production of synthetic natural gas from carbon dioxide and … · 2020. 4. 15. · hydrogen....

William L. BeckerDepartment of Mechanical Engineering,

Colorado School of Mines,

1610 Illinois Street,

Golden, CO 80401

e-mail: [email protected]

Michael PenevNational Renewable Energy Laboratory,

15013 Denver West Parkway,

Golden, CO 80401


Robert J. Braun1Department of Mechanical Engineering,

Colorado School of Mines,

1610 Illinois Street,

Golden, CO 80401


Production of Synthetic NaturalGas From Carbon Dioxide andRenewably Generated Hydrogen:A Techno-Economic Analysisof a Power-to-Gas StrategyPower-to-gas to energy systems are of increasing interest for low carbon fuels productionand as a low-cost grid-balancing solution for renewables penetration. However, such gasgeneration systems are typically focused on hydrogen production, which has compatibil-ity issues with the existing natural gas pipeline infrastructures. This study presents apower-to-synthetic natural gas (SNG) plant design and a techno-economic analysis of itsperformance for producing SNG by reacting renewably generated hydrogen from low-temperature electrolysis with captured carbon dioxide. The study presents a “bulk”methanation process that is unique due to the high concentration of carbon oxides andhydrogen. Carbon dioxide, as the only carbon feedstock, has much different reactioncharacteristics than carbon monoxide. Thermodynamic and kinetic considerations of themethanation reaction are explored to design a system of multistaged reactors for the con-version of hydrogen and carbon dioxide to SNG. Heat recuperation from the methanationreaction is accomplished using organic Rankine cycle (ORC) units to generate electricity.The product SNG has a Wobbe index of 47.5 MJ/m3 and the overall plant efficiency (H2/CO2 to SNG) is shown to be 78.1% LHV (83.2% HHV). The nominal production cost forSNG is estimated at 132 $/MWh (38.8 $/MMBTU) with 3 $/kg hydrogen and a 65%capacity factor. At U.S. DOE target hydrogen production costs (2.2 $/kg), SNG cost isestimated to be as low as 97.6 $/MWh (28.6 $/MMBtu or 1.46 $/kgSNG).[DOI: 10.1115/1.4041381]

Introduction

There has been much interest in hydrogen production to supportfuture energy and economic systems which rely on hydrogen asan energy carrier. Hydrogen can be produced by renewably pow-ered electrolysis, one of the most direct and technologicallymature ways to convert electrical energy to chemical energy(fuel). Hydrogen can potentially be used as a fuel for fuel cellvehicles, cooking appliances and space heating, and for powergeneration using stationary fuel cell systems or advanced gas tur-bines. In some energy supply paradigms, large-scale hydrogenwould be produced from a renewably powered electrolysis [1] orbiomass conversion process and used to replace natural gas andpetroleum for residential and transportation energy needs [2].However, there are numerous challenges to large-scale hydrogenproduction and use including lack of infrastructure, low volumet-ric energy density, inefficiency, and compression, storage, anddispensing costs which invite alternative methods for increasingthe utilization and implementation of renewable energy resources.

Natural gas, comprised mostly of methane, is the most commonfuel for residential and commercial building use, and there is alsoa growing transportation sector using compressed natural gas(CNG)-fueled vehicles. The natural gas industry has establishedan extensive transport network of piping to deliver natural gas tohomes, buildings, and to a lesser extent, CNG fueling stations.The U.S. has approximately 300 thousand miles of natural gaspipeline. An estimate for the cost of constructing an equivalentlength of hydrogen pipelines (assuming 12 in diameter) is $160 to

300 billion [3,4]; these costs are not inclusive of decommissioningthe existing natural gas pipelines. The task of building pipelines inhighly populated and developed cities is more complex comparedto construction of lengthy rural transport pipelines due to concernsof avoiding other infrastructure and general safety [5]. In additionto transport costs, the replacement of appliances which aredesigned to use natural gas (stoves, furnaces, etc.) is an indetermi-nate large cost for employing hydrogen at large scale.

While there are numerous challenges to wide-spread implemen-tation of hydrogen in the energy system at-large, its role as a ver-satile, clean energy carrier is compelling as reported elsewhere byPivovar et al. [6]. In part, the work presented here is relevanttowards evaluating the use of economic hydrogen production andsmaller (transitional) hydrogen distribution networks for the pro-duction of synthetic natural gas. CNG fueling stations are mostlyfound in urban areas where fleets of bus and taxi-cab NGVs arecommon. Using renewable sources of energy to eventually pro-duce a synthetic natural gas (SNG) would take advantage of theestablished transport and end-use infrastructure to leverage thevalue of the SNG product, particularly in countries where accessto cheap natural gas supplies is limited or nonexistent. Further-more, transporting energy via pipelines has 4.5� more energytransport capacity than high voltage transmission lines and has 4�lower transmission losses (4% in T&D versus 1% in pipelines)[7,8]. Such an energy supply paradigm is the basis of the increas-ingly popular power-to-gas (P2G) platform being pursued inEurope and led by Germany, however, it has mostly been focusedon hydrogen production and injection into pipelines [9,10].Indeed, P2G technology for hydrogen has been found to comparefavorably with batteries when viewed on either the basis of energyreturn on investment or energy stored on investment [11]. In par-ticular, Pellow et al. [11] observed that when using overgenerationfrom renewables (wind) to supply P2G systems, the energy return

1Corresponding author.Contributed by the Advanced Energy Systems Division of ASME for publication

in the JOURNAL OF ENERGY RESOURCES TECHNOLOGY. Manuscript received May 22,2018; final manuscript received August 14, 2018; published online September 26,2018. Assoc. Editor: Luis Serra.

Journal of Energy Resources Technology FEBRUARY 2019, Vol. 141 / 021901-1Copyright VC 2019 by ASME

on investment is on par with that of batteries despite the 30%roundtrip efficiency (compared to 75–90% for batteries). In con-trast to P2G with hydrogen, renewable SNG can be produced andinjected into the natural gas infrastructure without constraint aslong as it meets Wobbe Index and heating value specifications ofthe pipeline.

Renewably generated hydrogen can be produced by wind-powered low-temperature electrolysis, which is a commerciallyavailable technology. The hydrogen can then be converted toSNG with conventional technologies in a process called methana-tion, and the SNG can be used as a fuel in the current infrastruc-ture. Figure 1 illustrates an alternative pathway architecture inwhich the hydrogen is converted to SNG at the city-gate, wherebythe existing natural gas pipelines are used to distribute the fuel. Incomparison to direct hydrogen production and distribution, theadditional costs of SNG production replace the additional costs ofthe distribution network and end-use applications which wouldhave to be modified to function with hydrogen. The use of elec-trolysis technology for hydrogen generation has been accom-plished for decades dating back to early Haber–Bosch ammoniaplants [12]. Both low-temperature alkaline and proton exchangemembrane (PEM) electrolyzer technologies are the most commontoday, with PEM technology available at the MW-scale [10].Numerous pilot and demonstration plants for power-to-hydrogenusing PEM and alkaline technologies have been built, most ofthem below 100 kWe [10].

There are several possible hydrogen supply paradigms. Asshown in Fig. 1, the electrolysis plant is located at the wind farm,so the hydrogen must be piped to the city-gate, where the SNGplant is located. This approach could be considered one whichoffers a hydrogen economy transition via the build of shorterhydrogen pipelines. As renewable hydrogen can be intermittent inproduction, some storage in pressurized cylinders or undergroundcaverns is likely to be required. Alternatively, hydrogen could beproduced onsite with renewable or low-cost grid electricity. Thecarbon dioxide feedstock is derived from carbon capture in apower plant which is also located near the city-gate. The SNG iscompressed and transported using the NG infrastructure, and it isdistributed throughout the city to fueling stations and residential/commercial consumers.

It is recognized that natural gas is a relatively cheap utility inthe energy market, and renewable hydrogen is much more expen-sive; this will make the SNG more expensive than conventional

natural gas. In addition, the sources of carbon dioxide are limitedbecause there are few power plants which currently employ car-bon capture and sequestration (CCS). However, the advent ofincreasingly low cost electrical energy from renewable resources,such as wind and photovoltaics, is beginning to drive interest inH2O/CO2 electrolytic reduction processes in a host of applica-tions.2 Additionally, the recognition of CCS as a vital, least costpathway to climate change mitigation [13] likely means that theavailability of CO2 feedstocks will increase in the future at scalescommensurate with (or greater than) a transition to both syntheticfuels and chemical commodity production. The “power to SNG”study in this paper aims to quantify the performance of a renew-able pathway to SNG by way of converting renewably generatedhydrogen (from electrolysis) with recycled carbon dioxide fromnonrenewable resources.

Prior Work. Methanation has been a common practice foreliminating carbon monoxide and carbon dioxide in various chem-ical processes such as ammonia production and natural gas purifi-cation. For these processes, only small amounts (1–3% molarbasis) of carbon oxides need to be converted to methane. A “bulk”methanation process is unique due to the high concentration ofcarbon oxides and hydrogen. In addition, the carbon dioxide is theonly carbon source, and the reaction characteristics of carbondioxide are much different than carbon monoxide.

In the late 1970s and early 1980s, there was a strong effort forthe production of methane from the gasification of coal [14–17].The gasification of coal produces a syngas which is composedmainly of hydrogen and carbon monoxide. The methanation pro-cess was commonly designed using a series of adiabatic reactors,and the reactor temperature rise was controlled with the use ofrecycle and interstage cooling [16]. The predominant catalyst cho-sen for the methanation reactors was supported nickel due to itshigh activity, high selectivity to methane, and relatively low cost[17].

One example of large-scale SNG production from coal gasifica-tion is the Great Plains Synfuels Plant in North Dakota [18].Methanation of the coal syngas is accomplished over a pelleted-supported nickel catalyst employed within a packed bed reactorand interstage cooling to produce high pressure steam. The

Fig. 1 Hydrogen from wind-powered electrolysis to SNG production architecture

2https://www.topsoe.com/events/topsoe-catalysis-forum

021901-2 / Vol. 141, FEBRUARY 2019 Transactions of the ASME

https://www.topsoe.com/events/topsoe-catalysis-forum

process was found to operate well if sulfur was removed prior tothe reactor due to sulfur poisoning of the catalyst. Panek andGrasser [18] observed that carbon dioxide hydrogenated to formmethane in the reactor, but the reaction was not as complete com-pared to carbon monoxide hydrogenation.

