A multicriteria comparison of utilizing sugar cane bagasse for methanol to gasoline and butanol production

Stavros Michailos a, David Parker b, Colin Webb a,*

aSchool of Chemical Engineering and Analytical Science, The University of Manchester,

Oxford Road, Manchester, M13 9PL, UK

bSchool of Biosciences, University of Exeter, Stocker Road, Exeter, EX4 4QD, UK

*Corresponding Author

E-mail address: colin.webb@manchester.ac.uk (C. Webb)

Keywords: Synthetic fuels, Bagasse utilisation, Multicriteria analysis, Process simulation, Gasoline synthesis, Butanol production

Abstract: The present study makes a consistent and comparative assessment of the overall exergy, financial and environmental efficiencies of two biomass-to-fuels (utilised in internal combustion engines with spark ignition) conversion options and based on this result, gives a recommendation as to which of the options assessed is most desirable. These options are methanol to gasoline (MTG) and biochemical butanol, while as feedstock the solid residue of sugar cane, bagasse, was considered. For the work presented in this study, a base case scenario has first been developed for each pathway by employing either Aspen Plus or SuperPro Designer (as simulators) to perform mass and energy balance calculations while Matlab software has been used for modelling the reaction kinetics of each process. Based on the simulations, thermodynamic (exergy analysis), economic (financial and risk analysis) and environmental (CO2

emissions) evaluations were carried out. Afterwards, sensitivity analyses have been performed in order to define the key parameters of each conversion route. Exergy and economic analysis favour the gasoline production while butanol produces less CO2 emissions. The study concludes with multicriteria decision analysis (MCDA) where each process is issued a score according to the investigated criteria. This makes it possible for the investigated procedures to

be compared on the same basis. According to this analysis, the production of gasoline achieves a higher overall score than butanol production, i.e. 97% and 90% respectively.

1. Introduction

In the last three decades, the pressing issue of energy security, fossil fuel price volatility, increasing awareness of global warming, and prevailing legislations confining the use of non-renewable energy sources have warranted a tremendous interest in, and growth of, the bioenergy industry. Additionally the manufacture of biofuels may contribute to the local economic growth [1]. In view of these and the related, inevitable, depletion of fossil reserves, the biorefinery concept has recently emerged. The focal aim of biorefineries is the integration of biomass conversion processes for the sustainable production of biofuels with the aim of substituting petroleum derived fuels such as gasoline, diesel and kerosene [2]. Resultant technologies producing first generation (1G) biofuels are already well-established; however exploitation of lignocellulosic biomass derived from forestry or agricultural residues, including bagasse, can positively contribute to the renewable production of biofuels and building block chemicals without competing for land [3]. Several studies have already raised the issue of waste utilization for developing a sustainable biofuel sector [4, 5, 6].Sugar cane milling processes for ethanol or sugar production leave approximately 250 kg of solid residue bagasse for every tonne of raw sugar cane processed which can eventually be utilised as feedstock for biofuels production [7, 8].

Traditionally, ethanol from sugarcane or corn has been recognised as the principal biofuel for the gasoline market. Nevertheless ethanol has some properties that make it somewhat incompatible with existing fuel distribution and motor vehicle infrastructure. The properties of butanol make it a more attractive fuel for blending with gasoline or for use directly in place of it. The advantages it has over ethanol include lower vapour pressure (thus safer to handle), higher flash point, decreased corrosiveness and decreased miscibility with water. It can be shipped and distributed through existing pipelines and filling stations and has a higher energy density (closer to that of gasoline) [9]. Thus, in this study, due to their high energy densities and compatibility with

existing infrastructure, gasoline and butanol were chosen as the desired fuels to be produced in the bioenergy conversion routes evaluated. They are both advanced biofuels and as such they have been targeted to make a major contribution to the total amount of renewable fuels produced in the next 20 years [10].

Butanol is usually synthesised from fossil fuels. However, biomass can also be used as feedstock for butanol production. These feedstocks are the same as for ethanol and include corn, sugar beets, and lignocellulosic material [11]. The industry (for example DuPont, BP, or Cobalt Biofuels) has also shown interest in so-called ‘biobutanol’ generation and some facilities have already started operation [9]. The core stage of the process is the acetone-butanol-ethanol (ABE) fermentation of sugars catalysed by strains of Clostridium acetobutylicum In the case of lignocellulosic biomass processing, the addition of a pretreatment step to crack down the lignin structure is essential. During ABE fermentation, butanol, acetone, and ethanol are produced in a molar ratio of 6:3:1. This specification limits butanol productivity with researchers, nowadays, focusing on changing the metabolic pathway and selectively increase the butanol yields [12].

Methanol is one of the most significant platform chemicals, used as feedstock for the production of formaldehyde, propylene, dimethyl ether, plastics, acetic acid and other chemicals. Huge amounts of methanol are also used to produce gasoline additive methyl tert-butyl ether (MTBE). Currently, methanol is chiefly synthesised in low temperature (200-300°C), high pressure (5-10 MPa) packed bed reactors, using a syngas feed. The main global producer is Lurgi [13]. Methanol can be used as fuel additive but can also be converted to gasoline in fluidized bed reactors over a zeolite based catalyst. The methanol to gasoline (MTG) process was first developed by Mobil Oil in the late 1970s. Nowadays, ExxonMobil produces 7,000 barrels per day in 15 plants located in West Virginia, USA [14]. Syngas is principally derived from conventional sources such as coal and natural gas. In this framework, the design of alternative and based on renewable feedstocks MTG production processes is essential.

Recently, several studies conducted techoeconomic analysis on butanol production but they were limited to calculating the economic performance of the process without considering the energy efficiency and the environmental impact of the process [15, 16, 17]. The production cost for butanol is in the

range of 0.59–0.75 $ kg-1. Furthermore, butanol feasibility was mainly compared to ethanol. The main outcome from this comparison can be summarised as that butanol can be produced at higher energy efficiencies than ethanol but it provides lower profits [18, 19]. Regarding the MTG process, the National Renewable Energy Laboratory (NREL) and the Pacific Northwest National Laboratory (PNLL) have conducted feasibility studies on gasoline production via the MTG pathway from biomass derived syngas. The main focus of these studies was to design comprehensive process models and subsequently to calculate the gasoline production cost which according to the NREL was equal to 16.73 $ GJ-1

and 17.46 $ GJ-1 based on the PNNL [20, 21]. Kempegowda et al. [22] have also conducted a detailed technoeconomic analysis of biomethanol production which results in a positive net present value (NPV) of 600 $ t -1 but upgrade to gasoline was beyond the scope of that study.

As a result of a literature review, it was concluded that the assessment of biochemical butanol and MTG production process were carried out mainly based on economic criteria. Thus, the study presented here was focused on integrating exhaustive process simulations, thorough exergetic, economic and environmental calculations to evaluate and compare the sustainability of the investigated processes, and eventually suggest the best alternative. This methodology provides a robust mechanism and can be used as a reliable decision making tool.

