Hydrogen Production in Supercritical Water 
 
by 
 
Adam Joseph Byrd 
 
 
 
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:  Hydrogen Production, Supercritical Water,  
Renewable Energy, Hydrothermal Stability 
 
 
Copyright 2011 by Adam Joseph Byrd 
 
 
Approved by 
 
Ram B. Gupta, Chair, Professor of Chemical Engineering 
Sushil Adhikari, Assistant Professor of Biosystems Engineering 
Mario R. Eden, Associate Professor of Chemical Engineering 
Yoon Y. Lee, Professor of Chemical Engineering 
Christopher B. Roberts, Professor of Chemical Engineering 
?
 
 
 
 
 
 
 Abstract 
 
 
 Hydrogen has been considered as an environmentally friendly energy carrier for 
the future.  It is not found freely in nature but must be obtained by processing hydrogen 
containing compounds.  Current industrial-scale hydrogen production relies on the fossil 
fuel feedstocks.  The processing of fossil fuels for hydrogen production may represent a 
more efficient use, but ultimately still contributes greenhouse gases into the atmosphere.  
Fossil fuels are not the only possibility for a hydrogen containing feedstock, however.  
Biomass and biomass-derived materials also can be processed to generate hydrogen.  
Carbon dioxide generated during hydrogen production from biomass will later be fixed 
by plants during photosynthesis, thus creating a closed loop with no net increase in CO
2
. 
 The properties of water above its critical point (T: 374?C, P: 221 bar) are 
markedly different than under ambient conditions, making it interesting both as solvent 
and reactant.  In the supercritical phase the dielectric constant of water is greatly reduced 
and accordingly it behaves as an organic solvent, easily dissolving many organic species 
and gases while precipitating polar salts.  As a homogenous phase with low viscosity and 
high diffusivity, transport limitations can be overcome in supercritical water.  Physical 
properties of supercritical water such as density, heat capacity, and ion product can be 
tuned by small changes in temperature and pressure to enhance reaction rate and reduce 
volume requirement for reactors.  A further benefit conferred by the supercritical water 
 ii
process is that water is compressed in the liquid state allowing the produced hydrogen to 
be obtained directly at high pressure without the need for energy intensive compression. 
 Four biomass-derived compounds have been examined as feedstocks for 
hydrogen production in supercritical water by catalytic reforming in a continuous flow 
reactor.  The flow-type reactor allowed the attainment of short residence times of seconds 
unavailable to previous researchers operating batch reactors.  First, glucose was used as a 
model compound for biomass (Chapter 2).  The presence of the ruthenium catalyst 
greatly increased the conversion and hydrogen yield from glucose while significantly 
reducing char and tar formation.  Feed concentrations of up to 5 wt% glucose gave a 
hydrogen yield near the theoretical maximum at 700?C with a residence time of only two 
seconds.  Ethanol (Chapter 3) was investigated as a feedstock for hydrogen production as 
it is already produced for use as an automotive fuel additive, however its conversion to 
hydrogen for use in a fuel cell would greatly increase its efficiency.  Full conversion to 
gaseous products was seen above 700?C with no coke formation being observed below 
10 wt% ethanol feed.  Varying pressure from 221 to 276 bar had little effect on the gas 
yields.  The third biomass-derived feedstock used was glycerol (Chapter 4), which is 
obtained as a byproduct from biodiesel manufacturing by transesterification of vegetable 
oils.  Hydrogen yields near the theoretical limit were obtained for dilute solutions with a 
1s residence time at 800?C, while hydrogen yields dropped with longer residence times 
due to methanation.  Feed concentrations of up to 40 wt% glycerol were also gasified at 
800?C and 1 s residence time with no coke formation and the yield of product gases 
closely following equilibrium values.  Liquefied switchgrass biocrude was evaluated as 
the fourth feedstock for hydrogen production in Chapter 5.  Nickel, cobalt, and ruthenium 
 iii
catalysts were prepared on titania, zirconia, and magnesium aluminum spinel supports to 
create a suite of nine catalysts.  These were evaluated for hydrogen production by 
gasification of switchgrass biocrude in supercritical water at 600?C and 250 bar.  
Magnesium aluminum spinel was seen to be an inappropriate support as reactors quickly 
plugged.  Ni/ZrO
2
 gave 0.98 mol H
2
/mol C, the highest hydrogen yield of all tested 
catalysts; however, over time, increase in pressure drop lead to reactor plugging with all 
zirconia supported catalysts.  Titania supported catalysts gave lower conversions, 
however they did not plug during the course of the study.   Charring of all catalysts was 
seen to occur at the entrance of the reactor as the biocrude was heated.  All support 
materials suffered significant surface area loss due to sintering.   
The severity of water?s critical point can lead to sintering and phase 
transformations of catalyst support materials.  Cerium-coated ?-alumina (Chapter 6) and 
binary oxides of aluminum, titanium, and zirconium (Chapter 7) were synthesized as 
potential catalyst supports and evaluated for their stability in hot compressed water.  ?-
Al
2
O
3
 modified with 1-10 wt% Ce was examined, specifically in the temperature range of 
500 ? 700?C at 246 bar.  Transformations of the ? phase were slowed but not prevented.  
Based on X-ray analysis, the transformation of ? -Al
2
O
3
 proceeded through the ? phase 
toward the stable ? phase.  Reduced cerium species were seen to be oxidized in the 
supercritical water environment, and low Ce-loading supports maintained the highest 
BET surface areas.  The stabilization was greatest at 700?C, where Ce-modified aluminas 
retained significantly higher specific surface areas than unmodified alumina.  Binary 
oxides of aluminum, titanium, and zirconium with 1:1 mole ratios of the component 
metals were synthesized by a coprecipitation method. Their stability in sub- and 
 iv
supercritical water was evaluated at 25 MPa over a temperature range of 350 ? 650 ?C for 
a period of three hours by XRD and BET studies.  The compound ZrTiO
4
 was 
crystallographically stable at all conditions.  It maintained its surface area in subcritical 
water, although it sintered and lost much of its pore volume in supercritical water.  
ZrO
2
/Al
2
O
3
 maintained high surface area up to 450?C, but sintered above this 
temperature as a result of phase transformation of both ZrO
2
 and Al
2
O
3
.  The TiO
2
/Al
2
O
3
 
mixed oxide, while having the highest initial surface area, sintered extensively following 
all hydrothermal treatments.  Alumina in the TiO
2
/Al
2
O
3
 system hydrolyzed in subcritical 
water and transformed to corundum in supercritical water, while anatase titania was 
transformed to rutile only at 650?C. 
?
 
 v
 
 
 
 
 
Acknowledgements 
 
 
 I wish to offer my thanks to my advisor Dr.  Ram B. Gupta, whose continued 
guidance and advice has been inestimable in my dissertation research.  He has provided 
encouragement in the face of difficulty as well as a model of excellence in all endeavors.   
 I thank the members of my research group both past and present who have helped 
me along the way.  Dr. Ganesh Sanganwar, Dr. Sandeep Kumar, Dr. K.K. Pant, Dr, 
Melinda Hemingway, Dr. Lingzhao Kong, Hema Ramsurn, Courtney Ober, and 
Siddharth Gumuluru are all owed many thanks for long and helpful discussions of 
research topics.  I also thank Dr. Sachin Nair for assistance with BET measurements and 
Dr. Aimee Poda for her help with operation of TGA-MS. 
 I must also pay homage to the memory of my grandfather Joseph V. Saliga, who 
provided the inspiration and dedication to do research in the field of hydrogen 
production. 
 Finally I wish to acknowledge the financial support of the Department of Energy 
through the NETL Consortium of Fossil Fuel Science, DOD/TACOM/TARDEC, and the 
Alabama Centre for Paper and Bioresource Engineering. 
 vi
 
 
 
 
 
Table of Contents 
 
 
Abstract................................................................................................................................ii 
Acknowledgements.............................................................................................................vi 
List of Tables.....................................................................................................................xii 
List of Figures...................................................................................................................xiii 
1.  Introduction.....................................................................................................................1 
1.1  Hydrogen Properties.........................................................................................2 
1.2  Supercritical Water  Background and Properties..............................................3 
1.3  Hydrogen Production from Hydrocarbons........................................................5 
1.4  Hydrogen Production from Biomass and Model Compounds..........................8 
1.4.1  Aqueous Phase Reforming.................................................................8 
1.4.2  Gasification........................................................................................8 
1.4.3  Supercritical Water Gasification........................................................9 
1.5  Water Electrolysis...........................................................................................13 
1.6  Other Relevant Previous SCW work..............................................................13 
1.6.1  Supercritical water oxidation (SCWO) ............................................13 
1.6.2  Corrosivity in SCW environment....................................................17 
1.7  References.......................................................................................................20 
2.  Hydrogen Production from Glucose.............................................................................27 
2.1  Abstract...........................................................................................................27 
2.2  Introduction.....................................................................................................27 
2.3  Experimental Section......................................................................................30 
 2.3.1  Materials..........................................................................................30 
 2.3.2  Apparatus.........................................................................................30 
 vii
 2.3.3  Experimental Procedure...................................................................32 
2.4  Results and Discussion...................................................................................33 
2.4.1  Effect of Ru/Al
2
O
3
 Catalyst............................................................33 
2.4.2  Effect of Residence Time................................................................37 
2.4.3  Effect of Temperature......................................................................38 
2.4.4  Effect of Glucose Concentration......................................................40 
2.4.5  Reaction Mechanism........................................................................41 
2.5  Conclusions.....................................................................................................44 
2.6  Acknowledgement..........................................................................................45 
2.7  References.......................................................................................................46 
3.  Hydrogen Production from Ethanol..............................................................................49 
3.1  Abstract...........................................................................................................49 
3.2  Introduction.....................................................................................................49 
3.3  Experimental Section......................................................................................53 
3.4  Results and Discussion...................................................................................56 
3.4.1  Effect of Ru/Al
2
O
3
 Catalyst.............................................................56 
3.4.2  Effect of Residence Time.................................................................62 
3.4.3  Effect of Pressure.............................................................................65 
3.4.4  Effect of Temperature......................................................................65 
3.4.5  Effect of the Ethanol Concentration................................................67 
3.4.6  Reaction Mechanism and Kinetics...................................................69 
3.5  Conclusions.....................................................................................................73 
3.6  Acknowledgment............................................................................................74 
3.7  References.......................................................................................................75 
4.  Hydrogen Production from Glycerol............................................................................78 
4.1  Abstract...........................................................................................................78 
4.2  Introduction.....................................................................................................78 
 viii
4.3  Experimental Section......................................................................................81 
 4.3.1  Materials..........................................................................................81 
 4.3.2  Experimental Procedure...................................................................82 
4.4  Results and Discussion...................................................................................84 
4.4.1  Effect of Reaction Time...................................................................85 
4.4.2  Effect of Temperature......................................................................86 
4.4.3  Effect of Glycerol Feed Concentration............................................87 
4.4.4  Reaction Kinetics.............................................................................88 
4.5  Conclusions.....................................................................................................92 
4.6  Acknowledgment............................................................................................92 
4.7  References.......................................................................................................93 
5.  Hydrogen Production from Catalytic Gasification of Switchgrass Biocrude...............95 
5.1  Abstract...........................................................................................................95 
5.2  Introduction.....................................................................................................95 
5.3  Experimental...................................................................................................98 
5.3.1 Catalyst Preparation..........................................................................98 
5.3.2 Feedstock Preparation.......................................................................99 
5.3.3 Experimental Procedure..................................................................100 
5.4  Results...........................................................................................................102 
5.4.1  Crystalline structure.......................................................................105 
5.4.2  Surface area....................................................................................107 
5.4.3  Particle size and morphology.........................................................109 
5.4.4  Thermo-gravimetric Analysis........................................................111 
 5.5  Discussion.....................................................................................................112 
5.6  Conclusions...................................................................................................114 
5.7  References.....................................................................................................115 
 
 ix
6.  Stability of Cerium-modified ??-Alumina Catalyst Support in Supercritical Water....118 
6.1  Abstract.........................................................................................................118     
6.2.  Introduction..................................................................................................118 
6.3.  Experimental................................................................................................123 
6.3.1  Sample Preparation........................................................................123 
6.3.2 Experimental Procedure..................................................................124 
6.4.  Results and Discussion................................................................................126 
6.4.1  XRD...............................................................................................126 
6.4.2  BET Surface Area..........................................................................133 
6.4.3  ICP-AES........................................................................................134 
6.5  Conclusions...................................................................................................135 
6.6  References.....................................................................................................136 
7.  Stability of Binary Aluminum, Titanium, and Zirconium Oxides for Catalyst  
     Supports in Sub- and Supercritical Water...................................................................139 
 
 7.1  Abstract.........................................................................................................139 
 7.2  Introduction...................................................................................................139 
 7.3  Experimental.................................................................................................141 
  7.3.1  Sample Preparation........................................................................141 
  7.3.2  Experimental Procedure.................................................................142 
  7.3.3  Characterization.............................................................................144 
 7.4  Results and Discussion.................................................................................144 
  7.4.1  Crystalline structure.......................................................................144 
  7.4.2  Surface area and pore characterization ..........................................147 
 7.5  Conclusions...................................................................................................149 
 7.6  References.....................................................................................................150 
8.  Overall Conclusions....................................................................................................152 
9.  Directions for Future Work ........................................................................................156 
 x
Appendix 1:  Determination of Gas Composition...........................................................159 
Appendix 2:  Calculation of Gas Flow using Viscosity Correction.................................161 
Appendix 3:  List of Publications and Presentations.......................................................162 
 
 xi
 
 
 
