Thermodynamic Analysis and
Kinetics Model of H2S Sorption
Using Different Sorbents
X.M. Meng, W. De Jong, and A.H.M. Verkooijen
Faculty of Mechanical, Maritime and Materials Engineering, Process and Energy Department, Energy Technology Section,
Delft University of Technology, Leeghwaterstraat 44, 2628 CA, Delft, The Netherlands; X.Meng@tudelft.nl (for correspondence)
Published online 12 August 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ep.10386
Integrated gasification combined cycle (IGCC)
power systems are being advanced worldwide for gen-
erating electricity from various fuels due to their
high-energy efficiency, superior environmental per-
formance, and low cost electricity in comparison to
conventional power plants. Syngas produced from
gasification process, primarily consists of CO, CO2,
CH4, and H2, with certain content of contaminants,
like sulfur compounds. These need to be removed
prior to the utilization to avoid detrimental effects on
the system equipments. Hot gas desulfurization
(HGD) can improve the thermal efficiencies and
lower the capital costs of the IGCC system due to the
elimination of fuel gas cooling and associated heat
exchangers. For this reason, various sorbents have
been extensively studied to remove both inorganic
and organic sulfur-containing compounds in situ or
downstream of gasifiers. To efficiently evaluate the
performance of different sorbents and optimize the
desulfurization process, a kinetic and thermody-
namic analysis of the sorption reaction(s) for sulfur
removal from hot gases is necessary. In this work, do-
lomite, limestone, lime, and several metal oxides, like
CuO, ZnO, FeO, and MnO were chosen as desulfur-
ization sorbent candidates. Three simulated gases
based on gas compositions obtained from typical gas-
ifiers were used for the thermodynamic analysis of
sorption with various sorbents for H2S removal. The
H2S concentration in equilibrium was calculated
using the software package Factsage version 5.4.1
(GTT Technologies). The thermodynamic analysis
shows that the equilibrium H2S concentration associ-
ated with the sulfidation of different sorbent is a
Ó 2009 American Institute of Chemical Engineers
360 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
strong function of temperature compared with other
conditional parameters. Copper, zinc, and manga-
nese oxides have the most favorable thermodynamic
property of H2S removal when the temperature is
lower than 6508C. Compared with other sorbents, do-
lomite, limestone, and lime have fairly similar re-
moval abilities and exhibit quite different sorption
tendencies as a function of the temperature. Finally,
different sulfidation models were presented and the
kinetics of sulfidation with calcium-based sorbent by
using the unreacted shrinking core model with vari-
able effective diffusivity will be further examined and
covered in the future publications. Ó 2009 American
Institute of Chemical Engineers Environ Prog, 28: 360–371,
2009
Keywords: hot gas desulfurization, thermodynamic
analysis, sorbent, kinetics, sulfidation model
INTRODUCTION
Integrated gasification combined cycle (IGCC)
power systems offer the potential for high-cycle effi-
ciency and improved environmental performance
compared with conventional power generation sys-
tems [1]. In IGCC systems, the fuel gas is cleaned to
remove particulates and chemical contaminants to
protect downstream equipment and to minimize emis-
sions to the environment. The composition of a raw
fuel gas produced by gasification can vary over a
wide range, depending on the gasification process (re-
actor type, condition parameters) and the type of fuels
used [2–4]. Normally, the product gas consists of CO,
CO2, CH4, N2, and H2, with trace amounts of higher
hydrocarbons, inert gases present in the gasification
agent, various containments such as sulfur com-
pounds, particles, and alkaline. During combustion
and gasification, the formation of sulfur compounds
can be described by the following chemistry [5].
) fuel devolatilization :
Fuel
S þ heat ! H2 S þ COS þ . . . þ Char
S
(R-1)
) Char oxidation :
(R-2)
Char
S þ CO2 ! COS
Char
S þ H2 O ! H2 S (R-3)
Char
S þ O2 ! SO2 (R-4)
) Gas phase reactions :
 (R-5)
H2 S þ 3 2O2 ! SO2 þ H2 O
CO2 þ H2 S , COS þ H2 O (R-6)
H2 S þ CO , H2 þ COS (R-7)
COS þ H2 S , CS2 þ H2 O (R-8)
 thermodynamic calculations the metal oxides of Fe,
CS2 , C þ 2 xSx (R-9)
, Sulphuric oxide formation :
(R-10)
SO2 þ 0:5O2 , SO3
H2 O þ SO3 , H2 SO4 (R-11)
For gasification processes, the situation is more com-
plicated than combustion. Sulfur in the biomass feed-
stock, of which the content is often low, is converted
to mainly hydrogen sulfide (H2S), with some amounts
of carbonyl sulfide (COS), carbon disulphide (CS2), or
elemental sulfur Sx etc. during gasification via reaction
(R6, 7, 8, and 9) depending on the gasification condi-
tions. These compounds need to be removed prior to
the utilization to avoid detrimental effects on the sys-
tem equipments and infringement of emissions legisla-
tion. Hot gas desulfurization (HGD) can improve the
thermal efficiencies and lower the capital costs of the
IGCC system due to the elimination of fuel gas cooling
and associated heat exchangers and be accomplished
by using different sorbents. Bakker et al. [6] reported
that a good sorbent will allow for a deep desulfuriza-
tion to sub-ppm levels and has fast adsorption kinetics,
physically and chemically stable, and no evaporation
or sintering during regeneration. Although in a real
system, the sulfur capture efficiency by using different
sorbents depends on various kinetic factors, equilib-
rium calculations are considered as an advantageous
Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
tool to qualitatively evaluate the performance of dif-
ferent sorbent materials. Because the majority of sulfur
in the fuel is converted into H2S during gasification in
accordance with van der Drift et al. [7], in this work,
only the thermodynamic calculations of H2S sorption
were conducted.
The main objectives of this study are threefold: (1)
to use thermodynamic equilibrium analyses to iden-
tify the effective sorbent materials, which can chemi-
cally bind sulfur to reduce its emissions; (2) to deter-