2. Methodology

The scope of the study was to evaluate and compare two process scenarios for the exploitation of bagasse in a novel and sustainable manner with the aim of contributing to the development and establishment of a reliable biorefinery sector. Butanol and gasoline derived from biomass are direct biofuel competitors for the petrol gasoline market. These options were designed, evaluated and compared within an integrated framework. Sugarcane bagasse was selected as feedstock due to its availability and the fact that it is a waste and as such is readily accessible, provides no food or land competition (unlike first generation feedstocks) and reduce waste management problems. The synthesis of the study is illustrated in Fig. 1.

Figure 1 – Integrated framework developed to evaluate the feasibility of bagasse utilization

2.1 Process modelling

The Aspen Plus simulation package was used to model the thermochemical conversion route (MTG process) and SuperPro designer the production of biochemical butanol. The reactor models have been developed in the Matlab environment due to the insufficient kinetic options provided directly in the simulators. The outputs of the reactor kinetic analysis have been transferred as inputs to the simulators via a VBA Excel Macro by taking advantage of Microsoft’s COM technology for software interaction. The inlet mass flow rate for all the cases was set equal to 100 t h -1. User defined non-conventional solids were determined to symbolize bagasse and ash. Aimed at those modules, two Aspen models were allocated: one for the density (DCOALIGT) and the second one enthalpy (HCOALGEN) that necessitates awareness of proximate analysis and ultimate analysis of the bagasse [23].

2.2 Feedstock and non-conventional component properties

Lignocellulosic materials consist of complex polymers rather than easily accessible monosaccharides, thus they have to be hydrolysed so as to release the desired substances (sugars).The feedstock investigated in this research is the solid residue of the sugar cane milling process, bagasse. Typical ultimate and proximate analyses as well as the chemical composition of bagasse are illustrated in Table 1. Bagasse consists of cellulose, hemicellulose and lignin; it was assumed that cellulose and hemicellulose consist only of glucan and xylan respectively. For the thermochemical procedures, bagasse was defined in terms of the elements in the proximate and ultimate analysis, whereas for the biochemical process it was defined by its chemical composition.

Table 1 – Typical Bagasse composition [24]

Proximate analysis Ultimate analysis

Parameters Mass fraction (%) Element Dry Weight (%)Moisture 50 (wb) C 45.38Ash 3.2 (db) H 5.96Volatile matter (dry Basis)

83.65 (db) O 45.21

Fixed Carbon (dry basis)

13.15 (db) N 0.15

Chemical composition

Component Dry Weight (%)Cellulose (of which glucan = 100%) 45

25206.8

Hemicellulose (of which xylan = 100%)

Lignin

Extractives

The higher heating value (HHV) of bagasse is estimated from the following empirical equation [25]:

HHV=0.349∗C+1.1783∗H+0.105∗S−0.1034∗O−0.0151∗N (1)

Where C, H, S, O, N represent the mass fractions of the respective elements. The lower heating value can be estimated as follows [25]:

LHV =HHV−hg( 9 H100

+ M100 )

(2) Where H is the mass fraction of hydrogen (dry basis), M the moisture content and hg stands for latent heat of steam (MJ kg-1). Hence for this case HHV=18.7 MJ kg-1 and LHV=16.4 MJ kg-1. The mass flow rate of dry bagasse is equal to 15.1 kg s-1, so the LHV, in power units, is equivalent` to 247.6 MW. Subsequently the exergy content of sugar cane bagasse was calculated from the following empirical equation [26]:

ε ch,bagasse=β∗LHV bagasse (3)

β=1.044+0.0160 H

C−0.3439 O

C∗(1+0.0531 H

C )+0.0493 NC

1−0.4124 OC

(4)

Where H/C, O/C, N/C represent atomic ratios in the fuel. Thus on this occasion the exergy content of bagasse is equal to 280 MW. Information about physical properties for several of the key substances used in the simulation for the biochemical conversion routes is not available in the customary property database of SuperPro Designer. In fact, quite a few of the properties necessary to positively model these processes do not exist anywhere. For that reason, it is necessary to assess the literature, calculate properties where required, and define a group of reliable physical properties for all components of importance. The National Renewable Energy Laboratory (NREL) [27] has conducted a study defining the key physical properties of the required components and the outcomes are presented in Table 2. Solids are principally everything that can be combusted in the bagasse apart from cellulose, lignin, or hemicellulose.

Table 2 – Properties of the key components used in this study [27]

Compound

Name

Formula MW (g mol-1) HHV (MJ mol-1)

Cellulose C6H10O5 180.16 2.81

Hemicellulose C5H8O4 150.132 2.35

Lignin C7.3H13.9O1.3 122.493 3.26

Biomass (cell

mass)

C

H1.64O0.39N0.23S0.0035

23.238 0.53

Cellulase C

H1.57O0.31N0.29S0.007

22.834 0.55

Trichoderma

reesei

CH1.8O0.5N0.2 24.626 0.52

Solids C

H1.48O0.019N0.29S0.001

3

16.584 0.55

2.3 Exergy analysis

Quantitative assessment of energy in a chemical process can be conducted by employing the first law of thermodynamics. On the other hand, the direction of flow or work (qualitative assessment) can be done using the second law of thermodynamics and it is known as exergy analysis. Exergy analysis is more useful in measuring the efficiency of process since it identifies the causes, locations and magnitude of the system inefficiencies and includes irreversibility in the thermodynamic analysis. As exergy, the maximum useful work that can be obtained from a system at a given state in a given environment can be defined. By analysing the exergy destroyed by each component in a process, it is possible to identify the stages that need to be improved [28]. Exergy analysis can also be used to compare components or systems to help make informed

design decisions. In this study, the exergy efficiency, ψ, of each process was calculated using Eq. (5).

ψ=mfuel EFuels+Eout

Q +W out

mbagasse Ebagasse+E ¿Q+W ¿

(5)

Where mfuel and mbagasse are the mass flow rates of the produced fuels and

bagasse, respectively and subscripts in and out stand for produced exergy flows and external exergy flows respectively.

Work is considered as pure exergy while the exergy content of a heat stream is equal to [29]:

EQ=Q∗(1−T 0

T)

(6)

Where T is the temperature at which Q is available and T0 the reference temperature (298 K throughout this study). The total exergy of a material

stream has been calculated as the sum of the physical and chemical exergy multiplied by the mass flow rate [29]:

E=m(ε ph+ε ch) (7)

ε ph=h−h0−T 0∗( s−s0 ) (8)

ε ch=∑i

x i ε0 , i+R T 0∑i

xi lnxi (9)

Where h and s are the mass enthalpy and entropy respectively at specific temperature T, xi the mass fraction of each component and ε0i the standard chemical exergy of each substance. All the necessary thermodynamic data have been extracted from the simulators.