 
 
List of Tables 
 
 
Table 1.1:  Properties of Hydrogen......................................................................................3 
Table 2.1:  Product Distribution during Thermal and Catalytic Reforming of Glucose in         
                 Supercritical Water (T: 700?C, P: 241 bar, 1.0 wt% glucose)..........................34 
Table 3.1:  Details of Experimental Conditions.................................................................58 
Table 3.2:  Values of rate constants at different temperatures...........................................73 
Table 4.1:  Details of experimental conditions..................................................................83 
Table 5.1:  Oxygenated Hydrocarbons Present in Biocrude............................................100 
Table 5.2:  Gasification Efficiencies and Carbon Balance Closure.................................103 
Table 5.3:  Surface Area, Pore Radii and Volume for Calcined and Used Catalysts......108 
Table 5.4:  Weight Loss (%) of Carbonaceous Material During TGA............................112 
Table 6.1:  BET Surface Area of Samples.......................................................................134 
Table 7.1:  Surface Area, Average Pore Size, and Pore Volume of Binary Oxides after  
                  Calcination and Subsequent Hydrothermal Treatment at 25 MPa.................148 
Table A1.1:  Composition of Calibration Gas in mol%...................................................160 
 
 xii
 
 
 
 
 
List of Figures 
 
 
Figure 1.1:  Phase diagram for water ..................................................................................4 
Figure 1.2:  Physical properties of water at 24 MPa............................................................5 
Figure  2.1:  Schematic diagram of the experimental apparatus........................................32 
Figure  2.2:  Effect of residence time on product gas yields during non-catalytic  
                     reforming of glucose (T: 700?C, P: 241 bar, glucose conc.: 1 wt.%)...........35 
 
Figure  2.3:  Effect of residence time on product gas yields: (T: 700?C, P: 241 bar,  
                    glucose conc.: 1 wt.%, 2.0 g Ru/Al
2
O
3
 catalyst)...........................................36 
 
Figure  2.4:  Effect of residence time on product gas yields. (T: 700?C, P: 241 bar,  
                     glucose conc. 5 wt. %, 2.0 g Ru/Al
2
O
3
 catalyst)...........................................37 
 
Figure 2.5:  Effect of residence time on total organic carbon in liquid phase  
                   (T=700?C, P: 241 bar, residence time: 2 s).....................................................38 
 
Figure  2.6:  Effect of temperature on product yields (P: 241 bar, glucose conc.:  
                     4 wt.%, residence time: 2 s, 2.0 g Ru/Al
2
O
3
 catalyst)................................. 40 
 
Figure  2.7:  Effect of glucose concentration on product yields (T: 700?C, P: 241 bar,  
                    residence time: 2 s: 2.0 g Ru/Al
2
O
3
 catalyst).................................................41 
Figure  3.1:  Schematic diagram of experimental apparatus..............................................54 
Figure  3.2:  SEM images of fresh Ru/Al
2
O
3
 catalyst. (a) 5500x magnification (b)  
                    33,000x magnification...................................................................................60 
Figure  3.3:  SEM images of used Ru/Al
2
O
3
 catalyst. (a) 600x magnification  
                     (b) 30,000x magnification.............................................................................61 
 
Figure  3.4:  X-ray diffraction spectra for fresh and spent Ru/Al
2
O
3
 catalyst...................62 
Figure  3.5:  Effect of residence time on product gas yields (T, 600 ?C; P, 221 bar; feed  
                    conc., 10 wt% EtOH; 1.922g Ru/Al
2
O
3 
W/F
Ao
, 5 s:14,200g cat*s/g-mol   
                    EtOH, 10 s: 25,200 cat*s/g-mol EtOH).........................................................63 
 xiii
 
Figure  3.6:  Effect of residence time on product yields (T, 700?C; P, 221 bar; feed  
                    conc., 10 wt% EtOH; 1.922g Ru/Al
2
O
3
).......................................................64 
 
Figure  3.7:  Effect of residence time on product yields (T,  800?C; P, 221 bar; feed  
                    conc., 10 wt% EtOH; 1.922g Ru/Al
2
O
3
).......................................................64 
 
Figure  3.8:  Effect of reactor pressure (T, 700?C; feed conc., 2 s in Ru/Al
2
O
3 
bed;  
                     10wt% EtOH; 1.922 g Ru/Al
2
O
3
).................................................................65 
Figure  3.9:  Effect of reactor temperature (P, 221 bar; 2 s residence time in Ru/Al
2
O
3  
                               
bed; feed conc., 10wt% EtOH; 1.922 g Ru/Al
2
O
3
).......................................66 
Figure  3.10:  Effect of feed concentration of product gas yields (T, 800?C; P, 221 bar;  
                       residence time 4 s in Ru/Al
2
O
3 
bed, feed conc., 10 wt% EtOH; 1.922 g  
                       Ru/Al
2
O
3
)....................................................................................................68 
Figure  4.1:  Schematic of SCW hydrogen production apparatus......................................83 
Figure  4.2:  Effect of residence time on product gas yields. T: 800?C, P: 241 bar, feed  
                     conc.:5 wt% glycerol, 2.0 g Ru/Al
2
O
3 
catalyst.............................................86 
Figure  4.3:  Effect of temperature on product gas yields. P: 241 bar, 1 s residence time,  
                     feed conc.: 5 wt%, 2.0 g Ru/Al
2
O
3
...............................................................87 
Figure  4.4:  Effect of glycerol feed concentration on product gas yields.  T: 800?C, P:  
                     241 bar, 1 s residence time, 2.0 g Ru/Al
2
O
3
 catalyst....................................88 
Figure  4.5:  SEM images of fresh (a) and used (b) Ru/Al
2
O
3
 catalysts............................91 
Figure 5.1:  Schematic of experimental apparatus...........................................................101 
Figure 5.2:  Product gas yields for TiO
2
 and ZrO
2
 supported catalysts...........................104 
Figure 5.3:  XRD spectra of ZrO
2
 supported catalysts before and after SCW exposure.    
                   (a) fresh and (b) used Ni/ZrO
2
; (c) fresh and (d) used Ru/ZrO
2
; (e) fresh  
                   and (f) used Co/ZrO
2
.....................................................................................105 
 
Figure 5.4:  XRD spectra of TiO
2
 supported catalysts before and after SCW exposure.    
                   (a) fresh and (b) used Ni/TiO
2
; (c) fresh and (d) used Ru/TiO
2
; (e) fresh  
                   and (f) used Co/TiO
2
.....................................................................................106 
 
Figure 5.5:  XRD spectra of MgAl
2
O
4
 supported catalysts before and after SCW  
                   exposure.  (a) fresh and (b) used Ni/ MgAl
2
O
4
; (c) fresh and (d) used Ru/  
                   MgAl
2
O
4
; (e) fresh and (f) used Co/ MgAl
2
O
4
.............................................107 
 xiv
Figure 5.6:  SEM of used Ni/ZrO
2
 catalyst taken from the reactor entrance...................109 
Figure 5.7:  SEM of used Ni/ZrO
2
 catalyst taken from the reactor exit..........................110 
Figure 5.8:  SEM of used Ni/TiO
2
 catalyst taken from the reactor entrance...................110 
Figure 5.9:  SEM of used Ni/TiO
2
 catalyst taken from the reactor exit...........................111 
Figure 6.1:  Dehydration sequence of selected alumina hydrates in air at ambient                 
                   pressure.  Path a favored by d
p
 < 10 ?m; Path b favored by alkalinity and  
                   d
p
 > 100 ?m. Open areas indicate transitional zones. Adapted from  
                   reference [3]..................................................................................................120 
 
Figure 6.2:  Model of alumina surface dehydroxylation resulting in loss of surface  
                   area.  Adapted from reference [4]................................................................ 121 
 
Figure 6.3:  Experimental Apparatus...............................................................................125 
Figure 6.4:  XRD spectra of (a) ?-Al
2
O
3
 as received; freshly calcined (b) 1CeAl,  
                   (c) 2CeAl, (d) 4CeAl, (e) 6CeAl, (f) 10CeAl.????corresponds to CeO
2
........127 
 
Figure 6.5:  XRD spectra of unmodified Al
2
O
3
 after 3h exposure to supercritical water  
                    at (a) 500?C, (b) 600?C, and (c) 700?C.  Labels correspond to (?) ?-Al
2
O
3
,  
                   (?) ?-Al
2
O
3
, and (?) Boehmite.....................................................................128 
?
Figure 6.6:  XRD spectra of CeAl samples after exposure to SCW at 500?C (a) 2CeAl  
                    (b) 6CeAl (c) 10CeAl. Labels correspond to (?) ? -Al
2
O
3
, (?)? -Al
2
O
3
,  
                    (?) Boehmite, (?) CeO
2
................................................................................129 
 
Figure 6.7:  XRD spectra of CeAl samples after exposure to SCW at 600?C (a) 1CeAl  
                   (b) 2CeAl (c) 4CeAl (d) 6CeAl (e) 10CeAl. Labels correspond to  
               (?)? -Al
2
O
3
, (?)? -Al
2
O
3
, (?) CeO
2
.............................................................130 
 
Figure 6.8:  XRD spectra of CeAl samples after exposure to SCW at 700?C (a) 2CeAl  
                   (b) 4CeAl (c) 6CeAl. Labels correspond to (?)? -Al
2
O
3
, (?) ?-Al
2
O
3
,  
                   (?)  CeO
2
.......................................................................................................131 
 
Figure 6.9:  XRD spectra of 6CeAl samples after exposure to SCW at 600?C.   
                   (a) Reduced at 950?C, SCW exposure 3h (b) Reduced at 950?C, SCW  
                   exposure  6 h (c) Reduced at 750?C, SCW exposure 3h (d) Reduced at  
               750?C, SCW  exposure 6 h...........................................................................133 
 
Figure 7.1:  Experimental apparatus for hydrothermal treatment of binary oxides.........142 
 
 
 
 xv
 xvi
Figure 7.2:  XRD spectra of ZrTi (a) after calcination, and after hydrothermal  
                   treatment at 25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.   
                   all peaks correspond to ZrTiO
4
.....................................................................145 
 
Figure 7.3:  XRD spectra of TiAl (a) after calcination, and after hydrothermal treatment  
                    at 25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.  Labels   
                    correspond to?(?)?AlO(OH), (?) ?-Al
2
O
3
, ({) anatase TiO
2
, (z) rutile   
               TiO
2
..............................................................................................................146 
 
Figure 7.4:  XRD spectra of ZrAl (a) after calcination, and after hydrothermal treatment  
                    at 25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.  Labels   
                    correspond to (?) t-ZrO
2
, (?)m-ZrO
2
, (?)AlO(OH), (?) ?-Al
2
O
3
..........147 
 
 
 
 
 
 
1.  Introduction 
The total energy demand of the world is growing every year with increasing 
population and demands for higher standards of living.   Currently most of the world?s 
energy demands are met by using non-renewable fossil fuels, be it coal, liquid 
hydrocarbons, or natural gas.  In 2007 the nation?s annual energy consumption was 
approximately 100 quadrillion BTU.  Of that, 39% was petroleum, 23% natural gas, and 
22% coal.  Nuclear energy accounted for 8%, while all renewable energy technologies 
contributed only 7%
1
  Hydrocarbon fuels are favored for their high energy density and 
ease of transportation, but combustion of these fuels contributes as a major source of air 
pollution and greenhouse gases.  Moreover, the supply of fossil fuels is finite and 
exhaustible.  Hydrogen has been considered as an environmentally friendly energy carrier 
for the future, although current production technologies rely on fossil resources as a 
feedstock.  Hydrogen production from fossil fuels may represent a more efficient 
utilization of our reserves but still entails the use of a non-renewable resource, and in the 
absence of sequestration, the release of CO
2
 into the atmosphere.  It is also possible to 
produce hydrogen from renewable resources such as biomass and biomass derived 
materials with zero net emissions of carbon dioxide.  This refers to the fact that CO
2
 
produced during the processing of biomass materials will be fixed by plants during 
photosynthesis, thus creating a closed carbon loop.  The properties of water above its 
critical point make it an attractive reaction medium that provides good heat and mass 
transfer, as well as small reactor volumes.  Supercritical water?s non-polar nature allows 
it to solubilize organic compounds, thus reducing tar or coke formation.  Another benefit 
conferred is that hydrogen is produced directly at high pressure ready for storage.  For 
 1
these reasons this research explores hydrogen production from several biomass-derived 
materials in supercritical water. 
1.1  Hydrogen Properties 
Hydrogen is the simplest and smallest atom, with the most common isotope 
consisting of only one proton and one electron.  It is the most abundant element in the 
universe, but owing to its high reactivity it is not found in its molecular form H
2
.  Most of 
the hydrogen on Earth exists either in its oxidized form as water with no fuel value or 
bound to carbon atoms. For this reason hydrogen may be considered as an energy carrier, 
but not an energy source.  In order to obtain molecular hydrogen it must be produced 
from some other hydrogen-containing compound. 
Hydrogen gas is colorless, odorless, tasteless, and nontoxic.  It diffuses rapidly 
through air with a diffusion coefficient of 0.61cm
2
 s
-1
; 
however it has a wide flammability 
range in air of 4-75 vol%, and its ignition energy of 0.02 mJ is an order of magnitude 
lower than hydrocarbon fuels in air.  Liquid hydrogen has the highest energy density of 
any fuel on mass basis, but has a relatively low energy density on a volume basis.  For 
example, the LHV of liquid hydrogen (140.4 MJ kg
-1
) is almost three times that of 
gasoline (48.6 MJ kg
-1
) on a mass basis, which is reversed when considering energy 
density on a volume basis (8,491 MJ m
-3
 vs. 31,150 MJ m
-3
).
2
  Additional physical 
properties of hydrogen are presented in Table 1. 
   