More recently, there has been a strong push, particularly in Euro-pean countries, for the so-called power-to-gas (P2G), power-to-fuel,or power-to-X technologies which can make use of renewable electri-cal energy and carbon feedstocks to produce valuable chemical prod-ucts, such as liquid fuels for transportation, hydrogen for fuel cellvehicles, methane as substitute natural gas for both transportation andheating purposes, or chemical commodities, such as methanol andethylene [10,19]. In addition, P2G is viewed as a possible enablingpathway for increasing renewables penetration into the electric gridas an energy storage technology and low-cost grid-balancing solution[19–21].3 P2G technology is one that enables coupling of the electricgrid with either a natural gas grid or a large storage infrastructure,such as pressurized underground caverns (cf. Jensen et al. [22]).However, at present, most CO2 methanation process plants havebeen accomplished at only small scale with the exception of the Audie-gas process, currently the largest at 6 MW, and none with fuel cell/electrolysis technologies. Several design studies of P2G plants havebeen performed. Buchholz et al. [20] studied methanation reactordesign and integration with alkaline electrolysis (at 80 MWe) and alignite fired coal power plant to produce SNG for pipeline storage.The study was concerned with economic viability and the merit ofbeing able to more frequently baseload the power plant by supplyingelectrical energy to the P2G plant. The plant efficiency was estimatedat 53% and plant economics indicated an SNG production costbetween 10 and 15 $/kgSNG (720–1080 $/MWh). These productioncost estimates are some 20–100 times greater than natural gas marketvalues in the U.S., Canada, and Germany and are largely attributed tohigh alkaline electrolyzer costs which represent some 78% of theplant CAPEX [20]. Vandewalle et al. [21] examined P2G systemsbut from the viewpoint of how P2G systems affect the gas, the elec-tric power, and the CO2 storage sectors, and the interactions betweenthese sectors. Schaaf et al. [23] focus on methanation of CO2 for stor-age in the natural gas distribution system. Their work focuses onhigher temperature methanation (>400 �C) processes and subscalereactor experimentation. Two process plant designs were presentedfor multistage methanation, however, efficiency and economic resultswere not presented.

Recently, high temperature electrolysis for SNG production hasalso been explored. Giglio et al. have examined SNG productionplant concepts from 10 MWe high temperature electrolysis,reporting production costs of 48–94 $/MWh (0.67–1.33 $/kgSNG)of SNG depending on CO2 and electricity prices [24] and plantefficiencies as high as 80% [25]. The economic results assumed>91% capacity factor. Ancona et al. [26] also have examinedSNG production from high temperature solid oxide electrolyzersfrom an efficiency perspective, reporting efficiencies ranging from77 to 85% (electricity to fuel) depending on the concentration ofmethane, with lower efficiencies for >80% molar CH4. In addi-tion to techno-economic analysis, Stemberg and Bardow [27] per-formed environmental impact assessments for various power-to-Xtechnologies, including power-to-methane (i.e., SNG) and con-cluded that sinking electric power into heat pumping rather thanfuels appeared to be most advantageous from a life cycle assess-ments perspective. While other electrofuel (or power-to-fuel)studies have examined liquid fuels production (e.g., see Beckeret al. [28] and Schemme et al. [29]), large-scale low-temperatureelectrolysis to SNG production via Sabatier, as given in the pres-ent work, has not previously been reported upon.

Objectives. This study centers on the process design andtechno-economic evaluation of a novel SNG production process

via reaction of carbon dioxide with renewably generated hydrogensupplied from a 40,000 kg/day electrolysis plant. The carbon diox-ide feedstock is assumed to be captured and scrubbed from anexisting coal fired power plant at the city-gate, where the SNGplant is co-located. Unlike the previous studies, the uniqueness ofthe present work is the process integration with membrane purifi-cation and low-temperature waste heat recovery via organic Ran-kine cycle (ORC) technologies to achieve high efficiency SNGproduction. Process design includes desulfurization of CO2 feed-stock, CO2 recycle, and sequential CO2 reactor injection to opti-mize methane conversion. Sensitivity analysis of SNG productioncost to techno-economic factors is also of interest.

Background and Modeling Approach

Electrolysis. In the present work, the electrolytic production ofhydrogen is envisioned to be geographically co-located with theelectricity from renewable energy supplies (e.g., photovoltaic orwind turbines) as previously shown in Fig. 1, and therefore, noeffort is made to integrate hydrogen production processes with theSNG plant. This flexible approach enables some independence ofthe plant cost and performance results from the feedstock costs ofhydrogen supply (which may or may not include additional infra-structure costs, such as hydrogen pipelines). Mass manufacturingof state-of-the-art PEM electrolyzer technology is estimated toprovide a hydrogen cost of production of 4.2 $/kg (for 400 $/kWcapital cost), a 97% capacity factor, and a lower heating value(LHV) efficiency of 66% for grid-electricity feedstock costs of66 $/MWh.3 Grid-electricity costs represent over 80% of the prod-uct cost. In forward-looking estimates, intermittency (40%capacity factor) and lower electricity feedstock costs (�10 $/MWh) are expected to lower the projected hydrogen cost to 2.2 $/kg. Further reductions down to 1.1 $/kg are projected with R&Dadvances and lower capital costs (100 $/kW).3 Many of the Fore-court analyses performed by the U.S. Department of Energy illus-trate a base case of 5.1 $/kg for distributed, pipeline hydrogenproduction.4 Given the expected range of hydrogen productioncosts (1–5 $/kg), the approach taken here has been to assume ahydrogen supply from state-of-the-art electrolyzers and to focuson methanation plant design, performance, and sensitivity analy-ses of feedstock costs and capacity factor on SNG productioncosts.

Methanation Thermodynamics. Reactor design requires anunderstanding of both methanation thermodynamics and kinetics.The Sabatier process of methanation reacts one mole of carbon

Fig. 2 Temperature dependence of equilibrium molar compo-sition at 1 bar and H2:CO2 feed ratio of 4:1

3https://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_pivovar_2.pdf

4https://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_sarkar_satyapal_2.pdf

Journal of Energy Resources Technology FEBRUARY 2019, Vol. 141 / 021901-3

https://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_pivovar_2.pdfhttps://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_pivovar_2.pdfhttps://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_sarkar_satyapal_2.pdfhttps://www.energy.gov/sites/prod/files/2016/12/f34/fcto_h2atscale_workshop_sarkar_satyapal_2.pdf

dioxide (CO2) with four moles of hydrogen (H2) over a catalyst asdescribed by the following equation:

CO2 þ 4H2 !CH4 þ 2H2O DH ¼ �165kJ=mol (1)

The reaction produces one mole of methane (CH4) and twomoles of water (H2O). The reaction is highly exothermic and thenumber of moles decreases as the reaction proceeds to the right(forward). The forward direction of the Sabatier reaction isfavored at lower temperatures due to its exothermicity and highpressures due to its molar stoichiometry.

Figure 2 depicts a plot of chemical equilibrium for methanationas a function of temperature. At low temperatures, the equilibriumconcentration of CO2 is very small, which corresponds to a highdegree of forward reaction progression. The CH4 concentrationdecreases as the temperature increases, and CO forms at tempera-tures greater than 450 �C due to the reverse water gas shift reac-tion, shown in the following equation:

H2 þ CO2 $ COþ H2O DH ¼ 41kJ=mol (2)

The maximum in CO2 concentration at 550�C is indicative of

the thermodynamic preference for both the reverse water gas shiftreaction and the reverse of the methanation reaction (i.e., steamreforming). As the temperature exceeds 600 �C, the steam reform-ing reaction dominates, producing equal amounts of CO and H2O,while consuming H2 and CO2, as shown in Fig. 2.

The chemical equilibrium plots provided in Fig. 2 inform thedesired operating temperature conditions for methanation. Resultsfor CH4 concentration at 20 bar are also depicted and illustratedthat the CH4 concentration increases at higher pressure, however,this effect is not significant for reactors operating near 300 �C.The stoichiometric ratio for H2:CO2 of 4:1 results in the highestequilibrium concentration of CH4. While similar amounts of CO2react to form CH4 for feed ratios of 4:1 and 5:1, the unreacted H2dilutes the concentration of the gas mixture, which lowers the con-centration of CH4. The effect of increasing pressure on equilib-rium CH4 concentration is most observable at temperatures above400 �C (see Supplemental Fig. S1 which is available under the“Supplemental Data” tab for this paper on the ASME Digital Col-lection). The formation of CO does not occur until the temperatureexceeds 550 �C (as opposed to 450 �C for atmospheric pressure).

Methanation Kinetics. The Sabatier reaction is thermody-namically favored at temperatures between 300 and 400 �C, sincethe Gibbs free energy of reaction in Eq. (1) is largely negative(�113.5 kJ). However, the reduction of carbon dioxide to methanerequires a catalyst to promote the elemental steps of dissociationand formation required for the conversion [30]. For supportednickel catalysts, activation requires temperatures of greater than250 �C [31]. Catalysts suitable for methanation have high nickelcontent, and there is a temperature limit of around 600 �C to miti-gate catalyst sintering [32]. For molar contracting reactions, suchas methanation, high pressure is generally favorable for reactionkinetics and equilibrium conversion.

Several global reaction rate expressions are available for reactorand process design [17,33–36]. Saletore and Thomson [33]derived an irreversible power law rate valid for temperatures up to400 �C, shown in the following equation:

rCH4 ¼ kP0:79H2 P0:12CO2

P�0:62H2O (3)

The rate constant k¼ 0.725 gmol/(hr-gcat-atm0.29) was shown tobe independent of temperature if operated over 300 �C. The reac-tor was operated at pressures up to 8 bar. The negative exponentfor the water partial pressure indicates that water removal is favor-able on the reaction kinetics. The validity of this reaction raterequires the following conditions: temperatures greater than300 �C, high (>10%) concentrations of CO2, and large catalyst

particles (>1.6 mm diameter). These constraints on reactor condi-tions are compatible with those selected in the SNG productionplant concept studied herein.

Alternative rate expressions were also explored. The rateexpression of Chlang and Hopper [34] was an irreversible powerlaw formulation but derived from data over small temperatureranges. The expression did not have a water dependency, had amuch stronger PCO2 dependence relative to the literature, and wasnot deemed robust enough or compatible for this application.Seglin et al. [17] presented a reversible Langmuir-Hinshelwood-Hougen-Watson rate law based on experimental data taken from260 to 399 �C and constant 1 bar operation. The reversible term inthis rate equation indicates that the reaction can approach an equi-librium limited regime. However, the uncertainty in extrapolationto higher pressure and the complexity of this equation along withthe ambiguity of determining the temperature dependence fromthe provided results made it difficult to implement in an AspenPlusTM kinetic reactor model for simulation. Additional detailsand discussion can be found in Becker [37].

