Promoted ZnO Sorbents for Wide Temperature Range H2S/COS Removal for Applications 
in Fuel Cells 
 
by 
 
Priyanka P. Dhage 
 
 
 
 
A dissertation submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Doctor of Philosophy 
 
Auburn, Alabama 
August 6, 2011 
 
 
 
 
Keywords:  desulfurization, hydrogen sulfide, carbonyl sulfide, doped sorbents, 
Hydrolysis 
 
 
 
 
Approved by 
 
Bruce J. Tatarchuk, Chair, Professor of Chemical Engineering 
Yoon Y. Lee, Professor of Chemical Engineering 
William Ashurst, Professor of Chemical Engineering 
Aleksandr Simonian, Professor of Materials Engineering 
 
ii 
 
Abstract 
 
 
High efficiency desulfurization is critical to maintain the activity of fuel processing 
catalysts and high-value membrane electrode assemblies in logistic fuel cell systems. On-board 
fuel processing of liquid hydrocarbon fuel is being investigated to supply hydrogen for fuel cell-
based auxiliary power units. For such a system, if sulfur is not removed from the liquid phase, 
the removal of sulfur as H2S from the reformate becomes a key-step since downstream catalysts 
and the fuel cell itself can be poisoned by a small amount of H2S in the feed. Hydrogen sulfide is 
present in many high temperature gas streams during extraction and processing of fossil fuels, 
natural gas and geothermal brines. Steam reforming catalysts, PEM anode catalysts and also the 
shift catalysts are intolerant to sulfur and to ensure adequate lifetime of fuel processors the 
desulfurization step is very important.  
This dissertation presents results of R&D efforts to develop novel sorbents for efficient 
gas phase desulfurization. Promoted ZnO sorbents with formulation M0.05Zn0.95O (M = Mn, Fe, 
Co, Ni, Cu) were supported on silica and effect of support, various operating parameters and 
microfibrous entrapment was studied. The results of desulfurization tests on these sorbents at 
room temperature indicate that a copper doped ZnO (15% w/w)/MCM-41 sorbent 
(Cu0.05Zn0.95O/MCM-41) has the highest saturation sulfur capacity at 0.9 mol S/mol 
(Cu0.05Zn0.95O), which is approximately twice that of ZnO/SiO2 sorbent at similar loadings. the 
utilization of the reactant (M0.05Zn0.95O) toward H2S removal depended on the support employed 
 
iii 
 
in the order MCM-41 > MCM-48 > silica gel. This dependence was investigated in terms of the 
support: surface area, pore volume, and pore size; using N2 adsorption-desorption isotherms 
(Chapter III). 
The Cu-ZnO/SiO2 sorbent for ultradeep adsorptive removal of H2S from the reformate 
streams at room temperature was prepared, tested, and characterization of the active sites was 
performed. The Cu dopant significantly enhances desulfurization capacity of ZnO/SiO2 sorbent 
at room temperature (up to 92 % utilization of ZnO), and maintains a high sulfur uptake capacity 
upon multiple cycles (up to 10) of regeneration by a simple thermal oxidation in air. XRD 
suggests that both zinc and copper compounds of the CuO-ZnO/SiO2 sorbent are nano-dispersed. 
The ESR spectroscopy found that the ?calcined? and ?sulfided? CuO-ZnO/SiO2 sorbents contain 
Cu2+ in the single dispersion and coordination state and during H2S adsorption, partial reduction 
of Cu2+ to Cu1+ occurs (Chapter IV) 
The Fe- and Mn-promoted H2S sorbents Fex-Mny-Zn1-x-yO/SiO2 (x, y=0, 0.025) for the 
ultradeep desulfurization of model reformates at room temperature were prepared, tested and 
characterized. The role of Mn and Fe promoter cations in the ?calcined? and ?sulfided? forms of 
the FexMnyZnO (1-x-y)/SiO2 sorbent has been studied by the in-situ ESR, temperature dependent 
XPS. Operando ESR is used for the first time to study dynamics of reduction of Mn3+promoter 
sites simultaneously with measuring sulfidation dynamics of the Fex?Mny?
Zn1?x?yO/SiO2 sorbent. Fe cations are believed to occupy the surface of supported ZnO 
nanocrystallites, while Mn cations are distributed within ZnO (Chapter V) 
Removal of both H2S and COS from reformate streams is critical for maintaining the 
activity of fuel processing catalysts. At temperatures < 250 C, COS formation is effectively 
inhibited, but at temperatures above 250 C, significant amount of COS is formed in presence of 
 
iv 
 
CO2/CO and H2S. A layered bed approach was used with layer of Al2O3/Carbon for COS 
hydrolysis over the followed by a layer high efficiency H2S removal over bimetallic-promoted 
supported ZnO sorbent (Chapter VI). The objective of our work is developing the sorbents for an 
efficient, cost-effective and scalable removal of H2S and COS over the broad temperature range, 
without significant activity loss upon multiple regeneration cycles, and understanding the 
mechanism of sulfur sorption by the metal oxide-promoted ZnO-based sorbents. 
  
 
v 
 
Acknowledgements 
 
 
I would like to acknowledge the guidance and encouragement of my advisor Dr. Bruce 
Tatarchuk. I would like to express my sincere gratitude to Dr. Yoon Lee, Dr. Robert Ashurst and 
Dr. Aleksandr Simonian and Dr. Evert Duin for serving on my committee. This dissertation 
would not have been possible without the unwavering support of Dr. Hongyun Yang from 
Intramicron Inc. Also without the cooperation and support of my colleagues at the Center for 
Microfibrous Materials Manufacturing, especially Dwight Cahela, Dr. Don Cahela, Dr. 
Alexander Samokhvalov, Megan Schumacher, Kimberly Dennis, Benjamin Doty, Matt and 
Wendall from Glass Shop Sachin Nair, Hussain, Amogh Karwa, Abhijeet Phalle, Robert 
Henderdon, Min Sheng, Achintya Sujant and many other, this work would not have been 
possible. I am also grateful to Sue Abner and Karen Cochran for their administrative support 
throughout my tenure at Auburn.  
Most importantly, I would like to thank my family and especially my parents, sister- 
Supriya & brother - Pratik for their support and trust in my abilities. My sincere thanks goes to 
my friends especially Jola Jayselene and Saurabh Wadwalkar who made my stay at Auburn, one 
of the most memorable of all times. 
 
vi 
 
Table of Contents 
 
 
Abstract ........................................................................................................................................... ii 
Acknowledgements ......................................................................................................................... v 
List of Tables ................................................................................................................................. xi 
List of Figures .............................................................................................................................. xiii 
Nomenclature ............................................................................................................................... xix 
Chapter I: Introduction and Literature Survey ................................................................................ 1 
I.1 Introduction ......................................................................................................................... 1 
I.2 Literature Review................................................................................................................ 5 
I.2.1  Desulfurization Technologies ....................................................................................... 5 
I.2.2  ZnO based sorbents ...................................................................................................... 6 
I.3 Low Temperature Desulfurization ...................................................................................... 9 
I.4 COS Removal/Inhibition .................................................................................................. 10 
I.5 High Temperature Desulfurization ................................................................................... 11 
I.6 Microfibrous Entrapped Sorbents ..................................................................................... 14 
I.7 Advantages of supported sorbents .................................................................................... 16 
I.7.1  Novel Support ? Mobil composition of Matter- MCM-41 ......................................... 19 
I.8 Scope and Objective of the work: ..................................................................................... 22 
I.9 Objective of this work ....................................................................................................... 24 
I.10  Outline of this work ........................................................................................................ 25 
Chapter II: Experimental Setup and Characterization Techniques ............................................... 26 
 
vii 
 
II.1   Sorbent Preparation ......................................................................................................... 26 
II.1.1   Sorbent for Packed Bed ............................................................................................ 26 
 II.1.1.1  Preparation of doped supported sorbent ............................................................................ 26 
 II.1.1.2  Preparation of Mesoporous type silica (MCM) .............................................................. 28 
 II.1.1.3   Preparation of Al2O3/Carbon ............................................................................................... 30 
 II.1.1.4  Glass fiber entrapped Sorbent preparation ....................................................................... 30 
II.2   Pressure drop measurement set-up .................................................................................. 31 
II.3   Experimental Procedure .................................................................................................. 33 
II.4   Adsorption experiment.................................................................................................... 33 
II.5  Analytical/Characterization Techniques ......................................................................... 36 
Chapter III: Wide Temperature Range H2S Removal by Promoted ZnO/SiO2 ............................ 40 
III.1   Introduction ................................................................................................................... 41 
III.2  Experimental Section ..................................................................................................... 43 
III.2.1  Silica support ........................................................................................................... 43 
III.2.2  Sorbent impregnation ............................................................................................... 43 
III.2.3  Adsorption experiment............................................................................................. 43 
III.3   Sorbent Characterization ............................................................................................... 44 
III.4 Results and discussion .................................................................................................. 45 
III.4.1  Preparation and characterization of ZnO supported sorbents .................................. 45 
III.4.2  Effect of different types of metal oxides.................................................................. 54 
III.4.3  Comparison with the commercial ZnO .................................................................... 56 
III.4.4  Screening test for the metal oxide ............................................................................ 58 
III.4.5  Effect of promoter .................................................................................................... 59 
     III.4.5.1  Effect of change in concentration of the promoter ...................................................... 61 
 
viii 
 
III.4.6  Effect of pore volume .............................................................................................. 62 
III.4.7  Effect of calcination temperature ............................................................................. 68 
III.4.8  Effect of H2S sorption temperature .......................................................................... 70 
III.4.9  Comparative performance of different types of silica support ................................ 73 
III.4.10 Effect of moisture content ...................................................................................... 75 
III.4.11 Effect of regeneration temperature ......................................................................... 76 
 III.4.11.1  Desorption test during regeneration ............................................................................... 77 
III.4.12 Effect of CO and CO2 ............................................................................................. 77 
III.4.13 Novel Bimetallic Sorbents for H2S removal at room temperature ......................... 80 
III.4.14 Scale-up studies ...................................................................................................... 81 
III.5 Microfibrous Entrapped Sorbent................................................................................... 83 
III.5.1  Kinetic effects due to microfibrous entrapped ZnO sorbents( MFES) .................... 83 
III.5.2  Preparation of MFES ............................................................................................... 84 
III.5.3  Model Evaluation ..................................................................................................... 84 
III.5.4  Effect of face velocity .............................................................................................. 85 
III.5.5  Effect of Pressure: .................................................................................................... 89 
III.5.6  Composite bed design .............................................................................................. 91 
III.6  Removal of SO2: ............................................................................................................ 94 
III.7  Conclusions .................................................................................................................... 95 
Chapter IV: Copper Promoted ZnO/SiO2 Regenerable Sorbents for the Room Temperature 
Removal of H2S from Reformate Gas Streams.......................................................................... 97 
IV.1  Introduction.................................................................................................................... 99 
IV.2  Experimental ................................................................................................................ 101 
IV.3.  Results and Discussion ............................................................................................... 102 
 
ix 
 
IV.3.1  Desulfurization Performance of the Sorbents ........................................................ 102 
IV.3.2  Performance of the Sorbents upon Multiple Regeneration Cycles ........................ 105 
IV.3.3  Structural Characterization of the Sorbents ........................................................... 105 
IV.3.4  Characterization of the Sorbents by XPS .............................................................. 106 
IV.3.5  Characterization of the Sorbents by ESR .............................................................. 113 
IV.4   Conclusions................................................................................................................. 122 
Chapter V: Regenerable Fe-Mn-ZnO/SiO2 sorbents for Room Temperature Removal of 
H2S from Fuel Reformates: Performance, Active sites and Operando studies ........................ 123 
V.1   Introduction .................................................................................................................. 124 
V.2   Experimental ................................................................................................................ 128 
V.3   Results and Discussion................................................................................................. 129 
V.3.1  Performance of the FexMnyZn1-x-yO/SiO2 Sorbents ................................................ 129 
V.3.2  Structural Characterization of the Sorbents ............................................................ 132 
V.3.3  Performance of the Sorbents upon Multiple Regeneration Cycles ......................... 135 
V.3.4  Characterization of the Sorbents by XPS ................................................................ 137 
V.3.5  Characterization of the FexMnyZn1-x-yO/SiO2 Sorbents by ESR ............................. 142 
V.4   Conclusions ................................................................................................................... 149 
Chapter VI: RT Hydrolysis and Removal of COS from Fuel Reformate Streams using 
Al2O3/Carbon & Fe0.025Mn0.025ZnO0.95/SiO2 Layered Beds ..................................................... 150 
VI.1  Introduction.................................................................................................................. 151 
VI.2  Experimental ................................................................................................................ 155 
VI.3  Results and Discussions ............................................................................................... 157 
VI.3.1  Desulfurization Performance of the Sorbents ........................................................ 157 
VI.3.2  COS Removal & Hydrolysis ................................................................................. 159 
VI.4  Conclusions.................................................................................................................. 169 
 
x 
 
Chapter VII: Conclusions and Recommendations for Future Work ........................................... 171 
VII.1  Conclusions ................................................................................................................ 171 
References ................................................................................................................................... 175 
Appendix I ? Calculation formulae............................................................................................. 197 
Appendix II ? GC Chromatography Analytic Methods .............................................................. 199 
a. TCD Analysis Method ................................................................................................. 199 
b. PFPD Analysis Method................................................................................................ 200 
Appendix III ? Calibration of Gases ........................................................................................... 201 
a. Carbon Dioxide ............................................................................................................ 201 
b. Nitrogen ....................................................................................................................... 202 
c. Carbon Monoxide ........................................................................................................ 203 
d. Hydrogen Sulfide ......................................................................................................... 204 
e. Carbonyl Sulfide .......................................................................................................... 205 
f. Furnace ......................................................................................................................... 206 
Appendix IV ? Inventory of Chemicals used ............................................................................. 207 
 
 
 
 
 
 
 
 
 
xi 
 
 
List of Tables 
 
 
Table I. 1: The fuel requirements for the principal fuel cells .................................................2 
Table I. 2: Equilibrium data for ZnO+H2S = ZnS+H2O by HSC* software ..........................8 
Table I. 3: Properties of the Glass fibers ..............................................................................16 
Table I. 4: Comparison of the literature Review on preparation of MCM- 41 ....................20 
Table  III.1:  Structural characteristics of Silica sorbents determined by N2 adsorption .........48 
Table III.2:  Saturation capacity values of the doped sorbents and commercial 
sorbents ...............................................................................................................61 
Table III.3: Capacity values of the silica with varying pore volumes and their 
adsorption capacities ...........................................................................................63 
Table III.4: Theoretical utilization values for scale up of the sorbent ...................................81 
Table III.5:  Composition of the GFES ...................................................................................84 
Table III.6:  Operating conditions: change in m: v with face velocity and length of 
bed .......................................................................................................................86 
Table III.7a:  Lumped K values for Material 1 ........................................................................87 
Table III.7b:  Lumped K values for Material 2 ........................................................................87 
Table III.8a:  Pressure gradient and log reduction for Material 1 .............................................90 
Table III.8b:  Pressure Gradient and Log reduction values for Material 2 ..............................90 
Table III. 9:  Composition of the packed bed and polishing layer ...........................................93 
Table III.10:  Saturation capacity of the sorbents tested for SO2 removal ...............................95 
Table IV.1:  Sulfur capacities of the sorbents M0.05ZnO0.95/SiO2 .........................................102 
TableIV.2: Sulfur capacities and ZnO utilization of the doped sorbents 
Cu0.05ZnO0.95/SiO2 vs. the un-doped ZnO/SiO2 sorbent. ..................................104 
 
xii 
 
Table IV.3:  Surface Area and Pore Volume Data analyzed using N2 Adsorption-
Desorption Curves ............................................................................................106 
Table V.1:  Comparative breakthrough, saturation capacities and ZnO utilization 
data ....................................................................................................................131 
Table V.2:  Structural characterization of various H2S sorbents .........................................134 
Table VI.1:  Breakthrough and Saturation H2S Capacity and utilization of ZnO (%) 
for various sorbents ...........................................................................................158 
Table VI.2:  Saturation Capacity of Fe0.025Mn0.025ZnO0.95/SiO2 with and without CO2 
at room temperature and 400 C .........................................................................163 
 
xiii 
 
List of Figures 
 
 
Figure I.1:  The concepts and steps for fuel processing for gaseous , liquid and solid 
fuels for high temperature and low temperature fuel cell applications .................1 
Figure I.2:  Fuel processing steps with amounts of poisons tolerable for operation in 
PEMFCs ................................................................................................................4 
Figure I.3:  Desulfurization technologies classified by the nature of the key process 
to remove sulfur ....................................................................................................5 
Figure I. 4:  Equilibrium H2S concentration (ppmv) using HSC software. ..............................7 
Figure I. 5:  Background on commercial ZnO sorbent based on literature search ...................8 
Figure I. 6:  SEM Image of the S2 glass fiber entrapped SiO2 particle.  ................................16 
Figure I. 7:  Important properties for sorbent formulation .....................................................17 
Figure I. 8:  Overview of literature review on various supports ............................................18 
Figure I. 9:  Schematic diagram of the fuel processing for PEMFCs with average 
operating temperatures ........................................................................................22 
Figure II.1a:  Preparation method of MCM- 48[40, 48, 60] .....................................................29 
Figure II.1b:  Preparation method of MCM- 48[40, 48, 60] .....................................................30 
Figure II. 2:  Schematic diagram of the pressure drop measurement .......................................32 
Figure II. 3:  Schematic diagram of the experimental set-up ...................................................35 
Figure II.4:  Schematic Diagram of the Configuration of the Reactor Bed ............................36 
Figure III.1a:  XRD Pattern of the MCM-41 .............................................................................45 
Figure III.1b:  XRD Pattern of the Lab-made MCM-48 ............................................................46 
Figure III.2a:  N2 adsorption-desorption isotherm for MCM-41- Commercial ........................47 
Figure III.2b:  N2 adsorption-desorption isotherm for MCM-48 ? Lab made ...........................48 
 
xiv 
 
Figure III. 3a:  N2 adsorption isotherms for MCM-41, ZnO/MCM-41 and Cu-
ZnO/MCM-41 .....................................................................................................50 
Figure III. 3b:  Pore Size Distribution of MCM-41 (commercial ) ............................................51 
Figure III. 3c:  Pore size distribution of MCM-48 (Lab made)...................................................52 
Figure III. 4a:  SEM image of MCM-48 sample before impregnation - MCM-48 Blank ..........53 
Figure III. 4b:  SEM image of MCM-48 sample after impregnation - 15 % ZnO / MCM-
48.........................................................................................................................53 
Figure III. 5a:  Breakthrough capacity and % theoretical capacity valuesa for different 
metal oxides with same loading on silica tested at RT, Q= 100 cc/min, 
face velocity = 2.12 cm/s, Calcination condition = 350 C/1h, Wt. = 0.5 
g...........................................................................................................................55 
Figure III. 5b:  Equilibrium H2S concentration data generated using HSC software for 
various metal oxides ...........................................................................................56 
Figure III.6:  Breakthrough curves for commercial ZnO and ZnO/SiO2: T = 20 C, Co = 
1Vol% H2S/H2, Face Velocity = 2.12 cm/s, Wt. =0.5 gm ..................................57 
Figure III.7:  Saturation capacity of the doped sorbents .........................................................60 
Figure III. 8:  Saturation Capacity of the sorbents with varying Cu concentrations ................62 
Figure III.9a:  Saturation Capacities of the different silica with varying pore volume 
(i.e. varying ZnO loading) ..................................................................................64 
Figure III.9b:  Saturation capacity of ZnO/SiO2 with varying ZnO loading .............................65 
Figure III.9c:  XRD patterns for silica and ZnO/SiO2 with varying ZnO loadings ..................66 
Figure III.9d:  XRD Pattern showing effect of adding Cu (0.05-1) on ZnO/SiO2 ....................67 
Figure III.9e:  XRD Pattern of pure ZnO and Cu0.05Zn0.95O made from the calcination 
(350 C/1h/air) from nitrate precursors. ...............................................................68 
Figure III.10a: Saturation capacity of ZnO/SiO2 calcined at different temperatures ..................69 
Figure III.10b: XRD patterns of the ZnO( 36 wt.%) /SiO2 calcined at different 
temperatures. .......................................................................................................70 
Figure. III.11:  H2S saturation capacity for ZnO/SiO2 and ZnO/MCM-41 (15 wt% 
loading) tested from room temperature to 400 C, Q = 100 cc/min, Face 
velocity = 2.12 cm/s, Calcination condition = 350oC/1h ....................................71 
 
xv 
 
Figure III.12:  H2S saturation capacity for ZnO/MCM-41 and Cu0.05ZnO0.95/MCM-
41 (15 wt% ZnO) tested at RT and 400 C, Q = 100 cc/min, Face velocity 
= 2.12 cm/s, Calcination condition = 350oC/1h ..................................................72 
Figure III.13:  H2S Breakthrough curves for ZnO/SiO2 and ZnO/MCM (15 wt% ZnO) 
compared with Commercial ZnO (~90% ZnO) tested at RT, Q = 100 
cc/min, Face velocity = 2.12 cm/s, Calcination condition = 350oC/1h ..............74 
Figure III.14:  H2S saturation capacity for ZnO/SiO2 and ZnO/MCM-41 (15wt. % 
ZnO) tested at varying moisture content (0-10%) at RT, Q = 100 cc/min, 
Face velocity = 2.12 cm/s, Calcination ...............................................................76 
Figure III.15:  Breakthrough curves for Cu doped ZnO/SiO2 tested in the presence of 
CO and CO2 ........................................................................................................78 
Figure  III.16:  Breakthrough performance of  Fe0.025Mn0.025ZnO0.95/SiO2 with and 
without CO2 at 400 C,  Test conditions :Q (2%H2S/H2) = 100 cc/min, 
Q(100% CO2) = 100 cc/min,  T = 400 C, GHSV = 8800 h-1 , Wt= 0.5 g ...........79 
Figure III.17:  Saturation Capacities of the novel bimetallic doped sorbents for H2S 
removal ...............................................................................................................80 
Figure III.18:  V-blender used for impregnation of samples for scale-up studies .....................82 
Figure III.19: Breakthrough curves for scale-up studies ...........................................................82 
Figure III. 20:  Evaluation of modified Amundson model .........................................................85 
Figure III.21a: Relationship between lumped K and face velocity U for material 1 with 
glass fiber fraction = 3 vol% ...............................................................................88 
Figure III.21b:  Relationship between lumped K and face velocity U for material 2 with 
glass fiber fraction = 4.5 vol% ............................................................................89 
Figure III.22:  Pressure drop data for the packed bed and GFES (Material 1 and 2) at 
400oC ..................................................................................................................90 
Figure III.23:  a) Composite bed test using glass fiber entrapped sorbents as polishing 
layer. Performance of Polishing Sorbent and Packed Bed + Polishing 
Sorbent @ 1% H2S in H2, RT, 2.12 cm/s Breakthrough curves of a 2.5 
cm thick packed bed of ZnO/SiO2 extrudates and a composite bed (the 
packed bed followed with a 4mm polishing layer). b) Schematic diagram 
of the packed and composite bed. .......................................................................93 
Figure III.24:  Breakthrough performance of promoted ZnO/SiO2 and FeO/Al2O3 
(15wt.%). Test Conditions: T= 20 C, Co = 1vol% SO2/Air, Face velocity 
= 0.53 cm/s ..........................................................................................................95 
 
xvi 
 
Figure IV.1:   Breakthrough Curves for Commercial ZnO (BASF and Sud-Chemie) 
with 21 wt.% ZnO/SiO2 and Cu0.05ZnO0.95/SiO2. Test Conditions : Co = 
1 vol%H2S/H2, T= 20C, Face velocity = 2.12 cm/s ..........................................103 
Figure IV.2:  Breakthrough curves for Regeneration of Cu0.05ZnO0.95/SiO2.Test 
Conditions: Calcination Temp = 350 C/Air/1h, Adsorption at 20 C, Co = 
1vol% H2S/H2, Regeneration at : 550 C, Air/1h. ..............................................105 
Figure IV.3:  XPS Spectra of Calcined Cu0.05ZnO0.95/SiO2 ....................................................108 
Figure IV.4:  XPS Spectra of sulfided sorbents Cu0.05ZnO0.95/SiO2 .......................................112 
Figure IV.5:  ESR spectra of the ?calcined? sorbent CuxZn1-xO/SiO2 Figure 5A - 
Cu1.0Zn0.0O/SiO2, Figure 5B ? Cu0.1Zn0.9O/SiO2, Figure 5C - 
Cu0.01Zn0.99O/SiO2 and Figure 5D - Cu0.001Zn0.999O/SiO2. ...............................114 
Figure IV.6:  ESR spectrum of Cu2+ in Cu0.05Zn0.95O/SiO2 simulated as the single kind 
of Cu2+ species. .................................................................................................117 
Figure IV.7:   Figure 7A shows the relative signal intensity of Cu2+ proportional to 
molar concentration of Cu2+ in the ?calcined? vs. ?sulfided? sorbents 
CuxZn1-xO/SiO2 (x=0.001, 0.01, 0.05, 0.1 and 1). Figure 7B shows the 
yield Y of chemical reaction of the reduction of Cu2+ to Cu1+ upon the 
interaction with the H2S component of the reformate ......................................119 
Figure IV.8:  ESR spectrum of Cu2+ in the sorbent Cu0.05-Zn0.95O/SiO2, ?calcined? as-
prepared vs. ?calcined? upon 10 cycles of desulfurization-regeneration. ........121 
Figure V.1.  H2S Breakthrough curves of the commercial ZnO Sorbent from BASF 
(filled circles), Sud-Chemie (Squares), ZnO/SiO2 (open Circles) and 
Fe0.025/Mn0.025ZnO0.095/SiO2 sorbent (diamonds) ..............................................130 
Figure V.2:  H2S Capacity (mg Sulfur/ g Sorbent) and the total surface area vs. the 
loading of ZnO (wt. %) in the ZnO/SiO2 Sorbents. ..........................................133 
Figure V.3:  XRD Spectra of the ZnO/SiO2 sorbents at 36 wt% of ZnO (solid line) 
and 21 wt. % (dashed dotted line) vs. SiO2 support (dotted line) .....................134 
Figure V.4:   H2S breakthrough curves upon the multiple adsorption/regeneration of 
Fex-Mny-ZnO 1-x-y/SiO2 sorbent ........................................................................136 
Figure V.5:  The XPS lines of Fe 2p3/2 (Figure 5A), Zn Auger L3M45M45 (Figure 
5B), O 1s (Figure 5C) and Zn 2p (Figure 5D) of the ?calcined? sorbent 
Fe0.2Zn0.8O/SiO2. ...............................................................................................138 
Figure V.6:  The XPS lines of Zn Auger L3M45M45 (Figure 6A), Zn 2p (Figure 6B) 
and O 1s (Figure 6C) of the sulfided sorbent Fe0.2Zn0.8O/SiO2. .......................142 
 
xvii 
 
Figure V.7:  ESR spectrum of the ?calcined? sorbent Fe0.025Zn0.975O/SiO2 (dotted 
solid line) vs. Fe0.025Mn0.025Zn0.975O/SiO2 (solid line), Figure 7A. ESR 
spectrum of the ?sulfided? sorbent Mn0.025Zn0.975O/SiO2 (dotted solid 
line) vs. Mn0.025Fe0.025Zn0.975O/SiO2 (thick solid line), Figure 7B....................143 
Figure V.8:  Schematic diagram of the mechanism of distribution of the Mn, Fe 
active sites in ZnO/SiO2 ....................................................................................146 
Figure V.9:  Schematic representation of the structure of Fe0.025Mn0.025Zn0.95O/SiO2 
sorbents and sulfidation/regeneration reactions. ...............................................148 
Figure VI.1.  H2S Breakthrough curves of the commercial ZnO Sorbent from BASF, 
Sud-Chemie ZnO/SiO2 and Fe0.025/Mn0.025ZnO0.095/SiO2 sorbent. Test 
conditions: adsorption T= 20 C, Particle size = 100-200 microns, Co=1 
vol5 H2S/H2 .......................................................................................................157 
Figure VI.2:  COS hydrolysis at 400 C using Al2O3 based and SiO2 based sorbents. 
Inlet concentration: COS/N2 = 500 ppmv, 1% Steam, GHSV = 19000h-1 .......159 
Figure VI.3a:  Breakthrough performance of  Fe0.025Mn0.025ZnO0.95/SiO2 with and 
without CO2 at 400 C,  Test conditions :Q (2%H2S/H2) = 100 cc/min, 
Q(100% CO2) = 100 cc/min,  T = 400 C, GHSV = 8800 h-1 , Wt= 0.5 g .........160 
Figure VI.3b:  Equilibrium COS Concentrations. Reformate Composition (vol %): CO 
= 25 %, CO2 = 10%, N2 = 33 %, H2O = 7%, H2 = 25 % and H2S = 
0.03% ................................................................................................................161 
Figure VI.3c:  Equilibrium H2S Concentrations. Reformate Composition (vol %): CO 
= 25 %, CO2 = 10%, N2 = 33 %, H2O = 7%, H2 = 25 % and H2S = 
0.03% ................................................................................................................162 
Figure VI.3d:  Breakthrough performance of Fe0.025Mn0.025ZnO0.95/SiO2 at 20 C Test 
conditions: Q (2%H2S/H2) = 100 cc/min, Q(100% CO2) = 100 cc/min,  T 
= 20 C, GHSV = 3800 h-1 , Wt= 0.5 g ..............................................................164 
Figure VI.4:  Breakthrough curves of layered beds tested with 300 ppmv H2S-25% 
H2-25% CO-10% CO2-7% H2O-33% N2 at a face velocity=100 cm/s at 
400 C. Bed length: 22 cm .................................................................................165 
Figure VI.5:  COS Removal using layered bed. Test conditions: T = 400 C, GHSV = 
15000 h-1, Wt. of each layer = 0.5g Metal oxide loading of each layer= 
15%wt. Gas Composition (vol%) : CO2 = 28%, H2S = 0.5%, H2O = 1%, 
H2 = 70.5% ........................................................................................................166 
Figure VI.6:  COS Hydrolysis using Al2O3/C, Test conditions: Co = 1000 ppmv 
COS/N2, T= 20C, Particle Size = 100-200 microns. .........................................168 
 
xviii 
 
Figure VI.7:  COS Hydrolysis for extended time on Al2O3/C. Test conditions: Co = 
1000 ppmv COS/N2, T= 20C, Particle Size = 100-200 microns. ......................169 
 
xix 
 
Nomenclature 
 
 
Co - Initial H2S concentration, ppmv  
CAo - Initial challenge H2S molar concentration, mol/cc 
Cb - breakthrough concentration mol/cc 
K  - Lumped shape factor of breakthrough curve, s-1  
U -Face velocity, cm/s 
X - ZnO utilization of the accessible ZnO, dimensionless 
t -time, s 
t1/2 -time to reach 50% Cao, s 
? -saturation time s  
? - Void fraction, dimensionless 
 
 
1 
 
Chapter I: Introduction and Literature Survey 
 
I.1 Introduction  
With the very growing demands for fuels, and depleting natural resources it is the need of the 
day to find alternative fuel or equipments for futuristic technologies. Fuel cells are emerging 
technology with applications in transportation, stationary and portable power generation. 
Hydrogen is the real fuel for fuel cells, which can be obtained by fuel reformulation on-site for 
stationary applications or on-board for automotive applications. 
 
Figure I.1: The concepts and steps for fuel processing for gaseous , liquid and solid fuels for high 
temperature and low temperature fuel cell applications 
NG
G a s i fi c a t i o n
De s u l fu r i z a t i o n
D es u l fu r i z a t i o n
L i qui d Fu e l
S o l i d  f u e l
Syn ga s  
H
2
/ C O ~ 2
M e O H + D M E
S y nt he s i s  &  
P r e pa r a t i on
F T  s y n t h e s i s  &  
P r e p a r a t i o n
Re f o r mi n g  
H
2
+C O
W G S  
H
2
+C O +C O
2
C O  P r o x t o  
H
2
+C O
2
MC FC
650 - 700
o
C
P A F C  
< 2 %  C O
180 - 220
o
C
P E M F C  
< 1 0  ppm CO
70 - 90 
o
C
G a s C l e a ni ng  
D e sul f ur i z a t i o n
M e O H
S O FC
800 - 1000
o
C
Fu el  P r ep a r a ti o n Fu el  P r o c es s i n g Fu el  C el l
 
2 
 
Fig I.1. illustrates the general concept of processing gaseous, liquid and solid fuels for fuel cell 
applications. Reformate (syngas and other components such as steam and carbon dioxide) can be 
used as fuel for high temperature fuel cells such as Solid Oxide Fuel Cell (SOFC) and Molten 
Carbonate Fuel Cell (MCFC), for which the fuel needs to be reformulated. When natural gas or 
other hydrocarbon is used in Phosphoric Acid Fuel Cell (PAFC) system, reformate must be 
processed by water-gas-shift (WGS) reaction. 
 
Table I. 1 : The fuel requirements for the principal fuel cells 
Gas species PEMFC      AFC PAFC MCFC SOFC 
H2 Fuel Fuel Fuel Fuel Fuel 
CO Poison 
(>10ppm) 
Poison Poison 
(> 0.5%) 
Fuela Fuela 
CH4 Diluent Diluent Diluent Diluentb Diluentb 
CO2 and H2O Diluent Poisonc Diluent Diluent Diluent 
S 
(as H2S and COS ) 
Few studies Unknown Poison 
(>50ppm) 
Poison 
(>0.5ppm) 
Poison 
(>0.1 ppm) 
a. In reality CO reacts with H2O producing H2 and CO2 viz the shift reaction and CH4 and H2O reforms to 
H2 and CO faster than reacting as a fuel at the electrode 
b. A fuel in the internal reforming MCFC and SOFC  
c. The fact that CO2 is a poison for the AFC more or less rules out its use with reformed fuels. 
 
The lower the operating temperature of the stack, the more stringent are the requirements, 
and greater the demand placed on fuel processing as shown in Table I.1. The most promising and 
most widely researched, developed and demonstrated type of fuel cells is the proton exchange 
membrane (PEM) fuel cell, which operates at low temperatures (~ 80oC)  [1]. Hydrogen as a fuel 
is not readily available, particularly not for residential applications, except if the system is to be 
used as a backup power system, in which case it may be equipped with an electrolytic hydrogen 
generator. To facilitate market acceptance, fuel cell developers are forced to add a fuel 
 
3 
 
processing section to the fuel cell system. For residential and commercial applications, natural 
gas is a logical fuel choice because its distribution channel is widely developed. The majority of 
the stationary power fuel cell systems developed to date use natural gas as fuel[2].   
High efficiency desulfurization is critical to maintain the activity of fuel processing catalysts 
and high-value membrane electrode assemblies in logistic fuel cell systems. On-board fuel 
processing of liquid hydrocarbon fuel is being investigated to supply hydrogen for fuel cell-
based auxiliary power units. For such a system, if sulfur is not removed from the liquid phase, 
the removal of sulfur as H2S from the reformate becomes a key-step since downstream catalysts 
and the fuel cell itself can be poisoned by a small amount of H2S in the feed [3]. Hydrogen 
sulfide is present in many high temperature gas streams during extraction and processing of 
fossil fuels, natural gas and geothermal brines. H2S is also found in many industrial process 
gases, particularly in the mineral and metallurgical process industries. Because it is highly toxic, 
and corrosive, H2S must be removed completely as early in a process as possible. Depending on 
the fuel selection additional ancillary components are required for processing the fuel to meet the 
fuel requirement for fuel cell. Steam reforming catalysts, PEM anode catalysts and also the shift 
catalysts are intolerant to sulfur and to ensure adequate lifetime of fuel processors the 
desulfurization step is very important.  
 