 
 
 
 2
Table 1.1.  Properties of Hydrogen
3
 
Property       Value Unit 
Molecular Weight   2.01594  
Density   
 gas at 0?C, 1 atm  0.08987 kg m
-3
 
 liquid at -253?C 708 kg m
-3
 
 solid at -259?C  858 kg m
-3
 
Melting Temperature  -259 ?C 
Normal Boiling Point   -253 ?C 
Heat of Fusion at -259?C  58 kJ kg
-1
 
Heat of Vaporization at -253?C  447 kJ kg
-1
 
Heat Capacity (Cp)    
 gas at 25?C  14.3 kJ kg
-1
?C
-1
 
 liquid at -256?C 8.1 kJ kg
-1
?C
-1
 
 solid at -259.8?C  2.63 kJ kg
-1
?C
-1
 
Critical Temperature  -240 ?C 
Critical Pressure  12.8 atm 
Critical Density  31.2 kg m
-3
 
Viscosity at 25?C, 1 atm  0.00892 cP 
Thermal Conductivity 190 W m
-1
?C
-1
 
 
 
1.2  Supercritical Water Background and Properties 
 Liquid water at ambient conditions (25?C, 1 atm) is an excellent solvent for many 
polar compounds and inorganic salts due to the polar nature of water.  Under elevated 
temperature and pressure, however, the nature of water as a solvent changes 
considerably.  Above the critical point of water (T
c
 = 373.946?C, P
c
 = 22.046 MPa) the 
fluid exists in the supercritical phase, which is distinct from the liquid and gas phases.  
The supercritical region is illustrated in the phase diagram below.   
 
 3
Pressure  
 
 
 
Supercritical 
Region 
 
Temperature  
Solid 
Liquid 
Gas 
Critical Point
T
c
= 374 
o
C 
P
c
= 221 bar 
? 
c
= 0.375 g cm
-3
 
 
T
 
Figure. 1.1:  Phase diagram for water 
 
 Above the critical point, the dielectric constant of water drops sharply from about 
80 under ambient conditions to around 10, allowing it to behave like a nonpolar solvent.
4
  
This is associated with a decreased degree of hydrogen bonding as temperature is raised.  
At the critical point the degree of hydrogen bonding in water is half of what it is under 
ambient conditions.
5
  As the temperature of water increases from ambient to 200-300?C, 
its ion product K
w
 increases three orders of magnitude
6
, allowing it to participate in acid 
and base catalyzed chemistry;
7, 8
 however, beyond the critical point the ion product falls 
drastically, making it a nonionic solvent.
6
  The transport properties of supercritical water 
have some gas-like characteristics as well as some liquid-like characteristics.  The 
viscosity of supercritical water is an order of magnitude lower than that of liquid water, 
greatly enhancing mass transfer and diffusion controlled reactions.  Thermal conductivity 
 4
of supercritical water is also high, allowing for excellent heat transfer.
9
  Several physical 
properties of water are presented graphically in Fig. 1.2 at 24 MPa.
10
 
 
 
Figure 1.2:  Physical properties of water at 24 MPa
10
 
 
1.3  Hydrogen Production from Hydrocarbons 
The processing of hydrocarbons, especially methane, is the most popular 
production strategy today in the industrial production of hydrogen. 
11
  Collectively called 
oxyforming, the three most prominent technologies in hydrocarbon reforming are steam 
reforming (SR), partial oxidation (POX), and autothermal reforming (ATR), each of 
which has advantages and disadvantages.
12
  SR is most commonly used in hydrogen 
production as it gives the highest hydrogen yields and uses the lowest processing 
temperature.  However, the process is highly endothermic and requires an external heat 
source, and also has the highest air emissions.
13
  As its name implies, POX partially 
 5
oxidizes a portion of the feedstock to give a mixture of carbon monoxide and hydrogen, 
and can be performed without a catalyst.  Accordingly the feedstock may not require 
desulfurization, but POX has several major disadvantages including low hydrogen yields, 
high operating temperatures, and the requirement of co-feeding oxygen.  Partial oxidation 
performed with a catalyst is known as CPOX.  ATR combines aspects of SR and POX, 
where a portion of the feed is combusted to provide heat necessary for endothermic steam 
reforming.
14
 ATR operates at lower temperatures than POX and gives H
2
/CO ratios good 
for Fischer-Tropsch synthesis, but has limited commercial experience and also requires 
co-feeding oxygen.
13
 
 Reactions pertinent to hydrocarbon reforming, water-gas shift, and oxidation may 
be summarized as follows for methane and with generalized hydrocarbon formulaes:
15-17
 
 
Steam reforming 
C
m
H
n
 + mH
2
O = mCO + (m+ ?n)H
2
  ?H =  endothermic, hydrocarbon dependent    (1.1) 
CH
4
 + H
2
O = 3H
2
 + CO  ?H = +205 kJ mol
-1                                                      
(1.2) 
Partial oxidation 
C
m
H
n
 + ?O
2
 = mCO + ?H
2
  ?H =  exothermic, hydrocarbon dependent         (1.3) 
CH
4
 + ?O
2
 = CO + 2H
2
  ?H= -35.6 kJ mol
-1
                                     (1.4) 
Autothermal reforming 
C
m
H
n
 + ?mH
2
O  + ?mO
2
 = mCO + (?m+ ?n)H
2
  ?H = thermoneutral           (1.5) 
Water-gas shift 
CO + H
2
O = CO
2
 + H
2
 ?H = - 41.1 kJ mol
-1                                                     
(1.6) 
 
 6
Dry reforming 
CH
4
 + CO
2
 = 2CO + 2H
2
  ?H = +247.3 kJ mol
-1                                            
(1.7) 
Coke formation 
C
m
H
n
 = xC + C
m-x
H
n-2x
 + xH
2
 ?H = hydrocarbon dependent                    (1.8) 
2CO = C + CO
2
  ?H = +172.4 kJ mol
-1                                                       
(1.9) 
CO + H
2
 = C + H
2
O  ?H = -130 kJ mol
-1                                                  
(1.10) 
CO oxidation 
CO + ? O
2
 = CO
2
  ?H = -283 kJ mol
-1                                                         
(1.11) 
H
2
 oxidation 
H
2
 + ?O
2
 = H
2
O  ?H = -242 kJ mol
-1                                                       
(1.12) 
 
A final version of methane reforming to be considered is the so-called dry 
reforming of methane, where methane is fed to the reactor with CO
2
 instead of steam.  
Dry reforming uses two greenhouse gas feedstocks and can convert them to useful fuels, 
but is not well suited to hydrogen production as it is characterized by a low H
2
:CO of 
1:1.
17
  Dry reforming methane has been studied the least out of all of the other schemes 
presented above.   
Detailed kinetic and isotopic studies by Wei and Iglesia have shown that there is a 
common sequence of elementary steps in both the steam reforming and dry reforming of 
methane.
18
  Their studies showed that in the absence of transport or thermodynamic 
artifacts that the only kinetically relevant step in methane reforming was activation of the 
C-H bond.  Similar activation energies were reported for the CH
4
/CO
2
 and CH
4
/H
2
O 
systems, suggesting that the reactions are mechanistically equivalent. 
 7
1.4  Hydrogen Production from Biomass and Model Compounds 
1.4.1  Aqueous Phase Reforming 
Dumesic et al. have extensively studied the aqueous phase reforming of several 
oxygenated hydrocarbons including sorbitol, ethylene glycol, and glycerol in the 
production of hydrogen as well as alkanes.
19-21
  In this work sufficient pressure is 
maintained to ensure a homogenous liquid phase at around 500K over supported metal 
catalysts.  Their work with Pt/Al
2
O
3
 and Raney Ni/Sn catalysts highlights the importance 
of a catalyst active in C-C bond scission in hydrogen production, for a catalyst active in 
C-O cleavage leads to consumption of hydrogen through alkane production. 
1.4.2  Gasification 
An established process for conversion of coal or biomass to fuel gases is direct 
gasification.  The combustion of biomass has served man as an energy source since time 
immemorial, although gasification represents a more efficient utilization of this resource.  
Gasification of coal was practiced as early as the dawn of the 18
th
 century to provide gas 
to light streetlamps and later homes.  Gasification involves the reaction at high 
temperatures (1200-1400 K), moderate pressures (5-10 bar), a source of carbon, 
associated or not with hydrogen, with a source of hydrogen, usually steam, and/or oxygen 
to yield a gas product that contains CO, H
2
, CO
2
, CH
4
, and N
2 
in various proportions, as 
well as tars.
15
 Generally, it involves several reactions including cracking, partial 
oxidation, steam gasification, water gas shift, and methanation.   
One novel modern coal gasification technology is the so-called hydrogen 
production by the integrated novel gasification (HyPr-RING) method.
22-24
  This method 
 8
integrates hydrocarbon reactions, water-gas shift, and CO
2
 absorption into one single 
reactor, where CO
2
 is captured by the following reaction: 
CaO + CO
2
 ? CaCO
3
                                                      (1.13) 
CaCO
3
 is calcined in a separate regenerator to give CaO and a stream of CO
2
.  
Heteroatoms in the coal such as chlorine and sulfur are also captured by sorbents in the 
reactor, resulting in a clean fuel containing ~90% hydrogen and ~10% methane with a 
cold gas efficiency of 77%. 
Gasification of biomass has been demonstrated on a medium to large scale in 
several varieties of fluidized beds.
25
  Gasification performed at temperatures exceeding 
1000?C gives mainly syngas, and below 1000?C yields more hydrocarbons.
26
  The 
process may be air blown or oxygen blown.  The additional costs associated providing 
and using pure oxygen are compensated by providing a better quality fuel.
25
  Gasification 
requires energy intensive pretreatment of biomass including drying and particle size 
reduction. 
1.4.3  Supercritical Water Gasification 
 Several of the disadvantages of gasification may be obviated if the gasification is 
performed in supercritical water.  Reduced temperatures in the range of 400?-800?C may 
be employed, lowering energy requirements.  It is possible to utilize high moisture 
content biomass without additional drying.  The product gases are obtained at high 
pressure, and very little tar and char is formed.  
 Some of the earliest work done in supercritical water gasification was performed 
by Modell et al.  His work in gasification used residence times of 30 minutes near the 
critical point of water to form methane or syngas from biomass or coal slurries.
27, 28
   
 9
 The research group of Antal has published extensively on hydrogen production in 
supercritical water from several feedstocks.  Xu et al. investigated the activity of coconut 
shell activated carbon to catalyze the gasification of a wide variety of compounds, 
including glucose, glycerol, bagasse, and phenols.  They found that a wide range of 
activated carbons were able to catalyst the gasification of the compounds at 600?C, 34.5 
MPa, and WHSV of 0.1 to 0.5 h
-1
. 
 They reported that the surface area available on the 
catalyst did not have a large effect, which is likely the reason they were able to maintain 
high catalyst activity in light of the catalyst itself being slowly gasified in supercritical 
water.
29, 30
  Further investigation by Xu and Antal used 5% corn starch mixed with 11% 
sawdust or 2% sewage sludge in water to form a viscous paste which was fed in a 
continuous manner to be gasified over activated carbon.  A hydrogen-rich gas with 
significant methane was produced with no tar formation, however the reactors were seen 
to plug after several hours, especially with the high ash content sewage sludge.
31
  Later 
studies with potato wastes and corn starch showed similar behavior, with high hydrogen 
yields being favored at high reaction temperatures.
32
  This work also recognized the 
catalytic properties of the Hastelloy reactor wall, which was corroded during the reaction 
and deemed inappropriate for gasification in supercritical water.  In another work on the 
steam reforming of glucose in supercritical water,
33
 Antal et al. compared catalytic effects 
of the reactor wall by using Hastelloy and Inconel alloys.  Dilute 0.1 M glucose solutions 
were completely gasified at 600?C, 34.5 MPa, 30 s residence time, obtaining a hydrogen 
yield of ~75% of the theoretical maximum with the Inconel reactor.  Carbon dioxide was 
the other major product gas, with small amounts of methane and CO as well.  With 
increasing feed concentrations up to 0.8 M, the hydrogen yield steadily declined to less 
 10
than 10% of the theoretical maximum and the carbon balance decreased to 68%.  The 
Hastelloy reactor initially gave low hydrogen yields (~30% of maximum) and large 
amounts of CO, however as the wall was corroded its catalytic activity increased to be 
similar to Inconel, with H
2
 yields increasing while CO decreased.  From this they 
concluded that Inconel was active in the water-gas shift, whereas fresh Hastelloy does not 
catalyze the water-gas shift until sufficiently corroded. 
 Gasification of cellulose
34
 and lignin
35
 in supercritical water in the presence of 
ZrO
2
  and NaOH catalysts has been studied by Watanabe et al.  Hydrogen yields were 
doubled in the presence of zirconia and quadrupled in the presence of 1M NaOH.  The 
fresh zirconia catalyst was a mixture of tetragonal and monoclinic crystal systems which 
was transformed almost exclusively to the more stable monoclinic form after exposure to 
the high temperature/pressure hydrothermal environment. 
 Char formation during biomass gasification can be problematic as the reactive 
intermediates from lignin decomposition undergo cross-linking to reform high molecular 
weight products.
36
  Osada et al. have investigated gasification of lignin and models of its 
hydrolysis product alkylphenol in supercritical water over several supported noble metal 
catalysts.
37-39
  Pipe bomb-type reactors at 400?C, water densities ranging form 0.1-0.3 g 
cm
-3
, and reaction times of 15 - 60 minutes revealed an activity series of  Ru/ ?-Al
2
O
3
 > 
Ru/C, Rh/C > Pt/ ?-Al
2
O
3
 > Pd/C > Pd/?-Al
2
O
3
.  Higher water density was seen to 
promote the gasification, however the most active catalyst only gave a gas yield of 10-
15% in 15 minutes at the temperatures studied.
37
  Further investigations on gasification of 
actual lignin showed that Ru/TiO
2
 had a higher activity than Ru/ ?-Al
2
O
3
 in lignin 
 11
gasification.  Lignin decomposition was enhanced by higher water density (i.e. higher 
pressure).
38
   