Carbon Conversion. High carbon conversion is of criticalimportance for plant economics and efficiency. Hoekman et al.[38] conducted an experimental study of the carbon dioxidemethanation reaction in a tubular, packed-bed reactor and eval-uated CO2 conversion (XCO2) and CO2 conversion efficiency(XCO2, eff) as defined on a molar basis in Eqs. (4) and (5), respec-tively. The study used a modern (�2009), commercially availablecatalyst with a 20–25% nickel loading in a laboratory scale reactorand was operated at various temperatures and H2:CO2 feed com-positions to determine the optimum operating conditions. As

Fig. 3 CO2 conversion dependence on catalyst temperature;experimental data (symbols) from Hoekman et al. [38]


http://dx.doi.org/10.1115/1.4041381

shown in Fig. 3(a), the highest carbon dioxide conversion (XCO2)of about 80% occurred with H2:CO2 feed ratios of 6:1. Figure 3(a)shows that a near linear increase in conversion occurs at tempera-tures up to about 275 �C with a maximum conversion predictednear 325 �C. A slight decline of conversion as temperaturesexceed 325 �C is observed, but conversion performance in the280–340 �C range is relatively constant. These results were alsocongruent with the work of Lefebvre et al. [35] who studiedmethanation in slurry bubble column reactors.

XCO2 ¼CO2;in � CO2;out

CO2;in(4)

XCO2;eff ¼CO2;in � CO2;out

H2;in(5)

In contrast, Fig. 3(b) shows that the CO2 conversion efficiency(XCO2,eff), was found to be the highest at the lowest H2:CO2 ratiotested (2:1), meaning the reactor proportionally converted themost carbon dioxide relative to the hydrogen feed. The resultssuggest that the optimum molar ratio of hydrogen to carbon diox-ide depends on the overall system process. If hydrogen is the lim-iting and most valued reactant, and there is no recycle employed,then the highest efficiency for each reactor pass is desired (i.e., anH2:CO2 of 2:1). With a series of reactors that utilize recycle, thereis a system-level trade-off to consider between per pass conver-sion and the penalties for a separation process and recycle com-pression energy. There are also limitations on acceptable levels ofcarbon dioxide and hydrogen in the delivered SNG product whichinform such decisions.

In addition to the selection of H2:CO2 feed ratios, another pro-cess system engineering consideration is water removal and reac-tor staging. Consistent with increasing the forward reactiondirection given in Eq. (2), Habazaki et al. [36] have noted thatremoval of water in between reactor stages drastically increasesthe conversion of carbon dioxide, thereby enabling an increase inmethanation by hydrogenation of carbon monoxide. Further, thecarbon monoxide methanation reaction (i.e., reverse steamreforming) is enhanced by the presence of some carbon dioxide inthe feed which tends to reduce carbon deposition and catalystdeactivation [36].

For the reactor design effort herein, the kinetic reaction profilefrom the reaction rate found in Saletore and Thomson [33] is usedbased on its similarity in experimental reactor conditions to thisstudy and because it captures the negative effect of water partialpressure on the reaction as was shown in Habazaki et al. [36].

Synthetic Natural Gas System Processes

Overview. The SNG production plant is modeled using AspenPlusTM software, which simulates and integrates chemical reactors

and balance-of-plant components (heat exchangers, compressors,pumps, etc.) from user-defined performance specifications. Thesize of the plant is based on an estimate of hydrogen that couldpotentially be produced from a large wind-powered electrolyzer at40,000 kg/day (1.667 mt/h). Figure 4 illustrates a process flow dia-gram for the SNG production plant with a few sample statepointsgiven. The key aspects of the plant include (i) carbon dioxideclean-up, (ii) three methanation reactor stages, (iii) heat recoverybetween reactor stages with ORC units and water condensing andpurge, and (iv) bulk gas, carbon dioxide, and hydrogen separationand recycle.

A pure H2 feed is assumed to be piped into the plant, and CO2is delivered either locally from a power plant with CCS or pipedfrom an existing pipeline. Both feedstocks are assumed to be pres-surized such that no additional compression is necessary toachieve the reactor pressure of 20 bar. The CO2 goes through apurification stage before being distributed to each reactor. Themethanation reactor pressure of 20 bar is chosen based on the ther-modynamic and kinetic favorability of the methanation reaction.The plant feed of CO2 is chosen to be consistent with the 4:1H2:CO2 molar ratio from the stoichiometry of the reaction; theresulting CO2 feedstock amounts to about 80,000 ton/year whichis equivalent to roughly 2% of the CO2 produced from a 500 MWcoal-fired power plant. The CO2 is distributed to each reactor suchthat an H2:CO2 molar ratio of higher than 4:1 is maintained, assuggested by Hoekman et al. [38] for high conversion.

Three reactor stages are chosen for several reasons: (i) multiplestages enable the purging of water which drives the reverse of thewater-gas shift (WGS) reaction, (ii) two or less reactors requiredexcessive bulk recycle and over-strained the ability of the separa-tion processes to limit the content of CO2 in the product SNG, and(iii) three (versus four or more) stages achieved a high enoughoverall conversion of CO2 and H2 to CH4 while mitigating exces-sive capital costs for the reactor stage components. Each of thethree reactor stages allows for water condensing and purge afterheat recovery. The recovered heat from the methanation reactorand effluent gas stream drives an organic Rankine cycle for powergeneration. While not shown, a cooling tower is also modeled tosupply the cooling for the condensers, the ORC heat sink, andcompressor intercooling.

CO2 Feedstock and Cleanup. Many carbon dioxide capturetechnologies simultaneously capture sulfur molecules, such ashydrogen sulfide (H2S), which must be scrubbed to less than 1ppmv to avoid poisoning of the nickel catalyst in the methanationreactors. This scrubbing process is accomplished by simulatingthe LO-CATTM liquid oxidation process, which removes the bulkof the H2S down to 10 ppmv [39]. The process uses an iron cata-lytic solution to promote the oxidation of H2S to produce elemen-tal sulfur. The elemental sulfur is separated out of the gas stream.The LO-CATTM process cannot remove enough H2S to mitigate

Fig. 4 SNG plant system flowsheet design and state-points


nickel catalyst poisoning, so the gas is heated and sent to an iso-thermal zinc oxide (ZnO) bed to purify the CO2 gas to under 1ppmv [39] (see Supplemental Fig. S4 which is available under the“Supplemental Data” tab for this paper on the ASME DigitalCollection.).

Methanation Reactors. Based on the analysis of the methana-tion reaction and reactor design operation, the performance ofeach of the three methanation reactor stages is estimated. The firstreactor has the highest reactant concentrations and pressure, and acarbon dioxide conversion of 80% is chosen. The second reactorhas a slightly increased inert (methane) concentration and slightlydecreased pressure (�17 bar, from upstream component pressuredrops), so a carbon dioxide conversion of 70% is chosen. Thethird reactor has the greatest concentration of inert species and thelowest operating pressure (�14 bar), so a carbon dioxide conver-sion of 60% is chosen. These conversions are slightly more con-servative than the findings in Hoekman et al. [38] based on thehigh hydrogen to carbon dioxide ratio (>6:1) and operating tem-perature of around 325 �C.

The space velocity, heat transfer coefficient, and maximumtemperature specification are used in Aspen PlusTM to design ashell-and-tube reactor for the methanation reaction. A detailedmultitubular, water-cooled plug flow reactor model based on thereaction rate from Saletore and Thomson [33] is used at the reac-tor conditions to establish the design and geometry requirements.The multitubular reactor sizing is used to provide more realisticcooling performance and cost analysis. Additional details of thereactor design can be found in Supplemental Figs. S2, S3 andTable S1 which are available under the “Supplemental Data” tabfor this paper on the ASME Digital Collection.

Interstage Heat Recovery and Water Drain. The effluent ofthe methanation reactor is used to preheat the reactor feed gasstream in a gas-to-gas heat exchanger. The feed-effluent heatexchanger heats the feed gas up to 250 �C for catalyst activation.The second of three reactor stages utilizes feed-effluent heat recu-peration to preheat the second stage feed gas, where residual ther-mal energy is supplied to an ORC unit at near 175 �C as shown inFig. 4. The temperature rise across the methanation reactor is lim-ited to 100 �C by a coolant stream as described earlier. The hoteffluent exchanges heat to the feed gas before supplying heat to awater loop for ORC power production. The gas exiting the ORCheat exchanger at 135 �C must be cooled down to 50 �C to con-dense and remove the water.

The reactor effluent gases are hot enough to recover useful heatafter recuperation to the feed gas stream. Utilization of a

commercial ORC power generator is advantageous in this sce-nario because it requires only moderate temperatures for vaporiza-tion to produce electricity. The ORC water stream is heated to atemperature of 120 �C by the reactor effluent which is necessaryfor the operation of the ORC at an efficiency of 12% [40]. Eachunit can produce 280 kWe, and the total recovered heat from theplant amounts to about 1010 kWe resulting in a net surplus of 82kWe that can be exported to the electric grid.

CO2/H2 Separation and Recycle. The product pipeline SNGmust meet the energetic and compositional requirements set bythe natural gas industry to be transported in the natural gas pipe-lines. The most stringent compositional requirement is to achievea product molar composition of less than 2% CO2. The uncon-verted carbon dioxide must be separated and recycled to meet thisrequirement.

After the three methanation reactor stages, the product gases gothrough three recycle stages which are shown in Fig. 5 and are dis-cussed in detail below: (i) 40% bulk recycle (diverter), (ii) 90%CO2 separation and recycle, and (iii) 90% H2 separation andrecycle (via polysulfone membrane).

A bulk gas recycle of 40% is used to increase the overall reac-tant conversion. The extent to which the CO2 separation andhydrogen membrane units separate reactant species is limited, andthe bulk recycle allows the plant to accomplish the required prod-uct gas composition while accounting for these limitations. Thepenalty for this recycle is mainly the compression of the recyclegas stream, including some of the desired product methane. Thisis the easiest separation process of the three, and it can be accom-plished simply with a diverter valve.