mine the most effective sorbent among the various
materials; and (3) to summarize different sulfidation
kinetic models. Dolomite, limestone, lime, and sev-
eral metal oxides were chosen as desulfurization sor-
bent candidates. Three simulated gases based on gas
compositions obtained from typical gasifiers with dif-
ferent oxidizers were used for the thermodynamic
analysis of sorption using various sorbents for H2S
removal. The H2S equilibrium concentration, equilib-
rium products, and phase stability diagram for iron
and manganese oxide at various temperatures during
regeneration were conducted using software package
Factsage version 5.4.1 (GTT Technologies) based on
the minimization of the Gibbs energy.
DESULFURIZATION AND REGENERATION SIMULATION
Sorbents Selection
Many types of metal oxides can be used for HGD.
Westmoreland et al. [8, 9] reported that based on
Zn, Mn, Mo, V, Ca, Sr, Ba, Co, Cu, and W are feasible
for HGD. In the temperature range 300–8008C, desul-
furization performance (% of sulfur adsorbed from
the gas phase) of some metal oxides decreased in the
order MnO > CaO 5 ZnO > V2O3. Swisher et al. [10,
11] reported a comprehensive literature review of
metals and binary oxides as sorbents for removing
sulfur from coal-derived gases, where they included
the information on the historical background and
properties of metals and binary oxide compounds.
Gasper-Galvin et al. [12] recently gave a broad review
of metal oxide sorbent development for the desulfur-
ization of coal gases with particular focus on copper-
based sorbents. Elseviers et al. [13] performed ther-
modynamic equilibrium simulations to determine
suitable materials for intensive desulfurization of fuel
gases from an IGCC. The authors concluded that
Zinc-based materials are the most promising for high-
temperature intensive desulfurization, but cause evap-
oration and regenerability problems. Based on the
overview of literature, the sorbent materials consid-
ered in this work and a summary of equilibrium
products formed during sulfidation reaction are listed
in Table 1.
Main Reaction
Typically, the metal oxide sorbents are converted
to sulfides during a sulfur-loading stage under reduc-
ing hot gas conditions, subsequently regenerated
(completely or partially) after they have captured the
undesired contaminants by using steam, oxygen, or
October 2009 361
sulfide oxide. The general desulfurization reaction for Shell oxygen blown entrained flow gasifier and TU
H2S reacting with a metal oxide is given by R-12 and Delft oxygen-steam blown circulating fluidized bed
the general regeneration reactions for HGD are R13- gasifier (CFBG)] with different oxidizers were used
15. From R15, the regeneration reaction equilibrium for the thermodynamic analysis of sorption with vari-
strongly depends upon the oxygen and sulfide oxide ous sorbents for H2S removal. The gas composition
partial pressure. An unwanted side reaction is the for- of Texaco and Shell are cited from Hepworth et al. in
mation of sulfate via R16, which should be avoided the references [14, 15]. To find the product composi-
as the formed sulphate is inert with respect to the tion for a multicomponent reacting system for the
desulfurization and thus leads to a loss of material specified conditions, stoichiometric and nonstoichio-
activities. metric methods are broadly used. The former uses
equilibrium constants for all constituent reactions,
Mex Oy ðsÞ þ xH2 SðgÞ þ ðy
xÞH2 ðgÞ whereas the latter minimizes the Gibbs free energy
subject to mass balance and non-negativity con-
! xMeSðsÞ þ yH2 OðgÞ (R-12) straints, which is more flexible in handling multiple
feed streams of unclear or unknown molecular for-
MeSðsÞ þ H2 OðgÞ ! MeOðsÞ þ H2 SðgÞ (R-13) mulation, as in biomass gasification, and has been
studied and presented in many literatures [16–18]. In
this work, for the third simulated gases, the gas com-
ðxÞMeSðsÞ þ ðy=2ÞSO2 ðgÞ ! Mex Oy ðsÞ þ ðx þ y=2ÞSðgÞ positions were calculated using software package
Factsage version 5.4.1 where the input data based on
(R-14)
DDGS fuel gasified in CFBG TUD. An exhaustive ex-
planation of Factsage features is available in the man-
ðxÞMeSðsÞðx þ y=2ÞO2 ðgÞ ! Mex Oy ðsÞ þ xSO2 ðgÞ ual of the program itself [19]. About DDGS fuel
(R-15) (Dried Distillers Grains with Solubles), which is a
nonfermentable byproduct of ethanol production,
currently, is mainly fed to livestock. With the rapid
MeSðsÞ þ 2O2 ðgÞ ! MeSO4 ðsÞ (R-16) growth of the ethanol industry, the increasing supply
of DDGS may saturate the livestock feed market, thus
other useful applications are required. For this rea-
son, DDGS is chosen as one of biomass fuels in
Simulation Specification GreenSyngas Project and to investigate its potential
Because gasification gas composition can vary uses for gasification to produce high-quality gaseous
over a wide range depending on fuel type and gasifi- fuel.
cation operating conditions, three simulated gases Table 2 lists simulated gas composition in desulfur-
based on gas compositions obtained from typical gas- ization and regeneration (simply called Texaco gas,
ifiers [Texaco oxygen blown entrained flow gasifier, Shell gas, CFBG gas). Because sulfur concentration
varies significantly from percent levels in fossil fuels
Table 1. Sorbent considered in the desulfurization to ppmv levels in biomass fuels, calculations were
simulation and equilibrium products. not conducted for a specific application. For the
desulfurization simulation, the influence of tempera-
ture from 300 to 15008C on the equilibrium H2S con-
Reaction
centration at atmospheric pressure was studied for
product at
each case, using the simulated gases as input data.
Metal equilibrium
The amount of H2S used was set to be the basis, and
type Considered sorbent phase
the amount of sorbent used was the stoichiometric ra-
Ca CaCO3, CaO, CaCO3,MgCO3 CaS tio of possible H2S-sorbent products. For the regener-
Fe Fe3O4, FeO, Fe2O3 FeS ation simulation, spent desulfurization sorbent was
Mn MnO2, MnO, Mn3O4 MnS fed to an equilibrium reactor operating at 1 atm. For
Cu CuO Cu2S each case, the influence of temperature from 300 to
Zn ZnO ZnS 15008C on equilibrium product forms was studied as
a function of the oxygen partial pressure.