The stability of ruthenium catalysts on various supports was also investigated in a 
separate study.
39
  Anatase-type TiO
2
 was stable at 400?C, 35 MPa for 3h, and maintained 
its activity for multiple uses.  This is interesting to compare with earlier results from 
Elliott et al. who found anatase unstable in subcritical water (350?C, 25 MPa), 
transforming to the rutile phase.
40
  Recent results have shown that a sub-critical water 
treatment is effective in regenerating Ru/TiO
2
 poisoned by sulfur.
41
  Ru/C showed high 
activity which gradually decreased with continued use, owing to gasification of the 
support itself as well as possible blocking of pores due to coking.  After the first use, the 
activity of the Ru/ ?-Al
2
O
3
 catalyst was greatly decreased owing to two phenomena.  The 
support underwent a phase change to the alpha phase sintering and a decrease in surface 
area.  During the phase change ruthenium metals were leached and detected in the 
aqueous phase. 
Elliott et al. have published on a mobile unit to gasify wet biomass such as dairy 
waste and distiller?s dried grains over a Ru/C catalyst in near-critical water (350?C, 20 
MPa) to a gas consisting mostly of methane and carbon dioxide.
42
  Their scheme used a 
CSTR as preheater which also liquified a slurry which was then fed to a fixed bed 
catalytic reactor.  In scale-up, reactor plugging was a problem, as was catalyst 
deactivation from components of the biomass such as calcium, magnesium, phosphorous, 
and sulfur. 
 
 
 12
1.5  Water Electrolysis 
 The ability of electricity to decompose water into its hydrogen and oxygen gases 
has been known for over 200 years, and the first large-scale electrolysis unit capable of 
producing 10,000 m
3
 h
-1
 of hydrogen went into operation in 1939.
43
  This process can 
produce clean hydrogen without contaminants such as CO or H
2
S, however water does 
not dissociate easily; the enthalpy of the reaction H
2
O ? ?O
2
 + H
2
 is +285.6 kJ mol
-1
.   
Alkaline water electrolysis is the most developed commercial technology, however 
proton exchange and solid oxide electrolysers are expected to have higher efficiencies in 
the future.
44
  In the long term, electrolysers can be used in conjunction with renewable 
energy sources such as wind or solar power to store surplus energy and alleviate 
problems associated with the intermittent nature of power produced from these 
technologies.
44
 
 
1.6  Other Relevant Previous SCW work 
1.6.1  Supercritical Water Oxidation (SCWO) 
 The solubilities of organic compounds and oxygen in supercritical water enable a 
homogenous phase in which the complete oxidation of organic species to carbon dioxide 
can take place in short reaction times.  This is known as supercritical water oxidation 
(SCWO) and is one of the most studied research fields dealing with supercritical water. 
45
  
Oxygen can be supplied by thermal decomposition of hydrogen peroxide or by dissolving 
oxygen into SCW; however, the same results are achieved with both methods .
46
 
Some of the first work in SCWO focusing on the disposal of hazardous organic 
waste streams was performed by Modell et al. in the early 1980s,
47, 48
, and was 
 13
commercialized in 1994.
49
  The destruction of hazardous wastes such as energetic 
compounds and chemical warfare agents like mustard gas and VX nerve gas was 
explored by US Department of Defense contractors General Atomics, who reported 
destruction of those compounds in excess of 99.9999%.  
The kinetics of SCWO of phenol and substituted phenolic compounds as model 
industrial pollutants has been studied extensively by several groups.  Destruction of 
phenol is first order in phenol and sensitive to oxygen concentration.
50-53
  Reaction 
networks and rate laws have been proposed for substituted phenols.  The reactivity series 
of substituted phenols was found to be ortho > para > meta, while substituted phenols 
were found to have greater reactivity than unsubstituted phenols, with methoxy 
substitutions having greater reactivity in SCW than methyl substitutions.
30
  It has also 
been reported that chlorinated phenols are much less reactive than other phenolic 
compounds.
54
 
 The employment of a catalyst in SCWO is often desirable for the same reasons 
that it is desirable in other systems:  to lower reaction temperatures and increase reaction 
rates, thus reducing energy requirements and reducing required reactor volumes, 
ultimately improving process economics.  Limited work has been done in heterogeneous 
catalysis for SCWO via transition metal salts, alkali, or polyheteroacids.  These offer only 
modest rate increases while requiring additional processing steps to recover the catalyst.  
Accordingly, heterogeneous catalysis is often sought in SCWO where a catalyst is 
required.
55
   
 A number of heterogeneous catalysts have been explored for SCWO of phenols 
and other hazardous organic compounds.  Jin et al.
56
 found that use of a V
2
O
5
 catalyst led 
 14
to the formation of large amounts of carbonaceous char in the destruction of 1,4-
dichlorobenzene.  Ding et al. reported no char in the destruction of phenol over V
2
O
5
 
catalyst, and both V
2
O
5
 and MnO
2
/CeO gave reduced incomplete oxidation byproducts 
compared with homogeneous oxidation.  The MnO
2
/CeO catalyst was the most stable, 
and interestingly showed less byproducts at higher feed concentrations.
57
  Krajnc and 
Levec used a CuO/ZnO catalyst in the SCWO of phenol, 1-methyl-2 pyrrolidone, benzoic 
acid and several other compounds and also found reduced concentrations of byproducts 
increased reaction rates compared to the noncatalytic homogenous oxidation.
58
 
 Commercial MnO
2
-CuO/Al
2
O
3
 VOC oxidation catalysts were used by Zhang and 
Savage
59
 to completely oxidize phenol at 390?C and 1 s space time, while Armbruster et 
al.
60
 showed that catalyst to be stable for 200 h in supercritical water.  Later the 
components of the commercial catalyst were tested individually.  The ability of MnO
2
 to 
catalyze the decomposition of phenol was evaluate by Yu and Savage
61, 62
 as well as by 
Oshima et al.
63
  The results of these studies showed significant rate increases, and that the 
same reaction pathway for phenol oxidation over MnO
2
 was the same as for 
homogeneous SCWO.  The role of the catalyst was to hasten the rate of production of 
reactive intermediates, including undesired dimers.  Yu and Savage also reported on 
SCWO of phenol over CuO/Al
2
O
3
.
62
  This catalyst reduced selectivity to dimers while 
increasing the selectivity towards CO
2
, however it was unstable, with Cu and Al showing 
up in the reactor effluent and the surface area decreasing from 200 to 10 m
2
 g
-1
. 
TiO
2
 was able to suppress the formation of dimers in phenol SCWO and maintained its 
activity for 125 h, during which time it underwent a phase transformation from anatase to 
rutile.
62
 
 15
 Activated carbon has been evaluated for use as catalyst in phenol decomposition 
with varying results in the literature.  Xu et al. screened phenol as part of a suite of 
Department of Defense wastes while studying gasification of various compounds and 
reported 80% conversion of phenol with benzene as byproduct at 600?C, 34.5 MPa, and 
WHSV of 0.1 h
-1
.  The authors consider gasification of the catalyst itself, however they 
determined that the rate of gasification of the catalyst was quite small under the reaction 
conditions and was stable over 4h.
30
  Matsumura et al. later published a study specifically 
focusing on SCWO of phenol at 400?C and 25 MPa.  In this work the activated carbon 
catalyst achieved similar conversions but gave a variety of undesirable dimer products in 
low yield.  In the oxidative environment, however, activated carbon was consumed 
entirely within three hours on stream.
64
 
 Elliott et al. have published extensively on chemical processing in the high 
pressure aqueous environment.
40, 42, 65-70
  Their research has been focused in the 
subcritical region at temperatures of about 350?C and 20 MPa pressure, with residence 
times of roughly 1-2 hours.  Early works screened a wide variety of catalytic materials 
and supports in the destruction of phenolic compounds.  Ru, Rh, and Ni were identified 
as active catalytic metals and ZrO
2
, ?-Al
2
O
3
 were identified as stable supports in the hot 
compressed water environment.  Transition aluminas were hydrolyzed to AlO(OH), while 
tableted TiO
2
 lost its physical integrity.
40
  Nickel was found to have less catalytic activity 
than ruthenium as well as having a higher propensity for sintering.
40, 67
  A recent 
publication from Elliott et al. examines catalytic activity and stability of additional 
materials in hot compressed water.
70
  Several metals were tested on titania supports, 
including iridium, silver, rhenium, tin, and lead.  Of these, iridium demonstrated a carbon 
 16
gasification of 1% in a 2 h reaction period, while the others gasified less than 0.1%.  
Ruthenium supported on titania showed high activity, with 55-94% of carbon being 
gasified in a 2 h reaction period.  The anatase form of titania was more active, but was 
seen to convert to the rutile form even under the relatively mild temperatures employed.  
Molybdenum sulfide catalysts were tested as a sulfur tolerant catalyst, but showed low 
activity, with several variants all having gasification efficiencies of less than 2%.  
Ni/ZrO
2
 prepared by the proprietary rapid thermal decomposition of solutes (RTDS) 
method gasified less than 4% of carbon, owing to nickel metal domains that were too 
large.  A commercial nickel steam reforming catalyst was modified by doping with 
ruthenium.  This gave a small increase in gasification of carbon (88 to 90%), however the 
doping improved the longevity of the catalyst by stabilizing nickel crystallite size and 
preventing sintering.   
 
1.6.2  Corrosivity in SCW Environment 
The high critical pressure of water of 221 bar along with high temperatures used 
in SCWO or SCW gasification make for severe processing conditions and may require 
exotic materials to cope with the harsh conditions.  Early corrosion testing by General 
Atomics in the SCWO process led to their recommending titanium or platinum as 
candidate reactor materials in the presence of halides.
71
  Foy et al. also report on SCWO 
of chlorinated organic compounds and reported acceptably low corrosion of a Ti-lined 
reactor.
72
  They also proposed other novel schemes for dealing with the corrosive 
environment.  One scheme was to neutralize the acid with sodium bicarbonate and 
operate at 650 bar where NaCl would remain as a separate fluid phase and not deposit on 
 17
reactor walls.  Yet another scheme was to employ a ceramic reactor encased in steel to 
provide structural integrity. 
Corrosion of 316 stainless steel has been studied at 400 bar and temperature 
ranging from 250-420?C.  No significant variation of corrosion was seen between 
degassed and oxygenated water, but corrosion rates increased 10x with the addition of a 
3% H
2
O
2
 mixture.  The increased corrosion rate was due to transformation of the 
protective chromium (III) oxide layer to soluble Cr (VI).
73
 
High nickel content alloys are known for their strength at elevated temperatures, 
however they also catalyze reactions such as the water-gas shift.
74
  The catalytic activity 
of the reactor wall is also affected by its oxidation state.  Antal has found Hastelloy to be 
inactive in promoting the water-gas shift until it has been somewhat corroded while 
Inconel was active immediately.
33
   
The behavior of Inconel 625 (62 wt% Ni, 22 wt% Cr, 9.0 wt% Mo, 3.5 wt% Fe) is 
similar in the presence of the strong acids HNO
3
, HCl and H
2
SO
4
 at concentrations of 0.2 
mol kg
-1
.
75
  At 24 MPa the protective chromium (III) oxide layer is stable below 270?C.  
Severe corrosion was observed between 270-380?C, however above 380?C a film of NiO 
forms a protective layer and the corrosion rate is greatly decreased, although the higher 
solubility of nickel nitrate leads to a corrosion rate 10x higher than seen in the presence 
of hydrochloric or sulfuric acids.  The stability of a variety of alloys has been evaluated 
during SCWO of various chlorinated compounds, and Hastelloy G-30 and the alloy HR-
160 were identified as having better corrosion resistance than several Inconel alloys.
76
  In 
the presence of phosphoric acid, however, corrosive behavior of Inconel 625 is much 
different.
77
  Below 400?C a metal phosphate film protects the reactor.  Above 400?C acid 
 18
concentrations of less than 0.1 mol kg
-1
 caused no corrosion, but concentrations greater 
than that resulted in severe corrosion.   The stability of a variety of alloys has been 
evaluated during SCWO of various chlorinated compounds, and Hastelloy G-30 and the 
alloy HR-160 were identified as having better corrosion resistance than several Inconel 
alloys.
76
   
 More relevant to this research is corrosive behavior in a reducing supercritical 
water environment, where the protective oxide film may behave differently.  Fujisawa et 
al. have investigated the corrosion behavior of 316 stainless steel and several Ni-base 
alloys in supercritical water at 25 MPa, 350-450?C, and H
2
 partial pressures up to 6.5 
MPa.
78, 79
  316SS corroded faster than any of the Ni-based alloys under all conditions, 
and corrosion rate increased with increasing hydrogen pressure.  Alloys C-276 (59% Ni, 
16% Cr, 16% Mo) and MAT21 (60% Ni, 19% Cr, 19% Mo) showed the best corrosion 
resistance in the reducing environment due to the high Mo content, which is thought to 
act as a stabilizer to the protective chromium oxide film.  In a reducing supercritical 
water environment containing 0.01 M NaOH, corrosion rates of 316SS were again higher 
than Ni-based alloys.  Ni-based alloys containing both Cr and Mo had better corrosion 
resistance than binary Ni-Cr alloys, and increasing Cr content reduced the susceptibility 
of stress cracking. 
 
