biofuel facility, which is half the cost of a 1000 ton/day facility

which would be $44/dry-ton. This contrasts with other biofuel technologies, such as cellulosic ethanol, where

bigger is usually better due to economies of scale.

Another advantage of the distributed technology is that it may create more jobs, compared with other

large-scale biofuel technologies. The distributed technology has lower capital cost investment, but a higher

operating cost than other biofuel technologies such as cellulosic ethanol. The higher operating cost is because

more operators are needed than with a larger scale processing facility. The chemical industry is one of the

highest paying industries for skilled labor. The U.S. Bureau of Labor Statistics (2005) has stated that

production workers in chemical manufacturing (which would be the same as those who work in a bio-refinery)

earn more money ($820 weekly) than in all other manufacturing industries ($659/week) and than in private

industry ($529/week). They further state that in 2004 the median hourly wages were above $19/hour for

plant operators, maintenance and repair workers, chemical technicians, and chemical equipment operators

and tenders.

Utilization of small mobile or modular pyrolysis reactors is a concept that is currently being developed to

produce liquid biofuels close to the biofuel location. Figure S-1.1 illustrates a fast pyrolysis reactor that is

built on a truck bed. One of the most successful tests of this concepts was the Canadian Waterloo Fast

Pyrolysis Process comprising a mobile trailer with a production capacity of 200 kg/hr that was tested

successfully to process several fuel types. Several small

companies, including Renewable Oil International® LLC,

are focusing on development of smaller mobile or

modular pyrolysis reactors.

1.2.1 Fast Pyrolysis continued

F i g u re S-1.1 A Small scale port a bl ereactor for liquid fuel production by

distributed fast py ro l y s i s .

DISTRIBUTED BIOFUELS PRODUCTION USING FAST PYRO LYS I S



33

produce substantial yields of bio-oils that retain upto 70% of the energy present in the biomassfeedstock prior to processing .

The products from the fast pyrolysis processinclude gases, bio-oil, and char. A representativedistribution of products from a fast pyrolysisreactor, operated to maximize the bio-oil liquid, is:65 wt% organics, 10 wt% water from the reaction,12 wt% char, and 13 wt% gas. However, therelative ratios of three of these fractions is highlydependent on reaction conditions, reactor design,and biomass alkali content.

Bio-oil consists of water soluble and insolublefractions. The water-soluble fraction, which is highin oxygen content, is derived from the cellulose

F i g u re 1.2: A diagram of the key process components for a re p re s e n t at i ve fast py rolysis process utilizing a fluidized bed reactor

(used with permissions from Bridgwater et al., 2 0 0 1 )

FLUID BEDREACTOR

CHARTo reactoror export

BIOMASS

Quenchcooler

Electrostaticprecipitator

BIO-OIL

Recyclegas heater

and/oroxidiser

Gasrecycle

GASFor export

and hemicellulose fractions of the biomass. Thewater insoluble fraction, which has lower oxygencontent, is derived primarily from the ligninfraction. The chemical species in the bio-oil exitthe pyrolysis reactor in either the vapor form, as afree radical precursor, or in an aerosol form.A schematic of a fast pyrolysis process is shownFigure 1.2. The schematic shows a fluidized bedreactor, but a number of different reactor designs,which can provide high biomass heating rates andshort product residence times, have been studied(Scott et al., 1999). A key feature in thesubsequent utilization of the bio-oil is effectiveseparation of the solid char so as to minimize itspresence in the liquid bio-oil.


1.2.2 LiquefactionBio-oils can also be produced by liquefaction, whichoccurs in the presence of a solvent at highpressure (120-200 atm) and relatively mildtemperatures (300-400 ºC). Addition of catalyststo the liquefaction process can improve the qualityof the bio-oil produced. The most commonly usedsolvent in liquefaction studies is water (Moffatt andOverend, 1985; Naber et al., 1997; Goudriaan andPeferoen, 1990). A biomass solids loading of about40% by weight can be used in the process. One

attractive feature of this process is that wetbiomass can be used directly. In contrast to thebio-oil from fast pyrolysis, the liquefaction producthas lower oxygen content. The difference is due toa significant amount of decarboxylation that occursduring the liquefaction process. The removal ofoxygen via decarboxylation leaves a more favorableC/H ratio in the bio-oil product and makes theproduct water insoluble.

Property Pyrolysis oil Liquefaction oil Heavy fuel oil

Moisture content, wt% 15-30 5.1 0.1pH 2.5Specific gravity 1.2 1.1 0.94Elemental composition, wt%

-carbon 54-58 73 85-hydrogen 5.5-7.0 8 11-oxygen 35-40 16 1.0-nitrogen 0-0.2 0.3-ash 0-0.2 0.1

Higher heating value, MJ/kg 16-19 34 40Viscosity (50°C), cP 40-100 15,000 180

(at 61°C)Solids, wt% 0.2-1 1Distillation residue, wt% Up to 50 1

Ta ble 1.1 Pro p e rties of bio-oils and heavy fuel oil ( c o m p i l ation from Cze rnik and Bridgwat e r, 2 0 0 4 ; Elliott and Schiefelbein, 1 9 8 9 ;

Huber et al., 2 0 0 6 ) .

F i g u re 1.3 A process diagram for an aqueous liquefaction pilot plant c o n s t ructed by Biofuel B. V. (adapted from Naber and Goudriann, 2 0 0 5 ) .

High PressurePump

BIOMASS10-20kg/hr (db)

Preheater/Reactor 1

Reactor 2 Cooler

Cooler

330oC180bar

CO2

Condensor

Gases

Storage Storage

1bar

PressureReducer

Biocrude/WaterCollection

Gas/LiquidSeparator



35

composition of bio-oil resembles that of biomassrather than that of petroleum oils.

Most of the phenolic compounds are present asoligomers having a molecular weight ranging from900 to 2500. The presence of oxygen in many oilcomponents is the primary reason for differencesin the properties and behavior seen betweenhydrocarbon fuels and biomass pyrolysis oils.Thehigh oxygen content results in a low energy density(heating value) of 50% of conventional fuel oils.Aneven more important consequence of the organicoxygen is the instability of bio-oil.

Liquid bio-oil can be transported and stored.Czernik and Bridgwater (2004) reviewed theresearch in this field, which demonstrates thatdirect application of bio-oils for heat and powergeneration is possible requiring only minormodifications of the existing equipment. Bio-oilhas been successfully used as boiler fuel and hasalso showed promise in diesel engine and gasturbine applications (Elliott et al., 1991; Solantausta

Wt%

Water 20-30

Lignin fragments: insoluble

pyrolytic lignin 15-20

Aldehydes: formaldehyde, acetaldehyde,

hydroxyacetaldehyde, glyoxal 10-20

Carboxylic acids: formic, acetic, propionic,

butyric, pentanoic, hexanoic 10-15

Carbohydrates: cellobiosan, levoglucosan,

oligosaccharides 5-10

Phenols: phenol, cresol, guaiacols, syringols 2-5

Furfurals 1-4

Alcohols: methanol, ethanol 2-5

Ketones: acetol (1-hydroxy-2-propanone),

cyclopentanone 1-5

The products from the liquefaction process includegases, water-insoluble bio-oil, and an aqueous phasewith water soluble products. A representativeproduct composition by weight basis of the initialbiomass feedstock is: 45% bio-oil, 25% gas (with>90% of gas being CO2), and 30% aqueous phase(with the aqueous phase consiting of 66% water,and 33% soluble organics) (Naber and Goudriann,2005). Several aqueous liquefaction pilot plantswere operated until the 1980’s (Moffatt andOverend, 1985; Elliott et al., 1991), but are nolonger in use. Biofuel B.V. is a startup companylocated in the Netherlands that is still activelypursuing aqueous liquefaction of biomass. Shownin the Figure 1.3 is a schematic of the Biofuel B.V.liquefaction pilot plant known as the HTU®

process.

1 . 3 R E S U LTING FUELS

The biomass-derived fuels produced from both thefast pyrolysis and liquefaction processes are highlydependent on the feestock properties as well asthe process conditions, making precise definitionsof the resulting products problematic.Nonetheless, the chemistries associated with thesetwo processes is somewhat different and havedistinguishing characteristics. Table 1.1 comparesthe compositions and properties of representativebio-oils from fast pyrolysis and liquefaction versusthat of heavy fuel oil (Czernik and Bridgwater,2004; Elliott and Schiefelbein, 1989; Huber et al., 2006).

1.3.1 Fast Pyrolysis Product A n a l y s i sBio-oil has significantly different physical andchemical properties compared to the liquid fromslow pyrolysis processes, which is more like a tar.Bio-oils are multi-component mixtures comprisedof different sized molecules derived primarily fromdepolymerization and fragmentation reactions ofthe three key biomass building blocks: cellulose,hemicellulose, and lignin. Therefore, the elemental

Ta ble 1.2 A general bre a k d own of thecomponents in fast py rolysis bio-oil

( B r i d g water et al., 2 0 0 1 ) .


et al., 1992; Grassi and Bridgwater, 1993; Peacockeand Bridgwater, 1994;Wornat et al., 1994; Czerniket al., 1995; Solantausta et al., 1995; Oasmaa andCzernik, 1999; Boucher et al., 2000; Shihadeh andHochgreb, 2000; Ganesh and Banerjee, 2001;Vitoloet al., 2001; Chiaramonti et al., 2003; Gayubo et al.,2004; Gayubo et al., 2004;Yaman, 2004). Severaltechnical challenges need to be overcome in orderto upgrade bio-oil to a quality of transport liquidfuel that would be broadly adopted andeconomically attractive (Bridgwater and Cottam,1992; Cottam and Bridgwater,,1994;Williams andHorne, 1995;Williams and Horne, 1995; Horne andWilliams, 1996;Vitolo et al., 2001;Yaman, 2004).

Though over 300 compounds have been identifiedin wood fast pyrolysis oil, they are found in smallamounts.Therefore isolation of specific singlecompounds is seldom practical or economicalbecause it usually requires complex separation

techniques. Some chemicals produced from thewhole bio-oil or by its fractionation are alreadycommercial products, for example liquid smokeused as a food-flavoring component. A few others,such as pyrolytic lignin as phenol replacement inresins or bio-oil based slow-release fertilizer, have achance for short-term commercialization, especiallyif a bio-refinery concept based on a fast pyrolysisprocess is implemented.

1.3.2 Liquefaction Product A n a l y s i sIn contrast to fast pyrolysis, significantly less workhas been reported on the chemical species analysisof liquefaction oils (Elliott 1981, Elliott 1981,Schirmer et al, 1984, Elliott 1985). Since a largeamount of decarboxylation and dehydration occursunder liquefaction conditions, the bio-oil producthas much lower oxygen content, and consequentlya higher heating value, than bio-oils from fastpyrolysis. Of the total oxygen removed during

F i g u re 1.4 Pyrolysis pat h way model (adapted from Radlein, 1 9 9 9 ) .



37

liquefaction, more than 50% has been attributed todecarboxylation (Goudriaan et al., 2001). Thewater-insoluble bio-oil derived from the BiofuelB.V. HTU® process is a heavy organic liquid thatsolidifies at 80ºC and has a H/C of 1.1 (Naber andGoudriann, 2005). The lower oxygen content inliquefaction bio-oil makes it a more attractive fuelthan fast pyrolysis bio-oil from the standpoint ofenergy density, but the high viscosity associatedwith the liquefaction bio-oil limits its utility withoutsome amount of upgrading.

1.4 SUMMARY OF PREVIOUSRESEARCH

1.4.1 Fast Pyro l y s i sSeveral published reviews provide extensiveaccounts of previous research on fast pyrolysis(Bridgwater and Peacocke, 2000; Bridgwater et al.,2001; Mohan et al., 2006). The key to maximizingthe yield of bio-oils in a fast pyrolysis reactor are:rapid heating, high heat transfer rates, a reactoroperating temperature of ~500ºC, and rapidcooling of the pyrolysis vapors. In addition tothese processing variables, the biomass alkalicontent has significant impact on the bio-oilcomposition because the alkali salts catalyzecracking to lower molecular weight species in thepyrolysis reactor. Due to the complexity of thereaction system, most of the studies have beenempirical and highly dependent on the reactorsystem and specific feedstock used in a given study.As a result detailed comprehensive models for fastpyrolysis have not been developed. Nonetheless,some high level models exist that can explain gross trends.

Radlein (1999) has proposed the pyrolysispathways shown in Figure 1.4. There are severalimportant features of this model that significantlyimpact the products from the fast pyrolysis

reaction. As noted in the figure, acid catalyzedreactions are thought to depolymerize the biomassto oligomers that can be subsequently convertedto monomers. The acids in the reaction, which arecarboxylic acids, are generated from the reactionand then participate in the conversion of otherreaction intermediates. Alkaline cations, which arepresent in lignocellulosic feedstocks, catalyzereactions to lower molecular weight species as wellas catalyzing ring opening reactions. Somepretreatment work has been performed in whichthe alkaline cations are removed from thefeedstock prior to pyrolysis. The pretreatmentleads to increased anhydrosugars in the bio-oil product.

Products from the fast pyrolysis reaction leave thereactor in three states: volatile gases that containcondensable and noncondensable fractions, liquidsthat are present as aerosols, and solids that consistof the char and some condensed liquid. The char isfirst removed from the effluent stream with acyclone. While a cyclone will remove much of thechar, some solid particles are not removed at thispoint and end up in the bio-oil, which hasimplications on the resulting bio-oil quality. Thecondensable gases are then quenched and collectedas the bio-oil. The method used to quench thesegases influences the chemical species ultimatelypresent in the bio-oil.

The bio-oil produced from fast pyrolysis containshundreds of chemical species making fullcharacterization of the bio-oil quite challenging. Assuch, many studies have only characterized bio-oilsby properties such as those given in Table 1.1.Unfortunately, this level of detail provides limitedinsight into the reactions taking place through thefast pyrolysis process. Reactions influencing thefinal bio-oil product occur in the fast pyrolysisreactor through the range of processes given in


Figure 1.1 as well as through char-facilitatedreactions and alkali-facilitated reactions. Anotherset of reactions occur during the condensationprocess. Finally, bio-oil in the condensed state isstill reactive and undergoes further conversion.Diebold (2002) suggested that the range ofreactions that can occur in fast pyrolysis bio-oilafter it has been condensed include esterificationof organic acids with alcohols, transesterification ofesters, homopolymerization of aldehydes, hydrationof aldehydes or ketones, hemiacetal formation ofaldehydes and alcohols, acetalization andtransacetalization of aldehydes and alcohols,phenol/aldehyde reactions to form resins,polymerization of furan derivatives, anddimerization of organic nitrogen from proteins.The high level of acidity in bio-oil can catalyze anumber of these reactions. The addition ofmethanol or ethanol to fast pyrolysis bio-oil at the10 wt% level was found to be effective in largelyretarding these condensed phase reactions(Diebold and Czernik, 1997).

A number of different reactor types have beenexamined to optimize the fast pyrolysis reactionsincluding, fluidized beds, circulating fluidized beds,ablative pyrolyzers, and vacuum pyrolyzers (Scott etal., 1999). In all cases, the goal is to maximize heattransfer to the biomass and minimize the residencetime of the pyrolysis products in the reactor.While each of these reactor systems hasadvantages and disadvantages, fluidized bedreactors appear to be the most economical andreadily scalable option.

1.4.2 LiquefactionWhile Biofuel B.V. continues the commercialdevelopment of liquefaction, it has receivedsignificantly less attention in the technical literaturerecently (Elliott 1981, Elliott 1981, Schirmer et al,1984, Elliott 1985). Several reviews provide anoverview of the liquefaction process through 1990

(Moffatt and Overend, 1985; Elliott et al., 1991). Toan even greater extent than with fast pyrolysis,there is little information available to comparereaction products of different liquefactionconditions because the reaction products havetypically only been evaluated at a very general levelas given in Table 1.1. Even the limited chemicalspeciation work done with fast pyrolysis bio-oil has not been replicated in the literature forliquefaction bio-oil. Therefore, prior work hasfocused on the empirical correlation betweenliquefaction conditions, primarily temperature,pressure, and residence time, with high levelproperties of the resulting bio-oil.

1.4.3 Sub and Supercritical Catalytic Reforming of Biomass

One important subset of biomass liquefaction isthe catalytic reforming of biomass in sub andsupercritical water. Sub and supercritical reformingof sugars and biomass-derived feedstocks canproduce H2 and/or methane (Matsumura, Minowaet al. 2005). Supercritical water conditions occurat conditions above the supercritical point of water(temperatures above 375oC and pressures above217 atm). Modell showed that supercriticalreforming of wood sawdust was able to producegaseous products and avoid coke formation(Modell 1977; Modell, Reid et al. 1978). Thus,supercritical reactions can be used to efficientlygasify glucose (and other biomass components)without coke formation. Heterogeneous catalystshave been used in subcritical reactions and havebeen shown to greatly change the productselectivity. The Battelle single-step subcriticalgasification reactor produces gas with highmethane levels at temperatures around 350oC andpressures 21 MPa with Ru or Ni catalystssupported on TiO2, ZrO2 or carbon (Elliott,Neuenschwander et al. 2004; Elliott, et al, 2006).Higher reaction temperatures (600oC and 34.5MPa) for supercritical reactions have been able to



39

produce H2 from supercritical reforming ofglucose (Matsumura, Minowa et al. 2005). Xu et al.showed that activated carbon is an efficient catalystfor supercritical gasification of glucose (Xu,Matsumura et al. 1996). At a WHSV of around 20h-1 close to 100 % of the glucose feed was gasifiedwith a molar gas composition of 22% H2, 34 %CO, 21% CO2, 15 % CH4, 6% C2H6, and 2%C3H8.

Other biomass feedstocks including whole biomasscan also be used for supercritical gasification. Theadvantages of supercritical reforming are that highreaction rates are obtained, impure feedstocks canbe used, wet feedstocks can be processed withhigh thermal efficiencies, product gas is produced ina single reactor, and the product gas is available at

high pressure. The disadvantages of supercriticalreforming are the high capital cost of a high-pressure reactor and that H2 can be selectivelyproduced only at high temperatures where largeamounts of CO are also produced. Supercriticalreforming is an excellent way to produce productgases from aqueous biomass mixtures.

1.5 ECONOMICS AND POTENTIAL OF TECHNOLOGY

The most recent direct comparison of theeconomics of the fast pyrolysis versus theliquefaction process was presented by Elliott et al.,1990 and is given in Table 1.2. The results arepresented in 1990 U.S. dollars and assume a 1000

Fast pyrolysis Liquefaction in solventPresent Potential Present Potential

Total capital requirement ($US millions)primary liquefaction 49.8 26.4 84.2 48.4crude upgrading 46.6 34.3 26.8 26.0product finishing 14.5 0.7 15.3 0.7total 110.9 61.4 126.3 75.1

Production costs ($US million/year)fixed operating costs 14.48 10.77 14.48 10.03variable operating costs 25.74 23.67 33.44 33.60(feedstock costs) (20.00) (20.00) (20.00) (20.00)capital charges 12.96 7.17 14.75 8.78total production cost 53.18 41.61 62.67 52.39

Minimum selling price ($US/GJ)bio-oil 9.32 6.91 13.44 12.27refined bio-oil 16.24 12.99 19.54 14.77

Process thermal efficiency (energyliquid products/energyfeed+inputs)

primary liquefaction prod. 0.61 0.68 0.55 0.48finished product 0.52 0.53 0.48 0.49

* See the current version of D 1655 for complete requirements.

Ta ble 1.2 A comparison in 1990 U. S . dollars of the economics of pro d u c i n gbio-oil from wood chips using either a fast py rolysis or liquefaction

p rocess ( t a ble adapted from Huber et al., 2 0 0 6 ) .


The full potential of biomass as a resource for

transportation fuels, fine chemicals or syngas can

only be reached through improved understanding of

the underlying conversion chemistry. This in turn

means the development of predictive models, which

will support the process design and optimization of

a conversion process. Given the complexity of

biomass, highly detailed model predictions will be

impossible in the foreseeable future. However, taking

gasification applications as an example, such

mechanisms should at least be able to reliably

predict hydrogen to carbon monoxide ratios as well

key reactive species and byproducts. Similarly, a good

model should be able to support the search for the

reaction conditions that will result in more valuable

reaction products in bio-oil upgrading.

While development of such a kinetic mechanism is a

formidable challenge, several factors suggest

significant progress is possible: (1) Despite the

overall complexity of biomass, it consists largely of

just a few chemically relevant units (e.g., lignin has

many phenyl, alcoholic and ether groups and bio oil

has acid groups); (2) The reaction chemistry is

thought to be analogous to the chemistry of well-

studied hydrocarbon systems. Thus, one expects that

the same reaction types (H abstraction, -scission

and isomerization) will dominate, even though the

specifics will differ to reflect the weakening of C–H

bonds next to the oxygen functionalities in biomass.

(3) Well-established theoretical tools (e. g., electronic

structure methods) are readily available and can be

applied either directly or with minor modifications

for oxygenated substances such as biomass.

Building on Colorado School of Mine’s expertise in

developing accurate detailed kinetic models of

hydrocarbon systems, a joint CSM/NREL research

program is getting underway to apply this approach

to biomass conversion. This research project is

focused on a quantitative characterization of

oxygenated model species that have been selected to

represent the classes of compounds found in

biomass. It is based on high-level electronic

structure calculations combined with statistical

mechanics methods to calculate thermodynamic

properties of reactants, products and intermediate

species as well as transition states. Transition state

theory provides rate constants at the high pressure

limit. QRRK theory with the modified strong collision

assumption yields pressure-dependent rate

expressions. We can generalize these results in “rate

rules” that allow us to assign rate expressions to

reactions that cannot directly be treated by ab

initio methods. These results, validated against

experimental data if possible, will be used to extend

existing chemical kinetic mechanisms to include

biomass. Initial results for hydrogen abstraction

reactions by H-atoms (Figure S-1.2) suggests that

the connection between reaction exothermicity and

activation energy for the oxygenates is different

than for hydrocarbons. Furthermore, these results

also suggest the importance of steric effects (e.g.,

the exothermicity of abstraction of a tertiary H in

CH3OCHR2 is less than expected).

DEVELOPMENT OF DETAILED KINETIC MECHANISMS FOR MODEL BIOMASS COMPOUNDS

F i g u re S-1.2 Comparison of the computeda c t i vation energies for H abstraction reactions by

H atoms from (A) hy d rocarbons (black line,s q u a res) and (B) alcohols/ethers (red line; c i rc l e s )

as a function of the reaction energ e t i c s .



41

dry metric ton/day of biomass processed using abiomass basis of wood chips with 50% moisture ata cost of $30/metric ton. The best estimate for thestate of each technology in 1990 as well as theestimated potential for the technology with furtheradvancements is shown. The table includes thecost associated with fast pyrolysis and liquefactionportions of proposed processes as well as thecosts associated with upgrading the resulting bio-oil to a more useable refined state. Due to thehigh pressures associated with liquefaction, thecapital cost for a fast pyrolysis unit would besignificantly less expensive. The differencediminishes to some extent when bio-oil upgradingis also considered, since the bio-oil from fastpyrolysis will require more significant upgrading.However, the side-by-side assessment as presentedin Table 1.2 suggests that the fast pyrolysis processcurrently enjoys a significant economic advantageover liquefaction. The “potential” columns of thistable depict how additional technologicaladvancements are expected to increase theeconomic attractiveness of fast pyrolysis-derivedfuels even further. For this reason, liquefactionprocessing is generally focused on biomassfeedstocks that have high water content, which aremore problematic for processing through fastpyrolysis. Additional economic assessments of fastpyrolysis can be found elsewhere (Bridgwater et al.,2000; Bridgwater et al., 2002) as well as forliquefaction (Naber et al., 1999).

1.6 CURRENT T E C H N O L O G YL I M I TATIONS AND RESEARCH/DEVELOPMENT NEEDS

1.6.1 Fast Pyro l y s i sAn important challenge to the use of fast pyrolysisbio-oil for production of liquid transportation fuelsis its high oxygen content. To make the bio-oilmore compatible with the existing transportationfuels infrastructure, much of the oxygen needs tobe removed.

The high oxygen content issue is particularly acutefor the fraction of the bio-oil that comes from thecellulose and hemicelluloses fractions oflignocelluloses. The bio-oil fraction from lignin issignificantly lower in oxygen content.

In previous studies, the bio-oil was hydrotreated athigh pressures (2000 – 2500 psi) and low spacevelocities (0.1 – 0.2 LHSV). The resultinghydrotreated oil was then cracked in a fluidcatalytic cracker, or hydrocracker, to producegasoline. At these high pressures and low spacevelocities, hydrodeoxygenation predominates.However, large quantities of hydrogen are requiredduring hydrodeoxygenation due to the high level ofoxygen in the bio-oil. Alternative strategies thatrequire the use of less hydrogen to remove oxygenwould be quite attractive. It is possible that thesealternative strategies would work better with acertain product distribution in the bio-oil.Processing to upgrade bio-oil is intimately linkedwith how the fast pyrolysis unit is operated.Therefore, research is needed to establish the bio-oil characteristics that would be most desirable toobtain from the fast pyrolysis reactor system.

Another serious problem for fast pyrolysisprocessing is the high acid number of the bio-oils,which will cause corrosion in standard refineryunits. Although the bio-oils can probably beprocessed using 317 stainless steel cladding, thismaterial is not standard in refinery units making itdifficult to introduce bio-oil into the existingrefinery infrastructure. Therefore pyrolysis bio-oilsrequire pre-processing to reduce the acid numberbefore processing in typical refinery units.Research is needed to understand how toaccomplish this pre-processing in an efficientmanner. There is an opportunity to introducechemical transformations during this pre-processing that would not only reduce the acidnumber, but also decrease the oxygen content,


area than the fast pyrolysis area. Some work hasbeen performed with solvents other than water,which would potentially allow operation atsomewhat lower pressures. However, liquefactionsystems using alternative solvents create a myriadof new problems, so none are currently indevelopment.

As with fast pyrolysis, the chemistry in biomassliquefaction is complex. So, thus far, investigatorshave attempted to draw only empiricalrelationships between reaction conditions andbiomass composition. In addition, significantly lesswork has been reported on analyzing the chemicalcomposition of liquefaction bio-oil than fastpyrolysis bio-oil. Therefore, further advancement inthis process area needs to be supported by a morefundamental understanding of the chemicalreactions that occur during the liquefactionprocess. There is also an opportunity to modifythis chemistry by introducing a catalyst into the process.

1.7 RECOMMENDAT I O N S

Research needed to advance the technology ofselective thermal processing of lignocellulosicbiomass can be grouped into six topics: overarchingtechnical needs, plant characteristics, feedstockpreprocessing, deconstruction selectivity, bio-oilrecovery (fast pyrolysis), and alternativedeconstruction approaches.

1.7.1 Ove r a rching Technical Needs

•Development of more detailed thermaldeconstruction microkinetic models that canaccount for a broad range of chemical reactions.These can be used to provide a basis forchoosing which reactions should be enhancedto increase selectivity for desired chemical

thereby addressing two bio-oil challenges.Due to the complexity of the biomass pyrolysisreaction system, the underlying chemistry is notwell understood. Therefore, the correlations thathave been developed between bio-oil compositionand pyrolysis and condensation conditions areempirical. The lack of basic understanding of thereaction system limits the ability to draw generalrelationships between biomass composition,pyrolysis conditions, condensation conditions, andbio-oil composition. Therefore, research is needed to develop better foundational chemistryknowledge about the fast pyrolysis reaction, whichcan drive the process to more desirable productcompositions.

As mentioned above, the alkaline cations in the lignocellulose feedstocks serve as a catalyst forreactions in the pyrolysis reactor. While this resultis understood empirically, little work has beendone on the intentional manipulation of thesealkaline cations. Instead, research has largelyfocused on examining how their naturally varyinglevels in native lignocellulose result in different bio-oil properties or on their removal in a pre-processing step. There is a need for a moresystematic understanding of the role of differentalkaline cations, as well as the alkaline cationconcentration, on the pyrolysis reaction.Additionally, the intentional introduction ofdifferent catalytic moieties could be used to alterthe product bio-oil composition.

1.6.2 LiquefactionDue to its lower oxygen content, bio-oil fromliquefaction has more desirable properties thanbio-oil from fast pyrolysis. However, the highpressure conditions required in the liquefactionprocess makes the capital requirement significantlyhigher than for a fast pyrolysis process. Since thislimitation is intrinsic to the process, there issignificantly less work ongoing in the liquefaction



43

BIOMASS CATA LYTIC CRACKING

One of the key technical problems that must be

solved in order to achieve cost-effective conversion

of the non-edible biomass is the problem of how to

open up the inaccessible solid fibrous 'woody'

material, so that it can be effectively transformed by

catalysts. Cellulose, the main component of wood, is

a very pure polymer of glucose, and once 'unlocked'

the woody structure can be converted into high

quality bio-oils as shown in Figure S-1.3.

The challenge is to find a simple non-energy

intensive way to make the woody biomass accessible

to reactive media, catalysts and/or enzymes. Biomass

catalytic conversion (BCC), demonstrated in Figure

S-1.4, is an improvement to classical biomass

pyrolysis, whereby catalytically accessible biomass is

converted into a bio-crude suitable for further

processing in new or existing oil refineries to

gasoline and/or diesel fuel.

Because of the enhanced accessibility of the biomass

to the catalyst, the catalyst in BCC does not only

improve the secondary tar cracking, like in classical

catalytic pyrolysis, but the catalyst also enhances

the kinetics and selectivity of the primary

F i g u re S-1.3Unlocking the woody stru c t u re

conversion of the solid biomass. This allows for

conversion at lower temperatures (improved energy

efficiency), while producing improved quality

products. BCC can build on the long history of fluid

catalytic cracking (FCC) technology the low cost

workhorse of the oil refining industry. BCC is a new

innovation of the FCC process based on low cost and

environmental friendly synthetic catalyst technology

developed by KIOR, a privately funded venture.

F i g u re S-1.4 Pat h ways for conve r s i o nof lignocellulosic biomass to bio-oils,

g a s e s , and chars


species in the resulting bio-oil.•Better characterization of the catalytic pathwaysthat occur due to acid generation in the thermalprocessing and the presence of alkalis.

•To complement the kinetic modeling, there is aneed to determine the thermodynamic andphysical properties of key chemical speciespresent in thermal processing, since little dataare currently available for these species.

•The chemical complexity of the bio-oilsproduced by selective thermal processingnecessitates the development of analyticaltechniques that can give chemical speciation.Without this information, experimental workcan only focus on lumped property effects asgiven in Table 1-1, which provides little insightinto the underlying chemistry. The lack ofanalytical characterization has historically causedmany thermal processing studies to be highlyempirical.

1.7.2 Plant Characteristics for Thermal Pro c e s s i n g

•For thermal processing, the lignin fraction of theplant biomass is decomposed into chemicalspecies that are lower in oxygen content thanthose from cellulose or hemicelluloses.Therefore, plant feedstocks with higher lignincontent and lower hemicellulose content wouldlikely be preferred for selective thermalprocessing. The desire for higher lignin-containing feedstocks is in opposition to theplant composition desired for biologicaldeconstruction, which seeks to diminish thelignin content.

•For fast pyrolysis-type processing, there is adesire to decrease the water content in the biomass feedstocks.

•Diminished mineral content in the plant materialwould potentially simplify the selective

thermochemical processing.

•Designing plants with higher oil/wax contentcould lead to bio-oils that require less oxygenremoval during the subsequent bio-oil upgrading.

1.7.3 Biomass Feedstock P re p ro c e s s i n g

•Development of preprocessing methods toefficiently remove alkalis from the biomass couldprovide better control over thermaldeconstruction process.

•The lignin and cellulose/hemcellulose fractionsyield distinctly different chemical species afterthermal processing. Therefore, improved bio-oilchemical specificity could be accomplished if abiomass separation process was available.

•Catalyst incorporation as a preprocessing stepthat could improve selectivity to desired speciesin the subsequent thermal processing.

•Torrefaction is a low temperature process(300°C) that removes volatile species frombiomass. It would be useful to evaluate whetherthis preprocessing step, which also densifies thebiomass, leads to a feedstock with improvedthermal processing characteristics.

•There is a need to better understand therelationship between biomass particlemorphology and its thermal processingcharacteristics to determine what is the optimalsize and shape to create during preprocessing.

•The pretreatment of biomass, which is beingdeveloped for the biological conversion process,enhances the accessibility of the biomass toenzymatic hydrolysis. How would thispretreatment procedure affect biomassdeconstruction through thermal processing?



45

BIO-OIL FERMENTAT I O N

Under rapid pyrolytic conditions, pure cellulose yieldslevoglucosan, an anhydrosugar with the same

empirical formula as themonomeric building block ofcellulose: C6H10O5 (Evans andMilne, 1987). A “dehydrated”sugar, it can be hydrolyzed toglucose. Thus, fast pyrolysispresents an alternative

approach to enzymatic hydrolysis for liberatingsimple sugars from biomass. However, most biomasscontains small amounts of alkali, which promotes theformation of hydroxyacetaldehyde instead oflevoglucosan, with the result that bio-oil typicallycontains only a few percent of levoglucosan andother anhydrosugars. Scott and coworkers at theUniversity of Waterloo discovered that by dilute acidwashing of wood before pyrolysis levoglucosan can beproduced at very high yields (Scott et al., 1989).

Brown and his collaborators evaluated the effect ofseveral treatments to remove alkali on the pyrolyticproducts of cornstover (Brown et al., 2001). Thesewere able to increase the yield of anhydrosugarsfrom less than 3 wt-% to as high as 28 wt-%. Acidhydrolysis of this anhydrosugaryielded 5% solutions of glucose andother simple sugars. The resultingglucose solutions can be fermented(Prosen et al., 1993). However, theresulting substrate derived fromthe bio-oil contains fermentationinhibitors that must be removedor neutralized by chemical orbiological methods

A biorefinery based onfermentation of bio-oil (shown inFigure S-1.5) would include severalunit operations that arereminiscent of those used in abiorefinery based on enzymatichydrolysis. Fibrous biomass ispretreated with dilute acid to

simultaneously remove alkali and hydrolyzes thehemicellulose fraction to pentose. The remainingfraction, containing cellulose and lignin, is dried andpyrolyzed at 500°C to yield bio-oil, gas, and char.The bio-oil is separated into a levoglucosan-richaqueous phase and pyrolytic lignin through a simpleprecipitation process. The levoglucosan is acid-hydrolyzed into glucose. After detoxification, thehexose and pentose recovered from the biomass areseparately fermented to ethanol. The lignin, char, andgas are burned to generate steam for distillation andother process heat requirements of the plant.

So and Brown (1999) have compared the capitalexpense and operating costs for a bio-oilfermentation biorefinery. The total capitalinvestment for an ethanol plant based onfermentation of bio-oil producing 95 million liters ofethanol was estimated to be $69 million, while theannual operating cost was about $39.2 million,resulting in an ethanol production cost of $0.42/L.The production cost was comparable to that of acidhydrolysis and enzymatic hydrolysis of woodybiomass to produce ethanol for the year of the analysis.

Bio-Oil vapor

Fermenter

Hot WaterExtraction

Pentose

Pyrolyzer

Fiber

FiberByproduct

Drier Cyclone

Char

Bio-OilRecovery

PhaseSeparation

Detoxification

Fermenter

Distillation

ETHANOL

WATER

LIGNIN

Anhydrosugar & othercarbohydrate

F i g u re S-1.5: Cellulosic ethanol plant concept based onf e r m e n t ation of bio-oils.


1.7.4 Thermal Deconstruction S e l e c t i v i t y

•While some work on catalyst incorporationdirectly into the thermal deconstruction hasbeen performed, there is a need to moresystematically evaluate the product selectivitypossible with different types of catalyticmaterials.

•Process condition optimization has largely beenbased on empirical studies, to date. Improvedmicrokinetic and reactor models will allow amore detailed understanding of the optimalprocess conditions.

•In addition to microkinetic models for individualreactants, improved understanding of thedepolymerization reaction mechanisms is needed.

•Both hydrolysis and cracking reactions playimportant deconstruction roles. More researchis needed to understand the appropriatebalance between these reactions in thermalprocesses.

•Catalytic materials will need to be designed thatprovide high selectivity, but also maintain theirphysical integrity in the thermal process and bereadily recovered from the product.

•Aerosol formation is an important mechanismfor nonvolatile bio-oil components to pass outof the pyrolysis reactor, there is a need tobetter understand this mechanism.

•Research is needed to explore the possibilitythat the selectivity of the deconstructionprocess can be altered by the introduction of co-reactants.

•While torrefaction has the potential to be auseful preprocessing step, there might instead bethe possibility to improve the selectivity of fastpyrolysis by operating under staged reaction

conditions. 1.7.5 Bio-Oil Recove ry

( Fast Pyro l y s i s )

•In the fast pyrolysis process bio-oil is recoveredthrough condensation. The bio-oil compositionis influenced by the condensation approachused, so there is a need to better understandthe reactions occurring during bio-oil condensation.

•Improved downstream separation wouldimprove the capability to segregate bio-oilfractions derived from the lignin orcarbohydrate fractions, which have verydifferent chemical characteristics.

•It is desirable to remove as much of the charfrom the bio-oil as possible. Introduction of acatalytic filter downstream from the char-removal cyclone could serve the dual purposeof particulate removal and conversion of thebio-oil vapor to alternative products prior tocondensation.

•Bio-oil stabilization will be needed if the bio-oilis to be produced at one location but shippedto another for upgrading. Stabilization strategiescould include catalytic reactions.

1.7.6 A l t e rn at i ve Deconstruction A p p ro a c h e s

•Alternative solvents to water could beconsidered in the liquefaction process.

•Elimination of alkalis from lignocellulosicfeedstocks increases the production ofanhydrosugars in the fast pyrolysis process.These anhydrosugars could be fermented tofuels and chemicals thereby making a coupledchemical/biological process.

•Ionic liquids have potential for selective thermal



47

deconstruction of biomass.1 . 8 REFERENCES

Boucher, M. E.;A. Chaala, H. Pakdel, C. Roy; BiomassBioenergy 19 (2000) 337-350.

Bridgwater,A.V.; J Anal.Appl. Pyrolysis 51 (1999) 3-22.

Bridgwater,A.V.; M. L. Cottam; Energy Fuels 6(1992): 113-120.

Bridgwater,A.V.; G.V.C. Peacocke; RenewableSustainable Energy Rev. 4 (2000) 1-73.

Bridgwater,A.V.; S. Czernik, J. Piskorz; in Progress inThermochemical Biomass Conversion,A.V.Bridgwater, ed. (2001) 977-997.

Bridgwater,A.V.;A.J.Toft, J.G. Brammer; RenewableSustainable Energy Rev. 6 (2002) 181-248.

Brown, R. C.; Pometto,A. L.; Peeples,T. L.; Khiyami,M.;Voss, B.; Kim J.W.; Fischer, S.; Proceedings ofthe Ninth Biennial Bioenergy Conference,Buffalo, New York (2000).

Brown, R. C.; Radlein, D.; Piskorz, J.; Chemicals andMaterials from Renewable Resources:ACSSymposium Series No. 784,Washington, D.C.,American Chemical Society: (2001) 123-132.

Chiaramonti, D.,A. Bonini, E. Fratini, G.Tondi, K.Gartner,A.V. Bridgwater, H.P. Grimm, I. Soldaini,A.Webster, P. Baglioni; Biomass Bioenergy 25(2003) 101-111.

Cottam, M. L.;A.V. Bridgwater; Biomass Bioenergy7 (1994) 267-273.

Czernik, S.;A.V. Bridgwater; Energy Fuels 18 (2004)590-598.

Czernik, S.; J. Scahill, J. Diebold; J. Solar Energy Eng.117 (1995) 2-6.

Diebold, J.P. in Fast Pyrolysis of Biomass: HandbookVolume 2,A.V. Bridgwater, ed. (2002) 243-292.

Diebold, J.P.; S. Czernik; Energy Fuels 11 (1997)1081-1091.

Elliott, D.C. (1981) In: Fuels from Biomass andWastes, Chapter 24, pp. 435-450 (1981). D. L.Klass and G. H. Emert, eds.Ann Arbor Science

Publishers, Inc.,Ann Arbor, Michigan.Elliott, D.C. In: Biotech and Bioeng, Sym #11,

(1981) pp. 187-198, John Wiley & Sons, NewYork City.

Schirmer, R.E.,T.R. Pahl, and D.C. Elliott. Fuel 63(1984) 4-8.

Elliott, D.C. In: Fundamentals of ThermochemicalBiomass Conversion, (1985) pp. 1003-1018. R. P.Overend,T.A. Milne, and L. K. Mudge, eds.,Elsevier Applied Science Publishers, London.

Elliott, D.C.; E.G. Baker, D. Beckman,Y. Solantausta,V.Tolenhiemo, S.B. Gevert, C. Hornell,A.Ostman, B. Kjellstrom; Biomass 22 (1990) 251-269.

Elliott, D.C.; D. Beckman,A.V. Bridgwater, J.P.Diebold, S.B. Gevert,Y. Solantausta; Energy Fuels5 (1991) 399-410.

Elliott, D.C.; G.F. Schiefelbein; Preprints of Papers -American Chemical Society, Division of FuelChemistry 34 (1989) 1160.

Elliott, D.C.; Neuenschwander, G.G. et al. (2004).Ind. Eng. Chem. Res. 42 (2004) 1999-2004.

Elliott, D.C.; Hart,T.R.; Neuenschwander, G.G. Ind.Eng. Chem. Res. 45 (2006) 3776-3781.

Evans, R. J.; Milne,T.A.; Energy Fuels 1 (1987) 123-137.

Ganesh,A.; R. Banerjee; Renewable Energy 22(2001) 9-14.

Gayubo,A. G.;A.T.Aguayo,A.Atutxa, R.Aguado, J.Bilbao; Ind. Eng. Chem. Res. 43 (2004) 2610-2618.

Gayubo,A. G.;A.T.Aguayo,A.Atutxa, R.Aguado, M.Olazar, J. Bilbao; Ind. Eng. Chem. Res. 43 (2004)2619-2626.

Goudriaan, F.; D.G.R. Peferoen; Chem. Eng. Sci. 45(1990) 2729-2734.

Goudriaan, F.; B. van de Beld, F.R. Boerefijn, G.M.Bos, J.E. Naber, S. van der Wal, J.A. Zeevalkink; inProgress in Thermochemical BiomassConversion,A.V. Bridgwater, ed. (2001) 1312-


1325.Grassi, G.;A.V. Bridgwater; Biomass Bioenergy 4

(1993) 339-345.Horne, P.A.; P.T.Williams; Fuel 75 (1996) 1043-

1050.Huber, G.W.; S. Iborra,A. Corma; Chem. Rev. 106

(2006) 4044-4098.Khiyami, M.A.; Pometto III,A. L.; Brown, R. C.; J.

Agric. Food Chem. 53 (2005) 2969-2977.So, K.; Brown, R. C.;Appl. Biochem. Biotechnol. 77-

79 (1999) 633-640.Matsumura,Y.,T. Minowa, et al. Biomass Bioenergy

29 (2005) 269-292.Modell, M. (1977). Reforming of Glucose and Wood

at the Critical Conditions of Water.Am. Soc.Mech. Eng.

Modell, M., R. C. Reid, et al. (1978). GasificationProcess. US Patent 4,113,446. USA,Massachusetts Institute of Technology.

Moffat, J.M.; R.P. Overend; Biomass 7 (1985) 99-123.Mohan, D.; C.U. Pittman, P.H. Steele; Energy Fuels

20 (2006) 848-889.Naber, J.E.; F. Goudriaan,A.S. Louter; Proceedings

3rd Biomass of the Americas (1997) 1651-1659.Naber, J.E.; F. Goudriaan,A.S. Louter; Proceedings

4rd Biomass of the Americas (1999) 789-797.Naber, J.E.; F. Goudriaan;ACS National Meeting,

Washington, DC (2005).Oasmaa,A.; S. Czernik; Energy Fuels 13 (1999) 914-

921.Peacocke, G.V.C.;A.V. Bridgwater; Biomass

Bioenergy 7 (1994) 147-154.Prosen, E. M.; Radlein, D.; Piskorz, J.; Scott, D. S.;

Legge, R. L.; Biotechnol. Bioeng. 42 (1993) 538-541.

Radlein, D.; in Fast Pyrolysis of Biomass: HandbookVolume 1,A.V. Bridgwater, ed. (1999) 164-188.

Scott, D. S.; Czernik, S.; Piskorz, J.; Radlein, D.;Energy from Biomass and Wastes XIII, NewOrleans, LA, USA, Publ by Inst of Gas

Technology, Chicago, IL, USA (1989).Scott, D.S.; Majerski, P.; Piskorz, J.; Radlein, D.; J.Anal.

Appl. Pyrolysis 51 (1999) 23-32.Shihadeh,A.; S. Hochgreb; Energy Fuels 14 (2000)

260-274.Solantausta,Y.; D. Beckman,A.V. Bridgwater, J.P.

Diebold, D.C. Elliot; Biomass Bioenergy 2 (1992)279-297.

Solantausta,Y.;A.T. Bridgwater, D. Beckman;Biomass Bioenergy 9 (1995) 257-269.

US Department of Labor, "Chemical Manufacturing,Except Pharmaceutical and MedicineManufacturing", U.S. Department of Labor, 2005.

Vitolo, S.; B. Bresci, M. Seggiani, M.G. Gallo; Fuel 80(2000) 17-26.

Williams, P.T.; P.A. Horne; Fuel 74 (1995) 1839-1851.

Williams, P.T.; P.A. Horne; J.Anal.Appl. Pyrolysis 31(1995) 39-61.

Wornat, M. J.; B. G. Porter, N.Y.C.Yang; Energy Fuels8 (1994) 1131-1142.

Xu, X.,Y. Matsumura, et al. Ind. Eng. Chem. Res. 35(1996) 2522-2530.

Yaman, S.; Energy Conversion Management 45(2004) 651-671.

1 . 8 REFERENCES



49


2 . Utilization of Pe t roleum RefiningTe c h n o l ogies for Biofuel Pro d u c t i o n2.0 UTILIZATION OF PETRO L E U M

REFINING TECHNOLOGIES FORBIOFUEL PRO D U C T I O N

OVERVIEW: Liquid biofuel precursors producedthrough pyrolysis or hydrothermal liquefaction ofbiomass can be converted to hydrocarbon fuelsincluding “green” gasoline, diesel and jet fuel byutilizing the same refining technologies applied topetroleum and coal-liquid conversion. Specifictechnologies covered include hydrotreating,hydrocracking and catalytic cracking. A majorresearch challenge associated with this thrust ishow to efficiently remove oxygen from biomass-derived feedstocks while minimizing theconsumption of expensive hydrogen. The complexnature of biomass and the ill-defined issues relatedto biomass-to-biofuels conversion pose asubstantial, but not insurmountable, challenge tothe large-scale production and widespread use of

biofuels. The following chapter provides data,descriptions, and schematic comparisons of thevarious technologies that can be used to producefuels from biomass within the existinginfrastructure of petroleum refineries. We alsoidentify the outstanding research and engineeringissues related to biomass conversion that must beaddressed in order to realize the tremendouspotential economic and environmental rewards oflarge-scale domestic biofuels production.

2 . 1 I N T RO D U C T I O N

Biomass-derived fuels have the potential to helpmeet the increasing global demand for transpor-tation fuel while reducing greenhouse gasemissionsby closing the carbon cycle loop; CO2 emitted bycombustion of biofuels is subsequently removedfrom the atmosphere during biomass re-growth

F i g u re 2.1 A re p re s e n t at i ve flow diagram of a petroleum re f i n e ry [ E n e rg e t i c s , 2 0 0 6 ]


SECTIONS OF T H RUST 2 :

2.1 Introduction


2.3 Resulting Fuels




2.7 Recommendations

2.8 References

PA RT I C I PA N T S :

Phillip Badger; Calvin H. Bartholomew;André L. Boehman; Curt Conner;Steven Czernik;Anthony Dean; BurtDavis; Douglas Elliott*; Kristi Fjare;Manuel Francisco; Bruce Gates; JamesGoodwin; Harold Kung; Leo Manzer;Richard Marinangeli; Chunshan Song*;Don Stevens; Philip Steele;TiffanyWestendorf; Phil Westmoreland (*Co-chairs and contact for this chapter)

[Song 2006]. Bio-oils can be produced through theselective thermal processing of a wide variety oflignocellulosic biomass feedstocks, as described inChapter 1. Bio-oils can be used for the productionof both fuels and chemicals [Czernik andBridgwater 2004]. However, bio-oils are inherentlycomplex, acidic, and thermally instable. Therefore,they must be upgraded before they can be used asa transportation fuel.