Table 2. Simulated gas compositions in desulfurization and regeneration [15].

Type CO CO2 H2 N2 H2O H2S COS O2

Desulfurization (vol %,wet)
SHELL 64 0.8 31.6 0.7 1.5 1.4 —
Texaco 39.1 12.25 30.13 0 17.52 1 —
CFBG 14.1 14.9 32.5 1.43 36.8 0.17 —
Regeneration As balance 5, 20, 40, and 100%

362 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep

Figure 1. H2S concentration as a function of
temperature for Texaco gas in equilibrium with
various sorbents. [Color figure can be viewed in the
online issue, which is available at www.interscience.
wiley.com.]
Figure 2. H2S equilibrium concentration as a function
of temperature for Shell gas with various sorbents.
[Color figure can be viewed in the online issue,
which is available at www.interscience.wiley.com.]

RESULTS AND DISCUSSION

Sulfidation Simulation Analysis

The sulfidation simulation results of for Texaco,
Shell, and CFBG gas are shown in Figures 1–3, respec-
tively. If the H2S concentration of 100 ppmv
is considered allowable [13, 14], Figures 1–3 show that
iron, manganese, zinc, and copper oxides are suitable
for desulfurization when the temperature is lower than
5008C for all three simulated gases. The desulfurization
capacity of calcium-based materials depends more on
the simulated gas compositions than other oxides.

Zinc Oxide
For Texaco, Shell, and CFBG gas, zinc oxide was
capable of reducing a H2S concentration lower than
50 ppmv up to 7008C. With increasing the tempera-
ture from 700 to 8008C, the H2S concentration
increased sharply from 52.9 to 293 ppmv, 42 to 263
ppmv, and 53 to 294 ppmv, respectively. The rapid
increase in H2S concentration is attributed to the zinc Figure 3. H2S equilibrium concentration as a function
vaporization at temperatures above 7008C. These of temperature for CFBG gas with various sorbents.
results fairly agree with reported values in the litera- [Color figure can be viewed in the online issue,
ture. Lew et al. [20] reported that the kinetics of the which is available at www.interscience.wiley.com.]
ZnO–H2S reaction is rapid and ZnO can reduce the
H2S concentration from 5000 ppmv to less than 10
ppmv. However, zinc sorbents cannot be used at tem-
peratures above 6008C due to the subsequent vapori- Manganese Oxides
zation of zinc, which results in a rapid decrease in the For Texaco, Shell, and CFBG gas, manganese
sulfur capture capacity [21–23]. Westmoreland et al. [9] oxides can keep H2S concentration lower than 80
reported that zinc is limited to a maximum tempera- ppmv up to 700, 1000, and 5008C. These findings fit
ture of 7008C because of the formation of zinc the reported result by Hepworth et al. [14]. The
vapor. Experimental observations in their laboratory authors studied sulfidation thermodynamics of differ-
have confirmed the formation of zinc vapor in similar ent manganese oxide by using Shell gas. The equilib-
atmospheres at temperatures in excess of 7008C. rium calculations show that the lower the tempera-

Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep October 2009 363
ture the lower the equilibrium H2S concentration.
Shell gas can theoretically be desulfurized with MnO
to IGCC specification at temperatures exceeding
9008C. Ben-Slimane et al. [15, 24, 25] studied the
desulfurization of hot coal-derived fuel gases with
Mn-based sorbents. The kinetics studies have shown
that manganese oxide is stable for HGD up to around
7008C with high sulfur capacity between 600 and
7008C. Westmoreland et al. [9] also reported that
high-fractional desulfurization of manganese oxides is
predicted for temperatures in excess of 10008C and
has a great desulfurization potential in the tempera-
ture of 600–7008C.

Copper Oxide
For Texaco, Shell, and CFBG gas, at the tempera-
ture of 7008C, the fairly low-H2S concentration of
76.2, 58.6, and 75.2 ppmv was achieved by using
copper oxide. Higher temperatures led to an increase
of H2S equilibrium concentration, whereas less rap-
idly compared to zinc oxide. The predicted results
are similar to reported values in the literature. West-
moreland et al. [9] reported that copper can maintain
95% desulfurization capability to a temperature in
excess of 9008C. Abbasian et al. [26] pointed out that
a fortuitous feature of a copper-based sorbent is its
capability to achieve extremely low levels of H2S in
the cleaned fuel gas. The optimum desulfurization
temperature in terms of sorbent efficiency and utiliza-
tion for HGD is around 6008C. The increase of H2S
concentration at higher temperature is mainly due to
high temperatures favoring the complete reduction of
copper compounds even in slightly reducing atmos-
pheres. Swisher et al. [10] reported that for Cu-based
sorbents, the kinetics of the sulfidation reaction is fast
for Cu in the 12 or 11 oxidation states. However,
CuOx is easily reduced to Cu, which has a slow
kinetics for H2S removal.
Iron Oxides
Iron oxide desulfurization potential is somewhat
lower compared to the other materials. For Texaco,
Shell, and CFBG gas, FeO and Fe2O3 had the similar
sulfur capture capacity as a function of the tempera-
ture. A lower H2S concentration of 10, 25.4, and 29.1
ppmv was achieved using FeO and Fe2O3 up to
5008C for Texaco and CFBG gas, and 6008C for Shell
gas. However, Fe3O4 showed less-desulfurization
capability at a temperature of lower than 6008C com-
pared to FeO and Fe2O3. For all the gases, FeO,
Fe2O3, and Fe3O4 had similar low-sulfur removal
capacity at a temperature higher than 6008. Gupta
et al. [27] reported that for iron oxides sorbent higher
temperatures (>5508C) lead to severe sorbent decrep-
itation due to excessive reduction and iron carbide
formation. The formation of solid carbon deposition
inside sorbent pores and on the surface quickly limits
further sulfidation. Focht et al. [28] mentioned that at
7008C Fe3O4 is easily reduced to FeO and the reduc-
tion to FeO has been shown to have a detrimental
effect on sulfidation. So, although iron oxide has a
high-sulfur capture capacity and possesses high reac-
Figure 4. H2S equilibrium concentration as a function
of temperature with different H2O content using CaO.
[Color figure can be viewed in the online issue,
which is available at www.interscience.wiley.com.]
tivity for H2S, its applicability is limited to measured
temperatures of above 6008C.
Calcium-Based Materials
Compared to other materials, dolomite, limestone,
and lime exhibited fairly similar sulfur removal capa-
bilities and had quite different sorption tendency as a
function of the temperature. For all the gases, cal-
cium-based materials had very low-desulfurization
capability when the temperature was lower than
9008C, while for other metal oxides high-desulfuriza-
tion efficiency was observed at same temperature.
The H2S equilibrium concentration for Texaco and
CFBG gas was observed to be much higher than
other metal oxides. Nevertheless, for Shell gas, with
very low CO2 and H2O content, the calcium-based
materials exhibited high-desulfurization efficiency at
higher temperature. The reason for that can be
explained as follows. For calcium-based materials,
the possible reactions involved in sulfidation and the
equilibrium constant are shown in equations (Ca-1–
6). The concentration of H2S can be reduced to suffi-
ciently low values according to the reactions Ca-1, 2,
5, and 6, which are determined by the reaction tem-
perature and the partial pressure of CO2 and H2O in
the gas. Similar results were also reported in the liter-
ature [9, 15, 29]. Westmoreland et al. [9] reported the
maximum desulfurization of CaCO3 occurs near
8808C, the temperature at which CaCO3 decomposes
to CaO. Fenouil et al. [29] also pointed out that cal-
cium oxide is an attractive material, because it is
inexpensive, readily available, and it is an effective
sorbent at temperatures in the range of 800–12008C.
Van der Ham et al. [30] reported for Shell gas that
the sulfidation temperature should be at a level of
815–8458C to achieve H2S concentration of 20 ppmv.
Furthermore, the effect of steam content on H2S equi-

