Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production

Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production Bioenergy systems for the future 6 thermodynamic analysis of ethanol reforming for hydrogen production

Thermodynamic analysis

of ethanol reforming

for hydrogen production

G Tamasi, C Bonechi, A Magnani, G Leone, A Donati, S Pepi, C RossiUniversity of Siena, Siena, Italy

Abbreviations

ESR ethanol steam reforming

WGSR water-gas shift reaction

It is widely recognized that the modern lifestyles require high consumption ofenergy, the generation of which still relies heavily on the use of fossil fuels deriva-tives, a nonrenewable source A new approach to meet the global energy request ismandatory, and the need for renewable alternatives is urgent

Among the renewable resources for energy production, the solar, wind, and biomassare the most promising However, they are usually site-specific and seasonally intermit-tent Hydrogen has been identified as a good “energy carrier” to support sustainableenergy development (Goltsov et al., 2006; Ni et al., 2006) and can be used in fuel cells

to generate electricity with high efficiency The use of hydrogen is very clean as the onlyfinal by-product is water However, in order to support sustainable hydrogen economy,

it is mandatory to produce hydrogen in a clean and renewable way

At present, almost 90% of the hydrogen is commercially produced in an economicallycompetitive method, by gasification, partial oxidation reactions of fossil fuels (Das andVeziroglu, 2001; Haryanto et al., 2005), via steam reforming reactions of hydrocarbons,for example, coal, natural gas, liquefied petroleum gas, propane, methane (CH4), gaso-line, and light diesel The current worldwide production is around 5
1011N m3/year(Vaidya and Rodrigues, 2006) The hydrogen is mostly used as a feedstock in the chem-ical industry and in the manufacture of ammonia and methanol, in refinery reprocessingBioenergy Systems for the Future http://dx.doi.org/10.1016/B978-0-08-101031-0.00006-5

© 2017 Elsevier Ltd All rights reserved.

and conversion processes (Vaidya and Rodrigues, 2006; Sun et al., 2004), but it is alsoused as feedstock in the production of ethylene, acetaldehyde, acetone, etc.

Fossil-fuel-based (e.g., natural gas) production of hydrogen fails to provide asolution to deal with the huge amount of carbon dioxide emissions during thesteam reforming at high-temperature processes As a result, studies on possibleeffective alternatives to produce renewable hydrogen in a clean and safe wayhave attracted considerable attention Particularly interesting is the production

of hydrogen through pyrolysis, gasification, and steam reforming processes

of lignocellulosic biomass, a renewable resource (Wang et al., 1997; Garcia

et al., 2000; Stiegel and Maxwell, 2001) or through intermediate liquid biofuels(Fatsikostas et al., 2002)

Among the liquid biofuels, ethanol (C2H5OH) is a good candidate for severalreasons: (i) Ethanol is renewable and is becoming increasingly available; (ii) it is easy

to transport, biodegradable, low in toxicity, and low volatile; and (iii) it could beeasily decomposed in the presence of water to generate a hydrogen-rich mixture Thislatter process, steam reforming, is conducted at 200–800°C in the presence ofcatalysts

The main advantage of liquid biofuels in general and ethanol in particular is theirhigh energy density, ease of handling, and “on-demand” production to feed fuel cells,with applications in mobile and stationary grid-independent power systems

In addition, ethanol can be produced renewably from several biomass sources,including energy plants, lignocellulosics, waste materials from agricultural andagro-industrial processes, forestry residue materials, and organic fraction of municipalsolid waste The ethanol produced in this way is called bioethanol, which is a mixture

of ethanol and water with a molar ratio of 1:13 (about 18 wt% ethanol;Roh et al.,2006; Benito et al., 2005; D€om€ok et al., 2007)

Furthermore, in contrast to other proposed fossil-fuel-based systems (like anol and gasoline), bioethanol has the significant advantage of being a nearly CO2

meth-neutral process, since the carbon dioxide produced was created by biomass growth,thus offering a nearly closed carbon loop, with high efficiency The proposedoverall process for the hydrogen (and electricity) production from biomass is sche-matically shown inFig 6.1(Fatsikostas et al., 2002) The biomass from plant cul-tivation (1) and/or residues of agricultural and agro-industrial processes (2) areused for the production of bioethanol by saccharification/fermentation reactions(3) The aqueous mixture is then distilled to 45%–55% ethanol (4), meanwhilethe fermented solid can feed an anaerobic digestion unit (5) where biogas is pro-duced (mixture of CH4and CO2) In addition, the anaerobic digester could also

be fed by the organic fraction of municipal solid waste (6) Finally, a gas mixturerich in H2is produced by reformation of bioethanol (7) and biogas (8) A water-gasshift reactor (WGSR) is then used for the transformation of CO into H2and CO2(9) The final mixture would be further purified by selective oxidation of residual