The objective of this thrust is to identify promisingpathways towards clean liquid hydrocarbontransportation fuels using technologies similar inprinciple to those employed in fossil fuels refining.The advantage of this approach is that it exploitsknown technological processing schemes and takesadvantage of the existing petroleum refininginfrastructure [Huber and Corma 2007 andMarinangeli et al. 2006].

2 . 2 OVERALL PRO C E S SD E S C R I P T I O N

The petroleum refining technologies for biofuelproduction discussed in this chapter includehydrotreating, hydrocracking, catalytic cracking, andhydrodeoxygenation, which are similar to thetechnologies used in petroleum refining, coal liquidupgrading, and Fischer-Tropsch product refining(Figure 2.1).The reasons for considering theseprocesses are two-fold. First, these processes couldtake advantage of existing infrastructure. Second,such processes produce clean liquid hydrocarbonfuels that can be transported through exisitingpipelines and used in current and next-generationinternal combustion engines for land, sea and airtransportation. (See Sidebar on next page)

2 . 3 R E S U LTING FUELS

The technologies discussed in this chapter are ableto convert bio-oils and other biomass derivedfeedstocks into clean hydrocarbon-based liquidbiofuels for transportation including “green”gasoline, jet fuel and diesel as well as organicchemicals.The feedstocks that can be processedwith petroleum technology include organic liquidsderived from selective thermal processing oflignocellulosic biomass. Examples of such


O P P O RTUNITIES FOR BIORENEWABLES IN PETROLEUM REFINERIES

Use of existing infrastructure would significantly

decrease the ramp-up time for economical, large-scale

production of advanced biofuels. However, despite the

heightened interest in the potential for domestic

biofuels production, there has so far been little

integration of biofuels into the existing infrastructure

of petroleum refineries. Using the existing

infrastructure of petroleum refineries could

significantly reduce the cost of biofuels production.

Instead of the five years it normally takes to build a

new fuel plant, existing petroleum refineries could be

updated to accommodate biofuels production in less

than a year.

If economical opportunities for blending or co-

processing them within traditional petroleum

refineries were identified and developed biofuels

could quickly and substantially alleviate the

increasing demand for petroleum. Indeed, many

current petroleum refineries are already starting to

commercialize biofuels produced using existing

petroleum refinery technology. For example, Conoco-

Phillips and Tyson Foods recently signed an

agreement to produce diesel fuel through

hydoprocessing of waste vegetable oils. In addition

UOP, the National Renewable Energy Laboratory, and

the Pacific Northwest National Laboratory completed

a U.S. Department of Energy-funded evaluation of the

economics of producing biofuels within petroleum

refineries [Holmgren 2005; Marker, et al. 2005]. The

purpose of this project was to identify economically

attractive opportunities for biofuels production and

blending using petroleum refinery processes.

Economic analyses were conducted to assess a

variety of potential processes and configurations

using process modeling and proof-of-principle

experiments.

The DOE study identified

many promising and

profitable options for

integrating biorenewable

feeds and fuels into

existing refineries. Figure

S-2.1 shows a schematic

of several options for

biofuels production from

different biomass

sources. Some of the

routes are already in

commercial practice, such

as ethanol from the

fermentation of corn

or sugar cane.

Several routes have a

considerably more

distant commercialization

horizon due to the

Green Diesel

Natural Oils

C6 Sugars

Lights

CO2Feedstocks Products

H2O

Bio -oil-

Distiller ’s Grain

Sugars

Optional Petroleum

Starches

C5

/ C6

Sugar

s

Co-Feed

Lignin, Cellulose &

Hemicellulose

Enzyme Conversion

Fermentation Dehydration Ethanol

Acid or Enzyme Hydrolysis

Direct Conversion

Pyrolysis/Thermal Depolymerization

Gasification

FCC

Hydrotreating

Transesterification

Hydrotreating

Fischer Tropsch

Alcohol Synthesis

Green Gasoline

Glycerine

FAME or FAEE

H2OSyngas

H2O

H2O

F i g u re S-2.1 Ove rv i ew of Biofuel Pro d u c t i o n


2 . Utilization of Pe t roleum Refining Te c h n o l ogies for Biofuel Pro d u c t i o n

53

related scientific and technical challenges and

limited feedstock availability. However, new

nanotechnology, quantum chemical and molecular

engineering methods developed over the last 10

years that will help accelerate these processes

into reality.

One of the most promising options for

lignocellulosic-derived biofuels is the use of

petroleum refineries for upgrading bio-oils. Bio-oils

are produced by fast pyrolysis, a thermochemical

process with the potential to convert large volumes

of lignocellulosic biomass into liquid fuels. In this

process a solid biomass feedstock is injected into a

fluidized bed with high heat transfer capability for

short contact times, followed by a quenching process

to condense a liquid bio-oil in 50-75% yields, with

gas and char making up the balance. The bio-oil

contains the thermally cracked products of the

original cellulose, hemicellulose, and lignin fractions

present in the biomass. It also contains a high

percentage of water, often as high as 30%.

Many economically attractive opportunities for the

integration of biorenewable feedstocks and biofuels

in petroleum refineries were identified in this study,

including pyrolysis oil to produce “green gasoline”.

However, this technology is still in its infancy. In

order to become a viable production technology,

pyrolysis oil processing will require additional

scientific and technical development. Over the long

term, however, the potential volume of pyrolysis oil

could rapidly and economically replace global

shortages of petroleum fuel and help alleviate the

environmental, economic and social problems related

to petroleum dependence and greenhouse gas

emissions.

feedstocks include bio-oils produced through fastpyrolysis of biomass and liquids from hydrothermalliquefaction of lignin and cellulose in biomass.

To meet government regulations for fuel sulfur andaromatic contents, petroleum-based fuels must be“cleaned” through a process called deepdesulfurization [Song 2003]. In contrast, thecleaning process for hydrocarbon fuels producedfrom biomass feedstocks involves the deepremoval of oxygen, because bio-oils producedthrough pyrolysis and hydrothermal liquefaction ofbiomass are more oxygen-rich than conventionalpetroleum-based feedstocks [Elliott 2007, Czernikand Bridgwater 2004]. (See Thrust 1,Table 1.1.)

2.4 SUMMARY OF PREVIOUS RESEARCH

The technology for producing the biomass-derivedfeedstocks, such as bio-oils and liquefaction oils aredescribed in Thrust 1. However, the liquidproducts generated through selective thermalprocessing are not useful as fuels, other than directboiler firing and possibly for some types of turbineand large diesel applications, and then only aftersignificant modifications. In order for the biomassliquids to be useful as transportation fuels theymust be chemically transformed (i.e. upgraded) toincrease volatility and thermal stability and toreduce viscosity. These changes can beaccomplished through oxygen removal andmolecular weight reduction by methods similar tothose used in the petroleum refinery. A usefulsummary of the early work in this field has beenpublished [Elliott et al. 1991]. Upgrading biomass-derived oils to hydrocarbon fuels can beaccomplished by one of two catalytic methods:hydroprocessing or catalytic cracking.


Bio-oil hydroprocessing refers to the conversion

processing of liquids from biomass using pressurized

hydrogen. Work has been underway, primarily in the

U.S. and Europe, in catalytic hydrotreating and

hydrocracking of biomass-derived liquids (bio-oils) in

both batch-fed and continuous-flow bench-scale

reactor systems. A range of heterogeneous catalyst

materials have been tested, including conventional

sulfided catalysts developed for petroleum

hydroprocessing and precious metal catalysts. The

important processing differences identified require

adjustments to conventional hydroprocessing as

applied to petroleum feedstocks. This application of

hydroprocessing is seen as an extension of petroleum

processing and system requirements are not far

outside the range of conventional hydroprocessing.

BIO-OIL HYDRO P ROCESSING

The technology is still under development, but can

play a significant role in supplementing increasingly

expensive petroleum.

Upgrading fast pyrolysis bio-oils to hydrocarbon fuels

requires oxygen removal and molecular weight

reduction. As a result, there is typically a formation

of an oil-phase product and a separate aqueous

phase product by hydroprocessing. To minimize

hydrogen consumption in hydroprocessing,

hydrodeoxygenation (HDO) must be emphasized,

without saturation of the aromatic rings.

Hydroprocessing biomass-derived oils differs from

processing petroleum or coal liquids because of the

importance of deoxygenation, as opposed to nitrogen

or sulfur removal. At the time that researchers

Gas Product Recycle

FastPyrolizer

Bio-OilCollector

CatalyticReactor

Biomass

H2

Hydrogen Recycle

LightProducts

MediumProducts

HeavyProducts

Char

HeavyPhase

LightPhase

F i g u re S-2.2 Pyrolysis Bio-oils Upgrading A p p ro a c h



55

began to evaluate the HDO of biomass-derived

oils, it had received only limited attention in the

literature [Furimsky 1983]. In the course of more

than 20 years a wide range of efforts has been

reported, as described in a recent literature

review [Furimsky 2000]. A large portion of the

body of work addresses the catalytic chemistry

by hydrotreating model compounds containing

oxygen. Many of these models, such as phenolics

and aromatic ethers, are relevant to bio-oil

hydroprocessing.

Extensive work in the field of hydroprocessing

biomass-derived liquids has been undertaken

during the past 25 years [Elliott 2007]. This

work extends from small-batch reactor tests in

universities to demonstration-scale processing in

petroleum refining laboratories. The processing

potential, measured by these results, is shown to

be relatively more expensive than conventional

petroleum processing at traditional prices, as

shown by the most recent economic studies, now

10 years old [Grange, et al. 1996]. The same

study found that pyrolysis-derived oils are more

economical than other biomass-to-liquid fuel

processing, such as ethanol or biodiesel.

In light of recent inflation in petroleum prices

the increasing potential for biofuels to be cost-

competitive with conventional fuels, should be

carefully evaluated.

Hydroprocessing of biomass-derived oils differsfrom processing of petroleum or coal liquidsbecause of the importance of deoxygenation, asopposed to nitrogen or sulfur removal. At thetime that research began to evaluate thehydrodeoxygenation (HDO) of biomass-derivedoils, HDO had received only limited attention inthe literature [Furimsky 1983]. Over the past 20-plus years, a range of efforts in the field has beenreported in the literature, as described in a recentreview [Furimsky 2000]. A large fraction of thisbody of work addresses the catalytic chemistry ofhydrotreating by using model compoundscontaining oxygen. Many of these modelcompounds are relevant to bio-oilhydroprocessing, such as phenolics and aromaticethers, while others, such as aromatic heterocyclicslike benzofuran or benzoquinone, are not. Inaddition, there are many other oxygenatedfunctional types, important to the composition ofbio-oil, which have received much less attention,such as ketones, organic acids or mixed functional types.

Work is underway, primarily in Europe and a smallamount in the US, in catalytic hydrotreating andhydrocracking of bio-oil in both batch-fed andcontinuous-flow bench-scale reactor systems. Arecent review by Elliot collects and summarizes thedata from experiments involving the handling andupgrading bio-oils by catalytic hydroprocessing[Elliott, 2007]. A range of heterogeneous catalystmaterials have been tested, including conventionalsulfided catalysts developed for petroleumhydroprocessing and precious metal catalysts.Important processing differences were identified,which required adjustments to conventionalhydroprocessing as they would normally be appliedto petroleum feedstocks. This application of bio-oilhydroprocessing is seen as an extension ofpetroleum processing and system requirements arenot far outside the range of conventional


hydroprocessing. The technology is still underdevelopment, but will play a significant role infuture biofuels refining schemes.

A primary concern related to hydroprocessingfast-pyrolysis oil is its thermal instability. Thetendency of bio-oil to further react upon heating,to increase in viscosity, and even form coke,precludes conventional catalytic processingmethods of the kind applied to petroleum. Alower temperature hydrogenation can be used tostabilize the bio-oil and allow its subsequentcatalytic hydroprocessing to be performed at moreconventional temperatures [Elliott and Baker1989]. The subsequent processing step can resultin efficient conversion to high yields of liquidhydrocarbons within the distillation rangeappropriate for transportation fuels.

In parallel developments, several laboratoriesaround the world have investigated catalyticcracking of bio-oils based on the technologycurrently used for petroleum fractions. The use ofacidic catalyst for cracking petroleum hydrocarbonstructures to lower molecular weight material andreforming it into aromatic structures is well-known. This atmospheric pressure process hasbeen found to produce hydrocarbon liquids frombio-oils as well, but with high yields of coke. Ashort summary of the results can be found in arecent review of biomass conversion totransportation fuels [Huber, et al. 2006; Huber andCorma 2007].

The use of zeolitic acid catalysts can maximize theyield of aromatic hydrocarbons in the crackingprocess. The catalyst HZSM-5 has beendemonstrated to be particularly useful in thisregard, producing aromatic yields with up to 5times the aliphatic content. Other zeolitesproduce an aromatics to aliphatics ratio of about

1 to 1. Coke yields typically range from 30- to 50-weight percent. Coke formation by itself is notnecessarily bad because the petroleum catalyticcracker is traditionally operated in a circulatingfluidized bed in which the coke is periodicallyburned off to stimulate catalyst regeneration andprovide heat for the process. However, a goodheat balance has not yet been demonstrated forbio-oil catalytic cracking.

Compared to fast pyrolysis oil, hydrothermalliquefaction oil has lower oxygen content and, dueto differences in chemical composition, it is morethermally stable. These desirable properties arethe result of more severe processing conditionsinvolved with hydrothermal liquefaction (i.e., longerresidence time at high pressure). Subsequenthydrodeoxygenation can be accomplished througha catalytic hydrotreatment at conventionalconditions. This process also results in high yieldsof hydrocarbon products.

If the objective is to produce a finishedhydrocarbon product, eliminating the need for thefirst stabilization step gives an important economicadvantage to hydrothermal liquefaction, comparedwith fast pyrolysis. Nevertheless, detailed chemicalanalysis of a limited number of samples suggeststhat these products would both be useful asfeedstocks in a petroleum refinery setting.

Determination of catalyst stability is a keyshortcoming of the hydroprocessing researchundertaken to date. No long-term processing ofmore than a few days has been reported forhydroprocessing the product from hydrothermalliquefaction. In the longest test (8 days) catalystdeactivation was also identified as a limitation forhydroprocessing fast pyrolysis bio-oil. Catalystdeactivation has also been reported in the catalyticcracking application, such that the useful catalyst



57

lifetime would be limited to only a few cycles.The deactivation of the catalysts is not wellunderstood, but can be attributed to either theeffect of high amounts of water on the oxidestructures, the presence of trace contaminants (e.g. the minerals, derived from the biomassfeedstock), or both of these things.

2 . 5 ECONOMICS AND POTENTIAL OF TECHNOLOGY

The production of bio-oils by fast pyrolysis hasbeen conducted by several industrial groups.However, the economics of producing cleanhydrocarbon fuels from bio-oils remain uncertainsince little work has been done in a manner thatallows for reliable economic evaluation of a large-scale operation. The production cost of biofuelsbased on current technoloies, including feedstockcosts and processing costs, is estimated to be inthe range of $60–120/barrel of oil equivalent[Lange 2007]. This cost will significantly decreaseas new technologies for biomass conversion aredeveloped further.

The end cost of biofuels derived from vegetableoils are currently dictated by feed cost becausevegetable oils already command a premium in themarketplace, whereas the cost of fuels derivedfrom inexpensive lignocellulosic biomass isdominated by technology [Lange 2007]. Figure 2.2on the following page illustrates the economics ofvarious technologies for transportation fuelproduction with processing cost plotted against thefeed cost. The alternatives depicted includeproducing transportation fuels from biomass andpetroleum feedstocks (i.e. from lignocellulose,starch and vegetable oil as well as from crude oiland natural gas), as reported in a study by Shell[Lange 2007].The Shell report illustrates that thecost of crude oil refining is dominated by feed cost

whereas the costs of gas conversion (e.g. MeOH orFischer–Tropsch synthesis) are dominated bytechnology. The report describes the trade-offbetween feedstock cost and plant cost, which isalso applicable to analysis of the options forbiofuels production.

Solid lignocellulosic biomass is inexpensive($2–4/GJ or $34–70/t dry), but very difficult toconvert. Nevertheless, implementation of thestrategies outlined in this roadmap will significantlyreduce the cost of producing fuels from cellulosicbiomass. The low cost of processing crude oils hasbeen achieved thanks to advances in petroleumchemistry and catalysis over the last 80 years. It islikely that similar advances anticipated over thecoming years will also significantly decrease thecost of processing biomass. The factors that affectthe process economics include the cost and typeof the biomass feedstock at the plant gate, theconversion efficiency, the scale of the process, andthe value of the end product (e.g. fuel, electricity or chemicals).


2 . 6 CURRENT T E C H N O L O G YL I M I TATIONS AND RESEARCH/DEVELOPMENT NEEDS

There are no large-scale commercially availabletechnologies for biofuel production frompetroleum refineries. The physical and chemicalproperties of bio-oils are fundamentally differentfrom petroleum feedstocks, and these differencesaccount for the challenges associated withupgrading biomass-derived feedstocks usingpetroleum refining technologies. Bio-oils aregenerally characterized by the following features:

• Bio-oils from fast pyrolysis are acidic, oxygen-rich mixtures of organic components fromthermal decomposition of lignocellulosicbiomass at 450-550 oC under rapid heating-short residence-time conditions.

• Bio-oils produced through liquefactionprocesses are mixtures of organiccomponents from hydrothermal conversionof lignocellulosic biomass at 300-400 oCunder pressurized conditions, which are lessacidic and less oxygen-rich compared topyrolysis oils.

• The compositions and properties of both fastpyrolysis bio-oils and liquefaction bio-oilsdepend, to some degree, on the feedstocktype and the processing conditions.

2 . 6 . 1 . Fe e d s t o c k - re l ated Bio-oil P rocessing Issues

The complex and ill-defined issues related tobiomass conversion pose a huge challenge to theproduction and widespread use of biofuels. A widevariety of feedstocks are used to make the bio-oils(including hardwoods, softwoods, switchgrass, andagricultural residues). Each of the feedstocks

F i g u re 2.2. Feed and processing cost of transport ation fuels derived fro mlignocellulose and fossil re s o u rces (the biofuel plants are set at 400 MWi n t a ke , which corresponds to 680 kt/a of lignocellulose) [Lange 2007].



59

produces different levels of impurities and desiredproducts. Bio-oils are like heavy oils in appearance.The challenges associated with processing themlike heavy oils are their acidity, instability, andpartial volatility. About 50% of the bio-oils arenon-volatile.

Bio-oils may need treatments, such as hydrode-oxygenation, before they can be sent to therefinery for processing. Hydrogen is needed toremove oxygen in hydrotreatment. The limitedavailability and high cost of hydrogen can be alimiting factor. However, hydrogen could be madeby steam reforming of some fractions of bio-oils or biomass.

2 . 6 . 2 . Scientific Issues Related to Bio-oils Conversion

Not surprisingly, processing whole bio-oils is farmore challenging than processing modelcompounds. The major challenge is how to designa catalytic process such that it can selectivelyremove targeted oxygen and minimize hydrogenconsumption. Bio-oil conversion includes waterformation (hydrogen consumption), CO2 and COformation (carbon rejection), and retention of anoxygen fraction in the fuel.

Priorities for future research related to bio-oilconversion include the following important issues:

• Improvement of upstream processes (i.e.pyrolysis and liquefaction), to make theresulting bio-oils more amenable todownstream processing.

• Development of methods to increase theoverall efficiency of the process. Forexample, perhaps process chars can be usedas energy for the pyrolysis and the water-soluble products used for hydrogenproduction via reforming.

• Techniques for effective removal of acidity.Decarboxylation would be desirable, but themost efficacious route (e.g. catalysis, reactionchemistry) is currently unclear. It may bepossible to make use of the acids in bio-oilsby reacting them with alcohols, present in thebio-oils or produced during processing, tomake esters that could become usefulproducts.

• Identification of desirable oxygen compoundsthat improve the fuel properties.

• Development of technologies that producefuels compliant with sulfur specifications,while preserving only the desirable oxygencompounds in the biofuels. There may beadvantages in leaving certain oxygenatedcompounds in the fuel for lubricity. Thereare multiple ways to improve lubricity. Oneway is to use vegetable oil-based biodieseland another way is to use lubricity additives[Knothe and Steidley 2005].The addition of1% ethyl levulinic acid to fuel in Canada wasfound beneficial.The possible benefits ofsome oxygenates in fuels are increasedlubricity and higher cetane number and insome cases the improved reactivity in thesoot [Song, et al. 2006].

• With respect to C-O bond cleavage, thebond energy distributions for bio-oils mustbe determined.The less stable bonds must beidentified. Molecular-level detailedcomposition information will be necessary todetermine bond energy.

Molecular compositions of bio-oils from pyrolysisand liquefaction remain ambiguous. Detailedanalytical characterization work must beundertaken to understand the molecularcomposition and the type of bonds in the bio-oilsand the intermediate products. Mass balance will


also need to be established on different fractionsand on hydrogen consumption during theconversion processing of bio-oils.

2 . 6 . 3 . Te c h n o l ogy Routes for Biofuels from Bio-oils

There is a variety of options for theimplementation of biofuels-to-bio-oils technology.One option is the grassroots refinery, but mis-match in scale may be a major issue. Additionally, itmay not be economically feasible to build a large-scale grassroots refinery exclusively for bio-oils.

Bio-oils may or may not be processed to theexclusion of crude oil in the petroleum refinery,since the high oxygen content, high acidity andinstability of bio-oils present problems.Thus,blending bio-oils from pyrolysis with petroleumrefinery streams may be an option.

Bio-oils may possibly be produced directly throughhydroprocessing. However, this depends on thequality of the feedstock and the hydrogen balance.A common processing challenge is the miscibility ofbio-oils with petroleum streams.

The impacts of bio-oils processing on the totalenergy balance and hydrogen balance in therefinery, and on the consumption of importedpetroleum, is not straightforward, but must beconsidered. How bio-oils are processed depends,to a large degree, on their source. For example,some processing technologies, such asdecarboxylation, can reduce hydrogenconsumption, while hydrodeoxygenation wouldsaturate the feed and increase hydrogenconsumption.

Regulated fuel specifications include sulfur andnitrogen contents, not oxygen contents. Thusconsiderations for oxygen removal would bedifferent than considerations for the otherheteroatoms in biofuels.

2 . 6 . 4 . R e s e a rch Needs for Using Pe t roleum Refining Processes with Bio-oils

HYDROTEATING PROCESSES/FUNDAMENTALS

• Model reactions should be conducted toestablish a baseline for understanding thefundamentals of hydrotreating processes.The reactions of model mixtures should beundertaken to elucidate the reactionnetwork, including catalytic and non-catalyticreactions.The questions related to free-radical reactions and other reactions need tobe understood at a fundamental level.Computational analysis using models that arecurrently available could augment this understanding.

• In order to understand the fundamentals ofbio-oils processing, model compounds shouldinclude the following classes: acids, aldehydes,ketones, phenols, alcohols, anhydrosugars,furans, furfurals, olefins, and ethers.

• The intermediate and final products of modelreactions should be carefully characterized.

• Selected tests could be conducted byblending small amounts of bio-oils into theexisting refinery streams.

• To some extent, bio-oils are similar to heavyoils. So, tar sands processing and the issuesrelated to solid formation from heavy oilsshould be studied and characterized for theirrelevance to bio-oils processing.

• It is worthwhile to test whether bio-oilsshould be blended with refinery heavy oils orcrude oils with some surfactants to ensurethe miscibility. There has been some relatedindustrial testing. However, a publishedreport is not available.



61

• The competition between oxygencompounds with sulfur and nitrogen on thecatalytic surface should be examined.

CRACKING PROCESSESFluid catalytic cracking (FCC) converts bio-oils toliquid fuels without hydrogen consumption.However, catalytic cracking currently gives lowerliquid yield, higher gas yield, and a higher yield of chars.

• The higher water yield, higher char yield andthe trace alkali metals in the 100-1000 ppmrange may lead to serious catalystdeterioration or deactivation, which needs tobe evaluated.

• Compared to hydrotreating, FCC is aprocess that conserves hydrogen. Thissavings is related to the hydrogen and carbonbalance, since there are more oxygen atomsper carbon atom in liquid bio-oils (e.g., 1 Oper 4 C) when compared to liquids fromcoal (about 1 O per 6 or 7 C).

• NREL has conducted fixed-bed catalyticcracking. Catalytic cracking of bio-oils hasbeen reported from studies in Spain and theUniversity of Massachusetts, in theSustainable Energy Center, in the MississippiState University, and in the ChemicalEngineering department at University ofSaskatchewan, Canada. (REFS)

• UOP has compared the different processesin a DOE-supported study, and the reportfrom DOE [Marinangeli et al. 2006].

HYDROCRACKING PROCESSES• Bio-oils contain high molecular weight

(MW) substances, but do not have longhydrocarbon chains as in the petroleum feed.Hydrocracking technologies will have to beadapted to accommodate this difference.

• Hydrocracking has been conducted in someresearch labs, PNNL with metal sulfides and PGM.

• The proper balance in catalytic functionssuch as metal, acid sites, base sites, asreflected in hydrogenation and C-C and C-Obond breaking must be determined.

• Average MW of bio-oils is on the order of800. Components from lignin in the bio-oilsinclude monomers and oligomers ofsubstituted phenols.

FISCHER-TROPSCH RELATED PRIOR WORKIt may be important to consider prior work in thearea of Fischer-Tropsch Synthesis (FTS) since non-hydrocarbon products in FTS have somesimilarities to some biomass-derived oxygen-containing products.There is some similarity to thedeoxygenation of low-temperature FTS liquidswhich contain acids and alcohols.

2 . 6 . 5 . Issues in Development of F u t u re Fuels for Nex t -G e n e r ation Internal Combustion Engines

There is a need to consider fuel utilizationefficiency as we imagine how to incorporatelignocellulosic biofuels into refinery operations.The composition and properties of ideal fuels forfuture engines/vehicles are still unclear. So, thebiomass conversion processing schemes must havethe flexibility to meet uncertain future demandsand criteria.

Variable fractionation of liquid fuels into “green”diesel, jet fuel and gasoline can be envisioned aspart of biorefinery operations. Flexibility of abiorefinery to adjust the proportions of fuel typesproduced is advantageous. One can imagine ascenario in which the demand for diesel fuelgrows, driven by the public’s preference for the


efficiency of diesel engines. This, combined withadvanced combustion engine technology, wouldcompound the benefits of using biofuels, vis-a-visenergy security and sustainability.

2 . 6 . 6 . Some BarriersBio-oils are complex mixtures with moleculardiversity; molecules in biomass feedstocks containcellulose and hemicellulose in 6- and 5-memberedcarbon rings.To some degree, there are similaritiesin bio-oils from different source materials, but withdifferent distributions of compounds.To whatextent the types of molecules and specificcompounds of heteroatoms present in the bio-oilsand intermediate products from bio-oils impact theprocessing scheme and product quality is unclear.We do not currently have a sufficient supply of therange of bio-oils and biofuels needed for wide-spread studies. It is also problematic that we donot have enough bio-oil-derived biofuels for engine testing. Some fuels will have to beproduced at pilot scale to have sufficient samplesfor engine testing.

2 . 6 . 7 . Needs for Collaborat i ve R e s e a rch

We see a need for international collaborationinvolving industrial, academic, and nationallaboratories. Researchers in many countriesoutside the U.S. are actively involved in bio-oilsconversion studies.We see the need for morecollaboration between industrial, academic, andnational laboratories that play different roles inpromoting the development of science andtechnology.

2 . 7 R E C O M M E N DATIONS

A large-scale fundamental research programfocused on the chemistry and catalysis involvedwith upgrading bio-oils to clean liquid hydrocarbon-based transportation fuels using processingschemes that are in principle similar to petroleumrefining technologies should be initiatedimmediately.

We recommend that future research programsaddress the issues and needs discussed in Section2.6 for conversion of bio-oils to hydrocarbon fuelsusing petroleum refining technologies. Some of theunresolved cross-cutting issues related to biofuelsproduction science and engineering are listedbelow:

• The chemical reaction processes ofpetroleum refining that are most suitable forproducing hydrocarbon fuels from bio-oilsmust be identified. In particular, the bestprocesses for producing specifictransportation fuels, including gasoline, jetfuel, and diesel fuel, need to be determined.

• The types of chemistries and catalysts mosteffective for dexoygenation of bio-oils in apetroleum refinery need to be studied.Technologies for efficient removal of oxygenduring the process of bio-oils to hydrocarbonfuels conversion must be developed further

• Selective catalytic oxygen removal from bio-oils should be carefully studied.

• With respect to C-O bond energy and bondcleavage, the bond energy distributions forbio-oils, and identification of which bonds areleast stable, must be determined.

• Work is needed on a detailed analyticalcharacterization in order to understand themolecular composition, bond type, andbonding energy distribution of bio-oils and



63

the intermediate products of conversionprocessing.The type of sulfur and nitrogencompounds need to be identified andquantified.

• Detailed mass balance will need to beestablished on different fractions and onhydrogen consumption during the conversionprocessing of bio-oils. Hydrogen managementand good mass balance, including directmeasurement of hydrogen consumption,should be taken into consideration.

• Technologies for effective acid removal mustbe developed. Decarboxylation would bedesirable, but the most efficacious way (e.g.catalysis, reaction chemistry) to accomplishthis is unclear. If there are any desirableoxygen compounds that improve fuelproperties, they must be identified. Biofuelsupgrading techniques designed to meet fuelsulfur specifications while leaving thedesirable oxygen compounds would bedesirable.

• Mechanistic study of the chemicaltransformation from the bio-oils derivedfrom lignocellulosic biomass to thehydrocarbon-based liquid biofuels is needed.

• A range of catalytic materials need to betested to identify suitable catalysts.Asystematic investigation of catalystdeactivation, including guard bed design andassociated fundamental engineering issues,should be studied. In particular, usedcatalysts at different stages of deactivationneed to be characterized. The impacts thatmetals in bio-oils have on catalystdeactivation need to be evaluated.The typeof metals (alkali at 100 ppm range, sulfurfrom 10 ppm range in feeds from wood to500 ppm range from agricultural resides) inbio-oils is determined, to some degree, on

the type of bio-oils production process. Forexample, bio-oils from pyrolysis are moreacidic and can leach out some metals.

• It is more challenging to process whole bio-oils than to convert single model compounds,like those used in laboratory studies.Therefore, conversion processing using bothmodel compounds and the real bio-oilsshould be conducted in basic studies toensure relevance and applicability of theresearch results.

• Strategic planning for basic research isneeded in order to reach the stage wherefundamental understanding can drive thedevelopment for cost-effective and chemicallyefficient processes for biofuels productionusing petroleum refining technologies.

• Graduate fellowships in the bio-fuels area, ifestablished in government agencies, will be amajor step forward to facilitate the researchand training of talented young researchers atU.S. universities.

• Government-funded research programs andcollaborative programs involving academic,industrial and national laboratories areneeded to bridge fundamental gaps inknowledge in order to pave a path forward.We must elucidate the fundamental reactionprocesses, types of processing steps andoperations necessary to produce bio-oils in a manner that will make their use asbiorefinery feedstocks for transportationfuels cost competitive with petroleum feeds.


2 . 8 R E F E R E N C E S

Czernik, S; Bridgwater,A.V. Overview ofApplications of Biomass Fast Pyrolysis Oil. Energy& Fuels 18 (2): 590-598, 2004

Elliott, D.C. “Historical Developments inHydroprocessing Bio-oils.” Energy & Fuels,21:1792-1815, 2007.

Elliott, D.C.; Baker, E.G.“Process for UpgradingBiomass Pyrolyzates. U.S. Patent 4,795,841, January3, 1989.

Elliott, D.C.; Beckman, D.; Bridgwater,A.V.; Diebold,J.P.; Gevert, S.B.; Solantausta,Y. "Developments inDirect Thermochemical Liquefaction of Biomass:1983-1990." Energy & Fuels, 5(3):399-410, 1991.

Energetics Inc. Energy Bandwidth for PetroleumRefining Processes. Prepared by EnergeticsIncorporated for the US DOE, October 2006.

Furimsky, E. Chemistry of CatalyticHydrodeoxygenation. Catal. Rev.-Sci. Eng., 25(3):421-458, 1983.

Furimsky, E. Catalytic Hydrodeoxygenation,AppliedCatalysis A: General, 199:147-190, 2000.

Grange, P.; Laurent, E.;Maggi, R.; Centeno,A.;Delmon, B. Hydrotreatment of Pyrolysis Oils fromBiomass: Reactivity of the Various Categories ofOxygenated Compounds and Preliminary Techno-Economical Study. Catalysis Today, 29 297-301,1996.

Holmgren, J.“Opportunities for Biorenewables inOil Refineries” presented at the First InternationalBiorefinery Workshop,Washington, DC, July 2005.

Huber, G.W.; Iborra, S.; Corma,A. “Synthesis ofTransportation Fuels from Biomass: Chemistry,Catalysts, and Engineering.” Chem. Rev. 106:4044-4098, 2006.

Huber, G.W.; Corma,A.“Synergies between Bio-and Oil Refineries for the Production of Fuels fromBiomass.” Angew. Chem. Int. Ed. 46, 2007

Johansson, B, Homogeneous charge compressionignition: the future of IC engines? InternationalJournal of Vehicle Design 44 (1-2): 1-19, 2007.

Knothe, G.; Steidley, K.R. Lubricity of Componentsof Biodiesel and Petrodiesel.The Origin of BiodieselLubricity,” Energy & Fuels 19: 1192-1200, 2005.

Lange, J.-P. Lignocellulose conversion: anintroduction to chemistry, process and economics.Biofuels, Bioprod. Bioref. 1:39–48, 2007.

Marinangeli, R.; Marker,T.; Petri, J.; Kalnes,T.; McCall,M.; Mackowiak, D.; Jerosky, B.; Reagan, B.; Nemeth,L.; Krawczyk, M.; Czernik, S.; Elliott, D.; Shonnard, D;"Opportunities for Biorenewables in OilRefineries," Report No. DE-FG36-05GO15085,UOP, 2006.

Marker,T.L., Petri, J., Kalnes,T., Mackiowiak, D.,McCall, M., Czernik, S., Elliott, D.,“Opportunities forBiorenewables in Petroleum Refineries” presentedat the American Chemical Society Symposium, Useof Renewable Fuels in Refining Processes at thenational meeting in Washington, DC,August 28-September 1, 2005.

Metzger J.O. Production of Liquid Hydrocarbonsfrom Biomass.Angew. Chem. Int. Ed. 45: 696-698,2006 2007.



65

Song, C.S. An Overview of New Approaches toDeep Desulfurization for Ultra-Clean Gasoline,Diesel Fuel and Jet Fuel. Catalysis Today, 86: 211-263, 2003.

Song, C.S. Global Challenges and Strategies forControl, Conversion and Utilization of CO2 forSustainable Development Involving Energy,Catalysis,Adsorption and Chemical Processing.Catalysis Today, 115: 2–32, 2006.

Song, J.;Alam, M.; Boehman,A.L.; Kim, U.“Examination of the oxidation behavior of biodieselsoot,” Combustion and Flame 146: 589–604, 2006.


3 . Liquid-phase Catalytic Processing of Biomass-derived Compounds3.0 LIQUID-PHASE CATA LYTIC

P RO C E S S I N G

OVERVIEW: Biomass-derived oxygenates typicallyhave a high-degree of functionality and a low-thermal stability making it difficult to process themin the gas phase where traditional petroleumreactions occur. Catalytic processing in the liquid-phase allows thermally unstable molecules to beselectively converted into a range of fuels andchemicals. Liquid-phase processing has manyadvantages including higher thermal efficiencies,high reaction rates per reactor volume, andexclusion of an energy intensive distillation step.However, new catalyst and reactor systems needto be designed specifically for conversion ofbiomass-derived feedstocks in the liquid-phase inorder to realize the full potential of thistechnology.

3.1 INTRO D U C T I O N

The overarching goal of liquid-phase catalyticprocessing of biomass-derived compounds (e.g.,sugars) is to produce next-generation liquidtransportation fuels that: (i) can be used with theexisting infrastructure; (ii) do not involveenergetically-intense distillation steps; and (iii) havehigh rates of production per reactor volume. Onestrategy for achieving this goal is to convert sugarsto liquid alkanes, such that the liquid fuel derivedfrom biomass is identical, with respect to chemicalcomposition and energy density, to the currentliquid fuels currently derived from petroleum (e.g.,gasoline, diesel, jet fuel). Another strategy is toproduce unconventional types of fuels by preciselyengineering molecules with well-defined amountsof oxygen such that the desired volatility andcombustion properties are optimized withoutadversely affecting the energy density or thehydrophobicity of the fuel.

3.2 PROCESS DESCRIPTION A N DR E S U LTING BIOFUELS

Liquid-phase catalytic processing of biomass-derived compounds offers unique opportunities forachieving high yields of specific, and well-defined,liquid fuels from biomass. For example, liquid-phase catalytic processing of sugars is typicallycarried out at lower temperatures (e.g., 500 K)compared to biomass pyrolysis, liquefaction, orgasification. However, whereas these latter processcan operate with complex biomass feedstocks (e.g.,containing cellulose, hemicellulose, and lignincomponents), liquid-phase catalytic processingtypically involves feedstocks containing specificbiomass-derived compounds, such as sugars orpolyols. Thus, an advantage of liquid-phase catalyticprocessing is that high selectivities and yields totargeted fuel compounds can be achieved, but adisadvantage of such processing is that real biomassfeedstocks must be pretreated to prepare a feedsolution that for subsequent liquid-phase catalyticprocessing.

Because of the high level of functionality (e.g., -OH, -C=O, -COOH groups), biomass feedshave low volatility and high reactivity, and thesefeeds must typically be processed by liquid-phasetechnologies. In addition, in view of theirhydrophilic properties, liquid-phase processing ofcarbohydrate feeds is typically carried out in theaqueous phase, or under biphasic conditionsemploying an aqueous and an organic phase. Ingeneral, a variety of fuels and chemicalintermediates can be produced from these biomassfeeds by employing various types of reactionsincluding: hydrolysis, dehydration, isomerization,C-C coupling (e.g., aldol-condensation), reforming,hydrogenation, oxidation, and hydrogenolysis. Theheterogeneous catalysts used for these reactionscan include acid, base, metal, and metal-oxidecatalysts. Several types of reactions typically occur



3.1 Introduction

3.2 Process Description and ResultingBiofuels

3.2 Feedstocks

3.4 Review of Catalytic Reactions forLiquid Phase Processing

3.5 Advantages of Liquid-PhaseProcessing


3.7 Recommendations

3.8 References

PA RT I C I PA N T S :Scott Auerbach, Paul Blommel,A.A. (Kwesi) Boateng, DouglasCameron, James A. Dumesic, FrankGerry, John Holladay, George W. Huber,Christopher Jones,Alexander Katz,Leo Manzer, Simona Marincean, RaulMiranada,Valeria Reed, John Regalbuto,William D. Rhodes, Bob Saxton,Brent Shanks, Philip H. Steele, James FStevens, Galen Suppes,Thomas HenryVanderspurt,Yang Wang, Rosemarie D.Wesson,Ye Xu, Conrad Zhang

during a given process, allowing the opportunity forthe use of multi-functional catalysts.

The production of ethanol by fermentation ofglucose is also a liquid-phase process that producesa liquid transportation fuel. This process has beenpracticed for many years (e.g., by the brewingindustry), which has led to its wide-scaleimplementation in the production of bioethanol asa transportation fuel. However, fermentation hasseveral disadvantages. One disadvantage is therather high energy costs associated with thedistillation of ethanol from the aqueous solution inwhich it is produced. Another is the low rates ofproduction per volume of reaction vessel. Incontrast to bioethanol, next-generation fuelsderived from liquid-phase processing should havehigher energy densities per volume and propertiesthat are nearer to those of gasoline and diesel fuels.

3.3 FEEDSTO C K S

Biomass may be made amenable for liquid phasecatalytic processing by a variety of means. First,the raw biomass feedstock may be fed directly to a liquid phase catalytic reactor. In this case, usuallyrelatively simple feeds such as vegetable oils, starchor cellulose are utilized. For processing morecomplex, lignocellulosic feeds, the biomass is usuallypretreated to produce a liquid stream that isamenable to catalytic upgrading. Bio-oils and bio-oil components can also be used as feeds forliquid-phase processing. Bio-oil generation by fastpyrolysis or liquefaction processes has beendiscussed in detail in Thrust 1. The most commonbiomass feed sources along with a few of their keyadvantages and disadvantages are listed below:

CELLULOSE can be solvated by aqueous acidic,basic, or ionic liquid mediated processes. Glucosemonosaccharides and degradation products can be

sourced from cellulose for liquid phase upgrading.Hemicellulose can be easily liberated fromlignocellulose by mild aqueous acid treatment,yielding a mixture of C5 and C6 sugars anddegradation products. Pure hemicellulose as astarting feedstock is difficult to obtain.

LIGNIN can be separated from the sugarfractions of the holocellulose portion of biomassby multiple means including kraft pulping,organosolv pulping, or aqueous acid hydrolysis.


Different treatments leave the solid lignin fractionwith different degrees of cross-linking andcondensation. Black liquor is a waste streamproduced via kraft pulping that contains lignin andseveral inorganic chemicals from the pulpingprocess. The aromatic rich stream represents apossible feedstock for further catalytic upgrading.

STARCH and sucrose are common food sourcesbut are also possible feedstocks for fuels. Starch isa glucose polymer that can be treated to yieldglucose and degradation products. Sucrose is asimple disaccharide containing a glucose andfructose unit.

Plant and animal GLYCERIDES are amenable toliquid phase catalytic processing for the synthesis of gasoline and diesel range fuels. Both virgin oilsand waste oils can be feedstocks for suchprocesses. However, the relative amounts of thesefeeds are too limited to make a large impact ontransportation fuels needs in the US.

PROTEINS represent another class of renewableraw materials for liquid phase catalytic processing.However, they do not represent a viable feedstockfor fuels, but maybe useful for chemical production.

3.4 REVIEW OF CATA LY T I CREACTIONS FOR LIQUID-PHASEP RO C E S S I N G

3.4.1 Liquid-phase Processing vs.Gas-phase Pro c e s s i n g

Petroleum feeds usually have a low extent offunctionality (e.g., -OH, -C=O, -COOH groups),making these feeds directly suitable for use as fuelsafter appropriate catalytic processing (e.g., crackingto control molecular weight, isomerization tocontrol octane number). In contrast, functionalgroups must be added to petroleum-derived feedsto produce chemical intermediates, and thechallenge in this field is to be able to add thesegroups selectively (e.g., to add -C=O groupswithout complete oxidation of the organic reactantto CO2 and H2O). Unlike petroleum whichcontains limited functionality, biomass-derivedcarbohydrates contain excess functionality for useas fuels and chemicals, and the challenge in this fieldis to develop methods to control functionality inthe final product.

Figure 3.1 is a qualitative illustration of thetemperature and pressure regimes at whichpetroleum and carbohydrate feedstocks aretypically processed. Petroleum processes areusually conducted at elevated temperatures, andmany of these processes are carried out in thevapor phase. Thermochemical processing ofbiomass-derived feedstocks, such as gasification,liquefaction, pyrolysis, and supercritical treatments,also involve high temperature treatments ofbiomass. As an alternative, biomass-derivedcarbohydrate feeds can also be treated at mildtemperatures in the liquid phase. The potentialadvantage of this approach is to be able to controlthe catalytic chemistry and reaction conditions toachieve high yields of desired products, without theundesirable formation of chars and tars that occurs


3 . Liquid-phase Catalytic Pro c e s s i n gof Biomass-derived Compounds

69

at elevated temperatures. Reactions such ashydrolysis, dehydration, isomerization, oxidation,C-C coupling (e.g., aldol-condensation), andoxidations are often carried out at temperaturesnear or below 400 K. Hydrogenolysis andhydrogenation reactions are usually carried out athigher temperatures (e.g., 470 K), and aqueous-phase reforming is carried out at slightly highertemperatures (e.g., 500 K) and at higher pressuresto maintain the water in the liquid state (> 50 atm).Vapor-phase reforming of oxygenated hydrocarbonscan be carried out over a wide range oftemperatures and at modest pressures (e.g.,10 atm).

3.4.2 The Thermodynamics of Liquid-phase Thermal Pro c e s s i n g

The conversion of biomass-derived oxygenates tofuels and chemicals involves the combination and/orcoupling of various types of reactions, includinghydrolysis, dehydration, reforming, C-Chydrogenolysis, C-O hydrogenolysis, hydrogenation,C-C coupling (e.g., aldol condensation),isomerization, selective oxidation, and water-gasshift. Figure 3.2 presents a schematicrepresentation of the energy changes associatedwith the aforementioned reactions at 300 K and 1atm for selected examples, where exothermicreactions (i.e., negative changes in enthalpy) arerepresented as moving toward the bottom of thefigure and endothermic reactions (i.e., positivechanges in enthalpy) are represented as movingtoward the top of the figure.

Temperature / K

Pre

ssure

/ a

tm

PetrochemicalProcesses

Vapor-phase Reforming

Liquid-phase

Vapor-phase

Super-critical watergasification

GasificationPyrolysis

50

100

150

200

Liquefaction

473 673 873 1073 1273

Aqueous-Phase

ReformingIsomerizationAldol-

CondensationOxidation

HydrogenationHydrogenolysis

DehydrationHydrolysis

Temperature / K

Pre

ssure

/ a

tm

PetrochemicalProcesses

Vapor-phase Reforming

Liquid-phase

Vapor-phase

Super-critical watergasification

GasificationPyrolysis

50

100

150

200

Liquefaction

473 673 873 1073 1273

Aqueous-Phase

ReformingIsomerizationAldol-

CondensationOxidation

HydrogenationHydrogenolysis

DehydrationHydrolysis

F i g u re 3.1 Comparison of reaction conditions for petroleum versus biomass processing

(Reprinted with permission from Chedd a , H u b e r, Dumesic 2007. C o pyright 2007 Wiley ) .


Hydrolysis of a polysaccharide to monosaccharidesis nearly neutral energetically (e.g., equal to -5kcal/mol for the hydrolysis of sucrose to glucoseand fructose), as is the subsequent dehydration ofglucose to form hydroxymethylfurfural (HMF). Incontrast to these steps where water is added orremoved from biomass-derived carbohydrates, thereforming of glucose with water to form CO2 andH2 is a highly endothermic reaction (150 kcal/molof glucose). This formation of CO2 and H2 may beconsidered to be a catalytic decomposition of thesugar to form CO and H2, combined with theconversion of CO and H2O to produce CO2 andH2 (i.e., the water-gas shift reaction, CO + H2OCO2 + H2, which is exothermic by 10 kcal/mol).In this respect, half of the H2 is derived from the

sugar molecule and the other half is derived fromH2O. The change in Gibbs free energy for theoverall reforming reaction is favorable at modesttemperatures and above (e.g., -50 kcal/mol ofglucose at 400 K), because of the large increase inentropy for the reaction.

Hydrogenation reactions are typically exothermic.For example, the enthalpy change for thehydrogenation of glucose to sorbitol is equal toapproximately -10 kcal/mol; the hydrogenation ofHMF to 2,5-di(hydroxymethyl)furan (DHMF) has anenthalpy change of about -20 kcal/mol; and, theenthalpy change for the hydrogenation of the furanring in DHMF to produce 2,5-di(hydroxymethyl)-tetrahydrofuran (DHM-THF) is about -35 kcal/mol

O

OO

nO

HOOH

OH

OH

HO-5 kcal/mol

HO

OH

OH

OH

HO

OH -20 kcal/mol

hydrogenation

+H2

+ H2O

HO

OH

HO

-5 kcal/mol

+ H2

C-C hydrogenolysis

OH

OH + H 2O

C-O hydrogenolysis

-25 kcal/mol

+ H2

dehydration/hydrogenation

O

OOH

- 3 H2 O

-5 kcal/mol

O

OHOH

+H2

-35 kcal/mol

hydrogenation

O

O HOH

+2H2

-10 kcal/mol

hydrogenation

O

OO

oxidation

+ H2O

-50 kcal/mol

+ 1/2 O2

O

OH OH

-10 kcal/mol

+ H2O

aldol condensation+Acetone

O

OH OH

+3H2 -60 kcal/mol

hydrogenation

O

+ 2H2O

+2H2-50 kcal/mol

+2H2

-50 kcal/mol

+ H2O

hydrolysis

dehydration

3CO+4H2

Alkanes

+CO2+H2O

reforming &FT synthesis

6CO2+12H2

ref orming150 kcal/mol

+ 6 H2O

C-O hydrogenolysisdehydration-hydrogenation

C-O hydrogenolysis

dehydration-hydrogenation

a polysaccharide

glucose

HMF

DHMF

DHM-THF

DFF

sorbitolglycerol

BH-HMF

synthes is gas

propanediol

HO

O

HO

glyceraldehyde

HO

Olactic acid

OH

dehydrogenation

15 kcal/mol

isomerization-15 kcal/mol

reforming

80 kcal/mol

-30 kcal/mol

-H2

F i g u re 3.2 Standard enthalpy changes for typical reactions i nvo l ved in biomass pro c e s s i n g .