4 
 
 
Figure I.2: Fuel processing steps with amounts of poisons tolerable for operation in PEMFCs 
Irrespective of the approach adopted to remove sulfur, following are some of the common 
requirements for sorbents used for logistical fuel cell applications: 
a. Achieving high levels of sulfur removal. Packed beds are used commercially for 
desulfurization to attain lower breakthrough concentration, higher bed utilization. 
These packed beds have larger size due to possible channeling and lower intra-
particle mass/heat transfer. 
b. Regenerability of the sorbent: temperature, energy requirement, purging gas, safety 
concern, valves and other utilities 
c. Scalability of the sorbent, ease of availability and cost 
d. Minimization of the system mass/volume and complexity 
This work is focused on development of sorbents which are regenerable, scalable over wide 
temperature ranges with uses in fuel cell systems. Attempts have been made to device 
< 1 0  
ppbv S
1 0  %  
CO
2 ppm
CO
G as o l i ne
C N G
D i e s e l
A l c o ho l  +
W a te r
D e s ul f ur i z e r
( 2 5 0 - 350 
o
C )
R e f o r m e r  
( 7 0 0 - 1100 
o
C )
W G S  R e a c to r  
( 2 0 0 - 450 
o
C )
P r e f e r e nti al  O x i di z e r  
( 1 5 0 - 250 
o
C )
P ow e r
P t :  <  1 - 10 ppm
C O  <  1 0  pp b S  
 
5 
 
appropriate strategies to reduce sulfur concentration to ppb levels in the reformate streams. Use 
of microfibrous entrapped sorbents as employed for benefits in the composite bed design is used 
to help miniaturize the desulfurization unit; this design has added benefit of higher breakthrough 
time without adding to pressure drop. 
I.2 Literature Review 
I.2.1 Desulfurization Technologies 
 
Sulfur removal from feed stocks usually takes place in two stages. The first stage involves the 
hydro desulfurization of organic compounds in the presence of hydrogen typically at 370 oC, 40 
Bars over CoO/MoO3/Al2O3 catalyst to generate H2S. The H2S is then absorbed in a bed of 
highly porous zinc oxide catalyst at 350-450 oC. 
 
Figure I.3: Desulfurization technologies classified by the nature of the key process to remove 
sulfur 
 
P h ys i co - ch e m i ca l  
s e p a r a t i o n / t r a n s f o r m a t i o n  o f S 
co m p o u n d s
C o n v en ti o n a l  HD S
C a t a l yt i c  t r a n s f o r m a t i o n  
w i t h  S e l i m i n a t i o n
H D S  by  adv anc e d 
c ata l y s ts
H D S  by  ad v an c e d 
r e ac to r  de s i g n
H D S  wi th f ue l  
s pe c i f i c at i on r e c ov e r y
C a ta l y ti c  d i s ti l l a ti o n
A l k y l a ti o n , P r e c i pi ta ti o n
E x tr a c ti o n , O x i d a ti o n
A d s o r p ti o n
D E S U L FU R I Z A T I O N
 
6 
 
Desulfurization by adsorption (ADS) is based on the ability of the solid sorbent to 
selectively adsorb sulfur compound from refinery streams. ADS is divided in two groups: 
Adsorptive desulfurization and Reactive adsorption desulfurization. Adsorptive desulfurization is 
based on physical adsorption of sulfur compounds on a solid sorbent surface. Regeneration of the 
sorbent is usually done by flushing the spent sorbent with a desorbent, resulting in a high sulfur 
compound concentration flow. Reactive adsorption desulfurization employs chemical interaction 
of the sulfur compound and the sorbent. Sulfur is fixed in the sorbent, usually as sulfide, and the 
S-free hydrocarbon is released into the purified fuel stream. Regeneration of the spent sorbent 
results in sulfur elimination as H2S, S, or SOx depending on the process applied. Efficiency of 
desulfurization is mainly determined by the sorbent properties: its adsorption capacity, selectivity 
for the sulfur compounds, durability and regenerability [4]. 
I.2.2 ZnO based sorbents 
The removal of H2S can be performed by different routes such as adsorption in liquid 
alkanolamine, ammonia solution and alkaline salt solution, oxidation with Fe (III) oxide and 
activated carbon. ZnO has been in use for H2S removal for more than 30 years. Among the tested 
metal oxides ZnO has the highest equilibrium constant for sulfidation, yielding H2S removal 
down to a fraction of 1 ppmv. Its principal limitation is that in the highly reducing atmosphere of 
synthesis gas it is partially reduced to elemental Zinc. It is volatile above 600 oC, with 
consequent sorbent loss. For achieving maximum useful life of the PEMFCs, it is crucial to 
reduce the H2S concentration to < 0.1 ppmv. Zinc Oxide is highly efficient desulfurizer due to 
favorable thermodynamics in the temperature range of 350-550 oC. ZnO shows low equilibrium 
H2S concentration. H2S absorption by ZnO is considered to be controlled by the following 
reaction:  
 
7 
 
)()()()( 22 gOHsZ n SgSHsZ n O  
This is an exothermic reaction and the equilibrium H2S concentration is determined by the 
temperature, the H2S partial pressure and to a lesser extent the phase of the zinc oxide. 
Equilibrium H2S concentration for ZnO with no H2O is shown in Fig I.4. The data is generated 
using the HSC software. Thermodynamically, it is impossible to reduce the sulfur concentration 
to less than 100 ppbv at temperatures above 300 oC. At lower temperatures of (< 250 oC), 
absorption kinetics are slower but the ZnS equilibrium is more favorable. The data in the table 
indicates that if the kinetics of H2S absorption is sufficiently rapid, concentrations well below 
100 ppb should be achievable.  
 
Figure I. 4: Equilibrium H2S concentration (ppmv) using HSC software. 
*HSC Chemistry Ver.3.0 Copyright ? Outokumpu Research Oy Pori Finland A.Roine. 
Intial H2S = 300 ppmv 
ZnO+H2S(g) = ZnS 
+H2O(g)  
0.515 
ppmv 
 
8 
 
Table I. 2 : Equilibrium data for ZnO+H2S = ZnS+H2O by HSC* software 
T (oC) Equilibrium constant (K) H2S outlet (ppmv) 
0 5.32E+13 4.11E-05 
200 7.60E+07 3.44E-02 
400 3.39E+05 5.15E-01 
600 1.82E+04 2.20E+00 
800 2.92E+03 5.45E+00 
1000 8.30E+02 1.01E+01 
 
 
 
 
Figure I. 5: Background on commercial ZnO sorbent based on literature search  
 
Based on the Fig I.5, there is a need of sorbents that can effectively remove sulfur in the lower 
temperature regime (T < 350C), high temperature regime (T > 550 C), regenerable over multiple 
cycles, COS tolerant.  The following sections present the literature review on the low 
temperature desulfurization, COS removal, high temperature desulfurization and Microfibrous 
entrapment for enhanced contacting efficiency in a packed bed.  
Z n O 
P o o r  r e ge n e r a b i l i t y  
U s e d  co m m e r ci a l l y 
3 5 0  C  <T<5 5 0  C
N o t  e ffi ci e n t  fo r  C O S r e m o va l  3 - 8
F a v o r a b le s u lfid a t io n
t h e r m o d y n a m ic s 10
? L a r ge  p a r t i cl e  s i z e 9
? l o w  co n t a ct i n g e ffi ci e n cy 
? i n t r a - p a r t i cl e  &  l a t t i ce  
d i ffu s i o n  l i m i t a t i o n s  
 
9 
 
I.3 Low Temperature Desulfurization 
 
There is an increasing need to purify gases at low temperatures (< 200 C) to improve the energy 
efficiency. Although this requirement has been partially met by the recent development of a 
high-surface area ZnO, there is still a scope for improvement[5]. Proton Exchange membrane 
fuel cells (PEMFC) have become the focus of significant interest for the stationary and portable 
power systems. A recent challenge is the development of the on-board fuel processing 
technologies utilizing high energy density commercial grade hydrocarbon fuels.  For Natural gas, 
Liquefied Petroleum Gas (LPG), which is the most suited for PEMFC applications (due to the 
highly developed infrastructure that exists for their distribution), ambient temperature removal of 
sulfur compounds using a solid adsorbent is technically attractive. There has been little work 
done on the H2S removal by metal oxides at room temperature. Activated carbons are also used 
for the H2S removal from natural gas or municipal sewage treatment facilities, because of their 
developed high surface area and large pore volumes. At room temperature, activated carbon was 
found to be a better adsorbent after extensive humidification of its surface[6]. 
Stirling et al [7] investigated different adsorbents such a ZnO and ZnO doped with 5% oxides of 
Cu, Fe and Co, high surface are Zn/Co/Al oxides and ZnCo oxides with different Co/Zn ratios 
for the H2S removal at room temperature.  Among the oxides studied, they reported the Co2O4 
oxide to be the best one because it showed almost stoichiometric reaction with H2S. Davidson et 
al [8] studied the rate of reaction of H2S with high surface area of undoped and doped ZnO 
samples at 0-45 oC and they reported the fast rates appeared to depend upon the crystallite size, 
morphology and coexisting water. Addition of the dopant not only stabilizes the active sorbent 
by increasing active surface area and decreasing crystallite size, but may introduce defects with 
promoter behavior [5, 7]. Small particle size (100-200 microns) allows entrapment in the 
 
10 
 
microfibrous media viable for composite bed design. The microfibrous media (developed in 
Auburn University) offer enhanced contacting efficiency and mass transfer without significant 
pressure drop [9-11]. This approach towards small-scale regenerable continuous batch fuel 
processing in PEMFC applications is commercially feasible by using micro structured particulate 
carriers. 
I.4 COS Removal/Inhibition 
 
With the introduction of the strong legislation to reduce sulfur emissions, fresh impetus is 
being given to modifying improving existing desulfurization technology. However, 
dehydrodesulfurization does not remove or significantly affect sulfur containing compound, 
namely, Carbonyl Sulfide (COS). Various researches for H2S removal have been reported in 
details for the purification of gasified products derived from various feedstocks [12]; however, 
removal of COS is not a big concern yet, because it is not the major sulfur compounds produced 
from gasification. The absorption of H2S by ZnO is stoichiometric above 350 oC but it falls 
rapidly at lower temperatures. The removal of COS has been reported to be more difficult at low 
temperatures in the range from room temperature to 200 oC than H2S. ZnO is a preferred metal 
oxide because of favorable sulfidation thermodynamics, [13] but is not efficient to remove COS 
[7]. Most commercial H2 is produced form natural gas via steam methane reformation (SMR) 
followed by a water gas shift (WGR) reaction in which CO is oxidized to CO2 while water is 
reduced to H2. The gas effluent from the WGS varies from a few ppmv to 2% by volume of CO 
in excess of H2. This low concentration of CO in the H2 outlet stream from the WGS can be 
avoided. However, eliminating the CO is beneficial in increasing the PEM fuel cell performance. 
Over the last two decades it has become increasingly apparent that emissions of sulfur 
 
11 
 
compounds, including COS, into atmosphere have been unacceptably high. Since COS is rather 
inactive compared to H2S probably due to its neutrality and similarity to CO2, COS is sometimes 
produced through the reaction of H2S with CO2, although the reaction can be reversible to 
produce again H2S and CO2 from the reaction of COS and H2O depending upon the adsorption 
conditions[14].  
The formation of COS is primarily governed by the reversible hydrolysis reaction and 
equilibrium conditions present:  
222 COSHOHC O S  
Removal of sulfur-containing compounds is one of the most important technologies for the 
utilization of the gasified products derived from various feedstocks such as biomass, waste and 
solid fossil fuels. Gaseous sulfur compounds of H2S and COS are severe catalyst poisons against 
the following processes of steam reforming for hydrogen. COS can be formed by the conversion 
of H2S and CO2 in the absence of water. Natural gas is saturated with water and therefore COS 
does not usually occur in those streams. A relatively small volume of COS can combine with 
water to form H2S if suitable equilibrium conditions exist[14]. 
 
 
 
 
I.5 High Temperature Desulfurization 
 
The removal of hydrogen sulfide to sufficiently low levels from coal derived fuel gases at 
elevated temperatures is crucial for efficient and economic coal utilization in emerging advanced 
 
12 
 
power generation systems such as integrated gasification-combined cycle (IGCC) and the 
gasification molten carbonate fuel cell (MCFC). Gasification is expected to be among the most 
promising  conversion processes to produce synthesis gases[15]. Integrated coal gasification 
combined cycle (IGCC) is one of the most prospective coal-based power generation technologies 
in this century because of its high efficiency and low emission. Before the gas goes into the gas 
turbine combustor, the polluting species and other contaminants in the raw gas must be removed, 
including dust, sulfur species, nitrogen species (ammonia and cyanides), halides and trace 
metals,  the high temperature gas cleanup system is then introduced and the gas sensible heat can 
be fully utilized with and increase in the efficiency by 0.5-1.5%.[16] 
The important factors of hot gas desulfurization sorbent are 
1. The sorbent should have good sulfur removal capacity and fast adsorption kinetics  
2. The sorbent should be chemically stable,  i.e., it should not evaporate or sinter 
during regeneration 
3. The sorbent should be physically stable, i.e., it should withstand any attrition. 
4. The sorbent should catalyze formation of elemental sulfur upon reductive 
regeneration, hydrolyze carbonyl sulfide (COS), and react with other 
contaminants such as tars 
5. The sorbent should be regenerable and it should maintain its sulfur removal 
capacity for many cycles 
6. The sorbent replacement cost should be affordable.  
Hot gas desulfurization can be accomplished by using metal oxide based sorbents like zinc, 
manganese, iron and copper. Typically, metal oxides are converted to sulfides during a sulfur 
loading stage under reducing hot gas conditions.  
 
13 
 
 
For optimal IGCC performance, high temperature coal gas purification technology is 
necessary. The sulfur present in the coal is converted primarily to H2S in the gasifier with a small 
amount of COS. Several liquid scrubbing processes are available for H2S removal to achieve the 
20-ppmv target. These processes, however, do not integrate well with IGCC due to large 
temperature differences. There will be energy losses associated with cooling to scrubbing 
temperatures, which is almost ambient temperature. Therefore, hot gas desulfurization is critical 
to the optimal development of the IGCC and the other advanced coal gas processes.  
ZnS can be regenerated if sufficiently high temperatures or low oxygen concentrations are 
used to avoid zinc sulfate formation. Zinc loss in the form of vapors limits the application of the 
ZnO at higher temperature. Mixed metal oxides allow raising the operating temperature for fuel 
gas desulfurization as a result of their lower ZnO reduction rate. Zinc-based and ferrite based 
sorbents show superior reactivity to H2S. Structural stability and good mechanical strength are 
additional desired features of the sorbents.  
Process requirements taken as a basis for determining metal oxides suitability for high 
temperature desulfurization: 
1. Rate of desulfurization and stability of the sulfide under reducing gas conditions 
2. Potential for detrimental secondary reactions on the solid under reducing coal 
gases ( e.g. metal carbides, reduction to zero-oxidation state, formation of 
chlorides from HCl) 
3. Rate of regeneration and production of SO2 or elemental sulfur under oxidizing 
gas conditions 
 
14 
 
4. Potential for detrimental secondary reactions on the solid under oxidizing gases 
(e.g. sulfates) and hydrothermal stability during regeneration. 
 
I.6 Microfibrous Entrapped Sorbents 
 
 Microfibrous technology developed at the Center of Microfibrous Materials 
Manufacturing (CM3) at Auburn University [17-26] provides a novel approach for a versatile 
design of small, efficient, and lightweight fuel processors. This approach can also enhance 
heat/mass transfer, improve contacting efficiency and promote regenerability. Packed beds that 
provide enough volume to remove sulfur from several 1000 ppmv to sub-ppmv levels generally 
use sorbent particles sizes, ca. 1-5 mm. These demonstrate low sorbent utilization and poor 
regenerability, owing to low contacting efficiency, intra-particle and lattice diffusion limitations 
[27]. Small particle size (100-200 ?m) allows entrapment in the microfibrous media viable for 
composite bed design. These  microfibrous media offer enhanced contacting efficiency and mass 
transfer without significant pressure drop [11]. 
The fabrication of the microfibrous media is based on reliable, proven, high speed roll to roll 
papermaking and sintering processes, which substantially reduces the production costs and 
improves the product quality. This approach utilizes micro-sized fibers to entrap sorbent and/or 
catalyst particulates into a sinter-locked microfibrous structures with a high voidage and high 
contacting efficiency. With improved contacting efficiency, these materials can reduce both the 
reactor weight and volume, which is very important for logistic fuel processors[28]. For 
example, microfibrous entrapped Ni/Al2O3 catalysts for toluene hydrogenation in a trickle bed 
reactor demonstrated 2-6 times higher specific activities than conventional packed bed catalysts 
 
15 
 
on a gravimetric basis, while volumetric activities of 40 vol% composite catalysts were 80% 
higher than conventional extrudates[29] . Microfibrous entrapped promoted Pt /Al2O3 catalysts 
for PrOx provided 3 times higher bed utilization efficiency compared to packed beds of 2-3 mm 
(dia.) pellets [25, 30] at same CO conversion. Two-phase mass-transfer experiments indicated 
that microfibrous composite catalysts take advantage of both high gas-liquid contacting and bulk 
mixing at low pressure drop with the potential to provide enhanced catalyst utilization. 
Additionally, the microfibrous media can be made into thin sheets of large area and/or pleated to 
control pressure drop and contacting efficiency. As for H2S removal, Ni fiber entrapped 
ZnO/SiO2 was prepared and demonstrated 3 times longer breakthrough time than a commercial 
ZnO extrudates [31],[24]. However, Ni fiber cannot sustain the high oxidizing atmosphere 
during ZnO regeneration. Therefore, new microfibrous entrapped sorbents with microfibrous 
structures that are able to work in both reducing and oxidizing environments were developed 
using the sintered ceramic/glass carriers with micro-sized ZnO entrapped for regenerable use to 
scavenge bulk H2S from reformate streams in a continuous batch mode at 400 oC [20]. Based on 
the thermal properties of various types of glass fibers available (Advanced glass fiber yarns 
LLC) S2 fiber and E type fiber were chosen as shown in Table I.3 and the SEM image of the 
glass fibers entrapped silica is shown in Fig. I.8. 
 
 
 
 
 
 
16 
 
Table I. 3: Properties of the Glass fibers 
Glass 
fiber type 
Dimensions 
Dia (um) x length 
(mm) 
Density 
(g cm-3) 
Softening 
point (oC) 
Annealing point 
(oC) 
Strain point 
(oC) 
S2 8 x 6 2.46 1056 810 760 
E 10 x 6 2.58 846 657 615 
 
 
Figure I. 6: SEM Image of the S2 glass fiber entrapped SiO2 particle. [20] 
I.7 Advantages of supported sorbents 
 
Supported types of sorbents are generally preferred for their mechanical strength because 
sufficient amount of cycles of sulfidation and regeneration are desirable in either fixed-bed form 
or fluidized bed systems. The support materials of sorbent applied for hot coal gas 
desulfurization is presently composed primarily of Al2O3 and SiO2, activated carbon, or other 
materials.  
 
 
17 
 
 
Figure I. 7: Important properties for sorbent formulation 
 
1. In order to improve desulfurization performance, sorbents with high porosity and 
small grain sizes are preferred.  In this regard, metal oxide sorbents on inert supports 
are widely used for desulfurization.   
2. In supported sorbents, active sorbent substances are supported on secondary oxides 
to form high surface area and high porosity sorbent particles/extrudates.  These 
secondary compounds are mainly inert to sulfur.   
3. Supports are utilized to enhance the structural stability for the active sorbent and to 
adhere/hold the sorbent crystallites within the micropores of the support in the 
absence of grain size, agglomeration and sintering. 
4. Supports also serve to stabilize the active metal oxide component against chemical 
reduction and vaporization. The supported sorbent design also facilitates the 
incorporation of the sorbent into process system hardware. 
So r b en t w i th  i m p r o v ed  
p er f o r m a n c e
N a tu r e o f  
a c ti v e s p ec i es
Su p p o r t 
C h o i c e
P r ep a r a ti o n  
r o u te
 
18 
 
5. Due to the above noted advantages provided by supported sorbents, these systems 
provide stable performance with extended service lives.  
The most extensively used porous adsorbent materials for desulfurization include activated 
carbon, -alumina, modified zeolites, etc. [32, 33]. However, these materials suffer several 
drawbacks in practical application as shown in Fig.I.7.  
 
Figure I. 8: Overview of literature review on various supports 
Silica supported ZnO sorbents are preferred because of better ZnO utilization and thermal 
stability [1-4]. Recent developments on improvement in the diffusion of reactants to the catalyst 
sites have been focused on the increase in Zeolite pore sizes decrease in zeolite crystal size and 
providing an additional mesoporous system within the microporous crystals. 
The discovery of highly ordered mesoporous materials has received considerable attention 
in heterogeneous catalysis. These materials have much promise for the development of novel 
solid catalyst due to their structural characteristics such as ordered pore structure, high specific 
surface area (1000-2000 m2/g), uniform pore size distribution (varying from 1.5-10 nm), and 
high specific pore volume (1.0-2.0 cc/g) [34]-[35]. The mesoporous silica supports including 
? L o w  me c h a n i c a l  s t a b i l i t y
? F o r ma ti o n  o f  f i n e s
? H i g h  tor tu o s i ty  w i th  l a r g e  a mo u n t o f  mi c r o p o r e s  h i n d e r s  
f u l l  a c c e s s i b i l i ty  o f  th e  r e a c ta n ts  to a c ti v e  s i te [ 1 8 ]
A c t i v at e d  
C arb o n  
? C h em i c a l  i n t er a c t i o n  b et w een  t h e s u p p o r t  a n d  a c t i v e p h a s e,  
l ea d s  t o  d ec r ea s e i n  c a t a l y s t  p er f o r m a n c e.  
? N a r r o w  a n d  u n i f o r m  p o r e s i z e p o s e m a s s  t r a n s f er  l i m i t a t i o n s  
f o r  l a r g e r ea c t a n t  m o l ec u l es  [ 1 9 ]
A l u m i n a
? S en s i t i v e t o  m o i s t u r e c o n t en t
? D eg r a d a t i o n  a t  h i g h er  t em p er a t u r es
Z e o l i t e
 
19 
 
MCM-41, MCM-48, SBA-15 are found to be the superior base matrix for various surface 
modifications with amines and their subsequent application in the low temperature removal of 
acidic gases like H2S and CO2 [36, 37]. Mesoporous silica has been recently used as a support 
for metal catalysts, resulting in significant improvements when compared to commercial and 
conventional amorphous silica-alumina catalysts. For example, Corma et al. [38] have reported a 
superior hydrogenation activity and sulfur tolerance for Pt/MCM-41 in comparison to Pt/Zeolite, 
Pt/SiO2, and Pt/Al2O3. Song and Reddy [39] reported that Co-Mo/MCM-41 showed higher 
hydrogenation and hydrocracking activities than conventional Co-Mo/ -Al2O3. Fe2O3/MCM-41 
was found to exhibit a superior performance for the conversion of SO2 into SO3 compared to 
Fe2O3/silica [40]. However, on reviewing literature, it is worth to mention that only little 
information is reported so far on the MCM-41 and MCM-48 application for H2S uptake. It is in 
this context only that in this study the commercially procured MCM-41 and laboratory 
synthesized MCM-48 were selected as H2S sorbent and their performances were compared with 
conventional silica gel. MCM-41 is the most extensively studied member of the mesoporous 
silica materials family. It exhibits a hexagonal array of one dimensional mesoporous which can 
be tuned from 2-10 nm by the suitable choice of the structure directing agent and synthesis 
conditions [41]. The mesoporous materials are in general synthesized by supra-molecular self-
assembly process in the presence of (cat-) ionic surfactants as templates during the mesophase 
formation.  
 
I.7.1  Novel Support ? Mobil composition of Matter- MCM-41 
The Table I.4  lists the comparison of the various method used for preparation of the supported 
metal oxide on MCM.  
 
20 
 
Table I. 4. Comparison of the literature Review on preparation of MCM- 41 
Sorbent 
/Catalyst 
Method Application Remarks Author/Year 
Cr-MCM-41 In-situ Olefin Oligomerization to produce lube oil 
Pour points and 
Viscosity 
index 
improved 
Pelrine et 
al.,1992[42] 
Co- MCM-41 In-situ HDS and HDN High activity at Si/Al= 60 Souza etal., 1995[43] 
MCM-41 N/A Acid catalyst for Friedal Crafts alkylation Good catalytic activity Kloeststra et al.,1995[44, 45] 
Ti (V, Cr)-
MCM-41 In-situ 
Oxidation in presence of 
hydrogen peroxide 
Excellent catalytic 
oxidation 
Tanev et al., 
1994[46] 
Ti,V-MCM-
41 In-situ 
Selective catalytic 
reduction of NO 
Higher NOx 
conversion than for 
silica based catalyst 
Beck et al., 1992[47] 
Cu-MCM-41 Organofunctionalization Not studied Characterization for metal loading Hao etal., 2006[48, 49] 
Fe-MCM-41 
Incipient wetness, solid 
state impregnation, in-
situ 
SO2 oxidation at  high 
temp (>600K) 
Wetness 
impregnation was 
better 
(conversions~60-
70%) 
Wingen et al., 
2000[50] 
Ag-MCM-41 Direct Hydrothermal & thermal ion exchange CO oxidation 
Reduction at 
500oC? oxidation 
at 
RT(~95%conversio
n) 
Gac et al., 2007[51] 
Zn-MCM-41 HIP Hydrogenation of MB At 300-400
oC- 
100% conversion Lu et al., 2002[52] 
ZnO, CuO 
MCM-41,  
MCM-48 
Incipient wetness 
method 
Low temperature H2S 
Adsorption-desorption 
NOT EXPLORED 
MUCH NOT REPORTED 
 
 
.Incipient wetness impregnation method as used to make metal oxide supported on MCM 
sorbent. Very few studies are reported on the metal incorporation in mesoporous silica by 
incipient wetness impregnation. Incipient wetness impregnation[53] is a simple method with 
fewer steps; an adequate amount of active metal can be loaded on the support by changing the 
precursor concentration, the oxide formed after calcination is stable. The sorbents supports can 
disperse the active components and increase the surface area of sorbents. Some support materials 
such as carbon material may also play roles in converting sulfur species. The quantity of support 
 
21 
 
materials is large in industry scale systems; therefore, the support material must be economic and 
easy to be obtained.  
 
  
 
22 
 
I.8 Scope and Objective of the work:  
 
 
Figure I. 9: Schematic diagram of the fuel processing for PEMFCs with average operating 
temperatures 
Fig.I.9 shows the fuel processor system with average operating temperatures, the 
desulfurizer can be located either before or after the HTWGS unit[54]. In the first case, high 
temperature sorbent will be required for protection of the HTWGS catalysts against sulfur 
poisoning. However for the second case, a low temperature ZnO based sorbent for protection 
against the most sulfur sensitive catalysts of the fuel processor (LTWGS and CO-PROX 
catalysts).  Typical sulfur compound include RSH, R2S, H2S and COS. COS is particularly 
problematic to remove as commercial sulfur adsorbents generally show poor adsorption 
capacities for COS at ambient temperature, and thermodynamic constraints limit COS removal 
via conventional hydrotreating. Sulfur impurities can reduce the effectiveness of fuel-processor 
R e f or m i ng
L T W G SCO - P RO X
D e s ul f ur i z e r
HT W G S
Fu el  C el l
710 
o
C 470 
o
C 250
o
C
280 
o
C150 
o
C80
o
C
Fu el
E x ha us t
 
23 
 
catalysts and can poison the anode catalysts of both high- and low-temperature fuel cells. The 
problem is most severe in polymer electrolyte fuel cells (PEMFC); because they operate at low 
temperature and their Pt group catalysts are susceptible to sulfur poisoning. The poisoning 
effects of sulfur are irreversible. PEMFCs operate at low temperatures ~80oC, an inline filter can 
be developed which takes care of removal of sulfur from several ppmv level to sub-ppmv level. 
During the cold start-ups of FC system, the temperature drops to less than 30oC, at this time we 
need an efficient sorbent which can operate over wide temperature range.  Regardless of initial 
H2S concentration, subsequent replacement of the contaminated fuel stream with pure H2 does 
not allow full recovery of the catalyst. Sulfur also degrades the performance of the high-
temperature solid oxide fuel cells (SOFC). The performance of the SOFC drops about 15% in the 
presence of 1 ppmv sulfur. The cell voltage increases, and performance is recovered once the 
sulfur flow is stopped[55]. Although this poisoning effect is reversible in SOFCs, long-term 
stable electrochemical performance of both high- and low temperature fuel cells requires that the 
sulfur concentration to be reduced to sub-ppmv levels. 
Accordingly the objective of this work is divided in four major parts:  
1) Low temperature Desulfurization  
2) COS Removal 
3) High Temperature Desulfurization 
4) Microfibrous entrapped sorbents  
 
 
 
 
24 
 
I.9 Objective of this work  
 
 To develop a sorbent for wide temperature range ( 20 ? T ? 550 C) gas phase sulfur 
removal (H2S & COS) 
 To develop a process that is efficient, cost-effective and scalable.   
 To develop sorbents that work efficiently without significant activity loss upon multiple 
regeneration cycles for logistic Fuel Cell systems. 
 To attain high levels of sulfur removal by employing various support characteristics 
 To test the sorbent efficiency for use in hot gas desulfurization for applications in SOFCs  
 To remove COS by employing various strategies  
o hydrolysis of COS  
o inhibition of COS by varying test conditions 
 To propose various schemes to eliminate sulfur from the fuel stream by developing 
sorbents that are 
o regenerable  
o non ?regenerable  
 To characterize the sorbents synthesized  in lab to understand the reaction mechanisms 
To study the effect of kinetic parameters on MFES  
 To establish a composite bed design for miniaturization of the desulfurization unit 
 
 
 
 
 
25 
 
I.10 Outline of this work 
 
 Chapter II describes the general experimental section and the characterization 
techniques used in this study.  
 Chapter III discusses the study on the wide temperature range promoted 
ZnO/SiO2 sorbents and effect of various parameters like type of support, 
promoters, promoter concentration, temperature, moisture content, presence of 
reformate streams (with CO, CO2) and advantages of microfibrous entrapment 
over packed bed and study of kinetic parameters for the same[56].  
 Chapter IV discusses the adsorption and multiple cycle regeneration performance 
of Cu0.05Zn0.95O/SiO2 and its characterization to understand the role of active sites 
using techniques like XPS, ESR and N2 adsorption-desorption isotherms[57]. 
  Chapter V focuses on the preparation and performance of the novel bimetallic 
doped Mn0.025Fe0.025Zn0.95O/SiO2 for wide temperature range H2S removal from 
the fuel reformate streams coupled with the XPS and Operando ESR studies to 
better understand the role of the dopants in ZnO/SiO2 [58] 
 Chapter VI discusses the strategies to mitigate COS present/formed in reformate 
streams. The chapter focuses on preparation and performance on carbon and 
alumina based sorbents to remove, inhibit and hydrolyze COS over wide 
temperature range. It also discusses the room temperature hydrolysis and removal 
of COS from fuel reformates using Al2O3/carbon and Mn0.025Fe0.025Zn0.95O/SiO2 
layered beds [59]. 
  
 
26 
 
Chapter II: Experimental Setup and Characterization Techniques 
 
 
II.1 Sorbent Preparation 
II.1.1 Sorbent for Packed Bed 
 
II.1.1.1 Preparation of doped supported sorbent 
 
The doped ZnO-based sorbents with the formula M0.05ZnO0.95/SiO2 (M=Mn, Fe, Co, Ni, Cu) 
were prepared by an incipient impregnation of the commercial high surface area silica (Fischer 
Scientific Inc., surface area ~550 m2/g, powder 100-200 ?m), with metal nitrates as the 2 M 
solutions in water used as precursors. Total metal loading was 15, 21 or 25 wt. %. The metal 
oxide loading was confirmed by Inductively Coupled Plasma Spectrometer (ICP) analysis. After 
impregnation and drying, the samples were calcined in air at 350 oC; these are referred to as the 
?calcined? samples. Different types of silica supports including MCM-41, MCM-48 silica and 
conventional silica gel were used. The un-promoted and promoted silica supported catalysts were 
prepared by incipient wetness method. The metal nitrate solutions of different transition metals 
were used as precursors for impregnating the MCM (-41 and -48) and silica support with 
different ZnO and doped ZnO loadings. The impregnated supports samples were dried at 100oC   
 
27 
 
for 6h and subsequently calcined at 350oC for 1 h under air flow. All the samples were stored 
in desiccators for further use.   
 
28 
 
The promoted ZnO-based desulfurization sorbents of the nominal formula FexMnyZnO1-x-
y/SiO2 (x, y=0; 0.025) were prepared by incipient co-impregnation of high surface area (300-550 
m2/g) silica (Fischer Scientific Inc.) of grain size 100-200 ?m with solutions of nitrates of the 
respective metals in water, namely Zn(NO3)2, Mn(NO3)2 and Fe(NO3)3. Single step incipient 
impregnation was performed on the silica support to achieve metal oxide loading of 12-36% by 
varying the molarity of nitrate solutions. Upon incipient impregnation and drying, the samples 
were calcined in the flowing air at 350-550 oC; these are referred to as the ?calcined? specimens. 
The specimens prepared as above, excepting the calcination step, are referred to as the ?dried? 
sorbents. In the reference experiments, with the commercial H2S sorbents (BASF SG-901 and 
Sud Chemie G-72E), they are crushed to the same particle size as that of the silica (100-200 
microns) used to prepare the supported FexMnyZnO1-x-y/SiO2 sorbents. 
 
II.1.1.2 Preparation of Mesoporous type silica (MCM) 
 
The MCM-41 was procured from Sigma Aldrich and used as-received without any further 
purification. The MCM-48 was prepared as described by Schumacher and co-workers elsewhere 
[60], 10.4 g of cetyltrimethyl ammonium bromide (CTAB, Aldrich) was dissolved in 480 ml of 
water and 200 ml of absolute ethanol (99.5%, Aldrich). 48 ml of ammonia solution (32%, 
Aldrich) was added to the mixture and allowed to constantly stir for 15 min. Then, 13.6 g of 
tetraethoxysilane (TEOS, 98%, Aldrich) was added, and the whole mixture was constantly 
stirred at room temperature for 10 h. The obtained white suspension was then filtered, washed 
with hot distilled water and dried at 100 oC for 12 h. The white powder was then calcined at 550 
oC in air for 10 h to obtain the MCM-48 as shown in Fig II.2. 
 
29 
 
Different types of supports including Alumina, Titania, ACP-carbon, MCM-41, MCM-48 
silica and conventional silica gel were used in this study. The MCM-41 used in this study was 
procured from Sigma Aldrich and used as-received without any further purification.  
 
 
 
Figure II.1a: Preparation method of MCM- 48[40, 48, 60] 
 
 
30 
 
 
 
Figure II.1b: Preparation method of MCM- 48[40, 48, 60] 
II.1.1.3 Preparation of Al2O3/Carbon 
 
Activated PICA carbon of particle size 100-200 microns was dried in oven at 100 C. The 
dried Carbon was then impregnated with 2M Aluminum nitrate. The impregnated sample was 
then dried in air for 6hrs and then calcined at 300 C for 1h. The calcined sample Al2O3/C is ready 
to test after cooling it down to room temperature.                                                                                                          
II.1.1.4 Glass fiber entrapped Sorbent preparation 
 
 Glass fiber entrapped sorbents were made by the wet-lay paper making procedure. 
Sintered microfibrous carrier was used to entrap 150-200 um diameter support particulates, 
 
31 
 
where SiO2 was chosen as support and ZnO was then placed on the supports by incipient wetness 
impregnation. 6g of S2 glass fibers (8 microns diameter) and 2g cellulose were added in water 
and stirred vigorously to obtain uniform suspension. The suspension and 18g of silica particles 
were added into head box of 1ft2 M/K sheet former aeration. The preform (1ft2) was then formed 
by filtration and drying. The glass fiber sheet was pre-oxidized in airflow for 30 min at 450oC 
and then sintered for 1h at high temperature, ca, and 910oC. The prepared microfibrous 
entrapped SiO2 was immersed into zinc nitrate solution (2mol/L) for 15min, and then vacuum 
dried and naturally dried overnight and then calcined at 350oC for 1h in air.  
II.2  Pressure drop measurement set-up 
 
To study the pressure drop effect, the setup as shown in the Fig.II.2 was used. It consisted of 
the differential pressure cell. The setup shown in Fig II.2 was used to measure the pressure drop 
across the reactor bed. Two sets of measurements were conducted on the microfibrous media as 
indicated in Chapter III (Section III.5). Effect of change in face velocity on the pressure drop and 
effect of change of media (change in solid loading) on pressure drop was studied.   
 
 
 
32 
 
 
Figure II. 2: Schematic diagram of the pressure drop measurement 
 
MOC ? Material Of Construction  
S.S ? Stainless Steel  
Reactor tube ? Quartz ? I.D ? 0.45? O.D- 0.5? 
To G.C for analyses/Inlet gases ? 1/8? S.S  
a- Length of reactor tube ? 23? 
b- Distance between reactor and inlet pressure tap (ultra torr fittings) ? 2.5? 
c- Length of bed + Glass wool ? 1? 
d- Length of glass beads ? 9? 
(x+y) ? Length of inlet pressure tap- (MOS ? ?? S.S) ? 37? 
z- Length of outlet pressure tap ? (MOC- ??S.S) -7?  
 
