author's personal copy - california state polytechnic ...minghengli/pdfs/thermodynamic analysis...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Thermodynamic analysis of adsorption enhancedreforming of ethanol

Mingheng Li*

Department of Chemical and Materials Engineering, California State Polytechnic University, Pomona, CA 91768, USA

a r t i c l e i n f o

Article history:

Received 9 July 2009

Received in revised form

14 September 2009

Accepted 21 September 2009

Available online 17 October 2009

Keywords:

Hydrogen generation

Adsorption enhanced reforming

Thermodynamic analysis

Graphite formation

a b s t r a c t

Adsorption enhanced reforming (AER) is a novel hydrogen generation method. To study

hydrogen yield and graphite formation in the AER process, a recently developed thermo-

dynamic code for reaction–adsorption systems is extended to account for the formation of

condensed species. Parametric analysis of the steam reforming of ethanol is performed to

study the effect of adsorption and other parameters (temperature, pressure, steam/carbon

ratio and O2/ethanol ratio in the feed) on the yield of hydrogen and the formation of

graphite. It is shown that with CO2 adsorption, the hydrogen yield is enhanced while the

graphite formation is suppressed. Based on the thermodynamic analysis, AER is a potential

low-temperature method for hydrogen generation.

ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction

As reducing the demand on fossil resources has been

a public concern, hydrogen, being a potential carrier of clean

energy, is now an important topic that may lead into a new

era of energy research. Among various industrial tech-

niques, steam reforming of hydrocarbons accounts for about

50% of the hydrogen produced over the world. In the United

States, 95% of the hydrogen production is made through

steam reforming of methane [1]. In this sense, steam

reforming serves as a foundation in America’s transition to

a hydrogen economy.

The past decade witnessed a significant development in

steam reforming. On the feedstock side, the reforming of

biomass-derived hydrocarbons (e.g., bio-ethanol) is widely

acknowledged as a potentially viable, renewable and

carbon-neutral (or even carbon-negative in conjunction

with sequestration) process [2–4]. On the technology side,

the steam reforming combined with CO2 removal using

adsorption [5–10] or membrane [11] is shown to signifi-

cantly enhance the yield and purity of hydrogen. When CO2

is removed from the reforming process, the thermody-

namic limitation of the water gas shift reaction

(COþH2O¼CO2þH2) is circumvented and therefore, the

chemical equilibrium shifts to the right, resulting in an

enhancement in the extent of the forward reaction.

Removing CO2 from the reactive system enables a relatively

lower reforming temperature, thus alleviating the require-

ment for both the catalyst and the reactor. Moreover, the

higher purity in the syngas leaving the reformer makes it

easier for downstream hydrogen separation, which is

typically achieved through a pressure swing adsorption (or

PSA) unit. Extensive experimental and computational

studies indicate that reforming with simultaneous CO2

removal is a very promising technique for enhanced

hydrogen production [11–14]. Ongoing research is

* Tel.: þ1 909 869 3668; fax: þ1 909 869 6920.E-mail address: [email protected]

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev ie r . com/ loca te /he

0360-3199/$ – see front matter ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.ijhydene.2009.09.054

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 3 6 2 – 9 3 7 2


conducted by various companies such as Chevron, Shell

and Intelligent Energy to commercialize these novel

reforming technologies [8,15,16].

Despite their wide utilization for hydrogen generation,

the reforming processes (including the aforementioned

enhanced reforming processes) might be accompanied by

unfavorable formation of graphite under certain conditions,

which could lead to catalyst deactivation and reformer

malfunction due to its accumulative nature [12,17–27].

Studies have shown that the formation of graphite can be

suppressed by using high steam/carbon (or S/C) ratios in the

feedstock [17,20,24,28], adding oxygen to the feed [20],

doping with a trace amount of sulphur, tin, chromium,

molybdenum, potassium, and gold on the catalyst [18,23–

25,29], adjusting reforming temperatures [30], enhancing

support (e.g., alumina, silica, magnesia, zirconia, and ceria

[25,31]), and adding sodium [32] etc. However, due to the

inherent complexity of various coupled physicochemical

phenomena involved in the reforming process, no general

fundamental understanding is currently available towards

the graphite formation and catalyst deactivation [12,21,26].

Bridging such a gap could germinate solutions to a longer

catalyst lifetime, which will lead to significant economical

benefits in the reforming process. Thermodynamic

modeling has been a powerful tool in the analysis of

hydrogen generation and fuel cells [33–37] and may provide

insight into graphite-free operating windows for the AER

process. However, to the best knowledge of the author,

a generic thermodynamic model that accounts for gaseous

reaction, adsorption, and formation of condensed species is

not yet available. Motivated by this, a comprehensive ther-

modynamic model is developed in this work to calculate the

equilibrium compositions in both the gaseous and the

condensed phases of general reaction–adsorption systems.