(Reprinted with permission from Chheda et al 2007. C o pyright 2007 Wiley ) .



71

(or about -18 kcal per mol of H2). Forcomparison, we note that the enthalpy change forhydrogenation of a C=C bond in an olefin is about-25 to -30 kcal/mol. Thus, it is thermodynamicallymore favorable to hydrogenate the C=C bonds inolefins, compared to hydrogenation of C=O bondsor the C=C bonds in the furan ring, and it isthermodynamically more favorable to hydrogenatethese latter bonds than it is to hydrogenate acarbohydrate to its sugar-alcohol (e.g., glucose tosorbitol).

Cleavage of C-C bonds in the presence ofhydrogen is termed C-C hydrogenolysis, and suchreactions are nearly neutral energetically (e.g.,equal to – 5 kcal/mol for the hydrogenolysis ofsorbitol to produce two molecules of glycerol). Incontrast, C-O hydrogenolysis reactions are highlyexothermic. As an example, the enthalpy changefor C-O hydrogenolysis of glycerol to producepropanediol and water is about -25 kcal/mol. It isimportant to note that C-O hydrogenolysis canalso be accomplished by a two-step processinvolving dehydration (catalyzed by an acid or abase) followed by hydrogenation (catalyzed by ametal). Because dehydration reactions areenergetically neutral and hydrogenation reactionsare highly exothermic, the combination ofdehydration followed by hydrogenation is a highlyexothermic process.

Another type of reforming reaction is theproduction of CO:H2 gas mixtures (calledsynthesis gas) from biomass-derived oxygenatedhydrocarbons. This route for production ofsynthesis gas is carried out by minimizing theextent of the water-gas shift reaction, for example,by minimizing the concentration of water in thefeed. As seen in Figure 3.2, the formation ofsynthesis gas from glycerol is highly endothermic,with an enthalpy change of about 80 kcal/mol. Thisformation of synthesis gas is highly beneficial forthe biorefinery, because synthesis gas can be used

as a source for fuels and chemicals by varioussynthesis gas utilization steps, such as Fischer-Tropsch synthesis or methanol synthesis. As anexample, the conversion of synthesis gas to alkanes(along with CO2 and water) is highly exothermic(e.g., -110 kcal/mol), such that the overallconversion of glycerol to alkanes by thecombination of reforming and Fischer-Tropschsynthesis is mildly exothermic, with an enthalpychange of about -30 kcal/mol of glycerol (Soares,Simonetti et al. 2006).

A useful synthetic reaction that can be employedto produce C-C bonds between biomass-derivedmolecules is aldol-condensation. For example,aldol condensation of HMF with acetone leads to aC9 species (1-buten-3-hydroxylhydroxymethylfuran,BH-HMF) (Huber, Chheda et al. 2005).This step is mildly exothermic, with an enthalpychange of about -10 kcal/mol. The subsequenthydrogenation and C-O hydrogenolysis (ordehydration-hydrogenation) steps involved in theconversion of BH-HMF to a C9-alkane are highlyexothermic (with enthalpy changes of about -20 to-25 kcal/mol of H2).

While the production of fuels from biomass-derived carbohydrates involves reduction reactions,the production of chemical intermediates mayinvolve oxidation reactions, such as the conversionof alcohols to aldehydes and carboxylic acids.These oxidation reactions are highly exothermic, asillustrated in Figure 3.2 by the oxidation of HMF toform 2,5-diformylfuran (DFF), with an enthalpychange of -50 kcal/mol.


Another class of important reactions for biomassconversion involves isomerization processes. Forexample, the conversion of glucose to fructose isof importance for the production of HMF, becausehigher rates and selectivities for fructosedehydration to HMF are achieved compared to thecase for glucose dehydration. The conversionbetween these two sugars is nearly neutralenergetically. In contrast, the isomerization of ahydroxy-aldehyde (such as a sugar) to a carboxylicacid is highly exothermic. As shown in Figure 3.2,the dehydrogenation of glycerol to glyceraldehydeis highly endothermic with an enthalpy change ofabout 15 kcal/mol, and the isomerization ofglyceraldehyde to lactic acid is exothermic with anenthalpy change of -15 kcal/mol, such that theoverall conversion of glycerol to lactic acid isabout neutral energetically.

3.4.3 Reaction Classes for Catalytic C o nversion of Carbohy d r at e - d e r i ved Fe e d s t o c k s

The main types of reactions involved in theconversion of biomass-derived feeds to fuels andchemicals are: hydrolysis, isomerization, reforming,C-C coupling (e.g., aldol-condensation),hydrogenation, selective oxidation, hydrogenolysis,dehydration/hydrogenation and olefinoligomerization and metathesis.

3.4.3.1 Hydro l y s i sHydrolysis is one of the major processingreactions of polysaccharides in which the glycosidicbonds between the sugar units are cleaved to formsimple sugars like glucose, fructose and xylose, andpartially hydrolyzed dimer, trimers, and otheroligomers.The challenge is to identify the reactionconditions and catalysts to convert a diverse set ofpolysaccharides (such as cellulose, hemicellulose,starch, inulin, and xylan) obtained from a variety ofbiomass sources.

Hydrolysis reactions are typically carried out usingacid or base catalysts at temperatures ranging from370 – 570 K, depending on the structure andnature of the polysaccharides. Acid hydrolysis ismore commonly practiced because base hydrolysisleads to more side reactions and thus lower yields.Acid hydrolysis proceeds by C-O-C bond cleavageat the intermediate oxygen atom between twosugar molecules. Often the reaction conditions canlead to further degradation of sugars to productssuch as furfural and HMF that may be undesirable.Cellulose, the most abundant polysaccharide with-glycosidic linkages, is the most difficult material

to hydrolyze because of its high crystallinity. Bothmineral acids and enzymatic catalysts can be usedfor cellulose hydrolysis, with enzymatic catalystsbeing more selective. The highest glucose yieldsachieved for cellulose hydrolysis with concentratedmineral acids are typically less than 70%, whereasenzymatic hydrolysis of cellulose can produceglucose in yields close to 100% (Huber et al. 2006).Hemicelluose is more open to attack atintermediate positions to break down theoligomers to single sugar molecules, therebyrequiring modest temperatures and dilute acidconcentrations, which minimize further degradationof the simple sugars. Soluble starch (a polyglucanwith -glycosidic linkages obtained from corn andrice) and inulin (a polyfructan obtained fromchicory) can be hydrolyzed at modest conditions(340 – 420 K) to form glucose and fructose,respectively (Moreau et al. 1997; Nagamori andFunazukuri 2004).



73

3.4.3.2 Dehy d r at i o nDehydration reactions of carbohydrates andcarbohydrate-derived molecules comprise animportant class of reactions in the field of sugarchemistry. Sugars can be dehydrated to form furancompounds such as furfural and HMF that cansubsequently be converted to diesel fuel additives(by aldol-condensation and aqueous phasedehydration–hydrogenation) (Huber et al. 2005)industrial solvents (e.g., furan, tetrahydrofurfurylalcohol, furfuryl alcohol) (Lichtenthaler and Peters2004) various bio-derived polymers (by conversionof HMF to FDCA) (Werpy and Petersen 2004) andP-series fuel (by subsequent hydrogenolysis offurfural) (Paul 2001). Furfural is industriallyproduced from biomass rich in pentosan (e.g., oathulls, etc) using the Quaker Oats technologyemploying mineral acid as catalyst (Zeitsch 2000).

However, HMF is not yet a high-volume chemical inview of difficulties in cost-effective production, eventhough many researchers have shown promisingresults in a wide-range of potential applications(Kuster 1990; Gandini and Belgacem 1997;Lewkowski 2001; Moreau, Belgacem et al. 2004).Production of HMF from sugars is a problem thatillustrates the selectivity challenges involved in theprocessing of highly functionalized carbohydratemolecules. Dehydration of hexoses has beenstudied in water, organic solvents, biphasic systems,ionic liquids and near- or super-critical water, usinga variety of catalysts such as mineral and organicacids, organocatalysts, salts and solid acid catalystssuch as ion-exchange resins (Mercadier et al.1981) and zeolites (Moreau et al. 1996) in thetemperature range from 370 – 470K.

Although evidence exists supporting both theopen-chain and the cyclic fructofuransylintermediate pathways, it is clear that the reactionintermediates and the HMF product degrade viavarious processes (Antal Jr. et al. 1990; Antal Jr. et

al. 1990; Moreau et al. 1996; Qian et al. 2005).Similarly, glycerol can be dehydrated to acrolein animportant industrial chemical.Acrolein can also beconverted to other chemicals including: acrylicacid, used in super adsorbents, and potentially 1,3-propanediol, used in Sorona™ polyester. Ott, et al.have shown promising results using sub- and super-critical water with zinc sulfate salts as catalysts toachieve acrolein yields up to 80%. However,corrosion induced by water and salt at theseconditions necessitates use of expensive corrosion-resistant materials for the reaction (Ott et al.2006).

3.4.3.3 Isomerizat i o nIsomerization of carbohydrates is typically carriedout in the presence of base catalysts at mildtemperatures and in different solvents. Glucoseconversion to fructose is widely practiced forproduction of high fructose corn syrup. Inaddition, HMF selectivity from glucose can beimproved by isomerization of glucose to fructose.

Isomerization can be carried out in the presence ofa base catalyst such as magnesium-aluminumhydrotalcites (Lecomte et al. 2002) at temperaturesranging from 310 to 350 K. Carbohydrates insolution are present in open chain (acyclic) and inring structures such as -furanose, -furanose, -pyranose, -pyranose in varying proportions(Collins and Ferrier 1995).

The isomerization reaction involves formation ofintermediate enolate species through open chainforms to transform aldo-hexoses to keto-hexoses.The rate of glucose isomerization is thus dictatedby the fraction of the glucose molecules that are inthe open chain form, which is governed by thesolvent medium and temperature.Thus, thereaction rates are higher in aprotic solvents, suchas dimethylsulfoxide (DMSO), in which the


Liquid alkanes or green gasoline can be produced

directly from glycerol by an integrated process

involving catalytic conversion to H2/CO gas mixtures

(synthesis gas) combined with Fischer-Tropsch

synthesis (Figure S-3.3). Concentrated solutions of

glycerol (e.g., 80 wt%) in water are first passed at

low temperatures (548 K) and moderate pressures

(1-17 bar) over a catalyst consisting of PtRe

nanoparticles supported on carbon, and the products

from this catalyst are then contacted at the same

temperature and pressure with nanoparticles of Ru

supported on titania, leading to the formation of

liquid alkanes.

GREEN GASOLINE PRODUCTION BY INTEGRATED SYN-GAS PRO D U C T I O NAND FISCHER-TROPSCH SYNTHESIS

This integrated process has the potential to improve

the economics of “green gasoline” by reducing capital

costs and increasing thermal efficiency. Importantly,

the coupling of glycerol conversion with synthesis

gas and Fischer-Tropsch synthesis leads to synergies

in the operations of these processes, such as (i)

avoiding the highly endothermic and exothermic

steps that would result from the separate operation

of these processes and (ii) eliminating the need to

condense water and oxygenated hydrocarbon

byproducts between the catalyst beds.

F i g u re S-3.3 Integ r ated catalytic process for production of liquid alkanes (green gasoline).

(Reprinted with permission from Soare s , S i m o n e t t i , and Dumesic 2007. C o pyright 2007 Wiley ) .



75

abundance of the acyclic form is about 3% forfructose as compared to in water where it is lessthan 0.8%. In addition, an increase in temperatureto 350 K increases the open chain form, therebyincreasing the rate of isomerization (Dais 1987;Franks 1987; Bicker et al. 2005).

3.4.3.4 Reforming ReactionsThe production of hydrogen for fuel cells, ammoniasynthesis and other industrial operations is anessential element of future biorefineries, similar tocurrent petroleum refineries. Pyrolysis of solidbiomass followed by reforming of bio-oils andbiomass gasification are known technologies for H2production (Huber, Iborra et al. 2006). In addition,it has recently been shown that aqueous phasereforming (APR) can be used to convert sugars andsugar alcohols with water to produce H2 and CO2at temperature near 500 K over metal catalysts(Cortright et al. 2002). Importantly, the selectivitytowards H2 can be controlled by altering thenature of the catalytically active metal sites (e.g., Pt)and metal-alloy (e.g. Ni-Sn) (Huber et al. 2003)components, and by choice of catalyst support(Shabaker et al. 2003).

Various competing pathways are involved in thereforming process. A good catalyst should promoteC-C bond cleavage and water-gas shift to convertCO to CO2, but it should not facilitate furtherhydrogenation reactions of CO and CO2 to formalkanes or parallel reactions by C-O bond cleavageto form alcohols and acids (Davda et al. 2005). Ithas also been demonstrated that APR can betailored to convert sorbitol to a clean stream oflight alkanes (C4-C6) using a bi-functional metal-acid catalyst (e.g., Pt/SiO2-Al2O3), whereinformation of hydrogen and CO2 takes place on ametal catalyst and dehydration of sorbitol occurson a solid acid catalyst (Huber et al. 2004). Thecombination of catalytically active sites, support,solution pH, feed concentration, process conditionsand reactor design governs the selectivity of

hydrogen and alkane production using aqueousphase processing. It has recently been shown thatthe APR process can be used to produce H2 fromactual biomass; however, low yields (1.05 -1.41mmol/g of carbohydrates) were obtained due tocoke and byproduct formation (Valenzuela et al.2006).

3.4.3.5 Carbon-Carbon Bond Coupling Reactions

To convert carbohydrate derived molecules intoliquid range alkanes that fit within the diesel andjet fuel range C-C bonds must be formed. Aldolcondensation is a C-C bond forming reaction,generally carried out to form larger molecules atmild temperatures (300 – 370 K) in the presenceof a base or acid catalyst. It has been shown thatvarious carbohydrate-derived carbonyl compoundssuch as furfural, HMF, dihydroxyacetone, acetone,and tetrahydrofurfural can be condensed inaqueous and organic solvents to form largermolecules (C7-C15) that can subsequently beconverted to diesel fuel components (Huber et al.2005). Aldol condensation requires at least onecarbonyl compound having an -hydrogen atom,and the reaction is generally carried out in thepresence of a base catalyst. At first, the basecatalyst abstracts the -hydrogen from thecarbonyl compound to form an intermediatecarbanion (enolate ion) species, which can thenattack the carbon atom of a carbonyl group fromanother molecule, that may or may not have an

-hydrogen atom, to form a C-C bond.The aldol-adduct can further undergo dehydration toform an unsaturated aldehyde or ketone. Factorssuch as reaction temperature, solvent, reactantmolar ratio, structure of reactant molecules, andthe nature of the catalyst determine the selectivityof the process towards heavier compounds(Barrett et al. 2006).


3.4.3.6 Hydrog e n at i o nBiomass-derived oxygenates are intrinsicallyhydrogen deficient molecules. Hydrogenation isone of the key reactions for conversion of biomass-derived oxygenates to liquid alkanes. Aqueous-phase hydrogenation reactions are used forhydrogenation of fermentation products (lactic acid,acetic acid), production of polyols (sorbitol, etc),upgrading of bio-oils, and liquid alkane productionby aqueous-phase dehydration/hydrogenation. Inthese processes a number of functionalities arehydrogenated including acids, aldehydes, C=Cbonds, and C-O-C (ester) linkages. Hydrogenationreactions are carried out in presence of a metalcatalyst such as Pd, Pt, Ni, or Ru at moderatetemperatures (370 – 420 K) and moderatepressures (10 - 30 bar) to saturate C=C, C=O andC-O-C bonds. Selective hydrogenation reactionsare important to make biofuels and products.Electrocatalytic hydrogenations offers the means toreduce unsaturated bonds by hydrogen produced atthe surface of an electrode made of catalytic metal.In-situ production of hydrogen eliminates the needof high pressures to accumulate enough H2 insolution, splitting of H2 on catalyst surface and itsmass transport issues in water. Lactic acidreduction using a reticulated vitreous carbonelectrode suffused with a Ru/C catalyst, wasperformed at 70°C in aqueous environment inpresence of an electrolyte to selectively producelactaldehyde in an 80% yield. These conditions aremilder than the ones used for batch reactionswhich typically run at 130°C and 1000 psi H2.Also one must note the selectivity of theelectrochemical process that stops at the first stepin terms of reduction, lactaldehyde. The challengesassociated with this technique at present are: diluteconditions needed due to mixing issues, stability ofthe reactants in presence of the electrolyte,reaction rate due to surface limits (Dalavoy et al.2007).

More generally, C-C coupling reactions typicallyrequire an activated site next to the carbonylfunctionality or an unsaturated C-C bond whichwould be either conjugated with the carbonylfunctionality or isolated.The main challenges withC-C bond forming reactions are related to: thereagents needed for this chemistry, how tointroduce selectively unsaturated C-C bonds, howmuch of the oxygen content can be removed andhow to do that in a selective fashion. Anotheraspect is that the above mentioned chemistry hasmainly been studied in organic solvents usinghomogeneous catalysis. Several questions arisewhen these reactions are applied to biofuel relatedchemistry. How will this chemistry apply toaqueous systems? Will the reagents be soluble inorganic solvents? What will their structure be inthose solvents? (For example, in water the carbonylfunctionality is typically hydrated to a certainextent, and thus may hinder aldol chemistry.) Willthe reagents be stable in the conditions requiredfor the transformation to take place?

Other molecular weight enhancing reactionsshould also be explored. Linking together smallermolecules via C-O bonds also appears to be apromising route. Esterification and etherificationreactions appear to be useful approaches, althoughthese methods necessarily introduce (or maintain)oxygen in the end products. This may or may notbe feasible for some intended applications. Inaddition, C-C bonds could potentially be formed byalkylation reactions (aromatic and aliphatic) overacid catalysts. Another route to C-C bondformation would be first to produce olefins,followed by oligomerization. Alternatively, alcoholscould be produced from biomass-derivedcompounds, followed by the formation of highermolecular weight species by MTG (methanol togasoline) process variants.



77

3.4.3.7 Selective Oxidat i o nSelective oxidation is conducted to form chemicalintermediates having specific functionality, and thisreaction is carried out in the presence of aqueousor organic solvents at temperatures from 330 to420 K and oxygen pressures of 2 to 10 bar in thepresence of supported metals (Pt, Pd,Au,Ti, Zr,V)or metal oxides and metal derivatives such asvanadyl phosphate (Carlini, et al. 2005). Catalyticoxidation reactions can form multiple products, andthus the challenge is to direct the reactionpathways to desired products. Selective oxidationof HMF leads to formation of 2,5-diformylfuran(DFF) that has potential applications in thesynthesis of drugs, fungicides and in preparing newpolymeric materials (Halliday, Jr. et al. 2003). Theproduct distribution for this reaction depends onthe type of solvent, pH, partial pressure of oxygen,temperature and nature of the catalyst. Hightemperatures and almost neutral pH in thepresence of a Pt/C catalyst lead to oxidation of thehydroxymethyl group to give DFF, while lowtemperatures and basic pH lead to oxidation ofboth the formyl and the hydroxyl groups of HMFto form 2,5-furandicarboxylic acid (FDCA)(Carliniet al. 2005). Similarly, acidic conditions in glyceroloxidation favor oxidation of the secondaryalcoholic group to dihydroxyacetone (DHA), whileunder basic conditions the primary alcoholicgroups are oxidized to form glyceric acid. Additionof Bi promotes the Pt/C catalyst in the presence ofbase to a mechanism involving oxidation of thesecondary alcoholic group to dihydroxyacetone(Bianchi et al. 2005). In a recent study of glyceroloxidation, researchers found that bimetalliccatalysts (Au-Pt,Au-Pd) are more active thanmonometallic catalysts (Au, Pt, Pd), indicating asynergistic effect existing between Au and Pd or Pt.(Bianchi et al. 2005). A recent study has shownthat glycerol can be converted to dihydroxyacetone(DHA) by an electrochemical route (Ciriminna,2006 #127).

3.4.3.8 Hydrog e n o l y s i sThe hydrogenolysis of C-C and C-O bonds inpolyols occurs in the presence of hydrogen (14 –300 bar), at temperatures from 400 to 500 K,usually under basic conditions, and with supportedmetal catalysts including Ru, Pd, Pt, Ni, and Cu(Zartman and Adkins 1933;Tronconi, Ferlazzo et al.1992; Lahr and Shanks 2003; Chaminand, Djakovitchet al. 2004; Dasari et al. 2005; Lahr and Shanks2005; Saxena et al. 2005; Miyazawa et al. 2006). Theobjective of hydrogenolysis is to selectively breaktargeted C-C and/or C-O bonds, therebyproducing more valuable polyols and/or diols.These lower polyols such as ethylene glycol (EG),1,2-propanediol (1,2 PDO) and 1,3-propanediol(1,3 PDO) have potential applications in thepolymer industry (Chaminand, et al. 2004).

The hydrogenolysis of glycerol has received recentattention, (Lahr and Shanks 2003; Chaminand et al.2004; Dasari et al. 2005; Lahr and Shanks 2005;Miyazawai et al. 2006) because the cost of glycerolas a byproduct is projected to decrease significantlyas biodiesel production increases (McCoy 2005).

Glycerol can undergo dehydration reactions toform acetol or 3-hydroxypropionaldehyde, whichare then hydrogenated on the metal catalyst to 1,2propanediol and 1,3 propanediol, respectively. Ithas been proposed that OH-species on Ru catalyzethe dehydration reaction to produce 3-hydroxypropionaldehyde, whereas the productionof acetol occurs on amberlyst sites.(Miyazawa,Kusunoki et al. 2006) The reactivity of 1,3propanediol is high, and it undergoes C-O or C-Cbond cleavage, where C-O bond cleavage occursthrough a dehydration/hydrogenation pathway.Alternatively, glycerol can undergo C-C bondcleavage to produce ethylene glycol and methanol.


In contrast to hydrogenolysis of glycerol, C-C bondcleavage is a desirable reaction for hydrogenolysisof large polyols (such as sorbitol). The addition ofa base catalyst (e.g., NaOH) increases the rate ofC-C hydrogenolysis. Wang, et al. have proposedthat carbon-carbon bond cleavage occurs by retro-aldol condensation, and they have studiedhydrogenolysis of 1,3 diols with Raney Ni and Cucatalysts (Wang, Hawley et al. 1995). They proposethat the first step in C-C bond cleavage isdehydrogenation, followed by retro-aldolcondensation. The products from retro-aldolcondensation are then hydrogenated. The forwardaldol condensation can also occur under theseconditions.

3.4.3.9 Aqueous-Phase Dehy d r at i o n /H y d rog e n at i o n

The conversion of biomass-derived molecules intoliquid alkanes involves removing the oxygenfunctionality from the feedstock. This can be doneby aqueous-phase dehydration/hydrogenation(APD/H) which involve bifunctional catalysiscontaining both metal and acid sites. One exampleof APD/H is alkane production directly from sugaralcohols by APD/H (Equation 1) with a catalystcontaining metal (e.g., Pt or Pd) and acid (e.g.,SiO2-Al2O3) sites to catalyze dehydration andhydrogenation reactions, respectively. Equation 1requires the addition of hydrogen, which can beproduced from sorbitol by aqueous-phasereforming (APR) (Equation 2). The net reaction(Equation 3) is exothermic, in which approximately1.5 moles of sorbitol produce 1 mole of hexane.

(1)

(2)

(3)

The essential features of the bi-functional reactionscheme for production of alkanes from sorbitolinvolve: (1) hydrogen production on metal catalyticsites by cleavage of C-C bonds followed by thewater-gas shift reaction, (2) dehydration on acidsites, and (3) hydrogenation of the dehydratedspecies on metal sites. Repeated cycling ofdehydration and hydrogenation reactions in thepresence of H2 leads to heavier alkanes (such ashexane) from sorbitol. Formation of lighteralkanes takes place by cleavage of C-C bondscompared to hydrogenation of dehydrated reactionintermediates. One method to produce morevaluable compounds is to combine the APD/Hprocess with a base catalyzed aldol condensationstep to produce larger alkanes ranging from C7 toC15 as recently reported by Huber et al. (Huber etal. 2005). Optimization of the APD/H systemsinvolves choosing the proper: (1) metal catalyst, (2)acid catalysts, (3) ratio of metal to acid sites, (4)reaction conditions and (5) proper reactor design.This example illustrates how bifunctional catalysts can be used to selectively producetargeted products.

OHHCHHOC 214621466 66 ++

2221466 1366 HCOOHHOC ++

OHCOHCHOC 22146146613

42

13

36

13

19++

Aqueous-phase processing can be used to selectively

produce liquid alkanes from carbohydrate-derived

compounds that can be used as “green” diesel or jet

fuel (Huber et al., 2006). This is a multi-step

process that requires acid catalysts for dehydration

reactions, base catalysts for aldol-condensation,

metal catalysts for hydrogenation reactions, and

bifunctional metal/acid catalysts for dehydration/

hydrogenation reactions (Figure S-3.1)

The first step in this process is biomass

deconstruction to furfural and HMF. These

compounds are then reacted with acetone over

basecatalysts to form C-C bonds. The final step in

this process involves converted these large biomass-

derived oxygenates into liquid alkanes by

dehydration/hydrogenation. This requires a four-

phase reactor involving: (1) a gas phase containing

H2; (2) a solid phase composed of a bifunctional

heterogeneous catalyst containing metal and acid

sites; (3) an aqueous phase containing the sugar

reactant; and (4) a liquid alkane phase used to

remove hydrophobic species from the catalyst surface

before then react further to form carbonaceous

deposits that deactivate the solid catalyst. The

alkanes are easily separated from the aqueous feed,

significantly improving the process thermal efficiency.

This example illustrates how chemistry and catalysts

can be used to selectively make targeted fuels from

biomass-derived feedstocks.

GREEN DIESEL AND JET FUEL RANGE ALKANES BY AQUEOUS-PHASE PRO C E S S I N G

O

OO

O OO

O

OH

CH2

OHOHHOOH

HH

HH

H

O

H

HO

H

HO

H

HOHH

OH

OH

O

OOH

O

OH

O

OH

O

OO

OH

O

OO O

OHO

D-Glucose

-pyrano se

n

polyfr uctan

Hydrolysi s

( acid)

StarchPol ygluc an

In ulin

OHOHOHOH

OOH

C12-alkane

OHOH

C9-a lka ne

3H2O

2H2

4H2

7H 2

4H2

8H2

O

HMFHMF

HMTHFA HMTHFA6 H2O

3 H2 O

5 H2O7H2

Dehydrat ion(ac id )

Aldol c rosse d -condensa ti on

(base )Aldo l crossed-conde ns ati on

(bas e)

Hydroge na ti on(me tal )

Hydrogenat ion(m eta l)

Hydrogenat ion(metal )

Se lec tiveHydroge na tion

(me tal )

Aldo l sel f-conde ns ation

(bas e)

Dehydrat ion/Hydroge na tion(metal -a ci d)

De hydra tion/Hydrogenat ion(me tal -acid )

Dehydrat ion /Hydroge na tion(metal -a cid)

H2

H2O

H 2O

C15-alkane

D -F ructo se

-pyranose

Biomass-derived Carbohydrates

O

OH

O

O

OH OH

O

OH

OH

F i g u re S-3.1 Reaction pat h ways for conversion of polysaccharides into liquid alkanes by aqueous-phase pro c e s s i n g . A n a l ogous chemistry can be depicted for

c o nversion of C5 polysaccharides to C10, C8 and C13 alkanes, re s p e c t i ve l y, via furfural asreaction intermediat e . (Reprinted with permission from Chedd a , H u b e r, Dumesic 2007.

C o pyright 2007 Wiley ) .



79


3.5 A DVA N TAGES OF LIQUID-PHASEP RO C E S S I N G

3.5.1 The Potential to Design N ew Generations of N a n o s t ru c t u red Catalysts

Heterogeneous catalysis has largely beendeveloped in concert with the conversion of fossilfuel feedstocks, which are predominantlyprocessed at elevated temperatures. Thisprocessing paradigm has placed constraints on thecatalyst systems to be used. In particular,supported metal catalysts or metal oxide catalystsare almost exclusively used. However, liquid-phaseprocessing, which will generally occur attemperatures below 500 K, opens the potential forcatalytic moieties beyond metals and metal oxides.A number of the reactions to be performed onbiomass-derived molecules will exploit acid/basecatalysis. In addition to traditional catalyticmaterials, organic acids and bases can be used ascatalysts. These species can be immobilized onpolymer or metal oxide supports for use asheterogeneous catalysts. Use of these organicacids and bases will allow the synthesis of materialswith well-defined catalytic sites. Additionally, theproperties of these catalytic moieties can besystematically adjusted. For example, organic acidswith a range of pKa values can be explored (e.g.carboxylic, phosphonic, sulfonic) or the pKa valuefor say sulfonic acid can be modified by changingthe chemistry of the tethering group (e.g., alkyl,arene, perfluorinated).

Compared to petroleum feeds, the carbohydratesin biomass-derived feeds are highly reactive andcan be processed at mild temperatures (e.g., < 500K). Accordingly, the catalysts employed for theseprocesses need not be subjected to hightemperatures, potentially allowing a greater degreeof flexibility in the design and synthesis of thesematerials. In fact, this ability to operate at

relatively low temperatures offers uniqueopportunities for applications of nano-synthetictechniques. In particular, recent advances in nano-technology have provided the concepts and toolsrequired to synthesize new materials withunprecedented control at the nanometer lengthscale. However, while many of these materials maynot be stable under the harsh conditions requiredfor the processing of petrochemical feeds, it islikely that sophisticated materials prepared bynano-synthetic techniques may be moreappropriate for low-temperature catalytic reactionsemployed in the processing of biomass-derivedcarbohydrates.

3.5.2 Ability to Process Thermally U n s t a ble Molecules

A key advantage of liquid-phase catalytic processingis that thermally unstable reactant molecules, suchas sugars, can be processed at low temperatures,where undesirable thermal degradation reactionsare slow, leading potentially to high selectivities fortargeted products. Clearly, a variation on thistheme is that liquid-phase catalytic processing canbe used to convert reactant molecules having lowvolatility, such as sugars. In particular, by processingthese molecules in the liquid phase we eliminatethe need to employ high temperatures toevaporate these molecules into the vapor phase,and operation at lower temperature minimizes thecontributions from degradation reactions, leadingagain to high selectivities for targeted products.This effective processing of non-volatile reactantsalso applies to the processing of feeds containingnon-volatile impurities and contaminants (e.g., ash).For example, the presence of ash in a biomass feedwould deposit on and accumulate within the poresof a solid catalyst for a vapor-phase reaction,whereas it is possible that the ash would be sweptfrom the catalyst by the flowing solvent for aliquid-phase catalytic reaction.



81

3.5.3 Reaction Control through S o l vent Selection

A key attribute in liquid-phase processing is thechoice of solvent, and this choice is dictated byconsideration of how the solvent affects thesolubilities and reactivities of the reactants,products and intermediates of the liquid-phaseprocess, as well as upstream and downstreamprocesses that interface with the liquid-phaseprocess. For example, the upstream reactant forthe overall process may be a highly polar,hydrophilic molecule such as a sugar or a polyol,and a logical solvent for this molecule would bewater. In the overall processing of this molecule toa liquid fuel, oxygen atoms will be removed leadingto a less polar, hydrophobic molecule. Accordingly,this change in physical properties leads toopportunities for biphasic liquid-phase processing,wherein the hydrophobic product is transferred,without vaporization, from the aqueous phase toan organic solvent phase. In this respect, thesolvent properties are chosen to optimize thepartition coefficients of the reactants, products andintermediates between the aqueous and theorganic phase. Moreover, the interactions of thesolvent with liquid-phase species, with the catalystsurface, and importantly with the transition statesfor the rate-determining steps control the reactionactivity and selectivity.

The choice of the solvent for liquid-phase catalyticprocessing can be of critical importance, and thereis much that can be learned from the pulp andpaper industry, such as is involved in Organosolvand kraft pre-treatment processes. Also, the choiceof the solvent will be influenced by availability atthe biorefinery location, and solvents of particularmerit in this respect would be butanol, ethanol,biodiesel, ethyl acetate, acetone, and gasoline itself.It is noteworthy that ionic liquids have recentlybeen shown to have unique properties for theselective conversion of biomass-derived

compounds, such as the conversion of hexoses toHMF. Moreover, these ionic liquid solvents havevery low vapor pressures at elevated temperatures,offering effective strategies for separating volatileproducts from the reactive solution by distillation.In addition, the low volatilities of these solventsmake it possible to use vacuum techniques (e.g.,XPS,TEM) to characterize these reactive systemsunder reaction conditions. (see sidebar on next page)

3.5.4 Biphasic ReactionsThe possibility of conducting reactions havingmultiple phases offers challenges in reactionengineering, but it also opens new avenues fordeveloping new catalytic processes that takeadvantage of synergies between reaction kineticsand thermodynamics. For example, if a desiredreaction intermediate or product undergoes side-reactions with a reactant, then the production ofthis species can be optimized by conducting thereaction in a bi-phasic reactor containing thereactants in one liquid phase plus a second liquidphase that preferentially extracts the desiredintermediate or product. In a variation on thistheme, a four-phase reactor can be used to carryout a hydrodeoxygenation conversion of a sugar toproduce an alkane, involving (1) a gas phasecontaining H2, (2) a solid phase composed of abifunctional heterogeneous catalyst containingmetal and acid sites, (3) an aqueous phasecontaining the sugar reactant, and (4) a liquidalkane phase used to remove hydrophobic speciesfrom the catalyst surface before they react furtherto form carbonaceous deposits that deactivate thesolid catalyst. If the feed stream also contains solidbiomass components, then the aforementionedreactor becomes a five-phase system. The concept of a bi-phasic reactor can also be used tocouple an aqueous-phase reaction leading todeconstruction of polysaccharides to form reactionintermediates (e.g., H2, sugars, polyols, furans),combined with an organic phase in which these


Ionic liquids are often defined as salts that have a

melting point below 100°C; many are liquids at

normal ambient temperatures. The properties of

these salts present a wide range of opportunities for

their use as solvents and as catalysts.

Most ionic liquids exhibit practically no measurable

vapor pressure and are highly polar, and yet non-

coordinating, which is ideal as solvent for catalytic

reactions. Some ionic liquids are immiscible with

water; others are immiscible with organic solvents.

Therefore, many ionic liquids are suitable for

catalytic processes. Their application can improve

reaction rates and selectivity, and importantly

reduce the cost of product separation. Using ionic

liquids to replace water and organic compounds as

solvents offer potential advantages, such as reduced

waste disposal and improved process economics,

provided that ionic liquids are reused. The non-

volatile nature, thermal stability, and higher density

LIGNOCELLULOSIC BIOMASS DECONSTRUCTION TO FUEL PRECURSORS WITH IONIC LIQUIDS

of most ionic liquids compared with water and

organic solvents make their reuse readily achievable

in most processes.

For biomass processing, a highly attractive property

of some ionic liquids is their capability to swell and

disperse the polymeric structure of biomass

components, creating a desirable condition that

allows appropriate catalysts to perform

depolymerization and chemical transformation in the

same media. For example, lignocellulosic materials

have been shown to be soluble in ionic liquids with

strong hydrogen acceptor (Fort et al. 2007). Ionic

liquids that are capable of dissolving carbohydrates

have been demonstrated to enable catalytic

transformation of the carbohydrates to a versatile

platform chemical, like hydroxymethylfurfural (Zhao

et al. 2007). The process from biomass to end

products is illustrated in Figure S-3.2

F i g u re S-3.2 Schematic view of ionic liquid application in biomass



83

intermediates are used to form liquid fuelcomponents (e.g., by reactions such ashydrogenation, hydrogenolysis, aldol-condensation,and Fischer-Tropsch synthesis).

3.5.5 Ability to Va ry Ionic Stre n g t hAn advantage of processing a biomass-derivedreactant in the liquid-phase is that it is possible tocontrol the reaction chemistry by altering the pHand ionic strength of the solution. Furthermore, itis possible to take advantage of salt effects thatchange the bonding and electrostatic interactionsof the solvent with the reactants, products andintermediates of the reaction. In addition, it mayalso be possible to utilize electrochemical effectsto control the rates and selectivities of reactionstaking place in the liquid phase. Furthermore, itmay be possible to use microwave energy todeliver heat specifically to the active sites of acatalytic reaction in the liquid phase, withoutextensive heating of the bulk liquid phase, therebyaccentuating the contribution of the catalyticreaction and minimizing undesired side-reactionstaking place in the bulk liquid phase.

3.5.6 Faster Reaction Rat e sCompared to the production of bioethanol byfermentation, the production of next-generationliquid fuels by liquid-phase catalytic processingoffers the possible advantage of being carried outat faster rates. For example, while the time-scaleof a fermentation process may typically bemeasured in days, the time-scales of liquid-phasecatalytic reactions are typically of the order ofminutes. These higher rates translate to potentiallylower capital costs versus cellulosic ethanol. Inaddition, the very high selectivity for fermentationof glucose to ethanol comes at the cost of beingvery sensitive to the nature of the feed (e.g., highpreference for C6 compared to C5 sugars, andhigh sensitivity to the presence of contaminantsinhibitory to the organisms).

In contrast, whereas the catalysts used in liquid-phase catalytic processing typically achieve modestselectivities (e.g., 80%), they are not typicallydependent on the detailed structure of the reactant(e.g., glucose versus xylose, or sorbitol versusxylitol), and the negative effects caused bycontaminants are either small or can be reversedby appropriate catalyst regeneration procedures.This difference in specificity leads to potentiallyhigher versatility in accepting feedstocks for liquid-phase catalytic processing. As an example weconsider the hydrolysis of polysaccharides to formsugar monomers in aqueous solution. Comparedto fermentation, liquid-phase catalytic processingshould have the potential ability to handle abroader range of biomass, including C5 sugars.Also, while the presence of organic byproductsduring the hydrolysis step (such as furancompounds or levulinic acid) is highly detrimentalfor the subsequent fermentation of glucose toethanol, these species may play a small role in thecatalytic conversion of glucose to liquid alkanes bythe combination of dehydration and hydrogenationsteps over heterogeneous catalysts. In this respect,it may be possible to push the hydrolysispretreatment step to more severe conditions forsubsequent liquid-phase catalytic processing, whichmay reduce the need for expensive enzymesrequired for cellulose.

3.5.7 Lower Capital CostsFinally, we note that liquid-phase catalyticprocessing may lead to lower capital cost versus gasphase processing at elevated pressures, in view ofthe lower energy cost associated with compressinga liquid versus a gaseous feed. Also, liquid-phasecatalytic processing does not require vaporizationof the feed and typically takes place at lowertemperatures compared to gas-phase processes,leading to potential savings in heats of vaporizationand minimizing the need for heat integration.


3 . 6 CURRENT T E C H N O L O G YL I M I TATIONS AND RESEARCH/DEVELOPMENT NEEDS

3.6.1 Analytical Challenges with B i o m a s s - d e r i ved Fe e d s t o c k s

An evaluation of any reaction system or separationstep requires a knowledge of the componentspresent in the feed goint into and the productscoming out of the process. In some cases, a bulkcompositional analysis is sufficient while in othercases a detailed compositional breakdown ofindividual components is required. Incomingbiomass is typically characterized by bulkcompositional analysis, including the proportions ofthe feed that consist of cellulose, hemicellulose,lignin, ash, and extractable material. The ashfraction may be further characterized to determinethe concentrations of specific metals present. Ingeneral, the analytical techniques developed forevaluating solid biomass feeds are well established,if time consuming.

As biomass is routed through various liquid phaseprocessing steps, the degree of compositionalcomplexity may increase, primarily due to thecomplex chemistry of oxygen functional groups.Often this complexity may prohibit a detailedanalysis of what occurs during a reaction step.One way to address this issue is to use a modelcompound, such as purified sugar, polyol, or aprocess intermediate, as a feedstock to a givenreaction step. Even in these cases, the productsemanating from a reaction step may not be able tobe fully characterized due to non-specificreactions. Nonetheless, studies using modelcompounds offer insights into catalyst performancethat would not be possible to understand if acomplex, incompletely characterized feedstockwere used. As oxygen is rejected during thecourse of processing, the degree of compositionalcomplexity eventually decreases because the

elimination of oxygen reduces the number offunctional groups that are present. The complexityof the final liquid fuel product should be similar tothat of conventionally produced fuel. Decades ofresearch by the petroleum industry have led toanalytical techniques that are well suited to finalproduct analysis, although some methods will needto be updated to reflect the unique nature of these products.

The composition of the intermediate processstreams, then, poses the greatest analyticalchallenge for liquid phase processing. For aqueousstreams, high performance liquid chromatography(HPLC) can identify and quantify many of thecomponents present. However, HPLC has limitedcapability to identify unknown compounds and theresolution of HPLC methods lag behind that of gaschromatography (GC), making it difficult todeconvolve complex mixtures. GC can be used foraqueous streams containing non-volatile organiccomponents if a derivatization step is utilized toincrease the volatility of normally non-volatilecompounds (such as sugars). In addition, volatilecompounds present in aqueous streams such asalcohols, carbonyl compounds, esters, ethers, andfurans, can be analyzed directly by GC, as can anyorganic phases produced in the process. Thesemethods suffer from a problem with compoundidentification and quantitation. The use of gaschromatography – mass spectrometry (GC-MS)can improve matters by identifying many relativelylow molecular weight compounds. Even with thispowerful compound identification capability, highermolecular weight compounds are not easilyidentified by GC-MS due to inconclusivefragmentation patterns and, often, the absence ofcompounds from spectral databases.

Continued effort will be necessary to increase thecapability for analyzing intermediate processstreams. For kinetic and mechanistic studies, a



85

high degree of compositional identificationcompleteness is necessary to identify both therates of desired product formation and the ratesand selectivity towards less desirable components,often present at trace levels. When processdevelopment is at a more advanced stage, a lessdetailed analysis, with incomplete identification ofcomponents may be acceptable, or even necessary,given that complex feedstreams will lead to a moredifficult challenge for both the feed andintermediate product streams.

3.6.2 Developing Activity Stru c t u re R e l at i o n s h i p s

A fundamental challenge for liquid-phase processingof biomass is the ability to selectively and efficientlybreak and make C-H, C-O, and C-C bonds on acatalyst submerged in the solvent. The catalystsemployed for various liquid-phase reactions haveby and large been selected based on empiricalevidence. Although impressive activity andselectivity have been demonstrated severalquestions are unknown: Why do certain catalystswork well and others poorly? What effect if anywill the choice of solvent and the presence ofimpurities have? What room for improvement ispresent? This reduces the engineer’s ability to adaptto different feeds and to maximize the economicadvantage of liquid-phase processing. To begin toaddress these important questions, multiple,complementary sets of tools, including in situspectroscopic techniques and computationalmodeling techniques, need to be applied in concertto unraveling the secrets of chemical reactions atthe molecular scale, and electronic structure-basedtheoretical methods are particularly powerful toolsfor this purpose. These methods, particularlydensity functional theory, have been successfullyapplied to elucidating reaction mechanisms andincreasingly to designing new materials withdesired catalytic properties. Nonetheless, someimportant obstacles lie in the way of applying

theory to elucidate reactions at the liquid-solidinterface. One is to accurately capture the natureof the catalytic material. For instance, whiletransition metal catalysts tend to have simplestructures, solid acid and base catalysts as well asfunctional catalyst supports may be structuralcomplex. Another level of complexity entersthrough how catalytic materials respond to a liquidenvironment: It is conceivable that, when exposedto a dense layer of water and ionic species, evenunder mild temperature and pressure, the exteriorof the catalyst may no longer be as prepared, andthe transformation may be connected to itscatalytic activity. Another major obstacle is tounderstand whether and how the liquid phaseaffects reactions, for which there are knownexamples. There is yet no well-establishedmethodology for describing reactions across theliquid-solid interface under the influence of thedynamics of the liquid, with quantum chemicalaccuracy. Ad-hoc or highly simplified models canbe found in the literature that attempt to capturethe effect of water on adsorption and reaction.Some of them may well be effective to somedegree and can possibly be used beneficially, but nocalibration of results is yet available, and bettertheoretical development is very much needed.

3.6.3 Spectro s c o py and Imaging of C atalytic Materials under Liquid-Phase Conditions

An important challenge for advancing theunderstanding and practice of liquid-phase catalyticprocessing is to develop advanced in situ andoperando methods to observe the catalyststructure and surface properties under controlledconditions and preferably to do so under realreaction conditions. This task is made difficult bythe presence of the solvent. However, techniques such as attenuated total reflectance infraredspectroscopy, polarization modulation infraredspectroscopy, sum frequency generation, X-ray


spectroscopy, sum frequency generation, X-rayabsorption measurements (EXAFS, XANES), and X-ray diffraction, appear to be well positioned in thisrespect. Also, it is possible that high resolutionimaging techniques (e.g.,AFM, SEM,TEM) will makeadvances to contribute to this important area.

3.6.4 Design of Integ r ated Catalyst Reactor Systems

A current limitation in the development andoptimization of liquid-phase catalytic processing ofcarbohydrate feeds is that the sugar breakdownchemistry (with competitive reaction ratesunknown) is complicated and difficult to control,especially in aqueous solution. A strategy that hasproven to be effective in this respect is first toconvert the sugar to a sugar-alcohol by ahydrogenation step, thereby suppressing the ratesof degradation reactions at elevated temperaturesin solution. However, another approach would beto develop catalyst support packings and reactorconfigurations that minimize the volume of heatedsolution at elevated temperature and maximizecontact of the liquid phase with active sites on thesurface of the catalyst. These support packings willalso have to deal with the flow of insolublebiomass impurities (e.g., ash) through the catalyticreactor, or these impurities will have to be filteredfrom solution upstream of the reactor.

Related to catalyst engineering are reactorengineering challenges of working in aqueoussystems. One challenge is the low solubility ofreactive gasses such as hydrogen and carbondioxide. This results in the need for high reactorpressure and temperature. The ionic nature ofwater at elevated temperatures leads to reactorcorrosion issues. As corrosion occurs dissolvedmetals can plate out on the catalyst and impactperformance.

3.6.5 Engineering a New Generation of Liquid-phase Catalytic M at e r i a l s

The need to work in aqueous environmentsproduces a suite of unique opportunities andchallenges that have not yet been dealt with inconventional catalysis, which are often either gasphase or organic solvent-based. This field is rapidlyevolving – especially in the area of heterogeneouscatalysis in aqueous environments.The richness ofthe chemistry involved in liquid-phase processing isevident in aspects -such as choice of acid effectingproduct selectivity and reaction rate, affect oforganic additives such as DMSO, DCM, on thereaction (i.e. glycosidic hydrolysis, aldolcondensation, and dehydration).

These aspects offer unequivocal proof that effectsabove and beyond bulk effects such as pH arecritical for controlling liquid-phase processing ofbiofuels, and argue therefore, in turn, that themicroscopic environment surrounding the catalyticactive site is critical. There are emerging examplesin the realm of synthetic homogeneous catalysis ofhow control of the microenvironment around adivalent zinc cation via choice of alcoholsubstituents on the ligand can greatly accelerate(over 100 fold) the rate of phosphate estershydrolysis in water as solvent. This subsequentlyallows the operation of the process undersufficiently mild conditions for avoiding degradationreactions. These observations suggest an approachfor liquid-phase processing in which the active siteis surrounded by a preferable microsolvationenvironment. This approach is inspired by theintricacy observed in biological active sites, butreduces the complexity of these sites into thesimplest possible scenario of an active site and itsimmediate surroundings. This could employemerging methods of anchoring acid and basecatalytic active sites on silica, such as amines, alkyland aryl sulfonic acids developed by Shanks et al.,



87

while placing these active sites within an outer-sphere that surrounds each and every one of them,enforcing a favorable dielectric and bifunctionalcooperative environment for catalysis (Shanks, etal). Such an approach has been successfullydemonstrated, for example, in the nitroaldolcondensation using primary amines as active sites.Organizing amines within a nanoscale environmentof cyano-terminated functional groups synthesizeda catalyst that produced 99% of a beta-nitroalcoholproduct in the nitroaldol reaction, whereas acatalyst in which the cyano groups are replacedwith methyl groups produces the same product infifty-fold lower rate as well as a nitroolefin product,which arises from a completely differentmechanistic pathway. Finally, amines surrounded bya silanol-rich environment produce predominantlythe nitroolefin product. This example clearlydemonstrates how control of microenvironmentsurrounding a catalyst active site can significantlyimpact catalyst selectivity and activity for nitroaldolcatalysis. These concepts can be extended to themany other reactions involved in liquid-phaseprocessing for biofuel synthesis.