A process to remove CO2 is necessary to control the amount ofCO2 in the product gas and increase its conversion to methane. Inorder to accomplish the separation and recycle of CO2, chemicalabsorption via methyl diethanolamines (MDEAs) were deter-mined to be the best option. It is a proven process with the abilityto capture 90% of the CO2 in a gas stream containing 15% CO2 atrelatively low pressures [41]. Additives such as Selexol make theMDEA solution more selective to CO2 [44]. The main drawbackto this process is the energy required for the regeneration of theMDEA in the CO2 stripping column. The chemical reactioninvolved in the absorption of CO2 into the MDEA is highly exo-thermic, so a significant amount of heat (�845 kW) is required toreverse the reaction in the stripper (“reboiling”) [42].

The molar percentage of H2 in the CO2-scrubbed gas stream isabout 35%. This is well below the hydrogen composition of feed

Table 1 Plant mass and energy balance summary

Value

Plant inputsHydrogen (kg/s) 0.463Carbon dioxide (kg/s) 2.522Cooling water (gpm) 169Internal power use (kW)Recycle compressors (Bulk, CO2, H2) 570Membrane compressor 146MDEA unit 29Cooling tower 125ORC units �952Total internal power use �82Plant outputsSNG (kg/s) 0.939SNG Wobbe index 47.5Total output HHV (MW) 51.3EfficiencyORC efficiency 12.0%Plant efficiency HHV 78.0% Fig. 5 Bulk recycle, MDEA CO2 separation and recycle, and

polysulfone membrane H2 separation and recycle


http://dx.doi.org/10.1115/1.4041381http://dx.doi.org/10.1115/1.4041381

gas such as that from steam methane reforming (�75% hydrogen)used for many hydrogen separation processes. Pressure swingadsorption is a common method for hydrogen separation, but theprocess becomes ineffective and uneconomical with less than70% H2 in the feed stream because of the large amount of adsorb-ent area required to separate the nonhydrogen gas molecules.Therefore, a well-established membrane technology for hydrogenseparation was chosen and is especially suitable for scenarios inwhich the permeate does not require high purity [43]. The molarflux through a membrane is driven by the pressure difference ofthe separable molecule, i, as shown by the following equation:

Ji ¼qi Pi;feed � Pi;permeateð Þ

d(6)

Ji is the molar flux (mol/m2-s), qi is the permeability constant

(a function of the permeability of a certain molecule through acertain membrane), Pi is the partial pressure of molecule i in thefeed and permeate gas streams, and d is the thickness of the mem-brane. For a certain membrane, Eq. (6) is used to determine thesize of the membrane required to achieve a rate of hydrogen flux.The membrane performance is approximated by the literature val-ues for commercial polysulfone-based membranes which haveproven durability, permeability, and selectivity records [44]. Theestimated normalized permeability constant of 100 (10�6 Scm3/cm2-s-cmHg) based on polysulfone membranes is used to deter-mine the size of the membrane required to achieve a 90% hydro-gen separation.

Process Model Results

Synthetic Natural Gas Product. After three stages of metha-nation reactors, the methane content amounts to about 55% asshown in Fig. 6 and 40% of the gas is recycled (Bulk Recycle).The CO2 from the remaining gas stream is then separated with anamine (MDEA) absorption process and fed back to the first metha-nation reactor stage boosting the methane content to over 90%.The remaining gas stream is compressed to 20 bar (to make up forpressure drops in the system) before being fed into a hydrogen

permeable membrane process for separation and recycle. Theretentate of the membrane process contains the final product SNGfor pipeline transport. The resulting SNG product composition,achieved through sequential stages of reactant conversion and sep-aration, is 92.7% CH4, 6.3% H2, and 1.0% CO2. The SNG is pro-duced at 20 bar for transport in the pipeline.

The heating value must be suitable for end-use applications.The heating value in the natural gas industry is quantified by theWobbe index, as shown in the following equation:

Wobbe Index ¼ HHVSNGffiffiffiffiffiffiffiffiffiffiqSNGqair

r (7)

where HHVSNG is the higher heating value of the SNG, and qSNGand qair are the density of the product SNG and air, respectively,at STP. The importance of the Wobbe Index resides in the end-usestandardization of burner units; for a given orifice size throughwhich the gas mixture passes, gas mixtures with the same WobbeIndex will deliver the same amount of fuel energy. The Wobbeindex for the product SNG is calculated to be 47.5 MJ/Sm3 whichis slightly below the normal index of natural gas of around 48 MJ/Sm3, but it is not low enough to significantly impact burning per-formance. The SNG is slightly lower in volumetric energy densitybased on the Wobbe Index because of the low volumetric energydensity of hydrogen, and natural gas typically contains up to 10%ethane (C2H6), 5% propane (C3H8), 2% butane (C4H10), and 0.5%pentane or heavier (C5þ) by volume [45]; these heavier hydrocar-bons increase the volumetric heating value of the natural gas. Ona weight basis, the heating value of the SNG is higher than typicalnatural gas streams (54.6 MJ/kg versus 52.2 MJ/kg).

Studies have been conducted on issues concerning hydrogenmixing with natural gas (cf. Haeseldonckx and D’haeseleer [46]or Melaina et al. [5]). Potential problems associated with having apartial amount of H2 in the SNG composition have been evaluatedand deemed acceptable with H2 concentrations up to 17%. Pipe-line embrittlement caused by H2 is an issue that requires furtherintensive study to evaluate the long-term effects on the pipelinematerial. Leakage issues are negligible with the predominant useof polyethylene pipelines in the existing infrastructure [46].Although the physical limitations for H2 in the pipelines are esti-mated at about 17%, natural gas companies will allow lessthan this due to heating value requirements and safety standards.The 6% H2 composition employed here was determined to beacceptable for heating value and pipeline transport quality andsafety [5,45].

Plant Performance. Figure 7 shows the overall power andmaterial flow of the SNG production plant. The overall plant effi-ciency on an HHV basis is calculated from Eq. (8) to be 78.0%HHV. The plant loses 22.0% of the heating content of the hydro-gen by the conversion to SNG

gplant ¼HHVSNG þ _W elect

HHVH2¼ 78:0% (8)

The overall plant efficiency on an LHV basis results in 83.2%

Fig. 6 Molar gas composition at various reactor and recyclestreams. Numbers indicate methanation reactor inlet and outletstreams.

Fig. 7 SNG plant input/output summary (HHV basis)


gplant ¼LHVSNG þ _W elect

LHVH2¼ 83:2% (9)

Table 1 summarizes the SNG plant heat balance. The hydro-gen feedstock contains 66 MW of energy on an HHV basis, andthis is necessary to produce about 51 MW of SNG fuel. All ofthe internal power usage required by the compressors, theMDEA unit, and the cooling tower is supplied by the ORC units,and an additional 82 kW of power is produced. The largest inter-nal power load of 417 kW is demanded by the H2 recycle com-pressor, with the balance of about 150 kW for bulk and CO2compression. The membrane compressor is used for pressurizingthe gas to increase the pressure difference across the membranewhile simultaneously achieving the desired pressure for the prod-uct SNG; this compressor has the second largest electric demandof 146 kW.

Plant Economics

The economics of producing SNG from renewable hydrogenare determined by first evaluating the capital investment of theplant. Operating costs are estimated from the plant components,and a life cycle cost analysis is used to determine the levelizedcost of SNG. The effects of hydrogen feedstock cost and operatingcapacity factors are explored to determine the potential SNG pro-duction cost range. It should be noted that no credits are given forCO2 consumption, byproduct oxygen production from the electro-lyzer, or ancillary electric grid benefits of serving as load balanc-ing technology.

Capital Investment. The capital costing of the plant utilizesquotes from industry, literature references, and AspenTech Eco-nomic AnalyzerTM software for conventional plant components.Several of the components use a scaling method to determine theinstalled cost; the cost of each scaling unit (S) is based on the ref-erence scaling unit (So) and base cost (C0). The cost is thenadjusted for the time-dependent equipment cost changes by usingthe chemical engineering plant cost index (CEPCI). The installedcost (IC) is then calculated by using an installation factor (IF),which accounts for various costs associated with installing thecomponent; the installation factor is only necessary if the refer-enced base cost does not include installation. Equation (10) isused to calculate the installed cost based on the given parameters

IC ¼ C0S

S0

� �nCEPCI

CEPCI0

� �IF (10)

The superscript n is the scaling factor which accounts for theeconomy of scale of a particular item. A detailed explanation ofthe cost estimation for each component is given as a footnote ofTable 2. It should be noted that the capital cost analysis conductedfor this study has an estimated accuracy of 630% [47]. The instal-lation factor of 2.47 is based on Spath et al. [39], and this factor isused only when the reference for capital cost does not includeinstallation; the Economic AnalyzerTM software includes adetailed installation calculation, so no factor is used for thosecomponents.

The total capital investment (direct plus indirect costs) for theplant is estimated at 40.9 MM$2009 which equates to a unit cost of

Table 2 Component capital costs in k$

Base cost (Co)a Base scale (So)

b Scaling factor (n)c IFd Installed cost (IC)e

LO-CATTMf 1319 22.6 kg/s 0.65 2.47 783ZnO Bedf 49 22.6 kg/s 0.56 2.47 35Heat Exchangersg — — — 1957Reactorsg — — — — 2919Catalysth 1.2 1 ft3 — — 318Compressorsg — — — — 5126MDEA Uniti 1049 1 unit — 2.47 2591Membrane Unitj 263 1 unit — 2.47 650ORC Unitsk 275 1 unit — 2.47 2717Pumpsg — — — — 214Cooling Towerg — — — — 555TICl — — — — 17,864TDCm — — — — 26,796

aThese base costs have been scaled from the original cost index (CEPCIo) to the 2009 CEPCI (521.9).bThe base scale is the unit for the referenced parameter. Many of these component costs are estimated from AspenTech Economic AnalyzerTM which gen-erates the installed cost directly.cThe scaling factor is not used for the components which were evaluated in AspenTech Economic AnalyzerTM or discretely sized to the proper scale (e.g.,the MDEA Unit).dThe IF is only used if the base cost does not include installation (otherwise its value is 1).eThe installed cost is calculated using Eq. (10) or obtained directly from AspenTech Economic AnalyzerTM.fSpath et al. [39] uses the scaling method from Eq. (10).gAspenTech Economic AnalyzerTM software is used to calculate the installed cost of floating head shell and tube heat exchangers based on heat transferarea (heat transfer coefficients calculated from Peters et al. [47] and materials suitable for the operating temperature and pressure. Reactors were designedbased on the given methodology described earlier. Stainless steel is used to mitigate corrosion from the catalyst packed bed. Each reactor was about 1MM$2009. Several centrifugal compressors were installed including two multi-stage centrifugal compressors with intercooling based on number of stages,pressure ratio, and cooling requirements. Centrifugal pumps are used for the ORC heating water source; six pumps are required for methanation reactorcooling and the interstage heat recovery. Cooling tower is sized based on water flow rate and temperature approach to wet bulb.hCatalyst priced from industry quote [48] for SNG 1000 nickel supported catalyst.iMDEA system cost calculated from industry software [49]. The specified separation of CO2 resulted in a 55 gpm absorbent flow and a system of 60 gpmwas matched to supply the separation.jPeters et al. [47]: cost based on spiral wound membrane fibers for polysulfone-based membrane. The upper end of the cost range of 100$2002/ft

3 wasused.kIndustrial quote [40] at $375,000 uninstalled. Four units are needed to utilize all of the available heat from the plant. A 2.47 IF was applied based on themethod described later.lThis is the total calculated installed cost (TIC) of the plant components.mA factor of 30% of TIC is used for high pressure (20 bar) piping installation based on Peters et al. [47]. A factor of 20% of TIC is used for buildings andstructures of the plant. These costs are added to the TIC to determine the TDCs of the plant (TICþ 8.93 MM$).