The thermodynamic analysis is carried out for steam

reforming of ethanol under various operating conditions

and it shows that adsorption enhanced reforming (AER) is

advantageous over conventional reforming processes not

only in terms of hydrogen yield but also in graphite

suppression. Such an analysis might be used to guide the

design and operation of an AER process.

2. Equilibrium of reaction–adsorptionsystems with formation of condensed species

The aim of this work is to develop a comprehensive thermo-

dynamic model to analyze reaction–adsorption systems

coupled by formation of condensed species. In the context of

AER of hydrocarbons, the adsorbed species of interest is CO2,

and the condensed species is graphite [5,10,13,14,38,39]. Multi-

component adsorption and multi-component condensation

might also be included in the current model to handle other

condensed species and the adsorption of gaseous species

besides CO2.

The equilibrium algorithm is based on the minimization of

the total Gibbs free energy in an isothermal and isobaric

system. Different from the equilibrium constant method, the

Gibbs minimization approach does not require the specifica-

tion of a detailed reaction scheme. A list of reactants and

possible products and their thermodynamic properties suffice

for the equilibrium calculation. Interested readers may refer

to a previous work for a comparison of both approaches for

solving a generic adsorption–reaction system with no

condensed species [40]. The condensed species have different

forms in the Gibbs free energy and pose a mathematical

challenge to the optimization algorithm. In order to calculate

the Gibbs free energy, the concept of mole ratio between the

gaseous species and the adsorbed species [40] is employed

here. According to the Langmuir adsorption/desorption iso-

therm (and also the equilibrium chemisorption isotherms of

CO2 on hydrotalcites [14,35]), the adsorbed amount of a

species follows the relationship below:

nðadsÞj ¼ S

nðadsÞj;sat bjPj

1þXs

j¼1

bjPj

¼ SnðadsÞ

j;sat bjnj

nTP

1þXs

j¼1

bj

nj

nTP

ðj ¼ 1;.; sÞ (1)

0 10 20 30 40 50 600

10

20

30

40

50

60

Measured mole fraction (%)

Pred

icte

d m

ole

fract

ion

(%)

COCO2CH4H2H2O

Fig. 1 – Comparison of gas composition of steam reforming of ethanol solved using the current equilibrium code with

experimental data [42] under various conditions (S/C [ 2, 3, 4, P [ 2, 5, 9 bar, T [ 700 8C and no CO2 adsorption).

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 3 6 2 – 9 3 7 2 9363


where nj is the amount of species j in the gas phase ([mole]),

nj(ads) is the amount of species j on the adsorbent ([mole]), nj,

sat(ads) is the saturated surface concentration ([mole/m2]), S is

the total surface area of the adsorbent [m2], bj is a constant

for species j under isothermal conditions ([bar–1]), Pj is the

partial pressure of species j in the gas phase, and P is the

system pressure. nT ¼Ps

j¼1 nj, where s is the number of

gaseous species. For the adsorption of gaseous species on an

adsorbent, nj, sat(ads) and bj can be determined experimentally. S

is then derived as the product of the experimentally

measured specific surface area of the adsorbent and its

weight. From Eq. (1), the ratio of nj(ads) to nj is expressed as

follows:

rj ¼nðadsÞ

j

nj¼

SnðadsÞj;sat bjP

nT þXs

j¼1

bjnjP

ðj ¼ 1;.; sÞ (2)

Because the adsorbed species and its gaseous counterpart have

the same chemical potential under chemical equilibrium, the

optimization problem is formulated as follows:

minnj

G ¼Xs

j¼1

�1þ rj

�mjnj þ

Xt

j¼sþ1

mjnj

s:t:

0 ¼Xs

j¼1

�1þ rj

�aijnj þ

Xt

j¼sþ1

aijnj � bið0Þ ði ¼ 1;.; 3Þ

mj ¼

8<:

m0j þ RT ln

P

P0þ RT ln

nj

nTðj ¼ 1;.; sÞ

m0j ðj ¼ sþ 1;.; tÞ

nT ¼Xs

j¼1

nj

rj ¼SnðadsÞ

j;sat bjP

nT þXs

j¼1

bjnjP

ðj ¼ 1;.; sÞ

nj � 0 ðj ¼ 1;.; tÞ

(3)

where G is the Gibbs free energy, mj is the chemical potential

for species j, aij is the number of element i in species j, bi(0)

is the total number of element i in the feed mixture bið0Þ ¼Ptj¼1 aijnjð0Þ, where nj(0) is the amount of species j in the feed),

0 stands for standard conditions, and t is the total number of

species including the condensed material.

The known parameters are T, P, S and nj(0) and the

variables to be determined are rj ( j¼ 1,., s), nj ( j¼ 1,., t), li

(i¼ 1,., e) and nT. Due to the complexity of the problem

represented by Eq. 3, one could solve an equivalent

problem in which T, P, rj ( j¼ 1,., s) and nj(0) are known

while S, nj ( j¼ 1,., t), li (i¼ 1,., e) and nT are to be deter-

mined [40]. However, it is worth noting that the following

relationship:

ri

rj¼

nðadsÞi;sat bi

nðadsÞj;sat bj

(4)

should be satisfied when specifying rj ( j¼ 1,., s) if multiple

component adsorption occurs in the system.