3.6.6 Ove rcoming Catalyst Stability Pro bl e m s

When working with biomass feedstocks, catalyststability (deactivation) issues are substantial.Chemistry that works well with pristine systems(such as purified glucose) fails when using actualbiomass feeds because of the rich mixture ofvarious impurities present. Impurities includeinorganic salts and other products in ash; proteincomponents rich in sulfur; phosphorus containingimpurities; and other unknowns. Some impuritiesact as inhibitors; when the purity is removedcatalyst performance returns. Other impuritiespermanently damaged the catalyst. One challengefaced by researchers is that impurities have adifferent effect on a catalyst to catalyst basis.However, some impurities cause problems in a

more general or universal way. Engineeringapproaches to solve this problem can include newseparation techniques to process crude biomass.Selective adsorption of substrates and products(particularly on carbon supports which have astrong affinity for organics) is another issue thatshould be addressed.

Catalyst systems that have been developed forpetroleum and petrochemical refining in generalare unstable under aqueous conditions.The mostcommon supports used in petroleum-basedcatalytic processes are based on metal oxides ofalumina, silica and alumina silicates.These supportsare unstable under hydrothermal conditions.Theissue is compounded when reactions are run underhigh or low pH. For these reasons catalystsupports for processes run under hydrothermalconditions commonly consist of carbon, monocliniczirconia, rutile titania, and to a lesser extent onniobia, tin oxide, and barium sulfate.

One of the challenges is avoiding leaching of activesites. New techniques are necessary forpermanently anchcoring active sites so that theycan be used without leaching in a continuousprocess. One of the promising approaches thatcan be used for this purpose is based on theconcept of using hydrophobic calixarenes asscaffolds for active sites, because these moleculesare not soluble in water. Attaching promoters toanchored calixarenes is a promising way thereforeto permanently anchor promoters to the surface.These anchored sites could either be used in andof themselves to enhance the activity andselectivity of a reaction, or they could be usedwithin the context of an outer sphere surroundingan active site. Another challenge related toleaching has to do with loss of inorganic activesites, such as those involved in aldol condensationfor biofuel synthesis when using MgO/OHmaterials, in water. Choice of support and


changing the composition of the support may beimportant here. For example, emerging methodsof “hydrophobizing” the surface may prevent waterfrom interacting with these sites and can be usedto reduce their leaching.

A key issue in the use of heterogeneous catalystsfor liquid-phase processing is that the catalysts canbe recycled for re-use in batch-reactor applicationsor that can be regenerated readily for flow-reactorprocesses. In this respect an important part of theoverall process will be pretreatment of biomass toremove poisons and deactivators, and this removalwill require the development of new materials andprocesses for effective separations.

3.7 RECOMMENDATIONS

IN SUMMARY, the following issues appear to be ofcritical importance for the successful application ofliquid-phase catalytic processing of biomass-derivedcompounds:

• Achieving selective transformations withgood carbon balances

• Achieving good energy balances (e.g.,minimizing energy-intensive distillation steps)

• Controlling pathways for oxygen management(decarbonylation versus hydrogenation)

• Controlling hydrogen management (hydrogenproduction and captive use)

• Controlling carbon and coke management

• Addressing lignin utilization (heat productionversus chemical production, such as aromaticsand cyclohexanes)

• Utilizing solvent effects to achieve desiredcatalyst performance, especially in multi-phasic reactors

• Developing effective methods for feedstockconditioning

• Developing catalysts that are tolerant tobiomass impurities (inorganics and ash,protein-sulfur, phosphorus)



89

3.8 REFERENCES

Hydrolysis of cellulose and hemicellulose, C. E.Wyman, S. R. Decker, M. E. Himmel,J.W. Brady, C. E.Skopec, L.Viikari in "Polysaccharides", 2nd edition,(Eds.: S. Dumitriu), Marcel Dekker, Inc., New York,2005, pp. 995-1033.

Hydrothermal degradation and fractionation ofsaccharides and polysaccharides, O. Bobleter in"Polysaccharides", 2nd edition, (Eds.: S. Dumitriu),Marcel Dekker, Inc., New York, 2005, pp. 893-937.

Barrett, C. J., J. N. Chheda, et al. (2006).“Single-reactor process for sequential aldol-condensationand hydrogenation of biomass-derived compoundsin water.” Appl. Catal., B 66(1-2): 111-118.

Bianchi, C. L., P. Canton, et al. (2005).“Selectiveoxidation of glycerol with oxygen using mono andbimetallic catalysts based on Au, Pd and Pt.” Catal.Today 102-103: 203-212.

Bicker, M., D. Kaiser, et al. (2005).“Dehydration ofD-fructose of hydroxymethylfurfural in sub- andsupercritical fluids.” J. Supercrit. Fluids 36: 118-126.

Carlini, C., P. Patrono, et al. (2005).“Selectiveoxidation of 5-hydroxymethyl-2-furaldehyde tofuran-2,5-dicarboxaldehyde by catalytic systemsbased on vanadyl phosphate.” Appl. Catal.,A 289:197-204.

Chaminand, J., L. Djakovitch, et al. (2004).“Glycerolhydrogenolysis on heterogeneous catalysts.” GreenChem. 6: 359-361.Collins, P. and R. Ferrier “Monosaccharides.”Monosaccharides,Wiley,West Sussex, England,1995.

Cortright, R. D., R. R. Davda, et al. (2002).“Hydrogen from catalytic reforming of biomass-derived hydrocarbons in liquid water.” Nature 418:964-967.

Dais, P. (1987).“Intramolecular hydrogen-bondingand solvation contributions of the relative stabilityof the furanose form of fructose in dimethylsulfoxide.” Carbohydr. Res. 169: 159-169.Dalavoy,T., J. E. Jackson, et al. (2007). Journal ofCatalysis 246: 15-28.

Dasari, M.A., P.-P. Kiatsimkul, et al. (2005).“Low-pressure hydrogenolysis of glycerol to propyleneglycol.” Appl. Catal.,A 281: 225-231.

Davda, R. R., J.W. Shabaker, et al. (2005).“A reviewof catalytic issues and process conditions forrenewable hydrogen and alkanes by aqueous-phasereforming of oxygenated hydrocarbons oversupported metal catalysts.” Appl. Catal., B 56: 171-186.

D.A. Fort, R. C. Remsing, R. P. Swatloski, P. Moyna,G. Moyna, and R. D. Rogers, Green Chem., 2007, 9,63–69

Franks, F. (1987).“Physical Chemistry of smallcarbohydrates- equilibrium solution properties.”Pure Appl. Chem. 59(9): 1189-1202.

Gandini,A. and M. N. Belgacem (1997). “Furans inpolymer chemistry.” Prog. Polym. Sci. 22: 1203-1379.

Halliday, G.A., R. J. Jr.Young, et al. (2003).“One-Pot,Two-Step, Practical Catalytic Synthesis of 2,5-Diformylfuran from Fructose.” Org. Lett. 5: 2003-2005.


Huber, G.W., J. N. Chheda, et al. (2005).“Production of Liquid alkanes by Aqueous-PhaseProcessing of Biomass-Derived Carbohydrates.”Science 308: 1446-1450.

Huber, G.W., R. D. Cortright, et al. (2004).“Renewable alkanes by aqueous-phase reforming ofbiomass-derived oxygenates.” Angew. Chem., Int.Ed. 43: 1549-1551.

Huber, G.W., S. Iborra, et al. (2006).“Synthesis ofTransportation Fuels from Biomass: Chemistry,Catalysts, and Engineering.” Chem. Rev. 106: 4044.Huber, G.W., J.W. Shabaker, et al. (2003).“RaneyNi-Sn Catalyst for H2 Production from Biomass-Derived Hydrocarbons.” Science 300: 2075-2078.

Jr.Antal, M. J.,W. S. Mok, et al. (1990).“Kineticstudies of the reactions of ketoses and aldoses inwater at high temperature. 1. Mechanism offormation of 5-(hydroxymethyl)-2-furaldehyde fromD-fructose and sucrose.” Carbohydr. Res. 199: 91.

Jr.Antal, M. J.,W. S. L. Mok, et al. (1990).“Kineticstudies of the reactions of ketoses and aldoses inwater at high temperature. 2. Four-carbon modelcompounds for the reactions of sugars in water athigh temperature.” Carbohydr. Res. 199: 111.

Kuster, B. M. F. (1990).“5-Hydroxymethylfurfural(HMF).A Review Focussing on its Manufacture.”Starch 42: 314-321.

Lahr, D. G. and B. H. Shanks (2003). Ind. Eng. Chem.Res. 42: 5467.

Lahr, D. G. and B. H. Shanks (2005).“Effect of sulfurand temperature on ruthenium-catalyzed glycerolhydrogenolysis to glycols.” J. Catal. 232: 386-394.

Lecomte, J.,A. Finiels, et al. (2002).“Kinetic study ofthe isomerization of glucose into fructose in thepresence of anion-modified hydrotalcites.” Starch54: 75-79.

Lewkowski, J. (2001).“Synthesis, Chemistry andApplications of 5-Hydroxymethyl-furfural And ItsDerivatives.” Arkivoc Available electronically atwww.arkatusa.org/ark/journal/2001/I01_General/403/0113.pdf 17-54.

Lichtenthaler, F.W. and S. Peters (2004).“Carbohydrates as green raw materials for thechemical industry.” Comptes Rendus Chimie 7: 65-90.

McCoy, M. (2005). Chem. Eng. News 83: 19.

Mercadier, D., L. Rigal, et al. (1981).“Synthesis of 5-Hydroxymethyl-2-furancarboxaldehyde Catalysedby Cationic Exhange Resins. Part 1. Choice of theCatalyst and the Characteristics of the ReactionMedium.” J. Chem.Technol. Biotechnol. 31: 489-496.

Miyazawa,T.,Y. Kusunoki, et al. (2006). J. Catal. 240:213.

Moreau, C., M. N. Belgacem, et al. (2004).“Recentcatalytic advances in the chemistry of substitutedfurans from carbohydrates and in the ensuingpolymers.” Top. Catal. 27(1-4): 11-30.

Moreau, C., R. Durand, et al. (1997).“Hydrolysis offructose and glucose precursors in the presence ofH-form zeolites.” J. Carbohydr. Chem. 16: 709-714.

Moreau, C., R. Durand, et al. (1996).“Dehydrationof fructose to 5-hydroxymethylfurfural over H-mordenites.” Appl. Catal.,A 145(1-2): 211-224.



91

Nagamori, M. and T. Funazukuri (2004). “Glucoseproduction by hydrolysis of starch underhydrothermal conditions.” J. Chem.Technol.Biotechnol. 79: 229-233.

Ott, L., M. Bicker, et al. (2006).“Catalyticdehydration of glycerol in sub- and supercriticalwater: a new chemical process for acroleinproduction.” Green Chem. 8: 214-220.Paul, S. F. (2001).“Alternative Fuel.” US Pat.,6,309,430,.

Qian, X., M. R. Nimlos, et al. (2005).“Ab initiomolecular dynamics simulations of D-glucose andD-xylose degradration mechanisms in acidicaqueous solution.” Carbohydr. Res. 340: 2319-2327.

Saxena, U., N. Dwivedi, et al. (2005). Ind. Eng. Chem.Res. 44: 1466.

Shabaker, J.W., G.W. Huber, et al. (2003). “Aqueous-Phase Reforming of Ethylene Glycol OverSupported Platinum Catalysts.” Catal. Lett. 88: 1-8.

Soares, R. R., D.A. Simonetti, et al. (2006).“Glycerolas a source for fuels and chemicals by low-temperature catalytic processing.” Angew. Chem.,Int. Ed. 45: 3982-3985.

Tronconi, E., N. Ferlazzo, et al. (1992). Chem. Eng.Sci. 47: 2451.

Valenzuela, M. B., C.W. Jones, et al. (2006).“BatchAqueous-Phase Reforming of Woody Biomass.”Energy Fuels 20: 1744-1752.

Wang, K., M. C. Hawley, et al. (1995). Ind. Eng.Chem. Res. 34: 3766.

Werpy,T. and G. Petersen (2004).“Volume 1:Results of Screening for Potential Candidates fromSugars and Synthesis Gas.” Top Value AddedChemicals From Biomass Available electronically athttp://www.osti.gov/bridge.

Zartman,W. H. and H.Adkins (1933). J.Am. Chem.Soc. 55: 45559.

Zeitsch, K. J. (2000).The chemistry and technologyof furfural and its many by-products, Elsevier,Amsterdam, Sugar Series,Vol. 13, 1st edn., pp 34-69.


4 . C a t a lytic Conversion of Syngason CO hydrogenation to fuels and accompanyingorganic chemicals, which provide the sourcematerial for a broad range of synthetic products(e.g. vinyl, polyester, rubber, and plastic).

FTS reactions that produce paraffinic hydrocarbonsprimarily are a subset of potential COhydrogenation reactions. CO hydrogenationreactions can also produce synthetic natural gasand alcohols, as shown in Figure 4.1, as well asolefins and other compounds (not shown).Synthetic diesel and jet fuels produced using FTSare sulfur-free and have been shown to work wellin conventional engines with essentially no designmodifications. On the other hand, ethanol and

4.0 CATA LYTIC CONVERSION OF SYNGAS

OVERVIEW: Biomass gasification is a flexible,energy efficient, and “green” way to produce low-to-medium energy fuel gases, synthesis gas, andhydrogen for fuel cell applications. Syngas,produced by gasification, can be converted toelectric power, hydrogen, steam and/or a widevariety of fuels/chemicals depending on locationrequirements. This thrust area focuses on thecleanup and conversion of syngas derived fromlignocellulosic biomass. The actual gasificationprocess, while not the focus of this thrust area, isaddressed to a small degree to provide context fordiscussions of subsequent syngas conversion.Clean up of the gases exiting the gasifier andwater-gas-shift reactions are considered heavily.A major focal point of this thrust is on theconversion of CO and H2 to fuels (andaccompanying chemicals).

4 . 1 I N T RO D U C T I O N

Currently, the best defined and technically provenroute for producing alternative fuels fromlignocellulosic biomass involves gasification/reforming of biomass to produce syngas (CO +H2), followed by syngas cleaning, Fischer-Tropschsynthesis (FTS) or alcohol synthesis, and someproduct upgrading via hydro-processing.Thissection of the report focuses on cleanup andconversion of the syngas produced through thegasification of lignocellulosic biomass.The actualgasification process will be considered only insofaras necessary to provide context for thesubsequent cleanup and conversion reactions. Inparticular, clean-up of the raw syngas exiting thegasifier (consisting primarily of H2, CO, CO2, andH2O with significant amounts of tars, ammonia,and H2S as impurities) and water-gas-shift will beconsidered. A major focus of this section will be

BIOMASSGASIFICATION

GASCLEAN-UP

WATER-GASSHIFT

COHYDROGENATION

syntheticnatural gas

gasoline

jet fuel

diesel

heatingfuel

methanolDME

ethanol

higheralcohols

F i g u re 4.1 A typical process based onsyngas conversion would consist of a

biomass gasifier, a biomass gas clean-upu n i t , a water-gas-shift reactor in cert a i n

c a s e s , and finally a reactor for syngasc o nversion (i.e., CO hy d rog e n at i o n ) .



4.1 Introduction


4.3 Resulting Fuels




4.7 Recommendations

4.8 References


Bartholomew, Andre Boehman,Stephen Chuang, Davis, Calvin Feik,Santosh Gangwal, James Goodwin,Guo, John Holladay, Devinder Mahajan,Steven Phillips,Thomas HenryVanderspurt, Yong Wang, and Ye Xu

higher alcohols generated from biomass by alcoholsynthesis, are better suited for gasoline blending

The required H2/CO ratio is different for each ofthe target products named in Fig. 4.1 (e.g., about1.0 for Fe-catalyzed FTS, and 2.0 for Co-catalyzedFTS and methanol synthesis).The actual H2/COratio of the clean syngas produced using differentfeedstocks and processes can vary significantly. Forexample, the ration can be 0.7 to 1.5 with coal orbiomass based processes and > 3 with natural gas,The following techniques can be used to modifythe H2/CO ratio up or down as necessary. In thecase of syngas with a low H2/CO ratio, the ratiocan be adjusted up by combining the syngas fromthe gasification unit with a methane reformerhaving a high H2/CO ratio or alternatively byemploying a water-gas-shift (WGS) reactor. Whenthe H2/CO ratio is higher than required, it isusually lowered by passing the syngas through aH2-selective membrane unit. Physical separationusing a membrane or simple blending is usuallymore economical and impurity-tolerant than acatalytic WGS step.

4 . 2 OVERALL PRO C E S SD E S C R I P T I O N

4.2.1 Gasification of BiomassThe gasification technology involves hightemperature (600-900°C) partial combustion ofbiomass in the presence of a gasification medium(e.g., oxygen, air, steam) in a gasifier. Fluidized-bedbiomass gasifiers are best suited for large-scaleoperations and, for this reason, they are thegasifiers of choice. Biomass gas from a gasifierconsists of syngas containing mostly H2, CO, CO2,and H2O but with important amounts of tars,ammonia, H2S, and particulates as impurities[Torres et al., 2007].These impurities are theprimary barrier to direct use of gasification gas forproducing fuels, chemicals, or electricity. They must

be removed before the syngas can be used in anengine, turbine or fuel cell for producing power, orin a catalytic reactor for producing liquid fuels andchemicals. In addition to these contaminants, thegas may contain trace quantities of HCN, halogens(e.g., HCl), alkali metals, and other metals (Hg,As,Pb). Tars, in particular, can coat surfacesdownstream and gum up power producing devices[Reed et al., 1999].

4.2.2 Biomass Gas Clean-upSyngas must be cleaned up prior to COhydrogenation since H2S can poison metalcatalysts, tars can cause fouling of such catalysts,and ammonia can block catalyst sites as a result of

BIOMASS GASIFICAT I O N


Biomass gasification is a complex thermochemical

process that consists of a number of elementary

chemical reactions, beginning with the partial

oxidation of a lignocellulosic fuel with a gasifying

agent, usually air, oxygen, or steam. Volatile matter,

which is released as the biomass fuel is heated,

partially oxidizes to yield the combustion products

H2O and CO2, plus heat to continue the endothermic

gasification process. Water vaporizes and biomass

pyrolysis continues as the fuel is heated. Thermal

decomposition and partial oxidation of the pyrolysis

vapors occur at higher temperatures, and yield a

product gas composed of CO, CO2, H2O, H2, CH4,

other gaseous hydrocarbons (including oxygenated

hydrocarbons from some processes), tars, char,

inorganic constituents, and ash. A generalized

reaction describing biomass gasification is as follows:

(1) biomass + O2 (or H2O)

CO, CO2, H2O, H2, CH4 + other

hydrocarbons

tar + char + ash

HCN + NH3 + HCl+ H2S + other

sulfur gases

The actual composition of the biomass gasification

product depends heavily on the gasification process,

the gasifying agent, and the feedstock properties

[Beenackers and van Swaaij, 1984; Hos and

Groeneveld, 1987]. Various gasification technologies

have been under investigation for their potential to

convert biomass into a gaseous fuel. These include

gasifiers where the biomass is introduced at the top

of the reactor and the gasifying medium is either

directed co-currently down (downdraft) or counter-

currently up (updraft) through the packed bed.

Other gasifier designs incorporate circulating or

bubbling fluidized beds. Tar yields can range from

0.1% (downdraft) to 20% (updraft) or greater

(pyrolysis) in the product gases. The energy content

of the gasification product gas ranges from 5 MJ/Nm3

to 15 MJ/Nm3 and is considered a low to medium

energy content gas compared to natural gas (35

MJ/Nm3). If air is used as the gasifying agent, then

roughly half of the product gas is N2 [de Bari, et al.

2000]. The relative amount of CO, CO2, H2O, H2, and

hydrocarbons depends on the stoichiometry of the

gasification process. The air/fuel ratio in a gasification

process generally ranges from 0.2-0.35 and if steam

is the gasifying agent, the steam/biomass ratio is

around 1. The actual amount of CO, CO2, H2O, H2,

tars, and hydrocarbons depends on the partial

oxidation of the volatile products, as shown in

equation (2).

(2) CnHm + (n/2+m/4) O2 nCO + (m/2) H2O

The char yield in a gasification process can be

optimized to maximize carbon conversion or the

char can be thermally oxidized to provide heat for

the process. Char is partially oxidized or gasified

according to the following reactions:

(3) C + 1⁄2O2 CO

(4) C + H2O CO + H2

(5) C + CO2 2CO (Boudouard reaction)

The gasification product gas composition, particularly

the H2/CO ratio, can be further adjusted by

reforming and shift chemistry. Additional hydrogen

is formed when CO reacts with excess water vapor

according to the water-gas shift reaction

(6) CO + H2O CO2 + H2

Reforming the light hydrocarbons and tars formed

during biomass gasification also produces hydrogen.


4 . C a t a lytic Conversion of Syngas

95

competitive adsorption. The target level of cleanupdepends on the contaminant tolerance of thedownstream conversion catalyst. The firstapproach to minimize the impurities listed above isto optimize properties of the biomass gasifier andits operating conditions (primary treatment).Thesecond approach is to remove these impurities in adownstream cleaning system based on physical(scrubbers, filters) or catalytic strategies(secondary treatment).This secondary treatment isalso known as hot gas clean-up of biomassgasification gas.

4.2.3 Wat e r - G a s - S h i f tThe H2/CO ratio from a typical biomass gasifier isbelow 1. Therefore, a separate water-gas-shiftreactor may be required to bring the ratio in linewith the necessary stoichiometry for productionof alcohols or of hydrocarbons using a cobaltcatalyst. However, since iron FTS catalystssimultaneously catalyze both WGS and FTSreactions, a separate WGS reactor is in that caseunnecessary.

4.2.4 Syngas Conve r s i o nThe syngas is converted to liquid fuels via COhydrogenation. The type of liquid fuel produceddepends on the type of catalyst used and reactionconditions.

4.3 RESULTING FUELS

A wide variety of fuels can be produced fromsyngas, including gasoline, diesel fuel, heating fuel,jet fuel, synthetic natural gas, methanol, di-methylether, ethanol, and higher alcohols. With theexception of methanol, di-methyl ether andsynthetic natural gas (primarily methane), it isimpossible to make any of the other fuels with 100percent selectivity. Instead, it is best to planfurther processing for the range of fuels andchemicals that result from syngas conversion. Inthis biorefinery model the products of biomass

Steam reforming and so-called dry or CO2 reforming

occur according to the following reactions and are

usually promoted by the use of catalysts.

(7) CnHm + nH2O n CO + (n+m/2) H2

(8) CnHm + nCO2 (2n) CO + (m/2) H2

Catalytic steam reforming of hydrocarbons has been

extensively studied, especially in the context of

methane reforming to make syngas (H2/CO = 2:1) for

methanol and Fisher-Tropsch liquid synthesis. The

basic mechanism of steam reforming is the

dehydrogenation of a hydrocarbon fuel and the

associated carbon deposition on the active sites of a

catalyst. Gasification of the carbon deposits via

reactions (3)-(5) yields additional CO and maintains

the catalyst activity. Similar catalysts have been

applied to biomass gasifier tar reforming with varied

success. Catalytic conversion of unwanted

hydrocarbons is applied for both product gas

purification and to adjust the composition of the

product gases for a particular end use. Tar reforming

also maintains the chemical energy content of the

product gases because, instead of being physically

removed, tars are converted to H2 and CO.


gasification are “cleaned up” in a manner that isanalogous to petroleum upgrading. A range ofproducts are formed via gasification andsubsequent catalytic upgrading, although the rangeis typically less broad than the product slateresulting from petroleum refining. Some productsmay require multiple processing steps. Forexample, if diesel fuel is preferred, the typicalapproach is to make long carbon chain paraffins(C1-C100) and then to use cracking orhydrocracking, as in oil refining, to maximize the production of hydrocarbons in the diesel fuel range.

Biofuels produced by conversion of biomass-basedsynthesis gas, commonly referred to as biomass-to-liquids (BTL), are similar to the fuels producedfrom petroleum and very similar to thoseproduced by coal-to-liquids (CTL) or gas-to-liquids(GTL). Using any of these four routes, it ispossible to produce gasoline, diesel, and jet fuels.However, diesel fuel is more readily produced bylow-temperature BTL, CTL, and GTL. Moreover,this synthetic diesel is superior to petroleum-based diesel since it contains essentially no sulfur,is cleaner burning, and consists largely of linearparaffins. It is important to be clear that dieselproduced by BTL processes is not the same as so-called “biodiesel”, which is made from plant oils oranimal fats and consisting of methyl or ethyl estersof free fatty acids.

From a more comprehensive process-systemperspective, it is also feasible to combine BTL,CTL, or GTL with electric power production usingvarious combinations of low-temperature (LT) andhigh-temperature (HT) FTS reactors, and differentseparation and recycle streams to optimizethermal and carbon efficiencies and processeconomics for a desired product slate [Dry andSteynberg, 2004; Bartholomew and Farauto, 2006].

4 . 4 S U M M A RY OF PREVIOUSRESEARCH

4.4.1 Biomass Gas Clean-UpSyngas from a biomass gasifier typically contains10,000-15,000 ppm of tars, 2000-4000 ppm ofammonia, and 100-500 ppm of H2S (Torres et al.,2007). Tars are a complex mixture of condensablehydrocarbons, but a unique definition is lacking(Biomass Technology World, 2004). Somedefinitions include "the mixture of chemicalcompounds which condense on metal surfaces atroom temperature" and "the sum of componentswith boiling points higher than 150ºC." One of thebetter definitions is “organic contaminants with amolecular weight greater than that of benzene”[Abatzoglou et al., 2000].

The number of compounds in tar can range fromhundreds at high temperature to thousands at lowtemperature. The total amount of tar leaving withthe syngas is a function of fuel type and gasificationconditions. Increasing the air-to-fuel ratio, bedtemperature, and freeboard temperature in afluidized bed gasifier reduces the tar amount[Narvaez et al, 1996; Gil et al., 1999]. Also, theaddition of dolomite to the bed reduces theamount of tar made. However, tar cannot bereduced below 1 g/m3 by varying theseparameters. Increasing the bed temperature to>900oC is not a good option, especially forbiomass containing alkali metals as they canpromote slagging and result in increased alkalivapor in the gas. Also, increasing the temperaturerequires more O2, resulting in more fuel loss as CO2.

Substantial research has been conducted on theremoval of tars and ammonia from biomass gas attemperatures greater than 500°C. Cleaningsystems based on condensation alone or systemscombining hot-gas cleanup with downstream



97

process [Paisley et al., 2001] that produces up to16,000 mg/Nm3 of tar in the gas makes use of acracking catalyst (DN-34).

Dolomites (CaMg(CO3)2) are the most extensivelyresearched basic catalysts for the purpose of tarremoval mainly because of their low cost. Use ofdolomite in a reactor after the gasifier results in adecrease in the concentration of tars in theeffluent stream. CaO appears to also catalyzesteam reforming of higher hydrocarbons [Simell,1997]. However, steam reforming of tars isnormally accomplished using a Ni reformingcatalyst in 1-2 reactors after the gasifier [Caballeroet al., 2000; Hepola and Simell, 1997;Aznar et al.,1998]. Ni is supported on thermo-resistant Si-freesupports such as -alumina, MgAl spinel, or ZrO2.Ni catalysts for steam reforming may containpromoters of Fe, Mn, K, or Ba. Caballero et al.[2000] were able to get the tar content down toas low as 2 mg/Nm3 using dolomite in the gasifierand 2 reactors containing Ni catalysts at 900ºC to clean up the gas. Tar contents as low as 10mg/Nm3 were obtained without noticeabledeactivation of the Ni catalyst over a 50-hour run.Nickel catalysts, however, can be poisoned by sulfurin the syngas (requiring the level of sulfur be keptbelow 10 ppm) or by carbon deposition.

The main problem in catalytic tar removal is cokeaccumulation resulting in catalyst deactivation andsulfur poisoning of catalysts due to H2S present inthe stream. Currently, there are no cleaningsystems that have demonstrated consistent cleanup(in particular, tar removal) on a long-term basisfrom an updraft fluidized-bed biomass gasifier at itspractical exit temperature, which is less than800ºC (1472ºF).

Ammonia and hydrogen cyanide formation inbiomass gasification increases in proportion to theamount of fuel-bound nitrogen. [Zhou and

condensation have been researched in the past.The removal of tars and ammonia from biomassgasifier gas is typically approached in one of severalways. One way is by choosing operating conditionsthat minimize tar and ammonia formation. Thishelps but has never been able to reduce theproduction of those compounds to tolerable levels.

The second way of removing tars is to adddolomite to the gasifier to reduce the productionof tars. Gil et al. [1999] found that use of dolomitein the gasifier reduced tar from 20 to ca. 2 g/Nm3

at 800-820oC. They hypothesize that dolomiteacts as a base catalyst and catalyzes steamreforming of the tars. However, use of dolomitein the gasifier to reduce tars has been reported toincrease the formation of ammonia [Berg et al.,2001]. Also, because of poor attrition resistance,the dolomite can lead to production of dustentrained in the syngas. In addition, although theconcentration of tars is significantly decreased(down to 1-8 g/Nm3 depending on gasifieroperating conditions), it is still too high.

The third way of removing tars is by placingreactors after the gasifier for cracking or steamreforming of the tars. Steam reforming ofhydrocarbons to produce CO + H2 (i.e. syngas) ishighly endothermic [Kochloefl, 1997]. Materialsthat can effectively crack tars include nickel-basedcatalysts, limestone and dolomite. However, thesematerials require very high temperatures, on theorder of 871-899ºC (1600-1650ºF), that aretypically not available at the gasifier exit. Long-term durability of these materials is also notconfirmed. Attempts at scrubbing (cold cleaning)the tar without a cracking step have largely metwith failure because of the tendency of tar to forman aerosol. Catalysts studied for tar removalinclude acidic catalysts (e.g., zeolites, silica-alumina,sulfated metal oxides), iron catalysts, and supportedNi catalysts. The Battelle/FERCO gasification


Masutaui, 2000; Leppälahti and Koljonen, 1995].Using pressurized fluidized-bed gasification, de Jonget al. [2003] found fuel-nitrogen conversion toammonia to be >50% for a gasifier operated at 1to 7 atm and a maximum temperature of 900ºC.Again as mentioned above, such temperatures arenot practical for biomass gasification. The effect oftemperature on NH3 formation is complicated.Zhou et al. [2000] reported a significant decreasein NH3 formation with increasing temperature.Leppälahti et al. [1995] reported an increase inNH3 formation with temperature until most of thevolatile matter was released, followed by adecrease with increasing freeboard temperature. Apressurized top-fed fluidized-bed reactor appearedto convert a lower amount of fuel-nitrogencontent to NH3 than did bottom fed reactors[Chen, 1998]. This appears to be due todifferences in the environment in which the initialflash pyrolysis takes place.

The preferred methods for ammonia removal is todecompose it to N2 and H2.The traditionalcatalysts for ammonia decomposition includesupported Ru, Ni, Fe catalysts. Carbides andnitrides of W,V, and Mo can also be used for thispurpose. Calcined dolomites and supported Nicatalysts are efficient in both NH3 and tar removal.Also, if not removed, it is expected that more than50% of the ammonia will be converted to NOx inan engine or turbine.

There are a wide variety of metals and metaloxides/carbides/nitrides that can catalyze thedecomposition of ammonia. These include GroupIb (Cu,Ag), Group IIb (Zn), Group IIIa (La, Ce),Group IIIb (Al, Ca), Group IVa (Ti, Zr), Group IVb(Ge, Sn), Group Va (V, Nb), Group VIa (Cr, Mo,W),Group VIIa (Mn, Re), and all of Group VIII (Fe, Co,Ni, Ru, Rh, Pd, Os, Ir, Pt) [Bera and Hegde, 2002;Cholach et al., 1981; Grosman and Loeffler, 1983;Nakatsuji et al., 1996; Papapolymerou and

Bontozoglou, 1997; Sheu et al., 1997; Sugishima andMitsuharu, 1995]. Although the Group VIII metalstend to be more active than many other elements,carbides and nitrides of Groups Va and VIa can beespecially active for ammonia decomposition. Forexample, Mo2C is about twice as active as thevanadium carbides that are 2-3 times more activethan Pt/C [Choi et al., 1997]. Vanadium nitrideswere found to be comparable or superiorcatalytically to Ni supported on silica-alumina[Choi et al., 1997]. LaNi alloys are also very activedue to the formation of a nitride phase [Dinkovand Lazarov, 1993]. CaO [Chambers at al., 1996],MgO [Kagami et al., 1984], and dolomite (CaO-MgO) are all active. MgO will decomposeammonia to N2 and H2 at temperatures as low as300oC. Group VIII metals seem to be active mainlyin the metallic state [Friedlander et al., 1977]. Eventhough reaction is noted for oxides of Group VIIImetals in gases containing H2 or CO, this is likelyonly due after reduction of the metal surface.Activity for ammonia decomposition on smoothmetal surfaces has been reported to fall in thefollowing order: Co > Ni > Cu > Zr [Artyukh etal., 1963].

Gas phase composition can significantly impactcatalyst activity. For example, CaO is deactivatedalmost totally when CO, CO2, and H2 are present[Kong et al., 2001; Chambers, et al., 1996] –probably due to the reaction of CO2 with theCaO. Ni on the other hand, does not seem to beaffected by the presence of such gases [Hepola andSimell, 1997]. Small quantities (< 2000 ppm) ofH2S, were not found to lead to severe poisoning ofcalcined dolomite, CaO or, surprisingly, Fe fordecomposition of 2000 ppm quantities of ammonia[Kong et al., 2001; Hepola and Simell, 1996]. Notsurprisingly, since ammonia synthesis has beenfound to be highly structure sensitive, the ammoniadecomposition activity of Fe has also been foundto be highly dependent on particle size (20-50 nm)



99

[Ohtsuka et al., 2004]. NH3 decompositioncatalysts in the presence of syngas with andwithout H2S have been extensively evaluated in thepast [Gangwal et al., 1992; Gangwal et al., 1997;Jothimurugesan and Gangwal, 1998]. When H2S isless than 10 ppmv, commercial reforming catalystswork reasonably well, even at 650ºC, for NH3decomposition.

The sulfur content of biomass is typically very lowcompared to fossil fuels. Biomass gasificationresults in approximately 20 to 600 ppmv H2S +COS in the syngas. However, even this low sulfurcontent is too high because it can rapidlydeactivate downstream catalysts, typically nickel-based, used for tar removal [Hepola and Simell,1997] and ammonia decomposition [Krishnan etal., 1988]. Deactivation may be overcome by theuse of very high temperatures (< 900ºC) for thenickel-based hot-gas cleanup reactor. But such hightemperatures create material-of-constructionissues and also result in accelerated deactivation ofthe catalyst due to sintering. Also, as mentionedabove, such high temperatures are not practical insyngas leaving fluidized-bed biomass gasifiers.Removal of sulfur to very low levels is alsonecessary if the goal is to use the syngas in afuel/chemical synthesis reactor or a fuel cell.

Sulfur can be removed down to low levels using anamine scrubbing process or chilled methanol(Rectisol Process). These are expensive, equipmentintensive, low temperature processes. A moreattractive option would be to use a hightemperature regenerable sorbent or solid reactant.One such material is a zinc oxide-based sorbent[Gangwal et al., 2002] that functions as follows:

ZnO + H2S ZnS + H2O (sulfidation)

ZnS + 3/2 O2 ZnO + SO2 (regeneration)

Zinc ferrites and zinc titanates are also effective asH2S removal sorbents.

Although hot gas clean-up has been studied formany years, a recent review paper [Torres et al.,2007] and references cited therein indicate thatconsiderable additional R&D will be necessarybefore full commercialization of biomass gas clean-up.

4.4.2 Wat e r - G a s - S h i f tThe water-gas-shift (WGS) is a well-known andhighly practiced commercial process. This reactioncan be written as:

CO + H2O = CO2 +H2

Commercial catalysts are available for low, mediumand high temperature ranges - Cu/Zn basedcatalysts at 225 to 250oC, Co-Mo catalysts at 350to 375oC, and Fe-Cr catalysts at 450-475oC.The Cu-Zn catalyst is easily poisoned by syngascontaminants, especially sulfur; thus, removal ofH2S in syngas down to less than 60 ppb isrequired. On the other hand the high temperatureFe-Cr catalyst is known to withstand sulfur atmoderate levels of around 50-100 ppm. The Co-Mo catalyst, which is actually used in the sulfidedform, can be used in sour gas at H2S levels ofseveral thousand ppm. In fact, it requires thepresence of H2S in syngas to maintain itssulfidation level. The choice of the water-gas-shiftcatalyst for a biomass-derived syngas would be sitespecific depending on the sulfur level in thebiomass and the overall process design. In general,since the sulfur is typically low for woody biomass,a Fe-Cr catalyst may be a good option. Tars wouldperhaps remain vapor in this temperature rangeand may not affect the catalyst. It would be useful to evaluate the effect of tar on the Fe-Cr catalyst in order to determine what levels of tar itcould tolerate.


4.4.3 CO Hydrog e n at i o n

4.4.3.1 Synthetic Natural GasSynthetic natural gas (SNG) production ormethanation via CO hydrogenation using a nickel-based catalyst is commercially practiced in the U.S.(e.g. at the Great Plains Coal Gasificationcomplex). Since the currently employed catalysthas sufficient activity to achieve operation at ornear the diffusion limited rate, it would seemdifficult to produce additional effective catalystimprovements, apart from a longer life for analready robust catalyst. However, since biomassgasification can already produce significant levels ofmethane, a further increase in methane production

The distribution of hydrocarbon chain length in FTS

is governed by a parameter called the Anderson-

Schultz-Flory chain growth probability, . This

parameter is borrowed from a similar concept in

polymerization chemistry. The synthesis of FTS

products can be understood by looking at a

x

+ *Hx 4

y

+ *Hx y 2 6

y

CO + 2 * *C + *O

*C + x*H *CH + *

*CH * + CH Chain Termination

+ *CH Chain Growth

*CH -CH * + C H Chain Termination

+ *CH

�

�

+ *H2 2x y 3 8

+ *H

+ *H

Chain Growth

C H -CH * + C H Chain Termination

CARBON CHAIN GROWTH IN FISCHER-TROPSCH SYNTHESIS

simplified mechanism where * indicates a catalyst

site, and *CHx represents an adsorbed reaction

intermediate (in this case having a single carbon).

The mechanism proceeds in a way such that

intermediates are either terminated by

hydrogenation (chain termination) to a product

(represented here in the form of paraffins) or

continue to grow as a carbon chain (i.e., polymer)

(chain growth). The Anderson-Schultz-Flory chain

growth probability is thus a number between 0-1

giving the probability that chain growth will occur

rather than termination. This probability is affected

by the active catalytic material, the presence of

promoters, and the reaction conditions used. The

product distributions for different values of this

parameter are shown in Figure 4.2.

via modification of the gasification process (e.g.pyrolysis followed by catalytic gasification) may bea fruitful research area to pursue, depending onsite-specific demand for natural gas.

4.4.3.2 GasolineSasol uses a HT FTS process with a fused-ironcatalyst to produce gasoline range hydrocarbons ina bubbling fluidized-bed reactor. LT FTS (Co or Fecatalyst) generally produces a relatively small C5-C12 fraction of mostly low octane n-paraffins;hence this approach by itself would not be efficientfor producing gasoline. However, in a large FTcomplex using both LT and HT FTS, it is possiblethrough a combination of several series/parallel



101

350oC) FTS process is used. The wax (C20+)produced in the low temperature FTS process canbe subjected to mild hydrocracking in a separatereactor to further increase the desirable jet fueland diesel fraction.

Catalysts based on cobalt and iron find wideapplication in FTS because of their high activitiesand selectivities for production of higherhydrocarbons (C5+) and relatively low selectivitiesfor gaseous hydrocarbons (C1-C4). Co catalystsare generally 5-10 times more active than ironcatalysts for comparable conditions. Moreover,carbon selectivities to C5+ of Co catalysts are alsogenerally higher relative to Fe catalysts, since Coproduces very little or no CO2.The hydrocarbonchain growth in FTS is largely dictated by theAnderson-Schulz-Flory (ASF) distributionparameter, , typically ranging from 0.75 to 0.95and affected by both the catalyst and reactionconditions. A catalyst with of 0.9 or higher is

1.0

0.8

0.6

0.4

0.2

0.00.0 0.2 0.4 0.6 0.8

C1

C2 - C4C5 - C11

C19+

C12 - C18

Chain Growth Probability,1.0

(a)

classical catalystnew catalyst

0.75 0.80 0.85 0.90 0.95probability of chain growth,

0

20

40

60

80

100

naphtha

wax

C1 + C2 fuel gas

C3 +C4 LPG

kerosene

gas-oil

(b)

F i g u re 4.2 (a) Weight fraction of hy d rocarbon products as a function ofchain growth (pro p a g ation) probability during FTS. (b) Pe rcentage of

d i f f e rent hy d rocarbon product cuts as a function of chain grow t h( p ro p a g ation) probability showing a range of operation for classical and

d eveloping Fischer-Tropsch catalysts and synthesis [ B a rt h o l o m ew and Fa r r a u t o, 2 0 0 6 ] .

reactions, wax cracking, and separation stages toachieve selectivities of 70–90% for gasoline ordiesel products. Mobil developed a process for theconversion of methanol (produced from syngas) togasoline. However, even though it produced anacceptable octane number, that process is nolonger utilized because it made a gasoline with ahigh aromatic content, which is no longerpermitted.

4.4.3.4 Diesel FuelFTS is the most effective catalytic process forconverting syngas (H2 + CO) into a variety ofhydrocarbons including diesel fuel. FTS results in awide product slate ranging from C1 to C60+hydrocarbons via polymerization reactions (CO +2H2 -CH2 - + H2O) (see Figure 4-2).The so-called low temperature FTS process operating inthe 200-250oC temperature range is used for jetfuel and diesel production. To produce gasolinerange hydrocarbons the high temperature (300-


preferred as it produces less gaseous hydrocarbons(see Fig. 2b). Fe, due to its low cost and high water-gas-shift (WGS) activity, is the preferred catalystfor unshifted low H2/CO ratio (typically 0.6 to 1)syngas produced by gasifying coal or biomass.TheWGS reaction provides intrinsic hydrogen toovercome the hydrogen deficiency of the syngasfrom these feedstocks. Co, on the other hand, haslow WGS activity and is usually the preferredcatalyst for a high H2/CO ratio syngas (typically2.0 to 2.2) produced by reforming natural gas orshifting low H2/CO ratio syngas. However, thechoice between Fe and Co catalysts is not totallyclear-cut for a high H2/CO ratio syngas at low FTSconversion conditions.

Reaction mechanisms, kinetics, and design of Coand Fe FT catalysts have been vigorouslyinvestigated since1925. Results fromrepresentative previous studies of FT mechanisms,reaction kinetics, catalyst preparation,characterization, testing, and design aresummarized by Bartholomew and Farrauto [2006].

Since FTS reactions are highly exothermic, heatremoval and reactor temperature control arecrucial to control product selectivity and preventthermal degradation of the catalyst. Several reactortypes have been developed, including tubular fixed-bed (TFBR), fluidized bed, and slurry bubblecolumn (SBCR) reactors; details of theperformance characteristics of these reactors aresummarized elsewhere [Bartholomew andFarrauto, 2006]. Each reactor type has its own setof advantages and disadvantages. For example,SBCRs are said to have the following advantagesover TFBRs [Jager, 1997]: (i) simpler, cheaperconstruction; (ii) lower pressure drop; (iii) higherproduction rate for the same reactor dimensions;(iv) ease of heat removal; (v) lower catalystconsumption (20–30% of the TFBR); (vi) reducedmaintenance cost; (vii) on-line replacement ofcatalyst; and (viii) substantially lower capital cost.

For example, Jager [1997] indicated that the costof a single 10,000 bbl/day SBCR reactor train is25% of that of an equivalent TFBR system with sixreactors! Nevertheless, Shell workers havereported substantially improved performance fortheir TFBR achieving productivities 2–3 timeshigher than for their existing commercial plant inBintulu, Malaysia [Geerlings et al., 1999; Hoek andKersten, 2004].This is probably possible throughoperation at higher pressure (i.e., 40 bar instead of30 bar) favoring higher heat transfer rates and useof smaller catalyst tubes for which heat transfercoefficients are higher. In fact, Hoek and Kersten[2004] report that the performance of theirimproved TFBR rivals that of a large SBCR; i.e.,both reactors have a production capacity of 19,000bbl, but the TFBR is smaller (8 m diameter x 21 mheight relative to 8–10 m diameter x 40 m height)and has higher C5+ selectivity, lower CO2production, and lower catalyst consumption.Moreover, the 5-year life of the Shell catalysts,combined with the requirement of only one in situregeneration annually, sets a high standard forcatalyst longevity.

Refinery processes can be used for upgrading FTSliquid and wax products, such as oligomerization,catalytic reforming, hydrotreating, and mildhydro¬cracking/hydroisomerization but thecombination of these processes needs to beoptimized to produce optimum yield of diesel andjet fuel. Jet fuel and diesel produced from FTS areof high quality due to their low aromatic (diesel)and zero sulfur content (both diesel and jet fuel),although, as mentioned above, some aromatics areneeded in jet fuel to give it the desired properties.

While much is known regarding FTS catalyst designand performance, there are nevertheless significantimprovements that could be made through a moredetailed understanding of the nanoscale structureof the catalysts, reaction mechanism and kinetics.



103

4.4.3.4 Jet FuelAviation turbine fuels (ATF) (i.e., jet fuels) are acomplex mixture of C8 to C17 organiccompounds and are presently derived from thekerosene fraction of petroleum (crude oil)distillation as well as hydroprocessing of heavierfractions. Petroleum-based ATF (kerosene)represents the optimum fuel for aviation given itshigh energy density, good combustioncharacteristics, low cost, excellent thermal andoxidative stabilities, safety characteristics, andfluidity. Of particular importance, given the verylow temperatures experienced at high altitude is its fluidity as measured by freeze point (-40 to -47oC).

The petroleum shortages of 1970s led to thesearch for domestic sources of liquidtransportation fuels. Large US reserves of coal andoil shale and Canadian reserves of tar sandsspurred the development of conversion processesto produce fuels from these non-petroleumsources. In the 1980s, programs were initiated todemonstrate the suitability of fuel derived fromshale, coal and tar sands. Engine tests and flightdemonstrations of shale-derived JP-4 indicated nodeleterious effects. Jet fuels produced from syngas(CO + H2) via Fischer-Tropsch technology wereevaluated in the mid to late 1990s. A thoroughstudy by Southwest Research Institutedemonstrated that the properties of a 50/50 blendof petroleum-derived Jet A-1 and iso-paraffinic FTliquids fell well within the Jet A-1 specificationrange and should have no adverse impact onengine operation. Recent tests in 2005 of theshale-derived JP-4 and JP-8 fuels showed that thefuel still met specifications after 20 + years instorage.

Currently, the most promising route for producingalternative ATF or jet fuels involves gasification-/reforming of coal, natural gas, biomass or other

hydrocarbon feedstock to produce a syngas (CO +H2) followed by syngas cleaning, FTS, and FTSproduct upgrading via hydroprocessing.Thesynthetic ATF produced using this route is sulfur-free and has been shown to integrate well withpresent and future aircraft with essentially nomodifications to engine design [Dittrick, 2007].The synthesis of jet fuel is probably best doneusing a cobalt catalyst to make long chain paraffinsfollowed by the use of hydrocracking to produceparaffins in the desired range.While the price pointat which the FTS jet fuel becomes an economicallyviable alternative to petroleum-based jet fuelremains to be established, experts have indicatedcrude oil-price thresholds ranging from $38 to $50per barrel. However, production costs in theUnited States must be determined, including thefinancing and amortization cost of the FTS plant(estimated to be in $billions).