2.4 Economic analysis

Cost estimating based on recent data for actual prices paid for similar equipment is the most accurate method but access to large amounts of high quality data is required, which are not openly available. Thus for this study cost estimation based on historic data was used by utilizing the following equation [30]:

C=C0(SS0

)f

(10)

Where C is the estimated actual cost of the unit, C0 the base cost of the unit, S the actual size or capacity of the unit, S0 the base or capacity and f an empirical scaling factor. Values for these parameters can be found in the literature [17, 18, 20, 21, 6, 31]. All the costs have been brought forward using the Chemical Engineering Plant Cost Index (CEPCI). After the estimation of the equipment cost it is possible to proceed in calculating the direct and indirect costs of the project by following the methodology proposed by Peters and Timmerhaus [32]. According to this method the direct and indirect costs are calculated as a percentage of the basic equipment cost (BEC). Subsequently to the estimation of total capital investment, operating costs have been calculated and finally the estimation of net cash flows (CF) has been conducted as in Eq. (11). The cost of capital was set equal to 7% per annum, the declining-balance depreciation method – Eq. (12) [30] - was selected (assets depreciated over 10 years), the

annual operating hours were taken as 8000, project lifetime was 20 years, and the tax rate was selected as 40%.

CF=D+(1−t)(R−OC−D) (11)

Dm=C ¿ (12)

Where D is depreciation, t is tax rate, R refers to revenues, OC to operating costs, Dm depreciation after m years, C the capital cost, the fraction Fd equals to 2/n where n corresponds to the depreciable life in years (10 in this case). The operating costs comprise labour, catalyst, enzyme and utilities cost [6, 31, 33, 34]. Various economic indicators, including net present value (NPV), IRR, ROI, annualised capital cost (ACC) and payback period (PP), are estimated in order to assess each project’s economic performance.

NPV =−TCI+∑t=1

20 CF(1+i)t

(13)

ROI= Annual incomeCapital Investment

×100 %

(14)

Annualised CapitalCost=TCI× i×(1+i)n

−1+(1+i)n

(15)

The total annual cost (TAC) of each project derives from the sum of the ACC and the operating costs. Afterwards it is possible to calculate a crucial economic factor, the cost of production. The production cost of a product is a significant index especially when comparing the financial feasibility between different conversion pathways. It is extremely useful when the value of a product cannot be determined clearly, for instance when a known product is produced from a nonconventional feedstock (as in this case) [30].

ProductionCost= TACProductionrate

(16)

2.5 Carbon footprint

These days, there is an important driving force encouraging the transformation of manufacturing procedures towards more sustainable directions. This, in turn, inspires alternative process configurations that will eventually have lower environmental impact which meet ever more stringent legislation. In this study, the CO2 emissions of each alternative were calculated. In principle, even if the carbon footprint of biomass is considered neutral, there is still a substantial uncertainty on whether biofuels (e.g. cellulosic ethanol) generate less greenhouse gas (GHG) emissions than conventional petroleum fuels (e.g. gasoline), as discussed in recent studies [35]. As a result, bioprocesses with lower emissions will have a greater contribution to the development of a sustainable biofuel sector.

2.6 Multicriteria analysis

By using multicriteria decision analysis (MCDA), it becomes possible to evaluate alternatives on the basis of several interacting factors rather than just one (e.g. economic or environmental). This creates the opportunity of applying modifications so as to broaden the focus of the analysis, to the point where all attributes are evaluated on an equal basis. The purpose of MCDA is to quantify attributes related to a project and present them in a comprehensive and consistent format. In general, the multicriteria decision models consist of two processes; decomposition and aggregation. During the decomposition step, the problem is divided into several smaller problems equal to the number of distinct criteria. This allows the decision makers to thoroughly examine the data deriving from different sources. The aggregation step brings together all the individual pieces of information and concludes with a final decision. During aggregation, assumptions regarding the relative importance weightings of the various criteria, have to be made. Simple Additive Weighting (SAW) is the oldest, most widely known and practically used method for multivariable analysis [36]. The sum Sj of the weighted normalized values of all the criteria is calculated for the j-th object as follows:

S j=∑i=1

m

w ir ij

(17)

Where w i is weight of the i-th criterion (∑i=1

m

wi=1); rij is the normalized value of

the i-th criterion for the j-th object; i=1,…,n; j=1,…,m; n is the number of criteria used, m is the number of objects (alternatives) compared. The formulae that are used in order to maximize or minimize a criterion are the following [37]:

r ij=min

jrij

r ij

(18)

rij=r ij

maxj

r ij

(19)

The normalised importance weight, w i, for each criterion can be calculated

using Eq. 20:

w i=n−ri+1

∑i=1

n

n−r i+1

(20)

3. Process description

3.1 MTG process

An Aspen plus model has been developed in order to simulate the MTG process of bagasse. The physical properties of the conventional components have been estimated by using the Redlich-Kwong-Soave cubic equation of state with Boston-Mathias alpha function (RKS-BM). It is recommended for hydrocarbon processing applications, such as gas processing, refinery, and petrochemical processes [38]. The SOLIDS property option was employed for the biomass crushing and drying units as it is recommended for solids processing unit operations [23]. The MTG process consists of some certain steps, namely, 1) bagasse pre-treatment (crushing and drying), 2) gasifier island, 3) syngas quenching and cleaning, 4) methanol generation, 5) gasoline production and generation and 5) heat and power generation system. An Aspen plus crusher

block has been employed to simulate a gyratory crusher which chops bagasse to a final particle size of 2 mm. Afterwards bagasse enters a fluidized bed dryer in order to reduce its moisture content to 10%. The energy required for the drying process is provided by heat produced within the processes. For the pyrolysis zone two blocks have been utilized, one RYIELD reactor which decomposes bagasse to its constituent elements (using a FORTRAN statement according to ultimate analysis) and another one RYIELD reactor which gives the product distribution of the pyrolysis section according to the empirical correlations presented in Eqs (21-27) [39].

yChar=1.055× 10−4∗T 2−0.22016∗T+150.4764 (21)

yTar=1.582× 10−4∗T 2−0.28218∗T−92.1972

(22)

yCO=7.261× 10−5∗T2−0.10416∗T +42.089 (23)

yCO 2=2.534 ×10−5∗T 2−0.03036∗T+18.7576 (24)

y H2=6.198× 10−6∗T2−9.6545∗T +3.75195 (25)

yCH 4=7.954 ×10−6∗T 2−6.9416 ×10−3∗T +1.2026 (26)

y H2 O=7.954×10−6∗T 2−6.9416 ×10−3∗T +1.2026 (27)