 19
1.7  References 
1. http://www.eia.doe.gov/fuelrenewable.html EIA Renewable & Alternative Fuel 
Data, Reports, Analysis, Surveys. (July 15, 2009),  
 
2. Hydrogen Fuel Cell Engines and Related Technologies. College of the Desert: 
Palm Desert, CA, 2001. 
 
3. Fundamentals and Use of Hydrogen as a Fuel. In Kirk-Othmer Encyclopedia of 
Chemical Technology, 3rd ed.; Wiley: New York, 1992; Vol. 4, p 631. 
 
4. Fernandez, D. P. G., A. R. H.; Lemmon, E. W.; Sengers, J. M. H. Levelt; 
Williams, R. C., A formulation for the static permittivity of water and steam at 
temperatures from 238 K to 873 K at pressures up to 1200 MPa, including derivatives 
and Debye-Hueckel coefficients. Journal of Physical and Chemical Reference Data 
1997, 26, 1125-1166. 
 
5. Gorbaty, Y. E. G., R.B., The Structural Features of Liquid and Supercritical 
Water. Industrial and Engineering Chemistry Research 1998, 37, 3026-3035. 
 
6. Bandura, A. V. L., S.N., The Ionization Constant of Water over Wide Ranges of 
Temperature and Density. Journal of Physical and Chemical Reference Data 2006, 35, 
15-30. 
 
7. Watanabe, M. O., Mitsumasa; Inomata, Hiroshi; Arai, Kunio; Kruse, Andrea, 
Acidity and basicity of metal oxide catalysts for formaldehyde reaction in supercritical 
water at 673 K. Applied Catalysis A: General 2003, 245, 333-341. 
 
8. Ikushima, Y. H., Kiyotaka; Sato, Osamu; Yokoyama, Toshirou; Arai, Masahiko, 
Acceleration of Synthetic Organic Reactions Using Supercritical Water: Noncatalytic 
Beckmann and Pinacol Rearrangements. Journal of the American Chemical Society 2000, 
122, 1908-1918. 
 
9. Weingartner, H. F., E.U., Supercritical Water as a Solvent. Angewandte Chemie 
International Edition 2005, 44, 2672-2692. 
 
10. Kritzer, P. D., E., An assessment of supercritical water oxidation (SCWO) 
Existing problems, possible solutions and new reactor concepts. Chemical Engineering 
Journal 2001, 83, 207-214. 
 
11. Holladay, J. D. J., J.; King, D.L.; Wang, Y., An Overview of hydrogen production 
technologies. Catalysis Today 2009, 139, 244-260. 
 
12. York, A. P. E., Xiao, T.C.; Green, M.L.H.; Claridge, J.B., Methand Oxyforming 
for Synthesis Gas Production. Catalysis Reviews 2007, 49, 511-560. 
 20
13. Wilhelm, J. D. S., D.R.; Karp, A.D.; Dickenson, R.L., Syngas production for gas-
to-liquids application: technologies, issues and outlook. Fuel Processing Technology 
2001, 71, 139-148. 
 
14. Flytzani-Stephanopoulos, M. V., G.E., Autothermal reforming of aliphatic and 
aromatic hydrocarbon liquids. International Journal of Hydrogen Energy 1983, 8, (7), 
539-548. 
 
15. Navarro, R. M. P., M.A.; Fierro, J.L.G., Hydrogen Production Reactions from 
Carbon Feedstocks: Fossil Fuels and Biomass. Chemical Reviews 2007, 107, 3952-3991. 
 
16. Wender, I., Reactions of Synthesis Gas. Fuel Processing Technology 1996, 48, 
189-297. 
 
17. Verykios, X. E., Catalytic dry reforming of natural gas for the production of 
chemicals and hydrogen. International Journal of Hydrogen Energy 2003, 28, 1045-
1063. 
 
18. Wei, J. I., E., Reaction Pathways and Site Requirements for the Activation and 
Chemical Conversion of 
 
Methane on Ru-Based Catalysts. J. Phys. Chem. B 2004, 108, 7253-7262. 
19. Cortright, R. D. D., J.A. Low-temperature hydrogen production from oxygenated 
hydrocarbons 6,964,758, 2005. 
 
20. Huber, G. W. D., J.A., An overview of aqueous-phase catalytic processes for 
production of hydrogen and alkanes in a biorefinery. Catalysis Today 2006, 111, 119-
132. 
 
21. Shabaker, J. W. D., J.A., Kinetics of Aqueous-Phase Reforming of oxygenated 
Hydrocarbons: Pt/Al2O3 and Sn-Modified Ni Catalysts. Industrial and Engineering 
Chemistry Research 2004, 43, 3105-3112. 
 
22. Lin, S.-Y. S., Yoshizo; Hatano, Hiroyuki; Harada, Michaiki, Hydrogen 
Production from Hydrocarbon by Integration of Water-Carbon Reaction and Carbon 
Dioxide Removal (HyPr-RING Method). Energy and Fuels 2001, 15, 339-343. 
 
23. Lin, S.-Y. S., Yoshizo; Hatano, Hiroyuki; Harada, Michiaki, Developing an 
innovative method, HyPr-RING, to produce hydrogen from hydrocarbons. Energy 
Conversion and Management 2002, 43, 1283-1290. 
 
24. Lin, S.-Y. H., Michiaki; Suzuki, Yoshizo; Harada, Michaiki; Hatano, Hiroyuki, 
Process analysis for hydrogen production by reaction integrated novel gasification (HyPr-
RING). Energy Conversion and Management 2005, 46, 869-880. 
 
 21
25. Bridgwater, A. V., Renewable fuels and chemicals by thermal processing of 
biomass. Chemical Engineering Journal 2003, 91, 87-201. 
 
26. Klass, D. L., Biomass for Renewable Energy; Fuels and Chemicals. Academic 
Press: San Diego, CA, 1998. 
 
27. Modell, M. Reforming of glucose and wood at the critical conditions of water; 
Massachussetts Institute of Technology: Cambridge, MA, 1977; p 7. 
 
28. Modell, M. R., R.C.; Amin, S.I. Gasification Process. 4,113,446, 1978. 
 
29. Matsumura, Y. X., X.; Antal Jr., M.J., Gasification characteristics of an activated 
carbon in supercritical water. Carbon 1997, 35, 819-824. 
 
30. Xu, X. M., Y.; Stengerg, J.; Antal Jr., M.J., Carbon-Catalyzed Gasification of 
Organic Feedstocks in Supercritical Water. Industrial and Engineering Chemistry 
Research 1996, 35, 2522-2530. 
 
31. Xu, X. A. J., M.J., Gasification of Sewage Sludge and Other Biomass for 
Hydrogen Production in Supercritical Water. Environmental Progress 1998, 17, 215-220. 
 
32. Antal Jr., M. J. A., S.G.; Schulman, D.; Xu, X.; Divilio, R.J., Biomass 
Gasification in Supercritical Water. Industrial and Engineering Chemistry Research 
2000, 39, 4040-4053. 
 
33. Yu, D. A., M.; Antal Jr., M.J., Hydrogen Production by Steam Reforming 
Glucose in Supercritical Water. Energy and Fuels 1993, 7, 574-577. 
 
34. Watanabe, M. I., H.; Arai, K., Catalytic hydrogen generation from biomass 
(glucose and cellulose) with ZrO2 in supercritical water. Biomass and Bioenergy 2002, 
22, 405-410. 
 
35. Watanabe, M. I., H.; Osada, M.; Sato, T.; Adschiri, T.;  Arai, K., Catalytic effects 
of NaOH and ZrO2 for partial oxidative gasification of n-hexadecane and lignin in 
supercritical water. Fuel 2003, 82, 545-552. 
 
36. Saisu, M. S., T.; Watanabe, M.; Adschiri, T, Arai, K., Conversion of Lignin with 
Supercritical Water-Phenol Mixtures. Energy and Fuels 2003, 17, 922-928. 
 
37. Sato, T. O., M.; Watanabe, M.; Shirai, M.; Arai, K., Gasification of Alkylphenols 
with Supported Noble Metal Catalysts in Supercritical Water. Industrial and Engineering 
Chemistry Research 2003, 42, 4277-4282. 
 
38. Osada, M. S., O.; Watanabe, M.; Arai, K.; Shirai, M., Water Density Effect on 
Lignin Gasification over Supported Noble Metal Catalysts in Supercritical Water. Energy 
and Fuels 2006, 20, 930-953. 
 22
39. Osada, M. S., O.; Arai, K.; Shirai, M., Stability of Supported Ruthenium Catalysts 
for Lignin Gasification in Supercritical Water. Energy and Fuels 2006, 20, 2337-2343. 
 
40. Elliott, D. C. S. J., L.J.; Baker, E.G., Chemical Processing in High-pressure 
Aqueous Environments. 2. Development of Catalysts for Gasification. Industrial and 
Engineering Chemistry Research 1993, 32, 1542-1548. 
 
41. Osada, M. Y., A.; Hiyoshi, N.; Sato, O.; Shirai, M., Regeneration of catalyst 
poisoned by sulfur in subcritical water. Baiomasu Kagaku Kaigi Happyo Ronbunshu 
2009, 4, 124-125. 
 
42. Elliott, D. C. N., G.G.;  Hart, T.R.; Butner, R.S.; Zacher, A.H.; Englehard, M.H.; 
Young, J.S.; McCready, D.E., Chemical Processing in High-Pressure Aqueous 
Environments. 7. Process Development for Catalytic Gasification of Wet Biomass 
Feedstocks. Industrial and Engineering Chemistry Research 2004, 43, 1999-2004. 
 
43. Kreuter, W. H., H., Electrolysis: the important energy transformer in a world of 
sustainable energy. International Journal of Hydrogen Energy 1998, 23, 661-666. 
 
44. Mueller-Langer, F. T., E.; Kaltschmitt, M.; Peteves, S., Techno-economic 
assessment of hydrogen production processes for the hydrogen economy for the short and 
medium term International Journal of Hydrogen Energy 2007, 32, 3797-3810. 
 
45. Savage, P. E., Organic Chemical Reactions in Supercritical Water. Chemical 
Reviews 1999, 99, 603-622. 
 
46. Phenix, B. D. D., Joanna L.; Tester, Jefferson W.; Howard, Jack B.; Smith, 
Kenneth A., The Effects of Mixing and Oxidant Choice on Laboratory-Scale 
Measurements of Supercritical Water Oxidation Kinetics. Industrial and Engineering 
Chemistry Research 2002, 41, 624-631. 
 
47. Modell, M. Processing methods for the oxidation of organics in supercritical 
water 4,338,199, 1982. 
 
48. Thomason, T. B. M., M., Supercritical water destruction of aqueous wastes. 
Hazardous Waste 1984, 1, 453-467. 
 
49. McBrayer, R. In Design and Operation of the First Commercial Supercritical 
Water Oxidation Facility, First International Workshop on Supercritical Water Oxidation, 
Jacksonville, FL, 1995; Mohiuddin, J., Ed. Jacksonville, FL, 1995. 
 
50. Gopalan, S. S., P.E., A reaction network model for phenol oxidation in 
supercritical water. AIChE Journal 1995, 41, 1864-1873. 
 
51. Krajnc, M. L., J., On the kinetics of phenol oxidation in supercritical water. 
AIChE Journal 1996, 42, 1977-1984. 
 23
52. Koo, M. L., W.K.; Lee, C.H., New reactor system for supercritical water 
oxidation and its application on phenol destruction. Chemical Engineering Science 1997, 
52, 1201-1214. 
 
53. Oshima, Y. H., K.; Toda, M.; Chommanad, T.; Koda, S., Phenol oxidation 
kinetics in supercritical water. Journal of Supercritical Fluids 1998, 13, 241-246. 
 
54. Lin, K. S. W., H.P.; Li, M.C., Oxidation of 2,4-dichlorophenol in supercritical 
water. Chemosphere 1998, 36, 2075-2083. 
 
55. Savage, P. E. D., J.B.; Yu, J., Recent advances in catalytic oxidation in 
supercritical water. Combustion Science and Technology 2006, 178, 443-465. 
 
56. Jin, L. S., Y.T.; Martin, A, The effect of supercritical water on the catalytic 
oxidation of 1,4-dichlorobenzene. Journal of Supercritical Fluids 1990, 3, 233-239. 
 
57. Ding, Z.-Y. A., Sudhir N. V. K.; Abraham, Martin A, Catalytic Supercritical 
Water Oxidation: Phenol Conversion and Product Selectivity. Environmental Science and 
Technology 1995, 29, 2748-2753. 
 
58. Krajnc, M. L., J., Catalytic oxidation of toxic organics in supercritical water. 
Applied Catalysis B: Environmental 1994, 3, L101-L107. 
 
59. Zhang, X. S., P.E., Fast catalytic oxidation of phenol in supercritical water. 
Catalysis Today 1998, 40, 333-342. 
 
60. Armbruster, U. M., A.; Krepel, A., Hydrolysis and oxidative decomposition of 
ethyl acetate in sub- and super-critical water. Applied Catalysis B: Environmental 2001, 
31, 263-273. 
 