One of the principal concerns for the FTS-derivedATF fraction is that it cannot by itself meet thefreeze point specifications (-40 to -47 oC) ofcommercial ATF (Jet A) and military ATF (JP-8).Some aromatics and/or naphthenes need to bepresent in the FTS-derived ATF fraction in order tomeet freeze point and other importantspecifications. Research is needed to explorepossibilities for synthesizing ATF via FTS that meetsall the specifications. Such an in-spec ATF mightrequire secondary upgrading of FTS liquids.

4.4.3.5 Heating FuelThe synthesis of long-chain paraffins using a cobaltcatalyst followed by hydrockracking to produceparaffins in the desired range is most likely the bestmethod for producing heating oil from biomass.See the previous section on diesel fuel. See theprevious section on diesel fuel.


H2CA R : S U S TAINABLE T R A N S P O RTATION FUELS FROM BIOMASS

Fuel molecules with the energy density and

combustion characteristics of gasoline or diesel can

be produced using carbon atoms derived from

biomass combined with hydrogen atoms generated

through carbon-free energy generation processes –

such as wind, solar, or nuclear. Instead of looking to

biomass as a source of potential energy and carbon

atoms, this hydrogen-carbon (H2CAR) process

exploits only the presence of carbon in biomass,

which reduces the amount of biomass needed to

produce finished fuels. Even using photovoltaics

power systems with 15% efficiency, the annualized

average solar energy conversion efficiency for

hydrogen generation is more than an order of

magnitude greater than it is for biomass growth

[Agrawal, et al. 2007].

A unique aspect of this hydrogen-carbon, or

“H2CAR”, process, recently described by Agrawal, et

al. [2007] and illustrated in Figure S-4.1, is that the

biomass gasifier is not only co-fed with the

supplemental H2, but also with CO2 recycled from

the gas-to-liquid conversion reactor. The great

benefit of this approach is that CO2, which is

normally lost to the atmosphere as a greenhouse gas

or sequestered in conventional approaches, is

recycled back to the process. Thus, all the carbon

fed to the system is converted into fuel product,

minimizing the amount of biomass needed to

produce a given amount of fuel and reducing

greenhouse gas emissions. Furthermore, the added

hydrogen accounts for more than half of the energy

content of the liquid product. Thus, the carbon-free

production of hydrogen contributes significantly to

the overall efficiency of the H2CAR process.

Through computational modeling, Agrawal, et al

[2007] have identified several major advantages of

this biomass to liquids process: (1) 40% less land is

needed to grow the biomass for fuels produced by

F i g u re S-4.1 The H2CAR Process



105

the H2CAR process compared with other biofuels

production processes that derive all the fuel

energy from the biomass itself; (2) because of the

decreased land area requirements, the H2CAR

process could potentially meet 100% of the

demand for transportation fuels in the U.S. using

the 1.37 billion dry tons of domestic biomass

that can be substantially harvested, instead of

the estimated 30% of total fuel demand that

Perlack, et al [2005] calculated could be

produced from this amount of biomass using

other conversion processes; and (3) this fuel-

production solution makes use of existing

transportation fuel distribution infrastructure.

Long term implementation of the H2CAR process

will certainly be accompanied by improved

methods for carbon-free hydrogen production ,

but even in advance of these improved methods,

the comparatively higher efficiencies involved

with producing hydrogen by solar power, instead

of biomass-derived hydrogen, is a compelling

benchmark for the efficiency advantage of this

process. In light of the tremendous potential of

this process, new strategies for staged

implementation and further development of

H2CAR are urgently needed.


4.4.3.6 MethanolMethanol synthesis from syngas using a Cu-Zn-based catalyst at about 50-80 atm and 225-250oCis a well-known and highly practiced commercialprocess [Bartholomew, 2006]. Using this catalyst,methanol can be produced from syngas with nearly100 percent selectivity. The presence of a smallamount of CO2 (around 4 vol%) in the syngas hasbeen found to increase methanol productivity.Mechanistically, the synthesis reaction is believed toproceed via CO2 hydrogenation. Typical industrialproductivities range between 1.3-1.5 g methanol/gcatalyst/h at space velocities on the order of10,000 scc/cc catalyst/h. Due to thermodynamiclimitations, the per pass conversion is low and theunconverted syngas is recycled to increase theoverall conversion. Previous work suggestsqualitatively that pore diffusional resistance may besignificant under commercial reaction conditionsfor pellet diameters larger than 1.5 mm andtemperatures above 240°C.

While low-pressure methanol processes employ avariety of reactor types, all are of the gas-fluid,fixed-bed catalyst type [Bartholomew and Farrauto,2006]. Reactors are differentiated primarily by themethod used for removing heat generated byreaction. Methods include: (1) indirect cooling ofcatalyst-packed tubes with boiling water (e.g.,Lurgi); (2) interstage injection of cold feed (ICI); or(3) cooling with heat exchangers between 3–4stages of adiabatic reactors. Each of these methodsenables the outlet temperature to be controlled toabout 220–240°C.There are variations on each ofthese themes, each with its own set of advantagesand limitations.The near isothermal operation ofthe tubular boiling-water reactor (TBWR)facilitates: (1) high thermal efficiency (defined asenergy content of the products divided by theenergy content of the reactants); (2) control oftemperature; and (3) high conversion, yield, and

selectivity with a minimum of byproducts.Thetemperature profile in the second half of thecatalyst bed is close to the locus of maximum rates,thus minimizing the amount of catalyst needed.Thelow average operating temperature minimizessintering, leading to the longest catalyst life (5years) for any of the methanol processes.Thechoice of reactor for a specific plant dependsgreatly on plant size and the syngas productionprocess.The TBWR is economically favored forsmaller plants characteristic of BTL, while for large,world-class plants (> 2,500 TPD), the economics ofseries-adiabatic, spherical reactors are favored.

Like the low temperature WGS catalyst, themethanol synthesis catalyst is highly susceptible topoisons, particularly sulfur and arsine. Thus thesyngas needs to be cleaned prior to the methanolsynthesis reactor.

4.4.3.7 Dimethyl Ether (DME)Dimethyl ether (DME) is a potential ultra-cleandiesel fuel that offers excellent benefits as analternative fuel in that it can be used at highefficiency and with low emissions, is readily storedand transported, provides high well-to-wheelefficiency and is safe and reliable. DME is easilymade from methanol and can be used in a mixturewith regular diesel. DME is a compound that hasbeen targeted for future use as a fuel in severalcountries around the world [Phillips and Reader,1998;Verbeek and Van der Weide, 1997;Wakai etal., 1999].

Efforts to increase the use of DME as atransportation fuel are motivated by severalfactors. There has been confirmation that the fuelyields low particulate emissions and possibly lowerNOx emissions [Sorenson and Mikkelson, 1995;Fleisch et al., 1995]. Because of theseenvironmental advantages, DME may provide a farmore effective means of achieving the “Clean Diesel



107

Technology” mandated by the US EPA for cars andtrucks by 2010.

4.4.3.8 Ethanol and Higher A l c o h o l sCatalytic synthesis of ethanol and higher alcoholsfrom syngas has been under development since thebeginning of the 20th century. Substantial researchwork has been carried out for developingprocesses to convert syngas to higher alcoholsmainly for the purpose of synthesizing a mixture ofmethanol and isobutanol as precursors for methyltertiary butyl ether (MTBE), previously used widelyas an octane enhancer.

The recent debate surrounding the use of ethanolas a transportation fuel has been contentious.While there may be many opportunities, there arealso many unknowns [Johnson, 2007].Fermentation-derived ethanol from corn starch is aparticularly controversial and politicized issue. Onthe other hand, gasification of cellulosic biomassfrom biomass waste and non-food crops followedby catalytic synthesis of ethanol has potential forproviding an efficient and economically viable routeto transportation fuels or fueladditives. Some consider thedevelopment of an effectiveethanol synthesis catalyst to bethe “Holy Grail.”

Because MTBE use has recentlybeen phased out and is beingreplaced by ethanol, the interestin the synthesis of ethanol fromsyngas is growing. Ethanol iscurrently considered as the “fuelof the future” because it can beproduced domestically frombiomass. Although it offers alower chemical energy thangasoline, it produces much lesspollution and has a higher octane

F i g u re 4.3 Vo l vo ’s 2nd Generation Heav y - D u t yp rototype DME ve h i c l e , d eveloped under the EU-

s p o n s o red “ A F F O R H D ” ( A l t e rn at i ve Fuel Fo rH e avy-Duty) project in 2005.

number. It has been estimated that the use of a 10%blend of ethanol and gasoline as a fuel, known asE10 fuel that is commercially available in many ofU.S. gas stations, can reduce automobile greenhousegas emissions by 12-19%. Use of this fuel can alsoreduce CO emissions by 39% and particulatematter emission by 50%. Consequently, there is agrowing worldwide interest in producing ethanolfrom biomass and using it as an alternativetransportation fuel.

The conversion of syngas to ethanol via directsynthesis and methanol homologation pathways hasbeen performed using a wide range ofhomogeneous and heterogeneous catalysts.Twotypes of catalysts currently hold promise for thedirect synthesis of ethanol from syngas,: Rh-basedcatalysts and Cu-based catalysts. For Rh-basedcatalysts, there are reports in the literature ofselectivities as high as 50% at higher pressure (Huet al., 2007). However, most often this highselectivity is obtained at the expense of conversion(i.e., high selectivities are only seen at very lowconversions). Depending on the type of catalystused, both the direct synthesis and indirect


synthesis via methanol homologation areaccompanied by a host of side reactions leading tomethane, C2-C5 alkanes and olefins, ketones,aldehydes, esters, and acetic acid. Methanation canbe particularly significant via hydrogenation of CO.To increase ethanol selectivity, the catalyst and thereaction conditions need to be better designed tosuppress methanation activity.

While not yet commercialized, a few higher alcoholsynthesis (HAS) processes have advanced to thepilot scale stage, or conceptual processes based onpatented catalysts have been designed.

A review of the literature on the conversion ofsyngas into ethanol and higher alcohols indicatesthat:

• Currently, higher selectivity may be achievedwith homogeneous catalysts, but commercialprocesses based on these catalysts requireextremely high operating pressures, complexcatalyst recovery, and expensive catalysts,making their commercial applicationimpractical.

• Rh-based heterogeneous catalystspreferentially produce ethanol over otheralcohols. However, the limited availability ofrhodium, high cost of Rh, and insufficientethanol yield make these catalysts lessattractive for commercial application,especially if high metal loadings are required.

• Modified methanol synthesis catalysts basedon CuZn-, CuCo-, and Mo have beendeveloped and demonstrated in pilot planttesting.The alcohol production rates of 0.1 to0.6 g alcohol/g catalyst/h are significantly lessthan that achieved in methanol synthesis (1.3to 1.5 g methanol/g catalyst/h).Thus, significantimprovements, at least 2-3 fold, in alcoholproduction rate must be achieved.

• Direct synthesis of ethanol and higheralcohols from syngas is thermodynamicallyfeasible, but kinetically restricted.

Reactor designs employed in the higher alcoholsynthesis catalyst R&D have typically adaptedstandard fixed-bed reactor technology withspecialized cooling designs used for methanol orFT synthesis of hydrocarbons. Improved productyield and selectivity could be achieved byperforming the reactions in slurry reactors due toefficient heat removal and temperature control.However, for small-scale BTL plants, fixed-bedtechnology is likely to be more economicallyfavorable and enjoy the same advantages ofprocess intensification as FT and methanolsyntheses.

4 . 5 ECONOMICS AND POTENTIAL OF TECHNOLOGY

The overall syngas conversion plant is capitalintensive, consisting of a gasifier or reformerdepending on the feedstock, syngas cleanup system,the syngas conversion reactor, product separationequipment, and possibly a hydrocracker forupgrading wax into desired products. The majorityof the cost (60-70%) lies in the gasification of thebiomass.

The economics of most syngas conversionprocesses are dominated by the initial plantinvestment. For example, the investment for acoal-to-liquids FTS plant is now approaching$100,000/daily bbl. Once the FTS plant is built, theoperational costs are in the $10-25/bbl range. Theissue is how to provide protection for thesubstantial initial investment in plant construction.Despite the technical feasibility and attractivenessof FTS and the abundant availability of indigenouscoal and biomass resources, there is currently nocommercial FTS plant operating in the US, althougha number of plants are in the planning stages.Nevertheless, there is considerable potential for



109

improving the economics of small plants suitablefor BTL through process intensification andmodule production [Nehlsen et al., 2007]. andThrough integration with power generation theeconomics of CTL plants can also be improved[Robinson and Tatterson, 2007]. It may be possible,for example, to reduce the size and cost of an FBRfor FTS in a small BTL plant by a factor of 10,compared to conventional technology.

4 . 6 CURRENT T E C H N O L O G YL I M I TATIONS A N DR E S E A R C H / D E V E L O P M E N TN E E D S

4.6.1 Biomass Conversion to SyngasA major roadblock is the cost-effective productionof synthesis gas from biomass. Since the quantityof biomass is not concentrated at a site as fossilfuels are, gasification must be done on a local levelto eliminate the transportation costs of biomass.Due to the high cost of transporting raw biomass,conversion plants would probably service only anarea of radius 50-75 miles and would produceliquid fuels equivalent only to 10,000-20,000barrels/day of oil. Therefore, methods to reducethe cost of syngas generation on a small scale aredesperately needed. If this can be accomplished,the next priority will be to either developeconomical small-scale FTS plants or to develop apipeline system to transport the syngas generatedin the small plants.

In the present and projected environment ofenergy supply and demand, it is also reasonable toexamine these premises in light of the nation’slarge and growing need for clean electrical power.Biomass is a low energy density feed stock thatmay vary with the seasons. There is a need toconsider possible co-production of electricity andsyngas, with the excess electric power generatedexported to the grid while the syngas is convertedto liquid fuels.

4.6.2 Biomass Syngas Clean-upDue to the sensitivities of the equipment,processes, and catalysts downstream, biomasssyngas clean-up is critical. The concentrations andnature of impurities in the syngas are feedstockdependent because of the variety of biomassavailable. Thus, different catalyst formulations maybe required for biogas clean-up from differentfeedstocks.

Besides steam reforming, other ways of conversionof tars need to be researched. Cracking ofhydrocarbons is a well-established technology usingsolid acid catalysts such as silica-alumina andzeolites. In fact, any catalyst with strong acid siteswill crack hydrocarbons at temperatures above100-200ºC. Aromatics present during hydrocarboncracking can lead to higher molecular weighthydrocarbons and coke [Gates et al., 1979].

Since tars are polycyclic aromatics or theirprecursors, cracking the tars would be expected toextensively deposit coke on the catalyst.The cokecould be burned off the catalyst in an airregenerator, and the regenerated catalyst could bereturned to the cracker.

An alternate technique for tar elimination ishydrogenolysis and/or ring opening of thepolyaromatics. Unfortunately, hydrogenolysis and/orring opening activity are thermodynamically limited,with conversion decreasing with increasingtemperature. At a temperature even as low as200ºC, catalysts such as Rh, Pt, Ir, and Ru have beenfound to show little ring-opening activity fornaphthalene [Jacquin, 2003].

Little is known about the reaction mechanisms andkinetics of different contaminants including H2S,tars, and others on clean-up catalysts. Moreresearch in this area will be required to developbetter catalysts and strategies for biogas clean up.


It is known that FTS requires that there be < 60ppb of H2S, arsine, HCL, HCN, NH3, particulates,Se, Hg, alkali, and P. Thus, removal of all of theseimpurities needs to be addressed. Someimpurities may greatly interfere with catalyticremoval of other impurities. However, little iscurrently known about this issue.

4.6.3 Syngas Conversion to Liquid Fuels

The conversion of biomass derived syngas to liquidfuels represents a significant redirection directionfor a field that for the past 20 years hasconcentrated on the conversion of strandednatural gas to liquids, typically in the context of a>60,000 bbl/day plant with assured long term gassupplies. The syngas in these plants has a H2/COratio of about 2. Many of these plant conceptsfurther assume a desired product of paraffinic FTS-based “syncrude” with just enoughhydroisomerization treatment to facilitatetransportation to a refinery rather than being adirectly salable commercial fuel. Since theseconcepts were based on the idea of stranded gas,the export of electrical power was neverconsidered an option.

Assuming that one can develop a small gasifier, theproblem becomes one of defining the reactor typethat will be optimum for the small-scale process.Several companies are currently developingmicroreactor technologies that could lead to asubstantial reduction in capital cost for a smallplant. However, reactor design and economicstudies are not available in the public domain forthese alternatives.

A partial list of research needs with regards to thereaction of syngas is provided below.

4.6.3.1 Fischer-Tropsch SynthesisThere are many issues associated with the FTSreactor system that result in poor economics andprevent the widespread commercialization of FTS.Important topics that require significantly moreresearch are the following:

MECHANISMS AND THE ACTIVE CATALYTIC SITESThe exact mechanisms for syngas conversion tothe various possible products is still not completelyknown, especially the synthesis of oxygenates. Thestructures required for sites to be active forhydrocarbon vs. ethanol synthesis are not known.Both are needed in order to understand and designbetter catalysts for selective synthesis.Development of active, selective, stable, andattrition-resistant catalyst formulations and afundamental understanding of FTS reaction kineticscould lead to enhanced reaction yields, desirableproduct distribution, and improved reactor designand operation.

METHANE FORMATIONMethane is the least desired FTS product. There isa lack of understanding of the paths by whichmethane is formed during FTS and why themethane fraction of the hydrocarbon productsincreases as the catalyst ages.

LIMITATION OF CHAIN GROWTHDeviations from the Anderson-Schulz-Flory (ASF)distribution for a single catalyst are most likelydominated by the operation of the reactor and notby the reaction mechanism. Thus, it is highlyunlikely that limitation of chain growth can beachieved with a single catalyst. Bifunctionalcatalysis can do this, but to date no one has beenable to achieve a reasonable catalyst lifetime for atleast one of the functions to make it worthwhile.For example, Mobil abandoned the single reactorto adapt a two-reactor system, one for FTS and



111

one for cracking.While there may be combinedcatalyst/reactor design strategies that enablevarying product slates in FTS, very little has beendone in this area.

CATALYST DEACTIVATIONDeactivation of FTS catalysts can be significant dueto pore plugging by waxes, carbon deposition,partial oxidation, poisoning, and attrition. Forexample, cobalt catalysts deactivate significantlyduring the initial reaction period before reachingsteady operation. Better understanding of suchinitial deactivation could lead to better catalystdesign and less deactivation resulting in smallerreactors and less capital expenditures. Catalystattrition can be a big issue in the use of fluidizedbed reactors and of slurry phase reactors, whichare excellent for heat management. Majorproblems are observed with catalyst attrition inand wax removal from the catalyst slurry. Acurrently operational FTS plant in Qatar is havingproblems with development of fines during waxseparation even though they are using a cobalt-alumina catalyst. Relatively little is known about thebasic mechanisms of catalyst attrition. For example,how do sudden pressure variations impact thedisintegration of the catalyst particle? There isneed to define the factors that determine the easeof wax removal from the slurry. Some research oncatalyst deactivation has been done, but, in general,especially for FTS, this has not been the focus of many fundamental studies. (see Sidebar on next page).

4.6.3.2 Higher Alcohol “ S y n t h e s i s ”( H A S )

There are other alternatives to liquid fuels besidesFTS, such as the synthesis of C2-C4 alcohols.Ethanol can be used directly as a gasoline additive.Isobutanol can be de-hydrated and the resultingolefin dimerized yielding iso-octane, a valuablegasoline component. HAS catalysts can incorporate

light olefins back into the synthesis process. It isclear that economics and feed stock availability willbe essential in choosing the most fruitful route.Methanol synthesis has already reached a high levelof commercial development, and federally fundedresearch for methanol synthesis is notrecommended. However, the yield of methanol inthe commercial process could serve as a target toreach for HAS processes.

Catalytic synthesis of ethanol from syngas suffersfrom low yield and poor selectivity of the desiredalcohol product due to the slow kinetics of theC1-C2 linear chain growth and fast chain growth toform C2+ alcohols. R&D work to improve theethanol yield and selectivity should focus ondeveloping methodology for increasing the kineticsof the C1-C2 chain growth. Homogeneouscatalysts are expensive and perhaps should not beconsidered based on cost and selectivity.

Among the heterogeneous catalysts, the high temperature ZnO/Cr2O3 catalyst that uses severeconditions (high temperature and very high pressure) is unattractive because of the following factors:

• Operating pressure is not compatible withthe operating pressure envisioned forcommercial and developmental biomassgasifiers.

• High operating temperature results in highselectivity for methane and isobutanol butnot ethanol.

The low yield combined with high cost and limitedavailability of Rh makes it less attractive. However,Rh, though expensive, still offers the bestpossibilities for the design of a selective ethanolsynthesis catalyst. Since it is used only in lowloadings, if a Rh catalyst could be designed withboth high ethanol selectivity and long lifetime, theeconomics would be acceptable.


CATA LYST DEACTIVATION AND REGENERATION IN FISCHER-TROPSCH SYNTHESIS

Catalyst deactivation in Fischer-Tropsch Synthesis

(FTS) is a serious, economic- and process-limiting

problem [Dry, 2003; Bartholomew and Farrauto,

2006]. In processes using expensive Co catalysts,

catalyst life must be extended to a minimum of 3-5

years through purposeful catalyst, reactor, and

process design, periodic in situ rejuvenation, and

either online or offline regeneration. Similar

measures are required to extend the lifetime of less

expensive Fe catalysts to at least 8-16 months.

Primary catalyst deactivation problems in FTS

include: (1) poisoning of catalysts by sulfur and/or

nitrogen compounds; (2) fouling by hard waxes and

carbon; (3) formation of inactive catalytic phases

such as oxides, inactive carbides and metal-support

compounds; (4) hydrothermal sintering; and (5)

catalyst attrition. Causes of these catalyst

deactivation problems and their prevention and/or

treatment (e.g. regeneration) are addressed in detail

by Bartholomew and Farrauto [2006]. In the case of

Co catalysts two general types of deactivation are

observed [LeViness et al., 1998]: (i) a short-term,

easily-reversible deactivation with a typical half-life

of 20–40 days due to reversible oxidation and

accumulation of hard waxes, organic acids, and

reversible poisons and (ii) a long-term, difficult-to-

reverse deactivation with a typical half-life of

100–200 days resulting from irreversible metal-

support compound and carbide formation along with

accumulation of surface carbons and reversibly-

adsorbed poisons.

Rejuvenation, a mild, in situ treatment in H2 at

around 250-300°C is effective in removing waxes and

acids, while a more severe ex situ, high-temperature

regenerative treatments in O2 and H2 is required to

reverse formation of oxides and carbides. Irreversible

poisoning by sulfur compounds and reversible

poisoning by nitrogen compounds in reduced by

decreasing these poisons to ppb levels in guard beds

upstream of the FT reactor(s). Hydrothermal

sintering is avoided by (1) choice of hydrothermally

stabilized supports and (2) avoiding operation at

steam partial pressures above 5-6 atm. Attrition can

be a serious problem for catalysts used in slurry

bubble column reactors (SBCRs). High attrition

resistance is realized through design of strong, dense

catalyst spheres typically prepared by spray drying.

The serious consequences of severe attrition were

dramatically apparent in the recent startup of the

Sasol Oryx GTL plant in Qatar (Figure 4S.1).

Apparently, the attrition rate was found to be 5

times the design level, a situation that required the

plant to be operated at 20% of its full capacity for a

period with resulting losses of millions of dollars in

potential revenue.

F i g u re S-4.1 Sasol Oryx Plant in Qatar



113

Other heterogeneous catalyst classes for furtherconsideration in HAS are:

• Alkali modified low-temperature methanolsynthesis catalyst – Cu-ZnO/Al2O3

• Alkali-modified Cu-Co alloy catalysts

• Alkali-modified Mo-based catalysts

It is noteworthy that formulations based on thesethree classes of catalysts have been used in pilotplants for HAS or are being considered ascandidates for pilot plants to be constructed in thefuture.Among the three classes, the Cu-ZnO/Al2O3 catalyst shows the lowest yield ofhigher alcohols and the Mo-based catalysts showthe highest.Thus a logical order of emphasis couldbe: Mo-based catalysts > Cu-Co alloy catalysts >Cu-ZnO/Al2O3 catalysts.

Catalysts of particular interest for furtherimprovement include Cu-Co, unsulfided Co-Mo,and unpromoted and cobalt-promoted MoS2.Current total alcohol yields from these catalystsare in the 0.1 to 0.6 g/g catalyst/h range ascompared to the bench-mark 1.3 to 1.5 g/gcatalyst/h methanol yield in the commerciallypracticed methanol synthesis process.Also,hydrocarbons, especially methane, and CO2 areproduced thereby reducing total alcohol andethanol selectivities. Modifications could be madeto these baseline catalyst formulations usingpromoters that might improve yield and selectivityof ethanol.

High dispersion of the catalyst, that has beenshown to improve activity for HAS, is also a veryimportant consideration in catalystpreparation/modification. Catalysts need to beprepared with high dispersion and with structuralpromoters to prevent sintering at reactionconditions. Scalability and cost of catalystpreparation/modification is also an important

consideration. Exotic methods of catalystmodification/preparation that cannot be easilyscaled up using conventional commercialequipment should be avoided.

Since temperature is one of the most importantreaction parameters, the temperature of maximumselectivity needs to be determined throughexperimentation and then closely controlled at thisvalue in a commercial reactor.The requirement ofclose temperature control can be met by suitablereactor choice and design. Choices include fixed-bed, fluidized-bed and slurry bubble columnreactors (SBCR). Shell and tube type fixed-bedreactors can be employed to better controltemperature. However, catalyst extrudates (typically3/16-inch) have to be used for commercialapplications whose internal temperature can bequite different from the gas temperature forexothermic reactions.These reactors are also hardto scale up. Fluidized-bed reactors use smallcatalyst particles and can provide for heat removalusing boiler tubes placed in the bed. However, thecatalyst particles need to be highly attritionresistant. SBCR could have advantages reactors fora commercial embodiment of an ethanol synthesisand HAS process.

4.6.3.3 Dimethyl Ether “ S y n t h e s i s ”

Further development of fuel system and enginehardware for use with dimethyl ether (DME) inadvanced high efficiency clean combustion engine isneeded to remove barriers to deployment of DMEvehicles in the U.S.


4 . 6 . 3 . 4 D evelopment of Improved P roduct Upgrading M e t h o d o l ogy for O p t i m i z ation of Diesel and Jet Fuel Pro d u c t i o n

Further work is recommended for optimizing theupgrading processes to maximize the yield of dieseland/or jet fuel. A combination of upgradingprocesses should be considered includingoligomerization of lighter hydrocarbons to producejet fuel-range compounds, catalytic reforming toproduce aromatics to provide the necessaryproperties to jet fuels, hydrotreating to removeoxygenates and saturate olefins, and mildhydrocracking/isomerization to produce thedesired paraffins and iso-paraffins in the desiredboiling range.

4 . 6 . 3 . 5 Reactor Te c h n o l ogies for FTS Based on Process I n t e n s i f i c ation and Catalyst/ Reactor Integ r ation

Prevailing wisdom is that SBCRs will be thereactors of choice for future large-scale FTS plantsin the US and abroad. However, for smaller plants(< 500 bbl/d) applicable to BTL,TFBRs probablyhave the edge because they can be designed withcatalyst tubes of small diameter or with thin layersof structured catalyst, an approach whichsubstantially increases heat transfer rates andcatalyst productivity, while decreasing reactor sizeand capital/operating costs.The key here ismanaging the heat removal, the limiting factor inFTS reactors, through an integrated catalyst/reactordesign and modular design [Nehlsen et al., 2007].

The design of a reactor and catalyst go hand inhand. Relatively little has been done to supportresearch on integrated catalyst/reactor systems forFTS or structured catalysts for FTS.Very little hasbeen done to model integrated reactor/catalystsystems. Such systems require new approaches forcatalyst preparation and forming - issues that

SELECTIVE ETHANOL SYNTHESIS WITHM I C ROCHANNEL REACTO R S

Ethanol synthesis from syngas is a highly exothermic

reaction, meaning that there is significant heat

released during the reaction. Thus the reaction must

be tightly controlled to prevent a detrimental

temperature rise in the reactor, especially at high

conversions. Hu et al. [2007] recently demonstrated

that ethanol can be selectively produced in a

microchannel type reactor which allows for efficient

temperature control. This work demonstrates how

integrated catalyst/reactor design can be used to

more efficiently produce syngas products. Ethanol

selectivities of 61% at a conversion of 25% were

achieved using a RhMn/SiO2 catalyst operated at a

H2/CO ratio of 2, 265oC, and 50 atm pressure. The

only other significant product was methane (34%).

The structured catalyst prepared with the same

catalyst formulation on a high heat conducting

material were able to increase the space velocity up

to 7 times (to GHSV = 20,000 h-1) while keeping the

conversion at 20+% and the ethanol selectivity only

slightly below 60%. The ability to achieve high

conversion at this high flow rate was attributed to

an improvement in mass transfer, since the catalyst

particles were only 0.5-2 microns in diameter on the

structured catalyst. Such small particles could never

be used in a packed bed reactor because of the

pressure drop that they would cause. The results are

an indication of the possibilities for process

intensification using specially designed chemical

reactors such as this microchannel reactor. By

increasing the flow through the reactor while

keeping the conversion high, the size of the reactor

required and hence its cost can be significantly

reduced.



115

FT REACTO R S :THE ECONOMICS OF SCALING DOW N

Conventional wisdom teaches that large plants can

produce a product more economically than small

plants due to economies of scale (i.e., capital cost of

equipment and processes has been found over many

decades to scale with size raised to the 0.6-0.7

power). This empirical relation-ship has guided design

and construction of syngas conversion plants for

more than 50 years.

Prior to 2005, it was widely held that large

economies of scale were preferred for GTL plants, for

example, the 34,000 bbl/day Sasol Oryx Plant in

Qatar (Figure S-4.2). The capital cost for this plant,

which began operation in late 2006, was roughly

$1.0–1.5 billion or $30,000/bbl/d. By mid-2007,

however, escalation of worldwide material and

construction costs had driven capital costs upward

by 2–3 fold. Thus, the capital cost for a 100,000

bbl/d plant constructed in 2007 could exceed

$10 billion!

Given such enormous capital costs, interest has

recently shifted to the development of small-scale

BTL, CTL, and GTL plants based on structured

catalysts and mini-reactor technologies which

apparently enable substantial (3–5 fold) reductions

in FT reactor size and cost. Moreover, pressures to

reduce CO2 emissions have created incentives to

develop processes for converting renewable biomass

resources to fuel liquids in small, geographically-

dispersed modular plants, producing 500–2,000 bbl/d

of fuel liquids in shell-tube fixed-bed reactors. A

500 bbl/d GTL plant incorporating a compact GTL

reactor with small-diameter tubes and similar

process-intensified modules for syngas conversion,

gas purification, and product upgrading could cost as

little as $25 million [Bartholomew, 2007].F i g u re S-4.2 Reactor for

Sasol Oryx Plant in Qat a r

likewise have not been addressed. Novel catalystand reactor configurations need to be consideredthat may better fit such smaller plants and result inimproved heat management and improved overalloperation.

The promise of new step-out technologies usingprocess intensification and structured catalysts canonly be realized through additional research anddevelopment. Priorities for new research mightinclude studies of:

• Shaped catalysts and combinedcatalyst/reactor designs. For example,preparation, characterization, testing, andmodeling of shaped catalysts including coatedceramic or metal monolithic or skeletal formsor unique extruded forms with an emphasison heat and mass transport properties andexperimental validation of heat and masstransfer models.

• Development and validation ofreactor/catalyst models. For example,development of comprehensive robust 1D and 2D models that include reactionthermodynamics and kinetics, pore diffusion,heat transfer, pressure drop for steamreforming, methanol and FT synthesis and


A NEW PARADIGM IN SYNGAS CONVERSION PLANT DESIGN:I N T E G R AT E D, C O M PA C T, MODULAR REACTORS AND PLANTS

WITH STRUCTURED CATA LYS T S

During the past 25 years, improvements in Fischer-

Tropsch (FT) catalyst and reactor technologies have

enabled 30-40% improvement in the economics for

large gas-to-liquids (GTL) plants leading to an

estimated capital cost of around $30,000/bbl/d for a

10,000 bb/d plant. However, due to the wide

geographical dispersion of biomass or biogas sources,

conversion to liquid fuels is logistically favored in

small plants (500 to 1,000 bbl/d) close to regional

sources of the feedstocks. Scaling by the six-tenths

rule, the capital cost of a 500 bbl/day GTL plant

would be $180,000/bbl/d or $90 million, an

unacceptably high figure. Two important decade-old

trends, process intensification or miniaturization

based on integrated catalyst/reactor design [Nehlsen

et al, 2007] and modular plant construction [Shah,

2007] have considerable potential for reducing the

costs of small syngas-conversion plants 3-5 fold.

1. PROCESS INTENSIFICATION. Development of

process-intensified reactor technologies or ‘micro-

reactors’ promises to revolutionize many aspects of

syngas conversion including: catalyst, reactor and

process designs; catalyst and reactor manufacturing;

and chemical production/processing [Ehrfeld, 2002;

Matlosz and Commenge, 2002; Schouten et al., 2002;

Sharma, 2002; Tonkovich et al., 2004]. Such reactors

could be produced using high-throughput, low-cost

fabrication methods such as those developed in the

semiconductor industry. Development of thermally

efficient, high productivity, compact reactors is

possible for highly exothermic or endothermic

reactions due to more effective heat and mass

transfer in micro-channel reactors. Indeed, overall

reactor size can be reduced by 1-2 orders of

magnitude in water-gas-shift, steam reforming, and

FT processes. Developing compact reactors typically

relies on: (1) integrated design of catalyst and

reactor [Nehlsen et al, 2007]; (2) incorporation of

new catalysts forms, e.g. “shaped catalysts” such as

coated ceramic or metal monoliths or skeletal

“sponges”; and (3) targeting improvements in the

limiting technical and economic factors (e.g. heat

and mass transport or pressure drop).

2. MODULAR DESIGN. Construction of a plant using

truly modular mini-plants for BTL, CTL, and GTL

could lead to further economic efficiency. While

modular design principles are well known [Shah,

2007], they have not been applied widely in energy

or chemical industries. Truly modular, mass-

manufactured plants could revolutionize

manufacturing and refining of products and

substantially reduce construction and operating

costs of large and small plants. The ideal module is a

finished, pre-tested, standardized system which

performs a single unit operation (e.g. syngas

production, FT reaction, distillation, or

hydrocracking) and could in principle be substituted

interchangeably in any number of processes or could

be combined in parallel with identical modules to

enlarge a plant. Substantial savings in capital cost

and construction time can be realized due to

substantial reduction in equipment, piping, and

structural designs; on-site fabrication and

construction; and plant shakedown. The cost of

updating or enlarging plants can also be

substantially reduced.

The promise of these new step-out technologies can

only be realized through additional research and

development. The application of process

intensification is illustrated by recent developments

in FT catalyst/reactor technologies [Nehlsen et al.,

2007; Bartholomew, 2007] based on (1) heat removal

management and (2) incorporation of small catalyst



117

the validation of these models using pilotplant data obtained by industrial partners.Computer codes should be developed in user-friendly platforms such as Visual basic/Excel.Development of microkinetic models wouldalso be a high priority.

• Process design, synthesis, and simulation.Engineering and economic studies of processdesign, synthesis, and simulation using processintensified, modular technologies.

4.6.4 Molecular Computat i o nOne way to tackle key issues relating tomechanistic routes, modeling, and catalyst design isthrough the use of electronic structure-basedtheoretical methods. These methods, especiallydensity functional theory (DFT), have reached alevel of sophistication where they can be used toexplain diverse surface phenomena and describecomplete catalytic reactions. When combined withmodern surface characterization techniques, theorycan now afford us the ability to pinpoint the originof catalytic activity and, thus, open up the possibilityof rational catalyst design [Lauritsen et al., 2004;Zhang et al., 2005]. Syngas conversion technologiescan potentially benefit a great deal from theory.

Nonetheless, one needs to be aware of thepractical challenges facing the theoretical approachto heterogeneous catalysis right now. Chief amongthem is the difficulty in capturing the couplingbetween the chemical environment, the nature ofthe catalytic material, and the reaction mechanism.This coupling can lead to the dynamictransformation of the catalytic material andreaction mechanism during operation, often inunexpected ways.

pellets on the inside or outside of small tubes or

channels. Some advantages of this approach are

illustrated in Figure S-4.3, showing heat transfer

coefficient (i.e., normalized rate) plotted against

catalyst productivity for conventional technologies

and integrated catalyst/reactor technology. Heat

transfer in a slurry bubble column reactor (SBCR) is

highly efficient and enhanced by its well-mixed

characteristic, a factor that also reduces reaction

rate and productivity. On the other hand, a

conventional fixed-bed reactor (FBR) is characterized

by high reaction rates and high productivity but

poor heat transfer rate. The new integrated compact

catalyst/reactor technology facilitates both efficient

heat transfer and high productivity (see Figure S-

4.3), while also eliminating the catalyst-product

separation problem.

0

500

1000

1500

2000

2500

3000

3500

4000

0 200 400 600 800 1000 1200 1400

Catalyst Productivity, kg/m3-h

Hea

t-T

ra

nsf

er C

oeff

icie

nt,

W/m

2-K

Slurry Bubble Column Reactor Integrated Small-Channel

Catalyst/FBR

Conventional

Multi-Tube FBR

F i g u re S-4.3 Performance ofc o nventional and integ r at e d ,

compact FT reactors [ B a rt h o l o m ew, 2 0 0 7 ] .


temperature and pressure, generates various oxideand carbide phases or perhaps even mixtures withsupport materials, which likely possess complexstructures [Pentcheva et al., 2005] (many yet to beelucidated) and disparate catalytic properties.Considering how complicated the FTS reactionmechanism is itself, one can see how this addedlevel of complexity can potentially make thetheoretician’s task much harder. Therefore, at thecurrent state of the art, sustained, concentrated,and smart efforts will be required, most likelytightly integrated with experiments, in order tocomprehensively understand the catalyticproperties of these seemingly mundane materialsand develop an efficient methodology in theprocess.

At the same time, theory can help tackle some ofthe technological problems that have beenidentified in this thrust area now. For instance, thechain growth mechanism and kinetics on Fe andCo, the factors that limit selectivity to ethanol andhigher alcohols on transition metals, the effect ofpromoters, all remain to be better understood andmay well be within the reach of simplified modelsthat are expertly constructed and from which

F i g u re 4.5 Fe distributions in two fluorescence tomographic sections thro u g hFT wax showing the variability of catalyst concentration [Jones et al., 2 0 0 5 ] .

As a simple example, the active phases of severaltransition metal oxidation catalysts have recentlybeen identified as oxidic and not metallic. They arestabilized by the reaction conditions, and someexist only on the surface [Bocquet et al., 2003;Ackermann et al., 2005]. The deployment of basemetals such as Fe and Co, in an oxygen- andcarbon-rich environment under elevated

F i g u re 4.4: " E l e c t ron density plot ofthe transition state of ethy l e n e

( C2H4) formation from twom e t hylene groups (CH2) on the

Fe(110) surface. C o u rtesy of Pro f .Manos Mav r i k a k i s , U n i versity of

W i s c o n s i n - M a d i s o n . "



119

insights of practical value could be generated.Therefore, the possibility of theory making a moreimmediate impact on syngas conversion must notbe overlooked, and efforts along these lines shouldbe encouraged and fostered.

4 . 6 . 5 N ew Experimental M e t h o d o l og i e s

In order to address many of the issues raised, newor modified ex-situ and in-situ/operando experimentalmethodologies are required, such as micro-X-rayfluorescence (see Figure 4.5), to allow us todevelop more understanding of catalystcomposition and structure during reaction and thenature of the reaction at the site level.

4 . 7 R E C O M M E N DAT I O N S

The following is a summary of some of theresearch needed in order to move the conversionof biomass to liquid fuels via biomass gasificationtowards commercialization.

• Development of more efficient and lessexpensive biomass gasification processes iscrucial, especially given the smaller scale likelyto be required for BTL plants.

• Better analytical techniques are needed for thecharacterization of low levels of syngascontaminants.

• There is a need for better characterization ofthe composition of biomass feedstocksincluding the concentration of catalyst poisonprecursors for predictive purposes.

• Impurities are feed dependent. This may meanthat different catalyst formulations must beused for biogas clean-up from differentfeedstocks. Research is needed to understandthe optimal catalyst formulation for differentimpurities and concentrations.

• Fundamentals of reaction on clean-up catalystsneed to be studied using a combination ofcomputational chemistry and in operandotechniques for removal of

– H2S– Ammonia– Tars– Other contaminants

• Syngas needs to have < 60 ppb of H2S, arsine,HCL, HCN, NH3, particulates, Se, Hg, alkali,and P for FTS.– Some impurities may greatly interfere

with catalytic removal of other impurities.Both computation and modelexperimental studies are needed in this area.

– Can a one-stage biogas clean-up processbe achieved that meets all requirementsor are 2-3 stages required?

• There is a need to look at lower temperatureclean-up processes such as the venturiscrubber. Much can be learned from coalgasification clean-up studies, but thisknowledge will need to be extended tobiomass gasification.

• Investigate sorbents and membranes for hightemperature CO2 removal. CO2 may need tobe removed from syngas for catalyst orreaction reasons.

• Better characterization and understanding ofactive sites for CO hydrogenation and chaingrowth or termination are needed in order tobetter design new catalysts that are active,selective, and stable for biomass syngasconversion.

• Increase use of computation for multi-scalemodeling for catalyst design and mechanismdelineation. Use computation to improveactivity and/or selectivity of catalysts.


• Develop better mechanistic understanding ofand microkinetic models for FTS and higheralcohol synthesis.

• Improve FTS catalysts and product upgradingmethodologies for potentially newformulations of diesel and jet fuel based onnew insights gained from experiment andcomputation.

• Design higher-yield catalysts for ethanolsynthesis.

• Develop more understanding of catalystdeactivation and regeneration issues in syngasconversion.

• Attrition of currently available high Fe FTScatalysts in an SBCR results in plugging,fouling, difficulty in separating the catalystfrom the wax product, and catalyst loss.There is a need for more research into thedesign of attrition resistant catalysts.

• Develop economical reactor technologies forsyngas conversion based on processintensification, catalyst/reactor integration, andheat management.

• Develop reactor/catalyst design strategies thatenable varying product slates in FTS andhigher alcohol synthesis

• Investigate the use of structured catalysts andprocess intensification

• Develop better catalyst-reactor models.

4 . 8 REFERENCES

Abatzoglou, N.; Barker, N.; Hasler, P.; Knoef, H.;Biomass Bioenergy, 18 (2000) 5.

Abu El-Rub, Z.; Bramer, E.A.; and Brem, G.; Ind. Eng.Chem. Res. 43 (2004) 6911.

Ackermann, M. D.; Pedersen,T. M.; Hendriksen, B. L.M.; Robach, O.; Bobaru, S. C.; Popa, I.; Quiros, C.;Kim, H.; Hammer, B.; Ferrer, S.; and Frenken, J.W.M.; Physical Review Letters, 95 (2005) 255505.

Agrawal, R., Singh, N. R., Ribeiro, F. H. and Delgass,W. N. (2007) Proc. Natl.Acad. Sci. USA 104, 4828-4833.

Artyukh,Yu. N.; Rusov, M.T.; Strel'tsov, O.A.;Kinetika i Kataliz 4 (1963) 299.

Asadullah, M.; Miyazawa,T.; Ito, S.; Kunimori, K.;Yamada, M.; and Tomishige, K.;Appl. Catal.A: Gen.267 (2004) 95.

Aznar, M.P.; Caballero, M.A.; Gil, J.; Martin, J.A.; andCorella, J.; Ind. Eng. Chem. Res. 37 (1998) 2668-2680.

Bartholomew, C.H.; and Farrauto, R.J, Fundamentalsof Industrial Catalytic Processes,Wiley, 2006.

Bartholomew, C.H.;“Preliminary Design of aCompact, Small-Tube,Tube-Shell Fixed-Bed Reactorfor Fischer-Tropsch,” Unpublished study, July, 2007.

Beenackers,A.A.C.M.;Van Swaaij,W.P.M. (1984)."Gasification of Biomass, a State of the Art Review,"in Thermochemical Processing of Biomass,Bridgwater,A.V., Ed., London, UK: Butterworths, pp.91-136.

Bera, Parthasarathi; Hegde, M. S.; Indian J. Chem.,Section A 41A (8) (2002) 1554-1561.

Berg, M.;Vriesman, P.; Heginuz, E.; Sjostrom, K.; andEspenas, B.-G.; in Progress in ThermochemicalBiomass Conversion, Blackwell Science Ltd.,Oxford, 2001. Pp. 322-32.Biomass Technology World,www.btgworld.com/technologies/tar-removal.html,2004.



121

Bocquet, M.-L.; Sautet, P.; Cerda, J.; Carlisle, C. I. ;Webb, M. J., and King, D.A.; Journal of the AmericanChemical Society, 125 (2003) 3119-3125.

Caballero, M.A.; Corella, J.;Aznar, M.-P.; and Gil, J.;Ind. Eng. Chem. Res. 39 (2000) 1143-1154.

Chambers,Allan;Yoshii,Yasuo; Inada,Tooru;Miyamoto,Tomohiko; Can. J. Chem. Eng. 74(6)(1996) 929-934.

Chen, G., PhD Thesis, Royal Institute of Technology,Stockholm, Sweden, 1998.

Choi, Jeong-Gil; Jung, Moon-Ki; Choi, Saemin; Park,Tae-Keun; Kuk, Il Hiun;Yoo, Jae; Park, Hyun Soo; Lee,Han-Soo;Ahn, Do-Hee; and Chung, Hongsuk;Bulletin Chem. Soc. Japan 70(5) (1997) 993-996.

Cholach,A. R.; Sobyanin,V.A.; and Gorodetskii,V.V.;Reaction Kinetics Catal. Lett.18(3-4) (1981) 391-6.

de Jong,W.; Unal, O.;Andries, J.; Hein, K.R.G.; andSpliethoff, H.; Biomass Bioenergy 25 (2003) 59-83.

Dinkov, Shishman; and Lazarov, Dobri; Godishnik naSofiiskiya Universitet Sv. Kliment Okhridski,Khimicheski Fakultet 85 (1993) 169-74.

Dittrick, P.; Oil & Gas Journal 105 (8) (2007) 24-25.

Dry, M.E.;“Fischer-Tropsch Synthesis-Industrial”, inEncyclopedia of Catalysis, ed. I.T. Horvath. JohnWiley & Sons, 2003, vol 3, pp. 347–403.

Ehrfeld,W.; “Design Guidelines and ManufacturingMethods for Microreaction Devices,” CHIMIA, 56(2002) 598–604.

Fleisch,T.; McCarthy, C.; Basu,A.; Udovich, C.;Charbonneau, P.; Slodowske,W.; Mikkelsen, S.; andMcCandless, J.; Society of Automotive Engineers,1995 (950061).

Friedlander,A. G.; Courty, P.R.; Montarnal, R. E.; J.Catal. 48(1-3) (1977) 312-21.

Gangwal, S.K.;Turk, B.S.; and Gupta, R.P.;"Development of Fluidized-Bed Sorbent forDesulfurization of ChevronTexaco Quench GasifierSyngas", Pittsburgh Coal Conference Paper,September 2002.

Gangwal, S.K.; Portzer, J.W.; and Jothimurugesan, K.;"Simultaneous Removal of H2S and NH3 from Coal Gas", Final Report, Contract No. DE-AC21-92MC29011.

Gangwal, S.K.; Gupta, R.; Portzer, J.W.; and Turk, B.S.;"Catalytic Ammonia Decomposition for CoalDerived Gases",Topical Report, Contract No. DE-AC21-92MC29011, July 1997.

Gates, B.C., Katzer, J.R., and Schuit, G.C.A.;Chemistry of Catalytic Processes, McGraw-HillBook Company, 1979. P. 17.

Geerlings, J.J.C.;Wilson, J.H.; Kramer, G.J.; Kuipers,H.P.C.E.; Hoek,A.; Huisman, H.M.;,Appl. Cat.A 186(1999) 27-40.

Gil, J.; Caballero, M.A.; Martin, J.A.;Aznar, M.-P.; andCorella, J. ; I&EC Res. 38 (1999) 4226-4235.

Grosman, Monica; and Loeffler, Daniel G.; J. Catal.80(1) (1983) 188-93.

Hepola, J.; and Simell, P.;Appl. Catal. B 14, 287(1997).