A steam bubbling fluidised bed reactor, operating at 1200°C and 0.1 MPa, was considered in this study where the gasification reactions occur. The equilibrium model approach was adopted and thereby the gasifier was modelled using Gibbs free energy minimisation method in a RGIBBS reactor (by identifying the possible products). This approach has been used before in several studies [40, 41] and is generally suitable for feasibility studies, such as that undertaken here, but would not be suitable for reactor design [42]. Furthermore, it was assumed that the nitrogen and sulphur content of bagasse are completely converted to NH3 and H2S respectively. Then, the producer gas enters a tar reformer where tar and methane are converted to syngas (hydrogen and carbon monoxide). The next step is to remove ammonia, thus an RSTOICH reactor (efficiency of 100% was assumed) has been employed to simulate an ammonia scrubber where ammonia is removed using water and sulphuric acid. Ammonia

reacts with the sulphuric acid to form ammonium sulphate. Subsequently, H2S is removed by using a solution of methyldiethanolamine (MDEA) in an absorber. The solvent is recovered in a stripper and recycled to absorber unit. Afterwards the producer gas is cooled down to produce steam that enters the gasifier and a flash drum has been employed to remove water from the gas mixture. Purified syngas is then compressed to 4 MPa in a three stage compressor, heated to 240°C and enters the methanol reactor. Skrzypek et al. [43] have illustrated a Langmuir-Hinshelwood-type kinetic equation for low pressure methanol synthesis exploiting as catalyst a commercial Cu/ZnO/Al2O3. The main conclusion was that methanol synthesis utilizes as carbon source carbon dioxide rather than carbon monoxide. This statement is supported by several researchers [44, 45]. Hence the following two reactions were taken into account:

C O2+3 H 2↔CH 3OH+ H2O

(28)

CO+H 2O ↔C O2+H 2

(29)

The proposed kinetic model is described by the following kinetic rate equations:

r MeOH=k 1 K H 2

2 KCO2¿ (30)

rWGS=k2 KH 2KCO2

¿ (31)

The values of the above kinetic parameters can be found in the literature [40]. The reactor was modelled as a PFR at steady state. Fig. 2 presents the molar fraction of each component throughout the reactor. Produced methanol is then cooled down and separated from the unconverted gas in a flash drum. A portion of the unconverted gas (90%) is recycled to the reactor while the purge stream is sent to the power generation unit [46]. A combined gas-steam turbine unit was designed to generate electricity according to specifications found in the literature [47]. The methanol stream is then heated and compressed to 200°C and 1.5 MPa before entering the methanol to gasoline reactor. A fluidized bed reactor is employed so as to convert methanol to gasoline via a series of catalytic reactions over a zeolite based catalyst according to the succeeding mechanism:

n2[2C H3 OH ↔C H 3OC H3+H 2O ]→Cn H 2n→n[C H 2]

(32)

The reactor has been simulated as a RYIELD reactor with product distribution on a mass basis as follows [17]: gasoline 37.08%, LPG, 4.4%, ethane 1.42%, propane 2.1% and water 55% subsequently, gasoline needs to be recovered. For this purpose two distillation columns have been employed. The first one (de-ethanizer) separates gasoline and LPG from ethane and propane and finally pure gasoline is recovered at the bottom of the second distillation column (stabilizer) at a rate of 12 t h-1. The fuel gas is sent to the power generation unit where 23 MW of electricity are produced. Fig. 3 presents the process flow diagram of the MTG procedure.

Figure 2 – Molar composition of components in the simulated methanol synthesis reactor

Figure 3 – Process flow diagram for the MTG process

3.2 Butanol production process

SuperPro Designer software has been employed for the simulation of biobutanol production process due to its suitability of modelling bioprocesses and Matlab for modelling the fermentation kinetics. The steps of each process can be classified as: 1) Bagasse pre-treatment, 2) Cellulose Hydrolysis, 3) Sugars co-fermentation 4) Product recovery and 5) CHP unit. Dilute acid pretreatment was selected as pretreatment method where bagasse is mixed with dilute sulphuric acid solution (2%) and enters a reactor vessel (T=112°C, P=0.5 MPa, t=30min). The solid product (cellulose and lignin) are separated from the liquid stream (hydrolysates) and after being mixed with cellulases enters the cellulose hydrolysis reactor. The cycle time of the process is equal to 50 h and the temperature 50°C. The pretreatment and the hydrolysis reactor were simulated as batch conversion reactors and the occurred reactions as well as the conversion coefficients were adopted from the literature [48]. The required enzymes for the hydrolysis step were obtained by an external supplier and it was assumed that 0.02 kg of cellulases can hydrolyse 1 kg of cellulose [49]. Afterwards hemicellulose hydrolysates have to be detoxified mainly from the sulphuric acid which is partially eliminated after reacting with ammonia and Ca(OH)2. The produced gypsum is separated via a solid liquid separator. Separate hydrolysis and cofermentation approach has been applied for all the butanol production. This way processes can attain higher product yields since hydrolysis and fermentation operate at their own optimum conditions. However they come with increased volumes and residence time [48]. Before fermentation lignin is separated via filtration and is utilised in a CHP unit (a typical Rankine cycle unit [50]). ABE fermentation is strictly anaerobic and catalysed by strains of Clostridium acetobutylicum bacteria. Participated reactions can be summarized as follows:

Sugars→C3 H 6O+CO2+4 H 2

(33)

Sugars→C4 H10O+C O 2+H 2 O

(34)

Sugars→C2 H 6O+CO2

(35)

Sugars→C4 H 8 O2(butyric acid )+CO2+H 2

(36)

Sugars→C2 H 4O2(aceticacid)

(37)

Sugars→bacteria+C O2+H 2O

(38)

Reactions 36 and 37 have not been included in the model and conversion coefficients have been adopted from literature (2% and 2.5% respectively) [49]. The model for the batch fermentation is a combination of Haldane’s model [51] for substrate inhibition and Aiba’s [52] for product inhibition. Glucose and xylose have not been treated separately but as total sugars. In addition acetic acid inhibition on cell growth was taken into consideration. Thus growth rate can be considered as a modified Monod model and is expressed as follows.

μ=μmax∗S

K S+S+ S2

K I

∗∏i=1

3

exp(−Pi

KPIi)∗exp (−K AI∗A)

(39)

Biomass growth rate can subsequently be defined by equation 5:

dXdt

=μ∗X−kd

(40)

Luedking-Piret model has been used in order to determine ethanol production rate while for the substrate consumption an overall carbon balance has been applied.

d Pi

dt=ai

dXdt

+β i X

(41)

dSdt

=−∑

i=1

3 1Y Pi S

∗d Pi

dt−

1Y XS

∗dX

dt−mX

(42)

where i=butanol , acetone , ethanol

Fig. 4 illustrates the profiles of ABE fermentation substances against time. Kinetic parameters have been calculated by fitting model with respective published experimental data [53] by minimizing the sum of squared errors

between experimental and computed value (see Table 3). For this purpose an optimization routine based on genetic algorithm followed by a gradient based method (SQP) has been developed in Matlab [54]. Finally the code calculates as well the extent of the reactions and transfers the values to SPD.