364 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
Figure 5. Phase stability diagram for the Mn-S-O system at different temperature.

librium concentration is shown in Figure 4. So, for Regeneration Simulations Analysis

achieving a low-H2S equilibrium concentration of Considering that manganese and iron oxide
around 100 ppmv in the exit gas, the steam content have various states, phase stability diagrams of the
in the gas composition had better be lower than 5%. Fe

S

O and the Mn

S

O system at different
temperature is shown in Figures 5 and 6. For other
Direct sulfidation : CaCO3 þ H2 S , CaS þ CO2 þ H2 O metal oxides, the regeneration simulation results by
(Ca-1) using different volume content of oxygen are shown
in Figures 7–9.
CaCO3  MgO þ H2 S , CaS  MgO þ CO2 þ H2 O

) KSd ¼ PH2 O PCO2 PH2 S ¼ 1:01 3 108 expð
13426=T Þ Manganese Oxides
(Ca-2) According to Hepworth et al. [14] during the
regeneration of MnS with oxygen, the following reac-
tions that can occur in Eqs. Mn-1–4.
Calcination : CaCO3 , CaO þ CO2 (Ca-3) 
MnS þ 3 2O2 ! MnO þ SO2 (Mn-1)
CaCO3  MgO , CaO  MgO þ CO2
(Ca-4)  
) Kc ¼ PCO2 ¼ 4:14 3 1012 expð
20474=T Þ MnS þ 1 6O2 ! 1 3Mn3 O4 (Mn-2)

 
Sulfidation of calcined sorbents : Mn3 O4 þ 1 4O2 ! 3 2Mn2 O3 (Mn-3)
(Ca-5)
CaO þ H2 S , CaS þ H2 O
MnS þ O2 þ 2SO2 ! 2MnSO4 (Mn-4)
CaO  MgO þ H2 S , CaS  MgO þ H2 O
 (Ca-6) Equation Mn-1 is the most desirable reaction for the
) KSc ¼ PH2 O PH2 S ¼ 1:22 expð7208:3=T Þ regeneration. Further oxidation of manganese occurs

Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep October 2009 365
Figure 6. Phase stability diagram for the Fe-S-O system at different temperature.

Figure 8. CuO, Cu2O, and CuSO4 formation during

Figure 7. ZnO, Zn, and ZnSO4 formation during ZnO
CuO regeneration [O2 volume fraction in regenerator
regeneration [O2 volume fraction in regenerator feed
feed gas: (5, 40, and 100%)]. [Color figure can be
gas: (5, 40, and 100%)]. [Color figure can be viewed
in the online issue, which is available at www. viewed in the online issue, which is available at
www.interscience.wiley.com.]
interscience.wiley.com.]

based on Eqs. Mn-2 and 3. The sulfation of manga- cause the pellets to spall and crack. To prevent sulfa-
nese as shown in Eq. Mn-4 is especially undesirable, tion, the conditions under which it occurs must be
because the production of manganese sulfate will avoided. Figure 5 shows a phase stability diagram of

366 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep

Figure 9. CaO and CaSO4 formation during CaO
regeneration [O2 volume fraction in regenerator feed
gas: (5, 20, 60, and 100%)]. [Color figure can be
viewed in the online issue, which is available at
www.interscience.wiley.com.]
Mn

S

O at temperature of 600, 900, 1200, and
15008C. From Figure 5, we can see, Mn2O3 and
Mn3O4 were the most stable forms, whereas MnO
and MnO2 cannot co-exit. The regeneration should
be conducted at a temperature higher than 9008C
when regenerating with oxygen at ambient pressure
to prevent the formation of MnSO4. By lowering the
O2 and/or the SO2, partial pressure will result in
lower MnSO4 stability. By using a regeneration gas
with O2 volume content of 5, 40, and 100% at the
temperature of 900 and 15008C, the main species in
the regenerated products were Mn2O3, Mn2O3/
Mn3O4, Mn3O4, and MnO, Mn3O4, and Mn3O4,
respectively.
Iron Oxides
A phase-stability diagram of Fe