Hoek,A.; and Kerste, L.B.J.M.; Studies In SurfaceScience And Catalysis 147 (2004) 25-30.

Hos, J.J.; Groeneveld, M.J. (1987). “BiomassGasification.” In Biomass, Hall, D.O.; Overend, R.P.,eds. Chichester, UK:John Wiley & Sons, pp. 237-255.


Hu, J. L.;Wang,Y., Cao, C.S.; Elliott, D.C.; Stevens,D.J.; and White, J.F.; Catal.Today 120(1) (2007) 90-95.

Jacquin, M.; Jones, D.J.; Roziere, J.;Albertazzi, S.;Vaccari,A.; Lenarda, M.; Storaro, L.; and Ganzerla,R.;App. Cat.A 251 (2003) 131-141.

Jager, B., Studies In Surface Science And Catalysis107 (1997) 219-224.

Johnson, J.; “Ethanol—Is it Worth it?”, ChemicalEng.News, (Jan. 1, 2007) 19-21.

Jones, K.W.; Feng, H.; Lanzirotti,A.; Mahajan, D.;Topics in Catalysis 32 (2005) 263.

Jothimurugesan, K.; and Gangwal, S.;Adv. Environ.Research, 2(1), 116(1998).

Kagami, Setsuko;Takaharu Onishi; and KenziTamaru.; J. Chem. Soc., Faraday Trans. 1 80(1)(1984) 29-35.

Kochloefl, K., in Handbook of HeterogeneousCatalysis,Vol.4, ed. by G. Ertl, H. Knozinger, and J.Weitkamp,Wiley VCH,Weihiem, Germany, 1997.Pp.1819-1830.

Kong, O. Dapeng;Tsubouchi, Naoto; and Ohtsuka,Yasuo; Sekitan Kagaku Kaigi Happyo Ronbunshu38th (2001) 311-314.

Krishman, G.N., et al., "Study of NH3 Removal inCoal Gasification Processes", Final Report,Contract No. DE-AC21-86MC23087.

Lauritsen, J.V.; Bollinger, M.V.; Laegsgaard, E.;Jacobsen, K.W.; Nørskov, J. K.; Clausen, B.S.;Topsøe,H.; Besenbacher, F. J. Catal., 221 (2004) 510-522.

Leppalahti, J.; and Koljonen,T.; Fuel Proces.Tech. 43(1995) 1-45.

LeViness, S.C.; Mart, C.J.; Behrmann,W.C.; Hsia,S.J.; and Neskora, D.; “Slurry HydrocarbonSynthesis Process with Increased Catalyst Life,”WIPO Pat. 98/50487 (Nov. 12, 1998;Assigned toExxon Res. & Engr. Co).

Matlosz, M.; and Commenge, J.M.;“From ProcessMiniaturization to Structured Multiscale Design:The Innovative, High-Performance ChemicalReactors of Tomorrow,” CHIMIA, 56 (2002)654–656.

Nakatsuji,Tadao; Nagano, Kazuhiko; and Ikeda,Shiroji; Japanese Pat. 8332388 (1996).

Narvaez, I.; and Orio;A., Ind. Eng. Chem. Res., 35,2110 (1996).

Nehlsen, J.; Mukherjee, M.; and Porcelli, R.V.;Chemical Engineering Progress, February 2007,31-38.

Ohtsuka,Yasuo; Xu, Chunbao; Kong, Dapeng; andTsubouchi, Naoto; Fuel 83(6) (2004) 685-692.

Paisley, M.A.; Irving, J.M.; and Overend, R.P.; "APromising Power Option – The Ferco SilvagasBiomass Gasificaiton Process – OperatingExperience at the Burlington Gasifier", Proceedingsof ASME Turbo Expo 2001, June 4-7, 2001, NewOrleans.

Papapolymerou, George; and Bontozoglou,Vasilis; JMolecular Catal.A 120(1-3) (1997) 165-171.

Pentcheva, R.;Wendler, F.; Meyerheim, H. L.; Moritz,W.; Jedrecy, N.; and Scheffler, M. Physical ReviewLetters, 94 (2005) 126101.

Perlack, R. D.,Wright, L. L.,Turhollow,A. F., Graham,R. L., Stokes, B. J. & Erbach, D. C. (2005) Biomass asFeedstock for a Bioenergy and BioproductsIndustry:The Technical Feasibility of a Billion-TonAnnual Supply.



123

Phillips, J.G.; and Reader, G.T.; "The Use of DME asa Transportation fuel - A Canadian Perspective."ASME Fall Technical Conference, 1998. 31-3 (98-ICE-152): p. 65-71.

Reed,T.B.; and Guur, S.; "A survey of BiomassGasification 200 – Gasifier Projects andManufactures around the World", DOE ContractNo. DE-AC36-83CH10093, Subcontract No. ECG-6-16604-01(BEGF).

Robinson, K.K.; and Tatterson, D.F.; Oil and GasJournal, February 26, 2007, 20-31.

Schouten, J.C.; Rebrov, E.V.; and de Croon, M.H.J.M.;“Miniaturization of Heterogeneous CatalyticReactors: Prospects for New Developments inCatalysis and Process Engineering,” CHIMIA, 56(2002) 627–635.

Shah, S.;“Modular Mini-Plants:A New Paradigm,”Chem. Eng. Prog., March 2007, 36-41.

Sharma, M.M.;“Strategies of Conducting Reactionson a Small Scale. Selectivity Engineering andProcess Intensification,” Pure and AppliedChemistry, 74 (2002) 2265–2269.

Sheu, Shie-Ping; Karge, Hellmut G.; and Schlogl,Robert; J. Catal. 168(2) (1997) 278-291.

Simell, P.A.; Hepola, J.O.; and Krause,A.O.I.; Fuel76(12) (1997) 1117-1127.

Sorenson, S.C.; and Mikkelsen, S.; "Performance andEmissions of a DI Diesel Engine Fuelled with NeatDimethyl Ether." Society of Automotive Engineers,1995 (950064).

Sugishima, Noboru; Hagi, Mitsuharu; and Kobayashi,Motonobu; Japanese Pat. 07289897 A2 19951107Heisei (1995).

Tonkovich,A.Y.; Zilka, J.L.; LaMont, M.J.;Wang,Y.; andWegeng, R.S.;“Microchannel Reactors for FuelProcessing Applications. I.Water Gas Shift Reactor,”Chem. Eng. Sci., 54 (1999) 2947–2951.

Tonkovich,A.Y.; Perry, S.;Wang,Y.; Qiu, D.; LaPlante,T.; and Rogers,W.A.;“Microchannel ProcessTechnology for Compact Methane SteamReforming,” Chem.Eng.Sci., 59 (2004) 4819-4824.

Torres,W.; Pansare, S. S.; and Goodwin, J. G., Jr.;Catalysis Reviews: Science & Engineering, in press(2007).

Turner, M. L.; Marsih, N.; Mann, B. E.; Quyoum, R.;Long, H. C.; and Maitlis, P. M.;“Investigations by 13CNMR spectroscopy of ethene-initiated catalyticCO hydrogenation”, J.Amer. Chem. Soc., 124(2002) 10456-10472.

Verbeek, R.; and Van der Weide, J.; "GlobalAssessment of Dimethyl Ether : Comparison withOther Fuels." Society of Automotive Engineers,1997 (971607).

Wakai, K.; Nishida, K.;Yoshizaki,T.; and Hiroyasu, H.;"Ignition Delays of DME and Diesel Fuel SpraysInjected by a DI Diesel Injector." Society ofAutomotive Engineers, 1999 (1999-01-3600).

Zhang, J.;Vukmirovic, M. B.; Sasaki, K.; Nilekar,A. U.;and Adzic, R. R.; J.Amer. Chem. Soc., 127 (2005)12480-12481.

Zhou, J.; and Masutaui, S.M.; Ind. Eng. Chem. Res., 39(2000) 626.


5 . P rocess Engineering & DesignOV E RV I E W

Process optimization is critical to the success ofscaled-up biofuels production operations. Thisrequires the combination of process modelsderived from known feedstock and reactantproperties and application of property predictionmethods in tandem with laboratory-basedresearch. In this Thrust area we articulate the roleof process analysis for optimization of the biofuelsproduction process.

5.1 INTRO D U C T I O N

Like all large-volume petrochemical productionprocesses, the processes for biofuels productioncan be designed and optimized using advancedanalytical and modeling tools. This requires acombination of models implemented along withdatabases and development of property predictionmethods assembled in parallel with laboratoryresearch. Because this is a low-margin, high-volume business, optimization is critical for thesuccess of chemical or biological processes on alarge scale. Unlike the early days in thedevelopment of production processes forpetroleum-based fuels and chemicals, there arenow many excellent computer-aided design andoptimization tools available.

The application of design and optimization tools tounits and processes for the biofuels productiondemands new data on the physical, chemical andbiological molecules, mixtures, and steps in theprocesses. Many of these are described anddiscussed in other sections of the workshop.Furthermore, there are typically many options forthe arrangement of process units, as well as for theprocess conditions to be used. Consequently, acomprehensive and complete database sufficientfor the detailed design of all alternatives is notlikely to be developed within any reasonable

timeframe. Conceptual design is useful alternativeto bridge the gap between laboratory research onnew process steps and the detailed biofuelsproduction process design and optimization.Biofuels process plants must also be integratedwith models for supply and distribution of biofuelsand including life-cycle analysis.

5.2 PROCESS A N A LYS I S

In the context of lignocellulosic biofuels, theobjective of process analysis is to evaluate biofuelsand biochemical processes to determine materialsand energy balances, product yields, equipmentsize, capital and operating costs, environmentalimpacts, and cost of production. Process analysis isthe application of scientific methods to therecognition of problems or definition of processes(Himmelblau and Bischoff 1968) and thedevelopment of procedures for their solution; itinvolves an examination of the process, alternativeprocesses, and economics.The steps in processanalysis are 1) mathematical specification of theproblem or process, 2) detailed analysis to obtainmathematical models describing the problem orprocess, and 3) synthesis and presentation ofresults to ensure full understanding of the problemor process. Process analysis is frequently referredto as technoeconomic analysis.

D e f i n i t i o n sA number of definitions are useful in discussingprocess analysis and design:

A PROCESS is the actual series of operations ortreatment of materials of interest in the analysis.

A MODEL is the mathematical representation ofthe real process.



5.1 Introduction

5.2 Process Analysisa. Definitionsb. Disciplines and Applicationsc. Process Modeling Toolsd. Examples

5.3 Priorities for Research &Development

5.4 Summary

5.5 References


Boateng, Brown, Chueng, Fjare,Gaffney, Lucia, Mahajan, Malone,Phillips, Rabil,Turn,Vrana,Wyman

A SYSTEM is the group of process elements thatare tied together by common flows of materialsand/or information. It is always important tounderstand that the model and system are arepresentation of the physical process, and notalways an exact replication of the process. Inmany cases, the process representation permits amathematical solution where the exactreplication does not.

A PARAMETER is a property of the process orits environment that can be assigned arbitrarynumerical values. It is also a constant orcoefficient in an equation.

SIMULATION is the study of a system or itsparts by manipulation of the mathematical model.

D i s c i p l i n e sMany scientific disciplines are used in processanalysis, and include thermodynamics, transportphenomena, phase equilibria, mass transfer, heattransfer, chemical reaction kinetics, catalysis,mathematics such as statistics and multivariateanalysis, and economics.The applications of processanalysis include reactor design, separations,materials transfer and handling, life cycleassessment, and net energy analysis.

Examples of Process Modeling To o l sA number of processing modeling tools areavailable for process analysis. Examples of thesestools include:

Interpretation of R&D data – Unscrambler, Matlab

Process modeling and design – Chemkin (kinetics), Fluent (computational

fluid dynamics), FACT (thermodynamics andphase equilibria)

Material and energy balances– ASPEN

Process cost estimation – ASPEN Icarus Process Estimator

Data analysis, economics etc.– EXCEL

Risk analysis– Crystal Ball, At-Risk

Dynamic modeling – STELLA

Life cycle assessment – TEAM, SimaPro, GREET


5.3 PRIORITIES FOR RESEARCH &D E V E L O P M E N T

Based on the presentations for each thrust in thisworkshop, the participants in Thrust 5 developed alist of areas that are important for the processengineering and design needed to overcome thechemical and engineering barriers to advancingbiofuels technology. Many more detailed individualtopics were considered and discussed by theparticipants.These topics were then rank-orderedand grouped into higher-level areas for researchand development. It is important to note that all ofthe goals in the final list were consideredimportant.The priority is primarily dictated by theprecedence order for developing and implementingthe new chemical and physical steps considered inthis workshop. For example, detailed optimizationstudies cannot proceed without a proper basis ofchemical and physical properties. Table 5.1 listseach of these goals, grouped approximately intothree priority groups.

A brief discussion of each area is given in the following subsections:

5.3.1 Analytical MethodsRapid and reliable methods are needed tocharacterize not only biomass feeds, but alsointermediate streams in the process as well as theproducts and byproducts. In addition to watercontent, measurements of the bulk density, ash, andtrace elements would be very useful.

5.3.2 Physical Pro p e rt i e sQuantitative information for many physicalproperties is required for process design andoptimization. This should include data fromlaboratory studies as well as physical propertymodels. Physical property models that can build ona small set of data to describe the complexmaterials often encountered seem to be essential.Improved information for heats of reaction andphase equilibrium will be needed for many of theprocess engineering and design activities necessaryto reach the goals described here.

5.3.3 KineticsA significantly expanded and quantitativedescription of chemical kinetics is needed forreactor design and potentially for studies of hybridreaction-separation methods. It is especiallyimportant to have reduced-order models forcomplex reaction networks. These should besufficiently detailed to describe, not only the rateof the primary reaction(s), but also the productdistribution. In many cases, studies and laboratorytests are required to assess and remove transporteffects from data or empirical models.

Ta ble 5.1 Goals and Priority Gro u p sIdentified for Process Engineering & Design

Area Priority

1 Analytical Methods A

2 Physical Properties A

3 Kinetics A

4 Conceptual Design Methods A

5 Life Cycle Assessment A

6 Separations A

7 Economic Optimization Studies B

8 Pretreatments B

9 Educational Materials B

10 Hybrid Reaction-Separations B

11 Byproduct & Co-product & Markets B

12 Robustness & Process Control C

13 Reactor Design Models C

14 Pollution Prevention & Treatment C

15 Cost of Promoters, Catalysts & Solvents C


5 . P rocess Engineering & Design

127

Figure 5.1 is a schematic of the relationshipbetween conceptual design and the more detailedsteps in the implementation of a process, in termsof the Costs Committed to and the Money Spentto support the activities required to implement aproject. The costs committed represent the bestcase that can be accomplished with the choice ofchemical route, catalyst, conditions, separation andpurification, etc. The level of activity and peopleinvolved in the Conceptual Design stage is typicallysmall compared to the later stages. However, thedecisions made at the conceptual design stagedictate the ultimate performance and costs of theproject that can be achieved assuming that detailedengineering design and optimization are usedsubsequently. That is, even the best downstreamengineering can only reduce the costs to a certain

level (i.e., the Costs Committed).

Process Engineering and Design isoften considered after the basicresearch and development work toproduce a science and technologybasis for a process. However,experience shows that conceptualdesign activities conducted inparallel with the process researchand development often yield muchbetter processes and, in fact, willoften suggest new types andconditions for experiments oncatalysts, process conditions, etc.

F i g u re 5.1 Impact and Phases of Process andP roject Engineering

E a rly stage decisions about Conceptual Design, such as chemicalro u t e , c at a l y s t , and separation and purification technology invo l vere l at i vely few people and budget, but have a gre at impact on thep roject because they dictate the lowest cost that can be at t a i n e d .

Once the design concept is ava i l a bl e , p rocess development ando p t i m i z ation are essential to achieving these lowest costs in the

E n g i n e e r i n g , P ro c u rement and Construction stages.

5.3.4 Conceptual Design MethodsConceptual design refers to the engineeringdecision-making that is required to conceive aprocess from information on the products, feeds,and chemical and physical steps proposed as a basis[Douglas 1989]. There are frequently manyprocesses alternatives that can be chosen andmany design variables to specify. Conceptualdesign approaches practically always include bothassumptions and heuristics along with numericalmodeling and optimization techniques.The extentto which each of these is useful depends stronglyon the databases of physical and chemicalproperties and the level of detail in the processdescriptions. Whatever the combination, theeffective use of conceptual design methods in theearly stages of process design can have a large impact.


Life cycle assessment is a systematic evaluation of

the environmental impacts and resource utilization

associated with the production of biofuels, from

initial growth of biomass feedstocks (i.e. plants and

trees) all the way through the various processing

steps to end use of the fuels (Segundo, 2005a;

Segundo, 2005b). In other words, it is a “cradle to

grave” assessment of everything that goes into and

comes out of the production and subsequent

utilization of biofuels. Life cycle assessment is an

important tool for evaluating the impacts of

different biofuels production processes utilizing

various feedstock types. It provides a methodology

that can be used to account for all the energy and

resources consumed, as well as all the emissions and

waste generated, during the process of producing

and refining biofuels compared with the total energy

content of the finished fuels.

LIFE CYCLE A S S E S S M E N T

A comprehensive life cycle assessment includes at

least two components: 1) an inventory of raw

materials requirements, total energy production and

consumption, greenhouse gas emissions, water use

and discharge, nutrient return, element cycling, solid

waste and agricultural runoff, and release of

byproducts during the entire production cycle; and

2) an assessment of the economic, environmental,

and public health impacts of findings identified

during the inventory. A third component, the life

cycle improvement analysis, can be used as a tool to

identify opportunities to mitigate the impacts of the

production process (Svoboda 1995). Improvement

analysis may involve life cycle comparisons based on

changes in one or more production parameters, such

as: catalyst type; reaction conditions; raw material

usage; production scale; end use behavior; waste

management; or overall process design.

F i g u re S-5.1 Life Cycle Assessment for Biofuel Pro d u c t i o n(Oak Ridge Nationaly Laborat o ry ) .



129

5.3.5 Life Cycle A s s e s s m e n tThere is currently no validated database of lifecycle inventories, including water, nutrient return,total energy picture and element cycling suitable asa basis for decision-making for biofuelsdevelopment. Such a database would enablestudies of feedstock production and processingmodels, along with identifying and assessing policyissues and constraints. (See Sidebar).

5.3.6 Separat i o n sA systematic quantitative approach should bedeveloped to describe and explain physical andchemical property characteristics and differencesthat can be used as the basis for separations.Research and development for reduced energyseparation methods should be accelerated.These separation studies should include bothcontaminants and products in feed andintermediate streams. It is particularly importantto identify contaminants that can adversely impactcatalysts, as well as alternative catalysts that areless sensitive or insensitive to such compounds.Methods to separate catalysts from product orintermediate streams will be essential if the catalystis not immobilized.

5.3.7 Economic Optimization StudiesMany economic studies have been conducted foralternative biofuels routes. These should beupdated and made available in a comprehensiveform. A review of the existing studies, widelydisseminated, would be particularly useful andtimely. Gaps in the studies should also beidentified, especially for new technologies, and thestudies should be expanded as necessary to closethese gaps. Estimates of the optimum planttechnology choice, size scale, feed(s) and harvestingfootprint, would be very useful. Along with suchoptimization studies, an estimate of uncertainty inthe results to the assumptions and basis areimportant.

5.3.8 Pre t re at m e n t sMany pretreatment methods have been studied forbiofuels in general. But further work in support ofaqueous-phase processing is important. Improvedalternatives such as Organosolv and otherapproaches should be considered. Alternatives toexpensive enzymatic pretreatment are particularlyimportant.

5.3.9 Educational Mat e r i a l sChemists and engineers have the science andengineering fundamentals to contribute to thediscovery, invention, design and operation of abiofuels industry. A greater emphasis on theapplication of these fundamentals to addressparticular biofuels issues would be useful. Oneroute to the development of such educationalmaterials is to provide specific challenge problemsfor faculty to use in project-based classes. Forexample, the American Institute of ChemicalEngineers distributes a “Students Contest Problem”in a yearly competition that is often used in seniordesign courses. Student teams often provideexcellent solutions as a result of this competitionand this has potential not only as an educationalvehicle, but also to provide technical insights andsolutions.

5.3.10 Hybrid Reaction-Separation Methods & Process Intensificat i o n

Because the processing costs and sustainability forbiofuels are critical to implementing the technologysuccessfully, process intensification is an importantarea of potential impact. Jenck et al. [2004] discusssustainability and process intensification in thechemical industry. In fact, process intensification inthe chemical industry has found surprising successin the last two decades, especially in reactivedistillation as discussed recently by Harmsen[2007]. Some process intensification has beenrecognized as significant option for biofuelsproduction, e.g., Simultaneous Saccharificaiton and


Reactive distillation, sometimes called “catalytic

distillation” when a catalyst is present, combines one

or more chemical reaction steps with distillation of

the reacting mixture. Among many publications,

Harmsen [2007] gives a recent review of industrial

applications; conceptual design methods and

additional examples appear in Doherty and Malone

[2001].The best-studied and best-known example is

probably the Eastman Chemical Company process for

the production of methyl acetate [Agreda et al.

REACTIVE DISTILLAT I O N

1990]. The traditional process for methyl acetate

production used a separate catalytic reaction

followed by separation and purification in nine major

units as shown in Figure S-5.2. The appropriate

combination of chemical reaction with distillation

provides a process with a single major (hybrid) unit

shown in Figure S-5.3 that requires approximately

one fifth of the investment and consumes only one

fifth of the energy in the traditional process; see

Doherty and Malone [2001], pp. 7-12.

F i g u re S-5.2 Traditional Process of ReactionFo l l owed by Separation and Purification for the

P roduction of Methyl A c e t at e .Adapted from Siirola [1996].



131

F i g u re S-5.3. R e a c t i ve Distillat i o nP rocess for the Production of

M e t hyl A c e t at e .I nvestment and energy use are reduced to 1/5 the

values for the traditional process shown in theF i g u re . Adapted from A g reda and Pa rten

[1984] and A g reda et al. [ 1 9 9 0 ] .

The design, development and control of integrated

reaction-separation systems such as these are made

possible by sophisticated modeling tools, combined

with data on catalysts, equilibria and reaction rates.

Fermentation. Other combinations combiningreaction and separation have been considered, butmore can be done in this area.

5 . 3 . 1 1 B y p roduct and Co-Product I d e n t i f i c ation & Marke t s

Byproducts and potential co-products from thechemical and engineering steps discussed in thisworkshop will have a significant economic and

environmental impact. More should be done toidentify and develop applications for thesematerials. For example, there is currently nostandard for the incorporation of ash from biomassproduction into concrete.

5.3.12 Robustness & Process C o n t ro l

Particularly because of the solid feed streams tothe initial steps in biofuels production, and the factthat they are often variable in composition andsize, the process robustness to these variations,both chemical and mechanical, is an importantconsideration in process engineering and design. Asmall change in the feed stream should not cause asignificant upset in the process. So this variabilityshould be characterized early and included in theprocess design and optimization analysis.Appropriate process analytical methods (see 5.3.1)in support of process control are important andprocess control studies should be done early in thedesign activity to identify critical measurements.

5.3.13 Reactor Design ModelsInformation on the kinetics and physical propertiesdiscussed above, along with feed characteristics,should be integrated and used as the basis forreactor design models which include the capabilityto describe highly non-isothermal conditions, andthe presence of multiple phases. In addition to theinformation in 5.2.2 and 5.2.3, it will be necessaryto include transport effects in a manner that issuitable for scale-up from laboratory or pilot scaleto commercial operation. Some work has beendone in this area, but more attention is needed.

5.3.14 Pollution Prevention & Tre at m e n t

Pollution prevention and waste minimizationmethods should be integrated into process designstudies. New options for air pollution control,waste-water treatment and especially solids (e.g.,


5.4 SUMMARY

Processes for the production of biofuels can bedesigned and developed relying in part onadvanced modeling tools, databases, and propertyprediction methods.The development and use ofthese models for optimization from single plantthrough the supply chain, distribution and life-cycleanalysis is critical for the economic success ofbiofuels development. Many excellent softwaretools are available, but new models specific tocertain aspects of new chemical or biological stepswill also be needed. The application of thesemodels for the development of biofuels demandsnew data on the physical, chemical and biologicalmolecules, mixtures, and steps in the processes. Aneffective and efficient approach to biofuels processdevelopment demands that the models and thedata that drive them be developed in parallel and inclose coordination with laboratory studies. Oncedesigns for one or more good candidate processesare understood, these must also be integrated intosystems to support decisions concerning thesupply and distribution of biofuels.

There are typically many options for thearrangement and selection of process units, as wellas for the process conditions to be used.Consequently, a comprehensive and completedatabase sufficient for the detailed design of allalternatives is not likely to be developed within anyreasonable timeframe. In this context, conceptualdesign is useful to bridge the gap betweenlaboratory research on new process steps and thedetailed biofuels production process design andoptimization.A variety of open questions importantto the design and engineering of processes for theproduction of biofuels were identified in thisworkshop. Some of these take a naturalprecedence over others, and the consensus amongacademic and industry experts at the workshop issummarized in Table 5-1, on page 126.

ash) should be explored. It was pointed out thatthe fate of byproducts is also a criticalconsideration in Life Cycle Assessment.

5.3.15 Cost of Pro m o t e r s , C atalysts & Solve n t s

Process engineering and Design studies, especiallyconceptual design (See 5.3.4) can guide theselection of promoters, catalysts, and solvents foraqueous phase processing. For example, aquantitative estimate for the benefits ofimmobilizing catalysts would inform ongoinglaboratory studies. The relative costs and benefitsof solvent alternatives such as ionic liquids, ethanol,butanol, acetone, ethyl acetate, gasoline or bio-diesel, should be similarly useful.



133

5.5 REFERENCES

Agreda,V. H. and Parten, L. R., US Patent 4,435,595to Eastman Kodak Company (1984).

Agreda,V. H., Partin, L. R., Heise,W. H., ChemicalEngineering Progress, 86(2), 40-46 (1990).

Doherty, M. F. and Malone, M. F., Conceptual Designof Distillation Systems, Ch. 10, McGraw-Hill (2001).

Douglas, J. M., Conceptual Design of ChemicalProcesses, McGraw-Hill (1989).

Harmsen, G. J., Chemical Engineering andProcessing, 46, 774-780 (2007).

Jenck, J. F.Agterberg, F. and Droescher, M. J., GreenChemistry, 6, 544-556 (2004).

Seungdo Kim, Bruce E. Dale Biomass andBioenergy, 28, 475-489, 2005.

Seungdo Kim, Bruce E. Dale Biomass andBioenergy, 29, 426-439, 2005.

Siirola, J. J.,Advances in Chemical Engineering, 23, 1-62 (1996).

Svoboda, S. Note on Life Cycle Assessment. 1995.National Pollution Prevention Center for HigherEducation. University of Michigan. Ann Arbor.http://www.umich.edu/~nppcpub/resources/compendia/CORPpdfs/CORPlca.pdf


OV E RV I E W

Exhaustive analysis of biomass reactants,intermediates, and products will provide afoundation of data from which biorefineryprocessing methods can be developed, tested, andoptimized. With these data as inputs, powerfulcomputational models will be used to illuminate thethermodynamic and economic efficiencies ofbiomass conversion reactions, identify prototypicalreactant biomolecules and catalysts, and contributeto the discovery and development of novelmultiphase biomass conversion reactors. Thisapproach is derived, to a large extent, from thelessons of the petrochemical industry learned overthe course of the last century. An importantdifference is the tremendous variation amongbiomass feedstocks, compared with thecomparatively uniform properties of petroleumfeeds. New analytical tools and techniquescombined with exceedingly powerful computationalmethods promise to accelerate the development ofbiorefinery processing technologies at a rate thatwould have been inconceivable just a fewdecades ago.

I N T RO D U C T I O N

The concept of a biorefinery is still in its infancy.The ultimate success of biorefineries will dependfirst on the identification and development of novelprocesses for the conversion of biomass feedstocksto valuable fuels and chemicals, and second on theintegration of these processes such that the

6 . C rosscutting Scientific Issuesbiorefinery minimizes the production ofundesirable by-products and waste heat. Arecurring theme of this cross-cutting area is theparallels drawn between the concept of a 21stCentury biorefinery and the advances inproduction efficiencies achieved through theconcerted and sustained efforts of thepetrochemical industry over the course of the last century.

Progress in the petroleum and petrochemicalindustries has been made possible in large part bythe development of extensive database of physicalproperties of reactants, products, andintermediates involved in the conversion offeedstocks to valuable fuels and chemicals andpowerful computational approaches used to modeltheir thermodynamic and transport propertiesduring various reactions and under differentconditions. This information has been vitalthroughout all levels of research and development.As occurred during the century-long developmentand optimization of petroleum refineries, thedesign, control, and optimization of biorefineryoperations will involve detailed process analysesand simulations, which will require extensiveknowledge of the physical properties of reactants,products, and intermediates. While there are manyparallels, the range of new tools and techniquesavailable today promises to accelerate thedevelopment of biorefinery operations at a ratethat would have been unattainable petrochemicaltechnology developers just a few decades ago.We begin this thrust with a detailed discussion of

O

OO

OO

OO

OOHO

HOH2C

OH

HOH2CHO

HO

OH

OH

HO

HOH2C HOH2C

OH

ll l

F i g u re 6.1 Cellulose is a crystalline polymer of glucose connected via beta linkages.



6.1 Analytical Database forBiomolecules

6.2 Thermodynamics

6.3 Chemical Reaction Engineering

6.4 Catalyst Engineering

6.5 Catalyst Characterization

6.6 Computational Catalysis

6.7 Recommendations

6.8 References


Scott Auerbach , Douglas Cameron,Curt Conner . James Dumesic, ManuelA. Francisco, John Hewgley,Christopher Jones,Alexander Katz,Harold Kung, Simona Marincean, RaulMiranada, Don Stevens, Bruce Vrana,Yang Wang, Conrad Zhang, Luca Zullo

the analytical and engineering challenges involvedwith thermochemical conversion of biomass tobiofuels. Because of the central role that catalystsplay in biofuels production, the overarchingscientific issues associated with catalyst engineeringand characterization are prominently featured.Many aspects of traditional fuel production havebeen illuminated through computational chemistry.Clearly, the widely applicable cross-cutting tool ofchemical modeling will increasingly be used to shedlight on the underlying chemistry of biofuelsproduction as well. However, the multi-phasenature of chemical biofuels processing is drivingthe need for new, more powerful computationalmethods. Through exhaustive analysis of biomassreactants, intermediates, and products – combinedwith judicious choice of model systems –development of new computational methods canhelp us to understand more about conventionalbiomass refining processes and rapidly developincreasingly efficient processes for the future.

6.1 A N A LY T I CAL DATABASE FORB I O M O L E C U L E S

A pillar of modern petroleum refining technologyis a comprehensive database that characterizes theindividual components of many crude oils and theproducts of their conversion. This database isvaluable for making predictions about thereactivities of various feedstocks and theproperties of their refining products; it is aresource worth billions of dollars to refiners whodepend on it to guide the optimization of theiroperations.

The creation of an efficient, economical fuelindustry based on biomass will require thedevelopment of a similar database for the hundredsof biomass-derived compounds. Such a databasedoes not currently exist, and its absence

significantly hinders the fundamental researchprograms that are exploring new biofuelsprocessing routes.

Of the three main components in cellulosicbiomass (cellulose, hemicellulose, and lignin),hemicellulose and lignin are the most difficult tocharacterize structurally. Cellulose is thecrystalline fraction (Figure 6.1) of biomass, which iscomposed entirely of beta-linked D-glucose unitsthat can be characterized by a variety oftechniques including X-ray diffraction methods,13C MAS NMR spectroscopy, and various otherspectroscopies. Lignin is unique in that it is apolyaromatic structure, composed ofphenolpropanoids units (Figure 6.2). Although the


which is both crystalline and composed of a singlemonosaccharides unit, hemicellulose exists inbiomass in an amorphous structure that containsseveral different C5 and C6 sugars, includingxylose, mannose, galactose, glucose, and arabinose.A further structural feature of hemicellulose is thepresence of numerous acetyl groups on many ofthe saccharides. Because hemicellulose isamorphous, characterization methods that rely oncrystallinity (such as X-ray diffraction) are notuseful. Furthermore, in common spectroscopies(i.e. 13C MAS NMR), the signatures of the differentsugar residues are essentially identical, making thecellulose and hemicellulose residues appearidentical. Only the presence of acetyl groups is auseful marker for hemicellulose in raw biomasssamples. Unfortunately, most mild treatments ofcellulosic biomass in aqueous media liberate theseacetyl groups, making them essentially useless inthe characterization of hemicellulose fractionsduring chemical processing. Currently, no facilemethods exist to monitor the degradation ofhemicellulose during chemical treatments, otherthan to track the formation of non-glucosemonosaccharides during the process. Therefore,the development of spectroscopic, or othermethods, expressly designed to interrogate thestructure of hemicellulose in solid biomass samplesis an important research target.

The complex mixtures of hydrocarbons in crudeoils, including even many of the heaviestcompounds, have been characterized precisely bymass spectrometry. Gas chromatography/massspectrometry has complemented this method, andmany useful results have been obtained withelement-specific detectors.

Mass spectrometric analysis is applicable to all butthe most intractable components of petroleum,such as asphaltenes. Thus, it can be expected thatthese techniques would be suitable for analysis of

HC

CH

OHC [CH2OH]

H3CO

O CH

CH2OH

CHOH

O CH

CH2OH

HC O

OCH3H2C

OCH

H3CO

O CH

CH2OH

CHOH

H3CO

OH OH

OCH3

C

[O C]

O

HC O

CH2OH H3CO

CHOH

HC

CH2OH

O

H3COO

CHOH

CH

H3CO

HO

CH2OH

OCH3

HCO

CH2

CHHC

H2C CHO

OCH3

OH

O

CH

HOH2C C

O

H

H3CO

CHOH

HC O

CH2OH

H3CO OCH3

CHOH

HC

CH2OH

O

OCH3

CHOH

CH

OH

CHO

H3CO HC O

CH2OH

CH2

CH2

CH3

CH2OH

proposed structural segment of lignin

OH

CH

CH

CH2OH

OH

CH

CH

CH2OH

OCH3H3CO

p-coumaryl alcohol

syringyl alcoho l

coniferyl alcohol

OH

CH

CH

CH2OH

OCH3

F i g u re 6.2 Lignin is polya ro m atic ins t ru c t u re , consisting of linke d

p h e n o l p ropanoid units. T h ree lignolbuilding blocks include syringyla l c o h o l , c o n i f e ryl alcohol, and

p - c o u m a ryl alcohol.

structure of lignin at the level of atomicconnectivities is poorly known, due in part tointer-specific structural variations, the fact thatlignin is the only aromatic fraction of cellulosicbiomass allows for some characterization of itsstructure and the concomitant changes that occurduring chemical processing of the biomass.

In contrast, characterization of the structure andcomposition of hemicellulose (Figure 6.3) in thesolid state is exceedingly difficult. Unlike cellulose,


6 . C rosscutting Scientific Issues

137

O

OO

OO

OO

HOH2C

OHHO

OORO

RO

OH

ROOR

OH

OR

HO

HOH2C

HOH2C

OH2C

O

OR

OH

HOH2Cgalactoglucomannan

O

O

O

OO

OO

OO

O

H3CO

HOOH

HO

HO

OHHO

O

OH

O CH2OHHOH2C

O

OH

OH

OH

arabinoglucuronoxylan

F i g u re 6.3: Hemicellulose is an amorphous polysaccharide containing both5 and 6 carbon sugars. Examples of hemicelluloses include

g a l a c t oglucomannans and arabinog l u c u ro n ox y l a n s .

almost all the individual compounds formed frombiomass during conversion processes such aspyrolysis and liquefaction. However, many of thecompounds in biomass-derived liquid processessuch as those described in Chapter 3 are nothydrocarbons; these compounds requiredevelopment of procedures different from thosedeveloped for petroleum.

6 . 2 . T H E R M O DY N A M I C S

Major innovations in biorefining will require theidentification of new processes, many of these beingcatalytic in nature. If we follow the lead of the ofthe petroleum and petrochemical industries, thesenew processes will be invented by scientists andengineers working with reaction mechanisms,reaction networks, and novel reactors. In addition,these developments will require an understandingof how reactants, products, and intermediatesinteract with catalyst surfaces, partition betweenthe two phases of a biphasic reactor, are distributed

between the gas and liquid (or solvent) phases, orare extracted into specific solvents in separationprocesses. At present, however, this vitalinformation about the physical properties ofbiomass-derived molecules is lacking. This dearthof information impedes not only the discovery ofnew processes, but also the design, operation,optimization and control of emerging technologies.

The unprecedented efficiencies achieved by thepetroleum and petrochemical refineries of todaywould not have been achieved without detailedprocess analyses, wherein intermediates and heatgenerated in one process are used in otherprocesses, such that the most effective utilization ofthe petroleum feed is achieved, both on an atom-efficiency and energy basis. Moreover, the successof the petroleum and petrochemical industries hasbeen achieved through the development andcontinuous improvement of new chemicalprocesses, most of these processes being catalyticin nature. In this respect, the formulation of a new process or the elucidation of the reaction


chemistry for an existing process typically involvesconsideration, at the molecular level, of reactionmechanisms and/or reaction networks. Suchanalyses also depend on having knowledge of thephysical properties of the reactants, products, andimportantly the potential reaction intermediates.

Likewise, to advance the field of biofuels productionand provide a solid foundation for further work, itis essential that theorists develop new approachesfor prediction of the thermodynamic properties(enthalpy, entropy, heat capacity) of biomass-derivedmolecules in the gas and liquid states, as well as invarious solvents. This work should provide notonly reliable thermodynamic data for specificbiomass-derived molecules of perceived importancein a biorefinery, but should also yield new methodsthat can be used effectively by practitioners toestimate the thermodynamic properties of newcompounds that may appear, for example, inreaction mechanisms, reaction networks, andprocess flow sheets for new chemical processes.

It is important to note, however, that the vastknowledge on the physical properties of moleculesnow available from the past decades of researchand development in the petroleum andpetrochemical industries deals primarily withhydrocarbon molecules that have limited oxygenfunctionality. Furthermore, most of this informationdeals with hydrocarbons in the gas phase orperhaps as non-polar liquids. Unfortunately, manyof the reactants, intermediates and even some ofthe products involved in the processing of biomass-derived feeds are highly oxygenated species, andthese molecules may well be processed mosteffectively in water or in polar solvents.

In short, the starting point for any systematicdiscovery of a new process or for theimplementation of an emerging process in abiorefinery is an analysis of the overall

thermodynamics, independent of whether theprocess is a chemical transformation or aseparation step. It is for this reason that the lackof thermodynamic data for the biomass-derivedreactants, products and intermediates presents amajor impediment to further developments inbiorefining.

6 . 3 . C H E M I CAL REACTION ENGINEERING

6.3.1 Identification of P rototypical Reactants (model compounds)

Much has been learned about catalytic processingin petroleum refining as a result of experimentationwith individual compounds that are representativeof a larger class of compounds; these are usuallyreferred to as model compounds. Examples areisobutane in catalytic cracking and thiophene inhydrodesulfurization of gas oil. Extensive testing isrequired to distinguish good from unsatisfactorymodel compounds, because no model is withoutlimitations in representing a larger group ofcompounds.

The application of this approach to biomass-derived feedstocks is still in its infancy. The modelcompounds identified so far have barely beeninvestigated as reactants in conversions intended tomimic biomass conversion. Work is needed toidentify appropriate model compounds in biomass-derived feedstocks. A good example of the kind ofstudies that need to be undertaken is a series ofmodel compounds that were identified as pertinentto the upgrading of pyrolysis oil to hydrocarbons[Gayubo,Aguayo et al., 2004a, 2004b, 2005].

Certain compounds have been shown to bevaluable for investigations related to petroleumrefining because, being more than representative intheir reactivities, they are key components of the



139

conversion reaction because they are the leastreactive compounds in the feedstock and,therefore, need to be converted for the product tomeet the required purity standards. An example is4,6-dimethyldibenzothiophene, which must beconverted in large measure for the fuel to meetsulfur content specifications.

Work is needed to identify these kinds of keycompounds in biomass-derived feedstocks forvarious upgrading processes.

6.3.2 Lumping of Biomolecules A useful methodology for representation ofcomplex multicomponent feedstocks in petroleumrefining technology is called lumping, whereby agroup of compounds that are chemically similar arerepresented as if they were a single (perhapsfictitious) compound chosen to approximate theirreactivity characteristics. The enormoussimplification of this procedure (representing, forexample, hundreds of compounds in a petroleumresiduum with tens of lumps), which is based on anextensive set of analytical results characterizing thefeeds and products, makes it possible to accuratelypredict the reactivities of various feedstocks (eventhose not yet tested) and the properties of theirrefining products.

Successful application of the lumping methods tobiomass-derived feedstocks will require thoroughanalysis of these feedstocks and the products oftheir conversion, primarily by mass spectrometry.It will also require a more fundamentalunderstanding of the reactions that occur inbiomass upgrading, which may be obtained in partfrom investigations of representative modelcompounds, as described above.

Successful identification of model compounds forthe upgrading of biomass-derived feedstocks willalso require extensive experimentation so that

candidate models can be improved by comparisonwith experiment. The work will need to beextended to a range of specific catalysts so thatgeneralizations can be made about fundamentalcatalyst types, such as acids and metals.

6.3.3 Characterization of Activities of Prototypical Cat a l y s t s

At this early stage of understanding biomassconversion, there is only fragmentary knowledgeabout how various feedstocks (or the modelcompounds representing them) are converted inthe presence of different catalysts. To achieveaccurate predictions of the best catalysts andreaction conditions for candidate biomass-conversion processes, it will be necessary togenerate data that demonstrate the conversion ofindividual model compounds with catalysts thatmay be considered protypical, such as acidiczeolites, supported metals, and bimetallic catalystsconsisting of metals on acidic supports. Additionalwork will provide further improved ideas aboutwhich catalysts constitute the best prototypes. Theinitial choices should be based on performancedata, and candidate catalysts that are low in activity,poor in selectivity, or poor in stability should beexcluded. As testing proceeds, the data shouldindicate catalyst stability in longer-term tests inflow systems.

Since biomass contains much more oxygen thanfossil fuels such as coal, oil and natural gas, catalyticconversion of this renewable resource willnecessarily involve processing in aqueous phasesystems. Consequently, many of the heterogeneouscatalysts that have been developed for hydrocarbonprocessing are inappropriate for conversion ofbiomass in aqueous media. Moreover, the chemicalreactions and mechanisms involved in thetransformation of biomass to transportation fuelsare poorly understood.


Transformation of biomass-derived molecules totransportation fuels will involve the removal ofoxygen while maintaining, as much as possible, thecarbon and hydrogen inventory of the molecules.These transformations necessarily include theselective manipulation of C-C, C-O and C-Hbonds. Heterogeneous catalysts and enzymes mustbe developed to selectively carry out the desiredreactions while minimizing the unproductiveformation of CO2 and heavy tars.

The data used to identify prototypical catalystsshould include activity and selectivity data, as wellas enough information to determine the reactionnetworks of well-chosen model compounds asreactants. Determination of the networks willrequire analysis to identify reaction intermediatesand experiments with these intermediates as feeds.Experiments should also be done to identify classesof compounds in feedstocks that are significantinhibitors of the various reactions, as well ascomponents (even impurities) that may be catalystpoisons. This work should include long-term tests(some with model compounds as reactants andsome with full feedstocks) to determine catalystdeactivation. Analysis of used catalysts will beneeded to understand the mechanisms of catalystdeactivation.

Ultimately, data characterizing catalyst performanceshould be reduced to include fundamental reactionkinetics, with rates of reactions expressed asturnover frequencies. Such determinations requirecharacterization of the catalyst to quantify thenumbers of catalytic sites (see section 6.5).

6.3.4 Reactor EngineeringChemical reactor engineering crosscuts all of theareas through which chemical catalysis will addressthe conversion of biologically derived feedstocksfor fuels production. Basic studies will be requiredto provide the foundation for implementing these

emerging technologies. Mechanisms will need to bedetermined, if not in atomic detail, at least in theability to determine kinetic reaction pathways.Once determined, these mechanisms will provide aframework for optimization of yields andselectivities of conversion processes in practicalreactors. Only after the complex details of eachreaction step are worked out can these processesbe compared, and eventually analyzed in scaled-upprocesses.

Real feedstocks, as opposed to laboratory-gradefeeds, often comprise mixtures of reagents,transformed in parallel reaction networks; theformation of biodiesel by the transesterification ofvegetable oil is an example.Transesterification isthe reaction of triglyceride (or other esters) withalcohols to produce alkyl esters (biodiesel) andglycerol typically in the presence of acid or basecatalysts.The oils contain many glycerol tri-estersfrom C13-C19 and each transition of tri- to di- tomono- ester could require individual kineticexpressions, all of which would have to be solvedsimultaneously. It is far simpler to assume that eachof the tri-esters would react with similar rates andto lump these together.This might work well forthis sequence, while for other more complexreaction mixtures the rates for specific species, orgroups of species, could be considerably different.Therefore, it may be advantageous to identify andclassify the reactants and identify those for whichthe overall rates of reaction are crucial to processsuccess. (In cellulosic liquefaction for example, aclass of lignin might prove to be most difficult tosolubilize by a specific approach.) The identificationof the fundamental reaction classes (lumps) can bedetermined experimentally by labeling. A similarapproach is employed in the petrochemical fieldand in polymerization kinetic analyses.

In addition to influencing the overall yields, reactorconfigurations can provide a method for the



141

collection of basic kinetic data. Stirred tank orplug flow reactor configurations appear as the idealextremes of mixing. However, many of thereactions involved in the transformation of bio-derived fuels are multiphase. Mixing between thegas, liquid, and solid (e.g. the requirement for theaddition of gas to the liquid phase duringhydrogenolysis) phases and at the interfaces thencontrols overall reactivity.

6 . 4 . CATA LYST ENGINEERING

Engineering the catalyst itself can be consideredover several scales. First, there are rational waysto prepare well-dispersed metal particles over highsurface-area supports, based on an understanding

of the chemical fundamentals of catalystimpregnation (see sidebar 6.1). Going further,catalyst sites might be designed at the molecularlevel.

An example of catalyst activity is shown in Figure6.4. Elucidating how the outer-sphere environmentcontrols heterogeneous catalysis is an importantcross-cutting issue that spans both enzyme andabiological catalysts. This can be accomplished bybinding a chromaphore that reports on the acidityand dielectric constant surrounding the active site,and is demonstrated for primary amines anchoredon silica. By way of illustration, diffuse reflectanceUV-Vis spectra of materials treated withsalicylaldehyde (more than 80% of the sites arebound with salicylaldehyde) show the acidic outer

F i g u re 6.4 Tre atment of heterogeneous primary amine catalysts withs a l i c y l a l d e hyde synthesizes a bound spectrophotometric probe that re p o rt s

d i rectly on the acidity and dielectric constant of the env i ronment thati m m e d i ately surrounds the primary amine active site. This env i ronment in

t u rn strongly affects nitroaldol catalysis activity and selectivity.


Calls are often heard to “transform the art of

catalyst preparation into a science.” One arena in

which this is occurring is in fundamental studies of

catalyst impregnation, the process by which a

solution containing dissolved metal precursors is

contacted with the high-surface-area catalyst

support. In this vein of work, the chemical

fundamentals of impregnation are being understood

and exploited for simple, rational methods to make

metal nanoparticles. Most industrial catalysts are of

this sort and employ aluminum oxide or carbon as

the support.

The preparation strategy of “strong electrostatic

adsorption” (SEA) [Brunelle 1978; Schreier and

Regalbuto 2004; Miller, et al; 2004] can yield ultra-

small metal particles with a simple contacting

process. The electrostatic adsorption mechanism is

depicted in Figure S-6.1, part a. At the pH of “point

of zero charge” or PZC, which is characteristic of

each oxide, the hydroxyl groups populating an oxide

or carbon surface are essentially neutral. At pH

levels more acidic than the PZC, the hydroxyl groups

protonate and the surface charge is positive, while

THE NANOSCIENCE OF CATA LYST SYNTHESIS

above the PZC, the hydroxyl groups deprotonate, and

the surface charge is negative. As an example, silica

has a PZC of about 3.5 and preferentially charges

negatively at high pH and so strongly adsorbs

[(NH3)4Pt]+2 [Miller et al. 2004].