After fermentation the broth enters a gas striper where the organics are separated from the bulk water and the solids. Then the organic mixture is cooled down to 40 °C and via flash separation a gas mixture of H2 and CO2 can be obtained. This mixture can be sold in methanol production plants as feedstock. The condensate stream of the gas stripper is sent to centrifugation so as to recover cells and eventually recycle them. Afterwards a series of distillation columns separate the desired products. The substances that need to be distinguished in the ABE process are acetone, ethanol, butanol, water and azeotrope of ethanol and water. This can be accomplished by employing numerous of separation methods such as distillation, absorption and adsorption. During the dominant method [55], the broth enters the first distillation column where from the top of it a mixture of ethanol, acetone and water leave while the bottom stream contains principally butanol of 99.5 % wt. purity which is appropriate for transportation fuels. The top stream heads to a second distillation column where acetone is concentrated, purified and exits the column from the top stream. The bottom stream consists of water and ethanol and it is subjected to further separation in a third distillation where from the top stream the azeotrope of ethanol-water (94.4%) is sent out and enters an adsorption unit. The water stream from the bottom is sent back to the fermenter. Fig. 5 provides the PFD for the biobutanol production process and Table 4 a summary of the simulation results for both processes.

Figure 4 – Concentration of the main products versus time during ABE fermentation. Symbols represent data points from [53], lines generated from Eqs 40-42.

Table 3 – Computed values of the kinetic parameters for the ABE fermentation

model

Parameter Computed valueμmax (h-1) 0.47KS (kg m-3) 1.29KI (kg m-3) 162KPI (kg m-3) (butanol) 10,000KPI (kg m-3) (acetone) 10,000KPI (kg m-3) (ethanol) 177KAI (m3 kg-1) 0.02kd (h-1) 0.34αB (kg kg-1) (butanol) 4.8βB (kg kg-1 h-1) (butanol) 0.28αA (kg kg-1) (acetone) 4.7βA (kg kg-1 h-1) (acetone) 0.14αE (kg kg-1) (ethanol) 4.8βE (kg kg-1 h-1) (ethanol) 0.1m (kg kg-1 h-1) 0.373YPS (kg kg-1) (butanol) 0.41YPS(kg kg-1) (acetone) 0.64YPS(kg kg-1) (ethanol) 0.51YXS (kg kg-1) 0.24

Figure 5 – Process flow diagram for the biochemical butanol process

Table 4 – Simulation results MTG Butanol

Methanol reactor volume (m3)

238 -

Fermenter reactor volume (m3)

- 9,100

Heating duties (MW) 89 110Cooling duties (MW) 135 140Electricity generation (MW) 23 17Electricity consumption (MW)

18 6

Gasoline productivity (t h-1) 12 -LPG productivity (t h-1) 1.1 -Butanol productivity (t h-1) - 7.9Acetone productivity (t h-1) - 2.3Ethanol productivity (t h-1) - 0.8H2+CO2 (t h-1) - 3.7

4. Results

4.1 Exergy efficiency

The calculation of exergy efficiency is the most reliable way to evaluate the performance of one process and it was calculated by utilizing Eqs (5-9). All the necessary thermodynamic data was extracted from the simulators. Given the process technology designed and modelled and the assumptions made as described in the previous section, MTG process attains higher exergy efficiency than butanol production, i.e. 47% and 42% respectively. This can be attributed to the highly energy intense downstream process for butanol recovery and to the fact that during ABE fermentation apart from butanol, acetone and ethanol are also produced which have lower energy content than gasoline and LPG. As depicted in Fig. 6, the fermentation section is responsible for the higher exergy losses, mainly due to the fact that butanol production does not surpass 75% of the maximum theoretical yield. In fact, the low butanol productivity is a major drawback and is due to low cell density caused by butanol inhibition [56]. The implementation of continuous fermentation along with the use of cell immobilization and cell recycling, which results in increased cell density inside

the reactor, could increase volumetric productivity compared with batch fermentation. This way, reduction of reactor volume and operational period could be achieved and consequently process efficiency and economics are improved [57]. Nevertheless, the overall butanol yield in continuous processes is lower than in batch operations, i.e. 0.22-0.26 kg kg -1 and 0.28-0.33 kg kg-1

respectively [58]. Additionally, the prerequisite for expensive membranes or membrane fouling hinders the commercial establishment of cell immobilization technology [59]. With respect to the MTG process the chief contributors to the losses are the gasifier and the methanol synthesis reactor. The total exergy losses for butanol production are 167 MW and for the gasoline case 157 MW.

Figure 6 - Exergy losses per section for the investigated procedures (a-MTG, b-butanol)

(a)MTG

(b) Butanol

4.2 Exergy costing

Exergy analysis has been finalised by calculating the cost related to the exergy destruction throughout the process. This expenditure (CE) refers to the cost of producing a fuel which can provide useful exergy equal to the destructed exergy under the same conditions. It can be described as follows [60]:

CE=I ×h×q

ψ

(43)

Where I is the amount of the exergy losses (ireversibilities), h is the annual working hours of the plant, q the bagasse cost (= 17 $ t-1 [34] or 0.96 $ GJ-1) and ψ the exergetic efficiency. In order to precisely compare the outcomes derived from the irreversibility cost analysis, the total cost per exergy flow input

(c E) was calculated as the ratio of the irreversibility cost (CE) over the exergy

flow input to each process (see Table 5). The exergy input comprises the bagasse exergy content as well as heat and electricity imports. It is quite obvious that there is a strong connection between efficiency and cost – the more the efficiency the less the exergetic costs. As a result, the cost of producing butanol is higher than the gasoline production cost.

Table 5 – Annual costs of irreversibilities for the examined processesProcess ψ (%) CE (M$) Exergy input

(MW)c E (k$ MW-1)

Gasoline 47 9.2 323 28.5Butanol 42 11 328 33.5

4.3 Economic evaluation

According to the mass and energy balances presented in the previous chapter, the total fixed capital investment was calculated based on the capacities of the equipment components of each process. All the necessary equations were presented in the methodology section. The selling price of the biofuels was adopted from relevant published data [15, 21]. As depicted in Fig. 7 the MTG conversion route is more profitable than biochemical butanol production. This can be explained by the low butanol productivity compared to gasoline and the high costs related with the recovery section of butanol. Furthermore a major

bottleneck of all the biochemical processes of lignocellulosic biomass is the high enzyme (necessary for the cellulose hydrolysis step) cost which approximately counts for 15% of the total annual cost. A proposed solution to overcome this issue is the on-site production of enzymes utilizing part of cellulose as substrate. According to recent studies, though, this approach does not lead to cost reduction [61, 62]. Taking into account all the above, it is expected that the gasoline production cost would be lower than the butanol one, 0.6 $ L-1 and 0.66 $ L-1 or in terms of energy 18.38 $ GJ-1 and 23.8 $ GJ-1 respectively.

(b)

(a)

Figure 7 - Economic performance of the investigated biomass conversion routes

4.4 CO2 emissions

For both processes, CO2 is mainly generated in the CHP units, since the CO2

produced in the gasifier is later used as substrate for methanol synthesis and that produced during ABE fermentation is considered as by-product. Also, in the calculations, the emissions related with producing steam from conventional sources (to cover the heating demand of each process) were also included. Typical value for CO2 emitted from raising steam is 201 g kWh-1 [6]. Based on this analysis, the more environmentally benign process is the butanol production generating 60 kt y-1 of CO2 compared to 75 kt y-1 of CO2 emitted during the MTG process.