It is also worth noting that inequity constraints nj� 0

( j¼ 1,., s) are handled in the solution procedure by solving ln

nj instead of nj [40,41]. To avoid singularity problems when

taking logarithms of extremely small positive numbers, any nj

that is less than 10�50 will be replaced by this number during

the iteration process.

The constrained optimization problem represented by Eq. (3)

can be solved by the Lagrange multiplier method. Based on this

method, an auxiliary function is first defined as follows:

f ¼Ps

j¼1 ð1þ rjÞmjnj þPt

j¼sþ1 mjnj þPe

i¼1 li

hPsj¼1ð1þ rjÞaijnjþPt

j¼sþ1 aijnj � bið0Þi, where li are the Lagrange multipliers for the

atomic balance constraints. As a result, the optimal solution to

the optimization problem of Eq. (3) is determined by solving the

following equations:

0 ¼�1þ rj

�"mj þ

Xe

i¼1

aijli

#ðj ¼ 1;.; sÞ

0 ¼ mj þXe

i¼1

aijli ðj ¼ sþ 1;.; tÞ

0 ¼Xs

j¼1

�1þ rj

�aijnj þ

Xt

j¼sþ1

aijnj � bið0Þ ði ¼ 1;.; 3Þ

mj ¼(

m0j þ RT ln

PP0þ RT ln

nj

nTðj ¼ 1;.; sÞ


nT ¼Xs

j¼1

nj

(5)

The decent Newton–Raphson method is used to solve the

above problems, and the iterative variables are chosen to be D

ln nj ( j¼ 1,., s), D ln nj ( j¼ sþ 1,., t), D ln nT and pi¼�li/RT

(i¼ 1,., e) which are different from the NASA Chemical

Equilibrium with Applications (CEA) code [41]. The following

iterative formulas are then derived:

0 ¼ �mj

RTþXe

i¼1

aijpi þ D ln nT � D ln nj ðj ¼ 1;.; sÞ

0 ¼Xe

i¼1

aijpi �mj

RTðj ¼ sþ 1;.; tÞ

0 ¼Xs

j¼1

�1þ rj

�aijnj þ

Xt

j¼sþ1

aijnj � bið0Þ

þXs

j¼1

�1þ rj

�aijnjD ln nj þ

Xt

j¼sþ1

aijnjD ln nj ði ¼ 1;.; 3Þ

0 ¼Xs

j¼1

nj � nT þXs

j¼1

njD ln nj � nTD ln nT

mj ¼(

m0j þ RT ln

PP0þ RT ln

nj

nTðj ¼ 1;.; sÞ


(6)

The ‘‘reduced Gibbs iteration’’ scheme [41] is employed to

effectively track a large number of species. The idea is to

cancel D ln nj in Eq. (6). Consequently, the following equations

are derived:



Xe

k¼1

�Xs

j¼1

�1þ rj

�akjaijnj

�pk þ

Xt

j¼sþ1

aijnjDln nj

þ�Xs

j¼1

�1þ rj

�aijnj

�DlnnT ¼

�bið0Þ �

Xs

j¼1

�1þ rj

�aijnj

�Xt

j¼sþ1

aijnj þXs

j¼1

�1þ rj

�aijnj

mj

RT

�ði ¼ 1;.; eÞ

Xe

k¼1

akjpk ¼mj

RTðj ¼ sþ 1;.; tÞ

Xe

k¼1

�Xs

j¼1

akjnj

�pk þ

�Xs

j¼1

nj � nT

�Dln nT

¼�nT �

Xs

j¼1

nj þXs

j¼1

nj

mj

RT

�

mj ¼(

m0j þ RT ln

PP0þ RT ln

nj

nTðj ¼ 1;.; sÞ


(7)

Eq. (7) is an eþ t–sþ 1 dimensional linear equation which can

be written in a matrix form. The variables to be determined at

each iteration step are pk (k¼ 1,., e), D ln nj ( j¼ sþ 1,., t) and

nT. During the iteration process, if any column vector has

a norm close to zero such that the matrix becomes nearly

singular, the associated variable will be temporarily removed

from Eq. (7). The solution procedure is described as follows.

With initial guesses of nj ( j¼ 1,., t) and nT, Eq. (7) is solved to

determine D ln nj ( j¼ sþ 1,., t) and D ln nT. D ln nj ( j¼ 1,., s) is

then determined using the first equation in Eq. (6). Subse-

quently, nj ( j¼ 1,., t) and nT are updated. To avoid divergence

of the algorithm, the step size used for updating nj ( j¼ sþ 1,.,

t) is smaller than nj ( j¼ 1,., s) and nT. The iterative procedure

continues until the changes in all the correction variables in

Eq. (7) fall into specified regions [40]. The thermodynamic data

of each species at temperature T, such as the heat capacity,

enthalpy, entropy and chemical potential are calculated as

functions of temperature [41].