870 $/MWSNG (or �1045 e/MWSNG). Additional economic mod-eling details can be found in the Supplemental Information (seeSupplemental Tables S2–S5 which are available under the“Supplemental Data” tab for this paper on the ASME DigitalCollection).

Operating and Life Cycle Costs. The plant operating costs aredetermined by replacement cost estimations for the major plantcomponents and by estimating unplanned maintenance andreplacement as a fraction of the total direct costs (TDC). In addi-tion to these operating costs, other operating costs include staffand labor (12 staff members with a combined annual salary of1.39 MM$), property tax and insurance (2% of TIC), and anunplanned O&M and materials replacement factor of 1% totalinstalled cost (TIC). These estimations are consistent with theH2A analysis tool default values and H2A analysis case studiesfor similar sized fuel production plants [53]. The total annualO&M costs amount to 3.45 MM$ (14.2% of TDC, 8.4% of TIC).

The feedstock to the SNG plant includes hydrogen, carbondioxide, and water. The water consumption of the plant consistsof cooling tower make-up water at 169 gpm, and this cost repre-sents a very small fraction (0.01%) of the total feedstock cost. Thecost of the carbon dioxide feedstock is estimated based on capturefrom a coal-fired power plant from Rubin et al. [51] to be 40 $/ton, and this represents 7.4% of the total feedstock cost. Thehydrogen feedstock cost is the most significant input with regardto the cost of SNG production. Several analyses have been con-ducted on hydrogen production from various renewable sources,and the hydrogen costs associated with production and transport(0.87 $/kg H2 for a 100 mile pipeline delivery [50]) are given inTable 3.

From Table 3, it is clear that the hydrogen source has a largeimpact on the cost of SNG. Even if the SNG production plant hada 100% energetic efficiency, no capital investment costs, and usedthe future scenario for hydrogen production from central biomassgasification (2.34 $/kg H2), the resulting SNG would cost be 19.4$/MMBTU (66 $/MWh), which is just under double the rate thatresidential customers pay in the continental United States in 2016[52].

The total capital investment (direct plus indirect costs) andoperating costs for the plant are inputs to the H2A life cycle costanalysis tool, which is used to generate a levelized cost of SNGfuel. The H2A tool accounts for the capital investment, fixed andvariable operating costs, feedstock and utilities consumption rates,and the time span in which the plant operates [50]. Various eco-nomic parameters specified for the H2A program are given inTable S5 in the Supplemental Data for this paper.

Economic Results and Feedstock Cost Sensitivity. The eco-nomics of SNG production rely heavily on the hydrogen feedstockcost, and to a lesser extent, the plant operating capacity factor (thepercentage of time that the plant operates at full capacity). Thebreakdown of cost contributions toward SNG production cost isgiven in Fig. 8(a) at various hydrogen feedstock costs and acapacity factor of 90%.

The hydrogen feedstock cost dominates the contribution distri-bution, especially for costs greater than 3 $/kg. Even at the verylow hydrogen feedstock cost of 1 $/kg, it represents about half ofthe total SNG production cost. For a 3 $/kg hydrogen feedstock

cost, the hydrogen, capital, and O&M costs represent about 27, 5,and 2 $/MMBTU, respectively. The carbon dioxide feedstock costof 40 $/t represents only 2 $/MMBTU (6.8 $/MWh).

Figure 8(b) shows how the cost contributions shift at variouscapacity factors with a nominal hydrogen feedstock cost of 3 $/kg. At low capacity factors (40%), the contribution of capital andO&M costs becomes more significant; however, the hydrogenfeedstock cost contribution is still greater than 50% at 3 $/kg, rep-resenting 27.4 $/MMBTU (93.5 $/MWh).

A range of hydrogen feedstock costs was given in Table 3, andthe effects of this range on SNG production cost are explored inFig. 9. The steep slope on the plotted lines of Fig. 9(a) indicatesthe importance of obtaining a cheap source of hydrogen for pro-ducing economically competitive SNG. The range of costs forSNG production is 8.2 $/MMBTU (28 $/MWh) with free hydro-gen (and free delivery) and a 90% capacity factor, to 81.8 $/MMBTU (279 $/MWh) with 7 $/kg H2 and a 40% capacity factor.

Table 3 Hydrogen feedstock costs from various renewable sources [53]

$/kg $/MMBTU (HHV) $/MWh (HHV)

Current central wind farm electrolysis 6.71 49.8 170Future central wind farm electrolysis 2.88 21.4 73.0Current central biomass gasification 2.48 18.4 62.8Future central biomass gasification 2.34 17.4 59.4

Fig. 8 Cost contribution for SNG production at (a) three differ-ent hydrogen feedstock costs and a 90% capacity factor and (b)three different capacity factors and a 3 $/kg hydrogen feedstockcost


http://dx.doi.org/10.1115/1.4041381http://dx.doi.org/10.1115/1.4041381

For each 1 $/kg of hydrogen feedstock cost, the SNG productioncost increases 9.4 $/MMBTU (32.1 $/MWh). The DOE cost targetfor large central hydrogen production in 2017 is about 2.2 $/kg;using this feedstock cost results in an SNG production cost of28.6 $/MMBtu (97.6 $/MWh). The Energy Information Agency(EIA) states [52] that the average cost of natural gas at the city-gate in the continental U.S. is approximately 18.8 $/MWh (5.5 $/MMBTU), which is less than the SNG production cost using freehydrogen. However, production costs in other countries, such asEurope, can range from 47.8 to 81.9 $/MWh (14–24 $/MMBtu),with a nominal average value of about 50 $/MWh from 2010 to2016 [54].

The effects of overall plant capacity factor on the SNG produc-tion costs are shown in more detail in Fig. 9(b). While the operat-ing capacity factor greatly increases the SNG production costs iflower than 50%, the effect is relatively small at larger capacityfactors. The capital cost contribution increase is the main reasonfor this effect, and since it represents a relatively small fraction ofthe total cost, there are small SNG cost decreases for capacity fac-tors greater than 60%. The SNG production costs decrease 50%by increasing the capacity factor from 20% to 60% and there isonly a 12% cost decrease by increasing the capacity factor from60% to 100%.

Despite the relatively small benefits from operating 90% of thetime (e.g., versus 60%), every economic advantage must besought. Hydrogen storage at the producer’s site would enable highhydrogen feedstock availability, and this is important for

continuous operation (assuming the wind farm electrolysis plantproduces more than the maximum capacity of the SNG plant).The capacity factor should ideally be limited by general mainte-nance and replacements.

The range of hydrogen feedstock cost for this sensitivity analy-sis does not include large-scale hydrogen storage by the electroly-sis plant, so it should be assumed that a capacity factor in therange of 40–50% can be achieved using this feedstock cost range.A capacity factor of 40–50% is slightly above that of a wind farmcapacity factor, but since the plant is assumed to be under-sizedcompared to the wind-farm, a higher capacity factor can beachieved.

6 Conclusions

Thermodynamic and kinetic considerations of the methanationreaction were explored to model and simulate a system of reactorsfor the conversion of hydrogen and carbon dioxide to SNG. Thehydrogen is supplied to the SNG plant from a pipeline fed by low-temperature PEM electrolysis units. Carbon dioxide is assumed tobe supplied by a power plant utilizing CCS, and a clean-up pro-cess for the gas is necessary to remove sulfur. Reactor design isimportant to maximize the conversion of carbon dioxide by main-taining optimal reaction temperatures. Heat integration and recov-ery is a crucial part of plant operation when highly exothermicreactions, such as methanation, are involved. Multiple reactorstages are also necessary to condense and purge the producedwater from the process stream, thereby promoting the methanationreaction. Multiple reactors increase the per-pass conversion of thereactor train, thereby allowing the separation and recycle proc-esses to produce an SNG composition which is adequate for natu-ral gas pipeline transport and end-use applications. Inter-stageheat recovery is used, and organic Rankine cycles are ideal candi-dates for producing electricity from low temperature waste-heat.The ORC units supplied all of the internal electricity needs, whichenable the plant to operate without reliance on the grid. Separationprocesses were investigated to recycle the reactant species andincrease overall reactant conversion. Conventional separationtechnologies were found to adequately produce SNG with highmethane content while limiting the amount of carbon dioxide inthe product.

The product SNG has a Wobbe index of 47.5 MJ/Sm3 which isacceptable for natural gas pipeline transport and end-use applian-ces in the existing infrastructure. The overall plant efficiency is78.1% LHV and 83.2% HHV. The bottom-up plant cost of theSNG product is found to be highly dependent on the hydrogenfeedstock cost. The range of costs for SNG production is 28 $/MWh (8.2 $/MMBTU) with free hydrogen (and free delivery) anda 90% capacity factor, to 279 $/MWh (81.8 $/MMBTU) with 7 $/kg H2 and a 40% capacity factor. For each 1 $/kg of hydrogenfeedstock cost, the SNG production cost increases 32.1 $/MWh(9.4 $/MMBTU).