61. Yu, J. a. S., P.E., Catalytic oxidation of phenol over MnO2 in supercritical water. 
Industrial and Engineering Chemistry Research 1999, 38, 3793-3801. 
 
62. Yu, J. a. S., P.E., Catalyst activity, stability, and transformations during oxidation 
in supercritical water. Applied Catalysis B: Environmental 2001, 31, 123-132. 
 
63. Oshima, Y. T., K.; Koda, S., Kinetics of the catalytic oxidation of phenol over 
manganese oxide in supercritical water. Industrial and Engineering Chemistry Research 
1999, 38, 4183-4188. 
 
64. Matsumura, Y. U., T.; Yamamoto, K.; Nunoura, T., Carbon catalyzed 
supercritical water oxidation of phenol. Journal of Supercritical Fluids 2002, 22, 149-
156. 
 
 24
65. Sealock, J. L. E., D.C.; Baker, E.G.; Butner, R.S., Chemical Processing in High-
pressure Aqueous Environments. 1. Historical Perspective and Continuing 
Developments. Industrial and Engineering Chemistry Research 1993, 32, 1535-1541. 
 
66. Elliott, D. C. S. J., L.J.; Baker, E.G., Chemical Processing in High-pressure 
Aqueous Environments. 3. Batch Reactor Process Development Experiments for 
Organics Destruction. Industrial and Engineering Chemistry Research 1994, 33, 558-
565. 
 
67. Elliott, D. C. P., M.R.; Sealock Jr., L.J.; Baker, E.G., Chemical Processing in 
High-pressure Aqueous Environments. 4. Continuous-Flow Reactor Process 
Development Experiments for Organics Destruction. Industrial and Engineering 
Chemistry Research 1994, 33, 566-574. 
 
68. Sealock, J. L. E., D.C.; Baker, E.G.; Fassbender, A.G.; Silva, L.J., Chemical 
Processing in High-Pressure Aqueous Environments. 5. New Processing Concepts. 
Industrial and Engineering Chemistry Research 1996, 35, 4111-4118. 
 
69. Elliott, D. C. N., G.G.; Phelps, M.R.; Hart, T.R.; Zacher, A.H.; Silva, L.J., 
Chemical Processing in High-Pressure Aqueous Environments. 6. Demonstration of 
Catalytic Gasification for Chemical Manufacturing Wastewater Cleanup in Industrial 
Plants. Industrial and Engineering Chemistry Research 1999, 38, 879-883. 
 
70. Elliott, D. C. H., T.R.; Neuenschwander, G.G., Chemical Processing in High-
Pressure Aqueous Environments. 8. Improved Catalysts for Hydrothermal Gasification. 
Industrial and Engineering Chemistry Research 2006, 45, 3776-3781. 
 
71. Downey, K. W. S., R.H.; Hazlebeck D.A.; Roberts, A.J., Corrosion and chemical 
agent destruction. Research on supercritical water oxidation of hazardous military wastes. 
In ACS Symposium Series 608 - Innovations in Supercritical Fluids, Hutchenson, K. W. 
F., N.R., Ed. American Chemical Society: Washington, D.C., 1995; pp 313-326. 
 
72. Foy, B. R. W., Kurt; Sedillo, Michael A.; Buelow, Steven J., Hydrothermal 
Processing of Chlorinated Hydrocarbons in a Titanium Reactor. Environmental Science 
and Technology 1996, 30, 2790-2799. 
 
73. Hayward, T. M. S., T.M.; Makhija, R.C., Stainless steel flow reactor for 
supercritical water oxidation: corrosion tests. Journal of Supercritical Fluids 2003, 27, 
275-281. 
 
74. Bustamante, F. E., R. M.; Killmeyer, R. P.; Howard, B. H.; Rothenberger, K. S.; 
Cugini, A. V.; Morreale, B. D.; Ciocco, M. V., Uncatalyzed and wall-catalyzed forward 
water-gas shift reaction kinetics. AIChE Journal 2005, 51, 1440-1454. 
 
 25
75. Kritzer, P. B., N.; Dinjus, E., The corrosion of nickel?base alloy 625 in sub- and 
supercritical aqueous solutions of HNO3 in the presence of oxygen. Journal of Materials 
Science Letters 1999, 18, 771-773. 
 
76. Mitton, D. B. Y., J.-H.; Cline, J.A.; Kim, H.-S.; Eliaz, N. Latanision, R.M., 
Corrosion Behavior of Nickel-Based Alloys in Supercritical Water Oxidation Systems. 
Industrial and Engineering Chemistry Research 2000, 39, 4689-4696. 
 
77. Kritzer, P. B., N.; Dinjus, E., The corrosion of alloy 625 (NiCr22Mo9Nb; 2.4856) 
in high-temperature, high-pressure aqueous solutions of phosphoric acid and oxygen. 
Corrosion at sub- and supercritical temperatures. Materials and Corrosion 1998, 49, 831-
839. 
 
78. Fujisawa, R. S., M.; Kurata, Y.; Watanabe, Y., Corrosion behaviour of nickel base 
alloys and 316 stainless steel in supercritical water under alkaline conditions. Corrosion 
Engineering, Science, and Technology 2005, 40, 244-248. 
 
79. Fujisawa, R. N., K.; Nishida, T.; Sakaihara, M.; Kurata, Y.; Watanabe, Y., 
Corrosion Behavior of Nickel-Based Alloys and Type 316 Stainless Steel in Slightly 
Oxidizing or Reducing Supercritical Water. Corrosion 2006, 62, 270-274. 
 
 
 26
2.  Hydrogen Production from Glucose using Ru/Al
2
O
3
 Catalyst in Supercritical 
Water 
 
2.1  Abstract 
Glucose, as a model biomass compound, was catalytically reformed in 
supercritical water to produce hydrogen.  The reforming experiments were conducted in a 
continuous tubular reactor with and without Ru/Al
2
O
3
 catalyst, and with short residence 
time.  The addition of catalyst significantly enhanced the overall conversion and 
hydrogen yield, and reduced methane formation.  The gaseous products contained mainly 
hydrogen, carbon dioxide, methane and a small amount of carbon monoxide.  The effects 
of experimental conditions such as temperature, reaction time and weight percent of 
glucose in the feed water on formation of hydrogen product were investigated.  
Experimental hydrogen yield as high as 12 mol H
2
/ mole of glucose were obtained, which 
is the stoichiometric limit.  The gas yield was sensitive to temperature, residence time and 
feed concentration. High yield of H
2
 with low CO and CH
4
 yields were obtained at high 
reaction temperature and low glucose concentrations. Tar formation was observed at high 
glucose concentrations (> 5 wt. %). The catalytic conversion of glucose with ruthenium 
catalyst in supercritical water is an effective method for the hydrogen production directly 
at a high pressure, which can be extended to other biomass materials.  The reaction 
mechanism for catalytic reforming in supercritical water is also discussed. 
2.2  Introduction 
Due to the environmental impacts of the fossil fuel, hydrogen is now considered 
as an excellent replacement. Hydrogen produced from biomass has no net carbon dioxide 
 27
impact on the environment.  Thermochemical gasification of biomass has been identified 
as a possible way to produce renewable hydrogen.  Several researchers have investigated 
hydrogen production  from methanol, ethanol and biomass materials in subcritical and 
supercritical water.
1-8
  Compared to other biomass thermochemical reforming processes, 
supercritical water reforming has a high efficiency and operates at a lower temperature.  
Supercritical water (SCW) possesses properties very different from that of liquid water at 
ambient conditions.  The dielectric constant of SCW is much less than that of ambient 
water and hydrogen bonding is much weaker.  Therefore SCW behaves like an organic 
solvent and is completely miscible with organic materials.  Thus with SCW it is possible 
to conduct reactions with organic compounds in a single fluid phase which would 
otherwise occur in a multiphase system under conventional conditions.
1, 5, 8-10 
  
Heterogeneous reactions can also be performed in supercritical fluids.  The high 
diffusivity in SCW can significantly enhance mass transfer.  SCW can reduce coke 
formation on the catalyst as it is a good solvent for the intermediate coke precursors. 
Hence, gasification of biomass in SCW has many advantages including high gasification 
efficiency and a high yield of hydrogen.  In addition, the product hydrogen is obtained 
directly at a high pressure, hence further compression is not needed.   Several attempts 
have been made in recent years for the gasification of biomass model compounds and 
biomass in supercritical water
. 8-15
  Cellulose, glucose, and phenolic compounds (e.g., 
phenol, guaiacol) can be gasified in subcritical and SCW in the presence of a metal 
catalyst (platinum, nickel, etc.) or an alkali catalyst.
16-26
  Cortright et al.
16
 conducted 
glucose gasification at 265 ?C and 56 bar in water, using a platinum catalyst, and 
obtained high yields of CO
2
 and H
2
. Minowa et al.
18
 investigated  cellulose gasification in 
 28
near-critical water at 350 ?C and 165 bar with a reduced nickel catalyst and reported that 
70% of the carbon could be gasified.  
One of the main problems encountered during the biomass gasification is reactor 
plugging.  Yu et al.
12
 reported that the gasification of biomass is effective only at low 
concentrations of biomass, as at high concentrations, polymerization of the 
decomposition products occurs.  Kruse et al.
7,9,10
 observed that the hydrogen yield can be 
significantly increased by adding 5 wt% KOH in SCW.  Watanabe et al.
13,14
 observed a 
two fold increase in the gasification efficiency of glucose and cellulose in a batch  reactor 
by adding zirconia catalyst.  Similarly Nickel based catalysts have also been tested.  
Elliott et al.
21
 conducted a gasification of p-cresol in water at 350?C and 200 bar using 
various types of base and noble catalysts and reported that nickel and ruthenium were 
active for the reaction.   
Again using a batch reactor, Osada et al.
22
 reported formation of a high amount of 
methane in the presence of ruthenium catalyst during the gasification of cellulose and 
lignin in SCW.  In a recent study, Lu et al.
23 
studied the gasification of cellulose in the 
presence of metal catalyst including CeO
2
, Pd/C and Ru/C in a batch reactor, and  
observed that the maximum hydrogen yield was obtained with Ru/C catalyst.  
Most of the above studies were done in batch mode, in which the 
biomass/water/catalyst is loaded in a small steel tube reactor and then sealed and placed 
in an oven. After the reaction, the mixture is quenched and analysed. Typical reaction 
time varied from minutes to hours. In our recent work on methanol reforming, it was 
observed that high reaction time leads to the secondary reaction of methane formation.
1
 
To limit the methane formation reaction time needs to be limited to the order of seconds.  
 29
This work examines the reforming of glucose in SCW in a continuous reactor with a 
short reaction time (of the order of seconds). Ru/Al
2
O
3
 is selected as a catalyst. The effect 
of reaction time, temperature and feed concentrations is studied. 
2.3  Experimental Section 
2.3.1  Materials.   All the chemicals used were of high purity (99.9%) and of analytical 
grade.  Glucose and catalyst (5 wt.% Ru/amorphous Al
2
O
3
) were procured from Aldrich 
and used without any further treatment.  The surface area and pore volume of the catalyst 
is 100 m
2
/g and 0.30 cm
3
/g, respectively.  Distilled and deionized water was used. 
2.3.2  Apparatus.   The reforming of glucose was carried out in a tubular reactor (0.5 m 
long, 0.250? OD and 0.12? ID) made of Inconel 600 (Microgroup) having a composition 
of 73% Ni, 18% Cr, 9% Fe.  Initially a few experiments were also carried out in empty 
bed reactor to study the effect of reactor wall.  For the catalytic experiments, a known 
mass of the catalyst (2.0 g) was carefully poured into the tube reactor with gently tapping 
on the outside of the reactor wall to ensure no large voids in the bed.  The catalyst was 
retained in the reactor by placing stainless steel frits with a pore diameter of 0.5 ?m 
(Valco) at either end.  The same charge of catalyst was used for all of the experiments.  
Glucose feed stock was prepared by dissolving a known mass of glucose in water and 
from the feed tank it was pumped to the reactor using an HPLC pump (Waters 590).  The 
schematic diagram of the experimental set up is shown in Figure 2.1.  The reactor and 
preheater assembly were placed inside a tubular furnace equipped with a temperature 
controller (Thermolyne 21100).  The reactor temperature at the exit of the furnace was 
monitored using a type-K thermocouple.  The ends of the tube furnace were covered 
properly to avoid heat loss and achieve a uniform temperature.  The gaseous products 
 30
exiting the reactor were cooled using a water-cooled double pipe heat exchanger made of 
SS 316 tubing.  The reactor pressure was constantly monitored by a pressure gauge.  The 
pressure of the stream was reduced to ambient by means of a back pressure regulator 
(Straval BP00201T-SS).  The gas-liquid mixture was separated in a glass phase-separator 
having gas tight valves to prevent the escape of gases.   The gas flow rate was measured 
using a gas flow meter (Omega FMA-1600).  A six-port injection valve (Valco) having a 
100 ?L sample loop was used for the online sample injection.  The product gas 
composition was measured using a gas chromatograph (SRI 8610C) equipped with a 
TCD and 60-80 mesh Carboxen-1000 carbon molecular sieve column (Supelco) having 
dimensions of 15? x 1/8?.  Nitrogen was used as the carrier gas.  The gas chromatograph 
was calibrated using a standard gas mixture of known composition.  The liquid product 
coming out of the phase separator was collected at regular time intervals.  The total 
organic carbon (TOC) content of the liquid was analyzed using a TOC analyzer 
(Shimadzu TOC-V
CSN
).  The flow meter (FMA-1600) was used in the H
2
 mode and it 
generated 30 readings of the volumetric flow rate per second based the built-in properties 
of pure H
2
.  These instantaneous volumetric flow rates were acquired on computer via a 
RS-232 port and corrected for pressure and temperature.  The average volumetric flow 
rate was found out by totalizing the flow for a period of 15 minutes.  This average gas 
flow rate which corresponds to pure H
2
 was corrected to the actual gas coming out of the 
phase separator by estimating the viscosity of the gas mixture by Wilke?s semi-empirical 
formula
24  
using the gas composition obtained from the GC analysis.  This flow 
measurement method was checked and confirmed for accuracy by flowing the calibration 
gas consisting of H
2
, CO, CH
4 
and CO
2
 of known composition at several flow rates  and 
 31
measuring the actual flow by a soap-bubble flow meter.  All the experiments were carried 
out under isothermal conditions.  
 