3. Thermodynamic analysis of the effect ofCO2 adsorption on steam reforming of ethanol

A reactive system consisting of steam and ethanol is studied

in this section. The steam reforming of biomass-generated

ethanol would be a potentially viable and renewable process

for hydrogen production [2–4]. Experimental studies have

shown that for a system of hydrocarbon (e.g., methane,

methanol or ethanol) and steam, the yield and purity of

hydrogen can be significantly improved if CO2 is adsorbed [5–

10]. Calculations in this section will provide a quantitative

thermodynamic analysis of the AER process.

In the steam reforming of ethanol, the most common

products and by-products would be H2, CH4, CO, CO2 and C

300400

500600

0.51

1.52

2.50

1

2

3

Temperature (oC)S/C ratio

H2

a

300400

500600

0.51

1.52

2.50

1

2

3

4

5


H2

b

300400

500600

0.51

1.52

2.50

0.2

0.4

0.6

0.8


Gra

phite

c

300400

500600

0.51

1.52

2.50

0.2

0.4

0.6

0.8

Gra

phite


d

Fig.2 – Profiles ofH2 yield andgraphite formation insteam reforming ofethanol (a, c) withoutand (b, d)withCO2 adsorption (P [ 5

bar, rCO2 [102 for CO2 adsorption). Profiles of H2 and graphite are normalized based on the molar amount of ethanol in the feed.



(graphite). An inclusion of more species would not affect the

complexity of this problem because the developed algorithm

is based on the reduced Gibbs iteration. One possible reaction

scheme, specified based on the minimum number of inde-

pendent reactions, is listed as follows:

C2H5OH ¼ CH4þCOþH2 (8a)

CH4þH2O ¼ COþ 3H2 (8b)

COþH2O ¼ CO2þH2 (8c)

2CO ¼ CO2þC (8d)

As mentioned earlier, these reactions are not necessary in the

calculation of equilibrium composition using the Gibbs mini-

mization approach. However, they are helpful for the expla-

nation of gas composition change as operating conditions vary.

The developed algorithm is first applied to a conventional

steam reforming of ethanol process with no CO2 adsorption

300400

500600

0.51

1.52

2.5

0.8

1

1.2

1.4

1.6


CH

4

a

300400

500600

0.51

1.52

2.50

0.5

1

1.5


CH

4

b

300400

500600

0.51

1.52

2.50

0.1

0.2

0.3

0.4


CO

c

300400

500600

0.51

1.52

2.50

0.005

0.01

0.015

0.02


CO

d

300400

500600

0.51

1.52

2.50.2

0.4

0.6

0.8

1


CO

2

e

300400

500600

0.51

1.52

2.50.005

0.01

0.015

0.02


CO

2

f

Fig. 3 – Profiles of CH4, CO and CO2 in steam reforming of ethanol (a, c, e) without and (b, d, f) with CO2 adsorption (P [ 5 bar,

rCO2 [ 102 for CO2 adsorption). Profiles of CH4, CO and CO2 are normalized based on the molar amount of ethanol in the feed.



(i.e., rCO2 ¼ 0) [42]. In this process, the reforming occurs in

a tube reactor where Ni was used as the catalyst. All carbon

containing gases in the product gas mixture were analyzed

using a NDIR spectrometer. A comparison of model predicted

gas composition in the steam reforming of ethanol using Ni as

catalyst with experimental data under various pressures and

steam/carbon ratios is given in Fig. 1. The model prediction

matches the experimental data very well, which partly

validates the accuracy of the developed thermodynamic

model.

The equilibrium analysis is then made for steam reforming

of ethanol with/without CO2 adsorption at various S/C ratios,

temperatures and a fixed pressure of 5 bar. In the case of CO2

adsorption, rCO2 is set to be 102. It is shown in Fig. 2 that in both

cases, the yield of hydrogen increases as the S/C ratio or the

temperature increases. When CO2 is partly removed form the

system by adsorption, the hydrogen yield is improved (by 8–

178%). The enhancement in hydrogen yield is more prominent

at low temperatures and S/C ratios higher than 1.3. Moreover,

while the graphite formation is very severe when the S/C is

below 1.3 when no CO2 is adsorbed, it reduces to a minimal

level when adequate CO2 adsorption is taken into effect.

The profiles of CH4, CO and CO2 with/without CO2

adsorption are shown in Fig. 3. It is worth noting that

rCO2 ¼ 102 means the amount of CO2 in the gaseous phase is

1% of the one adsorbed on the surface of the adsorbent. If one

compares the CO2 in the gaseous phase in Fig. 3(e) (without

adsorption) and (f) (with adsorption), the latter is not exactly

1% but between 1% and 2% of the former. This is because the

entire reactive system behaves like a buffer solution [40].