Fu
r
n
a
c
e
Fu
r
n
a
c
e
x
y
z
a
b
c
d
P r e s s u r e  c e ll
T o  G C  fo r  a n a ly s is
I n le t  g a s e s
 
33 
 
II.3 Experimental Procedure 
In the desulfurization experiments, the challenge gas was the model reformate with an inlet 
concentration of 1 or 2 vol. % H2S, 33 vol. % CO or CO2, balance H2 (UHP grade from Airgas 
South, Inc.), at a face velocity of 1900 h-1 , corresponding to the volumetric gas flow rate of 0.1 
slpm. The challenge gas was passed through the sorbent in the packed bed inside a vertically-
mounted quartz tubular reactor (10 mm I.D. x 30 mm long), coaxially located inside a 200 mm 
long tubular furnace. In the reactor, sorbent weight was 0.5 g, bed size was 10 mm (dia.) ? mm 
(thickness). The samples upon adsorption of H2S are referred to as the ?sulfided? samples. H2S 
uptakes during the dynamic adsorption experiments were measured using a gas chromatography 
(GC) instrument (Varian CP3800) equipped with the thermal conductivity detector (TCD) and 
pulse flame photometric detector (PFPD). 
II.4 .Adsorption experiment 
 
 The adsorption experiments for desulfurization were carried out at ambient conditions 
(20 oC, 1 atm) as shown in Fig II.3. It is comprised of three major sections gas supply section, 
Reactor system and analysis section. A vertical quartz made reactor (10 mm I.D. x 30 mm L) 
coaxially mounted in a 200 mm long tubular furnace. The temperature of the furnace during 
desorption experiments was controlled using a PID temperature controller. The gas flow rates 
were controlled by mass flow controllers (Omega FMA 2405, Alaborg GFC1718). The face 
velocity (GHSV) of the stream is 1900 h-1, corresponding to volumetric gas flow rate of 0.1 
slpm. An inlet concentration of 1 % (v/v) H2S in H2 (ultra high purity grade; from Airgas South, 
Inc.) was used as sulfur source at a face velocity of 2.12 cm/s, corresponding to the volumetric 
gas flowrate of 1900 h-1 GHSV (0.1 slpm). The desulfurization reactor contained 0.500 g 
 
34 
 
sorbent; the sorbent bed size was 9 mm in diameter and 10 mm thick. Gas supply system consists 
of two H2S/H2 gas cylinders of 2vol% and 321ppmv concentrations. UHP H2 was utilized to 
dilute the H2S gas concentration. COS/N2 procured from Matheson Tri-gas was used in the 
experiments where COS was used as challenge gas. UHP N2 was used to eliminate traces oxygen 
in the reactor during the experiment and to dilute COS concentration. UHP He was used as a 
inert gas to eliminate traces of Oxygen in the reactor. CO (99.5%) and UHP CO2 were used as 
challenge gas to mimic the reformate streams composition, to investigate the COS formation and 
also to study their effect on the sorbent.  H2S uptakes during the dynamic adsorption experiments 
were measured using a gas chromatography (GC) instrument (Varian 3800) equipped with 
thermal conductivity detector (TCD) and pulse flame photometric detector (PFPD). . Varian GC 
3800C equipped with three detectors TCD, PFPD and FID was used. TCD was utilized to 
analyze outlet gases, specification and details are mentioned in the appendix II.  
A gas bubbler/ vaporizer was used to saturate the gas streams to study the effect of water or 
moisture on the system. There was also a provision to heat the bubbler to study the effect at the 
various moisture contents in the bed. Water was introduced in system by passing He or H2 
through the vaporizer with a temperature controller and was carried in a 1/8? stainless steel 
tubing wrapped with heating tape. This stream containing water was then mixed with H2S stream 
before entering the reactor.  In each adsorption run, 0.5 g sample was packed in the reactor. In 
this study, the breakthrough time was defined as the time from beginning of the desulfurization 
to the time when the H2S concentration at the exit reached 100 ppmv.  The specimens of the 
sorbents upon adsorption of H2S are referred to as the ?sulfided? samples. 
 
 
35 
 
 
Figure II. 3: Schematic diagram of the experimental set-up 
Regeneration of the ?sulfided,? i.e. ?spent? sorbent was performed in-situ in the sulfidation 
reactor at 550 oC in air at a flow rate of 950 h-1. Househood air was used to regenerate the 
sorbent bed. The temperature of the furnace during the experiments was maintained using a PID 
temperature set point controller. The gas flow rates were controlled by mass flow controllers.  
The Reactor system mainly consists of the quartz reactor tube, the dimensions and the 
structure is shown in the Fig.II.2. The dimensions of the reactor tube were 16-19? length and 
0.5? I.D. The glass beads of size 4mm diameter from Fischer scientific were used to support the 
bed. The bed consisted of two layers of glass wool about 0.25cm length on the upstream and 
downstream ends of the sorbent bed as shown in the Fig II.4. These layers of glass wool ensured 
uniform gas flow through the sorbent bed and supported the particles in the sorbent bed from 
moving. The sorbent was loaded 9? from the bottom of the tube.  
 
36 
 
 
                                 
 
Figure II.4: Schematic Diagram of the Configuration of the Reactor Bed 
 
Stainless steel tubing of ?-1/8? was used in the set-up. The tubes and the fittings were 
replaced every 6 months to ensure no clogging has taken place. Leak detection was always 
performed using snoop - soap solution to ensure adequate and desired flow of the gases into the 
system.  
 
II.4. Analytical/Characterization Techniques  
 
N2 Adsorption Desorption Isotherms 
Nitrogen adsorption/desorption isotherms at 77 K were measured using the Autosorb 1-C 
instrument from Quantachrome Instrument Corp., USA. Prior to the measurement, all samples 
were degassed for 10 h at 200 ?C. Specific surface area, SBET was calculated using the BET 
Glass beads support  
Sorbent Bed  
Quartz tube  
ID ? 10 mm length 30 mm  
Glass wool   
 
37 
 
equation. The total pore volume, VP was calculated at P/P0 = 0.95. The pore width, Pw 
distribution over the range of ~2?80 nm was generated from the adsorption branches of the 
isotherms via the BJH method, and the calculations were performed using the Autosorb 1-C 
software for Windows from Quantachrome Instruments. 
Scanning Electron Microscopy (SEM) 
The surface morphology of the MCM samples before and after metal impregnation was 
investigated with Scanning Electron Microscopy (SEM). Prior to SEM (Zeiss Digital Scanning 
Microscope DSM940), the samples were vacuum coated with gold (Pelco SC-7 auto sputter 
coater). 
X-Ray Diffraction(XRD) 
XRD patterns were obtained using a Rigaku Miniflex diffractometer at room temperature. 
Diffraction patterns were obtained with the Ni-filtered Cu K? radiation (  = 0.15418 nm) using a 
scanning speed of 1 o/min. The resultant XRD patterns were compared with those from the 
standard commercial XRD database. 
X-Ray Photoelectron Spectroscopy(XPS) 
X-ray Photoelectron Spectroscopy (XPS) was performed using the Leybold-Heraeus LHS-10 
instrument. The sample of the sorbent of ca. 200 mg was pressed into a pellet 16 mm dia. by a 
hydraulic press. The resultant pellet was loaded to the High Vacuum ?loadlock? chamber (base 
pressure ~10-6 Torr), with the subsequent transfer to the high-vacuum (HV) XPS measurement 
chamber (10-8-10-7 Torr). In XPS, the non-monochromated Mg K? line with hv=1253.6 eV or Al 
K? line with hv=1486.6 eV was used, and spectra were fitted by the XPSPEAK program. 
Sample charging effects were compensated by adjusting the XPS instrumental settings, until the 
Binding Energy (BE) of C 1s = 284.6 eV. 
 
38 
 
Electron Spin Resonance Spectroscopy (ESR) 
The CW ESR spectra of the sorbent taken ?as-is?, either ?calcined? or ?sulfided?, were recorded 
at the X-band on a Bruker EMX-6/1 EPR spectrometer composed of the EMX 1/3 console, an 
ER 041 X6 bridge with a built-in ER-0410-116 microwave frequency counter, an ER-070 
magnet, and an ER-4102st standard universal rectangular cavity. Samples of the ?sulfided? 
sorbent were transferred to the ESR test tube with a minimal exposure to ambient air. Samples 
were cooled to 77 K in a liquid nitrogen finger Dewar. All spectra were recorded with a field 
modulation frequency of 100 kHz, a modulation amplitude of 6 mT, a power incident to the 
cavity of 2 mW and a frequency of 9.37 GHz. Determination of the ESR spin concentrations 
were carried out under the nonsaturating conditions using 10 mM CuSO4 solution in water as 
standard. ESR measurements with samples of the sorbent that were carefully outgassed in the 
High Vacuum (HV) of ~1x10-6 Torr are consistent with those obtained upon the re-admission of 
air into the ESR test tubes. The BioEPR software was used for computer simulations of the ESR 
signals. 
N2 Adsorption and Desorption Isotherms: 
 Nitrogen adsorption/desorption isotherms at 77 K were measured by an Autosorb 1-C 
instrument  
. Before measuring the total surface area, samples were outgassed for 3 h at 200 ?C. The specific 
surface area, SBET was calculated via the Brunauer-Emmett-Teller (BET) equation, and the total 
pore volume was calculated at P/P0 = 0.95. 
X-Ray Diffraction  
 
39 
 
XRD at room temperature was performed by a Rigaku Miniflex instrument and the  
diffraction patterns were obtained with the Ni-filtered Cu K? radiation (  = 0.15418 nm), 
scanning speed of 1 o/min using commercial XRD libraries. 
 
 
40 
 
Chapter III: Wide Temperature Range H2S Removal by Promoted ZnO/SiO2 
: Effect of Support, Entrapment in Microfibrous Media and Scale-up 
Priyanka Dhage, Vivekanand Gaur1, Hongyun Yang2, Bruce J. Tatarchuk 
Department of Chemical Engineering, Auburn University, Auburn, AL 36849, USA 
1Filtrex Technologies Pvt Ltd. Bangalore - 560043, INDIA 
2IntraMicron Inc. 368 Industry Dr. Auburn, AL, 36832, USA 
 
Abstract 
 Promoted ZnO sorbents with formulation M0.05Zn0.95O (M = Mn, Fe, Co, Ni, Cu) were 
supported on mesoporous silica gels such as MCM-41 and MCM-48. H2S adsorption was 
conducted in temperatures ranging from room temperature to 400 oC in flowing 1%H2S/H2 at 
various moisture levels. The results of desulfurization tests on these sorbents at room 
temperature indicate that a copper doped ZnO (15% w/w)/MCM-41 sorbent 
(Cu0.05Zn0.95O/MCM-41) has the highest saturation sulfur capacity at 0.9 mol S/mol 
(Cu0.05Zn0.95O), which is approximately twice that of ZnO/SiO2 sorbent at similar loadings.  H2S 
adsorption at elevated temperatures (ca. 400 oC), resulted in near total sulfidization of the 
available reactant regardless of dopant. At intermediate temperatures, the utilization of the 
reactant (M0.05Zn0.95O) toward H2S removal depended on the support employed in the order 
MCM-41 > MCM-48 > silica gel. This dependence was investigated in terms of the support: 
surface area, pore volume, and pore size; using N2 adsorption-desorption isotherms. With an 
 
41 
 
increase in ZnO loading on the silica support, the surface area, pore volume and pore size 
decreased. At equivalent levels of support surface area and pore size, higher pore volumes 
provided greater low temperature H2S capacities, presumably as a result of the lower Zn 
(NO3)2.6H2O concentrations used during impregnation/drying and a diminution in Zn0.95M0.05O 
crystallite size.  
Keywords: H2S removal; ZnO catalyst; Mesoporous silica; MCM-41; Breakthrough analysis 
III.1. Introduction 
High efficiency desulfurization is critical to maintain the activity of fuel processing catalysts and 
high-value membrane electrode assemblies in logistic fuel cell systems. On-board fuel 
processing of liquid hydrocarbon fuel is being investigated to supply hydrogen for fuel cell-
based auxiliary power units. If sulfur is not removed from the liquid phase, the removal of sulfur 
as H2S from the reformate is a key step since downstream catalysts and the fuel cell can be 
poisoned by a small amount of H2S in the feed [3]. Depending on the fuel selection, additional 
ancillary components are required for processing the fuel to meet the fuel cell?s requirement. 
Steam reforming catalysts, PEM anode catalysts and the shift catalysts are intolerant to sulfur, 
and the desulfurization step is very important to ensure adequate lifetime of fuel processors [61]. 
Hydrogen sulfide is present in many high temperature gas streams during extraction and 
processing of fossil fuels, natural gas and geothermal brines. H2S is also found in many industrial 
process gases, particularly in the mineral and metallurgical process industries [62].  The sulfur 
compounds needs to be reduced to less than 1 ppmv for a clean environment because high 
concentration of sulfur compounds result in health hazards, air pollution, acid rain and corrosion 
of metallic materials. Hydrogen-rich fuel needs to be less than 100 ppb [63]. Metal oxides, in 
particular ZnO, are widely applied for gas desulfurization processes [64-66]. Westmoreland and 
 
42 
 
Harrison [13] have shown that the oxides of Fe, Mn, Zn, Ca, V, Cu, Co and W are the most 
suitable sorbents at temperatures above 300 oC. Among various metal oxides, the ZnO based 
sorbents have shown advantages of higher sulfur capacity and favorable sulfidation 
thermodynamics at moderate temperatures [67] . Silica supported with ZnO and/or doped with 
Cu is also a widely used catalyst for desulfurization [17, 20, 22]. In combination with high-
temperature stability, low-temperature activity is highly desirable for a new catalyst. The 
removal of H2S at high temperatures (ca. 350 oC) has been extensively studied, but little work 
has been reported in the literature for the development of low-temperature H2S adsorbents [6]. 
Concerns about the removal of H2S at low temperatures are growing because of its removal for 
fuel cell application and several other processes, including natural gas sweetening and the Claus 
process, wherein H2S is by-product at low temperatures.   
Mesoporous silica has been recently used as a support for metal catalysts, resulting in several 
cases in significant improvements when compared to commercial and conventional amorphous 
silica-alumina catalysts. Most of the literature on the mesoporous silica materials deals mainly 
with their synthesis and surface characterization. The mesoporous silica supports including 
MCM-41, MCM-48, SBA-15 are found to be the superior base matrix for various surface 
modifications with amines and their subsequent application in the low temperature removal of 
acidic gases like H2S and CO2 [36]-[37]. Very few studies are reported on the metal 
incorporation in mesoporous silica by incipient wetness impregnation. It is well-known that the 
surface area, porosity, and chemical nature of the oxide support can affect supported metal-
catalyzed reactions. Therefore, it is interesting to compare adsorption capacity of metal oxide 
supported on mesoporous silica with conventional silica. Since ZnO has a high equilibrium 
constant for H2S removal at ambient temperature, in the present work, mesoporous silica (MCM-
 
43 
 
41 and MCM-48) supports are impregnated with ZnO by incipient wetness method. Additionally, 
the ZnO supported sorbents were doped with Cu and examined for desulfurization performance 
at room temperature. These materials were characterized by XRD, N2 adsorption, and SEM to 
obtain detailed information in the development of new sorbents. The influence of moisture on 
H2S removal was also discussed. After the adsorption tests, the catalysts were thermally 
regenerated for multiple adsorption-desorption cycles.   
IV.2. Experimental Section 
 
III.2.1 Silica support 
 
 Different types of silica supports including MCM-41, MCM-48 silica and conventional 
silica gel were used in this study. The MCM-41 used in this study was procured from Sigma 
Aldrich and used as-received without any further purification. The MCM-48 was prepared as 
described in Chapter II by Schumacher and co-workers [60].  
III.2.2 Sorbent impregnation  
 
 The metal nitrate solutions of different transition metals were used as precursors for 
impregnating the MCM (-41 and -48) and silica support with different ZnO and doped ZnO 
loadings. The impregnated supports samples were dried at 100oC for 6h and subsequently 
calcined at 350oC for 1 h under air flow. All the samples were stored in desiccators for further 
use.   
 
III.2.3 Adsorption experiment 
 
 
44 
 
 The experimental set-up and procedure are described in Chapter II. The adsorption 
experiments for desulfurization were carried out at ambient conditions (20 oC, 1 atm). An inlet 
concentration of 1 % (v/v) H2S in H2 (ultra high purity grade; from Airgas South, Inc.) was used 
as sulfur source at a face velocity of 2.12 cm/s, corresponding to the volumetric gas flowrate of 
1900 h-1 GHSV (0.1 slpm). H2S uptakes during the dynamic adsorption experiments were 
measured using a gas chromatography (GC) instrument (Varian 3800) equipped with thermal 
conductivity detector (TCD).  In each adsorption run, 0.5 g sample was packed in the reactor. In 
this study, the breakthrough time was defined as the time from beginning of the desulfurization 
to the time when the H2S concentration at the exit reached 100 ppmv.   
IV.3. Sorbent Characterization 
 Nitrogen adsorption/desorption isotherms at 77 K were measured using Autosorb 1-C 
model from Quantachrome Instrument Corporation. Prior to measurement, all samples were 
degassed for 10h at 200 ?C. Specific surface area, SBET was calculated using the BET equation. 
Total pore volume, VP was calculated at P/P0 = 0.95. The pore width, Pw distribution over the 
range of (2?80 nm) was generated from the adsorption branches of the isotherms via the BJH 
method. Calculations were performed using Autosorb 1C software. XRD patterns were obtained 
using a Rigaku Miniflex diffractometer at room temperature using CuK? radiation. Diffraction 
patterns were obtained with Ni-filtered CuK? radiation (  = 0.15418 nm) using a scanning speed 
of 1o/min and an accelerating voltage of 30 kV. The resultant patterns matched with standard 
data for ZnO for the purpose of phase identification. The surface morphology of the MCM 
samples before and after metal impregnation was investigated with Scanning Electron 
Microscopy (SEM). Prior to SEM (Zeiss Digital Scanning Microscope DSM940), the samples 
were vacuum coated with gold (Pelco SC-7 auto sputter coater). 
 
45 
 
III.4 Results and discussion 
III.4.1 Preparation and characterization of ZnO supported sorbents 
  
XRD patterns of MCM-41 and MCM-48 are shown in Fig. III.1 (a-b).  
 
Figure III.1a: XRD Pattern of the MCM-41 
 
 
46 
 
 
 
Figure III.1b: XRD Pattern of the Lab-made MCM-48 
 
The MCM-48 was made in the lab and the XRD pattern was obtained to compare with the 
literature to ensure the ordered mesoporous structure. The diffraction peaks obtained at 2 angles 
of 1.8o, 3.6o, 4.5o, 5.5o for MCM-41 and 2.7o, 3.2o for MCM-48 confirmed the structure of the 
same as also reported elsewhere[68],[69]. The pore structure analysis obtained by nitrogen 
adsorption/desorption isotherms further confirmed the mesoporosity and that ZnO loading 
occurred inside the pore channels of the MCM-41 support. The degassed MCM samples showed 
a type IV isotherm as shown in Fig III.2(a-b). 
 
 
 
47 
 
 
Figure III.2a : N2 adsorption-desorption isotherm for MCM-41- Commercial 
 
 
48 
 
Figure III.2b : N2 adsorption-desorption isotherm for MCM-48 ? Lab made 
The isotherms also confirm that after impregnation with ZnO and/or doped with CuO, the 
mesoporous pores were not completely filled or blocked, resulting in the preserved type IV 
isotherm, allowing liquid nitrogen to access the pores the pores. The surface area, pore volume, 
and pore size of MCM-41, MCM-48 and SiO2 before and after impregnation with ZnO are 
shown in Table III.1.  
 
 
 
 
Table  III. 1. Structural characteristics of Silica sorbents determined by N2 adsorption 
  Sample ZnO 
loading 
(w/w%) 
Sg 
(m2/g) 
Vt  
(cc/g) 
Wavg 
(nm) 
Avg. Pore 
Size (nm) 
 
49 
 
 
 
MCM-41 
ZnO/MCM-41 
ZnO/MCM-41 
Cu/ZnO/MCM-41 
 
MCM-48 
ZnO/MCM-48 
ZnO/MCM-48 
Cu/ZnO/MCM-48 
 
 
 
 
0 
15 
25 
15 
 
0 
15 
25 
15 
 
 
 
 
1260 
850 
672 
524 
 
1420 
631 
592 
303 
 
 
 
 
1.30 
0.74 
0.71 
0.52 
 
1.10 
0.47 
0.45 
0.32 
 
 
 
 
2.45 
2.58 
2.25 
2.51 
 
2.53 
2.26 
2.26 
2.58 
 
 
4.1 
3.5 
3.5 
3.6 
 
3.1 
3.0 
3.0 
4.2 
Sg: Specific surface area calculated from the BET equation; Vg: Total pore volume; Vmicro: 
Micropore volume; Wavg: Average pore width determined from DR method  
 
 
The BET area as well as the pore volume of all the support samples decreases on impregnation 
with the metal oxide. As metal loading increases, the surface area and pore volume decrease. 
SiO2-supported catalysts have BET surface area between 200 and 300 m2/g and large pore 
volumes. As the loading of ZnO was increased from 0 to 15% (w/w), the BET area, pore volume, 
and the average pore size decreased. The decrease in the BET area in MCM-41 and MCM-48 
samples was observed to be from 1260 and 1420 m2/g to 850 and 592 m2/g, respectively. The 
 
50 
 
pore volumes were also decreased to almost half. Furthermore, on promoting the ZnO-based 
samples with 5 mol% CuO, the BET area and pore volume further decreased considerably. On 
the other hand, in the Cu-promoted samples, pore size was observed to increase significantly (in 
SiO2) or marginally (in MCM). This indicates that impregnation of silica supports with metal 
oxides may result in an decrease in micro porosity and an increase in macro or meso porosity. In 
other words, it may be concluded that the metal oxides are preferentially dispersed in the interior 
of the porous texture of Silica, which results in the blocking of mainly the micropores and in the 
development of pores opening.   
 The pore size, surface area and pore volume of MCM-41 before and after ZnO loading 
were obtained from the nitrogen adsorption/desorption isotherms. Likewise, for the adsorption 
isotherms of Cu-promoted and unpromoted ZnO/MCM-41 shown in Figure III.3.  
 
Figure III. 3a: N2 adsorption isotherms for MCM-41, ZnO/MCM-41 and Cu-ZnO/MCM-41 
 
51 
 
 
The PSD of the MCM-41 and MCM-48 are shown in Fig III.3(b-c), it indicates the pore size in 
the range of 2-4 nm for both the MCM samples.  
 
Figure III. 3b : Pore Size Distribution of MCM-41 (commercial ) 
 
52 
 
 
Figure III. 3c: Pore size distribution of MCM-48 (Lab made) 
The PSD data is shown in Table III.1 and Fig III.3(b-c). The pore size of the MCM-41 support 
was 4.14 nm. After the ZnO was loaded into its channels, the pore size decreased. The pore size 
of ZnO (8%)/MCM was 3.5 nm, smaller than that of the MCM-41 support, which confirmed that 
ZnO was dispersed into the MCM-41 pore channels. With increasing ZnO loadings, the pore size 
further decreased, but only marginally. The pore sizes were 2.29 and 2.25 nm for ZnO 
(15%)/MCM-41 and ZnO (25%)/MCM-41, respectively. The surface area and the pore volume 
of MCM-41, after ZnO loading, exhibited the same trends as the pore size. 
 
 
 
 
53 
 
The morphology of MCM-41, MCM-48 and ZnO loaded MCM-41 and MCM-48 was viewed by 
SEM as shown in Fig III 4.  
 
Figure III. 4a: SEM image of MCM-48 sample before impregnation - MCM-48 Blank 
 
Figure III. 4b: SEM image of MCM-48 sample after impregnation - 15 % ZnO / MCM-48 
The particle size of the MCM-41 and MCM-48 support was 5?10 m. The MCM-41 particle was 
made of loosely packed small particles with submicron size. After impregnation with ZnO, the 
particle size of MCM-41 remained unchanged, this indicated that ZnO was dispersed into the 
support pores and was not deposited on the outer surface of the particles.  
  
 
54 
 
 
III.4.2 Effect of different types of metal oxides 
 
In order to find an optimal sorbent for H2S removal at ambient temperature, a number of 
adsorption experiments have been carried out on silica dispersed with different transition metal 
(Zn, Cu, Mn, Fe, Co and Ni) oxides. Fig. III.5a compares the breakthrough capacities of H2S for 
various metals at room temperature. Different metal oxides supported on SiO2 at almost identical 
loading of 21% (w/w) showed distinct performances. ZnO/SiO2 showed the highest (~ 48 mg 
sulfur/g sorbent) capacity. On the other hand, iron, cobalt and nickel oxides supported silica 
samples are not effective candidate because they showed almost no capacity under identical 
operating conditions. As a result, the H2S adsorption performance of the supported metal oxides 
increased in the order: Fe  Co  Ni < Mn < Cu < Zn.    
 
55 
 
 
Figure III. 5a: Breakthrough capacity and % theoretical capacity valuesa for different metal 
oxides with same loading on silica tested at RT, Q= 100 cc/min, face velocity = 2.12 cm/s, 
Calcination condition = 350 C/1h, Wt. = 0.5 g. 
 
The thermodynamic data of the reaction of the metal oxides with H2S was obtained using the 
HSC software as shown in Fig III.5b. The formulae for saturation capacity and % theoretical 
capacity are given in appendix I. It shows that ZnO and CuO have favorable thermodynamics 
with lower outlet equilibrium concentrations (ppmv). This compliments the results obtained by 
the H2S adsorption study conducted at room temperature on these metal oxides supported on 
silica with approximately similar metal oxide loadings. CuO showed better sulfidation 
thermodynamics than other oxides but CuO is unstable over the range of temperature.  
 
56 
 
 
Figure III. 5b: Equilibrium H2S concentration data generated using HSC software for various 
metal oxides 
 
III.4.3 Comparison with the commercial ZnO 
 
The H2S adsorption of the ZnO/SiO2 (21 wt%) was compared with the commercial ZnO 
samples obtained from Sud-Chemie (G-72E) and BASF (SG-901).The breakthrough 
performance of the three sorbents,  tested at same conditions ? Vf= 2.12 cm/s, sample wt = 0.5 
gm, temperature = 20oC, is shown in the Fig. III.6.The nature of the breakthrough curve differs 
for these sorbents indicating different diffusion mechanisms in each case due to the difference in 
the sample preparation. The commercial sorbents contain over 90% of pure ZnO with small 
amounts of binder whereas the ZnO/SiO2 contains 21wt% impregnated on the silica and contains 
 
57 
 
uniform nanocrystals of the ZnO dispersed in the porous silica matrix. This leads to better ZnO 
utilization and adsorption capacity of the sorbent even at room temperature.  
 
 
Figure III.6. Breakthrough curves for commercial ZnO and ZnO/SiO2: T = 20 C, Co = 1Vol% 
H2S/H2, Face Velocity = 2.12 cm/s, Wt. =0.5 gm 
   
The shape of the breakthrough curve indicates that the adsorption (diffusion mechanism of 
the H2S to ZnO) is different in all these cases. This is mainly due to the method of preparation of 
the sorbent. The lab-made ZnO/SiO2 shows a desirable sharp breakthrough curve. The 
commercial ZnO were in the form of extrudates crushed to same size as SiO2 (150-200 microns) 
for comparison. The extrudates contain approx. 90 % pure ZnO and rest is binder. The ZnO/SiO2 
(21 wt% loading) shows maximum H2S capacity as compared to commercial extrudates with ~ 
 
58 
 
90% ZnO loading. The ZnO on the silica matrix is present in the form of nanocrystals with 
uniform dispersion, thus ensuring maximum accessibility to H2S and this leads to higher 
capacity. No XRD pattern was observed when the sample was tested indicating that the ZnO 
crystal size is < 4 nm. 
 
III.4.4 Screening test for the metal oxide 
 
Different metal oxides supported on SiO2 at almost identical loading of 21% (w/w) 
showed distinct performances. ZnO/SiO2 showed the highest (~53.12 mg sulfur/g sorbent) 
capacity. On the other hand, iron, cobalt and nickel oxides supported silica samples are not 
effective candidates because they showed almost no capacity under identical operating 
conditions. As a result, the H2S adsorption performance of the supported metal oxides increased 
in this order: Fe  Co  Ni < Mn < Cu < Zn.  The MCM materials exhibited a superior affinity to 
H2S, and the desulfurization capacity is up to 0.9 mol S/mol sorbent at ~15% ZnO loading. 
Likewise, the capacity decreased with further increases in ZnO loading. This suggests that an 
optimum loading of metal oxides exists for every silica support depending on the support pore 
volume. An excess of 15% (w/w) ZnO loading on MCM-41 may result in the formation of 
relatively larger metal crystallites which may cause the blockage of micro and mesopores of the 
silica support. A similar explanation of the excess of ZnO loading is reported elsewhere [2]. 
During the desulfurization reaction, the reaction-product may plug the pores and limit the gas 
diffusion, resulting in a decrease in H2S capture. The SEM images, as presented in Fig. III.4, 
showed spherical type morphology for the ZnO supported MCMs which is rather similar to their 
blank MCM-41 and MCM-48, and no zinc oxide aggregates were observed on the external 
surface of particles. Thus, the adopted method allowed the persistence of MCM-41 texture with 
 
59 
 
zinc oxide inserts in the MCM framework and/or forms finely divided zinc oxide nanoparticles 
in the pores of MCM samples.  
III.4.5 Effect of promoter 
 
To investigate the influence of the doped ZnO on the desulfurization activity, a series of metal 
(M) doped ZnO supported on silica (M-ZnO/SiO2) with M/Zn atomic ratio of 5/95 were prepared 
by incipient wetness method. Here, M includes transition metals, including Mn, Fe, Co, Ni and 
Cu oxides. In all the promoted sorbent samples, the total (Zn + M) metal oxide loading was kept 
at 21% (w/w). Table III.2 shows the comparative desulfurization capacity of promoted and un-
promoted ZnO-SiO2 at room temperature.  
The CuO0.05ZnO0.95/SiO2 showed the highest saturation capacity followed by 
FeO0.05ZnO0.95/SiO2. The decreasing order of H2S removal at saturation level may be expressed 
as: CuO0.05ZnO0.95/SiO2 > FeO0.05ZnO0.95/SiO2 > CoO0.05ZnO0.95/SiO2 > NiO0.05ZnO0.95/SiO2  
MnO0.05ZnO0.95/SiO2. The saturation capacity of ZnO/SiO2 increased by approximately 31% and 
23% on Cu- and Fe-promotion, respectively. To compare the relative effect of doping on 
saturation capacity during desulfurization, theoretical utilization of metal (Zn + Cu) oxides was 
also calculated. Table III.2 shows the results obtained for M-ZnO/SiO2. The percent metal 
utilization for H2S sorption was highest for Cu and followed the same trend as for the saturation 
capacity. Interestingly, approximately 90% Zn/Cu was utilized at room temperature as shown in 
Fig.III.7 . Similar experiments were performed for Cu-ZnO/MCM-41 and Cu-ZnO/MCM-48, 
and the obtained results showed the same trend. It is proposed that Cu-promoted ZnO/SiO2 may 
have increased defects on the ZnO surface and higher intra-particle diffusivity. The Cu doping 
may significantly change the crystallite size of ZnO.  
 
 
60 
 
 
 
Figure III.7. Saturation capacity of the doped sorbents 
 
 
 
 
Hypothesis:  
The dopant (usually added in small quantities) serves to  
 Reduce crystallite size  
 Increase surface area 
 Add defects to the structure  (thus enhancing the accessibility of the H2S to active 
metal for adsorption) 
Zn
O/
SiO
2
Mn
-Z
nO/
SiO
2
Fe
-Z
nO
/Si
O2
Co
-Z
nO
/Si
O2
Ni
-Z
nO/
SiO
2
Cu
-Z
nO/
SiO
2
0
10
20
30
40
50
60
70
80
90
Sat
ur
ation
 Capacit
y* ( m
g S
/g 
sorbe
nt)
Co=1vol% H2S/H2, sorbent wt= 0.5gm
ZnO loading = 21wt.%
T=20oC, U=2.12cm/s
Particle size = 100-200 um
 
61 
 
Table III.2. Saturation capacity values of the doped sorbents and commercial sorbents 
Dopant@5mol% 
M0.05ZnO0.95/SiO2 
Saturation Capacity 
( g S/g ZnO) % of Theo. Capacity 
MnOx(1<x<1.5) 0.24 59.31 
Fe2O3 0.33 81.25 
Co Ox(1<x<1.5) 0.30 75.00 
NiO 0.25 62.51 
CuO 0.37 90.63 
ZnO(un-doped) 0.25 62.51 
ZnO (BASF SG-901) 0.02 4.96 
ZnO 
(Sud-Chemie G-72E) 0.04 9.8 
 
 
III.4.5.1 Effect of change in concentration of the promoter 
 
Promoted ZnO sorbents, MxZnO(1-x)/SiO2 where 0?x?1 and (M, N = Cu, Ni, Mn, Fe, Co, Mg); 
were made by incipient impregnation method on SiO2 with nitrates (2 M conc.) as precursors. 
The challenge gas was chosen as 1 vol% H2S/H2, outlet gases were analyzed by TCD-GC 
(Varian CP3800). The concentration of the dopant was varied from 0-100% for the Cux-ZnO(1-
x)/SiO2, the sorbent shows highest adsorption at the Cu0.2ZnO0.8/SiO2 indicating that the dopant 
concentration can be changed in the range of 5-20 atomic % with effective sulfur adsorption as 
shown in Fig.III.8.  
 
 
62 
 
 
Figure III. 8. Saturation Capacity of the sorbents with varying Cu concentrations 
 
Hypothesis: When you have less dopant the dopant acts as adding more defects and hence 
adding this you are exposing more surface area , while if you keep adding a dopant that means 
you are replacing zinc with a less effective sorbent and hence after a certain optimum it keeps on 
dropping  
 
 
III.4.6 Effect of pore volume 
 
To study the effect of pore volume, surface area of the support on the adsorption capacities of the 
sorbent silica (including mesoporous silica) of varying pore volume (0.6-2 cc/g), surface area 
(300-1200 m2/g) were tested. Hypothesis: The single step wetness impregnation metal loading 
T=20oC, U=2.12cm/s, 
Sorbent wt. =0.25gm, 
Particle size = 100-200 
um 
 
34 wt.% (CuO+ZnO)/SiO2 
14 wt.% (CuO+ZnO)/SiO2 
 
63 
 
increases as the pore volume of the support increases and this leads to increased sulfur 
adsorption capacity. Table III.3 shows silica of varying pore volume compared with MCM-41 
and commercial ZnO. Higher the pore volume, higher is the ZnO that can be loaded in single 
step incipient wetness impregnation method and hence higher capacities are observed as shown 
Fig.III 9a [5].  
Table III. 3. Capacity values of the silica with varying pore volumes and their adsorption 
capacities 
Silica 
Pore volume 
(cc/g) 
Wt.% ZnO 
loaded 
Expt. Saturation 
capacity 
(mg S/g sorbent) 
Expt. Sat. Cap. % of 
theo. Cap. 
(392 mg S/g ZnO) 
0.8 11.50 18.65 41.4 
1.15 15.80 31.87 51.4 
1.65 21.20 50.46 60.7 
1.8 22.66 53.21 59.9 
0.23 
(Sud-Chemie-G 
72E) 
90 34.58 9.8 
1.0 
(MCM-41) 14.00 55.77 94.8 
 
 
 
 
 
64 
 
 
Figure III.9a. Saturation Capacities of the different silica with varying pore volume (i.e. varying 
ZnO loading) 
Higher ZnO loadings lead to higher sulfur adsorption capacities but then for a given silica, how 
much of ZnO can be loaded that can still effectively remove per unit ZnO. This was studied in 
the next study as shown in Fig III.9b, where the type of silica support was kept same and varying 
amounts of ZnO were loaded by changing the concentration of the precursor. The Fig III.9b. 
shows that there an optimum loading for a given silica ( pore volume and surface area), where 
the sorbent shows the maximum sulfur adsorption at a given operating condition. Single step 
impregnation ZnO loadings were varied on SiO2 to obtain a series of ZnO/SiO2 from 12 wt% - 
36 wt% (Surface area: 300 m2/g, Pore volume: 1.65 cc/g) by changing the concentration of the 
nitrate precursor from 0.5-3.5M. As the loading increases above 21wt% we expect the adsorption 
capacity to rise but after 21wt% loading the capacity goes down as indicated in Fig III.9b.  
 