4.5 Sensitivity analysis

Then, sensitivity analysis was conducted on biofuels production cost by varying the process related variables of each alternative. It was observed that the most crucial technical parameters for biobutanol production are the enzyme loading, sugars conversion during fermentation, the energy demand of the downstream process and cellulose conversion to glucose. As for gasoline production, the cost of the catalysts for methanol synthesis, and upgrading and the amount of the unconverted gas that is recycled to the methanol synthesis reactor. These sensitivities are illustrated in Fig. 8. It can be observed that at the most optimistic scenario the production cost of biobutanol could be reduced to 23 $ GJ-1 which is still reasonably higher than the 18.25 $ GJ-1 of gasoline.

(a)

Figure 8 - Sensitivity analysis on biofuels production cost for each alternative (a-MTG, b-butanol)

4.6 Risk analysis

So far, it was assumed that the financial analysis was deterministic and the values of each key factor are well known with absolute certainty. In real life many, if not all, of these parameters are subjected to changes. In this study, the behaviour of NPV was examined with the values of TCI, revenues and OC fluctuating as Table 6 illustrates. By employing Matlab and using probabilistic Monte Carlo simulation, meaning that the net present value was calculated for random combinations of the above factors within their limits, it can be observed which process is more tolerant to such changes. The following histograms (Fig. 9) depict the results of this analysis and additionally the mean value and standard deviation for each case were calculated. The risk analysis indicates that MTG process seems to be more reliable through the changes of key economic factors since its mean value of NPV is less than butanol, i.e. M$ 22 and M$ 12 respectively as well as the standard deviation is lower, i.e. 112 and 167.

(b)

Table 6 – Limits used for key financial variablesProfitability factor Variation (%)

TCI -10 to 25

OC -15 to 25

Revenues -30 to 25

Interest rate -25 to 25

-150 -100 -50 0 50 100 150 2000

5

10

15

20

25

-200 -150 -100 -50 0 50 100 150 2000

5

10

15

20

25

NPV (£M)

Freq

uenc

y

Figure 9 – Risk analysis of NPV for the investigated processes (a-MTG, b-butanol)

4.7 Multicriteria analysis

The criteria used in this study were exergy efficiency (rank 1), NPV (rank 2) and CO2 emissions (rank 3). The exergy efficiency, extracted by profound rigorous thermodynamic analysis, was considered as the most crucial. The NPV was used as a profitability criterion and the carbon dioxide emissions take into account

(b)(a)

Frequency

Frequency

NPV (M$)

NPV (M$)

the environmental impact of each alternative. After determining the criteria ranking, the normalised importance weights were calculated (see Table 7).

Table 7 - Computation of normalised weights using ranking methodCriteria Rank position Rank score ri n-ri+1 wi

Exergy efficiency

First 1 3 0.5

NPV Second 2 2 0.34

CO2 emissions Third 3 1 0.16

After defining the weight of each criterion it was possible to calculate the score, S j, attributed to each conversion route. As depicted in Table 8, the MTG

process achieves higher overall score than the butanol production, i.e. 97% and 90% respectively. It is quite obvious that a crucial factor, when MCDA is used, is the criteria ranking which can vary according to the goals of the decision maker. In this case, the CO2 emissions, as discussed in section 4.4, are not too high and thereby they were considered as the criterion of least importance. Nevertheless, even if the ranking was differentiated and the CO2 emissions were second and NPV third, the overall score of butanol would be still lower than gasoline, i.e. 92.7% and 93.3% respectively. The only case that butanol can outplay gasoline production is when the environmental performance is of the highest importance.

Table 8 – Overall scores for the investigated options MTG Butanol

Energy efficiency 1 0.89

NPV 1 0.88

CO2 emissions 0.8 1

Sj (%) 97 90

5. Concluding remarks

In this work, two different scenarios for the exploitation of bagasse were designed, evaluated and compared. The pathways under investigation are gasoline production via methanol (syngas derived) upgrade and biochemical butanol production. Exergy analysis clearly favours methanol to gasoline process achieving higher efficiencies than biobutanol production, i.e., 47% and 42% respectively and lower losses. According to the economic assessment, gasoline production outplays butanol achieving higher values for all economic indicators. The production cost of a biofuel is a significant index especially when comparing the financial feasibility between different conversion pathways. In this study, it was found that gasoline production cost is 18.38 $ GJ-1and biobutanol 23.8 $ GJ-1. On the other hand, butanol is responsible for lower CO2

emissions, but in general both processes are environmental friendly since the only contributors to those emissions are the CHP units. As a result, the MCDA attributes a higher score to gasoline production than butanol, i.e., 97% and 90% correspondingly. The MTG technology can provide an efficient way of adding value to methanol production. The catalysts developed for this purpose can achieve high selectivity to gasoline-range hydrocarbons (approximately 85%). MTG efficiently converts methanol to one liquid product: ultra-low-sulfur, low-benzene regular octane gasoline [63]. On the other hand, the major bottlenecks of butanol production are the low fermentation productivities, the high enzyme cost, the energy intensive product recovery section and the difficulty in handling lignocellulosic feedstock due to the high lignin content which is difficult to crack down and thereby does not contribute to fuel production. Significant advances have to be made to troubleshoot these obstacles and increase the efficiency of butanol production. Towards this direction, recent studies focus on addressing these challenges via research into the use of continuous fermentation reactors with recycle membrane reactors, improved bacteria strains, hybrid pretreatment methods, and novel downstream procedures [12, 64, 65]. Nevertheless, these technologies are yet to be established on a large scale and thereby at the current technological status the MTG process provides more benefits than butanol production.

REFERENCES

[1] Naik SN, Goud VV, Rout PK, Dalai AK. Production of first and second generation biofuels: A comprehensive review. Renewable and Sustainable Energy Reviews. 2010;14(2):578-97.

[2] Kamm B, Gruber PR, Kamm M. Biorefineries - industrial processes and products : status quo and future directions. Weinheim: Wiley-VCH; 2010.

[3] Sadhukhan J, Ng KS, Hernandez EM. Biorefineries and chemical processes : design, integration and sustainability analysis. John Wiley & Sons; 2014.

[4] Du C, Campbell GM, Misailidis N, Mateos-Salvador F, Sadhukhan J, Mustafa M, et al. Evaluating the feasibility of commercial arabinoxylan production in the context of a wheat biorefinery principally producing ethanol. Part 1. Experimental studies of arabinoxylan extraction from wheat bran. Chemical Engineering Research and Design. 2009;87(9):1232-8.

[5] Martinez-Hernandez E, Martinez-Herrera J, Campbell GM, Sadhukhan J. Process integration, energy and GHG emission analyses of Jatropha-based biorefinery systems. Biomass Conversion and Biorefinery. 2014;4(2):105-24.

[6] Vlysidis A, Binns M, Webb C, Theodoropoulos C. A techno-economic analysis of biodiesel biorefineries: Assessment of integrated designs for the co-production of fuels and chemicals. Energy. 2011; 36(8):4671-83.

[7] Chauhan MK, Varun, Chaudhary S, Kumar S, Samar. Life cycle assessment of sugar industry: A review. Renewable and Sustainable Energy Reviews. 2011;15(7):3445-53.