When CO2 is partly adsorbed on the surface of adsorbent, it is

generated by other reaction mechanisms, e.g., the water gas

shift reaction (Eq. 8c). Because the CO2 generation partly

compensates for the removal in the gaseous phase due to

adsorption, the total amount of CO2 (in both the gaseous

phase and on the surface) is more in the case of AER. This also

explains why the CO concentration in the gas phase is reduced

to a minimal level when substantial CO2 is removed by

adsorption (The CO in Fig. 3(d) is only 2–10% of the one in

Fig. 3(b)). The reduction in CO2 and CO favors the water gas

shift reaction (Eq. 8c) and the methane reforming reaction (Eq.

8b), leading to an enhancement in the hydrogen yield. Even

though the reduction in CO2 is more than the one in CO when

CO2 is substantially adsorbed, the carbon formation reaction

(Eq. 8d) moves backwards because its equilibrium constant is

related to the square of CO concentration.

The steam reforming of ethanol with/without CO2

adsorption at a lower pressure (P¼ 1 bar) is shown in Figs. 4

300400

500600

0.51

1.52

2.50

1

2

3

4

5


H2

a

300400

500600

0.51

1.52

2.50

2

4

6


H2

b

300400

500600

0.51

1.52

2.50

0.2

0.4

0.6

0.8


Gra

phite

c

300400

500600

0.51

1.52

2.50

0.1

0.2

0.3

0.4


Gra

phite

d

Fig. 4 – Profiles of H2 yield and graphite formation in steam reforming of ethanol (a, c) without and (b, d) with CO2 adsorption

(P [ 1 bar, rCO2 [102 for CO2 adsorption). Profiles of H2 and graphite are normalized based on the molar amount of ethanol in

the feed.



and 5. It is observed that without adsorption, the hydrogen

yield at chemical equilibrium at 1 bar is higher than the one at

5 bar. This can be explained using Le Chatelier’s principle

which suggests that the generation of light gaseous molecules

is favored at reduced pressures. However, the amount of

graphite formed is also higher at 1 bar. Under this low pres-

sure, both the hydrogen yield and graphite suppression are

significantly improved when AER is used. It is noticed that for

rCO2 ¼ 102, the graphite formation is still obvious at higher

temperatures (500–600�C) when the S/C ratio is low, which

might be improved by increasing rCO2 , or the amount of

adsorbent.

A further study of the reforming pressure and adsorption on

hydrogen yield and graphite formation is shown in Fig. 6. It

clearly shows the evolution of hydrogen yield (Fig. 6(a)) and

carbon formation (Fig. 6(b)) as the pressure and rCO2 vary

gradually. In the entire range of interest, the yield of hydrogen

increases when the pressure is reduced. However, the graphite

formation is also promoted. A combination of hydrogen

enhancement and graphite suppression can be achieved when

300400

500600

0.51

1.52

2.50

0.5

1

1.5


CH

4

a

300400

500600

0.51

1.52

2.50

0.5

1

1.5


CH

4

b

300400

500600

0.51

1.52

2.50

0.2

0.4

0.6

0.8


CO

c

300400

500600

0.51

1.52

2.50

0.02

0.04

0.06


CO

d

300400

500600

0.51

1.52

2.50.2

0.4

0.6

0.8

1

1.2


CO

2

e

300400

500600

0.51

1.52

2.50.005

0.01

0.015

0.02


CO

2

f

Fig. 5 – Profiles of CH4, CO and CO2 in steam reforming of ethanol (a, c, e) without and (b, d, f) with CO2 adsorption (P [ 1 bar,

rCO2 [102 for CO2 adsorption). Profiles of CH4, CO and CO2 are normalized based on the molar amount of ethanol in the feed.



CO2 is removed from the reactive system by adsorption. Based

on Fig. 6(b), a contour of graphite formation is plotted in

Fig. 6(c), which reveals that a graphite-free region is guaranteed

by sufficient CO2 adsorption. Based on this thermodynamic

analysis, CO2 adsorption could be an alternative or supplement

to the high S/C ratios or high pressures used in the reforming

processes to inhibit carbon formation. The required amount of

adsorbent per mol of ethanol in the feed at specified rCO2 and P

can be calculated based on Eq. (2), or

SnðadsÞCO2 ;sat ¼

rCO2

�Psj¼1 nj þ bCO2

nCO2P�

bCO2P

(9)

Because nðadsÞCO2 ;sat is a physical parameter of the adsorbent that is

measured experimentally, the surface area of the adsorbent S

can be readily determined if SnðadsÞCO2 ;sat is known. The group

SnðadsÞCO2 ;sat based on potassium promoted hydrotalcite (bCO2

is

around 22 bar–1 at 500�C [43]) is shown in Fig. 6(d). It is seen

that the amount of adsorbent reduces with increased pressure

if rCO2and the temperature are fixed. This is primarily due to

an increased partial pressure of CO2 in the gaseous phase,

which favors CO2 adsorption on the adsorbent.