The production process of SNG from hydrogen and carbondioxide as presented here is unique, and thus, there are no bench-mark studies from which to compare the results. The plant modelwas reviewed and validated by an expert in the natural gas proc-essing industry, as well as other sources with industrial chemicalprocessing experience. Many laboratory studies are presented inthe literature which attempt to characterize the reaction kinetics ofcarbon dioxide methanation, but there is a large gap between thesestudies and large-scale reactor performance for carbon dioxidemethanation. The SNG plant in this paper characterizes the syner-gies between large-scale reactors, thermal management, andsystem-level techno-economic considerations.

Acknowledgment

The authors would like to thank the National RenewableEnergy Laboratory for financial support under Award No. KXEA-3-33607-54.

Fig. 9 (a) Effect of hydrogen feedstock cost on SNG produc-tion cost for various capacity factors and (b) effect of operatingcapacity factor on SNG production costs for various hydrogenfeedstock costs


References

[1] Eichman, J., Harrison, K., and Peters, M., 2014, “Novel Electrolyzer Applica-tions: Providing More Than Just Hydrogen,” National Renewable Energy Labo-ratory, Golden, CO, Technical Report, No. NREL/TP-5400-61758.

[2] Larson, E. D., Jin, J., and Celik, F. E., 2009, “Large-Scale Gasification-BasedCoproduction of Fuels and Electricity From Switchgrass,” Biofuels Bioprod.Bioref., 3, pp. 174–94.

[3] Parker, N., 2004, “Using Natural Gas Transmission Pipeline Costs to EstimateHydrogen Pipeline Costs. Institute of Transportation Studies,” University ofCalifornia-Davis, Davis, CA, Report No. UCD-ITS-RR-04-35.

[4] Mintz, M., Folga, S., Molburg, J., and Gillette, J., 2002, “Cost of Some Hydro-gen Fuel Infrastructure Options,” Argonne National Laboratory, Lemont, IL.

[5] Melaina, M., Antonia, O., and Penev, M., March, 2013, “Blending Hydrogen IntoNatural Gas Pipeline Networks: A Review of Key Issues,” National RenewableEnergy Laboratory, Golden, CO, Report No. NREL/TP-5600-51995.

[6] Pivovar, B., Rustagi, N., Satyapal, S., 2018, “Hydrogen at Scale, Key to aClean, Economic, and Sustainable Energy System,” Interface, 27(1), pp. 47–52.

[7] Zakeri, B., and Syri, S., 2015, “Electrical Energy Storage Systems: A Compara-tive Life Cycle Cost Analysis,” Renewable Sustainable Energy Rev., 42, pp.569–596.

[8] Auer, J., and Keil, J., 2012, State-of-the-Art Electricity Storage Systems: Indis-pensable Elements of the Energy Revolution, Deutsche Bank AG, Frankfurt amMain, Germany.

[9] Winkler-Goldstein, R., and Rastetter, A., 2013, “Power to Gas: The FinalBreakthrough for the Hydrogen Economy?,” Green, 3(1), pp. 69–78.

[10] Gahleitner, G., 2013, “Hydrogen From Renewable Electricity: An InternationalReview of Power-to-Gas Pilot Plants for Stationary Applications,” Int. J.Hydrogen Energy, 38(5), pp. 2039–2061.

[11] Pellow, M. A., Emmott, C. J., Barnhart, C. J., and Benson, S. M., 2015,“Hydrogen or Batteries for Grid Storage? a Net Energy Analysis,” Energy Envi-ron. Sci., 8(7), p. 1938.

[12] Hager, T., 2008, The Alchemy of Air, Broadway Books, New York.[13] Bui, M., Adjiman, C. S., Bardow, A., Anthony, E . J., Boston, A., Brown, S.,

Fennell, P. S., Fuss, S., Galindo, A., Hackett, L. A., Hallett, J. P., Herzog, H. J.,Jackson, G., Kemper, J., Krevor, S., Maitland, G. C., Matuszewski, M., Met-calfe, I. S., Petit, C., Puxty, G., Reimer, J., Reiner, D. M., Rubin, E. S., Scott, S.A., Shah, N., Smit, B., Martin Trusler, J. P., Webley, P., Wilcox, J., and Dowell,N. M., 2018, “Carbon Capture and Storage (CCS): The Way Forward,” EnergyEnviron. Sci., 11, pp. 1062–1176.

[14] Kertamus, N. G., 1978, “Combined Shift-Methanation Processes,” a report pre-pared for the U.S. Department of Energy, by C. F. Braun & Co, Alhambra, CA,Technical Report No. FE-2240-97.

[15] Dirksen, H. A., and Linden, H. R., 1963, “Pipeline Gas From Coal by Methana-tion of Synthesis Gas,” Inst. Gas Technol., Res. Bull., Vol. 31.

[16] Zahnstecher, L. W., 1984, “Coal Gasification Via the Lurgi Process,” TopicalReport Volume 1: Production of SNG, prepared for the U.S. DOE under Con-tract No.: DE-AC01082FE0508.

[17] Seglin, L., Geosits, R., Franko, B. R., and Gruber, G., 1974, “Survey of Metha-nation Chemistry and Processes,” in Methanation of Synthesis Gas, Advancesin Chemistry Series, Vol. 146, L. Seglin, ed., American Chemistry Society,Washington, D.C.

[18] Panek, J. M., and Grasser, J., 2006, “Practical Experience Gained During the FirstTwenty Years of Operation of the Great Plains Gasification Plant and Implica-tions for Future Projects,” U.S. DOE Office of Fossil Energy, Technical Report.

[19] G€otz, M., Lefebvre, J., M€ors, F., McDaniel Koch, A., Graf, F., Bajohr, S.,Reimert, R., and Kolb, T., 2016, “Renewable Power-to-Gas: A Technologicaland Economic Review,” Renewable Energy, 85, pp. 1371–1390.

[20] Buchholz, O. S., van der Ham, A. G. J., Veneman, R., Brilman, D. W. F., andKersten, S. R. A., 2014, “Power-to-Gas: Storing Surplus Electrical Energy. ADesign Study,” Energy Procedia, 63, pp. 7993–8009.

[21] Vandewalle, J., Bruninx, K., and D’haeseleer, W., 2015, “Effects of Large-Scale Power to Gas Conversion on the Power, Gas and Carbon Sectors andTheir Interactions,” Energy Convers. Manage., 94, pp. 28–39.

[22] Jensen, S. H., Graves, C., Mogensen, M., Wendel, C., Braun, R., Hughes, G.,Gao, A., and Barnett, S. A., 2015, “Large-Scale Electricity Storage UtilizingReversible Solid Oxide Cells Combined With Underground Storage of CO2 andCH4,” Energy Environ. Sci., 8, pp. 2471–2479.

[23] Schaaf, T., Gruenig, J., Schuster, M. R., Rothenfluh, T., and Orth, A., 2014,“Methanation of CO2 - storage of renewable energy in a gas distribution sys-tem,” Energy, Sustainability, and Society, Springer.

[24] Giglio, E., Lanzini, A., Santarelli, M., and Leone, P., 2015, “Synthetic NaturalGas Via Integrated High-Temperature Electrolysis and Methanation: Part II—Economic Analysis,” J. Energy Storage, 2, pp. 64–79.

[25] Giglio, E., Lanzini, A., Santarelli, M., and Leone, P., 2015, “Synthetic NaturalGas Via Integrated High-Temperature Electrolysis and Methanation—Part I:Energy Performance,” J. Energy Storage, 1, pp. 22–37.

[26] Ancona, M. A., Antonioni, G., Branchini, L., De Pascale, A., Melino, F., Orlandini,V., Antonucci, V., and Ferraro, M., 2016, “Renewable Energy Storage SystemBased on a Power-to-Gas Conversion Process,” Energy Procedia, 101, pp. 854–861.

[27] Stemberg, A., and Bardow, A., 2015, “Power-to-What?—EnvironmentalAssessment of Energy Storage Systems,” Energy Environ. Sci., 8, p. 389.

[28] Becker, W. L., Braun, R. J., Melaina, M., and Penev, M., 2012, “Production ofFischer-Tropsch Liquid Fuels From High Temperature Solid Oxide Co-Electrolysis Units,” Energy, 47(1), pp. 99–115.

[29] Schemme, S., Samsun, R., Peters, R., and Stolten, D., 2017, “Power-to-Fuel asa Key to Sustainable Transport Systems—An Analysis of Diesel Fuels Pro-duced From CO2 and Renewable Electricity,” Fuel, 205, pp. 198–221.

[30] Wang W., and Gong J., 2011, “Methanation of carbon dioxide: an overview,”Front. Chem. Sci. Eng. 5(1), pp. 2–10.

[31] Twygg, M., 1989, Catalyst Handbook, 2nd ed., Wolfe Publishing Ltd., UnitedKingdom.

[32] Rostrup-Nielsen, J. R., Pedersen, K., and Sehested, J., 2007, “High TemperatureMethanation Sintering and Structure Sensitivity,” Appl. Catal. A: General, 330,pp. 134–138.

[33] Saletore, D. A., and Thomson, W. J., 1977, “Methanation Reaction Rates forRecycle Reactor Compositions,” Ind. Eng. Chem. Process Des. Dev., 16(1), pp.70–75.

[34] Chlang, J. H., and Hopper, J. R., 1983, “Kinetics of the Hydrogenation of Car-bon Dioxide Over Supported Nickel,” Ind. Eng. Chem. Prod. Res., 22(2), pp.225–228.

[35] Lefebvre, J., G€otz, M., Bajohr, S., Reimert, R., and Kolb, T., 2014,“Improvement of Three-Phase Methanation Reactor Performance for Steady-State and Transient Operation,” Fuel Proc. Technol., 132, pp. 83–90.

[36] Habazaki, H., Yamasaki, M., Zhang, B.-P., Kawashima, A., Kohno, S., Takai,T., and Hashimoto, K., 1998, “Co-Methanation of Carbon Monoxide and Car-bon Dioxide on Supported Nickel and Cobalt Catalysts Prepared From Amor-phous Alloys,” Applied Catalysis A: General, 172(1), pp. 131–140.

[37] Becker, W. L., 2011, “Design, Performance, and Economic Assessment ofRenewable and Alternative Fuel Production Pathways,” M.S. thesis, ColoradoSchool of Mines, Golden, CO.

[38] Hoekman, S. K., Broch, A., Robbins, C., and Purcell, R., 2010, “CO2 Recyclingby Reaction With Renewably-Generated Hydrogen,” Int. J. Greenhouse GasControl, 4(1), pp. 44–50.