65 
 
 
 
Figure III.9b. Saturation capacity of ZnO/SiO2 with varying ZnO loading 
 
0
10
20
30
40
50
60
70
10% 15% 20% 25% 30% 35% 40%
Sa
tura
tio
n C
apa
cit
y 
( m
g S/
 g so
rbent)
Wt. % ZnO loading
U =2.12 cm/s, T= 20oC, 
Particle Size = 100-200 um
Sorbent Wt. = 0.5 gm, Co=1 vol% H2S/H2
 
66 
 
 
Figure III.9c. XRD patterns for silica and ZnO/SiO2 with varying ZnO loadings 
 
Hypothesis: Use of higher concentration of the nitrate precursor leads to formation of larger 
ZnO crystallite size which can lead to blocking of the pores and hence limit the access of the gas 
to the  complete  pore.  
This was also verified by obtaining XRD of the samples, shown in Fig. III.9c. It was evident that 
until 21wt.% no XRD peaks were observed, indicating that small crystallite size which is 
probably uniformly distributed in the silica matrix but as the loading goes to 36wt. % significant 
ZnO peaks were observed by XRD. Fig III.9d. shows the effect of adding CuO into ZnO/SiO2. 
Here 36 wt% ZnO/SiO2 was used and this loading was kept constant only the concentration of 
the Cu was varied from 5-100%. The pattern shows that with 5% CuO in ZnO/SiO2 the XRD 
pattern is similar to that of ZnO/SiO2, indicating that CuO is still amorphous and is distributed 
 
67 
 
inside the matrix of ZnO/SiO2. The 36wt.% CuO/SiO2 shows clear peaks of existence of CuO. 
This can be compared to the XRD pattern obtained from pure Cu0.05Zn0.95O and ZnO obtained 
from pure powders made by calcined the nitrates of respective solutions at 350 C/1h in air as 
shown in Fig. III. 9e. 
 
Figure III.9d. XRD Pattern showing effect of adding Cu (0.05-1) on ZnO/SiO2 
 
68 
 
 
 
Figure III.9e. XRD Pattern of pure ZnO and Cu0.05Zn0.95O made from the calcination (350 
C/1h/air) from nitrate precursors. 
III.4.7 Effect of calcination temperature 
 
The impregnated samples, once dried in air for 6hrs were calcined. To study the effect of the 
calcination temperature the ZnO/SiO2 (21wt %) samples were calcined at different temperatures 
in the range of 250-550 oC. The calcination is carried out to decompose the nitrates (used as 
precursor) to oxide. Fig.III.10a shows the saturation capacity of the samples tested for H2S 
adsorption capacity with challenge concentration of 1 vol% H2S/H2 at room temperature. XRD 
of the samples calcined at different temperatures is shown in Fig. III.10a. The XRD shows that 
250 oC is lower temperature for decomposition of nitrates to oxides and hence the capacity is 
high, because even thought the entire nitrate is not converted to metal oxide, the nitrates can 
 
69 
 
absorb additional sulfur. On the other hand, 550 oC is higher temperature which results in lower 
adsorption capacities, possibly due to increased crystallite size of the particles. 350 oC was 
chosen to be the calcination temperature for all the samples made henceforth. Hypothesis: Lower 
calcination temperature can result in smaller more uniform and well dispersed crystals in the 
porous silica matrix.  
This was verified by obtaining the XRD patterns as shown in Fig. III.10b for these samples 
calcined at different temperatures. The XRD peak size increases as the calcination temperature 
went above 250 oC. At 250 oC we see a wider range of peaks (along with ZnO peaks) which are 
due to presence of nitrates in the sample. This also confirms that the temperature for calcination 
to ensure decomposition of nitrates to oxides should be above 250 oC and more precisely ~ 350 
oC in air. The XRD patterns were obtained with samples of higher loading (~36 wt %) and the 
adsorption results in Fig. III.10a was carried out on ZnO/SiO2 of 21wt%, since XRD patterns of 
ZnO (21wt %) /SiO2 is not observable.  
 
Figure III.10a. Saturation capacity of ZnO/SiO2 calcined at different temperatures 
0
10
20
30
40
50
60
70
80
250 350 450 550
Sat
ur
ation
 Capacit
y
(m
g of
 S/g 
of sorbent)
Temperature (oC)
ZnO(21wt%) /SiO2. Vf=2.12 cm/s, 
RT, Sorbent wt. = 0.5gm.  
 
70 
 
 
Figure III 10b.XRD patterns of the ZnO( 36 wt.%) /SiO2 calcined at different temperatures. 
 
III.4.8 Effect of H2S sorption temperature 
 
To understand the effect of temperature on the adsorption of H2S, ZnO supported silica was 
tested for its reactivity in the temperature range between 20 and 400 oC. Fig. III.11 compares the 
effect of sorption temperature on the H2S adsorption capacity and saturation level for ZnO 
(15%)/SiO2 and ZnO (15%)/MCM-41, when the inlet H2S concentration (Co) is 1% (v/v) in H2 at 
a face velocity of 2.12 cm/s.  
 
71 
 
 
Figure. III.11 H2S saturation capacity for ZnO/SiO2 and ZnO/MCM-41 (15 wt% loading) tested 
from room temperature to 400 C, Q = 100 cc/min, Face velocity = 2.12 cm/s, Calcination 
condition = 350oC/1h 
 
The desulfurization capacity increased from 20 and 38 mgS/g sorbent to 33 and 65 mg S/g 
sorbent, respectively for ZnO/SiO2 and ZnO/MCM-41 as the temperature is increased from 20 to 
300 oC.  The nature of the curves for both the sorbents (ZnO/SiO2 and ZnO/MCM-41 can be 
explained as follows, at room temperature it is predominantly the physisorption and reaction that 
leads to higher capacities (attractive forces) as the temperature rises, there will be reduction in 
physisorption (attractive forces) as the kinetics of reaction starts taking over. At 200 C, a 
decrease in the capacity was observed for both sorbents, this is because attractive forces 
responsible for the Physisorption have dropped but the temperature is not sufficient to enhance 
 
72 
 
the kinetics of the reaction yet. A significant increase in capacity (~60 mg S/g sorbent) can be 
observed for ZnO/SiO2 as the reaction temperature rises to 400 oC. However, the capacity 
remains unchanged for ZnO/MCM-41 as the temperature rises from 300 to 400 oC. As a result, 
the performance of ZnO supported on either support is significantly larger at a high operation 
temperature of 400 oC. The Cu-promoted ZnO/MCM-41 was also examined for its 
desulfurization capacity at 400 oC and room temperature. The result is shown in Fig III.12 
 
 
 
Figure III.12. H2S saturation capacity for ZnO/MCM-41 and Cu0.05ZnO0.95/MCM-41 (15 wt% 
ZnO) tested at RT and 400 C, Q = 100 cc/min, Face velocity = 2.12 cm/s, Calcination condition 
= 350oC/1h 
 
At room temperature, an approximately 30% increase in capacity was observed by Cu-promotion 
of both ZnO/SiO2 and ZnO/MCM-41. At 400 oC, the ZnO/SiO2 showed almost 46% increase in 
0
10
20
30
40
50
60
70
20 400
T e m p e r a t u r e  (
o
C)
S
a
t
u
r
a
t
io
n
 C
a
p
a
c
it
y
 /
 m
g
 S
 (
g
 s
o
r
b
e
n
t
)
-1
Z n O / M C M - 4 1
C u - Z n O / M C M 4 1
 
73 
 
saturation capacity on promoting with Cu. The Cu-promoted ZnO/SiO2 does not show 
improvement in capacity at 400oC as compared to 20oC indicating that the Cu dopant is only 
active at lower temperatures.  
 
 
III.4.9 Comparative performance of different types of silica support 
 
 To determine the influence of support properties, in particular the structural properties, on 
the H2S adsorption capacity, different types of silica materials, including the conventional and 
amorphous silica (SiO2) and highly ordered mesoporous silica (MCM-41 and MCM-48) were 
chosen as support for ZnO and tested for their desulfurization performance. Unlike SiO2, both 
MCM-41 and MCM-48 without any metal loading show some H2S adsorption capacity at room 
temperature. This indicates a very weak interaction between H2S and MCM-41/MCM-48 at 
room temperature. Both, the conventional silica gel and mesoporous silica were impregnated 
with ZnO at the identical loading of 15% (w/w). The desulfurization activity of these samples 
was examined using breakthrough capacity measurements described in previous section. Fig. 
III.13 shows the breakthrough curves obtained from the adsorption experiments on these ZnO-
based silica samples under identical conditions.  
 
74 
 
 
Figure III.13. H2S Breakthrough curves for ZnO/SiO2 and ZnO/MCM (15 wt% ZnO) compared 
with Commercial ZnO (~90% ZnO) tested at RT, Q = 100 cc/min, Face velocity = 2.12 cm/s, 
Calcination condition = 350oC/1h 
 
The adsorption behavior is almost similar for all three samples. The breakthrough of H2S was 
observed to occur at 5, 17, and 20 minutes for the SiO2, MCM-41 and MCM-48 samples. The 
longer breakthrough time and higher desulfurization capacity of MCM-sorbents suggests them to 
be superior candidates for H2S removal over silica at ambient temperature. The enhanced 
adsorption capacity of MCM samples may be explained on the basis of their high surface area 
and well-structured pores, which may allow more uniform dispersion of ZnO crystallites. As can 
be seen in Table III.1, the BET surface area and total pore volume of MCM samples are 
significantly larger than those of SiO2. Even after the loading of ZnO, MCM-41 and MCM-48 
 
75 
 
samples found to have significantly larger surface area as compared to that of the corresponding 
ZnO/SiO2 sample. For example, after 15% (w/w) loading of ZnO, MCM-41 and MCM-48 
samples have approximately four- and three- fold larger (850 and 631 m2/g) surface area in 
comparison to that (248 m2/g) for SiO2 with equal ZnO loading. This suggests that the 
crystallites size of ZnO dispersed on larger surface (MCM-41 and MCM-48) may be smaller 
than that on the smaller surface of SiO2. Further, the smaller pore size in MCM-41 (4.1 nm) and 
MCM-48 (3.1 nm) as compared with that in SiO2 (25 nm) favor the possibility of smaller ZnO 
crystallite size dispersed inside their pores.  
 
III.4.10  Effect of moisture content 
 
  The sorbents prepared by impregnation of silica and MCM-41 with Cu-doped ZnO were 
tested for their H2S adsorption capacity in the presence of moisture. The moisture content in the 
inlet gaseous stream was varied from 1-10% (v/v). All the sorbent samples had a similar loading 
of Cu and ZnO and were tested under identical conditions. .Fig. III.14 shows that the capacity of 
Cu-ZnO/SiO2 increased as the moisture content increased until 5% moisture level in the inlet 
stream. Indicating a maximum of 53 mg S/g sorbent breakthrough capacity in the stream and 
then the capacity was found to decrease. On the other hand, in the case of MCM-41 the capacity 
firstly decreased from 53.12 mg S/g sorbent to 17.67 mg S/g sorbent, respectively for moisture 
level from 0 to 2.5% and then remained constant until 10%. 
 
76 
 
 
Figure III.14. H2S saturation capacity for ZnO/SiO2 and ZnO/MCM-41 (15wt. % ZnO) tested at 
varying moisture content (0-10%) at RT, Q = 100 cc/min, Face velocity = 2.12 cm/s, Calcination 
 
This decrease in the capacity of the Cu-ZnO/MCM-41 may be due to the change in the structural 
integrity of MCM-41 support in the presence of moisture. In the case of silica, with the increase 
in moisture content above 5% the capacity drops due to competitive adsorption of H2O and H2S: 
the presence of moisture will block the pores of silica, thus reducing the saturation capacity.  
III.4.11 Effect of regeneration temperature 
 
 The sorbents used in this study were also attempted to be regenerated after the sulfidation 
runs by thermal heating under an oxygen flow. The sorbents were tested for their repeated use 
over a number of adsorption-desorption cycles. The sulfided ZnO/SiO2 samples were regenerated 
 
77 
 
at 600 oC in air for 1h (shown in Chapter IV). After regeneration the adsorption capacity of the 
ZnO/SiO2 samples dropped almost 74% and then remained constant for the next cycles. It was 
observed that the Cu-promoted ZnO/SiO2 showed similar behavior in regeneration. On the other 
hand, ZnO/MCM-41 and ZnO/MCM-48 showed no recovery of the H2S adsorption capacity 
under identical regeneration at 600 oC. Based on a series of experiments on varying regeneration 
temperature and time, it was observed that the regeneration of ZnO/MCM-41 samples occurred 
at 400 oC, but the recovered adsorption capacity was lower (11%) than that (36%) of the 
ZnO/SiO2. 
III.4.11.1 Desorption test during regeneration 
 
The reaction taking place  
2223 SOZ n OOZ n S
 
The desorption of the adsorbent bed was carried out to test the liberated gases using the 
PFPD detector.  The desorbed gas was tested at equal interval of temperature (20-600oC). The 
analysis showed lower levels of liberation of SO2. This liberation of gases was more in the 
temperature range of 250-350oC. It is important during the regeneration that presence of O2 does 
not lead to formation of SO2, if the regeneration gas contains more SO2 then further removal of 
sulfur during regeneration should be adopted. The amount of SO2 liberated in this case was less 
than 10% of the challenge sulfur concentration.  
 
 
III.4.12 Effect of CO and CO2    
 
 
78 
 
The reformate streams usually contains CO, CO2, H2O, H2, N2 and H2S. Above mentioned tests 
confirmed the excellent performance of the prepared sorbents at room temperature over 
commercial sorbents in the presence of dry H2S, therefore it was important to test the 
performance in the presence of CO and CO2. The two best sorbents Cu0.05ZnO0.95/SiO2 and 
Mn0.025Fe0.025ZnO0.95/SiO2 were tested in the presence of CO and CO2. No change in the capacity 
(both breakthrough and saturation) was observed for either of the sorbents in the presence of 
CO/CO2, when tested at room temperature. as shown in Fig..III.15. But similar study at high 
temperature indicated a loss in capacity at higher temperatures (T= 400C) as shown in Fig.III.16. 
This was due to the formation of the COS in the reformate streams due to the reaction between 
CO/CO2 with H2S. A detailed thermodynamic study to understand the formation of COS and 
strategies adopted to mitigate this issue of COS at high and low temperatures are described in 
Chapter VI.  
 
Figure III.15. Breakthrough curves for Cu doped ZnO/SiO2 tested in the presence of CO and CO2 
0
0.2
0.4
0.6
0.8
1
1.2
0 10 20 30 40 50
C/Co
Time (mins)
Absence 
Presence of CO
Presence of CO2
T: 20 C Challenge: 1 vol.% H2S-33 vol. % CO2/CO-H2 at 4.66 cm/s
Sorbent: 15wt. % Sorbent weight: 1.0 g 
Cu0.05ZnO0.95/SiO2
 
79 
 
 
 
Figure  III.16: Breakthrough performance of  Fe0.025Mn0.025ZnO0.95/SiO2 with and without CO2 at 
400 C,  Test conditions :Q (2%H2S/H2) = 100 cc/min, Q(100% CO2) = 100 
cc/min,  T = 400 C, GHSV = 8800 h-1 , Wt= 0.5 g 
Cu in Cu0.05Zn0.95OSiO2 is catalyst for the reaction between CO/CO2 and H2S and that is why 
there is significantly higher concentrations of COS formation at high temperature (T = 400 C). 
This led to a new set of novel bimetallic sorbents with higher sulfur adsorption capacities but a 
dopant other than Cu. Also, the capacity of the Cu sorbent dropped significantly in the presence 
of CO, CO2 at high temperatures (T =400 C). There was therefore a need to develop a doped 
supported ZnO/SiO2 sorbent with alternate dopant. It was shown in the previous study that 
among the single dopants, only showed higher adsorption capacity for sulfur removal and 
therefore bimetallic dopants were tested with various combinations.  
 
80 
 
III.4.13 Novel Bimetallic Sorbents for H2S removal at room temperature 
 
The aim of the study was to develop doped supported sorbents which can effectively remove in 
the reformate streams. The previous study focused on Cu as a dopant but Cu is a catalyst for the 
reaction between CO/CO2 and H2S at high temperatures and hence it was important to get rid of 
this dopant. A set of novel bimetallic dopant promoted ZnO/SiO2 sorbents with formulation 
Mx/2Nx/2ZnO (1-x)/SiO2, where 0?x?1, (M, N = Cu, Co, Ni, Fe, Mn) were prepared by wetness 
impregnation method. The nitrates (2M conc.) were used as precursors and the sorbent was dried 
for 6hrs followed by calcination. The sorbents showed better ZnO utilization and saturation 
capacity at room temperature in comparison to the single dopant promoted ZnO/SiO2.  The 
Mn0.025Fe0.025ZnO0.95/SiO2 showed highest saturation capacity as shown in Fig.III.17. Chapter V 
talks in details about this novel sorbent for H2S removal and their performance over multiple 
cyles as well as the characterization of this sorbent to understand the role of the active sites.  
 
 
Figure III.17. Saturation Capacities of the novel bimetallic doped sorbents for H2S removal 
Cu
-Mg
Fe
-Mg
Mn
-Mg
Ni
-Fe
Mn
-Fe
Cu
-Mg
Cu
-Fe
Cu
Sud
-Chem
ie 
 Z
nO 
0
10
20
30
40
50
60
70
80
90
100
t1/2 
Capacit
y (m
g S
/g 
sorbe
nt)
 
81 
 
In Chapter VI. the results for this novel Mn0.025Fe0.025Zn0.95/SiO2  tested at higher temperatures 
(T=400 C) in the reformate stream are listed. The results indicate that the formation of COS was 
lower than in the presence of Cu, and the sorbent maintained higher capacities at T = 400 C in 
the presence of CO/CO2. 
 
 
 
III.4.14 Scale-up studies 
 
The sorbents Cu0.05Zn0.95O/SiO2 and Mn0.025Fe0.025Zn0.95/SiO2 were scaled up (1 kg batch) using 
the V-blender, shown in Fig.III.18 for the impregnation and it shows consistent performance as 
shown in Fig.III.19. 
 
Table III.4.Theoretical utilization values for scale up of the sorbent 
 
 
 
 
Batch 
number Method 
Batch size 
(gms) 
% Theoretical 
Utilization 
B1 Hand-made 20 82.65 
B2 V-Blender 800 78.06 
B3 V-Blender 800 82.65 
 
82 
 
 
Figure III.18. V-blender used for impregnation of samples for scale-up studies 
 
 
Figure III.19.Breakthrough curves for scale-up studies 
V - bl e nd e r  d e s i gne d  t o pr oc e s s  1.5 K g of  S i l i c a  
( w i t h bul k d e ns i t y of  0.4kg/ l )
 
83 
 
 
III.5 Microfibrous Entrapped Sorbent 
III.5.1 Kinetic effects due to microfibrous entrapped ZnO sorbents( MFES) 
 
This work is done as a continuation of the work published in 2008 Chemical Engineering 
Science Journal [17], in collaboration with the Virginia Tech- Aerospace and Ocean Engineering 
department for CFD simulation study (VT-AOE), to better understand the purpose of microfibers 
in packed bed. 
The basic relationships between breakthrough curves and the kinetic behaviors of fixed bed 
reactors were studied. Mecklenburg model[70], Wheeler model[71] and Yoon model[72] 
Mathematical models developed to predict the breakthrough time of adsorption processes taking 
place in packed beds. The bed depth service time equation (1) derived from Amundson 
equation[73]:  
 
)(1ln tKCC
A
Ao    (1) 
Where lumped K is defined as:  
c
Aoa CkK      (2) 
 A sharp breakthrough curve always has a large lumped K. The breakthrough ZnO utilization (X) 
of packed bed increases with increase in lumped K value and the critical bed depth (Zc)as shown 
in the rearranged equations (3) and (4) [17].  
K
C
C
X b
o 1ln
1    (3) 
 
84 
 
 
c
Ao
b
Aoc KUCCCZ 1ln   (4) 
 
In the modified Amundson model[17], the lumped K is explicitly correlated to ka as shown in 
equation (2). In this work, the equation (1) has been verified experimentally and is used to 
investigate the performance of packed beds and MFES. 
III.5.2 Preparation of MFES 
 
Microfibrous media with two different target fiber fractions were made using the method 
described in detail elsewhere [31]. The method is also described in the glass fiber entrapped 
sorbent preparation section of this document. Following compositions as shown in table were 
obtained by varying the fiber loading and sintering conditions.  
Table III.5. Composition of the GFES 
Loading Material 1 Material 2 
Solid Fiber % 2.9 4.35 
Total Solid  % 28 29.8 
Void % 72 70.2 
 
 
III.5.3 Model Evaluation 
 
The GFES were tested at 400oC in the presence of 0.5 vol% H2S/H2. The result shown in the Fig. 
III.20 is for material 1 tested at U=1.2 cm/s. K and ? were calculated from linear regression and 
the values obtained were 0.535 min-1 and 22 min respectively. Similar method was used for 
 
85 
 
calculations of all the K and ? values for varying face velocities (1.2-9.6 cm/s) and glass fiber 
volume fractions (3-4.5 vol %). Because of the symmetry of the breakthrough curve the ? and t1/2 
are equal. Ln (CAo/CA-1) vs. t is linear for majority part of the breakthrough curve. The slope is 
indicative of the lumped K value and the intercept is the log reduction value for a given 
breakthrough curve.  
 
Figure III. 20. Evaluation of modified Amundson model 
 
III.5.4 Effect of face velocity  
 
The GFES with ~ 17 wt. % ZnO loading were tested at 400oC in the presence of 5000 ppmv of 
H2S/H2. In this experiment, the m: v (mass: volume) of the GFES was varied as the face velocity 
was doubled, in order to maintain the same ?.  
y = -0.5354x + 11.126
R? = 0.9959
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30
ln((C
o/C
)-1)
C/
Co
Time(mins)
U= 1.2 cm/s, T= 400oC 
GFES glass fiber = 3 vol%
ZnO loading = 17 wt.% 
m:v. 
Challenge gas = 0.5 vol% H2S/H2
 
86 
 
 
 
 
 
Table III.6. Operating conditions: change in m: v with face velocity and length of bed 
 
Ratio Face velocity U (cm/s) Length of bed (inches) 
m:v 1.2 0.5 
2m:2v 2.4 1.0 
4m:4v 4.8 1.5 
8m:8v 9.6 2.0 
 
Lumped K values were obtained from the breakthrough curves and are shown in Fig III.21. A 
linear regression suggests that lumped K increases with U0.56 for lower fiber (3 vol%) fraction in 
MFES and K increases with U0.6 for higher ( 4.5 vol% ) fiber fraction in MFES. The results are 
in agreement with the previous work done[17].    
 
 
 
 
 
 
 
 
 
87 
 
Material 1: Fiber vol= 3%       
Table III.7a. Lumped K values for Material 1 
 
 
 
 
 
 
 
Material 2: Fiber vol = 4.5 % 
Table III.7b. Lumped K values for Material 2 
 
 
 
  
 
 
 
 
U (cm/s) Wt. (gms) Lumped K (s-1) 
1.2 1.86 0.0071 
2.4 0.325 0.0101 
4.8 0.613 0.0156 
9.6 0.828 0.0216 
(cm/s) Wt. (gms) Lumped K (s-1) 
1.2 0.203 0.0089 
2.4 0.425 0.0077 
4.8 0.658 0.2005 
9.6 0.838 0.0313 
 
88 
 
 
Figure III.21a. Relationship between lumped K and face velocity U for material 1 with glass 
fiber fraction = 3 vol% 
 
 
y = 0.0065x0.5395
R? = 0.9974
0
0.005
0.01
0.015
0.02
0.025
0 2 4 6 8 10 12
K 
 (1/
s) 
U (cm/s)
Fiber vol %= 3 T= 400oC 
ZnO loading = 17 wt% . 
Challenge gas = 0.5 vol% H2S/H2
 
89 
 
 
Figure III.21b. Relationship between lumped K and face velocity U for material 2 with glass 
fiber fraction = 4.5 vol% 
 
III.5.5 Effect of Pressure: 
 
The pressure drop data for the two materials under changing face velocities was obtained using a 
pressure cell. The set-up used for the measurement of the pressure drop at 400oC is described in 
the experimental section. The packed bed of small particle size (100-200 um) and large particle 
size (1-2mm) ZnO/SiO2 with ~ 17wt. % loading were compared with the GFES. The 
microfibrous entrapped sorbents give an advantage of the lower pressure drop for significantly 
high breakthrough times via enhanced contact efficiency and thus can be used very effectively for 
the miniaturized desulfurization units. as shown in Fig.III.22 
y = 0.0079x0.6014
R? = 0.9994
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0 2 4 6 8 10 12
K 
(1/
s)
U (cm/s)
Fiber vol% = 4.5, T = 400oC
ZnO loading = 17 wt. %
Challenge gas = 0.5 vol% H2S/H2
 
90 
 
 
 
Figure III.22. Pressure drop data for the packed bed and GFES (Material 1 and 2) at 400oC 
 
Table III.8a. Pressure gradient and log reduction for Material 1 
 
Face Velocity 
(cm/s) 
Pressure 
Gradient ?P/L 
(Pa/m) 
Log10 
Reduction 
LR (t=0) 
?P/LR 
1.2 2022 4.83 418 
2.4 2973 3.45 861 
4.8 5917 8.62 1194 
9.6 11213 11.19 1001 
 
 
Table III.8b. Pressure Gradient and Log reduction values for Material 2 
0
20000
40000
60000
80000
100000
120000
0 2 4 6 8 10 12
Pr
ess
ur
e dr
op 
/Leng
th 
(P
a/m
)
U (cm/s)
packed bed of small particles
Material 1
Material 2
Packed bed of large particles
 
91 
 
 
Face Velocity 
(cm/s) 
Pressure 
Gradient ?P/L 
(Pa/m) 
Log10 
Reduction 
LR (t=0) 
?P/LR 
1.2 1543 2.92 527 
2.4 2579 4.81 536 
4.8 5323 6.59 807 
9.6 9997 9.46 1057 
 
III.5.6  Composite bed design  
 
With the enhanced mass transfer, Microfibrous Entrapped Sorbents (MFES) are targeted at high 
contacting efficiency, high ZnO utilization and high regenerability. It can be directly used in 
miniaturized desulfurizer with a thickness of several centimeters for applications, especially for 
those with low sulfur challenge concentrations. For desulfurization applications with high sulfur 
concentrations, MFES can be used as a polishing layer located at the downstream end of a 
conventional packed bed made of extrudates to form a composite bed.  
 
 
92 
 
 
 
 
 
 
 
 
 
 
 
 
 
a) Packed Bed 
Polishing Layer 
 
 
Polishing Layer 
b) Composite Bed 
 
93 
 
Figure III.23: a) Composite bed test using glass fiber entrapped sorbents as polishing layer. 
Performance of Polishing Sorbent and Packed Bed + Polishing Sorbent @ 1% 
H2S in H2, RT, 2.12 cm/s Breakthrough curves of a 2.5 cm thick packed bed of 
ZnO/SiO2 extrudates and a composite bed (the packed bed followed with a 4mm 
polishing layer). b) Schematic diagram of the packed and composite bed.  
 
 
 
Table III. 9: Composition of the packed bed and polishing layer 
Polishing Layer 
SiO2 = 23%, Void = 75% and Fiber = 2.5% 
Particle Size = 100-200 microns, Bed Thickness = 
4mm, ZnO loading = 19wt.% 
Packed Bed 15 wt.% ZnO/SiO2, Particle Size = 600-800 
microns, Bed thickness = 2.5 cm 
 
The composite bed test was carried out at RT using the Glass Fiber Entrapped Sorbent (GFES) as 
packed bed of larger particle size sorbent followed by the polishing layer of same sorbent of 
smaller particle size (100-200 ?m) entrapped in the fibrous matrix of glass fibers. The Fig. 35 
shows the breakthrough curve for packed bed and composite bed, the change in the breakthrough 
curve for composite bed can be attributed to the presence of the polishing layer at the 
downstream  of the bed, since both the packed beds were tested at similar conditions. This 
 
94 
 
approach enhances the breakthrough time as much as ~ 3 times without adding extra pressure 
drop. The GFES is prepared by wet ?lay paper making method with Silica particles (size: 150-
200 um) entrapped in the mesh of glass fibers (8um dia.). The individual performances of only 
the polishing layer and the packed bed are presented in the Fig. III.23  
III.6 Removal of SO2:  
Promoted ZnO/SiO2 sorbents were tested for removal of 1vol%SO2/Air at the ambient 
conditions. All the samples are made by incipient wetness impregnation method and calcined at 
350 C/1h in air. The Table III.10 Shows the saturation capacity values for all the sorbents and 
Figure III. 24 shows the breakthrough performance.  
 
 
95 
 
Figure III.24: Breakthrough performance of promoted ZnO/SiO2 and FeO/Al2O3 (15wt.%). Test 
Conditions: T= 20 C, Co = 1vol% SO2/Air, Face velocity = 0.53 cm/s  
Table III.10 : Saturation capacity of the sorbents tested for SO2 removal 
Sample Loading (wt.%) Saturation Capacity (mg S/g sorbent) 
Metal oxide 
utilization  
(%) 
Mn0.025Fe0.025Zn0.95O/SiO2 15 32 58 
Cu0.05Zn0.95O/SiO2 15 32 58 
FeO/Al2O3 15 19 42 
 
III.7 Conclusions 
 
Different types of silica materials, including the conventional silica gel (SiO2) and highly ordered 
mesoporous silica (MCM-41 and MCM-48), were impregnated with ZnO and tested for 
desulfurization performance. The lab made ZnO based samples were compared with the 
commercial ZnO. These ZnO-based sorbents were doped with several transition metals including 
Mn, Fe, Co, Ni, and Cu.  The Cu-dopant improved the desulfurization and regeneration 
performance of ZnO/SiO2 significantly. Among the doped ZnO based sorbents, Cu-doped 
ZnO/MCM-41 (Cu0.05ZnO0.95/MCM-41) and ZnO/SiO2 (Cu0.05Zn0.95O/SiO2) are promising 
sorbents for low temperature H2S removal for applications in PEMFCs. At room temperature, 
ZnO/MCM-41 demonstrated a high sulfur capacity (58.12 mg S/g sorbent), which is almost 
twice that of ZnO/SiO2 at similar ZnO loadings. In the presence of moisture, the breakthrough 
capacity of Cu-ZnO/SiO2 first increased up to a maximum of 53 mg S/g sorbent and then 
decreased, whereas, the breakthrough capacity of Cu-ZnO/MCM-41 decreased in the presence of 
moisture. It suggested that MCM-41 support is not suitable for moist gaseous streams. Due to the 
higher capacities achieved, the Cu promoted ZnO/SiO2 can efficiently be used as a non-
 
96 
 
regenerable inline filter before the reformate gases flow to PEMFC. Cu0.05ZnO0.95/MCM-41 are 
promising sorbents for room temperature H2S removal for PEMFC application.  
COS is formed in the reformate streams at high temperatures (T = 400 C). The effect of CO and 
CO2 in the challenge was studied and Cu0.05Zn0.95O/SiO2 showed significant drop in the capacity 
at high temperature (T= 400 C) due to COS formation and Cu is a catalyst for the reaction 
between CO/CO2 and H2S. Novel bimetallic doped sorbent Mn0.025Fe0.025Zn0.95O/SiO2 was 
developed to remove sulfur in the reformate streams with lower COS formations at T = 400 C.  
The pressure drop and kinetic parameters of the microfibrous entrapped ZnO/SiO2 at three 
different face velocities and solid loadings were studied. The trends in lumped K with respect to 
fiber volume fraction in the microfibrous media were studied. The composite bed design give an 
advantage of lower pressure drop for higher breakthrough times via enhanced contacting 
efficiency. 
 Acknowledgements 
This work was supported by the US Army under a U.S. Army contract at Auburn University 
(ARMY-W56HZV-05-C0686) administered through the US Army Tank-Automotive Research, 
Development and Engineering Center (TARDEC).  
 