Partial oxidation reforming is a method to heat up the

reforming gas by burning part of the fuel using oxygen, thus

reducing the energy demand for the reforming process [36].

The AER can be potentially combined with partial oxidation.

When oxygen is present in the feed, one more reaction should

be included in the reaction scheme described by Eq. (8). The

option is not unique. One acceptable reaction independent of

all the reactions in Eq. (8) is as follows:

2COþO2 ¼ 2CO2 (10)

The effect of O2 and adsorption on hydrogen yield and

graphite formation is shown in Fig. 7. It can be seen that

without adsorption, partial oxidation does not lead to

significant improvement in the hydrogen yield under the

given operating conditions. In fact, at high S/C ratios, the

thermodynamic model shows that partial oxidation might

result in a reduction in the hydrogen yield (not shown in

Fig. 7). However, when partial oxidation is combined with

CO2 adsorption, the hydrogen yield is enhanced. Moreover,

partial oxidation suppresses graphite formation. This is

because the addition of O2 in the system converts part of CO

into CO2 (see Eq. 10), and the carbon formation reaction (Eq.

0

5

10

0

50

1000.5

1

1.5

2

2.5

Pressure (bar)rCO2

H2

a

0

5

10

10−2

100

1020

0.1

0.2

0.3

0.4

Pressure (bar)rCO2

Gra

phite

b

0.05

0.05

0.050.1

0.1

0.150.2

Pressure (bar)

r CO

2

Grap

hite

supp

ress

ed

→

1 2 3 4 5 6 7 8 9 100

0.5

1

1.5c

0

5

10

0

50

1000

5

10

15

20

Pressure (bar)rCO2

SnC

O2,s

at(a

ds)

d

Fig. 6 – Profiles of (a) hydrogen yield and (b) graphite formation and (d) SnðadsÞCO2 ;sat and (c) contour of graphite formation as

functions of system pressure and CO2 adsorption (P [ 1–10 bar, T [ 500 8C, rCO2 [10L2–102, S/C [ 1 and bCO2[ 22 barL1

based on potassium promoted hydrotalcite [43]). The profiles of hydrogen, graphite and SnðadsÞCO2 ;sat are normalized based on

the molar amount of ethanol in the feed.



8d) moves backwards. When the partial oxidation is used in

the AER process, the graphite formation can be inhibited

with a less amount of adsorbent (see Fig. 6 for

a comparison).

4. Concluding remarks

Thermodynamic analyses have been carried out for AER of

ethanol. It is shown that adsorption of CO2 suppresses the

formation of carbon in the reforming process. When CO2

adsorption is combined with high S/C ratios, increased pres-

sures and partial oxidization, the formation of carbon can be

further reduced. Because adsorption also enhances hydrogen

yield and purity, AER would be an efficient way of making

hydrogen at a reduced temperature as compared to conven-

tional steam reforming.

It is worth noting that even though thermodynamic anal-

ysis can be used to explain certain phenomena, it only

provides the limit of the reactions. It may only reflect the

general trend of a reactor but is not suited for the detailed

modeling of a PFR-type reactor (see Ref. [40] for an effort of

thermodynamic modeling of gradual CO2 removal along

a reactor that is approximated by multiple equilibrium stages).

Due to the continuous operation nature of the AER process,

a spatiotemporal description of the reactor requires time-

dependent distributed kinetic models (including the kinetics

of reactions and of CO2 adsorption–desorption as well)

coupled by fluid flow and heat transfer. In this way the CO2

breakthrough curve can be captured. Interested readers may

refer to literature efforts in this direction [10,13].

Acknowledgements

This work is supported by the American Chemical Society

Petroleum Research Fund (PRF# 50095-UR5) and the Provost’s

Teacher-Scholar program at Cal Poly Pomona. Industrial

feedback from Dr. Durai Swamy at Intelligent Energy, Long

Beach, CA is gratefully acknowledged. The author would also

0

0.5

1

0

50

1000.5

1

1.5

2

O2rCO

2

H2

a

0

0.5

1

10−2

100

1020

0.05

0.1

0.15

0.2

O2rCO

2

Gra

phite

b

0.01

0.02

0.03

0.04

0.050.06

0.070.08

O2

r CO

2

Graphite suppresse

d

→

0 0.2 0.4 0.6 0.8 10

0.1

0.2

0.3

0.4c

0

0.5

1

0

50

1000

2

4

6

O2rCO

2

SnC

O2,s

at(a

ds)

d

Fig. 7 – Profiles of (a) hydrogen yield and (b) graphite formation and (d) SnðadsÞCO2 ;sat and (c) contour of graphite formation as

functions of O2/ethanol in the feed and CO2 adsorption (O2/ethanol [ 0–1, P [ 5 bar, T [ 500 8C, rCO2 [10L2–102, S/C [ 1 and

bCO2[ 22 barL1 based on potassium promoted hydrotalcite [43]). Hydrogen, oxygen, graphite and SnðadsÞ

CO2 ;sat are normalized

based on the molar amount of ethanol in the feed.



like to thank the anonymous reviewers for their valuable

comments and suggestions.