CO (10) for feeding a fuel cell (11) and produce electricity In addition, a combustion reactor (12) may be used to clean up the effluent of the fuel cell(Fatsikostas et al., 2002)

6.1.1 Bioethanol production

In this overview, it is interesting to underline that among the biomass materials that areavailable from ethanol production, sugar cane, switch grass, potatoes, corns, and otherstarch-rich materials can be effectively converted into ethanol by yeast-assisted fermen-tation, but the cost of this process is high because of the expensive feedstock plantation

Residues of agroindustries and

(5)

Biogas reformation

(8)

Aqueous broth 8%–10% ethanol

Thermodynamic analysis of ethanol reforming for hydrogen production 189

O O

O HO HO O

HO

O O O

HO

OH O

O

OH

O O

O

OH O HO

(C)

O

O HO OH

Fig 6.2 Cellulose (A), hemicelluloses (B), and lignin (C) structures

6.1.2 Ethanol steam reforming

The target reaction for the ethanol steam reforming (ESR) is the stoichiometric “ideal”reaction of ethanol (C2H5OH) that produces hydrogen (H2) that can be subsequentlyused for feeding fuel cells The ESR reaction is

However, it is very well known that the ESR “real” process is more complex and sists of several subreactions that lead to the formation of several intermediates andby-products in the final mixture, depending on the temperature, pressure, ethanol/water ratio, nature of the catalyst, specific plant characteristics, etc It is, therefore,fundamental to take into account the possible by-products that could inactivate thecatalyst used in the reforming process and/or the fuel cell themselves

con-Given this challenge, it is fundamental to explore the subreactions withoutdiscarding the many reactions on secondary species that could form Some of thosereactions (Fig 6.4;Haryanto et al., 2005) are summarized as

C2H5OH + 3H2O! 2CO2+ 6H2+ BY PRODUCTS (6.2)

Diluited acid hydrolysis

Enzyme production

Simultaneous saccharification and fermentation

Pretreatment

Cellulose/hemicellulose

Bioethanol

Fig 6.3 Possible processes for producing bioethanol from lignocellulosic biomass

Thermodynamic analysis of ethanol reforming for hydrogen production 191

A list of a selected group of the possible by-reaction is also reported inTable 6.1out considering reaction that can bring about NOxby-products, coming from the N2present in the air).

(with-Several experimental and theoretical studies have been performed in the last couple

of decades to shed light on this complex process (Sun et al., 2004; Freni et al., 1996;Fishtik et al., 2000; Ioanides, 2001; Comas et al., 2004; Goula et al., 2004; Fatsikostasand Verykios, 2004; Llorca et al., 2001; Diagne et al., 2002; Marin˜o et al., 2003;Batista et al., 2003; Sheng et al., 2004; Segal et al., 2003; Rasko et al., 2004)and the relation with the catalyst used (Sun et al., 2004; Fatsikostas andVerykios, 2004; Llorca et al., 2001, 2009; Diagne et al., 2002; Marin˜o et al., 2003;

Metal catalyst Steam reforming

Metal catalyst Polymerization Water-gas-shift

Water-gas-shift

Dehydration

Metal catalyst Metal catalyst

Metal catalyst Metal catalyst

O H

H H

Table 6.1 ESR reaction and selected subreactions taken

Eq.(6.27)

C2H5OH +½O2!2CO+3H2 Autothermal C2H5OH steam

reforming to H2

Eq.(6.28)

Subreaction group A: other possible steam reforming reactions for C2H5OH

C2H5OH + H2O!CH4+ CO2+ 2H2 C2H5OH steam reforming to CH4 Eq.(6.10)

C2H5OH + H2O!2CO+4H2 C2H5OH steam reforming to

Subreaction group B: reaction for CH4

CH4+ H2O!CO+3H2 CH4steam reforming to CO Eq.(6.16)

CH4+ 2H2O!CO2+ 4H2 CH4steam reforming to CO2 Eq.(6.17)

Subreaction group C: reaction for CO

CO + H2O!CO2+ H2 Water-gas shift reaction (WGSR) Eq.(6.19)