To locate the pH of the strongest interaction, the

metal uptake is surveyed as a function of pH at

constant metal concentration (Figure S-6.1 part b)

[Schreier and Regalbuto 2004]. At the pH of

strongest electrostatic adsorption (about 11), the

precursor is adsorbed as a monolayer; the high

degree of dispersion can be preserved through the

catalyst finishing steps as the metal is reduced to

its active state. The Pt/SiO2 material shown in

Figure S-6.1 part c contains 1-nanometer Pt

particles, in which virtually every Pt atom will

contribute to catalytic activity [Miller, et al 2004].

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10 12 14pH Final

Pt adsorb

ed (

mol/m

2) L90

M7D

EH5

VN3S

FK300

Model

a) b) c)

HOH2+

O-

pH<PZC

pH>PZC

OHPZC

K1

K2

[PtCl6]-2

[(NH3)4Pt]+2

[H]+ (pH shifts)

Kads

Kads

OH2+

O-

pH<PZC

pH>PZC

OHOHPZC

K1

K2

[PtCl6]-2

[(NH3)4Pt]+2

[H]+ (pH shifts)

Kads

Kads

F i g u re S-6.1 Pt/silica Catalyst Synthesis by SEA a) electro s t atic mechanism, b )P TA uptake – pH survey to locate optimal pH over silica, c) reduction of sample

p re p a red at optimum conditions yields ultra-small metal nanopart i c l e s(F i g u re courtesy of J. R eg a l b u t o ) .



143

sphere in the silanol-rich material (blue), whichfacilitates formation of an iminium cation uponsalicylaldehyde binding, and is manifested in aprominent band at around 400 nm [Bass, Katz,2003, Bass, Solovyov, et al., 2006]. Materials havingaprotic outer spheres with varying dielectricconstant, ranging from the polar cyano to the non-polar methyl, show a stronger intensity at 325 nm relative to 400 nm, and instead requirepredominantly neutral imine tautomer uponsalicylaldehyde binding (Bass and Katz, 2003). Thewavelength of the iminium cation band (~400 nmband) is also sensitive to the local dielectricconstant of the environment surrounding theactive site, because the excited state of this speciesis less polar than the ground state. Therefore, theiminium cation band of the methyl-rich outersphere material (red) appears at slightly higherwavelength than for the respective, higherdielectric, silanol- and cyano-rich (purple) outersphere materials. This data explains the differentreactivities observed in these heterogeneous aminecatalysts [Bass, Solovyov, et al., 2006].

F i g u re 6.5 pH effect on cru s hs t rength in water at 200oC

(James F. W h i t e , P N N L , with permission)

The physically largest scale on which to considercatalyst engineering is in the mechanical andchemical integrity of formed catalyst pellets, whichare formed by extruding or otherwise processingpowdered material into various shapes. Inparticular, engineering formed catalysts for theaqueous phase poses numerous fundamental cross-cutting technical challenges.The systems that havebeen developed for petroleum and petrochemicalrefining in general are unstable under aqueousconditions.The most common supports used inpetroleum-based catalytic processes are based onmetal oxides of alumina, silica and alumina silicates.These supports are unstable under hydrothermalconditions.The issue is compounded when rununder high or low pH (Figure 6.5).Thus, supportsfor hydrothermal catalytic processes commonlyconsist of carbon, monoclinic zirconia, rutile titania,as well as a few others such as niobia, tin oxide,and barium sulfate.

Carbon as a support has a broad range of qualities,which depend on carbon source, processing, andpost processing treatments. Catalyst-grade carbonsvary in surface properties such as pH or point ofzero charge (PZC), porosity, pore volume, andothers.These properties can influence carboninteractions with substrate and products. Organicmolecules have strong interactions with activatedcarbon and so selective adsorption of substrateand products can influence catalyst performance(selectivity and rate). In addition to support-substrate/product interactions, there are strongsupport-metal interactions.The non-covalentinteractions of metals on carbon are “less sticky”than those of metal oxides. Hence metal migration,crystallite formation and other forms of sinteringare often observed. Metal sintering can becomplicated at extreme pH. Catalysts arecommonly composed of multiple metals.Fundamental work should be done to understandthe metal-metal interactions within a catalyst. In

pH effect on crush strength (24 hrs soak in pH adjusted water at 200 °C)

pH effect on crush strength pH effect on crush strength (24 hrs soak in pH adjusted water at 200 (24 hrs soak in pH adjusted water at 200 °°C)C)

0

10

20

30

40

50

60

70

80

90

100

0 2 4 6 8 10 12 14

pH of Soak Water

% R

ete

nti

on

of

Ori

gin

al

Str

en

gth

Granular C

Extruded C

Rutile TiO2

Silica

Alumina


some cases, the second (or additional) metal maynot be playing a role in the catalyst at all, butinstead be responsible for stabilizing the first metal.In other instances, the second (or additional metal)has a clear role in catalysis by improving reactionrate and product selectivity.

Catalyst stability (deactivation) issues aresubstantial when working with biomass feedstocks.Chemistry that works well with pristine systems(such as purified glucose) fail when using actualbiomass feeds because of the rich mixture ofvarious impurities present. Impurities includeinorganic salts or other products in ash, proteincontaining components, and other unknowns.Some impurities act as inhibitors and catalystperformance returns when the impurity isremoved. In other examples, catalysts arepermanently damaged by the impurity. Onechallenge is that impurities have different effects ona catalyst-to-catalyst basis. However, someimpurities cause problems in a more general oruniversal way. Materials containing proteins havesufficient sulfur and similar components that placethem in this category.

Related to catalyst engineering are the challengesof reactor engineering in aqueous systems. Onechallenge is the low solubility of reactive gassessuch as hydrogen and carbon dioxide.This resultsin the need for high reactor pressure andtemperature.The ionic nature of water at elevatedtemperatures leads to reactor corrosion issues.Ascorrosion occurs dissolved metals can plate out onthe catalyst and impact its performance.

6 . 5 . CATA LYST CHARACTERIZAT I O NFundamental understanding of catalytic chemistryis obtained when the detailed, atomic scalestructure of a catalyst and its chemical compositionis known and can be correlated to its catalyticactivity and selectivity. These qualitative “structure-function” relationships can be derived throughdetailed characterization of the catalyst, precisemeasurement of catalytic rates, and in-situ or in-operando measurement of reaction intermediatesover the catalyst surface. A schematic of thesethree measurement regimes used for derivingstructure-function relationships are shown in Figure 6.6.

A

B

A B

C D

Support

Metal or alloy

Gas phase kinetics (GC, mass spec., etc.)

Operando observation of reaction intermediates (IR, Raman, etc.)

Catalyst characterization: - S u r f ace (XPS, ISS,

BET, Chemisorp., etc.) - Bulk (XRD, XAS, etc.)

F i g u re 6.6 The T h ree Regimes of Measurement for the Derivation of Cat a l y s tS t ru c t u re-Function Relat i o n s h i p s .

( Typical measurement methods in pare n t h e s e s . )



145

0

5

10

15

20

25

30

1 11 21 31 41 51 61 71 81 91 101 111 121

Particle Size (Å)

Dis

trib

uti

on

%

0

1

2

3

4

5

6

7

8

9

10

1 11 21 31 41 51 61 71 81 91 101 111 121

Particle Size (Å)

Dis

trib

uti

on

%

2.2% Pd/SiO2 Standard preparation

dav = 55 Å

2.2% Pd/SiO2 Optimized preparation

dav = 16 Å

One of the most utilized and direct means to

evaluate supported metal catalysts is with electron

microcopy. High Angle Annular Dark Field, or “Z-

contrast” imaging, is one of the workhorse

techniques. In this imaging mode, contrast is

proportional to the atomic weight of the

constituents, making a heavy metal such as

palladium on a light oxide such as silica an ideal

sample. Virtually all metal particles are imaged and

particle size distributions can be readily obtained.

The comparison below shows relatively large particles

obtained by a typical “dry” or “incipient wetness”

impregnation, versus much smaller particles obtained

with a method that optimizes metal precursor –

oxide interactions during impregnation. In most

catalyst applications, small particles are desired since

there is more exposed metal area (active area) per

mass of metal. This condition is referred to as ‘high

dispersion” and is particularly important for

expensive metals such as platinum and palladium.

E L E C T RON MICROSCOPY OF SUPPORTED METAL CATA LYS T S

Figure courtesy of J. Regalbuto.


Several comprehensive texts review thecharacterization methods for a host of catalyticmaterials and applications. These include theEncyclopedia of Catalyst Characterization [Brundleet al., 1992], Spectroscopy in Catalysis[Niementsverdriet, 2000], and Characterization ofCatalytic Materials [Wachs, 1992].

Sampling of the gas phase for reactant and productcompositions is easily accomplished for commonlaboratory reactors at atmospheric pressure usinggas of liquid chromatography. If the catalyst isdeployed in an ultrahigh vacuum setting, massspectrometers can be used. In a combinatorialsetting, a thermal imager may be employed to makeparallel measurements of catalytic activity for

exothermic or endothermic reactions. Massspectrometers can also be employed to rapidlyscan an array of samples.

A state-of-the-art system for in-operandocharacterization is shown in Figure 6.7. It consistsof simultaneous Fourier Transform IR, visible andUV of the catalyst surface, and mass spectroscopyof the reactor effluent. Applied to study themethanol conversion to formaldehyde, the systempermits the observation of, not only catalystintermediates and products, but also catalyst sites.In this case, the changes of the species and catalyticsites are observed as the temperature is increased.From such data, catalytic mechanisms such as thoseshown in Figure 6.7 can be derived. Raman

100 150 200 250 300 350 400

(Raman)V5+

(Raman)V3+/V4+

(UV-vis)

(MS)(UV-vis)

(IR)CH

3O*

H2COV5+

V3+/V4+

Inte

nsit

y (

a.u

.)Temperature (oC)

Visible Laser

Heating Plate

Quartz Microreactor

UV Laser

FT-IR

UV-Vis Fiber Optics

Real-time Online Spectroscopy

CCD Detector

Micro GC - Rapid Analysis Programmable Motorized Stage

EL

Real-time GC/MS system

O

V

O O O

O

V

O O OH

OCH3 V

O O O

+ CH3OH

O

V

O O OH

OCH2 H

OH

V

O O OH

+ H2CO + H2O

F i g u re 6.7. I n - o p e r a n d o S p e c t roscopic A p p roach to Establ i s h i n gM o l e c u l a r / E l e c t ronic Stru c t u re .

( C atalytic Activity/Selectivity Relat i o n s h i p, c o u rtesy of I.E. Wa c h s )



147

spectroscopy would enable simultaneousidentification of reactive intermediates andmeasurement of kinetic phenomena [Shao,Adzic, 2005].

Characterization in liquid phase reactions is muchmore difficult. In these systems, catalyticconversion of biomass-derived molecules to fuelproducts involves a combination of catalytic stepson solid surfaces and/or enzymes coupled tosolution-phase reactions facilitated by H+ or OH-.The lack of information regarding the reactiveintermediates involved in these transformationscomplicates characterization. Classical UHVsurface techniques that utilize electronspectroscopies are inappropriate for studies in theaqueous phase because of excessive absorption andscattering of the ejected electrons. Moreover, thevibrational technique IR spectroscopy can sufferfrom excessive absorption of IR radiation by water.Development of surface-sensitive in-situspectroscopic tools that are capable of identifyingreactive species on catalyst surfaces in aqueoussolution are needed. Techniques such as surfaceenhanced Raman spectroscopy and, in particular,attenuated total reflectance IR spectroscopy mayfind broad application in this arena [Ortiz-Hernandez,Williams, 2003, McQuillan, 2001, Ferri,Burgi et al., 2001, Ferri, Burgi et all, 2002]. Inaddition, transient methods to interrogate reactiondynamics at the solid-solution interface should bedeveloped to relate catalyst structure to reactivity.Electro-analytical methods combined with surfacespectroscopy would enable simultaneousidentification of reactive intermediates andmeasurement of kinetic phenomena (Shao,Adzic,2005).

The structure of heterogeneous catalysts andconformation of enzyme catalysts are affected bysolution pH, ionic strength, surfactants, impurities,temperature, etc. For example, a supported

bimetallic catalyst may experience segregation ofthe metals, dissolution of the support, disruption ofthe metal-support interface, and migration of metalparticles depending on the nature of the aqueousenvironment at the interface. Moreover, traceimpurities present in biomass feedstocks mayselectively poison catalyst surfaces or denatureenzyme catalysts thus deactivating the system. Forexample, polyphenolic structures of lignin andtannin may inhibit catalytic activity by binding tothe inorganic or biological catalysts. In-situ

0

0.5

1

1.5

2

2.5

3

3.5

4

22.08 22.12 22.16 22.2

Normalized Absorp

Energy (keV)

a

b

c

d

e

f

g

h

i

j

F i g u re 6.8 X-ray absorption nearedge spectra of Ru samples as

p re p a red in air (b-e), and follow i n gt re atment with 40 atm H2 s at u r at e d

water at 473 K (g-j); (a) RuO2p ow d e r, ( b,g) Ru/Al2O3, ( c,h) Ru/C,(d,i) Ru/TiO2, (e,j) Ru/SiO2, (f) 17.6

wt% Ru/Al2O3 re f e rence sample.Spectra are offset for clarity.

From Ke t c h i e , Maris and Davis [2007],with permission from Elsev i e r.


0

0.05

0.1

0.15

0.2

0 1 2 3 4 5 6

ExperimentalFit

Magnitude of

r (Å)

0 1 2 3 4 5 6

Pd foilA

0

0.02

0.04

0.06

0.08

0 1 2 3 4 5 6

ExperimentalFit

r (Å)

0 1 2 3 4 5 6

Au foilB

Magnitude of

1

2

34

56

1

2

3

4 5

0

50

100

150

200

0 1 2 3

X-ray counts

Energy (keV)

C K

Au M

Pd LPd L

5

4

3

2

1

1

2

3

4

5

6

1

2

3

4 51

2

3

4

5

6

1

2

3

4 5

0

50

100

150

200

0 1 2 3

X-ray counts

Energy (keV)

C K

Au M

Pd LPd L

5

4

3

2

1

F i g u re 6.9 Radial stru c t u re functions (Fourier transforms of EXAFS) fro ms u p p o rted Au-Pd nanoparticles (not corrected for phase shifts) with non-

linear least squares fits for bimetallic AuPd catalyst (A) Pd (Pd-Pd and Pd-Au shells) and (B) Au (Au-Au and Au-Pd shells). D ata we re collected atambient conditions. From Ke t c h i e , M u r ayama and Davis [2007], w i t h

permission from Elsev i e r.

F i g u re 6.10 Electron micrograph of AuPd bimetallic catalyst with EDSspectra of individual particles (numbers correspond to the part i c l e

d i rectly to the right of the number). Spectra are offset for clarity. Fro mKe t c h i e , M u r ayama and Davis [2007], with permission from Elsev i e r.



149

spectroscopic, scattering and imaging tools need tobe developed that allow for real-time monitoring ofcatalyst structure at the nanometer level. Highenergy X-ray methods are particularly suitable forstudying the structure of catalysts in water [Maris,Ketchie, et al., et al., 2006, Ketchie, Maris et al.,2007, Ketchie, Murayama , et al., 2007, Maris,Ketchie, et al., 2007]. For example, the near edgespectra associated with the K edge of varioussupported Ru catalysts in air and in liquid watercontaining hydrogen are shown in Figure 6.8(previous page). The spectra of the catalysts in airare similar to that of RuO2 powder, indicating thatthe supported ruthenium samples were completelyoxidized upon exposure to air, regardless of thesupport. The treatment in H2-saturated watersolution at 473oK was sufficient to reduce theoxide to Ru metal.

Combining catalyst characterization results fromX-ray spectroscopy with those from analyticalelectron microscopy is a powerful method toexamine multi-component materials. The analysisof a carbon-supported bimetallic Au-Pd catalystused for glycerol oxidation illustrates the utility ofthis approach. Figure 6.9 shows the radialstructure functions associated with Pd and Au insupported bimetallic catalysts prepared bydeposition of Au onto Pd nanoparticles. Theresults from analysis of the extended X-rayabsorption fine structure confirm the selectivedeposition of Au onto Pd without disruption of thePd particles. Figure 6.10 presents an electronmicrograph together with the EDS spectra for fiveindividual particles of the bimetallic sample, all ofwhich indicate the presence of both Au and Pd,which is consistent with the results from X-rayabsorption spectroscopy. Next-generation electronmicroscopes that allow for imaging of nanometer-sized features in thin liquid layers are needed toexplore the structural changes of catalysts inaqueous environments.

The number of active sites in a catalytic materialcan be determined by estimates of particle sizefrom electron microscopy or by chemical titrationmethods such as chemisorption or temperatureprogrammed desorption. From detailedcharacterization of the number of active sites,qualitative, and at times even quantitative,correlations can be derived between the structureand chemical composition of a catalyst and itsactivity and selectivity. Beyond these correlations,great advances have been made in computationalcatalysis, which impart an even more detailedunderstanding of catalyst function.

6.6 COMPUTATIONAL CHEMISTRY

Computational chemistry is a truly cross-cuttingtool for the development and optimization ofbiofuels production reactions. Computationalchemistry has shed light on many aspects oftraditional fuel production with heterogeneouscatalysts. Such calculations have providedmicroscopic insights into several important areas,including: reactant adsorption-diffusion; chemicalreaction and transition state control; and productdiffusion-desorption. All of these insights lead tonew understanding about the factors that controlselectivity in these processes. So too will it be with the burgeoning biofuel production industry.Computational chemistry methods (existing and to-be-developed) can provide new molecularunderstanding for virtually all stages of biomass-to-biofuel processing.

In the remainder of this section, we describe: thekinds of problems computational chemistry canaddress today (with examples); the special needsand challenges faced in the computational study ofbiomass refining; the near-term opportunities forcomputational chemistry in this area; and finally, thelong-term challenges/opportunities that should be


pursued in order to fully illuminate the molecularprocesses involved in making biofuels.

6.6.1 Present Computational A p p ro a c h e s

Quantum Chemistry: Computational chemistry canbe broken up into two broad categories: (i)quantum chemical calculations and (ii) classicalmolecular simulations. Quantum chemicalcalculations are important for accurately predictingthe reaction energies and barriers associated withbond-breaking and bond-making. One typicallyidentifies relatively few important configurations –reactant, transition state, product or intermediate –and computes accurate electronic energies andvibrational frequencies for use in harmonicestimates of equilibrium constants and ratecoefficients. These quantities are then fed intomicro-kinetic models that predict the time-dependent concentrations of observable species forcomparison with experiment [Kandoi, Greeley etal. 2006].

The Hartree-Fock approach provides a usefulbenchmark theory for quantum chemistry. Thisapproach applies the Pauli exclusion principleexactly, while treating electron-electron repulsionsonly within mean field theory. The differencebetween the exact (non-relativistic) energy and theHartree-Fock energy is defined as “correlationenergy,” which tends to be important for predictivetreatments of reaction kinetics. The most popularmethod for estimating correlation energy nowadaysis density functional theory (DFT), because theelectron density is a simpler property than the fullmany electron wavefunction. Also, theorists havemanaged to find ways to bundle correlation energyinto ever more sophisticated “exchange-correlationpotentials” such as that used in the B3LYPapproach [Kohn, Becke et al. 1996].

Molecular simulations are applied to systems suchas liquids, which exhibit many configurations withnearly equal energy, hence amounting to acalculation of the system entropy [Frenkel and Smit1996]. Molecular simulations have been applied tounderstanding the thermo-physical properties ofliquids and gases, on their own or in bulk mixtures[Stubbs, Chen et al. 2001]. Simulations have alsobeen performed to study inhomogeneous fluids incontact with solid surfaces such as carbons [Alba-Simionesco, Coasne et al. 2006] and zeolites[Saravanan and Auerbach 1998; Smit and Krishna2003]. Non-equilibrium simulations have beenperformed to model mass-transport in nanopores[Maginn, Bell et al. 1993] and microwave heating ofzeolites [Blanco and Auerbach 2002; Blanco andAuerbach 2003].

The two main simulation approaches are moleculardynamics (MD) and Monte Carlo (MC). MD issimpler to implement, although it requires not justthe potential energy, but also the derivative of thepotential (i.e. the force vector). This vector is usedto numerically integrate Newton’s classical equationof motion. Given the force vector and a systemtime step, updating the state of the system is trivial,using some integration algorithm such as “velocityVerlet.” This approach is particularly convenient formodeling the collective motions of fluid systems,because every particle in the system moves duringeach MD time step. However, MD suffers fromserious time-scale limitations because keeping thealgorithm stable requires relatively small time steps,on the order of femtoseconds, limiting most MDruns to nanoseconds.

MC is not limited by time steps in the same way asMD because MC does not numerically solve adifferential equation (as does MD). Instead, MCprovides stochastic sampling of the fluctuations in agiven ensemble such as the canonical or grand



151

canonical ensembles. The main problem with MC isthat it requires the modeler to concoct ways ofupdating the system, which for highly coupledsystems such as polymers can be difficult to divine.For example, MC moves of one atom at a time maywork well for an atomic fluid, but for a molecularor polymeric fluid, most such moves will berejected, leading to inefficient sampling of states.Methods such as configurational-bias MC wereinvented to improve sampling of alkanes in zeolitesand polymers in bulk [de Pablo,Yan et al. 1999].As yet, no method exists to sample the wholesalefluctuations of three-dimensional networks such as amorphous silica.

Auerbach and coworkers have applied non-equilibrium molecular simulations to exploreenergy distributions that arise during microwave(MW) heating of zeolite-guest systems [Blanco andAuerbach 2002; Blanco and Auerbach 2003].Experiments suggest that MW-heated zeolitesbehave in ways that are qualitatively different fromconventionally heated zeolites. Auerbach andcoworkers have found from their molecular

simulations that such “MW effects” can occurbecause of selective heating, wherein a portion ofthe system is excited by the MWs and becomeshot, while the rest of the system remains relativelycold (Fig. 6.11). Quasi-elastic neutron scatteringexperiments are underway to test the microscopicpredictions of Auerbach’s simulations.

6.6.2 Biomass Refining: Special Needs and Challenges

In comparison with atomistic modeling ofpetroleum refinement, biomass refining posesspecial needs and challenges. These can becategorized as pertaining to: (i) solid biomasssubstrate and (ii) oxygenated compounds.

SOLID BIOMASS SUBSTRATE: In contrast withpetroleum, which is a liquid albeit a viscous one,biomass is solid, which poses certain masstransport limitations on attempts to react biomass.If computational chemistry is to provide a deeperunderstanding of biomass refinement, the firstchallenge is to provide a microscopic picture of thesolid state of cellulose, hemicellulose, and lignin,

Fig 6.11 Microwave driven ze o l i t e . F i g u re courtesy of S. A u e r b a c h


and how these materials interact. A secondchallenge is to provide an understanding of how toliquefy, dissolve, and/or vaporize this solid phasewithout forming excess coke.

OXYGENATED COMPOUNDS: In contrast withpetroleum processing, which mostly involvestransformations of hydrocarbons (with some traceamounts of sulfur and lead), biomass processinginvolves the daunting task of removing largeamounts of oxygen from biomass-derivedfeedstocks. This fact has several consequences.First, many of the relevant reactants, intermediatesand products of biomass refinement do not havewidely tabulated thermodynamic properties, makingit difficult to apply the full power ofthermodynamic analyses on biomass refining.Second, because of the chemical and thermalinstabilities of carbohydrates (relative tohydrocarbons) the processing of biomass-derived

species might best be confined to the liquid phase[Chheda and Dumesic 2007]. Any deepunderstanding of such processing will have toaccount for solvent effects in addition to thealready challenging study of reaction kinetics andcatalysis. Third, to avoid complexities associatedwith liquid phase processing, some researchers arepursuing fast pyrolysis for gas phase processing(and concomitant gas phase modeling). But alas,there is no free lunch, as fast pyrolysis involvesstrongly non-equilibrium systems.

Gasification of biomass to syngas (CO/H2)followed by Fischer-Tropsch catalysis is yet anotheravenue for refining biomass. While still challengingto model by computational chemistry, thisapproach applied to biomass presents noqualitatively new challenges to molecular modeling,and as such is not discussed further in this section.

Simulating catalytic reactions in the liquid phase

presents a number of challenges due to poor

understanding of the structure at the aqueous/metal

interface, the number of potential reaction

intermediates and reaction pathways, and the

dynamics of the interface under reaction conditions.

Recent theoretical calculations by Desai, Sinha and

Neurock [Desai 2003; Desai and Neurock 2003; Sinha,

in preparation] have shown that the solution phase

can actually stabilize charged transition states in

some reactions and can even directly participate in

the reaction. Shown here is the lowest energy

transition state for the initial hydrogenation of the

C=O bond of formaldehyde over a palladium metal

surface in the presence of an aqueous solution.

Computational catalysis of this sort will complement

experimental efforts and begin to suggest potentially

more active and selective materials.

C O M P U TATIONAL CATA LYS I S

The transition state for thehy d rog e n ation of

f o r m a l d e hyde over Pd(111) inaqueous solution. Wat e r

molecules we re re m oved tohelp visualize the surface re a c t i o n .

Figure courtesy of M. Neurock.



153

In summary, modeling the atomic-level details ofbiomass processing will require methods that treat:(i) solid substrates; (ii) heretofore uncharacterizedoxygenated molecules; (iii) solvent effects inreactions; and (iv) strongly non-equilibriumsystems. We note that with the exception of item(ii), these tasks represent “grand challenges” forpresent-day methods in computational chemistry.As such, significant method development isrequired for computational chemistry to fullyilluminate the atomic-level details of biomassrefinement.

6.6.3 Near-Term Opportunities for C o m p u t ational Chemistry of Biomass Refining

Based on the capabilities of present-daycomputational chemistry methods, and theparticular technological challenges posed by thebiomass refining problem, several near-termopportunities exist for computational chemistry toprovide fundamental insights into biomass refining.In this section, we outline four specific areas.

6.6.3.1 Calculating Thermodynamic Properties

Thermodynamics provides the foundation fornearly all branches of science and engineering,specifying the constraints of energy conservationand maximal work. For the burgeoning field ofbiomass refining to be as successful as possible, allpossible thermodynamic information must beavailable for the wide variety of oxygenatedcompounds likely to arise as reactants, products,and intermediates of the refining process.Quantities such as enthalpies and entropies offormation, and also heat capacities over widetemperature ranges, must be tabulated and madeavailable to the community at large. The firstchallenge will be agreeing on a useful standardstate for these thermodynamic quantities. Forexample, glucose is a solid, dimethyl ether is a gas,and ethanol is a liquid at standard state. For

computational chemistry, gas phase is the mostconvenient standard state because one avoids thechallenge of computing liquid and/or solidequations of state. However, this may not be themost convenient for use by synthetic chemists,reaction engineers and biorefiners. Despite theseissues, computational chemistry should be able tocontribute important information to helpcharacterize the thermodynamic states of esotericoxygenated molecules, if only in the gas phase. Thismight take the form of an expanded “G2-like”database [Curtiss, Raghavachari et al. 1991], whichuses a hierarchy of basis sets and levels of theoryto provide heats of formation and Gibbs freeenergies of formation to within 1 kcal/mol (i.e.,“chemical”) accuracy.

6.6.3.2 Molecular Simulations of Biomass Components

Providing microscopic insights into the structure,dynamics and phase behavior of cellulose,hemicellulose and lignin may facilitatebreakthroughs in our ability to refine thesebiomass components. Because cellulose is acrystalline material, it is amenable to molecularmodeling with atomistic detail [Morrison, Chadwicket al. 2002] by analogy with metals and zeolites.Existing forcefields such as AMBER and CHARMM,used for modeling biomolecules such as proteinsand nucleic acids, should be readily applicable tocellulose. Computing solid-fluid phase diagrams inthe presence and absence of solubilizing specieswould yield useful information.

In contrast to cellulose, hemicellulose and lignin are both amorphous materials, making atomisticmodeling much more challenging. Because muchless is known about the structures of thesecomponents, and how they interact, coarse-grainedmodeling would still yield interesting insights intothe network connectivities and stabilities of thesebiomass components. Along these lines, coarse-grained modeling has recently shed some light on


silica polymerization to form nanoparticles that areprecursors to zeolite growth [Jorge,Auerbach et al.2005; Jorge,Auerbach et al. 2006]. The three-dimensional network formation in lignin bearsresemblance to that of amorphous silica.

6 . 6 . 3 . 3 Thermo-Physical Modeling of Bio-Oil

Bio-oil is a catch phrase for unrefined, liquefiedbiomass. It can result from fast pyrolysis, highpressure liquefaction, or other means. It is a verycomplex fluid with hundreds of components, and asa result, very poorly understood thermo-physicalproperties. Modeling of diffusion and phasebehavior of bio-oil could yield interesting insightsfacilitating its subsequent refinement. Simplermodels of bio-oil, containing many fewercomponents, would have to be developed for usein molecular simulations. Such models couldinspire the development of simple bio-oil standards,allowing various laboratories to comparerefinement results in meaningful ways.

6.6.3.4 Micro-Kinetic Modeling of Gas-Phase Refining of Biomass- Derived Feedstocks

As a first step towards modeling of solution-phasecatalytic processing of biomass-derived feedstocks,modeling the gas-phase refining would yield usefulbaselines of catalytic activity in the absence ofsolvent. Indeed, the best way to learn about theeffect of a solvent is first to omit it from themodel, then later include it and compare theeffects. The initial gas phase calculation would yieldmechanistic insights that later could be updated bynew methods that treat solvent effects as well.This could be applied to the catalytic processing ofbiomass-derived feedstocks reported by Dumesicand coworkers.

6.6.4 Long-Term Opportunities for C o m p u t ational Chemistry of Biomass Refining

The holy grail of molecular modeling for biomassrefining involves three main application areas: (i)modeling lignocellulose assembly/disassembly, (ii)modeling fast pyrolysis in the absence andpresence of catalysts, and (iii) modeling liquid-phasecatalytic refining of biomass.

6.6.4.1 Modeling Ligno-Cellulosic S t ru c t u re and Dynamics

Molecular simulations of lingocellulosic structureand dynamics will require the development of newmulti-scale approaches, able to resolve chemicalreactions at the atomic scale, while simultaneouslytreating amorphous structures at nano-scales. Thismay also require the development of novel reactivesimulation methods such as reactive Monte Carlo[Johnson 1999] to account for hydrolysis ofcarbohydrate chains.

Modeling fast pyrolysis is crucial for providingmicroscopic insights into the complex chemistryunder these rapid heating conditions. Thispresents another great challenge for computationalchemistry because of the strongly non-equilibriumnature of this process. In particular, it is notobvious that transition state theory (TST) appliesunder such conditions, or even if the concept oftemperature remains meaningful at all. TSTassumes that all system degrees of freedom remainin thermal equilibrium (i.e., in the canonicalensemble) as the reaction coordinate passes fromreactant, through the transition state, and on toproduct. For systems undergoing heating as rapidas 1000oCelsius per second, as happens during fastpyrolysis, it is unlikely that such an assumptionremains valid. Molecular simulations with reactiveforcefields, and ab initio (Car Parrinello) moleculardynamics [Kuo, Mundy et al. 2004], may yieldinsights into energy distributions that arise during



155

6.7 RECOMMENDAT I O N S

The cross-cutting scientific needs dealt with in thischapter are summarized in the followingrecommendations:

• Build an analytical database for biomolecules.

• Compile, by experiment and theory, the keythermodynamic properties of biomolecules.

• Identify prototypical reactant biomolecules aswell as lumps of molecules, and prototypicalcatalysts.

• Discover and develop novel multiphasereactors.

• Engineer catalysts on the micro and macroscales for improved activity and stability,particularly for the demanding liquid phaseenvironment.

• Develop in-operando and atomic-scale catalystcharacterization methods, particularly for theliquid phase.

• Develop computational methods for gas andliquid phase reactions involving solid biomasssubstrates and oxygenated compounds.

fast pyrolysis of biomass. Such simulations could beperformed on models of bare biomass, and onmodels that include catalysts such as metals orzeolites.

6 . 6 . 4 . 2 Modeling Solvent Effects in Liquid-Phase Pro c e s s i n g

As discussed above, present-day computationalchemistry breaks up into two kinds of calculations:accurate calculations of energies for fewconfigurations (chemical interactions), and rathermore approximate calculations of energies for manyconfigurations (physical interactions). It remainschallenging to model simultaneously coupledchemical and physical effects, such as liquid phasecatalytic processing of biomass studied by Dumesicand coworkers, involving biomass-derived moleculessuch as sugars in aqueous solution reacting in thepresence of metal catalysts. Multi-scalecomputational methods need to be developed toapproximate such complex catalytic systems. Forexample, one could imagine running a DFTcalculation on a reactive catalytic system embeddedin a classical model of a solvent. Such embeddingschemes are now becoming relatively commonplace[Sillar and Purk 2002; Fermann, Moniz et al. 2005].One could compute minimum energy reactionspaths for several fixed solvent configurations tosearch for solvent effects in the reaction. Thesecalculations appear daunting at present, but willbecome more feasible with both algorithmicdevelopments and improved computationalresources.

While biomass refining remains a series ofcomplicated and poorly understood processes,there is much room for computational chemistry toshed light on and improve the efficiency of suchrefining. Through the judicious choice of modelsystems, and the development of new methods,computational chemistry can shed much light onhow biomass is refined today, and how it should berefined in the future.


6.8 REFERENCES

Alba-Simionesco, C., B. Coasne, et al. (2006)."Effects of confinement on freezing and melting." J.Phys. Condensed Matter 18: R15-R68.

Bass, J. D. , and Katz,A., (2003),The Effect of Outer-Sphere Acidity on Chemical Reactivity in aSynthetic Base Heterogeneous Catalyst ,Angew.Chem. Int. Ed. Engl., 42, 5219-5222

Bass, J. D., Solovyov, et al., (2006),Acid-BaseBifunctional and Dielectric Outer-Sphere Effects inHeterogeneous Catalysis: A ComparativeInvestigation of Model Primary Amine Catalysts, J.Am. Chem. Soc. 2006, 128, 3737-3747

Blanco, C. and S. M.Auerbach (2002). "Microwave-Driven Zeolite-Guest Systems Show AthermalEffects from Nonequilibrium Molecular Dynamics."J.Am. Chem. Soc. 124: 6250-6251.

Blanco, C. and S. M.Auerbach (2003)."Nonequilibrium Molecular Dynamics ofMicrowave-driven Zeolite-Guest Systems: LoadingDependence of Athermal Effects." J. Phys. Chem. B107: 2490-2499.

Brundle, C.R., (1992), Encyclopedia of materialscharacterization, with Evans, C.A. Jr., and Wilson,S.,Butterworth-Heinemann, Boston.

Brunelle, J.P., “Preparation of catalysts by metalliccomplex adsorption on mineral oxides” Pure Appl.Chem. 50, 1978, 1211.

Chheda, J. N. and J.A. Dumesic (2007). "Anoverview of dehydration, aldol-condensation andhydrogenation processes for production of liquidalkanes from biomass-derived carbohydrates."Catalysis Today 123: 59-70.

Curtiss, L.A., K. Raghavachari, et al. (1991)."Gaussian-2 Theory for Molecular Energies of 1stRow and 2nd Row Compounds." J. Chem. Phys. 94:7221-7230.

de Pablo, J. J., Q. L.Yan, et al. (1999). "Simulation ofphase transitions in fluids." Annual Review ofPhysical Chemistry 50: 377-411.

Desai, S. K., Ph.D.Thesis University of Virginia, 2003.

Desai, S. K., and M. Neurock, Electrochim.Acta, 48(25-26), 3759-3773, 2003.

Fermann, J.T.,T. Moniz, et al. (2005). "ModelingProton Transfer in Zeolites: Convergence Behaviorof Embedded and Constrained ClusterCalculations." J. Chem.Theory Comput. 1: 1232-1239.

Frenkel, D. and B. Smit (1996). UnderstandingMolecular Simulations. San Diego,Academic Press.

Gayubo,A. G.,Aguayo,A.T. et al. (2004a),Transformation of Oxygenate Components ofBiomass Pyrolysis Oil on a HZSM-5 Zeolite. I.Alcohols and Phenols. Ind. Eng. Chem. Res. 43,2610-2618

Gayubo,A. G.,Aguayo,A.T. et al. (2004b),Transformation of Oxygenate Components ofBiomass Pyrolysis Oil on a HZSM-5 Zeolite. II.Aldehydes, Ketones, and Acids, Ind. Eng. Chem. Res.43, 2619-2626.

Gayubo,A. G.,Aguayo,A.T. et al. (2005), Undesiredcomponents in the transformation of biomasspyrolysis oil into hydrocarbons on an HZSM-5zeolite catalyst. J Chem Technol Biotechnol 80,1244.


Ferri, D., Burgi,T., and Baiker, A. (2001), Pt andPt/Al2O3 Thin Films for Investigation of CatalyticSolid-Liquid Interfaces by ATR-IR Spectroscopy: COAdsorption, H2-Induced Reconstruction andSurface Enhanced Absorption,” J. Phys. Chem. B105, 3187-3195.

Ferri, D., Burgi,T., and Baiker,A. (2002), ProbingBoundary Sites on a Pt/Al2O3 Model Catalyst byCO2 Hydrogenation and In Situ ATR-IRSpectroscopy of Catalytic Solid-Liquid Interfaces,Phys. Chem. Chem. Phys. 4, 2667-2672.

Johnson, J. K. (1999). "Reactive Canonical MonteCarlo." Advances in Chemical Physics 105: 461-481.

Jorge, M., S. M.Auerbach, et al. (2005). "ModelingSpontaneous Formation of PrecursorNanoparticles in Clear-Solution Zeolite Synthesis."J.Am. Chem. Soc. 127: 14388-14400.

Jorge, M., S. M.Auerbach, et al. (2006). "Modelingthe Thermal Stability of Precursor Nanoparticles inZeolite Synthesis." Mol. Phys. 104: 3513-3522.

Kandoi, S., J. Greeley, et al. (2006). "Prediction ofexperimental methanol decomposition rates onplatinum from first principles." Topics in Catalysis37: 17-28

Ketchie,W.C., Maris E.P., and Davis, R.J. (2007), In-situ X-ray Absorption Spectroscopy of SupportedRu Catalysts in the Aqueous Phase, Chem. Mater.19, 3406.

Ketchie,W.C., Murayama M., and Davis, R.J. (2007)Selective Oxidation of Glycerol over Carbon-Supported AuPd Catalysts” J. Catal., in press.

Kohn,W.,A. D. Becke, et al. (1996). "Densityfunctional theory of electronic structure." J. Phys.Chem. 100: 12974-12980.

Kuo, I. F.W., C. J. Mundy, et al. (2004). "Liquid waterfrom first principles: Investigation of differentsampling approaches." J. Phys. Chem. B 108: 12990-12998.

Maginn, E. J.,A.T. Bell, et al. (1993). "TransportDiffusivity of Methane in Silicalite from Equilibriumand Non-Equilibrium Simulations." J. Phys. Chem.97: 4173-4181.

Maris, E.P., Ketchie,W.C.,et al. (2006), J.“Metal Particle Growth during GlucoseHydrogenation over Ru/SiO2 Evaluated by X-rayAbsorption Spectroscopy and ElectronMicroscopy”Phys. Chem. B 110, 7869.

Maris, E.P., Ketchie,W.C., et al. (2007), GlycerolHydrogenolysis on Carbon-Supported PtRu andAuRu Bimetallic Catalysts” manuscript submittedfor publication.

McQuillan,A.J. (2001), Probing Solid-SolutionInterfacial Chemistry with ATR-IR Spectroscopy ofParticle Films,Adv. Mater. 13, 1034-1038.

Miller, J.T., Kropf,A. J., Schreier, M., and Regalbuto,J.R., J. “A fundamental study of platinumtetraammine impregnation of silica 2.The effect ofmethod of preparation, loading, and calcinationtemperature on (reduced) particle size” Catal. 225, 2004, 203.

Morrison, G.,A.V. Chadwick, et al. (2002)."Computational studies of the structural andtransport properties of the cellulose-water-amineoxide system." Phys. Chem. Chem. Phys. 4: 3407-3414.

Niemantsverdriet, J.W. (2000), Spectroscopy inCatalysis,Wiley-VCH,Weinheim.



Ortiz-Hernandez, I. and Williams, C.T. (2003), InSitu Investigation of Solid-Liquid CatalyticInterfaces by Attenuated Total Reflection InfraredSpectroscopy,” Langmuir 19 2956-2962.

Saravanan, C. and S. M.Auerbach (1998)."Simulations of High Tc Vapor-Liquid PhaseTransitions in Nanoporous Materials." J. Chem.Phys. 109: 8755-8758.

Schreier, M., and Regalbuto, J.R., J. Catal. 225, 2004,190.

Shao, M.H. and Adzic, R.R. (2005), SpectroscopicIdentification of the Reaction Intermediates inOxygen Reduction on Gold in Alkaline Solutions, J.Phys. Chem. B 109, 16563-16566.

Sillar, K. and P. Purk (2002). "Calculation of theproperties of acid sites of the zeolites ZSM-5 usingONIOM method." J. Mol. Struct.Theochem. 589:281-290.

Sinha, N. and Neurock, M. in preparation

Smit, B. and R. Krishna (2003). SimulatingAdsorption of Alkanes in Zeolites. Handbook ofZeolite Science and Technology. S. M.Auerbach, K.A. Carrado and P. K. Dutta. New York, MarcelDekker, Inc.: 317-340.

Stubbs, J. M., B. Chen, et al. (2001). "Monte Carlocalculations for the phase equilibria of alkanes,alcohols, water, and their mixtures." Fluid PhaseEquilibria 183: 301-309.

Wachs, I.E. (1992), Characterization of CatalyticMaterials, Israel E., Butterworth-Heinemann,Boston.


A p p e n d i c e sINVITED PA RT I C I PA N T S

Auerbach, ScottUniversity of Massachusetts [email protected]

Badger, PhillipRenewable Oil [email protected]

Bain, RichardNational Renewable Energy [email protected]

Bakshi, Bhavik R.Ohio State [email protected]

Bartholomew, CalvinBrigham Young [email protected]

Blommel, PaulVirent Energy [email protected]

Boateng,A.A. (Kwesi)United States Department of [email protected]

Boehman,André L.Pennsylvania State [email protected]

Boulard, David [email protected]

Brown, RobertIowa State [email protected]

Cameron, DouglasKhosla [email protected].

Chuang, StevenUniversity of [email protected]

Conner, CurtUniversity of Massachusetts [email protected]

Czernik, StevenNational Renewable Energy [email protected]

Davis, BurtUniversity of [email protected]

Dean,AnthonyColorado School of [email protected]

Dietenberger, Mark A.United States Department of [email protected]

Dumesic, James A.University of [email protected]

Elliott, DouglasPacific Northwest National [email protected]

Feik, CalvinNational Renewable Energy [email protected]

Fireovid, RobertUnited States Department of [email protected]

Appendix A • Pa rt i c i p a n t s


Fjare, [email protected]

Foust,ThomasNational Renewable Energy [email protected]

Francisco, Manuel [email protected]

Gaffney,AnneABB [email protected]

Gangwal, SantoshResearch Triangle Institute [email protected]

Gates, BruceUniversity of California- [email protected]

Gerry, FrankBritish [email protected]

Goodwin, James G., Jr.Clemson [email protected]

Grabowski, PaulDepartment of [email protected]

Hamilton, BruceNational Science [email protected]

Hewgley, John General Electric [email protected]

Holladay, JohnPacific Northwest National [email protected]

Huber, GeorgeUniversity of [email protected]

Jones, ChristopherGeorgia [email protected]

Jones, MarkDow [email protected]

Katz,AlexanderUniversity of [email protected]

Kung, HaroldNorthwestern [email protected]

Lucia,AngeloUniversity of Rhode [email protected]

Mahajan, DevinderBrookhaven National [email protected]

Malone, MikeUniversity of [email protected]

Manzer, LeoCatalytic [email protected]

Marinangeli, RichardUOP A Honeywell [email protected]

A p p e n d i c e s


Marincean, SimonaMichigan State [email protected]

Miranda, RaulDepartment of Energy- Basic [email protected]

Mulqueen, MichaelDefense Advanced Research Project [email protected]

Pezzullo, LeslieBCS, [email protected]

Steven D. PhillipsNational Renewable Energy [email protected]

Reed,ValeriaDepartment of Energy Office of [email protected]

Regalbuto, JohnUIC, National Science [email protected]

Rhodes,William D.Savannah River National [email protected]

Ro, Kyoung S.United States Department of [email protected]

Saxton, [email protected]

Schmidt, LannyUniversity of [email protected]

Shanks, BrentIowa State [email protected]

Song, ChunshanPennsylvania State [email protected]

Steele, Philip H.Mississippi State [email protected]

Stevens, DonPacific Northwest National [email protected]

Stevens, James [email protected]

Suppes, GalenUniversity of [email protected]

Turn, ScotUniversity of [email protected]

Vanderspurt,Thomas HenryUnited Technologies Research [email protected]

Vrana, [email protected]


Wang,YongPacific Northwest National [email protected]

Weatherwax, SharleneDepartment of Energy-Office of Biological andEnvironmental [email protected]

Wesson, Rosemarie D.National Science [email protected]

Westmoreland, PhilNational Science [email protected]

Wyman, CharlieUniversity of California [email protected]

Xu,YeOak Ridge National [email protected]

Zhang, ConradPacific Northwest National [email protected]

Zullo, [email protected]

Appendix A • Pa rt i c i p a n t s

A p p e n d i c e s


ORGANIZING COMMITTEE

B a i n , R i c h a rdNational Renewable Energy [email protected]

C ze rn i k , S t eve nNational Renewable Energy [email protected]

D u m e s i c, James A .University of [email protected]

E l l i o t t , D o u g l a sPacific Northwest National [email protected]

Fe i k , C a l v i nNational Renewable Energy [email protected]

G o o dw i n , James G. , J r.Clemson [email protected]

G r a b ow s k i , Pa u lDepartment of [email protected]

H a m i l t o n , B ru c eNational Science [email protected]

H o l l a d ay, J o h nPacific Northwest National [email protected]

H u b e r, G e o rg eUniversity of [email protected]

M a l o n e , M i keUniversity of [email protected]

M i r a n d a , R a u lDepartment of Energy- Basic [email protected]

R eg a l b u t o, J o h nUIC, National Science [email protected]

S h a n k s , B re n tIowa State [email protected]

S o n g , C h u n s h a nPennsylvania State [email protected]

S t eve n s , D o nPacific Northwest National [email protected]


Appendix B • Organizing Committee

A p p e n d i c e s


INVITED PA RT I C I PANT BIOGRAPHIES

Scott M. A u e r b a c hScott Auerbach is a theoretical chemist at UMass Amherst,having earned a PhD in physical chemistry at UC Berkeley in1993. He earned his BS (Summa cum Laude) in chemistryfrom Georgetown University in 1988, and was an NSFpostdoctoral fellow at UC Santa Barbara in chemistry andmaterials during 1994-1995. Since his arrival at UMassAmherst in 1995,Auerbach has been modeling adsorption,diffusion and reaction in zeolites, as well as self-assembly ofzeolites, all with a variety of atomistic modeling methods. Thegoal of Auerbach's research is to offer a powerful microscopeinto the factors that control selectivity in shape-selectiveprocesses mediated by zeolites. This effort has yielded twobooks, nearly 70 peer-reviewed articles, over 85 invitedlectures, and $1.8MM of extramural research funding.

Two of Auerbach's research areas that are most relevant tobiofuel production are microwave processing andfunctionalized materials. Microwave processing offers thepotential to revolutionize both the fabrication and use ofheterogeneous catalysts. Despite several reports ofremarkably short reaction times using microwaves, controversyremains regarding the actual effects of such microwaves.Auerbach applied molecular modeling to show that non-equilibrium energy distributions are possible and even likelyoutcomes of using microwaves, especially when ions anddipoles can absorb the radiation. This is especially promisingfor pre-treatment and reactions of biomass, which containssignificant oxygen, making biomass-derived feedstocks highlypolar species.

The second biofuel-related area of Auerbach's research isstudy of shape-selective basic catalysts. Such catalysts areimportant for controlling carbon-carbon bond-formingreactions of biomass-derived feedstocks. Auerbach has appliedperiodic density functional theory to show that amine-substituted zeolites may be good candidates for shape-selectivebasic catalysts. Future research will explore the mechanismsof addition and condensation reactions using such zeolites, andthe role of such zeolites during catalytic fast-pyrolysis.