Nomenclature

bj adsorption parameter of species j in the Langmuir

isotherm model, bar–1

e total number of chemical elements in the reactive

system

l Lagrange multiplier

m chemical potential, J/mol

p –l/RT(ads) properties of adsorbed species

aij number of chemical element i in species j

G Gibbs energy, J

i, j, k indices

nj molar amount of species j in the gas phase, mol

nj(ads) molar amount of species j on the surface, mol

nT total molar amount in the gas phase, mol

nj, sat(ads) saturated concentration of species j on the

adsorbent surface, mol/m2

P pressure, bar

R gas constant, 8.314 J/mol/K

ri molar ratio of a species in the gas phase to the

surface

S surface area of adsorbent, m2

s total number of gaseous species

T temperature, K

t total number of species

r e f e r e n c e s

[1] Lattin WC, Utgikar VP. Transition to hydrogen economy inthe united states: a 2006 status report. Int J Hydrogen Energy2007;32:3230–7.

[2] Arteaga LE, Peralta LM, Kafarov V, Casas Y, Gonzales E.Bioethanol steam reforming for ecological syngas andelectricity production using a fuel cell SOFC system. ChemEng J 2008;136:256–66.

[3] Deluga GA, Salge JR, Schmidt LD, Verykios XE. Renewablehydrogen from ethanol by autothermal reforming. Science2004;303:993–7.

[4] Florin NH, Harris AT. Enhanced hydrogen production frombiomass with in situ carbon dioxide capture using calciumoxide sorbents. Chem Eng Sci 2008;63:287–316.

[5] Hufton JR, Mayorga S, Sircar S. Sorption-enhancedreaction process for hydrogen production. AIChE J 1999;45:248–56.

[6] Koumpouras GC, Alpay E, Stepanek F. Mathematicalmodelling of low-temperature hydrogen production with insitu CO2 capture. Chem Eng Sci 2007;62:2833–41.

[7] Ortiz AL, Harrison DP. Hydrogen production using sorption-enhanced reaction. Ind Eng Chem Res 2001;40:5102–9.

[8] Stevens JF, Krishnamurthy B, Atanassova P, Spilker K.Development of 50 kW fuel processor for stationary fuel cellapplications. DOE Technical Report DOE/GO/13102–1.Chevron Techology Ventures, LLC; 2007.

[9] Wang YN, Rodrigues AE. Hydrogen production from steammethane reforming coupled with in situ CO2 capture:conceptual parametric study. Fuel 2005;84:1778–89.

[10] Xiu G, Li P, Rodrigues AE. New generalized strategy forimproving sorption-enhanced reaction process. Chem EngSci 2003;58:3425–37.

[11] Tiemersma TP, Patil CS, van Sint Annaland M, Kuipers JAM.Modelling of packed bed membrane reactors for autothermalproduction of ultrapure hydrogen. Chem Eng Sci 2006;61:1602–16.

[12] Chen Z, Yan Y, Elnashaie SS. Catalyst deactivation andengineering control for steam reforming of higherhydrocarbons in a novel membrane reformer. Chem Eng Sci2004;59:1965–78.

[13] Ding Y, Alpay E. Adsorption-enhanced steam-methanereforming. Chem Eng Sci 2000;55:3929–40.

[14] Lee KB, Beaver MG, Caram HS, Sircar S. Novel thermal-swingsorption-enhanced reaction process concept for hydrogenproduction by low-temperature steam-methane reforming.Ind Eng Chem Res 2007;46:5003–14.

[15] Engwall E, Saukaitis J, Joshi M, Matzakos A, Paggio AD.Hydrogen production with a membrane steam reformer. In:AIChE Annual Meeting, paper 479c, Salt Lake City, Utah;2007.

[16] Duraiswamy K, Mauzey JL, Pena DA, and Chellappa AS. Low-cost high-efficiency distributed hydrogen production. In:AIChE Spring National Meeting, paper 168b. New Orleans, LA;2008.

[17] Bartholomew CH. Mechanisms of catalyst deactivation. ApplCatal A 2001;212:17–60.

[18] Bengaarda HS, Norskova JK, Sehested J, Clausenb BS,Nielsenb LP, Molenbroekb AM, et al. Steam reforming andgraphite formation on Ni catalysts. J Catal 2002;209:365–84.

[19] Batista MS, Santos RKS, Assaf EM, Assaf JM, Ticianelli EA.Characterization of the activity and stability of supportedcobalt catalysts for the steam reforming of ethanol. J PowerSources 2003;124:99–103.