 
Glucose 
Solution  
T
P 
HPLC pumps 
Water-cooled 
heat exchanger 
 
GC with 
TCD 
6-port injection 
valve for online 
sampling 
Gas flow 
meter 
Tubular reactor 
Tube furnace with 
temperature controller 
T 
Back 
pressure 
regulator
Gas- Liquid 
separator 
 
 
Data acquisition 
system 
Data analysis with 
software 
 
TOC analyzer 
for liquid phase 
Gas flow 
totalization 
Water preheater 
Mix pt. 
 
Water  
 
                        Figure  2.1: Schematic diagram of the experimental apparatus 
 
2.3.3  Experimental Procedure.  First, distilled water was pumped through the system 
and pressurized to the desired pressure by adjusting the backpressure regulator.  After 
achieving a steady pressure, the tube furnace was switched on to heat the reactor.  After a 
steady exit temperature was achieved, glucose solution was introduced into the reactor.  
The steady state condition was marked by a constant temperature at the exit of the 
reactor.  The gas analysis was done at least three times to get a constant gas composition.  
The gas flow was totalized for a period of 15 min, and the average gas flow rate was 
 32
calculated using the method discussed before.  After completion of each experiment, the 
feed was switched back to distilled water to flush the reactor.   
All the experiments were conducted at a pressure of 241 bar (3540 psi) and at 
temperatures ranging from 700-800?C.  Experiments were also carried out in empty bed 
reactor with glucose feed stock in supercritical water to study the influence of reactor 
wall.  The effect of glucose feed concentration was studied by feeding glucose having a 
concentration ranging between 1.0 to 5.0 wt% calculated at the entrance of the reactor. 
The residence time in the catalyst bed was kept between 1 to 6 seconds. 
The molar flow rates of the product gases were calculated based on the volumetric 
gas flow rate and dry gas composition obtained from the GC.  The carbon content of the 
liquid stream was calculated knowing the TOC value of the liquid.  All the measurements 
were made in triplicate.  The accuracy of the run was checked by calculating the overall 
carbon balance for the system.  The error in the overall carbon balance was found to be 
less than 10%.  Scattering in the data of the totalized gas flow rate measured by the flow 
meter was less than 1%.  The error in the dry gas composition obtained by the GC 
analysis was typically less than 2%.  The overall error in the calculation of the gas yields 
due to the errors introduced by the individual analysis techniques and experimental error 
was found to be less than 5%. 
2.4  Results and Discussion 
2.4.1  Effect of Ru/Al
2
O
3
 Catalyst.  To study the effect of catalyst on hydrogen yield for 
reforming of glucose in supercritical water, the catalytic experiments were compared to 
the empty reactor experiments (without catalyst) under identical conditions.  The gas 
yield is defined as the moles of product gases divided by the moles of glucose fed to the 
 33
reactor.  Typical product distributions are shown in Table 2.1 for catalytic and non 
catalytic experiments at 700?C with 1 wt.% glucose feed.   
 
Table 2.1: Product Distribution during Thermal and Catalytic Reforming of Glucose in 
Supercritical Water (T: 700?C, P: 241 bar, 1.0 wt% glucose) 
 Typical product composition (mol%)  
H
2
 CO      CH
4
              CO
2
           H
2
 Yield 
 
Empty bed  54.0% 1.7% 10.2% 34.1%            6.8 
 
Catalyst 
(Ru/Al
2
O
3
) 68.9% 0.1% 1.3% 29.8%         ~12 
 
The hydrogen yield increased from 7 to 12 (moles of hydrogen formed /moles of 
glucose fed) upon adding the catalyst.  There was also a significant reduction in carbon 
monoxide and methane yields in the presence of the catalyst.  The main products of the 
reaction were hydrogen, methane, carbon dioxide and carbon monoxide.  Small amounts 
of higher molecular weight carbon compounds such as phenols and aldehydes could be 
seen in the liquid product in a few experiments.  Typical values of TOC in the liquid 
product were approximately 150 ppm. Although the detailed properties of the catalyst, 
such as catalyst life and strong metal- support interaction have not been reported for 
Ru/Al
2
O
3
 catalyst, the probable reason for higher gasification performance is that the 
intermediate agents formed during glucose decomposition, such as aldehydes and 
phenols, were gasified. 
The gasification of hydrocarbons in supercritical water proceeds via several 
complex reactions such as pyrolysis, hydrolysis, steam reforming, water gas shift, and 
methanation.  The product distribution mainly depends upon the relative extent of various 
 34
reactions.  During thermal gasification (without catalyst) of glucose, the product gas 
contained approximately 54 mol% hydrogen along with 1.7% CO, 34.1 % CO
2
 and 10.2 
% methane.  Due to the presence of catalyst the amount of hydrogen was significantly 
increased to the maximum theoretical value to approximately 69%.  The yield of major 
products for thermal and catalytic gasification of glucose is given in Figures 2.2 and 2.3, 
respectively.   
0
2
4
6
8
02468
Residence Time (s)
G
as Yie
ld 
(mo
l gas/mol gl
ucose fed)
H2
CO2
CH4
CO
 
Figure  2.2: Effect of residence time on product gas yields during non-catalytic reforming 
of glucose (T: 700?C, P: 241 bar, glucose conc.: 1 wt.%) 
 
 35
0
2
4
6
8
10
12
14
1234567
Residence Time (s)
Ga
s
 Y
i
e
l
d 
(mol gas/mo
l glucose fed)
H2
CO2
CH4
CO
 
Figure  2.3: Effect of residence time on product gas yields: (T: 700?C, P: 241 bar, glucose 
conc.: 1 wt.%, 2.0 g Ru/Al
2
O
3
 catalyst) 
 
In the catalytic experiments, there was only a negligible amount of CO in the 
product, and the methane concentration was also significantly reduced. Here glucose 
undergoes dehydrogenation on the metal surface to give adsorbed intermediates before 
the cleavage of C-C or C-O bonds. Subsequent cleavage of C-C bonds leads to formation 
of CO and H
2
. CO reacts with water to form CO
2 
and H
2
 by the water gas shift reaction. 
Carbon monoxide was present only in trace amounts, probably because the Ru catalyst 
promotes the water gas shift reaction to form carbon dioxide and hydrogen from carbon 
monoxide and water.  The shift reaction is initiated through interaction of CO with OH
-
, 
which are formed by ionic dissociation of supercritical water on the metal surface, and 
forming the formate ion which then decomposes in to CO
2
 and hydride anion.  The 
hydride anion further interacts with water, forming H
2
 and OH
-
 by electron transfer. 
25
 
OH 
-   
+  CO    HCOO
-
    H
- 
  +  CO
2                                                     
(2.1)
 
??? ???
    H
- 
  + H
2
O    H
2
 + OH
-                                                                                              
(2.2)
 
???
 36
2.4.2  Effect of Residence Time.  Effect of residence time inside a catalyst bed was 
studied by varying the contact time from 1.6 to 6. 2 seconds. This residence time was 
calculated as reactor void volume in the catalyst bed divided by the volumetric flow rate 
at experimental conditions.  The results are shown in Figure 2.4 for 5 wt.% glucose in the 
feed.  In the range of experiments carried out, hydrogen yield increased progressively as 
the contact time increased.  However, beyond a contact time of 4 s the increase was 
negligible.  Effect of increasing residence time on total organic carbon in the liquid 
product for catalytic and non-catalytic experiments is shown in Figure 2.5.  Increasing the 
residence time in the non-catalytic runs significantly reduced the total organic carbon in 
the liquid indicating that these were further converted to CO
2
 and H
2
 at higher residence 
time.  In the catalytic experiments, the low TOC values correspond to near complete 
conversion of carbon to gaseous products. It should be noted that the residence time in 
these experiments are much lower than the previous batch experiments.
22
 
0
2
4
6
8
10
12
14
012345
Residence Time (s)
G
as Yield (mol 
gas/mol glucose fed)
H2
CO2
CH4
CO
 
Figure  2.4: Effect of residence time on product gas yields. (T: 700?C, P: 241 bar, glucose 
conc. 5 wt. %, 2.0 g Ru/Al
2
O
3
 catalyst) 
 37
0
500
1000
1500
2000
2500
3000
01234567
Residence Time (s)
Tot
a
l O
r
ga
nic
 Ca
rb
on (
ppm
)
Without catalyst
With Ru/Al2O3 catalyst
 
Figure 2.5: Effect of residence time on total organic carbon in liquid phase (T=700?C, P: 
241 bar, residence time: 2 s) 
 
2.4.3  Effect of Temperature.  Reactor temperature has a strong influence on the 
gasification rate and hydrogen formation in biomass reactions.  To study the influence, 
experiments were carried out at temperatures ranging between 700 and  800?C.  The 
residence time inside the catalyst bed was 2 s while the glucose feed concentration was 4 
wt.%.  As shown in Figure 2.6, hydrogen yield was near the theoretical maximum of 12 
moles H
2
/mol of glucose fed.  Additional experiments carried out at 650?C showed 
incomplete conversion of glucose to gaseous products, as witness by a strong 
hydrocarbon smell in the liquid product, which was further confirmed by high TOC 
values.  Further investigation at low temperature was abandoned in the interest of 
preventing any tar formation on the surface of the catalyst.  Compared to the thermal 
gasification, significantly lower temperature was effective for the hydrogen production 
over Ru/Al
2
O
3
 catalyst.  
 38
With an increase in the temperature the hydrogen and carbon dioxide yields 
increase while the methane yield decreases. Thermodynamically at low reaction 
temperatures, H
2
 and CO
2
 readily react to form alkanes and water.  However, in the 
present study due to the presence of catalyst and low residence time (which avoids 
attainment of equilibrium), methane formation is suppressed even at low reaction 
temperature.  The low carbon monoxide yield indicates that the water gas shift reaction 
approaches completion.  The high water excess leads to a preference for the formation of 
hydrogen and carbon dioxide instead of carbon monoxide.  In addition, the intermediate 
products such as acids and aldehydes which may otherwise form in the conventional 
reactions are not found in supercritical water reforming in appreciable concentrations.  As 
compared with a typical temperature required (800? 1000?C) in conventional biomass 
gasification, supercritical water gasification can be carried out at a lower temperature. 
 
 
 39
0
2
4
6
8
10
12
14
800?C 700?C
Gas Y
i
el
d
 
(
m
o
l
 ga
s
 /
 m
o
l
 gl
uc
o
s
e
 
f
e
d)
 
H
2
 
H
2
 
CO
2
CH
4
CO
2
CO 
CO
CH
4
 
Figure  2.6: Effect of temperature on product yields (P: 241 bar, glucose conc.: 4 wt.%, 
residence time: 2 s, 2.0 g Ru/Al
2
O
3
 catalyst) 
 
2.4.4  Effect of Glucose Concentration.   Figure 2.7 shows the effect of glucose feed 
concentration on product gas yields at 700?C.  The hydrogen yield drops by 17% as the 
glucose concentration in feed was increased from 1 to 5 % wt.%.  Also, a decrease in the 
carbon dioxide yield and a small increase in methane and carbon monoxide yield were 
observed on increasing glucose concentration. The trends of experimental yield of 
hydrogen, carbon dioxide and carbon monoxide are similar to as those reported by other 
researchers.
28, 29, 30
  Further increase in the glucose concentration results in formation of 
heavier molecular weight hydrocarbons and coke which often plugs the reactor.  
However, the heavy molecules and coke can be reduced by increasing temperature in 
SCW.   
 40
0
2
4
6
8
10
12
14
0123456
wt.% Glucose in Feed
Gas Yi
e
l
d
 
(
m
ol ga
s
/m
ol g
luc
os
e
 f
e
d)
H2
CO2
CH4
CO
 
Figure  2.7: Effect of glucose concentration on product yields (T: 700?C, P: 241 bar, 
residence time: 2 s: 2.0 g Ru/Al
2
O
3
 catalyst) 
 
2.4.5  Reaction Mechanism.    As discussed earlier, biomass gasification proceeds via 
several complex reactions, such as pyrolysis, hydrolysis, steam reforming, water gas shift 
reaction and methanation.  For glucose reforming in supercritical water, a maximum 
theoretical yield of hydrogen can be calculated by converting all the feed carbon to 
carbon dioxide with water as 
 
C
6
H
12
O
6
 + 6 H
2
O    6 CO
2 
+ 12 H
2
                                                   (2.3) ???
 
In overall reaction scheme, CO
2
 also undergoes hydrogenation reaction to form CO and 
CH
4
 as 
CO
2 
+ H
2
    CO + H
2
O                                                                           (2.4) ???
and 
 41
CO + 3 H
2
    CH
4 
+ H
2
O                                                               (2.5) ???
 