[8] Luo L, van der Voet E, Huppes G. Life cycle assessment and life cycle costing of bioethanol from sugarcane in Brazil. Renewable and Sustainable Energy Reviews. 2009;13(6–7):1613-9.

[9] García V, Päkkilä J, Ojamo H, Muurinen E, Keiski RL. Challenges in biobutanol production: How to improve the efficiency? Renewable and Sustainable Energy Reviews. 2011;15(2):964-80.

[10] Cheng JJ, Timilsina GR. Status and barriers of advanced biofuel technologies: A review. Renewable Energy. 2011;36(12):3541-9.

[11] Kumar M, Gayen K. Developments in biobutanol production: New insights. Applied Energy. 2011;88(6):1999-2012.

[12] Li Z, Shi Z, Li X, Li L, Zheng J, Wang Z. Evaluation of high butanol/acetone ratios in ABE fermentations with cassava by graph theory and NADH regeneration analysis. Biotechnology and Bioprocess Engineering. 2013;18(4):759-69.

[13] Riaz A, Zahedi G, Klemeš JJ. A review of cleaner production methods for the manufacture of methanol. Journal of Cleaner Production. 2013;57:19-37.

[14] Galadima A, Muraza O. From synthesis gas production to methanol synthesis and potential upgrade to gasoline range hydrocarbons: A review. Journal of Natural Gas Science and Engineering. 2015;25:303-16.

[15] Kumar M, Goyal Y, Sarkar A, Gayen K. Comparative economic assessment of ABE fermentation based on cellulosic and non-cellulosic feedstocks. Applied Energy. 2012;93:193-204.

[16] Tao L, He X, Tan ECD, Zhang M, Aden A. Comparative techno-economic analysis and reviews of n-butanol production from corn grain and corn stover. Biofuels, Bioproducts and Biorefining. 2014;8(3):342-61.

[17] Qureshi N, Blaschek HP. ABE production from corn: a recent economic evaluation. J Ind Microbiol Biotechnol. 2001;27(5):292-7.

[18] Pfromm PH, Amanor-Boadu V, Nelson R, Vadlani P, Madl R. Bio-butanol vs. bio-ethanol: A technical and economic assessment for corn and switchgrass fermented by yeast or Clostridium acetobutylicum. Biomass and Bioenergy. 2010;34(4):515-24.

[19] Swana J, Yang Y, Behnam M, Thompson R. An analysis of net energy production and feedstock availability for biobutanol and bioethanol. Bioresource Technology. 2011;102(2):2112-7.

[20] Phillips SD, Tarud JK, Biddy MJ, Dutta A. Gasoline from wood via integrated gasification, synthesis, and methanol-to-gasoline technologies. Golden, Colo.: National Renewable Energy Laboratory; 2011. NREL/TP-5100-47594 .

[21] Jones SB, Zhu Y. Techno-economic Analysis for the Conversion of Lignocellulosic Biomass to Gasoline via the Methanol-to-Gasoline (MTG) Process; 2009. Richland, WA: Pacific Northwest National Laboratory. PNNL-18481.

[22] Kempegowda RS, Pannir Selvam PV, Skreiberg Ø, Tran K-Q. Process synthesis and economics of combined biomethanol and CHP energy production derived from biomass wastes. Journal of Chemical Technology & Biotechnology. 2012;87(7):897-902.

[23] Aspen Technology. Aspen Plus: Getting Started Modelling Processes with Solids. Burlington, MA; 2010.

[24] Botha T, von Blottnitz H. A comparison of the environmental benefits of bagasse-derived electricity and fuel ethanol on a life-cycle basis. Energy Policy. 2006;34(17):2654-61.

[25] Parikh J, Channiwala SA, Ghosal GK. A correlation for calculating HHV from proximate analysis of solid fuels. Fuel. 2005;84(5):487-94.

[26] Ptasinski KJ, Prins MJ, Pierik A. Exergetic evaluation of biomass gasification. Energy. 2007;32(4):568-74.

[27] Wooley R., Putsche V. Development of an ASPEN PLUS Physical Property Database for Biofuels Components. Golden, Colo.: National Renewable Energy Laboratory; 1996. NREL/MP-425-20685.

[28] Park SR, Pandey AK, Tyagi VV, Tyagi SK. Energy and exergy analysis of typical renewable energy systems. Renewable and Sustainable Energy Reviews. 2014;30:105-23.

[29] Ptasinski KJ, Prins MJ, Pierik A. Exergetic evaluation of biomass gasification. Energy. 2007;32(4):568-74.

[30] Towler GP, Sinnott RK. Chemical engineering design: principles, practice, and economics of plant and process design. 2nd ed. Boston, MA: Butterworth-Heinemann; 2013.

[31] Spath PL, Dayton DC. Preliminary screening -technical and economic assessment of synthesis gas to fuels and chemicals with emphasis on the

potential for biomass-derived syngas. Golden, Colo.: National Renewable Energy Laboratory; 2003. NREL/TP-510-34929.

[32] Peters MS, Timmerhaus KD, West RE. Plant design and economics for chemical engineers. 5th ed. New York: McGraw-Hill; 2003.

[33] Hamelinck CN, Hooijdonk Gv, Faaij APC. Ethanol from lignocellulosic biomass: techno-economic performance in short-, middle- and long-term. Biomass and Bioenergy. 2005;28(4):384-410.

[34] Furlan FF, Filho RT, Pinto FH, Costa CB, Cruz AJ, Giordano RL, et al. Bioelectricity versus bioethanol from sugarcane bagasse: is it worth being flexible? Biotechnology for Biofuels. 2013;6(1):1-12.

[35] Murphy CW, Kendall A. Life cycle analysis of biochemical cellulosic ethanol under multiple scenarios. GCB Bioenergy. 2015;7(5):1019-33.

[36] Duffy A, Rogers M, Ayompe L. Renewable energy and energy efficiency : assessment of projects and policies. Chicester: Wiley Blackwell; 2015.

[37] Podvezko V. The Comparative Analysis of MCDA Methods SAW and COPRAS. Inz Ekon. 2011;22(2):134-46.

[38] Aspen Technology. Aspen Physical Property System: Physical Property Models, Burlington, MA; 2012.

[39] Di Blasi C, Signorelli G, Di Russo C, Rea G. Product Distribution from Pyrolysis of Wood and Agricultural Residues. Industrial & Engineering Chemistry Research. 1999 1999/06/01;38(6):2216-24.

[40] Sudiro M, Bertucco A. Production of synthetic gasoline and diesel fuel by alternative processes using natural gas and coal: Process simulation and optimization. Energy. 2009;34(12):2206-14.

[41] Trippe F, Fröhling M, Schultmann F, Stahl R, Henrich E. Techno-economic assessment of gasification as a process step within biomass-to-liquid (BtL) fuel and chemicals production. Fuel Processing Technology. 2011;92(11):2169-84.