[20] Klouz V, Fierro V, Denton P, Katz H, Lisse JP, Bouvot-Mauduit S, et al. Ethanol reforming for hydrogen productionin a hybrid electric vehicle: process optimisation. J PowerSources 2002;105:26–34.

[21] Ren X-H, Bertmer M, Stapf S, Demco DE, Blumich B, Kern C,et al. Deactivation and regeneration of a naphtha reformingcatalyst. Appl Catal A 2002;228:39–52.

[22] Snoeck JW, Froment GF, Fowles M. Filamentous carbonformation and gasification: thermodynamics, driving force,nucleation, and steady-state growth. J Catal 1997;169:240–9.

[23] Snoeck JW, Froment GF, Fowles M. Steam/CO2 reforming ofmethane. carbon formation and gasification on catalystswith various potassium contents. Ind Eng Chem Res 2002;41:3548–56.

[24] Trimm DL. Coke formation and minimisation during steamreforming reactions. Catal Today 1997;37:233–8.

[25] Tang SB, Qiu FL, Lu SJ. Effect of supports on the carbondeposition of nickel catalysts for methane reforming withCO2. Catal Today 1995;24:253–5.

[26] Trimm DL. Catalysts for the control of coking during steamreforming. Catal Today 1999;49:3–10.

[27] Tomishige K, Chen YG, Fujimoto K. Studies on carbondeposition in CO2 reforming of CH4 over nickel–magnesiasolid solution catalysts. J Catal 1999;181:91–103.

[28] Mathure PV, Ganguly S, Patwardhan AV, Saha RK. Steamreforming of ethanol using a commercial nickel-basedcatalyst. Ind Eng Chem Res 2007;46:8471–9.

[29] Subramani V, Gangwal SK. A review of recent literature tosearch for an efficient catalytic process for the conversion ofsyngas to ethanol. Energy Fuels 2008;22:814–39.

[30] Gronchi P, Fumagalli D, Rosso RD, Centola P. Carbondeposition in methane reforming with carbon dioxide.J Therm Anal 1996;47:227–34.



[31] Duan S, Senkan S. Catalytic conversion of ethanol tohydrogen using combinatorial methods. Ind Eng Chem Res2005;44:6381–6.

[32] Llorca J, Homs N, Sales J, Fierro JLG, de la Piscina PR. Effect ofsodium addition on the performance of Co–ZnO-based catalystsfor hydrogen productionfrombioethanol. JCatal2004;222:470–80.

[33] Assabumrungrat S, Laosiripojana N, Pavarajarn V,Sangtongkitcharoen W, Tangjitmatee A, Praserthdam P.Thermodynamic analysis of carbon formation in a solidoxide fuel cell with a direct internal reformer fuelled bymethanol. J Power Sources 2005;139:55–60.

[34] Colpan CO, Dincer I, Hamdullahpur F. Thermodynamicmodeling of direct internal reforming solid oxide fuel cellsoperating with syngas. Int J Hydrogen Energy 2007;32:787–95.

[35] Ding Y, Alpay E. Equilibria and kinetics of CO2 adsorption onhydrotalcite adsorbent. Chem Eng Sci 2000;55:3461–74.

[36] Lutz AE, Bradshaw RW, Bromberg L, Rabinovich A.Thermodynamic analysis of hydrogen production by partialoxidation reforming. Int J Hydrogen Energy 2004;29:809–16.

[37] Lutz AE, Bradshaw RW, Keller JO, Witmer DE.Thermodynamic analysis of hydrogen production by steamreforming. Int J Hydrogen Energy 2003;28:159–67.

[38] Yong Z, Mata VG, Rodrigues AE. Adsorption of carbondioxide on chemically modified high surface area carbon-based adsorbents at high temperature. Adsorption 2001;7:41–50.

[39] Lee KB, Beaver MG, Caram HS, Sircar S. Chemisorption ofcarbon dioxide on sodium oxide promoted alumina. AIChE J2007;53:2824–31.

[40] Li M. Equilibrium calculation of gaseous reactive systemswith simultaneous species adsorption. Ind Eng Chem Res2008;47:9263–71.

[41] Gordon S, McBride BJ. Computer program for calculation ofcomplex chemical equilibrium compositions andapplications. Technical Report NASA RP-1311. NASA LewisResearch Center; 1996.

[42] Rampe T, Hubner P, Vogel B, Heinzel A. Hydrogengeneration from ethanol by allothermal reforming. In:proceedings of the first world conference on biomass forenergy and industry. James & James (Science Publishers)Ltd.; 2001. p. 1889–92.

[43] Lee KB, Verdooren A, Caram HS, Sircar S. Chemisorption ofcarbon dioxide on potassium-carbonate-promotedhydrotalcite. J Colloid Interf Sci 2007;308:30–9.


author's personal copy - california state polytechnic ...minghengli/pdfs/thermodynamic analysis...

Documents