 
97 
 
Chapter IV: Copper Promoted ZnO/SiO2 Regenerable Sorbents for the Room Temperature 
Removal of H2S from Reformate Gas Streams 
 
Priyanka Dhage, Alexander Samokhvalov, Divya Repala, Evert C. Duin1, and Bruce J. Tatarchuk 
Department of Chemical Engineering, Auburn University, Auburn, AL 36849 (USA)  
1Department of Chemistry and Biochemistry, Auburn University, Auburn, AL 36849  
Abstract 
The Cu-ZnO/SiO2 sorbent for ultradeep adsorptive removal of H2S from the reformate streams at 
room temperature was prepared, tested, and characterization of the active sites was performed. 
The Cu dopant significantly enhances desulfurization capacity of ZnO/SiO2 sorbent at room 
temperature (up to 92 % utilization of ZnO), and maintains a high sulfur uptake capacity upon 
multiple cycles (up to 10) of regeneration by a simple thermal oxidation in air. The ?as-prepared? 
(?calcined?) sorbent contains Cu in the Cu2+ form, O, Si, Zn as ZnO at the coverage <0.2 of 
monolayer, while the ?spent? (?sulfided?) sorbent contains Cu, O, Si, S and Zn as ZnS form, as 
found by XPS. XRD suggests that both zinc and copper compounds of the CuO-ZnO/SiO2 
sorbent are nano-dispersed. The ESR spectroscopy found that the ?calcined? and ?sulfided? 
CuO-ZnO/SiO2 sorbents contain Cu2+ in the single dispersion and coordination state. During H2S 
adsorption, partial reduction of Cu2+ to Cu1+ occurs: the higher Cu concentration in the sorbent, 
the lower the reduction yield of Cu2+ to Cu1+ thus correlating with sulfur uptake capacity. The 
 
98 
 
?deactivated? sorbent (10-11 adsorption/regeneration cycles) is enriched with a different 
chemical form of Cu2+, compared to the ?as-prepared? sorbent. 
Keywords: Desulfurization, H2S, Dopant, ZnO, Cu, XPS, ESR 
 
99 
 
IV.1 Introduction 
 
Electric power generation systems utilizing fuel cells, such as auxiliary power units (APUs) 
for diesel trucks and the military remote power supplies are the subjects of an intense research 
and development recently, due to their portability and high energy efficiency [74]. In these 
power-generating systems, the steam reformers, catalytic partial oxidation (CPO) or autothermal 
reformers (ATR) [75] are used that convert liquid hydrocarbon logistic fuels to H2-enriched 
gaseous reformates. The major chemical components of the reformates are H2, CO, CO2, 
hydrocarbons with low molecular weight, water and sulfur-containing compounds (mainly H2S 
and COS). The sulfur-containing gaseous reformates are known to be poisonous to the catalytic 
systems in fuel processing units (FPUs) and to electrolytes in fuel cells. Typically, to avoid 
poisoning, the feed to the fuel cell should contain less than 1 ppmv of sulfur; therefore, 
desulfurization systems need to be developed. In reformate streams, the most abundant sulfur-
containing compound is H2S. The modern adsorptive desulfurization technologies use the metal 
oxide-based sorbents that can reduce sulfur concentration from several thousand ppmv to the 
sub-ppmv levels [76, 77]. 
The most widely used sorbent for adsorptive removal of H2S from the gas streams is zinc 
oxide ZnO. Key advantages of ZnO are its high sulfur capacity and the favorable sulfidization 
thermodynamics. The non-supported oxide sorbents operating at 500?800 ?C that are based on 
Zn and Ti oxides are known to work efficiently for only a small number of the sulfidization?
regeneration cycles; as the number of cycles increases, the sorbent efficiency declines, and 
mechanical properties of the sorbent become unacceptable [78]. Based on the thermodynamics of 
sulfidization and phase separation [79], the choice of the regenerable sorbents has been often 
 
100 
 
directed towards copper oxides. The main advantage of the Cu-based sorbents is the highly 
favored sulfidization thermodynamics for copper in the oxidation states of +2 or +1, so that the 
equilibrium H2S concentration in the outlet gas of the typical fixed-bed reactor can be lowered 
down to the sub-ppmv levels [1]. However, copper compounds (oxides and sulfides) have 
relatively low melting points, and they are prone to re-growth of the crystallites and to thermally-
induced sintering that lowers the efficiency of the sorbent rapidly [1]. 
On the other hand, copper compounds are useful as dopants for the sorbents compared to 
other transition metal oxides. For instance, addition of the small amounts of copper oxide can 
significantly increase chemical reactivity of zinc ferrites [80], possibly due to copper migration 
from the ?bulk? to the active surface at high temperatures [81]. Copper oxide was used as a 
dopant in a highly-dispersed state, impregnated into the porous supports, such as alumina [82], 
chromia [83] and others. 
In earlier studies, it was shown that both surface and bulk dispersion and oxidation states of 
the Cu dopant are important factors in controlling the activity of the sorbent and its temporal 
stability upon multiple adsorption-regeneration cycles. There are few experimental techniques to 
study dispersion, oxidation and coordination state of the Cu dopant, such as ESR [84, 85] and 
XPS [85, 86]. To study the dopant that is the minority chemical component of the multi-
component sorbent or catalyst, the experimental technique(s) needs to offer: i) a rather high 
sensitivity, ii) the ability to analyze both surface and the ?bulk? of the specimen. Thus, ESR with 
its excellent sensitivity (>1011 spins/sample) and the capability to measure in the ?bulk? is the 
technique of choice. XPS can conveniently complement ESR, since it is the surface-sensitive 
technique. Moreover, upon introducing the Cu dopant ion into the lattice or onto the surface of 
 
101 
 
the ZnO crystallites, the formation of defects is reported, such as oxygen vacancies [87], and 
ESR is well-suited for detecting these [88].  
It is known that, in part, deactivation of the sorbents is due to destruction of the 3D structure 
of the multi-component material due to thermal factors. Therefore, desulfurization sorbents that 
operate at room temperature are expected to show the increased temporal stability as compared 
to their high-temperature analogs, due to the lowered clustering, phase separation and diffusion 
of the dopant ions. Earlier, we prepared and tested the novel ZnO/SiO2 sorbent for H2S and 
carbonyl sulfide COS with the minimized mass transfer resistance [77, 89-92].  
We report here preparation, desulfurization performance upon the multiple regeneration 
cycles, and experimental characterization of the Cu, Zn, O and S sites of the supported doped 
sorbent CuxZn1-xO/SiO2 for the ultradeep removal of H2S that i) achieves >90% of the theoretical 
sulfur uptake capacity at room temperature, ii) reduces sulfur concentration from ~1000 ppm to 
< 1 ppm, iii) can be easily regenerated multiple times by simple heating in air without a 
significant loss of performance. 
IV.2 Experimental 
 
The doped ZnO-based sorbent with the formula Cu0.05Zn0.95O/SiO2 was prepared by an 
incipient impregnation of the commercial high surface area silica (Fischer Scientific Inc., surface 
area ~550 m2/g, powder 100-200 ?m), with metal nitrates as the 2 M solutions in water used as 
precursors. Total metal loading was 15, 21 or 25 wt. %. After impregnation and drying, the 
samples were calcined in air at 350 oC; these are referred to as the ?calcined? samples. 
In the desulfurization experiments, the challenge gas was the model reformate with an inlet 
concentration of 1 or 2 vol. % H2S, 33 vol. % CO or CO2, balance H2 (UHP grade from Airgas 
 
102 
 
South, Inc.), at a face velocity of 2.12 cm/s, corresponding to the volumetric gas flow rate of 0.1 
slpm. The experimental setp-up and procedure are described in Chapter II. Regeneration of the 
?sulfided?, i.e. ?spent? sorbent was performed at 550 oC in the flowing air at a flow rate of 50 
cc/min. The temperature of the furnace during the experiments was maintained using a PID 
temperature controller. The gas flow rates were controlled by mass flow controllers (Omega 
FMA 2405 Alaborg GFC1718). XRD, N2 adsorption-desorption isotherms, XPS and ESR are 
used to characterize the sorbents. The description of the techniques and conditions at which the 
equipments are operated is given in Chapter II.  
IV.3. Results and Discussion 
 
IV.3.1. Desulfurization Performance of the Sorbents 
 
Figure IV.1 shows desulfurization performance of the undoped supported ZnO/SiO2 sorbent 
prepared by us vs. the commercial ZnO extrudates (BASF and Sud-Chemie). Table IV.1 shows 
sulfur uptake capacity (g sulfur / g sorbent) and utilization of ZnO in the sulfidization reaction 
(% of the theoretical value for the ZnS stoichiometry). 
Table IV.1: Sulfur capacities of the sorbents M0.05ZnO0.95/SiO2  
(at metal loading 21 wt. %). 
M = dopant, M0.05ZnO0.95/SiO2 Saturation Capacity, g S/g sorbent ZnO Utilization  at Saturation, % 
Mn 0.050 60 
Fe 0.069 83 
Co 0.064 77 
 
103 
 
Ni 0.053 64 
Cu 0.077 93 
None 0.053 64 
 
 
 
 
Figure IV.1:  Breakthrough Curves for Commercial ZnO (BASF and Sud-Chemie) with 21 wt.% 
ZnO/SiO2 and Cu0.05ZnO0.95/SiO2. Test Conditions : Co = 1 vol%H2S/H2, T= 20C, Face velocity 
= 2.12 cm/s 
 
 The supported ZnO/SiO2 sorbent showed better performance over both commercial ZnO-based 
sorbents. XRD of the ZnO/SiO2 sorbent did not detect lines of neither zinc silicate Zn2SiO4 [93] 
 
104 
 
nor zinc oxide ZnO. These findings and the high sulfur capacity of the ZnO/SiO2 sorbent indicate 
the nano-dispersed form of the supported ZnO, with a typical crystalline size of ~50 ? or less.  
Table IV.2: Sulfur capacities and ZnO utilization of the doped sorbents Cu0.05ZnO0.95/SiO2 vs. 
the un-doped ZnO/SiO2 sorbent.  
 
Sorbent ZnO Loading (w/w%) 
Saturation 
Capacity 
(g S/g sorbent) 
ZnO Utilization 
(%) 
BASF (SG-901)* ZnO 90 0.019 5.34 
Sud-Chemie*  
(G-72E) ZnO 90 0.032 9.00 
ZnO/SiO2 21 0.053 63.88 
ZnO/SiO2 15 0.032 54.00 
Cu0.05ZnO0.95/SiO2 21 0.077 92.81 
Cu0.05ZnO0.95/SiO2 15 0.043 72.56 
Cu0.2ZnO0.8/SiO2 21 0.078 94.02 
Cu0.2ZnO0.8/SiO2 15 0.045 75.94 
*Commercial ZnO is crushed to the same size 100-200 ?m as the supported sorbent 
 
The Cu-doped sorbent Cu0.05ZnO0.95/SiO2 showed an enhanced sulfur adsorption capacity 
over all other sorbents (M = Mn, Fe, Co, Ni), over the un-doped ZnO/SiO2 sorbent and over the 
un-supported commercial ZnO-based sorbents. Specifically, doped sorbent Cu-ZnO/SiO2 shows 
a ~45 % improvement in the sulfur capacity over the undoped ZnO/SiO2 sorbent. The XRD of 
the ?calcined? doped Cu0.05ZnO0.95/SiO2 sorbent was performed, and no lines due to any copper 
compound were found. This implies a high degree of dispersion of the Cu dopant in the 
Cu0.05ZnO0.95/SiO2 sorbent. In the XRD of the ?sulfided? sorbent, lines of CuS, Cu2S and 
metallic Cu were not identified as well, that indicates the high dispersion of the Cu dopant in the 
?sulfided? sorbent and the absence of phase separation upon desulfurization.  
 
105 
 
IV.3.2. Performance of the Sorbents upon Multiple Regeneration Cycles  
 
Figure IV.2 shows the breakthrough curves for the Cu0.05ZnO0.95/SiO2 sorbent upon H2S 
adsorption / regeneration cycles, as compared to the ?fresh? sorbent. Upon multiple cycles of 
?desulfurization-regeneration?, the sorbent retains up to 70 % of the initial sulfur capacity.  
 
 
Figure IV.2 Breakthrough curves for Regeneration of Cu0.05ZnO0.95/SiO2.Test Conditions: 
Calcination Temp = 350 C/Air/1h, Adsorption at 20 C, Co = 1vol% H2S/H2, Regeneration at : 
550 C, Air/1h.  
IV.3.3. Structural Characterization of the Sorbents 
 
Table IV.3 shows surface area and pore volume of the sorbents. At ZnO loading on SiO2 of 15 
wt. %, there is ~16 % and 30 % reduction of surface area and pore volume respectively, as 
 
106 
 
compared to silica, and with the further increase of ZnO loading (at 25 wt. %), the surface area 
and pore volume are further reduced. On the other hand, upon doping the ZnO/SiO2 sorbent (15 
wt. % of ZnO) with Cu to obtain the Cu0.05-Zn0.95/SiO2 sorbent, there is only a marginal change 
in surface area and pore volume. The latter finding indicates that structural characteristics of the 
ZnO/SiO2 sorbent do not significantly change when Cu dopant is added. 
Table IV.3: Surface Area and Pore Volume Data analyzed using N2 Adsorption-Desorption 
Curves 
Sample  ZnO Loading 
(w/w%)  
Sg  Vg 
(m2/g)  (cc/g)  
SiO2 0 550 0.792 
ZnO/SiO2 15 460 0.558 
ZnO/SiO2 25 330 0.486 
Cu-ZnO/SiO2 15 450 0.592 
SiO2 21 330 1.65 
ZnO/SiO2 21 244 1.04 
 
IV.3.4. Characterization of the Sorbents by XPS 
 
Figure IV.3 shows the XPS Zn 2p (Figure 3A), Zn L3M45M45 (Figure 3B) and O 1s (Figure 
IV.3C) lines of the ?calcined? sorbent Cu0.2Zn0.8O/SiO2. The following elements are identified in 
the XPS survey spectrum (data not shown): Cu, Zn, Si, O, and spurious carbon as expected. No 
residual nitrogen was detected that indicates the complete decomposition of metal nitrate 
precursors used. The samples show a strong electrostatic charging (~5 eV), as expected for the 
electrically insulating silica support. The sorbent of the formula Cu0.05Zn0.95O/SiO2 has the 
similar XPS spectrum, except that the XPS signal from the Cu dopant is too low to be reliably 
interpreted. The BE of the Zn 2p3/2 line is 1022.1 eV that is consistent with the reported BE of 
 
107 
 
1022.0-1022.1 eV for Zn2+ form [80] in zinc oxide catalyst [94] and of BE=1022.4 eV in pure 
ZnO [95]. The BE of the Zn L3M45M45 peak is found by us to be 265.6 eV (with Mg anode). The 
Auger Parameter (AP) is useful for processing XPS spectra of the electrically insulating samples 
such as supported sorbents and catalysts, since its value is independent on the electrostatic 
charging of the specimens [96]. We calculated the APZn to find the coordination state of Zn in 
the ?calcined? sorbent, by using the formula APZn = 1253.6 + BE(Zn 2p3/2) - BE(Zn L3M45M45) 
=  2010.1 eV. This corresponds to ZnO as expected whose APZn is 2010.25 eV [97]. On the other 
hand, for the ZnO-SiO2 nano-composites that were prepared by the sol-gel technique and that 
were shown to contain ZnO nanoparticles embedded into the SiO2 matrix with the significant 
concentration of Zn-O-Si bonds, APZn is as low as 2009.1 eV [97]. We conclude that in the 
?calcined? Cu-ZnO/SiO2 sorbent, Zn is present in the form of ZnO nanoparticles located on the 
SiO2 surface, rather than included into the lattice of SiO2. 
 
108 
 
 
Figure IV.3: XPS Spectra of Calcined Cu0.05ZnO0.95/SiO2 
Figure IV.3C shows the O 1s peak that can be well fitted as the singlet (our attempts to fit it as 
spectral doublet were unsuccessful). The BE is 531.9 eV that is close to the reported value of 
531.5 eV for silicon oxide SiO2 [98]. It was reported that for the pure silica that was calcined 
with  flowing oxygen at 673 K, the O 1s peak is the singlet [99], while for the SiO2 thin films, 
both bridging oxygen atoms (Si-O-Si, BOs) at 531.5 eV and the non-bridging atoms (Si-O-, 
NBOs) at the lower BE are found as shoulders of the O 1s peak [98]. On the other hand, the BE 
 
109 
 
of O 1s in zinc oxide ZnO is as low as 529.7 eV [100]. From these data, we conclude that the O 
1s peak in Figure IV.3C belongs mostly to the bridging oxygen of the silica support. 
We have calculated the atomic ratios Cu/Zn, Zn/Si and O/Si from our XPS data, following the 
standard formula that includes the areas of the XPS peaks, the photoionization cross-sections ? 
and the photoelectron mean free paths (MFPs) [95]. The following ratios are found for the 
?calcined? sorbent of the nominal formula Cu0.2Zn0.8O/SiO2: O/Si=2.00, Cu/Zn=0.30; 
Zn/Si=0.20. The atomic ratio O/Si=2.00 supports our conclusion above that the O 1s peak is 
mostly due to oxygen of the silica support. The atomic ratio Cu/Zn=0.30 is somewhat higher (by 
20 %) than the theoretical atomic ratio of 0.25 for the sorbent Cu0.2Zn0.8O/SiO2. The deviation of 
20% must be attributed to the standard error bar of the XPS measurement of ~10% and the 
respectively larger error bars for the atomic ratio; the error bar may also include the systematic 
errors due to the values of ? and MFPs used. On the other hand, the measured atomic ratio of 
Zn/Si=0.20 is higher (by 53 %) than the nominal atomic ratio of Zn/Si=0.13 of the sorbent 
containing 15 wt. % ZnO supported on SiO2. This deviation is significantly higher than the 
typical error of the XPS measurements, as mentioned above. Our explanation is that the atomic 
ratio determined by the surface sensitive XPS does not reflect the ?bulk? atomic ratio 
Zn/Si=0.13. If Zn is located on the surface of silica as the nano-islands (or nano-particles), the 
Zn/Si ratio determined by XPS should be higher than the ?bulk? ratio Zn/Si=0.13, due to the 
attenuation of the XPS signal of silicon support, consistently with our findings. However, only 
the small faction of the SiO2 surface is covered by the ZnO, since the O 1s XPS peak mostly 
belongs to SiO2, not to ZnO as shown by us above. From the combined XRD and XPS data is not 
possible to determine the exact coverage of the surface of SiO2 with ZnO and the size of the 
nano-crystallites formed. Assuming 100 % dispersion of ZnO, the uniform dispersion of ZnO 
 
110 
 
over all available surface area of SiO2 and the Zn-O bond length of 1.7 ?, the nominal coverage 
of ZnO is as low as 0.04 of a monolayer. The real coverage of ZnO is definitely higher, and is 
determined to be approximately 0.2 of a monolayer, as from our XPS data.  
We have measured the XPS Cu 2p1/2 and 2p3/2 lines (data not shown) of the ?calcined? sorbent 
Cu0.2Zn0.8O/SiO2. There are the ?shake-up? peaks in the spectra thus indicating the presence of 
Cu in the Cu2+ form. Based on the thermodynamic considerations and the ?history? of 
calcination in air, all Cu is expected to be present in the Cu2+ state as CuO, rather than in the 
Cu1+ state. For the XPS spectra of CuO, the ratio of the area of the Cu shake-up peak at ~942 eV 
to the area of the Cu 2p3/2 peak at 933.6 eV is 0.53 [101]. However, in our XPS spectra, this 
ratio is less thus indicating the presence of both Cu1+ and Cu2+ forms. We conclude that artificial 
XPS-induced reduction of the Cu2+ form to Cu1+ form occurred. Indeed,  XPS-induced reduction 
of Cu2+ in Cu-containing specimens due to the X-Rays, heat and secondary electrons was 
reported in the lietrature [102, 103]. Moreover, the conversion of the octahedral Cu2+ into the 
tetrahedral Cu2+ under the X-Rays radiation in the XPS experiments was reported in copper-
exchanged X- and Y-type sodium zeolites [101]. Thus, a  complementary non-destructive 
spectroscopic technique is needed to be used to learn about the speciation of the Cu dopant in the 
Cu-ZnO/SiO2 sorbents. 
Figure IV.4 shows the XPS Zn 2p (Figure 4A), Zn L3M45M45 (Figure 4B) and O 1s (Figure 4C) 
lines of the ?sulfided? sorbent of the formula Cu0.2Zn0.8O/SiO2. The following elements are 
identified in the XPS survey spectrum (data not shown): Cu, Zn, S, Si, O and spurious carbon as 
expected. No nitrogen was detected as expected. The samples show a strong electrostatic 
charging (~5 eV), as expected for the electrically insulating material, therefore there is no 
significant amount of metallic copper in the samples. The binding energy (BE) of Zn 2p3/2 line 
 
111 
 
is measured to be 1021.9 eV. Binding energy (BE) of Zn 2p3/2 line is not very characteristic of 
coordination environment of zinc in ZnO vs. ZnS, with the difference being less than 0.5 eV 
[103]. The binding energy of the Auger L3M45M45 line of zinc is 263.9 eV. We calculated the 
APZn to find the coordination state of Zn in the ?sulfided? samples, by using the formula APZn = 
1253.6 + BE(Zn 2p3/2) - BE(Zn L3M45M45) =  2011.6 eV. This corresponds to ZnS whose APZn 
is 2011.44 eV [104], while for ZnO, the APZn is as low as 2010.25 eV [97]. Formation of ZnS is 
consistent with the high sulphur uptake capacity of the CuxZn1-xSiO2 sorbents upon sulfidization, 
~92% of the theoretical value. 
 
112 
 
 
Figure IV.4: XPS Spectra of sulfided sorbents Cu0.05ZnO0.95/SiO2 
 
We have measured the XPS spectrum of the Cu 2p1/2 and 2p3/2 lines (data not shown) of the 
?sulfided? sorbent Cu0.2Zn0.8O/SiO2. In the Cu0.05Zn0.95O/SiO2 sorbent, the signal from Cu is too 
small to be reliably obtained. We have not found the XPS shake-up peaks of the Cu 2p lines in 
the spectra of the ?sulfided? sorbent that indicates the absence of CuO. Further, the literature 
states that the expected sulfidization product CuS has no XPS shake-up peaks [105]. Using the 
 
113 
 
Auger L3M45M45 line of Cu might be the choice, however, Auger lines are usually much broader 
that XPS lines, and fitting Auger line with the multiplet due to several components of Cu, from 
CuS, Cu2S and Cu2O is not reliable; in our measurements, the Cu Auger line was too small to be 
reliably interpreted. Moreover, the BE of the Cu 2p3/2 peak [106] in CuS (932.3 eV) is virtually 
identical to the one in Cu2S [107], so that these forms of Cu cannot be distinguished by XPS. In 
addition, the XPS-induced sample damage of the Cu-containing specimens due to the X-Rays, 
heat and secondary electrons was reported [103] as manifested by the reduction of Cu2+ to Cu1+. 
Thus, the complementary non-destructive spectroscopic technique was applied to learn more 
about the speciation of the Cu dopant. 
IV.3.4. Characterization of the Sorbents by ESR 
 
Figure IV.5 shows ESR spectra of the ?calcined? sorbent CuxZn1-xO/SiO2. Figure IV.5A 
corresponds to the Cu1.0Zn0.0O/SiO2, Figure IV.5B ? Cu0.1Zn0.9O/SiO2, Figure IV.5C - 
Cu0.01Zn0.99O/SiO2 and Figure IV.5D - Cu0.001Zn0.999O/SiO2. Silica support that was prepared 
similarly to the ?calcined? sorbent, except that Cu2+ salt was not used, shows no ESR spectrum, 
as expected. No spectral lines due to the paramagnetic Cu0 atoms are found in the spectra of the 
?calcined? sorbents CuxZn1-xO/SiO2, as expected. In addition, no spectral lines of any Reactive 
Oxygen Species (ROS) or oxygen vacancies [87] are present in the ESR spectra. The ?calcined? 
sorbent of the formula Cu0.0Zn1.0O/SiO2 shows no ESR spectrum, thus confirming that the 
spectral multiplet in Figure 5 belongs to Cu2+. 
 
114 
 
 
Figure IV.5: ESR spectra of the ?calcined? sorbent CuxZn1-xO/SiO2 Figure 5A - 
Cu1.0Zn0.0O/SiO2, Figure 5B ? Cu0.1Zn0.9O/SiO2, Figure 5C - Cu0.01Zn0.99O/SiO2 and Figure 5D - 
Cu0.001Zn0.999O/SiO2. 
The ESR spectral pattern of Cu2+ is rather complicated,  both due to hyperfine splitting [108] and 
presence of two major stable isotopes, 63Cu (mole fraction 0.6915, nuclear spin 3/2) and 65Cu 
(mole fraction 0.3085, nuclear spin 3/2) that both contribute to the multiplet observed. The ESR 
spectrum of Cu2+ in Cu0.05Zn0.95O/SiO2 was simulated as the single kind of Cu2+ species (Figure 
 
115 
 
6), and was found to have g values of 2.077, 2.051 and 2.349, consistently with the literature 
reports of the copper-zinc oxide catalysts [84]. The ESR spectra of the ?calcined? CuxZn1-
xO/SiO2 sorbent show broadening of the spectral features of the Cu2+, as concentration of Cu2+ 
increases (Figure IV.5). This behavior is well known [109], and it was attributed to interactions 
between isolated Cu2+ ions. 
Various forms of Cu species are found to exist in both supported and unsupported copper-
containing oxides: nanoclusters [110], isolated Cu2+ ions [110, 111], binuclear oxygen-bridged 
ion pairs [112] such as [Cu-O-Cu]2+. The high probability of formation of the Cu2+-OH-Cu1+ 
bridge structures was found by calculations [113]. Therefore, the straightforward interpretation 
of the ESR spectrum is difficult [111], thus some chemical tests needed to be performed in order 
to assign the spectrum to the certain Cu2+ species. 
First, we have checked if evacuation of the ?calcined? sorbent Cu0.05Zn0.95O/SiO2 in the ESR 
test-tube down to 10-6 Torr with the subsequent readmission of air affects the ESR spectrum of 
Cu2+. It was reported that Cu2+ ions present on surface of the Cu-Zn-Al mixed oxide catalysts 
cause the significant broadening of the ESR signal upon admission of air, due to interaction of 
Cu2+ with the adsorbate [84]. We have observed no line narrowing of the Cu2+ signal upon 
outgassing that indicates that the majority of Cu2+ ions in the ?calcined? sorbents CuxZn1-xO/ 
SiO2 are not on the surface of the sorbent. No other spectral lines appeared in the ESR spectrum 
of the sorbent upon evacuation and re-admission of air. The latter finding indicates that the ROS, 
including oxygen vacancies and superoxide radicals [114] are not present in the significant 
amounts in the ?calcined? sorbents CuxZn1-xO/ SiO2 and are unlikely to play a role in the surface 
chemistry of the subsequent H2S adsorption. 
 
116 
 
 Next, we have checked if reduction of Cu2+ ions with CO changes the ESR signal of the 
?calcined? sorbent Cu0.05Zn0.95O/SiO2. It was reported [84] that CO exhibit a high reactivity 
towards the surface Cu2+ ions in the Cu-containing catalysts at room temperature, reducing Cu2+ 
to Cu1+ and even to Cu0. We have not observed any changes in the ESR spectrum of Cu2+ after 
reduction of the ?calcined? sorbent Cu0.05Zn0.95O/ SiO2 with CO at room temperature. This 
finding is consistent with the conclusion that majority of Cu2+ ions are not on the surface. This 
excludes the possibility of the CuO-ZnO phase separation, formation of the ?core-shell? 
supported nanoparticles or the islands of Cu oxides. This finding also indicates that no reduction 
of the Cu dopant in the ?calcined? sorbent CuxZn1-xO/SiO2 occurs due to chemical reaction with 
the CO component of the H2S containing reformates. 
It was also reported that Cu2+ ions in the CuO-ZnO catalysts are not reduced by H2 at room 
temperature if Cu2+ ions are well-dispersed in the binary oxide [84]. We have not observed any 
significant changes in the ESR spectrum of Cu2+ after reduction of the ?calcined? sorbent 
Cu0.05Zn0.95O/SiO2 with H2 at room temperature. This finding indicates the following: i) The 
Cu2+ species present in the ?calcined? sorbent are likely to be the isolated Cu2+ ions; ii) Cu2+ ions 
are not preferentially located on surface of the sorbent; iii) no reduction of Cu2+ with H2 
component of the model reformate occurs upon H2S adsorption, iv) Cu2O is unlikely to be 
present in the ?sulfided? sorbent, and any Cu1+ found in the ?sulfided? sorbent is formed upon 
chemical reaction with H2S, not with H2 component of the reformate. The latter finding allows to 
expect that variations of the H2 concentration in the reformate would not affect the reactions of 
the Cu dopant in the CuxZn1-xO/SiO2 desulfurization sorbents.  
The overall shape of the ESR signal of Cu2+ in the ?calcined? CuxZn1-xO/SiO2 sorbent is similar 
to that of the polycrystalline sample containing isolated ions Cu2+ in a site with an axial 
 
117 
 
symmetry [114]. It is also similar to the ESR signal of Cu2+ ions in the site of octahedral 
symmetry with tetragonal distortions, namely, with axis lengthening and planar shortening [110]. 
We conclude that there is only one kind of Cu2+ ions in the ?calcined? sorbent CuxZn1-xO/SiO2 
that is Cu2+ ions well-dispersed in the ?bulk? of the sorbent. 
 
Figure IV.6: ESR spectrum of Cu2+ in Cu0.05Zn0.95O/SiO2 simulated as the single kind of Cu2+ 
species. 
The ESR spectra of the sulfide Cu (0,0.1,0.2 and 1) was also obtained. The spectral shapes of the 
signals of Cu2+ ions are similar to those of the ?calcined? sorbents (Figure IV.5), although the 
ESR intensities are lower for the ?sulfided? sorbents. We did not observe the ESR patterns of the 
Cu0 atoms, the ROS species or oxygen vacancies. The findings indicate a partial reduction of the 
ESR-active Cu2+ form to the ESR-silent Cu1+ form upon the interaction of the ?calcined? sorbent 
with the H2S component of the reformate. In the ESR spectra of the ?sulfided? sorbent, there was 
 
118 
 
an additional ESR triplet of the low intensity that was also found in the ESR spectrum of the 
silica support treated with H2S in hydrogen. This ESR triplet is sensitive to admission of air to 
the ESR test-tube, and is tentatively assigned to the HS- or S2- anion radical or similar species 
[115, 116], however, its exact structure is not known. Upon thermal oxidative regeneration of the 
?sulfided? sorbent, the ESR triplet disappears that supports its assignment to the reduced, rather 
than oxidized, form of radical species. In order to reliably determine the ESR signal of Cu2+ in 
the ?sulfided? sorbent, the triplet was subtracted from the spectra, and the spectral reminder that 
belongs only to Cu2+ was doubly-integrated as usual.  
Figure IV.7A shows the doubly integrated (DIN) ESR signal of Cu2+ that is proportional to molar 
concentration of Cu2+ in the ?calcined? vs. ?sulfided? sorbents CuxZn1-xO/SiO2 (x=0.001, 0.01, 
0.05, 0.1 and 1). Figure 7B shows the yield Y of chemical reaction of the reduction of Cu2+ to 
Cu1+ upon the interaction with the H2S component of the reformate. 
 
Y = [Cu2+calc.] - [Cu2+sulf.]   /   [Cu2+calc.]  (1) 
 
where [Cu2+calc] is molar concentration of Cu2+ in the ?calcined? sorbent; [Cu2+sulf.] is molar 
concentration of Cu2+ in the ?sulfided? sorbent. The Cu2+ reduction yield Y is dependent on 
concentration of copper in the ?calcined? specimens: the higher the concentration of copper, the 
less efficient the reduction of Cu2+ into Cu1+. This dependence correlates with the sulfur uptake 
capacity of the CuxZn1-xO/SiO2 sorbent, namely, sulfur capacity is significantly reduced for the 
samples with the high concentration of Cu, ~x>0.2. This correlation suggests that the highly 
dispersed Cu2+ ions in the CuxZn1-xO/SiO2 sorbent act as promoters of the adsorption of H2S by 
the host material ZnO and are themselves converted to copper sulfides.  
 
119 
 
 
Figure IV.7:  Figure 7A shows the relative signal intensity of Cu2+ proportional to molar 
concentration of Cu2+ in the ?calcined? vs. ?sulfided? sorbents CuxZn1-xO/SiO2 (x=0.001, 0.01, 
0.05, 0.1 and 1). Figure 7B shows the yield Y of chemical reaction of the reduction of Cu2+ to 
Cu1+ upon the interaction with the H2S component of the reformate 
 We have noted that upon the multiple desulfurization-regeneration cycles, there is a 
reduction of the sulfur uptake capacity (Figure IV.2). We have investigated if the reduction of 
the sulfur capacity upon multiple cycles is accompanied by the changes in the ESR signal of 
 
120 
 
Cu2+ dopant ions. Figure IV.8 shows ESR spectrum of Cu2+ in the sorbent Cu0.05-Zn0.95O/SiO2, 
?calcined? as-prepared vs. ?calcined? upon 10 cycles of desulfurization-regeneration. Several 
changes can be noted. First, the hyperfine structure of Cu2+ at 2600-3100 G is less pronounced 
for the multiply-regenerated sorbent. This suggests clustering of the isolated Cu2+ ions or the 
formation of a second kind of Cu2+. Second, the peak at ~3300 G shows different shapes, 
namely, the low-field shoulder at ~3285 G is stronger for the multiply-regenerated ?calcined? 
sorbent, and the high-field shoulder at ~3305 G is stronger for the ?fresh calcined? sorbent. 
Those differences indicate that some changes occur to the Cu2+ dopant ion upon multiple 
adsorption-regeneration cycles. Specifically, spectral changes could occur due to 1) aggregation 
of Cu2+ ions into nano-clusters or islands (phase separation); 2) diffusion of Cu2+ ions towards 
the surface of the supported sorbent and formation of the surface Cu2+; 3) diffusion towards the 
SiO2 interface, forming some kind of the interfacial copper silicate.  
 
 
121 
 
Figure IV.8: ESR spectrum of Cu2+ in the sorbent Cu0.05-Zn0.95O/SiO2, ?calcined? as-prepared 
vs. ?calcined? upon 10 cycles of desulfurization-regeneration. 
 
The surface complex with three Cu-O-Si bonds shows the largest shift of the ESR peak 
maximum towards the low field, compared to the surface complexes with two Cu-O-Si bonds. 
Moreover, the surface complex with three Cu-O-Si bonds shows the spectral shift compared to 
the Cu species that are not coordinated with surface of silicon oxide, i.e ?bulk? form of Cu2+. 
Based on the assignments from the literature [117], we propose that spectral change of the ESR 
signal of Cu2+ upon multiple desulfurization-regeneration (Figure IV.8) could be due to forming 
at least one Cu-O-Si bond, upon the thermally-induced diffusion of Cu2+ ions towards the 
interface with silica support. The alternative explanation originates, when we compare the 
spectral shape of the regenerated sorbent Cu0.05-Zn0.95O/SiO2 (Figure 8) with ESR spectrum of 
the ?calcined? sorbent that contains only copper and no zinc, i.e. Cu1.00-Zn0.00O/SiO2, Figure 5A. 
Both spectra show the same pattern, namely, the stronger shoulder at the low field, ~3285 G and 
the weaker shoulder at the higher field, ~3305 G. Such similarity implies that in the multiply-
regenerated sorbent, agglomeration of Cu2+ could also take place. Additional experiments are 
underway to determine in more detail the atomic level structure of the Cu2+ centers in the 
multiply-regenerated sorbent, as function of the ?aging? of the sorbent. One of the 
complementary approaches is to use the well-designed model sorbents, such as thin films of 
binary oxides, CuO-ZnO on the oxidized silicon wafers, in their ?calcined? vs. ?sulfided? form, 
to learn in the systematic fashion about the thermally-induced sintering, diffusion and surface 
chemical reactions upon sulfidization and regeneration. 
 
 
122 
 
IV.4 Conclusions 
 
The Cu dopant enhances utilization of the ZnO active phase of the novel ZnO/SiO2 sorbent 
during adsorptive desulfurization of the reformate streams at room temperature, from 64 % to 92 
%, and maintains a high sulfur uptake capacity upon multiple cycles of a simple thermal 
oxidative regeneration of the ?spent? sorbent in air (up to 10 cycles). Both zinc and copper 
oxides are nano-dispersed in the Cu-ZnO/SiO2 sorbent, and both the ?calcined? and ?sulfided? 
forms of the sorbent contain Cu2+ in the single dispersion and coordination state. The higher 
concentration of the Cu promoter in the Cu-ZnO/SiO2 sorbent, the lower the reduction yield of 
Cu2+ to Cu1+ upon adsorption of H2S that correlates with sulfur uptake capacity. The 
?deactivated? sorbent (upon 10 adsorption-regeneration cycles) is enriched with the different 
chemical form of Cu2+ dopant, as compared to the ?as-prepared? sorbent.  
 
Acknowledgement 
Authors would like to thank the US Army (TARDEC Contract W56HZV-05-C-0686) for the 
financial support of this work.  
 
 
 
 
 
 
 
 
 
123 
 
 
Chapter V: Regenerable Fe-Mn-ZnO/SiO2 sorbents for Room Temperature Removal of H2S 
from Fuel Reformates: Performance, Active sites and Operando studies 
 
Priyanka Dhage, Alexander Samokhvalov1, Divya Repala, Evert C. Duin2, and Bruce J. 
Tatarchuk1 
Department of Chemical Engineering, Auburn University, Auburn, AL 36849 
1Department of Chemistry, University of Rutgers, Camden, NJ 08102 
2Department of Chemistry and Biochemistry, Auburn University, Auburn, AL 36849  
Abstract 
The Fe- and Mn-promoted H2S sorbents Fex-Mny-Zn1-x-yO/SiO2 (x, y=0, 0.025) for the ultradeep 
desulfurization of model reformates at room temperature were prepared, tested and 
characterized. Their sulfur uptake capacity significantly exceeds that of both commercial 
unsupported ZnO sorbents (by 60 %) and of the un-promoted supported sorbent ZnO/SiO2 (by 
30 %). Sulfur sorption capacity and the breakthrough characteristics remain satisfactory after up 
to 10 cycles of adsorption/regeneration, with regeneration performed by a simple heating in air. 
XRD shows that both ?calcined? and ?spent? sorbents contain the nano-dispersed forms of ZnO, 
Fe and Mn and XPS confirms the conversion of the supported ZnO phase to ZnS. The ?calcined? 
sorbent contains Fe3+ and Mn3+ ions, while upon H2S adsorption, their reduction to Fe2+ and 
Mn2+ occurs. Fe3+ ions are believed to occupy the surface of the supported ZnO nanocrystallites, 
while Mn3+ ions are distributed uniformly within ZnO. 
 
124 
 
Keywords: Desulfurization, H2S, Promoter, ZnO, Fe, Mn, XPS, ESR 
V.1. Introduction 
 
Fuel cell-based systems for electric power generation, such as auxiliary power units (APUs) 
for diesel trucks and remote power supplies for the military, offer both portability and high 
energy efficiency, and they have been intensively studied in the past decade [74]. Recently, high 
power fuel cells were developed for non-transport applications as well, ranging from kilowatt 
[118] to megawatt [119] power generating systems. Intensively developed fuel cell technologies 
constitute the basis of a potential energy-efficient and environmentally benign ?hydrogen 
economy? [120]. The core components of the fuel cell-based power generation systems are: i) 
steam reformers, ii) catalytic partial oxidation (CPO) reformers and iii) autothermal reformers 
(ATR) [75] that convert liquid hydrocarbon logistic fuels to the H2-enriched gaseous reformates. 
H2S impurity in the reformates is known to be poisonous to the catalytic systems of fuel 
processing units (FPUs), fuel cell electrodes and electrolytes in the Poly Electrolyte Membrane 
fuel cells (PEMFCs). In order to avoid ?sulfur poisoning,? reformates should contain < 1 ppmv 
or even < 60 ppb sulfur as for the PEMFCs [121], and robust and inexpensive desulfurization 
materials and regimes need to be developed. Modern adsorptive desulfurization technologies that 
use metal oxide-based H2S sorbents can reduce sulfur concentration in the gas phase from 
several thousand ppmw down to the sub-ppmv levels [76, 77]. However, the majority of such 
sorbents were developed for the hot-gas cleanup (HGC) of the streams of the integrated 
gasification combined cycle (IGCC); therefore, such sorbents operate at high temperatures, 
~500-800 ?C [122]. 
 