[42] Sreejith CC, Muraleedharan C, Arun P. Thermo-Chemical Analysis of Biomass Gasification by Gibbs Free Energy Minimization Model-Part: II

(Optimization of Biomass Feed and Steam to Biomass Ratio). International Journal of Green Energy. 2013;10(6):610-39.

[43] Skrzypek J, Lachowska M, Moroz H. Kinetics of methanol synthesis over commercial copper/zinc oxide/alumina catalysts. Chemical Engineering Science. 1991;46(11):2809-13.

[44] Chinchen GC, Waugh KC, Whan DA. The activity and state of the copper surface in methanol synthesis catalysts. Applied Catalysis. 1986;25(1):101-7.

[45] Bussche KMV, Froment GF. A Steady-State Kinetic Model for Methanol Synthesis and the Water Gas Shift Reaction on a Commercial Cu/ZnO/Al2O3Catalyst. Journal of Catalysis. 1996;161(1):1-10.

[46] Kempegowda RS, Pannir Selvam PV, Skreiberg Ø, Tran K-Q. Process synthesis and economics of combined biomethanol and CHP energy production derived from biomass wastes. Journal of Chemical Technology & Biotechnology. 2012;87(7):897-902.

[47] De Kam MJ, Vance Morey R, Tiffany DG. Biomass Integrated Gasification Combined Cycle for heat and power at ethanol plants. Energy Conversion and Management. 2009;50(7):1682-90.

[48] Alfani F, Gallifuoco A, Saporosi A, Spera A, Cantarella M. Comparison of SHF and SSF processes for the bioconversion of steam-exploded wheat straw. Journal of Industrial Microbiology and Biotechnology. [journal article]. 2000;25(4):184-92.

[49] Liu B, Yang X, Song W, Lin W. Process Simulation Development of Coal Combustion in a Circulating Fluidized Bed Combustor Based on Aspen Plus. Energy & Fuels. 2011;25(4):1721-30.

[50] Wu M, Wang M, Liu J, Huo H. Assessment of potential life-cycle energy and greenhouse gas emission effects from using corn-based butanol as a transportation fuel. Biotechnol Prog. 2008;24(6):1204-14.

[51] Haldane JBS. Enzymes. Cambridge: M.I.T. Press; 1965.

[52] Aiba S, Nagatani M. Kinetics of Product Inhibition in Alcohol Fermentation. Biotechnology and Bioengineering. 1968;10(6):845-64.

[53] Chen Y, Zhou T, Liu D, Li A, Xu S, Liu Q, et al. Production of butanol from glucose and xylose with immobilized cells of Clostridium acetobutylicum. Biotechnology and Bioprocess Engineering. 2013;18(2):234-41.

[54] Rigaki A, Webb C, Theodoropoulos C. Double Substrate Limitation Model for the Experimental Scale-up of Succinic Acid Production from Biorefinery Glycerol. Chem Engineer Trans. 2013;35:1033-8.

[55] Liu J, Fan LT, Seib P, Friedler F, Bertok B. Downstream process synthesis for biochemical production of butanol, ethanol, and acetone from grains: generation of optimal and near-optimal flowsheets with conventional operating units. Biotechnol Prog. 2004;20(5):1518-27.

[56] Li Q, Cai H, Hao B, Zhang C, Yu Z, Zhou S, et al. Enhancing clostridial acetone-butanol-ethanol (ABE) production and improving fuel properties of ABE-enriched biodiesel by extractive fermentation with biodiesel. Appl Biochem Biotechnol. 2010;162(8):2381-6.

[57] Zheng J, Tashiro Y, Wang Q, Sonomoto K. Recent advances to improve fermentative butanol production: Genetic engineering and fermentation technology. J Biosci Bioeng. 2015;119(1):1-9.

[58] Qureshi N, Lai LL, Blaschek HP. Scale-Up of a High Productivity Continuous Biofilm Reactor to Produce Butanol by Adsorbed Cells of Clostridium Beijerinckii. Food and Bioproducts Processing. 2004;82(2):164-73.

[59] Tashiro Y, Takeda K, Kobayashi G, Sonomoto K. High production of acetone–butanol–ethanol with high cell density culture by cell-recycling and bleeding. Journal of Biotechnology. 2005;120(2):197-206.

[60] Martín C, Villamañán MA, Chamorro CR, Otero J, Cabanillas A, Segovia JJ. Low-grade coal and biomass co-combustion on fluidized bed: exergy analysis. Energy. 2006;31(2–3):330-44.

[61] Klein-Marcuschamer D, Oleskowicz-Popiel P, Simmons BA, Blanch HW. The challenge of enzyme cost in the production of lignocellulosic biofuels. Biotechnology and Bioengineering. 2012;109(4):1083-7.

[62] Michailos S, Parker D, Webb C. Simulation Studies on Ethanol Production from Sugar Cane Residues. Industrial & Engineering Chemistry Research. 2016;55(18):5173-9.

[63] Gayubo AG, Ortega JM, Aguayo AT, Arandes JM, Bilbao J. MTG fluidized bed reactor–regenerator unit with catalyst circulation: process simulation and operation of an experimental setup. Chemical Engineering Science. 2000;55(16):3223-35.

[64] Lee SY, Park JH, Jang SH, Nielsen LK, Kim J, Jung KS. Fermentative butanol production by Clostridia. Biotechnol Bioeng. 2008;101(2):209-28.

[65] Kraemer K, Harwardt A, Bronneberg R, Marquardt W. Separation of butanol from acetone–butanol–ethanol fermentation by a hybrid extraction–distillation process. Comput Chem Eng. 2011;35(5):949-63.

NomenclatureAbbreviationsACC Annualised capital costBEC Basic equipment costC CostCF Cash flowD DepreciationHHV Higher heating valueIRR Internal rate of returnLHV Lower heating valueMCDA Multicriteria decision analysisMTG Methanol to gasolineMW Molecular weight

NPV Net present valueOC Operating costPP Payback periodR RevenuesROI Return on investmentTAC Total annual costTCI Total capital investment

VariablesE exergy content (MW)EQ exergy content of a heat stream (MW)εch chemical exergy (MJ kg-1)εph physical exergy (MJ kg-1)kd specific cell death rate (h-1)KAI acetic acid inhibition constant (m3 kg-1)KI substrate growth inhibition term (kg m-3)KN nitrogen saturation constant (kg m-3)KPI product inhibition term for growth (kg m-3)KS saturation growth constant (kg m-3)m maintenance parameter (kg kg-1 h-1)P product concentration (kg m-3)Pm product concentration above which there is no growth (kg m-3)R ideal gas constant (J mol-1 K-1)S substrate concentration (kg m-3)yi mass fraction of component i after pyrolysis YP/S product yield coefficient (kg kg-1)YX/S biomass yield coefficient (kg kg-1)X biomass concentration (kg m-3)Greek Lettersα growth-related product formation coefficient (kg kg-1)β non-growth-related product formation coefficient (kg kg-1 h-1)η energy efficiencyμ specific growth rate (h-1)μmax maximum specific growth rate (h-1)ψ exergy efficiency