S

O system at dif-
ferent temperatures was also plotted to have a better
understanding of chemical changes occurring during
regeneration reaction. As is shown in Figure 6, FeS,
FeSO4, Fe2(SO4)3, Fe2O3, Fe3O4, and FeO different
forms may occur as the reaction products. FeSO4 and
Fe2(SO4)3 were only stable in a very limited area with a
fairly high-SO2 partial pressure. Thus during the regen-
eration process, increasing regeneration temperature
or lower O2 concentration is assumed to be the possi-
ble way to overcome sulfate formation. Figure 6 also
indicates that the partial pressure of O2 had less effect
than temperature on the regeneration process. At the
regeneration temperature of 900 and 12008C, Fe2O3
was be the main species in the regenerated sorbents,
which is also reported by Fuda et al. [31]. By using the
regeneration gas with 5 vol % O2 at the temperature of
15008C, Fe3O4 was being the main regenerated prod-
uct, whereas at 6008C, some Fe2(SO4)3 would be pro-
duced, which should be avoided in the operation.
Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
Zinc Oxide
Figure 7 shows the ZnSO4, ZnS, and ZnO forma-
tion as a function of temperature and O2 partial pres-
sure for the regeneration of pure ZnO. Stable ZnSO4
was produced when regeneration was carried out at
lower temperatures or using high-partial pressures of
oxygen. It can be seen that temperatures in excess of
7008C are required to avoid ZnSO4 formation and
that the minimum regeneration temperature strongly
increases with the partial pressure of oxygen. How-
ever, regeneration at higher temperatures leads to
severed excessive thermal sintering, which is reported
by Garcia et al. [23]. Also, from an economic point of
view, in order to obtain SO2 concentrations suitable
for H2SO4 production, the oxygen molar fraction in
the feed gas should be higher than 10%. This implies
that regeneration should be carried out at 8008C at
least, whereas desulfurization had better to be carried
out at much lower temperatures to avoid the volatili-
zation of zinc. Thus, implementation on an industrial
scale will yield a more complex unit and also the dif-
ferent operation temperature will result a loss of ther-
mal efficiency for the overall process.
Copper Oxide
The predicted optimum temperature of operation
for HGD by using copper oxide is around 8008C.
And, from Figure 8, we can see that the best regener-
ation temperature is also with the range of 600–
9008C, which favorably improves thermal efficiency
for the overall process. At 9008C, the main regener-
ated product was CuO even with different O2 volume
content, while at a temperature higher than 10008C,
Cu2O and CuO were both found in the regenerated
sorbents, and Cu2O gradually changed to the main
product at a temperature higher than 12008C. During
the regeneration process, Cu in Cu-based sorbents
undergoes crystal growth and thermal sintering,
decreasing the sulfur capture capacity. Therefore, in
order to increase the thermal stability and to maintain
copper in the 12 or 11 oxidation states, and to keep
the copper oxide in a highly dispersed state, Alonso
et al. [32] suggested that copper-based sorbents are
mixed with other metal oxides.
Calcium-Based Materials
Dolomite, limestone, and lime calcium-based mate-
rials have been widely used as in bed as nonregener-
able desulfurization sorbent for decades due to their
low cost and naturally availability. However, the large
quantities of spent sorbents produced with limestone
and dolomite contain calcium sulfide, which is not
environmentally stable and requires a further stabili-
zation step to convert calcium sulfide to environmen-
tally acceptable material for landfill disposal. Because
of this reason, the regeneration of calcium-based
materials is also currently considered.
From Figure 9, we can see that the regeneration of
CaS need to be performed at temperatures higher
than 11008C, because CaSO4 is the main product at
lower temperature. A similar result was also pre-
sented by Van der Ham et al. [30], who reported that
October 2009 367
for thermodynamic reasons the regeneration tempera-
ture of calcium-based sorbent of at least 1100–12008C
should be applied. Considering the different tempera-
ture between sulfidation and regeneration and also
the overall thermal efficiency and cost, the regenera-
tion of calcium-based materials still need to be further
taken into account.
From sulfidation and regeneration simulations, it
can be concluded that for a deep desulfurization,
copper and zinc oxides with their superior properties
are the most favorable metals, capable of reaching
even ppb level at appropriately temperature. How-
ever, evaporation of metal, loss of the surface area
and porosity, sintering, and mechanical disintegra-
tions affect their performance and life time adversely,
especially at higher temperatures of 7008C. Manga-
nese oxide also seems to be a most promising
material as a result of its resistance to reduction, vola-
tilization, and carbide formation. However, during
regeneration with oxygen, undesirable excessive tem-
perature increase is a problem. For high-temperature
desulfurization, calcium-based materials exhibit a bet-
ter potential than other metal oxides; however, the
desulfurization capability of calcium-based materials
is strongly limited by the temperature range and gas
composition especially the H2O and CO2 content. So,
the development of a low cost, highly efficient, rap-
idly reacting and regenerable sorbent, with high sul-
fur capture capacity and durability still eludes the art.
For solving the volatilization, sintering, and regenera-
tion problems, supporting the oxides on a carrier ma-
terial or also through the use of mixed metal oxides
is considered as a good approach for all the singe
metal oxide sorbent. The support materials, such as
alumina (Al2O3), titania (TiO2), silica (SiO2), and
chromia (Cr2O3), are added to create and maintain
the desirable sorbent structure. Over past decades,
the extensive research work has been dedicated to
this field [12, 14, 15, 20, 24–28, 33–38], such as zinc-