125 
 
The best material for adsorptive removal of H2S is ZnO [123] because of its favorable 
sulfidation thermodynamics and high sulfur capacity (by weight). However, a serious problem of 
the high temperature (> 500 ?C) H2S adsorbents is the reduction of ZnO by hydrogen into 
metallic zinc and evaporation of the latter [122]. Several oxides of other metals such as iron, 
vanadium, zinc, copper, manganese and molybdenum have been proposed as high-temperature 
desulfurization sorbents since the 1970s [124]. Chemical and structural transformations of those 
oxides upon desulfurization/regeneration were investigated; for instance, it is known that in the 
environment of the IGCC gasifier, Mn3O4 form is readily reduced to MnO and the latter reacts 
with H2S at the high temperatures [122]. Iron oxides have also been extensively investigated 
since the 1970s; iron oxide-based H2S sorbents have high sulfur capacity and reactivity towards 
H2S. However the equilibrium concentration of H2S is as high as 100 ppmw. In addition, a  
number of the degradation processes occur above ~ 500 ?C, most importantly reduction of Fe3O4 
to FeO [122]. Mixed metal oxide sorbents for high temperature desulfurization of coal gases 
were extensively reviewed in the past [125, 126]. 
Recently, active research and development efforts have been directed towards ?mid-
temperature? H2S adsorbents [123]. For instance, iron oxide sorbents supported on silica provide 
improved stability vs. unsupported iron oxides for adsorptive desulfurization at the ?mid-
temperature? range, ~400 ?C [122]. The major research objective of the studies of the 
?promoted? desulfurization sorbents is to provide better attrition resistance, higher sulfidization 
capacity, lower equilibrium concentrations of H2S and COS and an ability to remove multiple 
gas contaminants at the same time [123]. 
It is known that, in part, temporal deactivation of the sorbents is due to the destruction of the 
unique 3D structure of the material due to thermal factors. Therefore, desulfurization sorbents 
 
126 
 
that operate at room or slightly elevated temperatures are expected to show increased temporal 
stability as compared to their high- and mid-temperature analogs. Recently, there is increased 
interest in the ?low temperature? H2S adsorbents that operate between room temperature and 
~100 ?C [127-129]. For instance, we reported preparation and testing of novel ZnO/SiO2 sorbents 
for H2S and carbonyl sulfide COS with the minimized mass transfer resistance [77, 89-92] that 
operate at room temperature and retain their high desulfurization capacity after >10 
desulfurization/regeneration cycles, with the regeneration performed by the inexpensive and 
robust calcination in the flowing air. 
The typical desulfurization promoters of the ZnO-based H2S sorbents are cations of transition 
metals (TMs). The multi-component desulfurization sorbents are expected to demonstrate either 
additive or synergetic effects, similar to those reported for the heterogeneous catalysis, as found, 
for instance, by a high throughput synthesis and screening routine [130]. Both surface and bulk 
dispersion and oxidation states of the promoter ions are important factors controlling both 
reactivity of the sorbent and its temporal stability upon the multiple adsorption-regeneration 
cycles. Therefore, mechanistic studies of the effects of the promoter ions are needed.  
To study the desulfurization promoter, i.e. the minority chemical component of the multi-
component sorbent (or catalyst), suitable experimental technique(s) needs to offer: i) a rather 
high sensitivity, ii) the ability to analyze both surface and the ?bulk? of the specimen, iii) the 
ability to study the local structure of the promoter site. There are few experimental techniques 
available to study the dispersion, oxidation and coordination state of the TM promoters, namely 
Electron Spin Resonance (ESR) [84, 85] and X-Ray Photoelectron Spectroscopy (XPS) [85, 86]. 
ESR has an excellent sensitivity (>1011 spins/sample), and it provides information on the 
oxidation and coordination state of the typical TM dopant ions [12-14]. ESR is the typical ?bulk-
 
127 
 
sampling? technique, due to the large penetration depth of the gigahertz radio-frequency used; 
however, it can be effectively used as well to study the surface-localized radicals and the radical 
ions in the solid materials [131, 132]. On the other hand, XPS can conveniently complement 
ESR as pertinent to the studies of heterogeneous chemical systems, such as sorbents and 
catalysts [133], since it is the surface-sensitive technique that analyzes the topmost ca. 10 nm of 
the material only. The main limitation of XPS is its relatively low sensitivity (> 5 % of the 
monolayer) [103]. 
We report here the preparation of FexMnyZn1-x-yO/SiO2 and measurements of H2S uptake at 
room temperature and desulfurization performance upon the multiple regeneration cycles of 
tthese sorbents. The novel desulfurization sorbents FexMnyZn1-x-yO/SiO2 can i) achieve >90 % of 
theoretical sulfur uptake capacity at room temperature, ii) reduce sulfur concentration in the 
gaseous stream from ~1000 ppm to < 1 ppm, iii) and be easily regenerated > 10 times by simple 
heating in air without a significant loss of performance. We report the characterization of the Zn, 
Mn, Fe, S sites in those sorbents by ESR and XPS. 
 
128 
 
V.2.  Experimental 
 
The promoted ZnO-based desulfurization sorbents of the nominal formula FexMnyZnO1-x-
y/SiO2 (x, y=0; 0.025) were prepared by incipient co-impregnation of high surface area (300-550 
m2/g) silica (Fischer Scientific Inc.) of grain size 100-200 ?m with solutions of nitrates of the 
respective metals in water, namely Zn(NO3)2, Mn(NO3)2 and Fe(NO3)3. Single step incipient 
impregnation was performed on the silica support to achieve metal oxide loading of 12-36% by 
varying the molarity of nitrate solutions. Upon incipient impregnation and drying, the samples 
were calcined in the flowing air at 350-550 oC; these are referred to as the ?calcined? specimens. 
The specimens prepared as above, excepting the calcination step, are referred to as the ?dried? 
sorbents. In the reference experiments, with the commercial H2S sorbents (BASF SG-901 and 
Sud Chemie G-72E), they are crushed to the same particle size as that of the silica (100-200 
microns) used to prepare the supported FexMnyZnO1-x-y/SiO2 sorbents. 
Breakthrough curves for both commercial sorbents and FexMnyZnO1-x-y/SiO2 sorbents were 
measured at 20 ?C. In the desulfurization experiments, the challenge gas was the model 
reformate with an inlet concentration of 1 vol. % H2S in H2. Gases were purchased from Airgas 
Inc. The face velocity (GHSV) of the stream is 1900 h-1, corresponding to volumetric gas flow 
rate of 0.1 slpm. The desulfurization reactor contained 0.500 g sorbent; the sorbent bed size was 
9 mm in diameter and 10 mm thick. H2S uptakes during adsorption experiments were measured 
using a gas chromatography (GC) instrument (Varian CP3800) equipped with the thermal 
conductivity detector (TCD) and pulse flame photometric detector (PFPD). The specimens of the 
sorbents upon adsorption of H2S are referred to as the ?sulfided? samples. 
 
129 
 
Regeneration of the ?sulfided,? i.e. ?spent? sorbent was performed in-situ in the sulfidation 
reactor at 550 oC in air at a flow rate of 950 h-1. The sorbent FexMnyZnO1-x-y/SiO2 of 15 wt. % 
loading of ZnO was regenerated for over 10 cycles, with the regeneration temperature being the 
same as that of the sample calcination before the 1-st desulfurization cycle. The temperature of 
the furnace during the experiments was maintained using a PID temperature setpoint controller.  
The samples were characterized using the N2 adsorption desorption isotherms to study the 
changes in surface area, pore volume and pore size before and after metal oxide loading. Also, 
XPS, ESR and XRD are used to characterize the sorbent. The techniques and the conditions at 
which the equipments were operated is described in Chapter II.  
 
V.3. Results and Discussion 
V.3.1  Performance of the FexMnyZn1-x-yO/SiO2 Sorbents 
 
Figure V.1 shows the H2S sorption performance of the commercial ZnO sorbents from Sud 
Chemie and BASF, of the supported sorbent ZnO/SiO2 prepared in our lab (21 wt. % loading of 
ZnO) and of the promoted Fe0.025Mn0.025ZnO0.975/SiO2 sorbent (21 wt. % loading of ZnO). The 
Fe0.025Mn0.025ZnO0.975/SiO2 sorbent shows a superior H2S uptake compared to the others.  
 
130 
 
 
Figure V.1. H2S Breakthrough curves of the commercial ZnO Sorbent from BASF (filled 
circles), Sud-Chemie (Squares), ZnO/SiO2 (open Circles) and Fe0.025/Mn0.025ZnO0.095/SiO2 
sorbent (diamonds) 
 
 
Table V.1 shows the sulfur uptake capacity (g sulfur / g sorbent) and utilization of ZnO in the 
sulfidization reaction (% of the theoretical value for the ZnS stoichiometry) attained at the 
breakthrough and the saturation regimes. The breakthrough is defined as 2% of inlet 
concentration. The supported ZnO/SiO2 sorbent has shown better performance over both 
commercial ZnO-based sorbents. XRD of the ZnO/SiO2 sorbent at 15% wt. loading of ZnO did 
not detect lines of either zinc silicate Zn2SiO4 [93] nor zinc oxide ZnO. These findings and the 
 
131 
 
high sulfur capacity of the ZnO/SiO2 sorbent indicate that the nano-dispersed form of ZnO is 
present in the supported sorbent, with the typical ZnO crystalline size ~40 ? or less. 
Table V.1: Comparative breakthrough, saturation capacities and ZnO utilization data  
 
The adsorption capacity among the promoted sorbents of the formula FexMnyZnO1-x-y/SiO2 
follows the trends: Fe0.025Mn0.025 ~ Mn0.025 > Fe0.025 and Fe0.025Mn0.025 > Mn0.05 > Fe0.05. XRD of 
the Fe0.025Mn0.025ZnO0.975/SiO2 sorbent in both ?calcined? and ?sulfided? forms was performed, 
and no lines due to any Fe or Mn compound were found that indicates a high degree of 
dispersion of the Fe and Mn promoters. Moreover, the observed promoter effects of Mn and Fe 
cations on the ZnO/SiO2 sorbent are of the synergetic, rather than additive nature. Indeed, the 
increase of H2S uptake of the promoted sorbent due to the additive effect would be insignificant 
Sorbent 
Loadin
g 
Sat 
Cap 
ZnO 
Utilzation 
Sat. Cap 
Breakthroug
h 
Cap 
ZnO 
Utilization  
at 
Breakthrough 
BASF ZnO (SG-901) 90 0.019 5 0.011 3 
Sud-Chemie (G-72E) 90 0.032 9 0.024 7 
ZnO/SiO2 15 0.032 54 0.026 45 
Fe0.025ZnO0.975/SiO2 15 0.043 72 0.035 58 
Mn0.025ZnO0.975/SiO2 15 0.043 72 0.037 62 
Fe0.025Mn0.025ZnO0.95/Si
O2 15 0.045 76 0.037 62 
ZnO/SiO2 21 0.053 64 0.051 61 
Fe0.025Mn0.025ZnO0.95/Si
O2 21 0.075 90 0.069 83 
 
132 
 
within the error bars of determining the outlet concentration of H2S, given the low concentration 
of both Mn and Fe cations vs. concentration of ZnO in the promoted Fe0.025Mn0.025ZnO0.975/SiO2 
sorbent. The synergetic mechanism of Mn and Fe H2S sorption promoters implies that Fe and 
Mn cations are dispersed on top or within the ZnO supported nano-phase, rather than forming 
their own phases on the SiO2 support. 
 
V.3.2 Structural Characterization of the Sorbents 
 
Figure V.2 shows the saturation capacity of the ZnO/SiO2 sorbents and the total surface area vs. 
the wt. % loading of ZnO. It can be seen that the total surface area decreases linearly with the 
ZnO loading. On the other hand, saturation sulfur capacity is not linear vs. loading of ZnO within 
the whole range: it increases rather sharply at 0-25% loading and plateaus at the higher loadings. 
Similar phenomena were reported in the literature; for instance, H2S uptake by Fe-Zn mixed 
metal oxides at room temperature is not proportional to their (active) surface area [134]. The 
non-linearity is due to the different chemical reactivity of the active sites of the sorbents of the 
different surface area.  
 
133 
 
 
Figure V.2: H2S Capacity (mg Sulfur/ g Sorbent) and the total surface area vs. the loading of 
ZnO (wt. %) in the ZnO/SiO2 Sorbents.  
Figure V.3 shows XRD of the ZnO/SiO2 sorbents at high loadings of ZnO (21 and 36 %). The 
pattern observed for the 36 % wt. loading of ZnO belongs to the XRD spectrum of ZnO of the 
wurtzite (hexagonal) structure [135]. Therefore, at such high loading of the supported ZnO, the 
latter is present as large crystallites whose desulfurization behavior corresponds to the plateau of 
the sulfur uptake (Figure 2). At the lower loading of ZnO of 21 %, there are no XRD lines except 
those of the silica support (Figure V.3). Therefore, an increase of ZnO loading beyond ca. 21 % 
when ZnO is of the large crystal size does not lead to an increase of the sulfur capacity. 
Therefore, the conclusion is that H2S sorption occurs in the surface layer of the nano-dispersed 
ZnO, rather than proceeds within its ?bulk.? This conclusion is consistent with earlier reports that 
concluded that the surface reactivity of the ZnO-based H2S sorbents as room temperature, when 
 
134 
 
only the outermost 0.6 nm of ZnO reacts with H2S [134]. Indeed, for the ZnO crystallites at the 
limit of XRD detection (4 nm), as little as > 40% of the atoms are located on surface [136], while 
for the smaller ZnO crystallites (2 nm), as many as > 80% of all the atoms are on the surface.  
      
Figure V.3.: XRD Spectra of the ZnO/SiO2 sorbents at 36 wt% of ZnO (solid line) and 21 wt. % 
(dashed dotted line) vs. SiO2 support (dotted line)  
 
 
 
 
Table V.2: Structural characterization of various H2S sorbents 
 
Sorbent Loading Surface Area Pore Volume 
 
135 
 
(m2/g) (cc/g) 
SiO2 0 550 0.79 
ZnO/SiO2 15 460 0.55 
ZnO/SiO2 25 330 0.48 
Fe0.025ZnO0.975/SiO2 15 325 0.52 
Mn0.025ZnO0.975/SiO2 15 314 0.53 
Fe0.025Mn0.025ZnO0.95/SiO2 15 375 0.59 
SiO2 0 330 1.65 
ZnO/SiO2 21 244 1.04 
Fe0.025Mn0.025ZnO0.950/SiO2 21 160 1.02 
 
 
The structural characteristics of the H2S sorbents as determined by N2 adsorption are shown in 
Table V.2. Upon promoting the ZnO/SiO2 sorbent (15 wt. % of ZnO) with Fe and Mn to obtain 
the FexMnyZnO1-x-y/SiO2 sorbent, there is only a marginal change in surface area and pore 
volume. We have used the other silica support (with the pore volume of 1.65 cc/g) to prepare the 
sorbent with ZnO loading of 21 wt. %, and have obtained similar results. The latter finding 
indicates that structural characteristics of the ZnO/SiO2 sorbent at those loadings do not 
significantly change when Mn and/or Fe promoter cations are added. Therefore, the promoted 
sorbents with ZnO loading of either 15 % or 21 % are the most effective, and the multiple-cycle 
adsorption/regeneration was conducted on the sorbent with ZnO loading of 15 %.  
V.3.3 Performance of the Sorbents upon Multiple Regeneration Cycles  
 
 
136 
 
Figure V.4 shows the breakthrough curves for the FexMnyZnO1-x-y/SiO2 sorbents upon H2S 
adsorption / regeneration cycles, as compared to the ?fresh? sorbent (loading of ZnO is 15 wt %). 
The ?fresh? sample is prepared at the calcination temperature of 550 ?C and the regeneration was 
performed at the same temperature. The sorption capacity fluctuates within the first 10 
sulfidation/regeneration cycles; however, at the 10-th cycle, the capacity is as high as >80 of the 
?fresh? sorbent.  
 
Figure V.4:  H2S breakthrough curves upon the multiple adsorption/regeneration of Fex-Mny-
ZnO 1-x-y/SiO2 sorbent 
 
The loss of activity of the H2S sorbents upon multiple sulfidation/regeneration cycles is well 
known; it is also known that the sulphur capacity does not always steadily decay vs. the number 
of cycles, but may undergo some fluctuations. For instance, in the multi-cycle study of H2S 
 
137 
 
breakthrough curves obtained with the CuO/SiO2 sorbent, the breakthrough capacity upon the 2-
nd cycle is only ~50% as compared to the 1-st cycle. However, the capacity increases with the 
further cycles, with the occasional reduction of the breakthrough capacity on the 12-th cycle 
[137]. This type of behavior is likely to be due to the ?hysteresis? effects, as the result of the 
dynamic changes in the materials system used under the non-equilibrium conditions of the 
repeated cycling of both temperature and the redox regime used (reductive conditions of the 
sulfidation vs. the oxidative conditions of the regeneration). The study investigating the sulfur   
breakthrough capacity within the multicycle (up to 100) tests were carried out for zinc titanate 
sorbents sulfided at the ?middle-temperature? range of 480 ?C [138]. The breakthrough capacity 
was shown to change following the complex pattern: increase of capacity up to the 3-rd cycle, 
then the fluctuating decaying trend of capacity.   
 
V.3.4 Characterization of the Sorbents by XPS 
 
Figure V.5 shows the XPS spectra of the calcined sorbent Fe0.2Zn0.8O/SiO2. The strong 
electrostatic charging of up to 6 eV was observed, as expected for the electrically insulating 
specimen containing silica. Oxygen, silicon, iron, zinc and spurious carbon were detected in the 
survey spectra (not shown) as expected. The absence of the N 1s line in the XPS spectra 
indicates the complete decomposition of the nitrate precursors as expected. The sorbents with the 
lower concentration of Fe or Mn, such as Fe0.05Zn0.95O/SiO2 and Mn0.05Zn0.95O/SiO2 did not 
show the reliable XPS signals to determine Fe or Mn, consistently with the known limited 
sensitivity of XPS towards the species present at a small fraction of the monolayer. We note that 
the Fe0.2Zn0.8O/SiO2 sorbent (Figure V.5) shows the satisfactory H2S capacity, and the shape of 
its breakthrough curve reminds the one of the Fe0.05Zn0.95O/SiO2 sorbent (data not shown). 
 
138 
 
Therefore, the structural and chemical information obtained from the analysis of the XPS data of 
the transition metal (TM) enriched XPS samples such as Fe0.2Zn0.8O/SiO2 is relevant to the 
FexMnyZn1-x-yO/SiO2 (x,y=0, 0.05) sorbents that show the best desulfurization performance. 
 
Figure V.5: The XPS lines of Fe 2p3/2 (Figure 5A), Zn Auger L3M45M45 (Figure 5B), O 1s 
(Figure 5C) and Zn 2p (Figure 5D) of the ?calcined? sorbent Fe0.2Zn0.8O/SiO2. 
The Binding Energy (BE) of the Zn 2p3/2 line (Figure V.5D) is 1022.1 eV that is consistent with 
the reported BE of 1022.0-1022.1 eV for the Zn2+ form in the zinc oxide catalyst [80]. The BE of 
the Zn L3M45M45 peak (Figure V.5B) is 498.9 eV. The XPS Auger Parameter (AP) is more 
 
139 
 
useful than the BE for the determination of the oxidation state and coordination environment of 
atoms in the electrically insulating samples such as supported sorbents and catalysts, since its 
value is independent on the electrostatic charging [96]. The APZn is calculated to find the 
oxidation and coordination state of Zn in the ?calcined? sorbent, by using the formula APZn = 
K.E.(ZnAuger) + B.E.(Zn) = 2010 eV. As expected, this corresponds to ZnO whose APZn is 2009.8 
eV [97]. It must be further mentioned that atomic ratio of Zn/Si = 0.26 as measured by XPS is 
higher than the nominal value for the 21 % wt. loading of ZnO on silica. Hence, it should be 
concluded that in the ?calcined? Fe0.2Zn0.8O/SiO2 sorbent, Zn is present in the form of the ZnO 
nanoparticles located on the SiO2 surface, rather than those included into the lattice of SiO2. 
Figure V.5C shows the O 1s peak that can be well fitted as the singlet, and its BE of 531.6 eV is 
close to the value [98] of 531.5 eV that was reported for O 1s line in silicon oxide SiO2. For the 
pure silica that has been calcined in the flowing oxygen at 673 K, the O 1s peak is the singlet, 
while for the SiO2 thin films, both bridging oxygen atoms (Si-O-Si, BOs) at 531.5 eV and the 
non-bridging atoms (Si-O-, NBOs) at the lower BE are found as shoulders of the O 1s peak [98]. 
On the other hand, the BE of O 1s in zinc oxide ZnO is as low as 529.7 eV [100]. From these 
data, we conclude that the O 1s peak in Figure V.5C belongs mostly to the bridging oxygen of 
the silica support, consistently with the low coverage of the ZnO as determined above by us. 
Figure V.5A shows the Fe 2p3/2 XPS line that can be fitted with two spectral components. In the 
XPS studies of the Fe-containing sorbents and catalysts prepared or exploited in air, it was 
reported that Fe2+ and Fe3+ are the most typical oxidation states of iron [139]. Thus we conclude 
that Fe in the Fe0.2Zn0.8O/SiO2 sorbent exists in both +2 and +3 oxidation states, with the BEs of  
the 2p3/2 line for Fe2+ and Fe3+at 709.5 and 711.1 eV, respectively. 
 
140 
 
Atomic ratios in the calcined Fe0.2Zn0.8O/SiO2 sorbent have been calculated from the XPS data. 
The ratio O/Si is found to be 2, as expected. The Fe/Zn ratio as determined by XPS is 0.6; that is 
significantly higher than the value of 0.25 as expected for the sorbent of the nominal formula 
Fe0.2Zn0.8O/SiO2. This discrepancy should be attributed to the structure of the sorbent with its 
surface enriched with iron. Phase separation in the unsupported binary oxides of zinc and iron 
was reported earlier. For instance, upon the calcination of the binary oxide ZnO-Fe2O3 at 350 ?C 
that was prepared by the co-precipitation from the solutions of metal salts, Fe2O3 was found by 
XRD as the separate phase [134]. These findings are consistent with our model of the partial 
exclusion of Fe3+ cations from the lattice of the nanocrystalline supported ZnO upon calcination. 
On the other hand, in the XPS spectrum of the Mn0.2Zn0.8O/SiO2 sorbent (data not shown), there 
are no XPS lines of manganese, apparently due to the too small concentration of manganese 
within the probing depth of XPS. Given the similar values of both photoemission cross-sections 
and the electron mean free paths (MFP) for the Mn2p and Fe 2p lines [140], we conclude that in 
the MxNyZn1-xO/SiO2 sorbents (M,N = Fe, Mn), the surface of the supported ZnO nano-
crystallites is enriched with Fe, while Mn is dispersed relatively uniformly within the ZnO 
nanocrystallites.  
Figure V.6 shows the XPS lines of the sulfided sorbent Fe0.2Zn0.8O/SiO2. C, O, S, Si, Zn and Fe 
are found in the survey XPS spectrum. The Zn 2p3/2 line is found to be at 1021.6 eV, which can 
be attributed to ZnS. It is further confirmed by calculating the Auger parameter; namely, the A.P. 
of Zn in the sufided sorbent is calculated to be at 2011.5, which closely matches that of the ZnS 
form [104], as expected. The assignment of the XPS lines of sulphur is not possible, due to the 
their weak intensity, consistent with the smaller photoemission cross-section of the S 2p and S 2s 
lines, as compared with those for metals, such as Zn, Fe or Mn [140]. The XPS atomic ratio O/Si 
 
141 
 
is 2, as expected for the SiO2 being the majority chemical compound of the surface of the 
sorbent. Therefore, there is no significant changes of the morphology of the supported 
nanocrystalline ZnO upon sulfidization, as opposite to the re-crystallization and coarsening of 
the ZnO nanoparticles in the unsupported mixed metal oxides after the reaction with H2S at room 
temperature [128]. 
 
 
 
 
 
142 
 
Figure V.6: The XPS lines of Zn Auger L3M45M45 (Figure 6A), Zn 2p (Figure 6B) and O 1s 
(Figure 6C) of the sulfided sorbent Fe0.2Zn0.8O/SiO2.  
 
 
V.3.5 Characterization of the FexMnyZn1-x-yO/SiO2 Sorbents by ESR 
 
Figure V.7A shows the ESR spectra of the ?calcined? sorbents Fe0.025Zn0.975O/SiO2 vs. 
Fe0.025Mn0.025Zn0.950O/SiO2, and Figure V.7B shows the spectra of the ?sulfided? sorbents 
Mn0.025Zn0.975O/SiO2 vs. Fe0.025Mn0.025Zn0.950O/SiO2. In the calcined sorbents, no ESR signal of 
Mn cations are seen that indicates the presence of Mn as the ?ESR-silent? Mn3+ state only; even 
though Mn3+ with 3d4 electronic configuration is paramagnetic, but it is not ESR-detectable at 
room temperature due to the rapid spin-lattice relaxation [141]. The only ESR signal in the 
spectra of the calcined sorbent is the signal at g~4.28 (Figure V.7A) due to Fe3+ ions [142, 143]. 
According to the literature, the ESR signal at g~4.28 is due to the isolated Fe3+ cations in the 
tetrahedral coordination with rhombic distortion [143]. The ESR spectra of the ?dried? sorbents 
of neither chemical composition could be recorded, likely due to the strong adsorption of the 
ESR radio-frequency by the water that is chemisorbed on the surface of the sorbents.  
 
143 
 
 
 
Figure V.7: ESR spectrum of the ?calcined? sorbent Fe0.025Zn0.975O/SiO2 (dotted solid line) vs. 
Fe0.025Mn0.025Zn0.975O/SiO2 (solid line), Figure 7A. ESR spectrum of the ?sulfided? sorbent 
Mn0.025Zn0.975O/SiO2 (dotted solid line) vs. Mn0.025Fe0.025Zn0.975O/SiO2 (thick solid line), Figure 
7B.  
In the spectra of the ?sulfided? sorbents, the signal of Mn2+ appears (Figure V.7B), as the sharp 
sextuplet present on top of the broad spectral envelope at ~2500-4500 Gauss. The sixtuplet is 
 
144 
 
observed due to the hyperfine splitting (h.f.s.) of the isolated 55Mn2+ ions in the tetrahedral or 
octahedral coordination geometry [144, 145]. For instance, Mn cations in the Mn-MCM-41 vs. 
Zn/Mn-MCM-41 zeolites were studied by ESR, and values of g-factor of about 2.002 and A 
values of ca. 90-95 Gauss were reported that correspond to the isolated Mn2+ centers [144]. 
Therefore, the reduction of Mn3+ cations to Mn2+ occurs upon reaction of the Mn-containing 
sorbent with H2S in hydrogen. The broad spectral ?envelope? is due to the interacting Mn2+ 
cations [145]. In the reference experiments with pure H2, no spectral lines of Mn2+ appear in the 
ESR spectra; therefore, the reduction of Mn3+ to Mn2+ proceeds due to reduction by H2S, not by 
the H2 component of the challenge gas. The ESR signal of the isolated Fe3+ cations (Figure 5A) 
is much weaker (factor of 100) than the ESR signal of Mn2+ in the sulfided sorbent (Figure 5B), 
even though the stoichiometric amounts of the Fe and Mn are the same. Therefore, the Fe3+ 
isolated ions represent the minority form of iron. As a result, the quantitative or even semi-
quantitative determination of various forms of Fe3+ by ESR cannot be performed.  
The silica support that was prepared similarly to the ?calcined? sorbent, except that Fe, Zn and 
Mn salts were not used, shows no ESR spectrum, as expected. The ?calcined? sorbent of the 
formula Fe0.000Mn0.000Zn1.000O/SiO2 shows no ESR spectrum as well, thus confirming that the 
spectral multiplets in Figure V.5A and V.5B belong to Fe3+ and Mn2+, respectively. In addition, 
no spectral lines of any Reactive Oxygen Species (ROS) or oxygen vacancies [87] are present in 
the ESR spectra of neither calcined nor sulfided sorbents. 
It is known that when two paramagnetic ions are within the close distance, ~10 ?, the interaction 
of their spins can be observed in the ESR spectra. For instance, it was reported for the binuclear 
enzymes that the amplitude of the ESR signal of the Mn2+ cation was reduced when the extra 
Mn2+ cation was added within 8-11 ? distance [146]. We have compared the ESR spectra of 
 
145 
 
Mn2+ in the sorbent with and without Fe promoter present (Figure 5B). The differences are 
minor, and upon the processing of the spectra by the standard double-integration (DIN), the 
differences are within the accuracy of the quantitative ESR measurement (~5 %). Therefore, Mn 
cations do not interact with Fe cations in the Fe0.025Mn0.025Zn0.950O/SiO2 sorbent that indicates 
that those cations are, on average, >10 ? away from each other. The data of the structural 
characterization of the sorbents are consistent with their relative H2S uptake capacity: 
Fe0.025Mn0.025 ~ Mn0.025 > Fe0.025 and Fe0.025Mn0.025 > Mn0.05 > Mn0.05. 
The ESR spectra of the sorbent with the high content of Fe and Mn, namely, FeO/SiO2 and 
MnO/SiO2 are quite broad, consistently with the signal broadening due to the strongly interacting 
paramagnetic TM cations. ESR spectrum of Fe/SiO2 has the broad peak with g value of about 2. 
This corresponds to the clustered form of the Fe3+ ions, and the spectrum is similar to the 
spectrum of the Fe2O3 phase in the Fe-containing zeolites [143] and to the FexOy clusters [142].  
The spectrum of the MnO/SiO2 is also very broad, consistently with the spectrum of the 
interacting Mn2+ ions [144]. Loss of the h.f.s. due to broadening is typical for the samples 
containing more than 4.5 wt % of Mn [145], consistent with our data . n-situ ESR tracking 
change in Mn relates to the breakthrough curve of the Mn0.025Fe0.025Zn0.95O/SiO2 tested at same 
conditions. This indicates that the MnO is probably in the solid solution with ZnO and the 
presence of Fe lines as shown in XPS, Fe ions and clusters are predominantly distributed on the 
surface. The Figure V.8 shows the schematic representation of the possible locations of the 
active sites Mn and Fe in the promoted ZnO/SiO2 depending on all the characterization 
techniques. More work needs to be done to prove the mechanism shown in Fig V.8.  
 
146 
 
 
Figure V.8 Schematic diagram of the mechanism of distribution of the Mn, Fe active sites in 
ZnO/SiO2  
The important question is what are the mechanisms of the promoting effect of Fe and Mn cations 
on the desulfurization capacity of the Fex-Mny-Zn1-x-yO/SiO2 sorbents. In the literature, three 
mechanisms are discussed [127]: i) the enhancement of the active surface area of ZnO due to the 
presence of the TM oxide; ii) metal cation diffusion; iii) diffusion of HS? and S2? ions towards 
the bulk of the solid particles of ZnO. In the above referenced study of the unsupported Fe?Mn?
Zn?Ti?O mixed metal oxides, the enhancement of H2S uptake as explained due to the increase of 
ZnO active area is experimentally proven by the XRD and BET measurements. On the other 
hand, in the supported Fex-Mny-Zn1-x-yO/SiO2 sorbents studied by us, the changes of the surface 
area of the sorbent upon addition of the minor amounts of Fe or Mn promoters are too small to 
be measured by BET, and no XRD lines of any metal compound can be recorded for those nano-
dispersed materials.  
R oo m  T e m p .  H
2
S / H
2
T h e r m a l  
R e g e n e r a t i on  , O
2
XR D :  N o  XR D  p a tte r n  ? F e ,  Mn ,  Z n  c r y s tal l i te s  
( < 4 n m ) .  F e  an d  Mn a r e  d i s p e r s e d  i n  Z n O i n  
s o l u ti o n ,  h i g h l y  s tr ai n e d ,  o r  s u r f ac e  c ati o n s
E S R :  Mn tr a c k s  Z n O b e h a v i o r  ?
S i g n i f i c a n t a m o u n t o f  Mn i s  as s o c i a te d  
w i th  Mn - Z n O s o l i d  s o l u ti o n
X P S :  F e  o b s e r v ab l e  an d  M n  i s  n o t i n  
s am p l e s  o f  s am e  p r o m o ti o n al  
c o m p o s i ti o n ? M o s t o f  th e  F e  i s  o n  th e  
Z n O  o r  S i O
2
s u r f ac e  an d  M n  i s  n o t
E S R :  Mn
2 + 
tr a c k s  Z n S ? m a j o r  
c o m p o n e n t n o t o n  S i O
2 
s u r f ac e
Fe
3+
? Fe
2 + 
d u e  t o  s u l f i d e d
F e   i o n s / c l u s te r s .  F e  
p r o m o ti o n a l  e f f e c t  i s   
o b s e r v e d  s o  s o m e  m u s t b e  
c l o s e l y  a s s o c i a te d  w i th  Z n O
S i O
2
S u l f i d e d
XR D :  N o  XR D  p a tte r n  ? F e , Mn , Z n
C r y s tal l i te  s i z e  ( < 4 n m ) ,  e tc . .
S i O
2
 
147 
 
One has to note that both the Mn and Fe are the minor impurities in the 
Fe0.025Mn0.025Zn0.95O/SiO2 sorbents and they are the minor components of the supported 
nanophase of mixed metal oxides Fe- Mn-ZnO. Therefore, it is very challenging analytical and 
solid state chemistry task to study these sulfidation promoter sites. The promoter mechanisms of 
Mn and Fe cations could be proposed if the localization of those cations is determined or vice 
versa. From the XPS data, we conclude that the surface of Fe and Mn promoted ZnO/SiO2 
sorbents has an increased concentration of Fe, but decreased concentration of Mn. Therefore, Fe 
ions are likely to be located on the surface of ZnO, while Mn ions are likely to be located within 
ZnO crystallites. The mechanism of the promoter effect of Mn cations can be tentatively 
elaborated as follows. 1) If Mn cations were on the interface between ZnO and silica support, no 
or little promoter effects could be expected, due to the hindered diffusion of H2S towards the 
ZnO-SiO2 interface. Similarly, if Mn oxide formed its own nano-dispersed phase supported on 
silica, its effect on the H2S uptake of ZnO would have been additive, i.e. negligibly small. The 
schematic representation for the proposed structure of Fe0.025Mn0.025Zn0.95O/SiO2 is as shown in 
Fig.V.9. Therefore, we propose that Mn cations are located within the nano-crystallites of the 
supported ZnO. The proposed localization of the Fe and Mn cations is consistent with our ESR 
data shat show no spin-spin interactions between Fe and Mn cations, i.e. their localization at least 
10 ? from each other. This distance is of the same order of magnitude as the size of the ZnO 
nanocrystallites that are smaller than the XRD limit, i.e. < 40 ?. From the XPS data, we 
conclude that the surface of the Fex-Mny-Zn1-x-yO/SiO2 sorbent is enriched with Fe ions. We thus 
conclude that the surface of the Fex-Mny-Zn1-x-yO/SiO2 sorbent is enriched with Fe3+ ions, while 
Mn3+ ions are located within the ZnO supported nanocrystallites. Therefore, the promoter effect 
of Fe cations is likely to be the ?local? enhancement of the reactivity of ZnO towards H2S, while 
 
148 
 
the promoter effect of the Mn cations could be to decrease the size of the ZnO nanocrystallites 
[127].  
 
Figure V.9: Schematic representation of the structure of Fe0.025Mn0.025Zn0.95O/SiO2 sorbents and 
sulfidation/regeneration reactions. 
 
The mechanisms of the promoter effects in the Fex-Mny-Zn1-x-yO/SiO2 sorbents cannot be 
directly determined from the structural or spectroscopic characterization, and they need to be 
understood from the complementary advanced spectroscopic studies, including the real-time 
Operando spectroscopy that is currently underway in our laboratory. 
 
 
 
 
 
Si O
2
O
2 , 
H ea t
Si O
2
Mn
3+
Mn 2 +
Fe 3+
Fe 3+
H
2
S + H
2
H
2
 
149 
 
V.3  Conclusions 
 
The Mn and Fe promoter cations significantly enhance the utilization of the ZnO active phase in 
reaction with H2S of the novel sorbents Fex-Mny-Zn1-x-yO/SiO2 (x, y=0, 0.025), during 
desulfurization of the mixture of H2S and H2 at room temperature. The Mn- and Fe-promoted 
sorbents maintain a high sulfur uptake capacity upon the multiple cycles of a simple thermal 
oxidative regeneration of the ?spent? sorbent in air (up to 10 cycles). ZnO and cations of Fe and 
Mn are nano-dispersed in the Fex-Mny-Zn1-x-yO/SiO2 sorbents, both the ?calcined? and ?sulfided? 
forms. As judged by XPS and ESR, the surface of the sorbent is enriched with Fe promoter 
cations, while Mn3+ promoter cations are located within the supported ZnO nanocrystallites. 
Acknowledgement 
The authors would like to thank the US Army (TARDEC Contract W56HZV-05-C-0686) for the 
financial support of this work. A.S. thanks Prof. Michael Bowman (Department of Chemistry of 
the University of Alabama at Tuscaloosa) for useful discussions. 
 