Subreaction group D: reaction for CH3COCH3, CH3COH, CH3COOH

CH3COCH3+ 5H2O!3CO2+ 8H2 CH3COCH3steam reforming

CH3COH + 3H2O!2CO2+ 5H2 CH3COH steam reforming to CO2 a

CH3COH + H2O!2CO+3H2 CH3COH steam reforming to CO a

Sub-reaction group E: other reactions for C2H5OH

Batista et al., 2003; Iwasa et al., 1999; Kaddouri and Mazzocchia, 2004; Liguras et al.,2003; Frusteri et al., 2004) Several experimental studies showed effect of the temper-ature on the conversion and selectivity toward the main products and by-products.Studies at different temperatures have allowed optimizing the experimental conditions

to maximize hydrogen yield, to limit by-product formation, and to propose reactionschemes (Sun et al., 2004; Fishtik et al., 2000; Ioanides, 2001; Comas et al., 2004;Goula et al., 2004; Fatsikostas and Verykios, 2004; Llorca et al., 2001; Diagne

et al., 2002; Batista et al., 2003) The behavior of the catalysts versus time has alsobeen considered (Sun et al., 2004; Goula et al., 2004)

6.1.3 Brief overview on the catalyst for the ESR process

The main types of reforming could be classified as (i) steam reforming (reaction6.1,

Table 6.1), (ii) autothermal reforming (see below, reaction 6.27), and (iii) partialoxidation (see below, reaction6.28) All these processes have common aspects of aprimary reforming reaction that brings about a mixture of gases rich in H2, startingfrom the reactants, and processes of purification/separation of the stream mixture.Such catalytic reforming processes of ethanol can occur through various catalystswith a metallic phase that acts as the catalytic activator and of a supporting phase onwhich the metal microparticles are loaded and spread out at a variable weightconcentration

From analyses by laboratory research studies and experimental data from plants, theinitial step of the reforming process consists of the adsorption of ethanol molecules onthe surface of the catalyst where the breakage of the bonds CdC, CdO, CdH, andOdH takes place and intermediate species form (ethylene, acetone, and acetaldehyde).The concentrations of these latter species depend on the catalyst (nature of metal and thesupport, the concentration of the metal, and other experimental parameters) Com-monly, the catalysts can be noble metals (Pt, Pd, Rh, and Ru) or other “block-d” metals(Ir, Cu, Co, and Ni), loaded on supporting phase by metal oxides (Al2O3, CeO2, SiO2,ZrO2, TiO2, MgO, La2O, and Y2O3) or mixed metal oxides As examples, on usingRh/CeO2catalyst, not only the yields are excellent, as high as 95% for H2(Deluga

et al., 2004), but also catalysts based on iridium and cobalt have been reported to bringabout quite good yields (Wang et al., 2009; Iulianelli and Basile, 2010)

Another important step of the overall process is the purification of the gas from thereforming reactions for the production of H2 ESR can be carried out in traditionalcatalytic reactors or in inorganic membrane reactors with an inner tubular-shapedinorganic membrane that is selective for the permeation of H2 These membranesoften consist of metals like Pd that are extremely selective toward H2permeation(but unfortunately very costly), and/or its alloys (like Pd/-Ag;Basile et al., 2015)

As an example, on a laboratory scale, an autosupported tubular membrane reactor sisting of Pd-Ag allows to get a conversion proximal to 100%, for ethanol reformingand a recovery of H2of c.90%, when operating at 400–500°C and relatively low reac-tion pressure by 1–3 bar (Iulianelli and Basile, 2010; Iulianelli et al., 2009, 2010a,b;Basile et al., 2008a,b)

In conclusion, the chemical nature of the species produced needs to be determined,and the reaction paths identified Subsequently, a thermodynamic analysis can be car-ried out to evaluate the maximum yield reachable in terms of the characteristic of thereactor The literature offers several thermodynamic models for the ESR process(Wang and Wang, 2008, 2009; Graschinsky et al., 2012; Mas et al., 2006), and a recentstudy based on an exergetic approach was published (Casas-Ledo´n et al., 2012).