based materials: zinc ferrites (ZnFe2O4) zinc–copper
ferrite, and zinc titanates (ZnTiO3, Zn2TiO4, and
Zn3TiO8); copper-based materials: copper aluminate
(CuO/Al2O3), copper ferrite aluminate (CuO/Fe2O3/
Al2O3), copper ferrite (CuO/Fe2O3), copper manga-
nate (CuO/MnO2), copper chromite (CuO/Cr2O3);
manganese-based materials: MnO/g-Al2O3. Here, no
more detailed information will be presented.
Sulfidation Kinetics Models
The above thermodynamic calculation exhibited us
a clear view toward the potential evaluation of differ-
ent metal oxides as sulfur sorbent with respect to
their desulfurization and regeneration. However, only
satisfying with the established thermodynamic criteria
is insufficient to determine the desulfurization feasi-
bility. Kinetic analysis and experimental study should
also be considered following the thermodynamic
analysis, because the sorbent utilization depends crit-
ically on the sulfidation kinetics and physical proper-
ties of the sorbent. With this aim, different mathemat-
ical kinetics models are developed to modeling and
quantifying the sulfidation by using different sorb-
368 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
ents. The basic model was proposed by Yagi et al.
and Levenspiel et al. [39, 40], which is called
unreacted shrinking core model (USCM). By taking
into account the solid structural properties (the spe-
cific surface area), specifically porosity and grain ra-
dius, grain model (GM), and pore model (PM) have
been gradually developed [41–43]. The investigation
of sulfidation by using different kinetics models has
been undertaken sporadically. Zevenhoven and co-
workers [44, 45] studied H2S capture by limestone
and dolomite under high pressure by using USCM
with variable effective diffusivity, which accounts for
all diffusion effects of the gaseous species inside the
solid reactant. Iisa and co-workers [46–48] studied the
sulphation of uncalcined calcium carbonate in a pres-
surized thermobalance using USCM to determine the
rate limiting process of the reaction. They found that
all particle size of 125–180 lm was limited by diffu-
sion through CaSO4 layer in the temperature range of
750–9008C and the activation energies of effective dif-
fusivity through the CaSO4 layer were 110–280 kJ/
mol, whereas 1–5-lm particles the sulphation was
limited by chemical reaction kinetics above 8508C.
Adanez and co-workers [49, 50] used the changing
grain size model (CGSM) proposed by Georgakis
et al. [43] combined with the molecular scale hypoth-
eses of Attar et al. [51] to determine the kinetic pa-
rameter data for the sulfidation reaction over a broad
range of particle sizes and temperatures. Good agree-
ment between measured and predicted conversion
time curves and sulfur distributions inside the par-
ticles was observed. Heesink and co-workers [52]
developed a new grain size distribution model
(GSDM) based on the grain model firstly proposed
by Szekely et al. [41], but additionally takes the size
distribution of the reacting grains into account, in
which they presumes that the porous particle is a col-
lection of grains of various sizes, and its size distribu-
tion can be converted into grain size distribution by
pore-to-sphere factor whose value can be derived
from the porosimetry measurements. This model also
was used to study the mechanism of the sulfidation
of calcined limestone with hydrogen sulfide and car-
bonyl and fairly satisfactory simulation results were
achieved. Lew et al. [53] presented overlapping grain
model to describe the sulfidation reaction with zinc
titanates and oxides, where the porous solid is simu-
lated as an assemblage of grains randomly distributed
in space with overlapping of the grains permitted.
Sun et al. [54] reported that the sulfidation process of
porous zinc oxide sorbent with hydrogen sulfide hap-
pened in five steps after external mass transfer and
pore diffusion. A reversible, adsorption, and ion-
migration (RAIM) model was used and predicted
results were consistent with selected experiments of
single-particle sulfidation. Flytzani-Stephanopoulos
et al. [55] presented a detailed overview of kinetics
studies of reduction and sulfidation of single and
mixed oxide sorbents, which have appeared in the
recent literature, especially for zinc and copper-based
materials. For all the models, the more sophisticated
kinetic models require more parameters to solve the
reaction kinetics. Generally, there are several regimes
such as film mass transfer, pore diffusion, and chemi-
cal reaction in the activation process, which affects
on the overall reaction progressing, one or more of
which can be rate determining: (1) external mass
transfer from the gas phase to the sorbent particles’
sample, (2) diffusion of the gaseous reactant inside
the sample holder to a sorbent particle, (3) diffusion
of the gaseous reactant inside a sorbent particle
(pores and product layer), and (4) chemical reaction.
About the development of mathematical expression
of reaction mechanisms and the kinetics of sulfidation
with calcium-based materials by using USCM with
variable effective diffusivity will be covered in future
publications.
CONCLUSION
Sulfidation and regeneration simulations were per-
formed using the Factsage equilibrium and phase sta-
bility diagram model by applying the free energy
minimization method, which seems to be useful tool
for predicting and analyzing performance of sulfur
sorbent materials. Even though in this paper, only
single metal oxide systems are examined, the results
will favor in the further study of mixed-metal sorbent
compounds. Copper, manganese, and zinc oxides are
the most favorable metals, which are capable of
reaching even ppb level at a temperature of about
6508C. However, evaporation, sintering, excessive
temperature rising during regeneration, or mechanical
disintegration have a negative impact on their per-
formance, especially at higher temperatures. For
desulfurization at the temperature over 9008C, cal-
cium-based oxide exhibits a better potential than
other metal oxides. However, their desulfurization
capability is strongly limited by the appropriate tem-