 
 
150 
 
Chapter VI: RT Hydrolysis and Removal of COS from Fuel Reformate Streams using 
Al2O3/Carbon & Fe0.025Mn0.025ZnO0.95/SiO2 Layered Beds 
 
Priyanka Dhage, Hongyun Yang1 and Bruce J. Tatarchuk 
Department of Chemical Engineering, Auburn University, Auburn, AL 36849 (USA) 
1Intramicron Inc. 368 Industrial Pkwy, Auburn AL 36830(USA) 
 
Abstract 
Removal of both H2S and COS from reformate streams is critical for maintaining the activity of 
fuel processing catalysts. The objective of our work is developing sorbents for efficient, cost-
effective and scalable removal of H2S and COS over various temperatures, without significant 
activity loss upon multiple regeneration cycles. Bimetallic sorbents Mx/2Nx/2Zn(1-x)O supported 
on SiO2/Al2O3 (M, N = Mn, Fe, Ni, Mg, Cu and 0?x?1) prepared by impregnation/calcination 
were studied in packed bed, with model reformate gases (1 vol% H2S, 33% CO/CO2 inH-2, H-
2O), room temperature to 400 C. Their sulfur uptake capacity at room temperature significantly 
exceeds that of both commercial unsupported ZnO sorbents (by 60 %) and of the un-promoted 
supported sorbent ZnO/SiO2 (by 30 %), but showed no adsorption of COS.  Sulfur sorption 
capacity and the breakthrough characteristics remain satisfactory after up to 10 cycles of 
adsorption/regeneration, with regeneration performed by a simple heating in air. At 
temperatures< 250oC, COS formation is inhibited but significant amount of COS is formed in the  
 
 
 
151 
 
presence of CO2/CO and H2S. Al2O3/Carbon is a good catalyst for high temperature (T>100 C) 
COS hydrolysis. For room temperature COS hydrolysis, layered bed approach with COS 
hydrolysis on Al2O3/Carbon, followed by H2S removal on Fe0.025Mn0.025ZnO0.95/SiO2 was 
adopted.  
Keywords: COS hydrolysis, H2S, Al2O3, ZnO, Fe, Mn, XPS 
 
VI.1 Introduction 
 
With the introduction of the strong legislation to reduce sulfur emissions, fresh impetus is being 
given to modifying improving existing desulfurization technology. However, 
dehydrodesulfurization does not remove or significantly affect sulfur containing compound, 
namely, carbonyl sulfide. Removal of sulfur containing compounds is one of the most important 
technologies for utilization of gasified products derived from various feedstocks such as 
biomass, waste and solid fossil fuels[147]. Especially, gaseous sulfur compounds of H2S and 
COS are severe poisons against the following processing of steam reforming for hydrogen 
production or Fischer Tropsch synthesis [148]. Various researches for H2S removal have been 
reported in details for the purification of gasified products derived from various feed stocks; 
however, removal of COS is not a big concern yet, because it is not the major sulfur compounds 
produced from gasification [12]. The absorption of H2S by ZnO is stoichiometric above 350 oC 
but it falls rapidly at lower temperatures. The removal of COS has been reported to be more 
difficult at low temperatures in the range from room temperature to 200 oC than H2S. ZnO is a 
preferred metal oxide because of favorable sulfidation thermodynamics, [13] but is not efficient 
to remove COS [7]. Gaseous sulfur compounds of H2S and COS are severe catalyst poisons 
 
152 
 
against the following processes of steam reforming for hydrogen.COS can be formed by the 
conversion of H2S and CO2 in the absence of water. The conventional way to remove COS is 
hydrogenation and hydrolysis [149] COS is rather inactive compared to H2S probably due to its 
neutrality and similarity to CO2, COS is sometimes produced through the reaction of H2S with 
CO2, although the reaction can be reversible to produce again H2S and CO2 from the reaction of 
COS and H2O depending upon the adsorption conditions[14, 149, 150].  
The formation of COS is primarily governed by the reversible hydrolysis reaction and 
equilibrium conditions present:  
222 COSHOHC O S  
Parallel to the necessity for safe operations, the removal of trace sulfur components, such as 
mercaptans or carbonyl sulphide (COS), is a major challenge in designing the gas conditioning 
process. The pros and cons of several design options for deep COS removal are discussed in a 
case study, where the results of a hybrid solvent are compared with the performance of a BASF 
solvent.  These are compared to measurements from an operating plant. In a water-saturated 
reservoir, hydrogen sulphide (H2S) and carbon dioxide (CO2) are in thermodynamic equilibrium 
with COS. Thus a concentration of up to several hundred ppmv COS in the feed gas is not 
unusual. A relatively small volume of COS can combine with water to form H2S if suitable 
equilibrium conditions exist [14]. Molecular Sieves (e.g. Zeolite A) present a new problem for 
H2S removal because H2S and CO2 can react within the framework of the zeolite to produce COS 
and H2O. The problem is amplified further by the ability of molecular sieves, such as zeolites A 
and X, to absorb water and force the reaction far to the right; increasing COS concentration[151]. 
Most of the studies are concentrated on COS removal at operation temperature in excess of 100 
C and operational cost and energy consumption will be high[152] The studies on COS removal 
 
153 
 
and hydrolysis are divided in two parts: one part focuses on COS hydrolysis at low temperature 
and the other part on simultaneously removing both COS and H2S. A mathematical model was 
developed for COS removal using coupling reactions on a bi-functional catalyst. The 
temperature favors the effectiveness of the reaction rate constant, H2O adsorption equilibrium 
constant decrease in these conditions as expected [153]. The study of reaction mechanism for 
alumina as catalyst for COS hydrolysis in the temperature range 30-250 C was investigated 
[154]. Addition of Ni and Zn can efficiently promote COS hydrolysis on alumina at 30 C [155]. 
COS hydrolysis at low temperature (45-100 C) on alkali metal oxides and alkali earth metal 
oxides was studied[156].Single COS Removal process using an iron oxide catalyst around 50 C, 
where Fe2O3 was the catalysts for COS hydrolysis and the adsorbent for H2S removal [157].   
ZnO is reported to be best sorbent for sulfur adsorption [123] because of its favorable 
sulfidation thermodynamics and high sulfur capacity (by weight). However, a serious problem of 
the high temperature (> 500 ?C) H2S adsorbents is the reduction of ZnO by hydrogen into 
metallic zinc and evaporation of the latter [122]. Several oxides of other metals such as iron, 
vanadium, zinc, copper, manganese and molybdenum have been proposed as high-temperature 
desulfurization sorbents since the 1970s [124]. Chemical and structural transformations of those 
oxides upon desulfurization/regeneration were investigated; for instance, it is known that in the 
environment of the IGCC gasifier, Mn3O4 form is readily reduced to MnO and the latter reacts 
with H2S at the high temperatures [122]. Iron oxides have also been extensively investigated 
since the 1970s; iron oxide-based H2S sorbents have high sulfur capacity and reactivity towards 
H2S. However the equilibrium concentration of H2S is as high as 100 ppmw. In addition, the 
number of the degradation processes occur above ~ 500 ?C, namely the reduction of Fe3O4 to 
 
154 
 
FeO [122]. Mixed metal oxide sorbents for the high temperature desulfurization of coal gases 
were extensively reviewed in the past [125, 126]. 
Recently, increasing interest has been paid to the ?low temperature? H2S adsorbents that 
operate between room temperature and ~100 ?C [127-129]. For instance, we reported preparation 
and testing of novel ZnO/SiO2 sorbents for H2S and carbonyl sulfide COS with the minimized 
mass transfer resistance [77, 89-92] that operate at room temperature and retain their high 
desulfurization capacity after >10 desulfurization/regeneration cycles, with the regeneration 
performed by the inexpensive and robust calcination in the flowing air. 
To study the desulfurization promoter, i.e. the minority chemical component of the multi-
component sorbent (or catalyst), suitable experimental technique(s) needs to offer: i) a rather 
high sensitivity, ii) the ability to analyze both surface and the ?bulk? of the specimen, iii) the 
ability to study the local structure of the promoter site. The main limitation of XPS is its 
relatively low sensitivity (> 5 % of the monolayer) [103]. 
A previous study on the preparation of FexMnyZn1-x-yO/SiO2 and characterization of the 
active sites Zn, Mn, Fe, S sites in those sorbents by XRD and XPS has already been published 
[58] We report here the strategies to remove COS present/formed in the fuel reformate streams. 
And the preparation and performance of the novel Al2O3/Carbon for hydrolysis of COS and use 
of layered beds to remove both COS and H2S. The measurements of H2S uptake at room 
temperature, desulfurization performance upon the multiple regeneration cycles of the 
FexMnyZn1-x-yO/SiO2 sorbents are also discussed.  
 
 
 
155 
 
  
VI.2 Experimental 
 
Activated PICA Carbon of particle size 100-200 microns was dried in oven at 100 C. The dried 
Carbon was then impregnated via incipient wetness impregnation method with 2M Aluminium 
nitrate. The impregnated sample was then dried in air for 6hrs and then calcined at 350 C for 30 
mins. The calcined sample Al2O3/C is ready to test after cooling it down to room temperature.    
The ZnO/C and CuO/C were prepared by impregnating acetates as precursors and calcining at 
120 C in air for 1 hr.                                                                                                    
The promoted ZnO-based desulfurization sorbents of the nominal formula FexMnyZnO1-x-
y/SiO2 (x, y=0; 0.025) were prepared by incipient co-impregnation of high surface area (300-550 
m2/g) silica (Fischer Scientific Inc.) of grain size 100-200 ?m with solutions of nitrates of the 
respective metals in water, namely Zn(NO3)2, Mn(NO3)2 and Fe(NO3)3. Single step incipient 
impregnation was performed on the silica support to achieve metal oxide loading of 12-36% by 
varying the molarity of nitrate solutions. Upon incipient impregnation and drying, the samples 
were calcined in the flowing air at 350-550 oC; these are referred to as the ?calcined? specimens. 
The specimens prepared as above, excepting the calcination step, are referred to as the ?dried? 
sorbents. In the reference experiments, with the commercial H2S sorbents (BASF SG-901 and 
Sud Chemie G-72E), they are crushed to the same particle size as that of the silica (100-200 
microns) used to prepare the supported FexMnyZnO1-x-y/SiO2 sorbents. 
Breakthrough curves for both commercial sorbents and FexMnyZnO1-x-y/SiO2 sorbents were 
measured at 20 ?C. In the desulfurization experiments, the challenge gases were the model 
reformates with an inlet concentration of 1 vol. % H2S in H2 and 0.1% COS in N2. Model 
Reformate streams composition was chosen CO2 = 30 %, CO = 30%, H2O= 1% and H2S=1% 
 
156 
 
balance H2.  Gases were purchased from Airgas Inc and Matheson Tri-Gas. The face velocity 
(GHSV) of the stream is 2000-20000 h-1. Challenge gas was passed through the sorbent in a 
vertically-mounted packed bed tubular reactor (10 mm I.D. x 30 mm long) made of quartz that 
was coaxially located inside a 200 mm long tubular furnace. The desulfurization reactor 
contained 0.5-1.0 g sorbent; the sorbent bed size was 9 mm in diameter and 10 mm thick. H2S 
uptakes during adsorption experiments were measured using a gas chromatography (GC) 
instrument (Varian CP3800) equipped with the thermal conductivity detector (TCD) and pulse 
flame photometric detector (PFPD). The specimens of the sorbents upon adsorption of H2S are 
referred to as the ?sulfided? samples. 
Regeneration of the ?sulfided,? i.e. ?spent? sorbent was performed in-situ in the sulfidation 
reactor at 550 oC in air at a flow rate of 950 h-1. The sorbent FexMnyZnO1-x-y/SiO2 of 15 wt. % 
loading of ZnO was regenerated for over 10 cycles, with the regeneration temperature being the 
same as that of the sample calcination before the 1-st desulfurization cycle. The temperature of 
the furnace during the experiments was maintained using a PID temperature setpoint controller. 
The gas flow rates were controlled by mass flow controllers (Omega FMA 2405 Alaborg 
GFC1718). 
Nitrogen adsorption/desorption isotherms at 77 K were measured by an Autosorb 1-C 
instrument (Quantachrome Instrument Corp., USA). Before measuring the total surface area, 
samples were outgassed for 3 h at 200 ?C. The specific surface area, SBET was calculated via the 
Brunauer-Emmett-Teller (BET) equation, and the total pore volume (VP) was calculated at P/P0 
= 0.95. 
 
 
 
157 
 
VI.3 Results and Discussions 
VI.3.1   Desulfurization Performance of the Sorbents 
Figure VI.1 shows the H2S sorption performance of the commercial ZnO sorbents from Sud 
Chemie and BASF, of the supported sorbent ZnO/SiO2 prepared in our lab (21 wt. % loading of 
ZnO) and of the promoted Fe0.025Mn0.025ZnO0.95/SiO2 sorbent (21 wt. % loading of ZnO). The 
Fe0.025Mn0.025ZnO0.975/SiO2 sorbent shows a superior H2S uptake compared to the others.  
 
Figure VI.1. H2S Breakthrough curves of the commercial ZnO Sorbent from BASF, Sud-Chemie 
ZnO/SiO2 and Fe0.025/Mn0.025ZnO0.095/SiO2 sorbent. Test conditions: adsorption T= 20 C, Particle 
size = 100-200 microns, Co=1 vol5 H2S/H2 
Table VI.1 shows the sulfur uptake capacity (g sulfur / g sorbent) and utilization of ZnO in the 
sulfidization reaction (% of the theoretical value for the ZnS stoichiometry) attained at the 
breakthrough and the saturation regimes. The breakthrough is defined as 2% of inlet 
 
158 
 
concentration. The supported ZnO/SiO2 sorbent has shown better performance over both 
commercial ZnO-based sorbents. XRD of the Fe0.025Mn0.025ZnO0.975/SiO2 sorbent in both 
?calcined? and ?sulfided? forms was performed, and no lines due to any Fe or Mn compound 
were found that indicates a high degree of dispersion of the Fe and Mn promoters.  
Table VI.1: Breakthrough and Saturation H2S Capacity and utilization of ZnO (%) for various 
sorbents 
 
Figure VI.2 shows the COS adsorption at 400 C on the SiO2 and Al2O3 based samples, tested 
with inlet challenge gas varying from 500ppmv COS/N2. The silica based samples include SiO2, 
ZnO/SiO2 and Fe0.025Mn0.025ZnO0.95/SiO2 and alumina based samples include commercial BASF 
ZnO (SG-901) and Al2O3 and FeO/Al2O3 (15wt% loading). Alumina is a catalyst for COS 
hydrolysis at high temperature. The result shows that catalytic activity of alumina did not 
degrade for 6 hrs at the specified test conditions but silica based sample didn?t show any affinity 
or catalytic activity for COS hydrolysis.  
Sorbent Loading Sat 
Cap 
ZnO 
Utilization 
Sat. Cap 
Breakthrough 
Cap 
ZnO 
Utilization at 
Breakthrough(%) 
BASF ZnO (SG-901) 90 0.019 5 0.011 3 
Sud-Chemie (G-72E) 90 0.032 9 0.024 7 
ZnO/SiO2 15 0.032 54 0.026 45 
Fe0.025Mn0.025ZnO0.95/SiO2 15 0.045 76 0.037 62 
 
159 
 
 
Figure VI.2: COS hydrolysis at 400 C using Al2O3 based and SiO2 based sorbents. Inlet 
concentration: COS/N2 = 500 ppmv, 1% Steam, GHSV = 19000h-1 
 
VI.3.2 COS Removal & Hydrolysis  
Figure VI.3A shows the breakthrough performance of Fe0.025Mn0.025ZnO0.95/SiO2 tested at 400 C 
in the presence of CO2 and H2S. The result indicates that COS is formed in the presence of CO2 
and H2S. The stream contains 50% CO2 and 1%H2S and rest H2. At 400C, it is evident that COS 
is formed and because the sorbent has no reactivity for COS. The TCD detector was used to 
analyze the outlet gases. The chromatograph of the COS and H2S was recorded in the same run 
as TCD can detect both the gases simultaneously. Keeping the test conditions same, the  
Mn0.025Fe0.025Zn0.95O/SiO2 was tested without CO2, it showed almost 90% theoretical capacity 
without the formation of COS.  
 
160 
 
To understand how COS was formed, equilibrium COS concentrations for the reactions (1-3) 
running simultaneously were obtained using the HSC* software. The outlet COS concentrations 
are shown in Figure VI.3B and the outlet H2S concentration for reaction (4) is shown in Figure 
VI.3C.  
  
 
 
Figure VI. 3A: Breakthrough performance of  Fe0.025Mn0.025ZnO0.95/SiO2 with and without CO2 
at 400 C,  Test conditions :Q (2%H2S/H2) = 100 cc/min, Q(100% CO2) = 100 cc/min,  T = 400 C, 
GHSV = 8800 h-1 , Wt= 0.5 g 
The graph shows that when CO2 is present in the gas stream CO2 reacts with H2S to form COS.  
CO2+H2S ? COS+H2O 
 
161 
 
The reaction is homogenous and leads to significant formation of COS at high temperatures (T> 
250 C).  
 
Figure VI.3B: Equilibrium COS Concentrations. Reformate Composition (vol %): CO = 25 %, 
CO2 = 10%, N2 = 33 %, H2O = 7%, H2 = 25 % and H2S = 0.03% 
 
(1) 
 
162 
 
 
Figure VI.3C: Equilibrium H2S Concentrations. Reformate Composition (vol %): CO = 25 %, 
CO2 = 10%, N2 = 33 %, H2O = 7%, H2 = 25 % and H2S = 0.03% 
The equilibrium concentrations of COS produced by CO and CO2 on reacting with H2S are 
shown in Figure VI.3B.  This led to the hypothesis that COS formation in the reformate stream 
can be inhibited if the temperature is restricted to less than 250 C. Figure VI.3D shows the same 
test carried out at room temperature. This result shows that room temperature even in the 
presence of CO2 along with H2S, negligible amount of COS was formed. The Table VI.2 shows 
the comparison of the capacities at various temperatures, with and without CO2.  
 
163 
 
 
Table VI.2: Saturation Capacity of Fe0.025Mn0.025ZnO0.95/SiO2 with and without CO2 at room 
temperature and 400 C 
 
 
 
 
 
 
T (oC) Composition of Stream (vol %) 
Saturation Capacity 
(mol S/mol ZnO) 
20 CO2 =50, H2S =1, H2 = 49 89 
400 CO2 =50, H2S =1, H2 = 49 74 
400 CO2 =  0, H2S =1, H2 = 99 98 
1) CO2(g)+ H2S(g) ? COS(g) + H2O(g)  (Homogeneous) 
 
2) CO(g) +H2S(g) ? COS(g) +H2(g)        (Heterogeneous)  
 
3) CO(g) + H2O (g)? CO2(g) + H2(g)            (WGS) 
 
4) ZnO(g)+H2S(g) ? ZnS(s) + H2O(g)            
 
5)  
 
6) ZnO(s) +H2S(g) ? ZnS(s) + H2O(g) 
 
 
164 
 
 
 
 Figure VI. 3D. Breakthrough performance of Fe0.025Mn0.025ZnO0.95/SiO2 at 20 C Test conditions: 
Q (2%H2S/H2) = 100 cc/min, Q(100% CO2) = 100 cc/min,  T = 20 C, GHSV = 3800 h-1 , Wt= 
0.5 g  
  
Figure VI.4 shows the relative effect of adding CO/CO2 to the bed at 400 C with inlet 
concentration of 300 ppmv. Initially the bed is operated with only H2S/H2 in the stream and after 
40 minutes 10% CO was introduced into the bed, the PFPD detector shows an increase of about 
3.3 ppmv. This steadily decreases upto 2 ppmv. At 175 minutes 7% CO2 was introduced in the 
system and it shows about 0.2 ppmv increase in the outlet sulfur concentration. After 250 
minutes both CO and CO2 were removed from the stream and the concentration of sulfur goes 
 
165 
 
down to 0.2 ppmv.  The rise in concentration after adding CO and CO2 is due to formation of 
COS.  The figure shows that addition of CO/CO2 into the system leads to COS formation.  
 
 
Figure VI.4: Breakthrough curves of layered beds tested with 300 ppmv H2S-25% H2-25% CO-
10% CO2-7% H2O-33% N2 at a face velocity=100 cm/s at 400 C. Bed length: 22 cm 
Figure VI.5 shows breakthrough performance of Al2O3, FeO/Al2O3 and 
Fe0.025Mn0.025ZnO0.95/SiO2 tested at 400 C with CO2 and H2S in the inlet gas. At the test 
conditions, presence of CO2 and H2S at the given composition leads to formation of COS on 
Fe0.025Mn0.025ZnO0.95/SiO2 sorbent as shown by filled circles and H2S breakthrough is shown by 
triangles. The FeO/Al2O3 shows no adsorption for H2S as well as COS at these conditions.  
 
166 
 
 
Figure VI.5: COS Removal using layered bed. Test conditions: T = 400 C, GHSV = 15000 h-1, 
Wt. of each layer = 0.5g Metal oxide loading of each layer= 15%wt. Gas Composition (vol%) : 
CO2 = 28%, H2S = 0.5%, H2O = 1%, H2 = 70.5% 
H2S is more active than COS and therefore it reacts with the active sites of FeO/Al2O3 and hence 
the catalytic activity of Al2O3 were diminished, as opposed to the result in Figure VI.2 where 
COS is the challenge gas and Al2O3 works efficiently well as a catalyst. The layered bed in 
Figure VI.5 is the bed of Fe0.025Mn0.025ZnO0.95/SiO2 followed by a layer of ?guard bed? 
FeO/Al2O3. This design ensures adsorption of H2S by Fe0.025Mn0.025ZnO0.95/SiO2 and the COS 
formed by reaction of CO2 and H2S is taken care by second layer FeO/Al2O3.As the result shows, 
there is no significant reduction in capacity in layered bed, also no detectable amounts of COS 
were seen in the outlet gas composition.  
0 10 20 30 40 50
0. 0
0. 2
0. 4
0. 6
0. 8
1. 0
[
S]
/
[
S]
o
Time  (min s )
 COS  Bre akt h rou gh  
 H
2
S  Bre akt h rou gh
L a ye r e d  B e d
O n l y M n - F e  B e d
O n l y F e O / A l
2
O
3 
B e d
 
167 
 
 
 Figure VI.6 shows COS hydrolysis on carbon and Al2O3/carbon tested at room temperature with 
1000 ppmv COS as challenge gas. The breakthrough curve for bare carbon shows that COS is 
adsorbed with breakthrough time of ~ 7 minutes. The breakthrough curve for Al2O3/C indicates 
that COS was adsorbed upto C/Co= 0.5 and it remains constant at that concentration, formation 
of H2S was evident as shown in the figure. Part of COS is hydrolyzed to H2S.  This indicates that 
Al2O3/C can effectively hydrolyze COS even at room temperature with ambient moisture. In 
another study, commercial (a) Al2O3 (Fischer ~ 90% pure alumina), the (b)pure alumina made by 
calcination (350C/1h) of aluminium nitrate and (c) the physical mixture of this alumina with the 
carbon support were tested for COS hydrolysis to verify if the hydrolysis was due to Al2O3 or 
because of the unique method of preparation. Above mentioned (a-c) sorbents did not show 
hydrolysis of COS to H2S at room temperature. Hence indicating that the alumina loaded via 
impregnation on carbon support hydrolyzes the COS at room temperature. In Figure VI. 7, 
layered bed of Al2O3/C followed by Fe0.025Mn0.025ZnO0.95/SiO2 was used. The composite bed 
design ensures that the H2S formed by COS hydrolysis over Al2O3/C can be taken care by 
Fe0.025Mn0.025ZnO0.95/SiO2. As shown in the figure, the H2S breakthrough curve goes through 
maxima, indicating the H2S formed via COS hydrolysis is taken care by 
Fe0.025Mn0.025ZnO0.95/SiO2. The results indicated here are preliminary results and need further 
work to ascertain the role of alumina, carbon for COS hydrolysis. The authors recommend 
further work for characterization of the sorbent to understand the mechanism of hydrolysis. At 
this point it is speculated that the surface hydroxyls are responsible for the conversion of COS to 
H2S.  
 
168 
 
 
 
Figure VI.6: COS Hydrolysis using Al2O3/C, Test conditions: Co = 1000 ppmv COS/N2, T= 20C, 
Particle Size = 100-200 microns. 
 
 
 
 
 
 
 
 
 
 
169 
 
 
 
Figure VI.7. COS Hydrolysis for extended time on Al2O3/C. Test conditions: Co = 1000 ppmv 
COS/N2, T= 20C, Particle Size = 100-200 microns.  
V.4. Conclusions 
 
The removal of COS from the reformate streams via hydrolysis was the focus of this work. 
Promoted Carbon with alumina shows enhanced catalytic activity for conversion of COS at room 
temperature taking place via surface hydroxyl group. The Mn and Fe promoter cations 
significantly enhance the utilization of the ZnO active phase in reaction with H2S of the novel 
sorbents Fex-Mny-Zn1-x-yO/SiO2 (x, y=0, 0.025), during desulfurization of the mixture of H2S and 
H2 at room temperature. The Mn- and Fe-promoted sorbents maintain a high sulfur uptake 
capacity upon the multiple cycles of a simple thermal oxidative regeneration of the ?spent? 
 
170 
 
sorbent in air (up to 10 cycles). Thus, COS formation can be inhibited by restricting to low 
temperatures (T<250 C). Al2O3 based sorbents are good catalysts for COS hydrolysis at high 
temperature.  COS formed at high temperatures (T=400C) by the presence of CO2 and H2S is 
taken care by the layered bed design.  
 
Acknowledgement 
The authors would like to thank the US Army (TARDEC Contract W56HZV-05-C-0686) for the 
financial support of this work. A.S. thanks Prof. Michael Bowman (Department of Chemistry of 
the University of Alabama at Tuscaloosa) for useful discussions. 
 
 
  
 
171 
 
Chapter VII: Conclusions and Recommendations for Future Work 
 
VII.1. Conclusions 
Conclusions for the work on sulfur removal using promoted ZnO/SiO2 are presented at the 
end of the chapters (III-VI). An overview of all research activities conducted will be mentioned 
here. This study has led to development of novel materials, synthesis and analysis methods for 
effective sulfur removal over wide temperature range. Some of the notable achievements of this 
work are listed below:  
1) Novel Cu promoted ZnO/MCM41 and MCM-48 (Mesoporous silica) was developed for 
H2S removal over wide temperature range 
2) The promoted ZnO/MCM-41 and ZnO/MCM-48 showed highest H2S adsorption capacity 
at room temperature ever reported (~95% mol S/mol ZnO) 
3) A study of effect of temperature, moisture content, metal oxide loading and support 
properties (surface area and pore volume) was performed 
4) The composition of Cu promoted ZnO/SiO2 was optimized (by varying Cu concentration 
0-100%) Cu0.2ZnO0.8/SiO2 showed highest H2S adsorption capacity ( ~92% mol S/mol 
ZnO) 
5) Novel Mn-Fe promoted ZnO/SiO2 sorbent for removal of H2S from the fuel reformate 
streams at ambient conditions developed 
 
172 
 
6) Characterization of the doped sorbents was carried out using XPS, XRD, ESR and N2 
adsorption and in-situ ESR studies to understand the role of the dopants 
7) XRD suggests that both zinc and copper compounds of promoted ZnO/SiO2 sorbents are 
nano-dispersed 
8) The ESR spectroscopy found that the ?calcined? and ?sulfided? CuO-ZnO/SiO2 sorbents 
contain Cu2+ in the single dispersion and coordination state. During H2S adsorption, 
partial reduction of Cu2+ to Cu1+ occurs: the higher Cu concentration in the sorbent, the 
lower the reduction yield of Cu2+ to Cu1+ thus correlating with sulfur uptake capacity. 
9) The ?calcined? sorbent contains Fe3+ and Mn3+ ions, while upon H2S adsorption, their 
reduction to Fe2+ and Mn2+ occurs. Fe3+ ions are believed to occupy the surface of the 
supported ZnO nanocrystallites, while Mn3+ ions are distributed uniformly within ZnO. 
10) Thermal regeneration in air for the sorbent was established and use of promoted 
ZnO/SiO2 over multiple regeneration cycles was demonstrated  
11) The ?deactivated? Cu-ZnO/SiO2 sorbent (10-11 adsorption/regeneration cycles) is 
enriched with the different chemical form of Cu2+, compared to the ?as-prepared? 
sorbent.  
12)  The sorbents were scaled up (20g ? 4kgs)  and the batches showed consistency in the 
sulfur adsorption capacities 
13)  Formation of COS in the reformate streams was reported by understanding the 
thermodynamics of the reactions taking place  
14)  Methods to inhibit COS, removal COS and hydrolyze COS were demonstrated  
15)  COS can be inhibited by operating in the lower temperature regime in the fuel reformate 
streams  
 
173 
 
16) Activated Carbons can remove COS present in the fuel reformates at ambient conditions 
17) FeO/Al2O3 catalyses hydrolysis of COS at higher temperatures (~400 C)  
18) Novel Sorbent Al2O3/Carbon was developed to hydrolyze COS at ambient conditions 
19)  Room temperature hydrolysis and removal of sulfur (COS + H2S) from fuel reformate 
streams was carried out using Al2O3/carbon and promoted ZnO/SiO2 layered beds. 
 
An effective desulfurization composition for fuel reformate streams was developed. 
Performance comparisons with other sorbents indicated high sulfur capacity. The sorbent 
composition was regenerable over multiple cycles. Characterization of these novel sorbents was 
studied. Routes to mitigate COS present/formed in the reformate streams were studied. Novel 
room temperature removal of gas phase sulfur (COS+H2S) using layered beds of Al2O3/Carbon 
and Mn0.025Fe0.025Zn0.95O/SiO2 for COS hydrolysis and removal of H2S respectively was 
demonstrated.  
 
VII.2. Recommendations for Future Work 
Scope of promoted ZnO/SiO2 
Most of the results in this study were focused on the fuel reformate gas compositions. It 
would be recommended to study if the sorbents work efficiently in different stream 
compositions like natural gas, syn-gas in the presence of different sulfur impurities like 
mercaptans, sulfides and aromatic sulfur to improve the scope of use of these sorbents.  
Characterization of promoted ZnO/SiO2 
 
174 
 
Although many characterization techniques were used to understand the exact role of the 
dopants in the ZnO/SiO2, a clear picture of the mechanism is still not obtained. Use of some 
of the advanced techniques like EXAFS and Diffuse reflectance can help depict the defect 
structure of the ZnO/SiO2 matrix. Because the dopant amount is significantly less, the 
techniques chosen should be highly sensitive.  
Use of ESR and in-situ ESR has helped in more than many ways to realize the oxidations 
states of the Cu and Mn-Fe dopants because of its high sensitivity. Further studies might 
reveal more results. 
Also use of chemisorption technique with an appropriate probe molecule should help 
know the dispersion and active surface area of the samples with and without dopants. This 
will give quantitative results change in crystallite size on addition of dopant. There is a 
speculation that addition of dopants adds defects into the structure and reduces the crystallite 
size hence leading to higher saturation capacities [64, 67, 158-162]. This can be verified 
using techniques like chemisorption.  
Characterization of the Al2O3/Carbon sorbent is also essential to understand the 
mechanism of the hydrolysis of COS at room temperature. It is speculated that the surface 
hydroxyls[153] are the reason for this hydrolysis, but it is important to know the nature of the 
hydroxyls present, their role in hydrolysis, their concentration and thermal stability. 
Understanding these aspects would be important in order to develop a more stable and 
effective sorbent for COS hydrolysis.  
 
 
175 
 
References 
 
1. Barbir, F. and S. Yazici, Status and development of PEM fuel cell technology. 
International Journal of Energy Research, 2008. 32(5): p. 369-378. 
2. Barbir, F., PEM Fuel Cells: Theory and Practice: Elsevier Academic Press. 
3. Song, C., An Overview of new approches to deep desulfurization or ultra-clean gasoline, 
diesel fuel and jet fuel. Catalysis Today, 2003. 86. 
4. I.V. Babich, J.A.M., Science and Technology of novel processes for deep desulfurization 
of the oil refinery streams: a review. Fuel, 2003. 82. 
5. Baird, T., Denny, P.J., Hoyle,R., McMonage, F., Stirling,D., Tweedy, J., Modified Zinc 
Oxide absorbents for Low-Temperature gas desulfurisation. J.Chem. Soc. Faraday Trans., 
1992. 88(2). 
6. Xue, M., et al., Screening of adsorbents for removal of H2S at room temperature. Green 
Chemistry, 2003. 5(5): p. 529-534. 
7. Baird, T., et al., Modified Zinc-Oxide Absorbents for Low-Temperature Gas 
Desulfurization. Journal of the Chemical Society-Faraday Transactions, 1992. 88(22): p. 
3375-3382. 
 
176 
 
8. Davidson, J.M., C.H. Lawrie, and K. Sohail, Kinetics of the Absorption of Hydrogen-
Sulfide by High-Purity and Doped High-Surface-Area Zinc-Oxide. Industrial & 
Engineering Chemistry Research, 1995. 34(9): p. 2981-2989. 
9. Yang, H.Y., et al., Breakthrough Characteristics of Reformate Desulfurization Using ZnO 
Sorbents for Logistic Fuel Cell Power Systems. Industrial & Engineering Chemistry 
Research, 2008. 47(24): p. 10064-10070. 
10. Tatarchuk, B.J., Yang,H. and Dhage,P., Doped Zinc oxide sorbents for regenerable 
desulfurization applications US Patent WO 2008/137463 A2, 2008. 
11. B.Tatarchuk, B.C., Y.Lu, L.Chen, E.Luna, D.Cahela, A. University, Editor. 2005: USA. 
12. Sakanishi, K., et al., Simultaneous removal of H2S and COS using activated carbons and 
their supported catalysts. Catalysis Today, 2005. 104(1): p. 94-100. 
13. Westermoreland P.R, H.D.P., Environmental Science and Technology, 1976. 10. 
14. Watson, S., R. Kimmitt, and R.B. Rhinesmith, Study compares COS-removal processes. 
Oil & Gas Journal, 2003. 101(36): p. 66-+. 
15. Husna Atakul, j.P.W., Albert W. Gerritsen and Pieter J.Van den Berg, Regeneration of 
MnO/Alumina used for high temperature desulfurization for fuel gases. Fuel, 1996. 
75(3): p. 6. 
16. Xuepeng Bu, Y.Y., Xuguo ji,  and W.P. Cuiqing Zhang, New development of Zinc-based 
sorbents for hot gas desulfurization Fuel processing Technology 2007. 88: p. 5. 
 
177 
 
17. Yang, H.Y., D.R. Cahela, and B.J. Tatarchuk, A study of kinetic effects due to using 
microfibrous entrapped zinc oxide sorbents for hydrogen sulfide removal. Chemical 
Engineering Science, 2008. 63(10): p. 2707-2716. 
18. Harris, D.K., D.R. Cahela, and B.J. Tatarchuk, Wet layup and sintering of metal-
containing microfibrous composites for chemical processing opportunities. Composites 
Part a-Applied Science and Manufacturing, 2001. 32(8): p. 1117-1126. 
19. Kalluri, R.R., D.R. Cahela, and B.J. Tatarchuk, Microfibrous entrapped small particle 
adsorbents for high efficiency heterogeneous contacting. Separation and Purification 
Technology, 2008. 62(2): p. 304-316. 
20. Yang, H., Y. Lu, and B.J. Tatarchuk, Glass fiber entrapped sorbent for reformates 
desulfurization for logistic PEM fuel cell power systems. Journal of Power Sources, 
2007. 174(1): p. 302-311. 
21. Chang, B.K. and B.J. Tatarchuk, Microfibrous entrapment of small catalyst particulates 
for high contacting efficiency removal of trace CO from practical reformates for PEM H-
2-O-2 fuel cells. Journal of Materials Engineering and Performance, 2006. 15(4): p. 453-
456. 
22. Chang, B.K., et al., Facile regeneration vitreous microfibrous entrapped supported ZnO 
sorbent with high contacting efficiency for bulk H2S removal from reformate streams in 
fuel cell applications. Journal of Materials Engineering and Performance, 2006. 15(4): p. 
439-441. 
 