On the basis of this reasoning, this chapter reports on the state of the art of ESRprocesses and a theoretical thermodynamic analysis of process paths with the goal

to determine the optimal working conditions for high yield production of H2from ethanol and, from it, to produce electric power, that is, through fuel cells

The analysis utilized classical thermodynamic properties (ΔH, ΔS, ΔG, and Keq) for aselected number of subreactions (Table 6.1) to compare values calculated for the mainESR, reaction(6.1), and for hypothesizing the optimal working conditions as regardstemperature (T, K) and pressure (P, atm) The autothermal ESR reactions(6.27)and

(6.28)were also considered for comparative purposes

The calculation was performed applying the basic rules of the thermodynamics:Enthalpy of formation, as a function ofT and P for each species

Entropy of reaction, as a function ofT and P

ΔSreaction¼ ΣproductsΔSð T, P Þ ΣreagentsΔSð T, P Þ (6.6)Gibbs free energy, as a function ofT and P

Equilibrium constant as a function ofT and P

Thermodynamic analysis of ethanol reforming for hydrogen production 195

TheΔH, ΔS, ΔG, and Keqparameters for selected reactions (Table 6.1) were studied

by varying the temperature T in the range 274–973 K (steps by 1 K) and pressure P inthe range 0.5–10 atm (steps by 0.25 atm) Particular attention was devoted to focus onvalues obtained for parameters P¼5 atm and T>450 K, comparable with typicalexperimental working conditions for ESR plants

The data for standard condition parameters (T°¼298 K and P°¼1 atm) for ΔH°,ΔS°, and Cp° were obtained from the CRC Handbook of Chemistry and Physics(CRC, 2005–2006), whereas the theoretical data as a function ofT and P were calcu-lated by using the REFerence fluid PROPerties software (Lemmon et al., 2013) andthe PRODE software (PRODE software, 2014) for comparison

REFPROP is a program developed by the National Institute of Standards and nology (NIST) that calculates the thermodynamic properties of selected fluids andmixtures It is not a database and does not contain any experimental information, asidefrom the critical and triple points of the pure fluids REFPROP is based on the mostaccurate pure fluid and mixture models currently available: equations of state explicit

Tech-in Helmholtz energy, the modified Benedict-Webb-RubTech-in equation of state, and

an extended corresponding states (ECS) model High accuracy was obtained by usingmany coefficients in the equations that are generally valid over the entire vapor andliquid regions of the fluid, including supercritical states; the upper temperature limitwas usually near the point of decomposition of the fluid, and the upper pressure limitwas defined by the melting line of the substance Table 6.2 reports the referencesrelevant to the equation of state used for each single fluid, and selected parametersrelevant to them, like critical point parameters (temperature, Tc, K; pressure, Pc,MPa), triple-point temperature (TT, K), liquid-gas transition temperature (TLG, K),decomposition temperature (TD, K), and maximum temperature as limit of the model(Tmax, K) Selected details about uncertainties in the equation of state are also reported.Furthermore, it is important to note that not all the fluids are implemented inREFPROP software; thus, some of the by-reactions reported inTable 6.1were notconsidered in this work (i.e., reactions in which the formation of carbon powderappear) N2(from air) reactions producing NOxwere not considered at that stage.Each reaction was singularly studied as without taking into account any synergicand direct competitive effect among them, as in real experimental systems ESR plantsare usually catalytically assisted, and the present study did not consider any preferen-tial path on the basis of possible kinetic effects

6.3 Analysis of thermodynamic properties

for the single reactions

6.3.1 Reaction (6.9) : Ideal ESR

Table 6.2 Fluids used in this work and relevant equation of state reference, critical point temperature

C2H5OH Schroeder, 2011 514.71 6.268 159.0 398–399 (5 atm)

351–352 (1 atm)

650 (5 atm) 973 (5 atm)The fundamental equation can compute heat capacities within 1%–2% The uncertainty is higher in the critical region and

near the triple point

(5 atm)

973 (5 atm)

At pressures up to 30 MPa and temperatures up to 523 K, the estimated uncertainty ranges 0.15% (in the vapor)–1.5% (in theliquid) in heat capacity Special interest was focused on the description of the critical region and the extrapolation behavior of theformulation (to the limits of chemical stability)

The equation of state is valid from the triple point to 500 K with pressures to 100 MPa The uncertainties in the equation are2% in heat capacities

CH4 Setzmann and Wagner, 1991 190.56 4.5992 90.694 –– 625 (5 atm) 938 (5 atm)

The heat capacities may be calculated within an uncertainty of 1%

CH3COCH3 Lemmon and Span, 2006 508.1 4.7 178.5 385–386 (5 atm) 550 (5 atm) 825 (5 atm)