[39] Spath, P., Aden, A., Eggeman, T., Ringer, M., Wallace, B., and Jechura, J., 2005,“Biomass to Hydrogen Production Detailed Design and Economics Utilizing theBattelle Columbus Laboratory Indirectly-Heated Gasifier,” National RenewableEnergy Laboratory, Golden, CO, Technical Report No. TP-510-37408:13.

[40] Stone, J., 2010, “UTC Power, personal communication, Industrial Quote forOrganic Rankine Cycle Unit: Efficiency and Cost,” South Windsor, CT.

[41] Alie, C., Backham, L., Croiset, E., and Douglas, P. L., 2005, “Simulation ofCO2 Capture Using MEA Scrubbing: A Flowsheet Decomposition Method,”Energy Convers. Manage., 46(3), pp. 475–487.

[42] Nextant, 2009, “Task 2: Detailed MDEA Process Analysis,” Prepared for theNational Renewable Energy Laboratory, Golden, CO, NREL Task Order No.KAFT-8-882786-01.

[43] Ho, W., and Sirkar, K., 1992, Membrane Handbook, Springer Publishing, New York.[44] Baker, R. W., 2000, Membrane Technology and Applications, McGraw-Hill,

New York.[45] Kinay, A., and Parrish, W., 2006, Fundamentals of Natural Gas Processing,

Taylor and Francis Group, LLC, Didcot, United Kingdom.[46] Haeseldonckx, D., and D’haeseleer, W., 2006, “The Use of the Natural-Gas

Pipeline Infrastructure for Hydrogen Transport in a Changing MarketStructure,” Int. J. Hydrogen Energy, 32(10–11), pp. 1381–1386.

[47] Peters, M. S., Timmerhaus, K. D., and West, R. E., 2003, Plant Design andEconomics for Chemical Engineers, 5th ed., McGraw-Hill, New York.

[48] Helfrich, D., 2010, “Sud-Chemie, Personal Communication, Industrial Quotefor SNG1000 Methanation Catalyst Composition and Cost,” Charlotte, NC.

[49] Newpoint Gas, LP, 2010, “AmicalcTM,” Innovative Gas and Oil Solutions,Oklahoma City, OK, accessed June 2010, http://www.newpointgas.com/downloads.php

[50] Steward, D., Ramsden, T., and Zuboy, J., “H2A Production Model, 2008, Ver-sion 2 User Guide,” National Renewable Energy Laboratory, Golden, CO,Report No. NREL/TP-560-43983.

[51] Rubin, E. S., Chen, C., and Rao, A. B., 2007, “Cost and Performance of Fossil FuelPower Plants With CO2 Capture and Storage,” Energy Policy, 35(9), pp. 4444–4454.

[52] U.S. Energy Information Administration, 2018, Annual Energy Outlook,Appendix A, Table A3, Washington, DC.

[53] U.S. DOE Hydrogen and Fuel Cells Program, 2010, “Central and Forecourt Hydro-gen Production Case Studies,” Hydrogen & Fuel Cells Program, Washington, DC,accessed Sept. 20, 2018, http://www.hydrogen.energy.gov/h2a_analysis.html

[54] Hou, P., Enevoldsen, P., Eichman, J., Hu, W., Jacobson, M. Z., and Chen, Z.,2017, “Optimizing Investments in Coupled Offshore Wind–Electrolytic Hydro-gen Storage Systems in Denmark,” J. Power Sources, 259, pp. 186–197.


http://dx.doi.org/10.2172/1159377http://dx.doi.org/10.1002/bbb.137http://dx.doi.org/10.1002/bbb.137http://dx.doi.org/10.2172/1068610http://dx.doi.org/10.1016/j.rser.2014.10.011http://dx.doi.org/10.1515/green-2013-0001http://dx.doi.org/10.1016/j.ijhydene.2012.12.010http://dx.doi.org/10.1016/j.ijhydene.2012.12.010http://dx.doi.org/10.1039/C4EE04041Dhttp://dx.doi.org/10.1039/C4EE04041Dhttp://dx.doi.org/10.1039/C7EE02342Ahttp://dx.doi.org/10.1039/C7EE02342Ahttp://dx.doi.org/10.1016/j.renene.2015.07.066http://dx.doi.org/10.1016/j.egypro.2014.11.836http://dx.doi.org/10.1016/j.enconman.2015.01.038http://dx.doi.org/10.1039/C5EE01485Ahttp://dx.doi.org/10.1186/s13705-014-0029-1http://dx.doi.org/10.1016/j.est.2015.06.004http://dx.doi.org/10.1016/j.est.2015.04.002http://dx.doi.org/10.1016/j.egypro.2016.11.108http://dx.doi.org/10.1039/C4EE03051Fhttp://dx.doi.org/10.1016/j.energy.2012.08.047http://dx.doi.org/10.1016/j.fuel.2017.05.061http://dx.doi.org/10.1007/s11705-010-0528-3http://dx.doi.org/10.1016/j.apcata.2007.07.015http://dx.doi.org/10.1021/i260061a011http://dx.doi.org/10.1016/j.fuproc.2014.10.040http://dx.doi.org/10.1016/S0926-860X(98)00121-5http://dx.doi.org/10.1016/j.ijggc.2009.09.012http://dx.doi.org/10.1016/j.ijggc.2009.09.012http://dx.doi.org/10.1016/j.enconman.2004.03.003http://dx.doi.org/10.1016/j.ijhydene.2006.10.018http://www.newpointgas.com/downloads.phphttp://www.newpointgas.com/downloads.phphttp://dx.doi.org/10.1016/j.enpol.2007.03.009http://www.hydrogen.energy.gov/h2a_analysis.htmlhttp://dx.doi.org/10.1016/j.jpowsour.2017.05.048

JERT-18-1362 Braun S1

†Supporting Information (SI) to:

Production of Synthetic Natural Gas from Carbon Dioxide and Renewably Generated Hydrogen: A Techno-Economic Analysis of a Power-to-Gas Strategy

W.L. Becker1, M. Penev2, and R.J. Braun1,*

1 Department of Mechanical Engineering, Colorado School of Mines, Golden, CO USA 2 National Renewable Energy Laboratory, Golden, CO USA

INTRODUCTION

This Supporting Information (SI) provides additional details to the paper “Production of Synthetic Natural Gas from Carbon Dioxide and Renewably Generated Hydrogen: A Techno-Economic Analysis of a Power-to-Gas Strategy.” The sections contain additional documentation related to Sabatier methanation thermodynamics, reactor design, process system design, and process economics.

Thermodynamics. The pressure dependence on the CH4 concentration is shown in Figure S1. The CH4 concentration increases at higher pressure, but there is not a significant increase in CH4 concentration at pressures near 20 bar. The stoichiometric ratio for H2:CO2 of 4:1 results in the highest equilibrium concentration of CH4. While similar amounts of CO2 react to form CH4 for a feed ratio of 4:1 and 5:1, the unreacted H2 dilutes the concentration of the gas mixture, which lowers the concentration of CH4.

Figure S1. Pressure dependence of equilibrium CH4 molar fraction at T=400 °C and various H2:CO2 feed ratios

Methanation Reactor Modeling & Design. Reactor temperature control is a critical aspect of methanation reactor design and performance. The exothermic nature of bulk methanation is a concern for large-scale reactor temperature increase, and the design of reactor with cooling has multiple advantages over adiabatic reactors. While isothermal tubular reactors are nearly impossible to achieve in practice, cooling can be used to limit the temperature rise. One of the

0 4 8 12 16 200.15

0.18

0.21

0.24

0.27

0.30

P (bar)

Equi

libriu

m C

H4 M

olar

Fra

ctio

n

T=400°C

2:1

3:1

4:1

5:1

H2:CO2 Initial Ratio


main considerations for the extent of the reaction is the temperature at which the catalyst is activated to achieve optimum kinetics. Governed by the catalyst type and the incorporation of the catalyst within the reactor bed, the temperature which maximizes the activity of the catalyst would be held constant in an isothermal reactor. The amount of reactor stages and catalyst required decreases when internal cooling is employed due to the increased allowable extent of reaction; this alleviates the concern of heating the reactor to temperatures which would sinter and deactivate the catalyst, while maintaining a temperature high enough for desired catalyst activation.

Another consequence of high temperatures is catalyst poisoning due to carbon deposition (coking). The latter concern is not as critical as sintering in this study because of the high oxygen content from carbon dioxide. Aspen Plus™ was used to calculate the thermodynamic equilibrium for carbon deposition, and it is found to be negligible for the reactor conditions in this study (e.g. an atom composition of 7% C, 82% H, 11% O and temperatures of 250-400 °C).

A reactor design, similar to a shell-and-tube heat exchanger, is used to limit the temperature rise. The catalyst is packed in the tube side, while water flows through the shell to provide cooling. A co-flow arrangement is suitable for this design because the reaction is expected to activate near the front of the reactor, which allows for the largest temperature difference in the entry region. The catalyst for the bulk methanation reaction was chosen based on the literature to have high nickel content which promotes conversion of high concentrations of CO2 in the feedstock stream. One potential catalyst, SNG1000 [1], was selected for the reactor with a composition of 56.6% Ni-6%MnO-5%Si-21%Al. The calculated volume for the catalyst is given in Equation S4 and represents the required tube volume for the heat exchanger.

The packed bed tubes have unique heat transfer properties due to the two-phases of the reacting gases and the solid catalyst particles. Heat transfer phenomena include convection from the gas to the catalyst and the wall, and conduction between the catalyst particles and to the wall (radiation was neglected for simplification). The effective heat transfer coefficient is modeled differently depending on whether the bulk temperature inside a plug in the tube can be considered constant or variable with respect to the two phases. It is shown from Borkink et al. [2] that the temperatures of the packed particles and the surrounding fluid temperature are very similar for small tube diameters (less than 5 cm). Therefore, a homogenous model was implemented to determine the heat transfer parameters of the tube and external coolant.

To determine the overall heat transfer coefficient, the conduction from the packed bed tube, along with a wall heat transfer coefficient, must be determined. Equations (S1) – (S3) [2] give heat transfer properties for a packed bed with alumina cylinder catalyst particles having an effective diameter of 6 mm.