perature range and gas composition especially the
H2O and CO2 content. Supporting the oxides on a
carrier material or also through the use of mixed
metal oxides will be the most suitable approaches to
reduce or solve the volatilization, sintering problem
during sulfidation and regeneration reaction. So, for
the development of a low cost, highly efficient, rap-
idly reacting, and regenerable sorbent, with high sul-
fur capture capacity and durability still will be a long
way to go. As a conclusion, the simulation results
achieved in this study show a good agreement with
predicted and experimental one in the literature. Fur-
thermore, the reaction involved in the desulfurization
process of a real gasifier is a much more complicated
than just equilibrium reaction. Also, a pure-equilib-
rium model has its limitations, and because the reac-
tions and mass transfer in the real gasification process
are usually controlled by nonequilibrium factors. The
simulation results presented in this paper can only be
a reference for improving the efficiency of sorbent
utilization and optimize desulfurization process.
ACKNOWLEDGMENTS
The authors gratefully thank financial support from
the European Commission in the framework of the
FP7 Integrated Project ‘‘GreenSyngas’’ (Project NO.
213628).
Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
LITERATURE CITED
1. Huang, Y., Rezvani, S., McIlveen-Wright, D., Min-
chener, A. & Hewitt, N. (2008). Techno-economic
study of CO2 capture and storage in coal fired
oxygen fed entrained flow IGCC power plants,
Fuel Processing Technology, 89, 916–925.
2. Gil, J., Corella, J., Aznar, M.P. & Caballero, M.A.
(1999). Biomass gasification in atmospheric and
bubbling fluidized bed: Effect of the type of gasi-
fying agent on the product distribution, Biomass
and Bioenergy, 17, 389–403.
3. Gabra, M., Pettersson, E., Backman, R. & Kjell-
strom, B. (2001). Evaluation of cyclone gasifier
performance for gasification of sugar cane resi-
due, Part 1: Gasification of bagasse. Biomass and
Bioenergy, 21, 351–369.
4. Garcia, L., Salvador, M.L., Arauzo, J. & Bilbao, R.
(2001). CO2 as a gasifying agent for gas produc-
tion from pine sawdust at low temperature using
Ni/Al co-precipitated catalyst. Fuel Processing
Technology, 69, 157–174.
5. Zevenhoven, R. & Kilpinen, P. (2004). Control of
pollutants in flue gases and fuel gases (3rd Edi-
tion), Espoo: Energy Engineering and Environ-
mental Protection Publications.
6. Bakker, J.W., Kapteijn, H. & Moulijn, J.A. (2003).
A high capacity manganese-based sorbent for re-
generative high temperature desulfurization with
direct sulfur production conceptual process appli-
cation to coal gas cleaning. Chemical Engineering
Journal, 96, 223–235.
7. Van der Drift, A., van Doorn, J. & Vermeulen,
J.W. (2001). Ten residual biomass fuels for circu-
lating fluidized-bed gasification. Biomass and
Bioenergy, 20, 45–56.
8. Westmoreland, P.R. & Harrison, D.P. (1976). Eval-
uation of candidate solids for high-temperature
desulfurization of low-Btu gases. Environmental
Science and Technology, 10, 659–661.
9. Westmoreland, P.R., Gibson, J.B. & Harrison, D.P.
(1977). Comparative kinetics of high temperature
reaction between H2S and selected metal oxides.
Environmental Science and Technology, 11, 488–
491.
10. Swisher, J.H. & Schwerdtfeger, K. (1992). Review
of metals and binary oxides as sorbents for
removing sulfur from coal-derived gases. Journal
of Materials Engineering and Performance, 1,
399–408.
11. Swisher, J.H. & Schwerdtfeger, K. (1992). Ther-
modynamic analysis of sorption reactions for the
removal of sulfur from hot gases. Journal of
Materials Engineering and Performance, 1, 565–
571.
12. Gasper-Galvin, L.D. Zeolite-supported metal oxide
sorbents for hot-gas desulfurization. Industrial and
Engineering Chemistry Research, 1998; 37: 4157–
4166.
13. Elseviers, W.F. & Verelst, H. (1999). Transition
metal oxides for hot gas desulfurization. Fuel, 78,
601–612.
14. Hepworth, M.T., Berns, J. J. & Sadecki, K.A.
(1997). Kinetics of Mn-based sorbents for hot
October 2009 369
 coal gas desulfurization, Final Technical Report,
DE-FG22-94PC94212–11.
15. Ben-Slimane, R. & Hepworth, M.T. (1994). Desul-
furization of hot coal-derived fuel gases with
manganese-based regenerable sorbents. I. Loading
(sulfidation) tests. Energy and Fuels, 8, 1175–1183.
16. Li, X., Grace, J.R., Watkinson, A.P., Lim, C.J. &
Ergudenler, A. (2001). Equilibrium modeling of
gasification: A free energy minimization approach
and its application to a circulating fluidized bed
coal gasifier. Fuel, 80, 195–207.
17. Dı́az-Somoano, M. & Martı́nez-Tarazona, M.R.
(2003). Trace element evaporation during coal
gasification based on a thermodynamic equilib-
rium calculation approach, Fuel, 82, 137–145.
18. Thompson, D. & Argent, B.B. (2002). Prediction
of the distribution of trace elements between the
product streams of the Prenflo gasifier and com-
parison with reported data. Fuel, 81, 555–570.
19. Bale, C.W., Chartrand, P., Degterov, S.A., Eriks-
son, G., Hack, K., Ben Mahfoud, R., Melançon, J.,
Pelton, A.D., Peterson, S. (2002). FactSage Ther-
mochemical Software and Databases. Calphad
Journal, 26, 189–228.
20. Lew, S., Jothimurugesan, K. & Flytzani-Stephano-
poulos, M. (1989). High temperature H2S removal
from fuel gases by regenerable zinc oxide-tita-
nium dioxide sorbents. Industrial and Engineer-
ing Chemistry Research, 28, 535–541.
21. Zeng, Y., Kaytakoglu, S. & Harrison, D.P. (2000).
Reduced cerium oxide as an efficient and durable
high temperature desulfurization sorbent. Chemi-
cal Engineering Science, 55: 4893–4900.
22. Ko, T.H., Chu, H. & Chaung, L.K. (2005). The
sorption of hydrogen sulfide from hot syngas by
metal oxides over supports. Chemosphere, 58,
467–474.
23. Garcia, E., Cilleruelo, C., Ibarra, J.V., Pineda, M.
& Palacios, J.M. (1997). Kinetic study of high tem-
perature removal of H2S by novel metal oxide
sorbents. Industrial and Engineering Chemistry
Research, 1997; 36: 846–853.
24. Ben-Slimane, R. & Hepworth, M.T. (1994). Desul-
furization of hot coal-derived fuel gases with