178 
 
23. Yang, H.Y., et al., Mass transfer study of hydrogen sulfide removal from reformats with 
microfiber entrapped ZnO/SiO2 sorbents for fuel cell application. Abstracts of Papers of 
the American Chemical Society, 2005. 230: p. U1655-U1656. 
24. Lu, Y. and B.J. Tatarchuk, Microfibrous entrapped supported-ZNO sorbents with high 
contacting efficiency for trace H2S removal in PEMFC applications. Abstracts of Papers 
of the American Chemical Society, 2003. 226: p. U48-U48. 
25. Karanjikar, M.R., et al., Logistical fuel to hydrogen; An integrated processing approach 
augmented by microfibrous entrapped polishing catalysts and sorbents for PEM fuel 
cells. Abstracts of Papers of the American Chemical Society, 2004. 228: p. U683-U683. 
26. Cahela, D.R. and B.J. Tatarchuk, Permeability of sintered microfibrous composites for 
heterogeneous catalysis and other chemical processing opportunities. Catalysis Today, 
2001. 69(1-4): p. 33-39. 
27. Twigg, M.V., Catayst Handbook Wolfe, London. 1989. 
28. Ahn, S. and B.J. Tatarchuk, Fibrous metal-carbon composite structures as gas diffusion 
electrodes for use in alkaline electrolyte. Journal of Applied Electrochemistry, 1997. 
27(1): p. 9-17. 
29. Meffert, M.W., PhD Thesis, Auburn University. 1998. 
30. Chang, B.K. and B.J. Tatarchuk, High contacting efficiency microfibrous entrapped 
PROX catalysts for reformate cleanup and PEM fuel cell applications. Abstracts of 
Papers of the American Chemical Society, 2004. 228: p. U686-U686. 
 
179 
 
31. Lu, Y., et al., Microfibrous entrapped ZnO-support sorbents for high contacting 
efficiency H2S removal from reformate streams in PEMFC applications. Microreactor 
Technology and Process Intensification, 2005. 914: p. 406-422. 
32. M.A. Daley, C.L.M., J.A.D. Barr, S. Riha, A. Lizzio, G. Donnals, J. Economy, 
Adsorption of SO2 onto oxidized and heat treated activated carbon fibers (ACFs). 
Carbon, 1997. 35(3): p. 7. 
33. M.N. Bae, M.K.S., Y. Kim, K. Seff, Crystal structure of Mn46Si100Al92O384. 89H2S, a 
hydrogen sulfide sorption complex of fully dehydrated Mn2?-exchanged zeolite X. 
Microporous and Mesoporous Materials, 2003. 63: p. 10. 
34. A. B. Bourlinos, M.A.K., and D. Petridis, Synthesis and Characterization of Iron-
Containing MCM-41 Porous Silica by the Exchange Method of the Template. J. Phys. 
Chem. B, 2000. 104. 
35. X-Y. Hao, Y.-Q.Z., J-W. Wang, W. Zhou, C. Zhang, S. Liu,, A novel approach to prepare 
MCM-41 supported CuO catalyst with high metal loading and dispersion. Microporous 
and Mesoporous Materials, 2005. 88(1-3). 
36. Huang, H.Y., et al., Amine-grafted MCM-48 and silica xerogel as superior sorbents for 
acidic gas removal from natural gas. Industrial & Engineering Chemistry Research, 2003. 
42(12): p. 2427-2433. 
 
180 
 
37. Xiaochun Xu, C.S., , Bruce G. Miller and Alan W. Scaroni, Adsorption separation of 
carbon dioxide from flue gas of natural gas-fired boiler by a novel nanoporous 
?molecular basket? adsorbent. Fuel Processing Technology, 2005. 86(14-15). 
38. A. Corma, A.M.a.V.M.-S., Hydrogenation of Aromatics in Diesel Fuels on Pt/MCM-41 
Catalysts. journal of Catalysis. 169(2). 
39. Reddy, C.S.a.K.M., Mesoporous molecular sieve MCM-41 supported Co?Mo catalyst for 
hydrodesulfurization of dibenzothiophene in distillate fuels Applied Catalysis A: General, 
1999. 176(1). 
40. F. Sch?th, A. Wingen and J. Sauer, Oxide loaded ordered mesoporous oxides for catalytic 
applications. Microporous and Mesoporous Materials, 2001. 44-45: p. 465-476. 
41. Xiu S. Zhao, G.Q.M.L., and Graeme J. Millar, Advances in Mesoporous Molecular Sieve 
MCM-41. industrial & Engineering Chemistry Research, 1996. 35(7). 
42. Pelrine B, P., Schmitt, K. D., Vartuli, J.C., Production of olefin oligomer. 1992. 
43. Souza, M.J.B., et al., HDS of thiophene over CoMo/AlMCM-41 with different Si/Al 
ratios. Applied Catalysis a-General, 2007. 316(2): p. 212-218. 
44. Kloetstra, K.R. and H. Vanbekkum, Base and Acid Catalysis by the Alkali-Containing 
Mcm-41 Mesoporous Molecular-Sieve. Journal of the Chemical Society-Chemical 
Communications, 1995(10): p. 1005-1006. 
 
181 
 
45. Kloetstra, K.R. and H. Vanbekkum, Catalysis of the Tetrahydropyranylation of Alcohols 
and Phenols by the H-Mcm-41 Mesoporous Molecular-Sieve. Journal of Chemical 
Research-S, 1995(1): p. 26-27. 
46. Tanev, P.T., M. Chibwe, and T.J. Pinnavaia, Titanium-Containing Mesoporous 
Molecular-Sieves for Catalytic-Oxidation of Aromatic-Compounds. Nature, 1994. 
368(6469): p. 321-323. 
47. J. S. Beck, J.C.V., W. J. Roth, M. E. Leonowicz, C. T. Kresge, K. D. Schmitt, C. T. W. 
Chu, D. H. Olson, E. W. Sheppard, A new family of mesoporous molecular sieves 
prepared with liquid crystal templates. J. Am. Chem. Soc, 1992. 114(27). 
48. Hao, X.Y., et al., A novel approach to prepare MCM-41 supported CuO catalyst with 
high metal loading and dispersion. Microporous and Mesoporous Materials, 2006. 88(1-
3): p. 38-47. 
49. Ozaydin, Z., S. Yasyerli, and G. Dogu, Synthesis and activity comparison of copper-
incorporated MCM-41-type sorbents prepared by one-pot and impregnation procedures 
for H2S removal. Industrial & Engineering Chemistry Research, 2008. 47(4): p. 1035-
1042. 
50. Wingen, A., et al., Fe-MCM-41 as a catalyst for sulfur dioxide oxidation in highly 
concentrated gases. Journal of Catalysis, 2000. 193(2): p. 248-254. 
 
182 
 
51. Gac, W., et al., The influence of the preparation methods and pretreatment conditions on 
the properties of Ag-MCM-41 catalysts. Journal of Molecular Catalysis a-Chemical, 
2007. 268(1-2): p. 15-23. 
52. Lu, W.J., et al., A novel preparation method of ZnO/MCM-41 for hydrogenation of 
methyl benzoate. Journal of Molecular Catalysis a-Chemical, 2002. 188(1-2): p. 225-231. 
53. Haber J., B.H.a.D.B., Manual of methods and procedures for catalyst characterization. 
Pure and applied Chemistry, 1995. 67. 
54. Ilaria Rosso, C.G., Massimo Bizzi, Guido Saracco, and Vito Specchia, Zinc oxide 
sorbents for the removal of hydrogen sulfide from syngas. Industrial & Engineering 
Chemistry Research, 2003. 42. 
55. Israelson, G., Results of testing various natural gas desulfurization sorbents Journal of 
Materials Engineering and Performance, 2007. 13(3). 
56. P.Dhage, V.G., H.Yang and B. Tatarchuk Wide Temperature Range H2S Removal from 
Fuel Reformate streams using MCM-41 and MCM-48. Microporous and Mesoporous 
Materials, In Press. 
57. Dhage, P., et al., Copper-Promoted ZnO/SiO2 Regenerable Sorbents for the Room 
Temperature Removal of H2S from Reformate Gas Streams. Industrial & Engineering 
Chemistry Research, 2010. 49(18): p. 8388-8396. 
 
183 
 
58. Dhage, P., et al., Regenerable Fe-Mn-ZnO/SiO2 sorbents for room temperature removal 
of H2S from fuel reformates: performance, active sites, Operando studies. Physical 
Chemistry Chemical Physics, 2011. 13(6): p. 2179-2187. 
59. P.Dhage, H.Y.a.B.T., RT Hydrolysis and Removal of COS from Fuel Reformate Streams 
using Al2O3/Carbon+Fmn0.025Fe0.025Zn0.95O/SiO2 Layered Beds. In Press. 
60. K. Schumacher, P.I.R., A.D. Chesne, A.V. Neimark, K.K. Unger, Characterization of 
MCM-48 Materials. Langmuir, 2000. 16(10). 
61. James Larminie , A.D., Fuel cell systems explained. 2003. 
62. Li, L.a.K., D., H2S Removal with ZnO during fuel Processing for PEM fuel cell 
applications. Catalysis Today, 2006. 116. 
63. Song, C., Fuel processing for low-temperature and high-temeperature fuel cells 
Challenges and opportunitiesfor sustainable development in the 21st century. Catalysis 
Today, 2002. 77. 
64. Baird, T., et al., Cobalt-zinc oxide absorbents for low temperature gas desulfurisation. 
Journal of Materials Chemistry, 1999. 9(2): p. 599-605. 
65. Polychronopoulou, K., et al., Novel Fe-Mn-Zn-Ti-O mixed-metal oxides for the low-
temperature removal of H2S from gas streams in the presence of H-2, CO2, and H2O. 
Journal of Catalysis, 2005. 236(2): p. 205-220. 
 
184 
 
66. Laperdrix, E., et al., Selective oxidation of H2S over CuO/Al2O3: Identification and role 
of the sulfurated species formed on the catalyst during the reaction. Journal of Catalysis, 
2000. 189(1): p. 63-69. 
67. Sasaoka, E., et al., Characterization of Reaction between Zinc-Oxide and Hydrogen-
Sulfide. Energy & Fuels, 1994. 8(5): p. 1100-1105. 
68. Quanchang Li, S.E.B., Linda J. Broadbelt, Jian-Guo Zheng and N. Q. Wu, Synthesis and 
characterization of MCM-41-supported Ba2SiO4 base catalyst. Microporous and 
Mesoporous Materials, 2003. 59(2-3). 
69. Jae Wook Lee, D.L.C., Wang Geun Shim, and Hee Moon, Application of mesoporous 
MCM-48 and SBA-15 materials for the separation of biochemicals dissolved in aqueous 
solution. Korean Journal of Chemical Engineering, 2003. 21(1). 
70. Klotz, I.M., The Adsorption Wave. Chemical Reviews, 1946. 39(2): p. 241-268. 
71. Wheeler, A. and A.J. Robell, Performance of Fixed-Bed Catalytic Reactors with Poison 
in Feed. Journal of Catalysis, 1969. 13(3): p. 299-&. 
72. Yoon, Y.H. and J.H. Nelson, Application of Gas-Adsorption Kinetics .1. A Theoretical-
Model for Respirator Cartridge Service Life. American Industrial Hygiene Association 
Journal, 1984. 45(8): p. 509-516. 
73. Amundson, N.R., A Note on the Mathematics of Adsorption in Beds. Journal of Physical 
and Colloid Chemistry, 1948. 52(7): p. 1153-1157. 
 
185 
 
74. A. Erzos, H. Olgun, and S. Ozdogan, Reforming Options for Hydrogen Production from 
Fossil Fuels for PEM Fuel Cells. J. Power Sources, 2006(154): p. 67. 
75. A. Erzos, et al., Autothermal Reforming as a Hydrocarbon Fuel Processing Option for 
PEM Fuel Cell. Journal of Power Sources, 2003. 118: p. 384-392. 
76. I. I. Novochinskii, et al., Low Temperature H2S Removal from Steam-Containing Gas 
Mixtures with ZnO for Fuel Cell Application. 1. ZnO Particles and Extrudates. Energy 
Fuels, 2004. 18: p. 576. 
77. Y. Lu, et al. Microfibrous Entrapped ZnO-Supported Sorbent for High Contacting 
Efficiency H2S Removal from Reformate Streams in PEMFC Applications. in ACS 
Symposium Series 914. 2005. Washington, DC: American Chemical Society. 
78. L. Alonso, et al., Characterization of Mn and Cu oxides as regenerable sorbents for hot 
coal gas desulfurization. Fuel Processing Technology, 2000. 62: p. 31-44. 
79. J.H. Swisher and K. Schwerdtfeger, Review of metals and binary oxides as sorbents for 
removing sulfur from coal-derived gases. J. Mater. Eng. Perform., 1992. 1: p. 399. 
80. E. Garc?a, et al., Thermogravimetric study of regenerable sulphur sorbents for H2S 
retention at high temperature. Thermochim. Acta, 1997. 306: p. 23. 
81. R.V. Siriwardane and J. Poston, Characterization of copper oxides, iron oxides and zinc 
copper ferrite desulfurization sorbents by X-ray photoelectron spectroscopy and scanning 
electron microscopy. Appl. Surf. Sci., 1993. 68: p. 65. 
 
186 
 
82. S.S. Tamhankar, M. Bagajewicz, and G.R. Gavalas, Mixed-oxide sorbents for high-
temperature of hydrogen sulfide. Ind. Eng. Chem. Process Des. Dev., 1986. 25: p. 429. 
83. Z. Li and M. Flytzani-Stephanopoulos, Cu?Cr?O and Cu?Ce?O regenerable sorbents for 
hot gas desulfurization. Ind. Eng. Chem. Res., 1997. 36: p. 187. 
84. E. Giamello, B. Fubini, and P. Lauro, Investigation of the State of Copper in Copper-Zinc 
Oxice Biphasic Catalyst by EPR and Adsorption Microcalorimetry. Applied Catalysis, 
1986. 21: p. 133-147. 
85. H.Y. Chen, et al., Synergism between Cu and Zn sites in Cu/Zn catalysts for methanol 
synthesis. Applied Surface Science, 1999. 152: p. 193-199. 
86. Z. Wang, W. Wang, and G. Li, Studies on the active species and on dispersion of Cu in 
Cu/SiO2 and Cu/Zn/SiO2 for hydrogen production via methanol partial oxidation. 
International Journal of Hydrogen Energy, 2003. 28: p. 151-158. 
87. J.-D. Grunwaldt, et al., In Situ Investigations of structural changes in Cu/ZnO Catalysts. 
Journal of Catalysis, 2000. 194: p. 452-460. 
88. Isao Nakamura, et al., Role of oxygen vacancy in the plasma-treated TiO2 photocatalyst 
with visible light activity for NO removal. Journal of Molecular Catalysis A: Chemical, 
2000. 161: p. 205?212. 
89. B.J. Tatarchuk, H. Yang, and P. Dhage, Doped zinc oxide sorbents for regenerable 
desulfurization applications. US Patent WO 2008/137463 A2, 2008. 
 
187 
 
90. H. Yang, et al., Breakthrough Characteristics of Reformate Desulfurization Using ZnO 
Sorbents for Logistic Fuel Cell Power Systems. Ind. Eng. Chem. Res., 2008. 47: p. 
10064?10070. 
91. H. Yang, Y. Lu, and B.J. Tatarchuk, Glass fiber entrapped sorbent for reformates 
desulfurization for logistic PEM fuel cell power systems. Journal of Power Sources, 
2007. 174(1): p. 302-311. 
92. Y. Lu, et al., Microfibrous entrapped ZnO-support sorbents for high contacting efficiency 
H2S removal from reformate streams in PEMFC applications. Microreactor Technology 
and Process Intensification, 2005. 914: p. 406-422. 
93. J. H. Zeng, et al., Precursor, base concentration and solvent behavior on the formation of 
zinc silicate. Materials Research Bulletin, 2009. 44: p. 1106?1110. 
94. J.A. Lercher, H. Vinek, and H. Noller, TiO2/ZnO mixed oxide catalyst, characterization 
by X-ray photoelectron and infrared spectroscopy and reactions with propanol and 
butanol. Appl. Cat., 1984. 12: p. 293-307. 
95. Md Nurul Islam, et al., XPS and X-ray diffraction studies of aluminum-doped zinc oxide 
transparent conducting films. Thin Solid Films, 1996. 280: p. 20-25. 
96. D. Briggs and J.T. Grant, Surface Analysis by Auger and Electron Spectroscopy. 2003, 
Trowbridge: IM Publications and SurfaceSpectra Ltd. 
97. Zhengping Fu, et al., An intense ultraviolet photoluminescence in sol?gel ZnO?SiO2 
nanocomposites. J. Phys.: Condens. Matter, 2003. 15: p. 2867?2873. 
 
188 
 
98. S. Takeda and M. Fukawa, Surface OH groups governing surface chemical properties of 
SiO2 thin films deposited by RF magnetron sputtering. Thin Solid Films, 2003. 444: p. 
153-157. 
99. A. Bensalem, et al., Preparation and characterization of highly dispersed silica-supported 
ceria. Applied Catalysis A: General, 1995. 121: p. 81-93. 
100. S. K. Nandi, et al., Structural and optical properties of ZnO films grown on silicon and 
their applications in MOS devices in conjunction with ZrO2 as a gate dielectric. Bull. 
Mater. Sci., 2007. 30(3): p. 247-254. 
101. S. Contarini and L. Kevan, X-ray Photoelectron Spectroscopic Study of Copper-
Exchanged X- and Y-Type Sodium Zeolites: Resolution of Two Cupric Ion Components 
and Dependence on Dehydration and X-Irradiation. J. Phys. Chem., 1986. 90(8): p. 1630-
1632. 
102. R. Schlesinger, H. Klewe-Nebenius, and M. Bruns, Characterization of Artificially 
Produced Copper and Bronze Patina by XPS. Surf. Interface Anal., 2000. 30: p. 135-139. 
103. Handbook of Surface and Interface Analysis: Methods for Problem-Solving, ed. J. C. 
Riviere and S. Myhra. 1998, New York: Marcel Dekker, Inc. 
104. J.W.L. Wong, et al., XPS studies of Auger parameter shift of ZnS Te alloys. Journal of 
Electron Spectroscopy and Related Phenomena, 2001. 113: p. 215?220. 
105. Corrosion of Electronic and Magnetic Materials, ed. P.J. Peterson. 1992, Philadelphia: 
American Society for Testing and Materials. 
 
189 
 
106. Brion, D., Study by photoelectron-spectroscopy of surface degradation of FeS2, CuFeS2, 
ZnS and PbS exposed to air and water. Appl. Surf. Sci., 1980. 5: p. 133-152. 
107. D.L. Perry and J.A. Taylor, X-Ray photoelectron and Auger spectroscopic studies of 
Cu2S and CuS. Journal of Materials Science Letters, 1986. 5: p. 384-386. 
108. J.-S. Yu and L. Kevan, Cu(II)-Adsorbate Interactions in Cu(II)-Exchanged K-L Zeolite. 
J. Phys. Chem., 1994. 98: p. 12436-12441. 
109. M.G. Kalchev, A.A. Andreev, and N.S. Zotov, Water-Gas Shift Reaction on CuO-ZnO 
Catalysts. I. Structure and Catalytic Activity. Kinetics and Catalysis, 1995. 36(6): p. 821-
827. 
110. R.T. Figueriedo, et al., Spectroscopic Evidence of Cu-Al Interactions in Cu-Zn-Al Mixed 
Oxide Catalysts Used in CO Hydrogenation. Journal of Catalysis, 1998. 178: p. 146-152. 
111. A. Martinez-Arias, et al., Effect of Copper-Ceria Interactions on Copper Reduction in a 
Cu/CeO2/Al2O3 Catalyst Subjected to Thermal Treatments in CO. J. Phys. Chem. B, 
1998. 102: p. 809-817. 
112. Surface Chemistry and Physics, ed. A. F. Carley, et al. 2002, New York: Kluwer 
Academic/Plenum Publishers. 
113. A. L. Yakovlev, et al., DFT Study of oxygen-bridged Zn2+ ion pairs in Zn/ZSM-5 
zeolites. Catalysis Letters, 2000. 70: p. 175-181. 
 
190 
 
114. K. Hashimot and N. Toukai, Methanol decomposition over copper particles incorporated 
in the interlayer regions of zinc aluminum silicate hydroxide. Journal of Molecular 
Catalysis: A, 2002. 186: p. 79-88. 
115. Kolosov, ESR Study of Adsorption of Hydrogen-sulfide, Propanethiol, and Thiophene on 
Molybdenum Oxide and Sulfide Deposited on Silica-gel. Kinetics and Catalysis, 1975. 
16(1): p. 161-164. 
116. A. A. Spozhakina, et al., The Interaction of Hydrogen Sulfide with Silica-Supported 
Phosphorus-Molybdenum Heteropolyacid. Kinetics and Catalysis, 1993. 34(6): p. 976-
979. 
117. L. Trouillet, et al., In situ characterization of the coordination sphere of Cu(II) complexes 
supported on silica during the preparation of Cu/SiO2 catalysts by cation exchange. Phys. 
Chem. Chem. Phys., 2000. 2: p. 2005-2014. 
118. Sangho Yoon, et al., Self-sustained operation of a kWe-class kerosene-reforming 
processor for solid oxide fuel cells. Journal of Power Sources, 2009. 192: p. 360-366. 
119. Loredana Magistri, et al., Design and Off-Design Analysis of a MW Hybrid System 
Based on Rolls-Royce Integrated Planar Solid Oxide Fuel Cells. Transactions of the 
ASME, 2007. 129: p. 792-797. 
120. Song, C., Fuel Processing for Low-temperature and High-temperature Fuel Cells: 
Challenges, and Opportunities for Sustainable Development in the 21st Century. Catal. 
Today, 2002. 77(1-2): p. 17-49. 
 
191 
 
121. Ke Liu, Chushan Song, and V. Subramani, Hydrogen and Syngas Production and 
Purification Technologies. 2010, Hoboken, NJ: John Wiley & Sons, Inc. 545. 
122. Lee D. Gasper-Galvin, Aysel T. Atimtay, and R.P. Gupta, Zeolite-Supported Metal Oxide 
Sorbents for Hot-Gas Desulfurization. Ind. Eng. Chem. Res., 1998. 37: p. 4157-4166. 
123. Singfoong Cheah, Daniel L. Carpenter, and K.A. Magrini-Bair, Review of Mid- to High-
Temperature Sulfur Sorbents for Desulfurization of Biomass- and Coal-derived Syngas. 
Energy Fuels, 2009. 23(11): p. 5291?5307. 
124. P. R. Westmoreland and D.P. Harrison, Evaluation of Candidate Solids for High-
Temperature Desulfurization of Low Btu Gases. Environ. Sci. Technol., 1976. 10: p. 659. 
125. J. H. Swisher and K. Schwerdtfeger, Thermodynamic Analysis of Sorption Reactions for 
the Removal of Sulfur from Hot Gases. J. Mater. Eng. Perform., 1992. 1(4): p. 565-571. 
126. J.H. Swisher and K. Schwerdtfeger, Review of Metals and Binary Oxides as Sorbents for 
Removing Sulfur from Coal-Derived Gases. Journal of Materials Engineering and 
Performance, 1992. 1(3): p. 399-407. 
127. K. Polychronopoulou, et al., Novel Fe?Mn?Zn?Ti?O mixed-metal oxides for the low-
temperature removal of H2S from gas streams in the presence of H2, CO2, and H2O. 
Journal of Catalysis, 2005. 236: p. 205?220. 
128. K. Polychronopoulou, J.L.G. Fierro, and A.M. Efstathiou, Novel Zn?Ti-based mixed 
metal oxides for low-temperature adsorption of H2S from industrial gas streams. Applied 
Catalysis B: Environmental, 2005. 57: p. 125-137. 
 
192 
 
129. Kyriaki Polychronopoulou and A.M. Efstathiou, Effects of Sol-Gel Synthesis on 5Fe-
15Mn-40Zn-40Ti-O Mixed Oxide Structure and its H2S Removal Efficiency from 
Industrial Gas Streams. Environ. Sci. Technol., 2009. 43: p. 4367-4372. 
130. Sam Bergh, et al., Gas phase oxidation of ethane to acetic acid using high-throughput 
screening in a massively parallel microfluidic reactor system. Appl. Catal. A: General, 
2003. 254: p. 67-76. 
131. Mario Chiesa, Elio Giamello, and M. Che, EPR Characterization and Reactivity of 
Surface-Localized Inorganic Radicals and Radical Ions. Chem. Rev., 2010. 110: p. 1310-
1347. 
132. U.J. Katter, et al., Electron Spin-Resonance (ESR) Studies of Adsorbate Dynamics on 
Single Crustal Surfaces: Possibilities and Limitations. Ber. Bunsenges. Phys. Chem., 
1993. 97(3): p. 341-353. 
133. Venezia, A.M., X-ray photoelectron spectroscopy (XPS) for catalysts characterization. 
Catalysis Today, 2003. 77: p. 359?370. 
134. Thomas Baird, et al., Modified Zinc Oxide Absorbents for Low-temperature Gas Desulf 
ur isat ion. J. Chem. Soc. Faraday Trans., 1992. 88(22): p. 3375-3382. 
135. No-Kuk Park, et al., The preparation of a high surface area metal oxide prepared by a 
matrix-assisted method for hot gas desulphurization. Fuel Processing Technology, 2005. 
84: p. 2165-2171. 
 
193 
 
136. Corrie L. Carnes and K.J. Klabunde, Unique Chemical Reactivities of Nanocrystalline 
Metal Oxides toward Hydrogen Sulfide. Chem. Mater., 2002. 14: p. 1806-1811. 
137. Yi-Keun Song, et al., Reactivity of Copper Oxide-Based Sorbent in Coal Gas 
Desulfurization. Korean J. Chem. Eng., 2000. 17(6): p. 691-695. 
138. Si Ok Ryu, et al., Multicyclic Study on Improved Zn/Ti-Based Desulfurization Sorbents 
in Mid-Temperature Conditions. Ind. Eng. Chem. Res., 2004. 43: p. 1466-1471. 
139. Zhao-Tie Liu, et al., XPS Study of Iron Catalysts for Fischer-Tropsch Synthesis. Fuel 
Science and Technology Int'l., 1995. 13(5): p. 559-567. 
140. Scofield, J.H., Hartree-Slater Subshell Photoionization Cross-Sections at 1254 and 1487 
eV. Journal of Electron Spectroscopy and Related Phenomena, 1976. 8: p. 129-137. 
141. Maasaki Yanagisawa, et al., Discoloration of Alumina Ceramics with Ultraviolet 
Radiation and Elucidation of its Mechanism. Journal of the Ceramic Society of Japan. 
101(10): p. 1189-1191. 
142. Reinhard Stoesser, et al., A Magnetic Resonance Investigation of the Process of 
Corundum Formation Starting from Sol-Gel Precursors. J. Amer. Ceram. Soc., 2005. 
88(10): p. 2913-2922. 
143. El-M. El-Malki, R. A. van Santen, and W.M.H. Sachtler, Introduction of Zn, Ga, and Fe 
into HZSM-5 Cavities by Sublimation:  Identification of Acid Sites. J. Phys. Chem. B, 
1999. 103(22): p. 4611?4622. 
 
194 
 
144. Blake J. Aronson, Christopher F. Blanford, and A. Stein, Synthesis, Characterization, and 
Ion-Exchange Properties of Zinc and Magnesium Manganese Oxides Confined within 
MCM-41 Channels. J. Phys. Chem. B, 2000. 104: p. 449-459. 
145. W. Sjoerd Kijlstra, et al., Characterization of Al2O3-Supported Manganese Oxides by 
Electron Spin Resonance and Diffuse Reflectance Spectroscopy. J. Phys. Chem. B, 1997. 
101: p. 309-316. 
146. J. J. Villafranca, et al., Metal-Metal Interactions in Enzymes: EPR and NMR 
Investigations. Inorganica Chimica Acta, 1983. 79: p. 18. 
147. Colin Rhodes, S.A.R., John West , B. Peter Williams , Graham J. Hutchings, The low-
temperature hydrolysis of carbonyl sulfide 
and carbon disulfide: a review. Catalysis Today, 2000. 59. 
148. Hasler, P. and T. Nussbaumer, Gas cleaning for IC engine applications from fixed bed 
biomass gasification. Biomass and Bioenergy, 1999. 16(6): p. 385-395. 
149. Bruno, P.D.N.S.a.T.J., Carbonyl Sulfide: A Review of Its Chemistry and Properties. 
Industrial & Engineering Chemistry Research, 2002. 41(22). 
150. Liu, Z.-T., et al., Deactivation model of Fischer-Tropsch synthesis over an FeCuK 
commercial catalyst. Applied Catalysis A: General, 1997. 161(1-2): p. 137-151. 
 
195 
 
151. John D. Sherman, A.T.K., Suppression of COS formation in Molecular sieve purification 
of hydrocarbon gas streams. 1982, Union Carbide Corporation, Danbury , Conn. : United 
States. 
152. Li, W., W. Shudong, and Y. Quan, Removal of carbonyl sulfide at low temperature: 
Experiment and modeling. Fuel processing Technology, 2010. 91(7): p. 777-782. 
153. Li, W., W. Shudong, and Y. Quan, Removal of carbonyl sulfide at low temperature: 
Experiment and modeling. Fuel processing Technology, 2010 91(7): p. 777-782. 
154. Williams, B.P., et al., Carbonyl sulphide hydrolysis using alumina catalysts. Catalysis 
Today, 1999. 49(1-3): p. 99-104. 
155. West, J., et al., Ni- and Zn-promotion of [gamma]-Al2O3 for the hydrolysis of COS 
under mild conditions. Catalysis Communications, 2001. 2(3-4): p. 135-138. 
156. Shishao, T., et al., Compensation effect in catalytic hydrolysis of carbonyl sulfide at 
lower temperature compensation effect in COS hydrolysis. Catalysis Letters, 1991. 8(2): 
p. 155-167. 
157. Miura, K., et al., Simultaneous removal of carbonyl sulfide and hydrogen sulfide from 
coke oven gas at low temperature by iron oxide. Industrial & Engineering Chemistry 
Research, 1992. 31(1): p. 415-419. 
158. Baird, T., et al., Mixed cobalt-iron oxide absorbents for low-temperature gas 
desulfurisation. Journal of Materials Chemistry, 2003. 13(9): p. 2341-2347. 
 
196 
 
159. Baird, T., et al., Structural and morphological studies of iron sulfide. Journal of the 
Chemical Society-Faraday Transactions, 1996. 92(3): p. 445-450. 
160. Baird, T., et al., Characterisation of cobalt-zinc hydroxycarbonates and their products of 
decomposition. Journal of Materials Chemistry, 1997. 7(2): p. 319-330. 
161. Baird, T., et al., Mixed Co-Zn-Al Oxides as Absorbents for Law-Temperature Gas 
Desulfurization. Journal of the Chemical Society-Faraday Transactions, 1995. 91(18): p. 
3219-3230. 
162. Baird, T., Characterization of Carbons Formed on Nickel Surfaces. Fuel, 1984. 63(8): p. 
1081-1088. 
 
 
 
 
197 
 
Appendix I ? Calculation formulae 
 
Calculation of breakthrough capacity, saturation capacity and % theoretical capacity 
1. Breakthrough capacity: Breakthrough is taken as 2% of CO 
)(.
1032( m i n )m i n 3332
gms o r b e n tWt
mgm o lg m Sg h t i m eb r e a k t h r o umm o lRT PmSHf l o w r a t e o fv o l u m e t r i c
 
2. Saturation Capacity:  
The nature of the curve decides the method that can be used to calculate saturation capacity. In 
this work, we use the t1/2 method since the breakthrough curves look very sharp and S-shaped.  
The time at which the C/Co =0.5 is taken as t1/2. 
 
)(.
1032( m i n )m i n 3
213
3
2
gms o r b e n tWt
mgm o lg m Stmm o lRTPmSHf l o w r a t e o fv o l u m e t r i c
 
 
 
198 
 
3. % of Theoretical capacity: 
The theoretical capacity for ZnO is ~ 392 g S/g sorbent. The % with the saturation capacity 
obtained is known as the % of theoretical capacity.  
 
4. ZnO utilization: 
 
( % ))(.
1081( m i n )m i n 3
213
3
2
Z n O l o a d i n ggms o r b e n tWt
mgm o lg m Z n Otmm o lRTPmSHf l o w r a t e o fv o l u m e t r i c
 
 
 
 
 
 
 
 
 
 
 
 
199 
 
Appendix II ? GC Chromatography Analytic Methods 
 
a. TCD Analysis Method 
 
 Gas Chromatography Model  Varian CP3800 
 Column type Packed Column  HayeSep Q, 80/100 8?  1/8? SS 
 Column Stabilization time   2.00 min 
 Oven Temperature   80oC 
 Injector Temperature   80oC  
 Detector Temperature  175 oC 
 Filament Temperature  375 oC  
 Carrier Gas   H2  
 Carrier Gas Flow Rate  60mL/min 
 6 ?port valve is switched to ?inject? mode at the beginning of every minute and switched 
back to ?fill? mode 2 seconds after injection 
 
 
 
 
 
200 
 
b. PFPD Analysis Method 
 
 Gas Chromatography Model  Varian CP3800 
 Column  Restek XTI (30mm 0.25mm 0.5 m) 
 Oven Temperature Program:  60 oC for 1 min, Ramp to 90 oC at the rate of 20 oC/min and 
stay at 90 oC for 3.5 min. Total runtime is 6 minutes. 
 Injector Temperature  80oC 
 Capillary Flow Rate  1.2 oC 
 Air Flow Rate  17 cm3/min 
 H2 Flow Rate  13 cm3/min 
 Split Ratio  200 
 Tube Voltage  510 V 
 Trigger level  200 mA 
 Sample Delay  4 ms 
 Sample Width   10 ms 
 Gain Factor  2 
 Syringe Size  250 L 
 
 
 
 
 
 
201 
 
Appendix III ? Calibration of Gases 
 
a. Carbon Dioxide 
 
 
 
 
 
 
 
202 
 
 
 
b. Nitrogen  
 
= 0.9997 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
50 100 150 200 250 300
50
100
150
200
250
300
A
c
tua
l
 Rea
di
ng 
(
c
c
/mi
n)
MFC Readi ng ( c c /mi n)
 
203 
 
 
c. Carbon Monoxide 
 
 
 
 
 
 
 
 
 
204 
 
 
d. Hydrogen Sulfide 
 
 
 
 
 
 
 
 
 
205 
 
 
e. Carbonyl Sulfide 
 
 
 
 
 
 
 
 
 
 
206 
 
f. Furnace  
 
 
 
 
 
 
 
 
 
207 
 
Appendix IV ? Inventory of Chemicals used 
 
 
Chemical Vendor/Company 
Act.Carbon Centaur HSL Calgon Carbon Corporation 
Act.Carbon Minotaur OC Calgon Carbon Corporation 
Act.Carbon Centaur 4 x 6 Calgon Carbon Corporation 
Act.Carbon BPL 4 x 6 Calgon Carbon Corporation 
Selexsorb COS BASF 
ZnO SG9201 BASF 
Cr(NO3)3 Aldrich 
Zn(NO3)2.6H2O Aldrich 
Ni(NO3)2 Aldrich 
ZnO Alfa 
Ag(NO3)2 Alfa 
Fe(NO3)3.9H2O Alfa 
Zn(NO3)2.6H2O Fluka 
Cu(NO3)2.3H2O Fluka 
Cs(NO3) Alfa 
 
208 
 
Fe(NO3)3.9H2O Aldrich 
Zn(NO3)2.6H2O Fischer 
Cu(NO3)2.3H2O Fischer 
Mg(NO3)2.xH2O Fischer 
Mn(NO3)2.xH2O Adrich 
Cu(NO3)2.2.5H2O Aldrich 
Al2(NO3)3 Aldrich 
Fe(NO3)3 Aldrich 
ZrO(NO3)2.xH2O Aldrich 
Ni(NO3)3 Aldrich 
Cd(NO3)3 Aldrich 
Ce(NO3)3 Fluka /Alfa 
Co(NO3)2 Aldrich 
Mol. Sieves Strem Chemicals 
NaY Zeolite Strem Chemicals/ Grace Davison 
SiO2 Grade 10181 Aldrich 
Act. Al2O3 Aldrich 
Silica 60A Alfa 
TiO2 Alfa/Saint Gobain 
Silica Grade 10184 Sigma 
Silica Strem Chemicals 
Silica Alfa 
Zeolite Zeolyst 
 
209 
 
Silica Grade 40 Aldrich 
Mol. Sieves 13X Aldrich 
Silica 126724 Fischer 
Selexorb CDX BASF 
PICA Carbon PICA 
Cu-ZnO Actisorb Sud-Chemie 
Silica Grace Davison 
TEOS Aldrich 
Mol. Sieves  3A Aldrich 
Mol. Sieves  4A Aldrich 
Mol. Sieves  5A Aldrich 
Glass Beads Fischer ? 11-312D 4mm