∗ ∗,

S1

∗ S2

1∗

1∗ ∗ S3


In Equations (S1)-(S3), ∗ is the dimensionless effective radial heat conductivity, ∗ is the dimensionless effective heat conductivity of the packed bed with a stagnant fluid, is the molecular Peclet number based on the superficial velocity of the gas, , is the Bodenstein number for heat at fully developed turbulent flow, ∗ is the dimensionless wall heat transfer coefficient, A0 and a are fitting parameters determined for the specified catalyst, is a “lumped factor” determined to be 7.4, and N is the number of catalyst particles that can span the diameter of the tube. For the reactor conditions in this model, the resulting overall heat transfer coefficient is calculated to be 413 W/m2K, and this value is used for catalyst packed tube-side reactor heat transfer coefficient.

The variability in previous kinetic studies on bulk methanation of carbon dioxide demonstrates a level of uncertainty for large-scale reactor design. While the required catalyst volume and weight for a certain conversion can be determined from rate laws, the distribution of the catalyst in the reactor and the amount of time that the reactants “see” the active catalyst are determining factors for reactor design. The rate equations can be used to calculate the required catalyst volume from a given inlet flow rate, composition, and conversion using Equation (S4).

S4

Vcat is the required catalyst volume (m3), Fio is the molar flow rate (mol/s) of the reacting species i, cat is the bulk catalyst density (kg/m3), X is the specified conversion of species i, and ri is the reaction rate (mol/s-kgcat) for species i. It is found that the resulting catalyst volume calculations from the rate equations differed greatly, and they underestimate the catalyst requirement in comparison to space velocity suggestions by Hoekman et al. [3], from which the reactor can also be sized using Equation (S5).

S5

The volume of the reactor is calculated using v, the standard volumetric flow rate of the feed gas (Sm3/s), and SV (hr-1), the chosen space velocity. As an example calculation, a volumetric flow of 7 Sm3/s (similar to the flow rate for a methanation reactor in this study) and a chosen space velocity of 10,000 hr-1 results in a reactor volume of 2.5 m3.

Kinetic rate equations are used to explore the possible reaction profiles in the reactor which is necessary to determine appropriate cooling methods. The kinetic based profile gives insight into the “hot spots” by identifying possible maximum temperature gradients in certain locations of the reactor. The Hoekman et al. [3] methanation study employing a modern commercial catalyst is used to estimate the conversion of carbon dioxide based on the feed gas composition and flow rate; the suggested space velocity required to accomplish the conversion is used for reactor sizing.

The plug flow reactor geometry is designed to achieve the desired space velocity with a constraint on the tube diameter for adequate plug-flow assumptions. Table S1 summarizes the


design parameters for the methanation reactors. The heat transfer area is constrained by the volume and tube diameter requirements, but the co-flow coolant water flow rate through the shell is adjusted to accomplish the desired temperature control [2]. The same process is used for each of the three reactors, and the resulting geometry is similar for all three. Further details of the reactor modeling can be found elsewhere [4]. Table S1. Methanation fixed bed tubular reactor design

Parameter Value

Type Shell-and-tube HX, 1 tube pass

Length 3 m

Inner Tube Diam. 0.05 m

Number of Tubes 424

Heat Transfer Area 200 m2

Reaction Volume 2.5 m3 Figure S2 shows that the CO2 conversion in the methanation reactor has a negative linear

correlation with space velocity. Space velocity variations from 4000-18000 hr-1 result in conversions of 70-55%. This implies a trade-off between reactor capital and catalyst costs (high for low space velocities) and the desired per pass conversion (low for high space velocities).

Figure S2. CO2 conversion dependence on space velocity for a H2:CO2 feed ratio of 4:1; data from Hoekman et al. [3]

Figure S3 illustrates a temperature vs. position plot of the process gas and water coolant. The water coolant water is pressurized to 2.5 bar, enters the shell side at 50 °C, and boils for most of the length of the reactor at 130°C. The coolant flow is adjusted to provide cooling to the reactor such that the exit temperature of the process gas is 350 °C. This heated water can then be used for generating electricity via ORC units. This reactor design and coolant flow enables the reaction to


take place at a temperature of around 330 °C for the majority of the reactor length; this temperature approximately corresponds with the maximum conversion from which the reaction extent is based (cf. Hoekman et al. [3]).

Figure S3. Temperature profile of methanation reactor process gas and coolant stream

CO2 Feedgas Cleanup. Figure S4 illustrates the H2S removal process for the CO2 feedstock.

Following the LO-CAT™ process, the gas must be heated to 375 °C for catalyst activation of the ZnO bed. This heating is supplied by the effluent of the first methanation reactor; the temperature of the first reactor effluent is 400 °C (350 °C for the other two reactors) to allow sufficient temperature difference for the ZnO bed feed gas heat exchanger. The ZnO bed is approximately isothermal due to the relatively small amount of H2S removed.

ZnO Bed

Sulfur

AirCO2

Sulfur

LO-CATProcess

Exhaust

Heat from 1st reactor stage effluent

30 °C22.3 bar

30 °C21.3 bar

375 °C21.0 bar

375 °C20.0 bar

837 kW

Scrubbed CO2

Figure S4. CO2 feedstock clean-up: LO-CAT™ liquid oxidation followed by ZnO bed


Interstage Cooling & Water Removal. From the reaction mechanism, water removal favors the forward reaction to produce methane. Lower water content favors the reverse water gas shift reaction which enables the highly reactive carbon monoxide methanation reaction to take place. The condenser cooling source for removing the water out of the gas stream is a cooling tower water loop. The cooling tower also provides the cooling water for the ORC condenser. The water loss from the evaporative cooling is replaced with a water makeup source.

Heat exchanger coefficients are extrapolated from data in Peters et al. [5] for gas at pressure, liquid water, vaporization, and condensing. A pressure drop of 33 kPa across heat exchangers and 100 kPa across reactors is assumed throughout the process. These pressure drops represents a 1-1.5% pressure decrease for heat exchangers and 4-5% pressure decrease for reactors compared to the incoming gas stream pressure.

Process Economics. Summary information for installation, indirect, equipment operating

costs, and H2A tool input parameters for life cycle costing are given below. Installation factor cost allocation is summarized in Table S2.

Table S2. Installation factor cost allocation based on Spath et al. [6].

% of TPEC Total Purchased Equipment Cost (TPEC) 100

Purchased equipment installation 39 Instrumentation and controls 26 Piping 31 Electrical systems 10 Buildings (including services) 29 Yard improvements 12

Total Installed Cost (TIC) 247

The indirect costs associated with the plant are determined as a percentage of the TDC based on Spath et al. [6], and the allocation of various costs that contribute to the indirect costs of the plant are given in Table S3.

Table S3. Indirect costs allocated as a percentage of the TDC and actual value in MM$.

% of TDC MM$ Engineering 13 3.5 Construction 14 3.8 Legal and contractors fees 9 2.4 Project contingency 15 4.1

Total Indirect Costs 51 13.8


Table S4. Operating costs for major plant components

Annual k$2009 LOCAT™a 62

ZnO Bedb 34

Methanation Catalystc 159

MDEAd 68

Membranee 130

Cooling Towerf 28

Totalg 482 a Spath et al. [6]: 150 $2002/tonne sulfur removed b Spath et al. [6]: reactor sizing from 4,000 GHSV with 4.67 $2002/lb-cat. Assumed catalyst density of 1200 kg/m3 c Methanation catalyst replaced every 2 years d Nextant [7]: operating costs scaled from 160 gpm to 60 gpm (this plant) MDEA flow rate using a 0.7 scaling factor. This includes make-up MDEA solution. e Membrane replacement every 5 years. f 5% of installed cost for water chemicals and pump maintenance. g In addition to this annual O&M total, the ORC units must be replaced after 20 years.

H2A Life Cycle Cost Analysis Tool. The dollar value of 2005 is used for the H2A analysis tool, so all of the plant costs reported in this study were adjusted from 2009 to 2005 using the CEPCI. Due to the variability of fuel and electricity escalation rates, the production costs for SNG were not adjusted from the output value of the H2A program, which are in 2005 dollars, but inflation was taken into account for the lifetime of the plant. It is assumed that the increase in fuel cost from inflation is negligible compared to the fluctuating cost of fuels, which does not necessarily coincide with the time-value of money.

Table S5. Economic parameter inputs to the H2A tool

Value Constant dollar value 2005 Internal rate of return (after-tax) 10% Debt/Equity 0%/100% Plant life 40 years Depreciation MACRS

Depreciation Recovery period 20 years

Construction period 2 years 1st year 75% 2nd year 25%

Start-up time 12 months Revenues 50% Variable costs 75% Fixed costs 100%


Working capital 15% of total capital investment Inflation rate 1.90% Total taxes 38.90%

Decommissioning costs 10% of depreciable capital

Salvage value 10% of total capital investment

CO2 feedstock cost $40/tonne

Hydrogen feedstock cost Varied Plant capacity factor Varied

REFERENCES

[1] Helfrich, D., Sud-Chemie: industrial quote for methanation catalyst composition and cost: SNG 1000,

April, 2010.

[2] Borkink, J.G.H., Westerterp, K.R., Determination of effective heat transport coefficients for wall-

cooled packed beds, Chemical Engineering Science 47(9), 1992, 2337-2342.

[3] Hoekman, S.K., Broch, A., Robbins, C., Purcell, R., CO2 recycling by reaction with renewably-

generated hydrogen, Int. J. Greenhouse Gas Control 4(1), 2010, 44-50.

[4] W.L. Becker, Design, Performance, And Economic Assessment Of Renewable And Alternative Fuel

Production Pathways, 2011, Master’s Thesis, Colorado School of Mines, Golden, Colorado.

[5] Peters, M.S., Timmerhaus, K.D., West, R.E., Plant Design and Economics for Chemical Engineers,

5th ed., 2003, McGraw-Hill, New York. Ch. 6, 14.

[6] Spath P., Aden A., Eggeman T., Ringer M., Wallace B., Jechura J., Biomass to Hydrogen Production

Detailed Design and Economics Utilizing the Battelle Columbus Laboratory Indirectly-Heated

Gasifier, 2005, National Renewable Energy Laboratory, Golden, CO, NREL Technical Report: TP-

510-37408:13.

[7] Nextant, Task 2: Detailed MDEA Process Analysis, 2009, Prepared for the National Renewable

Energy Laboratory, Golden, CO, NREL Task Order No. KAFT-8-882786-01

aff1l1FN12FN2FN3FD1FD2FD33FD4FD5415FD6FD7FD8FD967FD102T2n1T2n2T2n3T2n4T2n5T2n6T2n7T2n8T2n9T2n10T2n11T2n12T2n1338s69123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354

production of synthetic natural gas from carbon dioxide and … · 2020. 4. 15. · hydrogen....

Documents