manganese-based regenerable sorbents. II.
Regeneration and multicycle tests. Energy and
Fuels, 8, 1184–1191.
25. Ben-Slimane, R. & Hepworth, M.T. (1992). Desul-
furization of hot coal-derived fuel gases with
manganese-based regenerable sorbents. III. Fixed
bed testing. Energy and Fuels, 9, 372–378.
26. Abbasian, J. & Slimane, R.B. (1998). A regener-
able copper-based sorbent for H2S removal from
coal gases. Industrial and Engineering Chemistry
Research, 37, 2775–2782.
27. Gupta, R.P., Gangwal, S.K. & Jain, S.C. (1992).
Development of zinc ferrite sorbents for desulfur-
ization of hot coal gas in a fluid-bed reactor.
Energy Fuels, 6, 21–27.
28. Focht, G.D., Ranade, P.V. & Harrison, D.P. (1988).
High-temperature desulfurization using zinc fer-
rite: Reduction and sulfidation kinetics. Chemical
Engineering Science, 43, 3005–3013.
370 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
29. Fenouil, L.A. (1994). Kinetic and structural studies
of the sulfidation of large particles of lime and
limestone in coal gas, Paper presented at AIChE
Annual Meeting, San Francisco, CA, November 16.
30. van der Ham, A.G.J., Heesink, A.B.M., Prins, W.
& van Swaaij, W.P.M. (1996). Proposal for a re-
generative high-temperature process for coal gas
cleanup with calcined limestone. Industrial and
Engineering Chemistry Research, 35, 1487–1495.
31. Fuda, K., Palmer, A.D. & Sears, P. (1991). Chemi-
cal changes occurring during sulphidation and
regeneration of iron containing sorbents. Fuel 70,
100–106.
32. Alonso, L., Palacios, J.M., Garcıa, E. & Moliner, R.
(2000). Characterization of Mn and Cu oxides as
regenerable sorbents for hot coal gas desulfuriza-
tion. Fuel Processing Technology, 62, 31–44.
33. Akyurtlu, J.F. & Akyurtlu, A. (1995). Hot gas
desulfurization with vanadium-promoted zinc fer-
rite sorbents. Gas Seperation and Purification, 9,
17–25.
34. Jothimurugesan, K. & Gangwal, S.K. (1998).
Regeneration of zinc titanate H2S sorbents. Indus-
trial and Engineering Chemistry Research, 37:
1929–1933.
35. Atakul, H., Wakker, J.P., Gerritsen, A.W. & van
den Berg, P.J. (1996). Regeneration of MnO/g-
Al2O3 used for high-temperature desulfurization
of fuel gases. Fuel, 75, 373–378.
36. Yasyerli, S. (2008). Cerium-manganese mixed
oxides for high temperature H2S removal and
activity comparisons with V-Mn, Zn-Mn, Fe-Mn
sorbents. Chemical Engineering and Processing,
4, 577–584.
37. Bjoernbom, E.N., Druesne, S., Zwinkels, M.F.M.
& Jaeraas, S.G. (1995). Study on the regenera-
tion of copper-manganese sorbent for removal
of sulfur dioxide from flue gases. Industrial
and Engineering Chemistry Research, 34: 1853–
1858.
38. Tamhankar, S.S., Bagajewicz, M., Gavalas, G.R.,
Sharma, P.K. & Flytzani-Stephanopoulos, M.
(1986). Mixed-oxide sorbents for high-tempera-
ture removal of hydrogen sulfide. Industrial and
Engineering Chemistry Process Design and De-
velopment, 25: 429–437.
39. Yagi, S. & Kunii, D. (1955). Studies on combus-
tion of carbon particles in flames and fluidized
beds. In 5th Symposium (International) on Com-
bustion (pp.231–244), Reinhold, New York.
40. Levenspiel, O. (1972). Chemical reaction engineer-
ing (2nd Edition, Chapter 12), New York: Wiley.
41. Szekely, J. & Evans, J.W. (1970). A structural
model for gas-solid reactions with a moving
boundary. Chemical Engineering Science, 25:
1091–1107.
42. Bhatia, S.K. & Perlmutter, D.D. (1980). A random
pore model for fluid-solid reactions. I. Isothermal
kinetic control. AIChE J, 26, 379–385.
43. Georgakis, C., Chang, C.W. & Szekely, J. 1979.
A changing grain size model for gas-solid reac-
tions. Chemical Engineering Science, 34, 1072–
1075.
44. Zevenhoven, C.A. P., Yrjas, K.P. & Hupa M.M.
(1995). How does sorbent particle structure influ-
ence sulfur capture under PFBC conditions. Pro-
ceedings of the 13th international conference
on fluidized bed combustion (pp. 1381–1392),
Orlando FL, ASME: Fairfield, NJ.
45. Zevenhoven, C.A.P., Yrjas, K.P. & Hupa, M.M.
(1996). Hydrogen sulfide capture by lime-
stone and dolomite at elevated pressure. II.
Sorbent particle conversion modeling. Indus-
trial and Engineering Chemistry Research, 35,
943–949.
46. Iisa, K. & Hupa, M. (1992). Rate limiting proc-
esses for the desulphurisation reaction at elevated
pressures. Journal of the Institute of Energy, 65,
201–205.
47. Iisa, K., Hupa, M. & Yrjas, P. (1992). Product
layer diffusion in the sulphation of calcium
carbonate. Twenty-Fourth Symposium (Interna-
tional) on Combustion, 24, 1349–1356.
48. Iisa, K. & Hupa, M. (1990). Sulphur absorption
by limestone at pressurized fluidized bed condi-
tions. Twenty-Third Symposium (International)
on Combustion, 23, 943–948.
49. Adánez, J., de Diego, L.F., Garcı́a-Labiano, F. &
Abad, A. (1998). Kinetics of H2S reaction with cal-
cined calcium-based sorbents. Energy Fuels, 12,
617–625.
Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep
50. De Diego, L.F., Abad, A., Garcı́a-Labiano, F., Adá-
nez, J. & Gayán, P. (2004). Simultaneous calcina-
tion and sulfidation of calcium-based sorbents.
Industrial and Engineering Chemistry Research,
43, 3261–3269.
51. Attar, A. & Dupuis, F. (1979). The rate and the
fundamental mechanisms of the reaction of
hydrogen sulfide with the basic minerals in coal.
Industrial and Engineering Chemistry Process
Design and Development, 18, 607–618.
52. Heesink, A.B.M., Prins, W. & van Swaaij, W.P.M.
(1993). A grain size distribution model for non-
catalytic gas—Solid reactions. The Chemical Engi-
neering Journal and the Biochemical Engineering
Journal, 53: 25–37.
53. Lew, S., Sarofim, A.F. & Stephanopoulos, M.F.
(1992). Modeling of the sulfidation of zinc-tita-
nium oxide sorbents with hydrogen sulfide.
AIChE J 38, 1161–1169.
54. Sun, J., Modi, S., Liu, K., Lesieur, R. & Buglass, J.
(2007). Kinetics of zinc oxide sulfidation for
packed-bed desulfurizer modeling. Energy Fuels,
21, 1863–1871.
55. Flytzani-Stephanopoulos, M. & Li, Z. (1998).
Kinetics of sulfidation reactions between H2S and
bulk oxide sorbents. In A.T. Atimtay, D.P. Harrison,
(Eds), NATO ASI ‘‘Desulfurization of Hot Coal Gas,’’
Springer-Verlag: Berlin Heidelberg, 42, 179–211.
October 2009 371