Paul BlommelPaul Blommel is a lead catalysis researcher at Virent EnergySystems, Inc. He received a bachelor's degree in chemicalengineering in 1993 and a doctorate in biophysics in 2007 fromthe University of Wisconsin, Madison. Between 1993 and 2001,Blommel worked for UOP, conducting research and starting upa diverse number of oil refining, petrochemical, and hydrogengeneration processes. His doctoral work focused on high

throughput protein production and the characterization ofenzyme catalysts. At Virent, he directs research on catalyticupgrading of bio-sourced feedstocks into valuable products.

A . A . B o at e n gA.A. Boateng is a research scientist with the AgriculturalResearch Service (ARS), the intramural research arm of theUSDA. He started the ARS thermo chemical program in 2003and is a lead chemical engineer on their efforts to develop on-farm biorefinery systems for agricultural residues includingenergy crops. His prior experiences include academic positionsat the University of Guyana in South America and atSwarthmore College in Pennsylvania. In 1988 he was facultyFulbright Fellow from the Caribbean region at the Departmentof Chemical Engineering, Kansas State University, where hecontinued his research on the combustion and gasification ofrice hulls including the use of rice hull ash as a pozzolan forcement extension. His industrial experiences include processdesign and operation of high-temperature industrial furnacesused for minerals and materials processing. He spent 10 yearsin that industry optimizing confined turbulent diffusion flamesand the modeling of particulate flows in such operations. He isthe author of the book: Rotary Kilns – Transport Phenomenaand Transport Processes by Butterworth-Heinemannpublishers. He holds a PhD from the University of BritishColumbia.

André BoehmanAndré Boehman is a Professor of Fuel Science and MaterialsScience and Engineering in the Department of Energy &Mineral Engineering in the College of Earth and MineralSciences at the Pennsylvania State University, where he hastaught courses on Energy, Fuels, Combustion and theEnvironment since 1994. He holds a BS in MechanicalEngineering from the University of Dayton (1986) and an MS(1987) and PhD (1993) in Mechanical Engineering fromStanford University. He held a two-year postdoctoralfellowship in the Molecular Physics Laboratory at SRIInternational, Menlo Park, CA.

Prof. Boehman’s research interests are in alternative andreformulated fuels, combustion and pollution control. Hispresent research activities are focused on alternative dieselfuels, diesel combustion and diesel exhaust aftertreatment. Hewas recently appointed the Editor of the journal FuelProcessing Technology and has held executive positions withthe American Chemical Society Division of Fuel Chemistry andwith the International DME Association. He has received the2007 Faculty Mentoring Award from the Penn State College ofEarth and Mineral Sciences, the 1999 Alumni AchievementAward from the University of Dayton School of Engineering,the 1999 Matthew and Anne Wilson Award for Outstanding

Appendix C • Pa rticipant Biog r ap h i e s


Teaching from the Penn State College of Earth and MineralSciences and the Philip L.Walker Jr. Faculty Fellowship inMaterials Science and Engineering, from 1995-97. He alsoreceived the 1986 Charles T. Main Bronze Medal from theAmerican Society of Mechanical Engineers. He has supervisedseventeen MS theses and six doctoral theses at Penn State andhe has published more than 40 refereed papers and bookchapters. At the Penn State Energy Institute, Prof. Boehmanmanages the Diesel Combustion and Emissions Laboratory.

R o b e rt C. B row nDr. Robert C. Brown is the Bergles Professor in ThermalScience at Iowa State University. He holds the rank ofProfessor in the Departments of Mechanical Engineering,Chemical and Biological Engineering, and Agricultural andBiosystems Engineering. He is the director of theBiorenewable Programs and the Center for SustainableEnvironmental Technologies. Dr. Brown is an expert inthermochemical processing of biomass into energy, fuels, andchemicals. His research in biomass gasification includes studiesof carbon conversion, tar measurement and control, hot gasclean-up (particulate matter and inorganic contaminants),hydrogen production, and synthesis of renewable fuels andother biobased products using catalytic and biocatalyticprocesses. His research in fast pyrolysis includes studies onthe evolution and transport of pyrolysis vapors and aerosols,selective condensation of pyrolysis liquids, catalytic andbiocatalytic conversion of pyrolysis liquids into fuels andfertilizer, utilization of char byproduct as soil amendment andcarbon sequestration agent, and power systems based onpyrolysis liquids. Dr. Brown also performs techno-economicanalysis of bioenergy and biofuel systems. In 2003 hepublished Biorenewable Resources: Engineering New Productsfrom Agriculture, a textbook for students interested in theBioeconomy. Dr. Brown is a Fellow of the American Society of Mechanical Engineers.

S t even Chuang Steven Chaung is a professor in the Department of Chemicaland Biomolecular Engineering at the University of Akron,Akron, Ohio. He received a Ph.D. in Chemical Engineeringfrom University of Pittsburgh and joined the University ofAkron in 1986. His research interest is in the area of catalysisand reaction engineering. His research primarily involves useof in situ infrared spectroscopy and mass spectroscopy tocharacterize adsorbed species and intermediates underreaction conditions. His current research effort focuses on thedevelopment of anode catalysts for the direct hydrocarbon andcoal solid oxide fuel cells as well mechanistic studies ofphotocatalytic reactions, partial oxidation, and CH4 reformingreactions.

A n t h o ny M. D e a nTony Dean is the W. K. Coors Distinguished Professor in theChemical Engineering Department of the Colorado School ofMines. He received his bachelor’s degree from Spring HillCollege and his masters and doctorate in physical chemistryfrom Harvard University. He joined the Chemistry Departmentof the University of Missouri-Columbia in 1970, where hisresearch focused on shock tube studies of elementarycombustion-related reactions. In 1979 he moved to theCorporate Research Labs of Exxon Research and Engineering,where his efforts focused on the quantitative kineticcharacterization of gas-phase reaction systems. In particular,methods were developed to quantitatively characterize boththe temperature and pressure dependence of chemically-activated reactions. Detailed kinetic mechanisms weredeveloped and used to develop improved-fuel/advanced-engineconcepts to improve efficiency and decrease emissions ofinternal combustion engines. He joined the CSM faculty in2000.

At CSM the efforts of his research team focus on thequantitative kinetic characterization of a variety of reactionsystems. One such area involves the reactions that occur inhigh-temperature solid-oxide fuel cells (SOFCs). Gas-phasereactions in the fuel-channel produce intermediates that canprofoundly influence fuel cell performance.We seek tounderstand the detailed chemical pathways that lead toproduction of the smaller hydrocarbons as well as themolecular weight growth processes that can lead to depositformation within the anode channels.We also characterize thecatalytic reforming kinetics occurring within the porous anode.These reactions are strongly influenced by multiple transportprocesses including diffusion and reaction of the gaseoushydrocarbons as they migrate through the porous anode.Wethen include these gas-phase and catalytic reactions, as well asthe accompanying electrochemical reactions, into complete fuelcell models that can be used to characterize and optimizeSOFC performance.

Another research area, in conjunction with NREL researchers,concerns the production of fuels and power from biomass.Thespecific thrust of much of this research is to transfer ourlearnings of free radical chemistry in hydrocarbon systems tobiomass systems, especially in the context of the quantitativeimpact of the weakening of adjacent C–H bonds by the oxygenfunctionalities in biomass.

A p p e n d i c e s


Mark Dietenberg e rMark Dietenberger is a Research General Engineer with USDAForest Service at Forest Products Laboratory in MadisonWisconsin. He received a bachelor degree in AppliedMathematics and Physics at UW-Milwaukee in 1974 and Ph.D.degree in Mechanical Engineering at University of Dayton in1991, with an emphasis in heat and mass transfer includingspecialization in combustion. During 1975 to 1992 with theUniversity of Dayton Research Institute he developedspecialized computer models that include ice/frost formationon aircraft and space shuttle, air blast propagation overcomplex terrain, reconstruction of airplane crashes in severeweather, impact damage of various metals and ceramics duringarmor piercing, heat pipe performance for use in space, andthree-dimensional fire growth on an upholstered furniture.After receiving his Ph.D., he won an AFOSR grant to studyactive control of combustion approaching stoichiometricconditions in a turbine engine. He also developed a computerprogram to analyze kinetically the thermal destruction of toxicgases.

Since 1992 he became responsible for fire growth research atForest Products Laboratory. In this capacity he worked withvarious fire test equipments, improving their performance,developing new hardware, and modifying test protocols. Manyin wood industry now consider him as their foremost expertin fire growth modeling and experimentation involving woodproducts, particularly with fire retardant treatments. Hisrecent work on fundamentals of wood pyrolysis andcombustion using specialized tests as performed in the conecalorimeter and oxygen bomb calorimeter resulted in invitedpresentations within the fire safety science and the thermalanalysis communities. Most recently he is applying his firegrowth modeling expertise to fire development on exteriorbuilding surfaces and ornamental vegetations for the wildlandurban interface problems. His publications is provided on-lineat http://www.fpl.fs.fed.us

His recent assignment to lead the thermo-chemicallignocellulosic conversion research acknowledges Dr.Dietenberger’s expertise and reasserts the laboratory’shistorical leading role in the field. He has a patent pending ona new gasifier technology that promises to double or eventriple the current productivity of syngas. He is currentlymember of the Int.Association of Fire Safety Science andSociety of Fire Protection Engineers.

James A . D u m e s i cJames A. Dumesic earned his B.S. degree from UW-Madisonand his M.S. and Ph.D. degrees from Stanford University. HisPh.D. work was conducted under the supervision of ProfessorMichel Boudart. Dumesic then conducted post-doctoral

research as a U.S.-U.S.S.R. Exchange Fellow at the Institute ofChemical Physics in Moscow and as a NATO PostdoctoralFellow at the Centre de Cinetique Physique et Chimique ofFrance. Dumesic joined the Department of ChemicalEngineering in 1976. He served two terms as departmentchair. He has been the Shoemaker Professor of ChemicalEngineering, and he is currently the Steenbock Chair in theCollege of Engineering.Throughout his career, Dumesic hasused spectroscopic, microcalorimetric, and kinetic techniquesto study the surface and dynamic properties of heterogeneouscatalysts. Dumesic pioneered the field of microkinetic analysis,in which diverse information from experimental andtheoretical studies is combined to elucidate the essentialsurface chemistry that controls catalyst performance. He hasdeveloped microcalorimetric techniques to measure surfacechemical bond strengths for adsorbates on metal, oxide, andacidic catalysts. He is actively involved in the use of electronicstructure calculations to study the structures and reactivity’sof adsorbed species on metal and metal oxide surfaces.Dumesic’s research group is currently studying thefundamental and applied aspects of generating of hydrogen andliquid alkanes by aqueous-phase reforming of oxygenatedhydrocarbons derived from biomass, as well as the productionof intermediates for the chemical industry.

Dumesic has received a variety of awards and honors in thefield of catalysis and chemical engineering. He has beenrecognized with the Colburn Award and Wilhelm Award fromthe American Institute of Chemical Engineers, the EmmettAward from the North American Catalysis Society, andresearch excellent awards from the New York and Michigancatalysis societies. In 1998, he was elected to the NationalAcademy of Engineering. He has also been recognized for hisexcellence in teaching at the University of Wisconsin with aPolygon Award and the 1995 Benjamin Smith Reynolds Award.In 2002, he was given the Byron Bird Award in the College ofEngineering for Excellence in a Research Publication, citing hiswork in the microkinetics of heterogeneous catalysis. Hisresearch accomplishments were recognized in 2003 by theHerman Pines Award of the Chicago Catalysis Society. He wasnamed one of the Top 50 Technology Leaders of 2003 byScientific American, and he received the 2005 Cross CanadaLectureship Award of the Canadian Catalysis Society. In 2006,he received the Somorjai Award for Creative Research inCatalysis from the American Chemical Society, and thePhiladelphia Catalysis Club Award for excellence in catalysisresearch. In 2007, he was awarded the Burwell Lectureship inCatalysis by the North American Catalysis Society.

Dumesic has published more than 300 papers in peer-reviewed journals. Various information about researchconducted by the Dumesic group can be found at thefollowing link: http://www.engr.wisc.edu/che/



Douglas C. E l l i o t tMr. Elliott has 33 years of research and project managementexperience in the Battelle system at the Pacific NorthwestNational Laboratory (PNNL). His work has mainly beendirected toward development of fuels and chemicals frombiomass and waste. His experience is primarily in high-pressure batch and continuous-flow processing reactorsystems. This research has also involved him in extensivestudy of catalyst systems. In addition to process development,chemical and physical analysis has also been a significant partof his work. While at Battelle, Mr. Elliott’s research hasinvolved such subject areas as biomass liquefaction andhydroprocessing of product oils, catalytic hydrothermalgasification of wet biomass and wastewaters, and chemicalsproduction from renewable sources. His work in biomassliquefaction has involved him in the International EnergyAgency as a Task representative for the U.S. under theBioenergy Agreement. He also spent the summer of 1989under contract working at the Technical Research Centre ofFinland in Espoo on oil production from black liquor.

Mr. Elliott is a listed inventor on 15 U.S. patents and numerousrelated foreign patents. In 2004 he was designated a BattelleDistinguished Inventor. He has been recognized two timeswith R&D 100 awards for development of notable newtechnologies and an award from the Federal LaboratoryConsortium for Technology Transfer. He was a recipient of aGreen Chemistry Challenge Award in 1999. He is the authorof over 70 peer-reviewed journal publications and bookchapters. In addition, he has made over 30 presentations atnational and international meetings and conferences.

Calvin J. Fe i kCalvin Feik is a senior engineer with the National RenewableEnergy Laboratory (NREL) in Golden, Colorado and currentlymanages operation of the Thermochemical User Facility(TCUF). He has worked in thermochemical biomassconversion since starting at NREL in 1992 after receiving hisBachelors in Engineering from the Colorado School of Mines.His responsibilities have been in the design of unit operationsand integration with automated control and data acquisition.He was instrumental in the development and upgrading of theThermochemical Process Development Unit (TCPDU), a stateof the art 1/2-ton/day pilot plant for the production of syngasand pyrolysis oils from biomass. Current research in the TCUFis focused on developing a catalytic process to remove tarsfrom syngas formed during biomass gasification. Calvin has alsobeen involved in the development of several bench scalesystems for biomass conversion research.

Kristi FjareKristi Fjare got her undergraduate degree in Chemistry fromSt. Olaf college in Minnesota and PhD in Inorganic Chemistryfrom the University of Minnesota working with Professor JohnEllis.

Kristi has spent over twenty years in the petroleum andpetrochemical industry, doing research on catalysis, support tomanufacturing, and business development at Amoco, BP andConocoPhillips. She is currently a Principal Scientist atConocoPhillips in BioFuels R&D, with responsibility forthermochemical projects.

Manuel Fr a n c i s c oManuel Francisco received his BS from the Ohio StateUniversity in 1977. He attended the Massachusetts Instituteof Technology from 1977 to 1981 as an NSF fellow and earnedhis PhD in Synthetic Organic Chemistrywith Professor GeorgeBuchi. After graduation, he joined Exxon CorporateResearch Laboratory. Manuel's first seven years were focusedon basic research in the conversion and upgrading of heavyoils, resids and bitumens. In 1988 he took a loan assignmentat Exxon's Products Research Division and spent three and ahalf years in basic lubricant research.

Manuel returned to Corporate Research in 1992 then tookanother loan assignment to the Product Research andTechnology Division where he spent a year and a half inapplied lubricant research. He returned to CorporateStrategic Research in 2001 and since has worked on highthroughput lubricant research and basic research in conversionand upgrading of heavy oils, resids and bitumens.

A p p e n d i c e s


Anne M. G a f f n eyAnne Gaffney joined ABB Lummus Global in 2005 and iscurrently the Vice President of the Technology DevelopmentCenter in Bloomfield, N.J. In this role,Anne leads programs toimprove upon existing technologies and to provide newinnovations in support of the refining, petrochemicals, olefinsand JV business groups.

Before joining ABB Lummus Global,Anne held seniortechnology roles at Rohm and Haas, DuPont and Lyondell inthe areas of catalysis, process chemistry, selective oxidation,and materials synthesis & characterization. Anne is theauthor/co-author of over 100 patents/patent applications andover 50 technical publications. She has held leadership rolesin ACS (Chair of the Petroleum Chemistry Division and theCatalysis Secretariat), the North American Catalysis Society(Chair of the 19th NAM and National Representative for thePhiladelphia Catalysis Club) and the Organic ReactionCatalysis Society (Director and Editorial Board Member).

Anne received her bachelor’s degree in chemistry andmathematics from Mount Holyoke College and herdoctorate’s degree in physical organic chemistry from theUniversity of Delaware. She is the 2007 recipient of theTribute to Women in Industry Award, the 2006 DOW/UnionCarbide Reaction Engineering and Catalysis Seminar SeriesAward and the 1999 Catalysis Club of Philadelphia Award.

D r. Santosh K Gangwa lDr. Santosh K Gangwal is presently a senior chemical engineerwith the Center for Energy Technology at Research TriangleInstitute International, Research Triangle Park, NC. Hereceived his PhD in chemical engineering from the Universityof Waterloo in Canada in 1977. Over the past 30 years, hehas procured and successfully managed projects totaling morethan $30 million. He is presently responsible for developingand managing projects in cleanup and conversion of biomass-and coal-derived syngas to fuels and alcohols. He is directingexperimental studies for the synthesis of novel attritionresistant Fischer-Tropsch (FT) catalysts using a spray dryer andtesting these catalysts in microreactor, CSTR and SBCR. He isalso scaling up the FT catalyst to a barrel/day plant. He hasevaluated modified base-metal oxide catalysts for synthesis ofethanol from methanol and/or syngas. He is also synthesizingnovel FCC catalysts and developing a novel triple functionFCC-type catalyst and reactor system to remove tar ,ammonia, and H2S from biomass gasifier gas. He is alsodeveloping a nano-particle iron material for production ofhydrogen from syngas via the steam-iron process and a nano-particle nickel hydrogenation catalyst. He is also developingCO2 sorbents for syngas as well as flue gas. He is alsoassisting in the development of a heterogeneous catalyst-based

process for conversion of free fatty acids to biodiesel and hasinvestigated the development of a process to convertvegetable oil to jet fuel.

Previously, he spearheaded the development of aninternationally recognized syngas desulfurization program atRTI International that grew into the Center for EnergyTechnology. He developed fixed-bed, moving-bed and fluidized-bed desulfurization sorbents and ammonia decompositioncatalysts. He scaled up a patented zinc oxidedesulfurization sorbent to 10,000 lb for testing using a syngasslipstream from the Eastman Chemical Company gasifier. Healso developed and tested mixed-metal oxide and carbon-based catalysts for demercaptanization and desulfurization ofdiesel and jet fuel and prepared and evaluated cobalt- andiron-based catalysts for Fischer-Tropsch synthesis at industriallyrelevant conditions for more than 20,000 hours. He alsoevaluated catalysts for deep catalytic oxidation of VOCmixtures and conducted sulfur balances around PC boilers todetermine sulfur capture using lime-based materials. He hasalso been active in University teaching, having taught bothundergraduate and advanced chemical engineering courses atUniversity of Maine, University of New Hampshire, and NorthCarolina A&T State University. He won the prestigious R&D100 award in 2004 for inventing and scaling up adesulfurization sorbent for syngas. He holds 10 US patents.He has published 45 refereed papers and made over 150presentations in scientific conferences.

James G. G o o dw i n , J r.Professor James G. Goodwin, Jr., Professor and Chairman ofthe Department of Chemical Engineering at ClemsonUniversity, is an internationally recognized expert in the fieldsof heterogeneous catalysis and reaction kinetics. He is bestknown for his research in the areas of Fischer-Tropschsynthesis, the use of isotopic tracing to study reactions at thesite level, and now biodiesel synthesis using solid acid and basecatalysts. In 1976 he received his PhD in chemical engineeringat the University of Michigan. After spending 18 months as anNSF-CNRS exchange scientist at the French Institute ofCatalysis and 1 year as an assistant professor of engineering atthe University of South Carolina, Professor Goodwin joinedthe University of Pittsburgh in 1979 where he rose in theranks to become William Kepler Whiteford Professor ofChemical Engineering. In 2000, he moved to ClemsonUniversity to become Chairman of Chemical Engineering.Most of Professor Goodwin's 173 refereed scientificpublications are related to supported metal and solid acidcatalysis. He is also an author of 12 U.S. (1 pending) and 19international (3 pending) patents in the area of Fischer-Tropschsynthesis. He has consulted extensively over the past 19 yearswith Altamira Instruments, Gas-to-Oil (Statoil), Energy



International (Williams), Hampton University, and the U.S.Department of Justice. Dr. Goodwin's research group isinternationally known for the study of hydrocarbon catalysisand for the use of isotopic tracing (SSITKA) to study surfacereactions. His current research focuses on biomass conversion(biomass gas clean-up, biodiesel synthesis, Fe Fischer-Tropschsynthesis), coal conversion (selective synthesis of ethanol fromsyngas), and the effect of impurities on PEM Fuel Celloperation.

G e o rge W. H u b e rGeorge W. Huber is an Assistant Professor of ChemicalEngineering at University of Massachusetts-Amherst whoseresearch focus is on Breaking the Chemical and EngineeringBarriers to Lignocellulosic Biofuels. He has authored 25 peer-reviewed publications including two papers in Science andthree articles in Angewandte Chemie International Edition. Hehas five patent applications pending in the area of biofuels andhas worked on catalysis projects for Exxon-Mobil, Conoco-Phillips, and Cargill. George’s discovery of Raney-NiSn catalystfor hydrogen production from biomass-derived oxygenateswas named as one of top 50 technology breakthroughs of2003 by Scientific America. His research is beingcommercialized by two different start-up companies (VirentEnergy Systems and Bio-e-con).

Prior to his appointment at UMass-Amherst George did apost-doctoral stay with Avelino Corma at the TechnicalChemical Institute at the Polytechnical University of Valencia,Spain (UPV-CSIC) where he studied bio-fuels production usingpetroleum refining technologies. He obtained his Ph.D. inChemical Engineering from University of Wisconsin-Madison(2005) where he helped develop aqueous-phase catalyticprocesses for biofuels production under the guidance of JamesA. Dumesic. He obtained his B.S. (1999) and M.S.(2000)degrees from Brigham Young University, where he studiedFischer-Tropsch Synthesis under the direction of Calvin H.Bartholomew.

Christopher W. J o n e sChristopher Jones obtained his BSE degree in ChemicalEngineering from the University of Michigan in 1995, followedby graduate studies at Caltech leading to MS and PhD degreesin 1997 and 1999, respectively. During his graduate studies atCaltech, his research activities spanned from the synthesis andcharacterization of new materials to zeolite catalysis. Followinga postdoctoral year at Caltech studying organometallicchemistry and catalysis, he joined Georgia Tech as an AssistantProfessor in Chemical Engineering in 2000.Today, he is anAssociate Professor and the J. Carl & Sheila Pirkle FacultyFellow in the School of Chemical & Biomolecular Engineering.

At Georgia Tech, Dr. Jones leads a research group that worksin the broad areas of materials, catalysis and separations. Inparticular, his group currently studies (i) the generation of amolecular-level understanding of supported organic andorganometallic catalysts, (ii) the conversion of biomass intofuels and chemicals, and (iii) the engineering of materials forlow energy adsorption or membrane separations. Working atthe interface of synthetic chemistry and chemical engineering,the Jones Group uses advanced organic, organometallic andinorganic synthetic methods to create unique, new functionalmaterials.

In the area of biomass conversion, Dr. Jones’ work focuses onthe use lignocellulosic biomass as a feedstock for liquidtransportation fuels or hydrogen. These biomass resources,particularly softwoods, are being systematically studied toassess their suitability as raw materials for downstreamcatalytic upgrading. Modern homogeneous and heterogeneouscatalytic methods are being used to develop reaction paths toenergy-dense, deoxygenated liquids.

A l exander Kat zAlexander Katz received his BS and MS degrees in chemicalengineering at the University of Minnesota, and was awarded aFannie and John Hertz Foundation Fellowship for doctoralwork at California Institute of Technology. He conductedpostdoctoral studies in supramolecular chemistry inStrasbourg, France as a NSF International Awards PostdoctoralFellow, and was subsequently appointed Assistant Professor ofChemical Engineering at the University of California, Berkeleyin 2000.

Since that time, he has begun an interdisciplinary researchprogram focused on the design and synthesis of functionalnanoscale interfaces in hybrid organic-inorganic materials,relying on molecular templating strategies. He has beenawarded four patents, covering his different research areas, anda Young Scientist Prize from the International Association ofCatalysis Societies for his discovery of grafted calixarene (Cal-Silica) materials. Katz can be reached [email protected].

A p p e n d i c e s


H a rold KungHarold H. Kung is professor of chemical and biologicalengineering at Northwestern University. He received his B.S.degree in chemical engineering from the University ofWisconsin, Madison, and his Ph.D. degree in chemistry fromNorthwestern University.After spending two years at theCentral Research and Development Department at DuPont, hestarted his academic career at the Chemical EngineeringDepartment at Northwestern University. His main researchinterest has been in the area of heterogeneous catalysis, and isknown for work on catalysis for selective hydrocarbonoxidation, deNOx, and hydrocarbon cracking, and catalysis bynanosize Au particles and oxide materials. Recently, his groupacquired significant experience in solution preparation ofvarious nanostructures such as dendrimers, vesicles, and 2-Dpatterns (by lithography), and has synthesized functionalizednanocages as new catalytic structures. He has authored amonograph,Transition Metal Oxides: Surface Chemistry andCatalysis. He is an editor of Applied Catalysis A: General, andrecipient of the P.H. Emmett Award and the Robert BurwellLectureship Award (North American Catalysis Society), theHerman Pines Award (Chicago Catalysis Club), CatalysisSociety of South Africa Eminent Visitor Award, and Cross-Canada Lectureship of the Catalysis Division of the ChemicalInstitute of Canada. He is a fellow of American Association forthe Advancement of Science, and has published over 200papers.

Angelo LuciaAngelo Lucia earned a Ph.D. in Chemical Engineering at theUniversity of Connecticut and is presently the Chester H. KirkProfessor of Chemical Engineering at the University of RhodeIsland, a position he has held since 1996. Prior to that,Professor Lucia was a member of the faculty at ClarksonUniversity in Potsdam, NY for fifteen years. His researchinterests are in the general area of process engineering, withparticular focuses in process modeling, synthesis, design, andoptimization and molecular modeling of fuels. Dr. Lucia alsohas a strong interest in mathematical analysis and currentfunded research interests in energy conservation and energyefficient design.

Specific areas of applied interest most relevant to the BioFuelsWorkshop include the design of energy efficient separations bydistillation and hybrid separation strategies as well as themolecular understanding and phase equilibrium oflignocellulosic materials.

D evinder MahajanProfessor Mahajan holds one of the ten joint appointmentsbetween Brookhaven National Laboratory and SUNY at StonyBrook. Dr. Mahajan’s professional goal is to bridge science andtechnology for the benefit of mankind.To achieve this goal, hisresearch interests focus on Energy issues that includes aportfolio of projects on Methane hydrates, H2 production, FuelCells, natural gas, Biomass and coal utilization via Fischer-Tropsch, Methanol, and mixed alcohol synthesis using soluble(single-site) or slurried (nano heterogeneous or colloidalphase) based catalysts, and extraction of value minerals fromgeothermal brines. Scientifically, his work in the synthesis ofclean fuels is internationally recognized. He has organizedsymposia and international workshops on issues such as CleanFuels, Methane Hydrates, and Biomass and has served or nowserving as a Guest Editor of three recent special volumes:Topics In Catalysis (2005), Journal of Petroleum Science &Engineering (2006) and Industrial & Engineering ChemistryResearch (2006-07). He is the author of over 75 publicationsincluding book chapters and encyclopedia articles, 10 patents,and presented over 110 invited lectures at various institutions,conferences and workshops around the globe. His work isconstantly covered through press releases. He serves onseveral national and international energy-related committees.In March 2006, he was recognized with a membership to theprestigious Russian Academy of Natural Sciences (RANS)-USSection and is a recipient of the RANS Crown and Eagle Medalof Honor for service to the field of “Petroleum Engineering”.

At Stony Brook University, he was instrumental in setting upthe Chemical & Molecular Engineering (CME) Undergraduateprogram. He is a member of various external and internalcommittees including the ABET committee at SBU. ProfessorMahajan’s is playing an active role in setting up the newlyfunded New York State Advanced Energy Research andTechnology Center (AERTC) at Stony Brook.As a Professorand Co-Director of the CME program at Stony Brook U., hispriority is to further integrate education and research at bothundergraduate and graduate level, foster collaboration withinthe university with a goal to train students in the next-generation energy technologies.



Leo Manze rLeo E. Manzer is founder and President of Catalytic InsightsLLC, a consulting company in the field of catalysis and processresearch. His clients cover a number of large and smallcompanies in Europe and North America and he serves on theScientific Advisory Boards of several startup companies such asRange Fuels and Segetis.

Manzer was born and educated in Canada and after receivinghis Ph.D. in chemistry from the University of Western Ontario,he joined the DuPont Company in Wilmington, DE. He retiredfrom the DuPont Company in 2005 as a DuPont Fellow, aposition currently held by only 15 out of 3,000 scientists andengineers. During his 30+ year career at DuPont, he foundedand directed the Corporate Catalysis Center, led DuPont’sR&D effort to replace chlorofluorocarbons (CFCs) with ozonefriendly substitutes, led Conoco’s Fischer-Tropsch catalysisprogram, and played a key role in DuPont’s efforts to developchemicals from renewable feedstocks.

Manzer is the author of 89 publications and >114 US patents(> 500 international patents). He has received a number ofawards, including: the 1995 ACS Earle B. Barnes Award; the1997 Catalysis Club of Philadelphia Award; the 1997 ACSHeroes of Chemistry Award; the 1998 Cross-Canada LectureTour Award by the Catalysis Division of the Chemical Instituteof Canada; the 2001 Eugene J. Houdry Award for AppliedCatalysis from the North American Catalysis Society; and the2003 ACS E.V. Murphree Award. He was also a member of theDuPont team recognized for the 2002 Presidential NationalMedal of Technology Award for his work in developing CFCAlternatives. He has served on the advisory board on manyjournals, including the Journal of Catalysis, Catalysis Today andApplied Catalysis.

R i c h a rd MarinangeliRichard Marinangeli has a Ph.D. in Chemical Engineering fromPrinceton University and a B.S. in Chemical Engineering fromthe University of Notre Dame. He was a CNRS Fellow inVilleurbanne, France prior to joining UOP. He has 28 years ofexperience at UOP in Process Development and application ofmaterials for catalysis and adsorption This work has resultedin 18 U.S. patents. Currently, Dr. Marinangeli is Manager of theRenewable Energy and Chemicals Group in the RefiningConversion Development Group.

S t even PhillipsMr. Phillips has over 22 years of experience in thermochemicalprocess research primarily in the conversion of biomass intofuels and power. In 1993, Steven joined NREL working inbiomass conversion to and upgrading of pyrolysis oil. In 1994,he and the Thermochemical Users Facility (TCUF) Team atNREL started design and construction of the ThermochemicalProcess Development Unit (TCPDU), a state-of-the-art, ?-ton/day pilot plant for producing syngas and pyrolysis oils froma wide variety of biomass feedstocks. Research emphasis hasbeen on developing and demonstrating biomass conversiontechnologies that will overcome the barriers to successfulcommercial deployment of renewable biomass energy. In late2005, Steven moved to the Systems Analysis Team in theNational Bioenergy Center. His emphasis over the past twoyears has been the technoeconomic modeling of biomass-to-liquid fuels processes. Most recently, he was lead author on abiomass to ethanol design report for a 2000 tonne/dayprocess.

Prior to joining NREL, Steven worked at the Dow ChemicalCompany investigating catalytic fluidized bed processes formaking styrene, and at National Semiconductor Corp.developing photolithographic processes for semiconductorproduction. His graduate school research involved collecting,organizing and characterizing combustion data for evaluatingthree-dimensional CFD models. He received both a B.S. andM.S. in Chemical Engineering from Brigham Young University.

Kyoung S. R o, P h . D. , P. E .Dr. Ro is an environmental engineer at the USDA-ARS CoastalPlains Soil,Water & Plant Research Center, Florence, SC.Before joining the USDA-ARS, he had worked as a facultymember for the Louisiana State University (LSU) and the CityCollege of New York (CCNY) for 13 years. His research focusincludes agricultural and municipal wastewater treatment, fateand transport of pollutants and greenhouse gases, and thermo-chemical/biological conversion of biomass and wastes intobioenergy.

A p p e n d i c e s


L a n ny D. S c h m i d tLanny D. Schmidt was born on May 6, 1938 in Waukegan,Illinois. He is married and has two children. He received aBachelor of Science degree in Chemistry in 1960 fromWheaton College and a Ph.D. degree in Physical Chemistry in1964 from the University of Chicago, where he was awarded aNational Science Foundation Graduate Fellowship. His thesison alkali metal adsorption was supervised by Robert Gomer.

After a postdoctoral year at the University of Chicago, hejoined the Chemical Engineering Department at the Universityof Minnesota where he is now Regents Professor in theDepartment of Chemical Engineering and Materials Science.

Professor Schmidt's research focuses on various aspects of thechemistry and engineering of chemical reactions on solid surfaces. Reaction systems ofrecent interest are catalytic combustion processes to produce products such as syngas,olefins, oxygenates by partial oxidation and NOx removal andincineration by total oxidation. One topic of research is characterization of adsorption and reactions on well definedsingle crystal surfaces. The second topic is steady state andtransient reaction kinetics under conditions from ultrahighvacuum to atmospheric pressure. The third topic ischaracterization of small particles and the correlation ofcatalytic activity with particle microstructure. The fourth topicis catalytic reaction engineering in which detailed models ofreactors are constructed to simulate industrial reactorperformance, with particular emphasis on chemical synthesis and on catalytic combustion.

Professor Schmidt has published over 300 papers in refereedjournals. He has supervised approximately 70 Ph.D. theses and 15 M.S. theses at Minnesota,and 11 of his former students hold university teachingpositions. He is a member of the National Academy ofEngineering.

B rent ShanksBrent Shanks is a professor in the Department of Chemicaland Biological Engineering at Iowa State University. Hereceived his bachelor’s degree in Chemical Engineering at IowaState University. In 1988, he received a doctoral degree inChemical Engineering from the California Institute ofTechnology. Following graduate school, he worked as aResearch Engineer and then Department Manager in catalystresearch and development for Shell Chemical Company. Hejoined the faculty at Iowa State University in 1999.

The Shanks research group is working on the design ofmaterials for use as heterogeneous catalysts with particularemphasis on their application to the conversion ofbiorenewable feedstocks to chemicals and fuels. He is afounding member of the Office of Biorenewable Programs atIowa State University and has been actively involved indeveloping interdisciplinary research and education programsfocused on biorenewables.

Chunshan SongDr. Chunshan Song is a Professor of Fuel Science and Directorof the Energy Institute at Pennsylvania State University. Hisresearch interests include catalysis and adsorption for fuelprocessing, desulfurization of fuels, reforming of alcohols andhydrocarbons for fuel cells, shape-selective catalysis forchemicals, CO2 capture and utilization, and conversion of coal,petroleum and biomass.

He has published 160 refereed papers, edited 11 books andspecial issues of catalysis journals, delivered over 35 plenary orkeynote lectures at international conferences, and given 150invited lectures worldwide. He has won some awardsincluding Herman Pines Award for Outstanding Research inCatalysis from Catalysis Club of Chicago and UOP; FulbrightDistinguished Scholar from the US-UK Commission; ChangJiang Scholar from Ministry of Education of China; Most CitedAuthors 2002-2006 and Top Cited Article Awards in catalysisfrom Elsevier;Wilson Award for Excellence in Research, andFaculty Mentoring Award at Penn State; Outstanding ScholarOverseas from Chinese Academy of Sciences; DistinguishedCatalysis Researcher Lectureship from Pacific NorthwestNational Lab; Outstanding Service Awards from InternationalPittsburgh Coal Conference (PCC) and from AmericanChemical Society (ACS) Fuel Chemistry Division; NEDOFellowship and AIST Fellowship Awards from Japan; InventorIncentive Awards and Materials Science and EngineeringService Award at Penn State.

He served as the Chair of ACS Petroleum Division, Chair-Elect(2007) and Program Chair for ACS Fuel Division, Chair of thePCC Advisory Board, Organizing Committee for NorthAmerican Catalysis Society Meetings, and Chair/Co-chair for25 national/international symposia. He is on the advisory boardfor ACS journal Energy & Fuels, international journals CatalysisToday, Research on Chemical Intermediates;Acta PetroleiSinica-Petroleum Processing, Journal of Fuel Chemistry andTechnology, and Coal Conversion. He held VisitingProfessorships in Imperial College London, University of ParisVI,Tsinghua University, Dalian University of Technology, andChinese Academy of Sciences.



Philip SteeleDr. Philip Steele has been a Professor in the Dept. of ForestProducts, College of Forest Resources, Mississippi StateUniversity (MSU) for 17 years with both research and teachingduties. He joined MSU after receiving his doctorate in WoodScience and Technology at MSU in 1986. Dr. Steele is Leader ofBio-Oil Research Thrust Area in the MSU Sustainable EnergyResearch Center (SERC) and is manager of the Bio-OilLaboratory at MSU. He has directed the development of aunique auger-fed laboratory-scale bio-oil reactor in acooperative development with the Renewable OilsInternational, LLC a manufacturer of auger pyrolysis reactors.Based on development of the auger bio-oil reactor researchersin the Bio-Oil Research Thrust Area have attracted substantialfunding from USDA and DOE grants to develop fuels andchemical products from bio-oil. The MSU Bio-Oil ResearchThrust Area is comprised of 10 on-campus facultycollaborators in the departments of Agricultural and BiologicalEngineering, Chemistry, Chemical engineering, Forest Productsand the Institute for Clean Energy Technology.This group isdeveloping technology for the production of fuels and specialtychemicals from bio-oils made from various types of forestryand agricultural feed stocks.As a researcher in the SERC Dr.Steele is project leader of the research group investigatingcatalytic upgrading of bio-oils by hydrogenation totransportation fuels and chemicals.

Dr. Steele has won several research awards including the MSUCollege of Forest Resources Outstanding Research Award andawards for exceptional research papers from both theHardwood Research Council and the Forest Products Society.He is the author or co-author of over 100 research papers.Dr. Steele holds three patents with one of his patented devicescommercialized and marketed internationally.

Jim Steve n sJim Stevens has worked for Chevron since 1981 and iscurrently the pathway manager for emerging technologyroutes for biomass conversion. He received his bachelor’s(1973) and doctoral degrees (1977) in chemistry from RiceUniversity. His work in the petroleum industry includesdevelopment of processes for production from mined heavyoil sands via solvent extraction and pyrolysis. He hasconducted R&D and managed groups involved in carbondioxide and steam enhanced oil recovery. For the last severalyears he has been involved in the development of technologyneeded to commercialize distributed production of hydrogen.Work in this area included catalyst development andevaluation, integration of reforming and separation processes,and sensor development. He is the holder of many U.S. andforeign patents in the area of oil recovery and hydrogenproduction.

Thomas Henry Va n d e r s p u rtThomas Henry Vanderspurt is a Fellow at United TechnologiesResearch Center. He earned his B.S., Chemistry, at LowellTechnological Institute, (now U. Massachusetts –Lowell) in1967 then studied under Prof. John Turkevich and completedhis Chemistry Ph. D at Princeton University, in 1972.After theU.S.Army Chemical Officers training and Post-doctoralresearch at Princeton he joined the Catalyst research group ofthe Celanese Research Co. in Summit, NJ in 1973. He leftCelanese 1979 for Oxirane International; Princeton NJ to builda world class catalysis group, however when Oxirane was soldin 1980 he joined Exxon Corporate Research Laboratories.There as a Research associate/Group Head he was responsiblefor advanced catalytic routes, including new Fischer-Tropschcatalysts, for converting coal derived syn-gas to hydrocarbonstransportation fuels and ammonia.

With the changing economy he then led the development oflight virgin naphtha aromatization over Pt/KL zeolite. He wasTechnical Lead and Section Head for Exxon Basic Chemicals1985-1988 during the commercializing of EXAR aromatizationtechnology. Returning to Exxon’s corporate labs in 1988 hetook on numerous assignments investigating non-FischerTropsch conversion of stranded natural gas, methane, toliquids.This included methane to light alcohols/isobutanoleffort. While at Exxon he co-chaired the 16th Meeting of theNorth American Catalysis Society in Boston in 1999, servedon the Council for Chemical Research Vision 2020 CatalysisImplementation Team, and organization a multi-year, multi-research group, academic study of sulfur tolerant three waycatalysts for the Auto-Oil Cooperative Research Council.

In 2000 the rebirth of the Fuel Cell effort at UnitedTechnologies Research Center brought him to East Hartford.Presently he is concerned with efforts to discover andcommercialize advanced catalytic systems for Biomassconversion, fuel processing, fuel cell electro-catalysts, sulfurtolerant hydrogen separation membranes, UV-photocatalyticoxidation of organic contaminants in air and relatedtechnology. He has 41 US Patents, 15 patent applicationspending, numerous presentations, publications, and foreignpatents.

A p p e n d i c e s


B ruce V r a n aBruce Vrana is a Senior Consultant in the EngineeringEvaluations and Sustainability Group at DuPont. He earnedhis B.S. (1980) and M.S. (1982) degrees in ChemicalEngineering from the University of Pennsylvania, and hisM.B.A. from the University of Houston (1984). He is aregistered Professional Engineer in the state of Delaware.Since joining DuPont in 1980, he has worked in research,manufacturing, and in the corporateengineering technology function. For the past 23 years as aninternal consultant, he has worked with research andmanufacturing teams from a wide variety of DuPontbusinesses, aimed at transforming research discoveries intocommercial successes through technoeconomic evaluations,process conceptualization and process synthesis techniques.For over a decade, he worked on nearly every importantprogram in the nylon intermediates business. In recent years,he has worked on the DuPont cellulosic ethanol(ICBR -Integrated Corn BioRefinery) and biobutanol research anddevelopment programs, playing a key role in the engineeringteam for both projects.

Bruce works with the University of Pennsylvania senior designcourse in CBE, proposing projects for groups of students towork on, as well as advising student groups working on bothhis and other projects. He has also served at RowanUniversity in a similar role. He teaches the internal DuPontcourse on engineering economics as well as a continuingeducation course on the same topic at the University ofDelaware.

Yong Wa n gYong Wang received his M.S. and PhD degrees in ChE fromWashington State University in 1992, and 1993. He joinedPacific Northwest National Laboratory (PNNL) in 1994 as apostdoc fellow. He became a research engineer in 1996, asenior research engineer in 1997, a chief scientist in 2000, anda laboratory fellow in 2005. He currently manages Catalysis &Reaction Engineering team of >20 technical staffs. Hisresearch interest is in the development of novel catalyticmaterials and innovative reaction engineering such asmicrochannel reactors, structured monolith reactors andmembrane reactors for hydrocarbon and biomass conversionsto fuels/chemicals; fundamental studies of structural andfunctional relationship of early transition metal oxide andbimetallic catalysts; novel and durable cathode materials forPEM fuel cell applications. His discoveries in microchannelreaction technology led to the spin-off of Velocys, a whollyowned subsidiary of Battelle in the commercialization ofmicrochannel technology.

Dr.Wang was named as the 2006 Asian American Engineer ofthe Year by Chinese Institute of Engineers. He has receivedtwo R&D100 awards (1997 and 1999) and a PresidentialGreen Chemistry Award in 1999. He was twice named PNNLInventor of the Year in 2004 and 2006. He was honored as aBattelle Distinguished Inventor in 2004. He is also a first-timerecipient of PNNL Laboratory Director’s Award forExceptional Scientific Achievement in 2005. He is an AdjunctProfessor in ChE at Washington State University, a GuestProfessor at Tianjin University (China), Sichuan University,Dalian University of Technology, and Dalian Institute ofChemical Physics. He currently serves editorial board ofCatalysis Today and Journal of Nanomaterials. He also servesas the Program Committee Chair of ACS (American ChemicalSociety) Petroleum Division (2006-2008). He has organizednumerous international and national conferences. He hasmore than 100 peer reviewed publications, 48 issued U.S.patents, and one book edited on Microreactor and ProcessIntensification (published in 2005).

Phil We s t m o re l a n dPhil Westmoreland is Professor of Chemical Engineering at theUniversity of Massachusetts Amherst, temporarily serving atNSF as Program Director for Combustion, Fire, and PlasmaSystems in the ENG/CBET division and as co-leader of thedivision's cyber infrastructure initiatives.

His research activities focus on elementary reaction kinetics inflames and polymer pyrolysis using molecular-beam massspectrometry, computational quantum chemistry, reactive-flowmodeling, and thermal analysis methods. He is using thesetools to examine fast pyrolysis methods on biomass and tostudy combustion chemistry of biofuels. His Pyroprobe/GC-MS instrument developed for studying pyrolysis of syntheticpolymers also reveals clues to mechanistic chemistry ofbiomass pyrolysis. Likewise, his group's Reactive MolecularDynamics method is very promising for predicting thesepathways. In a collaborative project at Lawrence BerkeleyNational Laboratory, he and his colleagues have mappedconcentrations of stable species and free radicals in flat flamesburning small-molecule alcohols, ethers, aldehydes, and esters.These results and their modeling provide qualitative andquantitative insights into their reaction kinetics.

He is a Fellow of AIChE; was founding president of AIChE'sComputational Molecular Science and Engineering Forum(CoMSEF); serves on the Board of Directors of theCombustion Institute, the Council for Chemical Research, andthe nonprofit CACHE Corporation; and has been recognizedby AIChE's Lappin Award, Lawrence Berkeley Laboratory'sShirley Award,ASEE Corcoran Award, and the UMass AmherstCollege of Engineering's Outstanding Senior Faculty Award.



C h a rles Wy m a nCharles Wyman has devoted most of his career to leading theadvancement of technology for biological conversion ofcellulosic biomass to ethanol and other products. In the fall of2005, he joined the University of California at Riverside as theFord Motor Company Chair in Environmental Engineering.Prior to that, he was the Paul E. and Joan H. QueneauDistinguished Professor in Environmental Engineering Designat the Thayer School of Engineering at Dartmouth Collegewhere he continues as an Adjunct Professor. Dr.Wyman isalso the Chief Development Officer, cofounder, and chair ofthe Scientific Advisory Board for Mascoma Corporation, a newstartup company focused on biomass conversion to ethanoland other products. Before joining Dartmouth College in thefall of 1998, Dr.Wyman was Director of Technology for BCInternational and led process development for the firstcellulosic ethanol plant planned for Jennings, Louisiana.Between 1978 and 1997, he served as Director of theBiotechnology Center for Fuels and Chemicals at the NationalRenewable Energy Laboratory (NREL) in Golden, Colorado;was Director of the NREL Alternative Fuels Division andManager of the Biotechnology Research Branch; and heldseveral other leadership positions at NREL, mostly focused onR&D for biological conversion of cellulosic biomass to fuelsand chemicals. He has also been Manager of ProcessDevelopment for Badger Engineers, an Assistant Professor ofChemical Engineering at the University of New Hampshire, anda Senior Chemical Engineer with Monsanto Company. Wymanhas a BS degree in chemical engineering from the University ofMassachusetts, MA and PhD degrees in chemical engineeringfrom Princeton University, and an MBA from the University ofDenver. He has authored over 80 peer-reviewed papers andbook chapters, made more than 50 presentations forpublication and more than 150 other presentations, manyinvited, written over 30 technical reports, chaired numeroustechnical meetings and sessions, edited 9 symposiumproceedings, edited a book on biomass ethanol technology, andbeen awarded 12 patents. He is also on the editorial board ofseveral technical journals and the board of directors or boardof advisors for several organizations and institutions.

Ye XuDr. Ye Xu is a member of the research staff of the Center forNanophase Materials Sciences and the Chemical SciencesDivision at Oak Ridge National Laboratory, where heinvestigates chemical reactions occurring on metals, metalcompounds and clusters by using first-principles theoreticalmethods. His research is focused on gaining fundamentalunderstanding of heterogeneous catalytic processes in orderto design and improve catalytic materials.

RESEARCH ROADMAP REPORTBASED ON THE

JUNE 25-26, 2007 WORKSHOPWASHINGTON, D.C.

TECHNICAL WRITING AND EDITING:LOREN WALKERL2W RESEARCHAMHERST, MA

DESIGN:LORI LYNN HOFFERWATERLILY DESIGN

LEVERETT, MA

roadmap2-08 next generation hydrocarbon bio refineries

Documents