The equation of state has the uncertainties in heat capacities of 1% The uncertainties in caloric properties may be higher atpressures above the saturation pressure and at temperatures above 320 K in the liquid phase and at supercritical conditions

Table 6.2 Continued

C2H4 Smukala et al., 2000 282.35 5.0418 103.99 –– 450 (5 atm) 675 (5 atm)

The uncertainty in heat capacity is 3% in the liquid phase, 0.2% in the vapor phase, and as high as 5% in the supercritical

region at higher pressures

The estimated uncertainty for heat capacities is 1.0%

H2O Wagner and Pruss, 2002 647.1 22.064 273.16 425–426 (5 atm)

373–374 (1 atm) ––

973 (5 atm)The uncertainty in isobaric heat capacity is 0.2% in the vapor and 0.1% in the liquid, with increasing values in the critical

region and at high pressures The uncertainties of saturation conditions are 0.025% in vapor pressure, 0.0025% in saturated

liquid density, and 0.1% in saturated vapor density The uncertainties in the saturated densities increase substantially as the criticalregion is approached

O2 Schmidt and Wagner, 1985;

without considering the formation of by-products, in the ranges of temperature

274–973 K and pressure 0.50–10.00 atm (see calculation methods for more details).The data are plotted inFigs 6.5–6.7, and selected parameters are reported inTable 6.3

Fig 6.5reports values ofΔHreactionas a function of temperature and pressure ditions and shows a regular trend, influenced by the temperature with respect to thepressure conditions The process is endothermic (ΔH>0) for all the consideredconditions and brings about an increase of mole number (Δn¼4) of gaseous species.Therefore, it is thermodynamically favored by high temperatures and low pressurevalues (ΔH298,1¼173.4530, ΔH673,1¼126.8319, ΔH973,1¼75.1636 kJ/mol, and

con-ΔH298,5¼173.4530, ΔH673,5¼126.0849, ΔH973,5¼74.8996 kJ/mol) It is worthy ofnote that the plotted surface shows an irregular area in the range of transitions phasetemperature for C2H5OH and H2O On the basis of this reasoning and also consideringthe usual working condition in experimental plant production, the coprocessby-reactions were studied in the range of temperature between 450 K and 973 K

Fig 6.6plots the trend ofΔGreactionas a function of temperature and pressure ditions showing that the reaction was not spontaneous at low temperatures(ΔGreaction>0, ΔG298,1¼65.7260 kJ/mol), but became spontaneous at T419 K,

con-T425 K, and T426 K for pressure of 1, 5, and 10 atm, respectively Also, in this case,

Fig 6.5 ΔHreaction(kJ/mol) as a function of temperature (K) and pressure (atm) conditionsfor the ideal ESR, reaction(6.9)

Thermodynamic analysis of ethanol reforming for hydrogen production 199

an increase in pressure was not confirmed, which may slightly disadvantage the process(ΔG673,1¼141.4295, ΔG673,5¼141.4843, and ΔG673,10¼141.0365 kJ/mol) Thisagrees with the value calculated for theKeq(reaction6.9, seeFig 6.7) confirming a con-siderable dependence with the temperature and very low dependence with pressure values,and indicating that at low temperatures, the reaction was shifted toward the reagents(Keq298,1¼3.012
1012,Keq298,5¼3.012
1012, andKeq298,10¼3.012
1012).

Selected computed data were compared with experimental data reported in theliterature showing a great agreement, that is, at standard conditions,ΔH298,1(calcd)¼173.4530 kJ/mol andΔH298,1(exp)¼174 kJ/mol (Vaidya and Rodrigues, 2006).Summarizing, from a thermodynamic point of view, the ideal ESR (reaction 6.9)would be strongly endothermic (ΔHreaction>0), and produces an increase in number

of moles, increasing the temperature and/or lowering the pressure would favor ESR sible competing reactions were considered, in a range of temperature, on the basis of thefluids experimental characteristics that are implemented in the computing software

Pos-6.3.2 Subreaction group A: Other possible steam reforming

reactions for C2H5OH

Other steam reforming reactions are possible for ethanol (seeTable 6.1, group A), andmost of these bring about undesired products, especially if the process would be used

to produce a gas mixture for fuel cells feeding Selected computed data forΔH, ΔS,

ΔG, and K , for group A subreactions, are reported inTable 6.4

–300 –250 –200 –150 –100

0 2 4 6 8

900 800 700 600 500 400 300

(B)

20 15

5

10 8 6 4 2

–10 –15 –5

10

0

900 800 700 600 500 400