However, mechanistically glucose reforming via reaction intermediates can be shown as 
follows
25
: 
 
Glucose  Acids / Aldehydes  Gases                              (2.6) ??? ???
 
With the presence of ruthenium catalyst in supercritical water, extremely high 
gasification rate and a high hydrogen yield was obtained.  Also, negligible coke was 
formed on the catalyst surface for up to 4 wt.% glucose feed, which indicates that  all the 
solid products such as chars and aldehydes are decomposed by the supercritical water in 
the presence of catalyst.  The dominant reactions in the gas phase are those for water gas 
shift and methanation reaction. 
The presence of methane and some traces of liquid hydrocarbon indicate that 
reforming of glucose in supercritical water occurs via above reaction intermediates.  
Reforming of these intermediates by water gas shift reaction is highly likely since water 
is in high excess. Although the exact role of ruthenium is not established in supercritical 
water reforming, however it may be assumed that some complex is formed as a result of 
the reaction between adsorbed glucose and water supercritical water.  The formation of 
adsorbed intermediates by the decomposition of this complex is assumed to be the rate 
determining step.  The adsorbed intermediates further react with water in the gas phase to 
give CO
2 
and H
2
.  Assuming S denotes an active site on the catalyst and A and B denote 
glucose and water, respectively, the overall reaction steps may be represented as follows: 
 42
 
A + S    AS                                                              (2.7) ????
?11
,kk
AS+ B  ABS                                                                           (2.8) ???
2
k
ABS  Intermediates  CO
2
 + H
2
                                        (2.9) ???
3
k
???
4
k
 
where, Equation 2.7 denotes the reversible adsorption of glucose on the catalyst surface, 
and Equation 2.8 represents the reaction of adsorbed glucose molecule with water to form 
a complex molecule ABS.  Assuming the steady state hypothesis for the intermediate 
complex ABS and AS, the dependence of the rate (r) on reactant concentration (c) can be 
expressed as: 
)(
3
21
211
21
k
cckk
ckckk
cckk
r
BA
BA
BA
+++
=
?
                                                     (2.10) 
 
Since in our experiments, water concentration is significantly high, c
B
 is assumed 
constant and the above expression is simplified as: 
A
AR
bc
ck
r
+
=
1
                                                                                                   (2.11) 
 
where, k
R
 and b are lumped parameters defined as 
 
Bo
Bo
R
bck
ckk
k
+
=
?1
21
                                                                                          (2.12)  
and 
 43
Bo
Bo
ckk
k
ckk
k
b
21
3
21
1
)(
+
+
=
?
                                                                                      (2.13) 
 
When  (b c
A
) << 1, Equation 2.11 reduces to 
 
r = k
R
c
A
                                                                                                          (2.14) 
 
The above equation was solved using a non linear regression technique on rate of 
hydrogen production at different glucose concentrations at a constant pressure and 
temperature. The value of rate constant (k
R
) at 700?C was calculated to be 0.33 s
-1
. The 
lowest rate of hydrogen production in the present study is more than  4x10
4
 ?moles/g. h, 
which is significantly higher than the maximum value of hydrogen production rate 
reported  for glucose reforming by enzymatic route (7x10
2
 ?moles/g. h) and also higher 
than the values reported for catalytic reforming of glucose over Pt/Al
2
O
3
 catalyst .
16
 
2.5  Conclusions 
     The supercritical water reforming of biomass materials appears to be a suitable 
technique for the production of high-pressure hydrogen with short residence times.  
Presence of 5 wt.% Ru/Al
2
O
3
 catalyst significantly enhanced the conversion and 
hydrogen yield from glucose, while significantly  reducing char and tar formation.  The 
product gases mainly consist of hydrogen and carbon dioxide, and a small amount of 
methane and carbon monoxide.  There are effectively no tarry products in the liquid 
effluent and negligible losses in catalytic activity for up to 5 wt% glucose in the feed. 
 44
However, the hydrogen yield decreases as the concentration of glucose in the feed 
increase beyond 5wt.%. 
 
2.6  Acknowledgement:  This research was supported by the U.S. Department of Energy 
through the Office of Fossil Energy (FE), National Energy Technology Laboratory 
(NETL), under DOE Contract DE-FC26-05424.56 via the Consortium of Fossil Fuel 
Science. 
 45
2.7  References 
 1. Gadhe, J.B.; Gupta, R.B.; ?Hydrogen production by methanol reforming in 
supercritical water: suppression of methane formation? Ind. Eng. Chem. Res. 2005,  (44), 
4577-4585. 
2. Patel Sanjay and K.K.Pant;  Production of hydrogen with low carbon monoxide 
formation via catalytic steam reforming of methanol, Journal of Fuel Cell Science and 
Technology  2006, Vol. 3 / 369 
3. Matsumura, Y.; Xu, X.; Antal, M. J. Gasification Characteristics of an Activated  
Carbon in Supercritical Water. Carbon 1997, 35, 819. 
4. Sahoo,D.R, Vajpai, S;   Patel,  S.;  Pant  K.K; Kinetic modeling of steam 
reforming of ethanol for the production of hydrogen over Co/Al2O3 catalyst Chemical 
Eng Journal 2006,(in press) 10.1016/j.cej.2006.08.011 
 
5.         Xu, X.; Matsumura, Y.; Stenberg, J.; Antal, M. J., Jr. Carbon-catalyzed 
gasification of organic feedstock in supercritical water. Ind. Eng. Chem. Res. 1996, 35, 
2522. 
 
6.       Xu, X. D.; Antal, M. J., Jr. Gasification of sewage sludge and other biomass for 
hydrogen production in supercritical water. Environ. Prog. 1998, 17, 215. 
 
7.        Antal, M. J., Jr; Allen, S. G.; Schulman, D.; Xu, X. D.; Divilio, R. J. Biomass 
gasification in supercritical water. Ind. Eng. Chem. Res. 2000, 39, 4040. 
 
8.        Kruse, A.; Meier, D.; Rimbrecht, P.; Schacht, M. Gasification of pyrocatechol in   
supercritical water in the presence of potassium hydroxide. Ind. Eng. Chem. Res. 2000, 
39, 4842. 
 
9.        Kruse, A.; Gawlik, A. Biomass conversion in water at 330- 410 ?C and 30-50 
MPa.     identification of key compounds for indicating different chemical reaction 
pathways. Ind. Eng. Chem. Res. 2003, 42, 267. 
  
10.      Kruse, A.; Henningsen, T.; Sinag, A.; Pfeiffer, J. Biomass gasification in 
supercritical  water: Influence of the dry matter content and the formation of phenol. Ind. 
Eng. Chem. Res. 2003, 42, 3711. 
 
11.      Yoshida T., Oshima Y. ?Partial Oxidative and catalytic biomass gasification in      
supercritical water: A promising Flow reactor system? Ind. Eng. Chem. Res. 2004,  43, 
4097. 
 
12.      Yu, D.; Aihara, M.; Antal, M. J., Jr. Hydrogen production by steam reforming 
glucose in supercritical water. Energy Fuels 1993, 7, 574. 
 
 46
13.  Watanabe, M.; Inomata, H.; Arai, K. Catalytic hydrogen generation from 
biomass  (glucose and cellulose) with ZrO2 in supercritical water. Biomass Bioenergy 
2002,    22, 405. 
 
14. Watanabe, M.; Inomata, H.; Osada, M.; Sato, T.; Adschiri, T.; Arai, K. Catalytic  
effects of NaOH and ZrO2 for partial oxidative gasification of n-hexadecane an  lignin in  
supercritical water. Fuel 2003, 82, 545. 
 
15. Holgate, H. R.; Meyer, J. C.; Tester, W. J. Glucose hydrolysis and oxidation in 
supercritical water. AIChE J. 1995, 41, 637. 
 
16. Cortright, R. D.; Davda, R. R.; Dumesic, J. A.  Hydrogen from catalytic 
reforming of biomass-derived hydrocarbons in liquid water.   Nature 2002, 418, 964-967. 
 
 17.  Minowa, T.; Ogi, T.; Yokoyama, S. Hydrogen production from wet cellulose by 
low temperature gasification using a reduced nickel catalyst. Chem. Lett. 1995, 286, 937. 
 
 18. Minowa, T.; Ogi, T.; Yokoyama, S. Effect of pressure on low-temperature 
gasification of wet cellulose into methane using reduced nickel catalyst and sodium 
carbonate. Chem. Lett. 1995, 280, 285. 
 
19. Yoshida, T, Matsumura, Y. Gasification of Cellulose, Xylan, and Lignin Mixtures 
in Supercritical Water. Ind. Eng. Chem. Res. 2001, 40, 5469. 
 
20. Yoshida, T.; Oshima, Y.; Matsumura, Y. Gasification of biomass model 
compounds  and real biomass in supercritical water. Biomass Bioenergy 2004, 26, 71. 
 
21. Elliott, D. C.; Sealock, L. J.; Backer, E. G.  Chemical processing in high pressure 
aqueous environment: development of catalyst for gasification Ind. Eng. Chem. Res    
1993, 32, 1542-1548. 
 
22. Osada M. ; Sato, T.; Wantabe M.; Adschiri, T.; Arai, K.;  Low temperature 
catalytic gasification of lignin and Cellulose with a Ruthenium Catalyst in Supercritical 
water. Energy and Fuels 2004, 18 327-333. 
 
23.  Lu Y.J.; Guo l.j.; Ji, C.M. , Zhang,  X.M.; Hao, X.H.; Yan Q.H.; Hydrogen 
production by biomass gasification in supercritical water: A parametric study Int. J. of 
Hydrogen Energy 31 2006, 822-831. 
 
24. Saisu, M.; Sato, T.; Watanabe, M.; Adschiri, T.; Arai, K. Conversion of lignin 
with   supercritical water-phenol mixtures. Energy Fuels 2003, 17, 922. 
 
 25. Penninger J. M.L; Rep. M.; Reforming of aqueous wood pyrolysis condensate in 
supercritical water. Int. J. of Hydrogen Energy  31 (2006) 1597-1606 
 
 47
26. Courson, C.; Makaga, E.; Petit, C.; Kiennemann, A. Development of Ni catalysts 
for gas production from biomass gasification. Reactivity in steam- and dry-reforming. 
Catal. Today. 2000, 63, 427. 
 
27. Wilke, C. R. ?A viscosity equation for gas mixtures? Journal of Chemical Physics 
(1950), 18, 517-19.   
 
28. Sinag, A.; Kruse, A.; Schwarzkoph, V. Key compounds of the hydropyrolysis of 
glucose in supercritical water in the presence of  K2CO3. Ind. Eng. Chem. Res.  2003, 42, 
3516. 
 
29. Yan, Q.; Guo, L. ; Lu, Y.;  Thermodynamic analysis of hydrogen production from 
biomass gasification in supercritical water. Energy Conversion and Management 2006, 
47, 1515-1528. 
 
30. Hashaikeh, R.; Butler,S.;  Kozinski, J.A.; Selective promotion of catalytic 
Reactions during biomass gasification to hydrogen.  Energy & Fuels 2006, 20, 2743-
2747 
 48
3.  Hydrogen Production from Ethanol by Reforming in Supercritical Water using 
Ru/Al
2
O
3
 Catalyst 
 
3.1  Abstract 
Supercritical water is a promising reforming media for the direct production of 
hydrogen at high pressures with a short reaction time.  In addition to being a dense 
solvent, supercritical water also participates in reforming reaction.  In this work, high 
pressure hydrogen is produced from ethanol by reforming over a Ru/Al
2
O
3 
catalyst with 
low methane and carbon monoxide formation.  Experiments were conducted in a 
continuous tubular reactor to study the effects of temperature, pressure, residence time, 
and water-to-carbon ratio on the H
2
 yield.  Hydrogen formation is favored at higher 
temperatures and at high water-to-ethanol ratios. The formation of methane can be 
suppressed by operating at an optimal residence time, high reactor temperature and a low 
feed concentration of ethanol. Excellent conversion in reaction time as short as 4 seconds 
is achieved.  Pressure has negligible effect on hydrogen yield above the critical pressure, 
and for less than 10 wt% ethanol concentration in the feed, there was negligible coke   
activation energy of 65.3 kJ mol
-1
 was observed. 
3.2  Introduction 
As the world supply of fossil fuels depletes there is a growing need for renewable 
energy sources.  The current use of fossil fuels is responsible for pollution due to carbon 
dioxide, carbon monoxide, hydrocarbons, sulfur dioxide and nitrogen oxide.  Every 
gallon of gasoline burned in an automobile produces approximately 20 pounds of the 
greenhouse gas CO
2
, and the transportation sector is responsible for one third of all CO
2
 
 49
emissions. Considerable efforts are currently under way to minimize the emissions.  One 
promising alternative to fossil fuels is the use of hydrogen as an energy carrier, which 
provides zero emission of pollutants and high energy efficiency when used in polymer 
electrolyte membrane (PEM) fuel cell.   
Hydrogen can be produced from a variety of starting materials, both renewable 
and nonrenewable, via several different processes.  Current technologies for the 
commercial production of H
2
 include steam reforming, partial oxidation, and electrolysis 
of water.  Out of these, steam reforming is most commonly used, which catalytically 
converts hydrocarbons or oxygenated hydrocarbons to produce a mixture of H
2
 and CO 
followed by a water gas shift reaction to produce a mixture of H
2 
and CO
2
 along with 
small amount of un-reacted CO.
1-3
  Supercritical water is an environmentally friendly 
fluid and is gaining in popularity as a reaction medium owing to fast heat and mass 
transfers and adjustable density.  The thermochemical properties of water at various 
pressures are summarized elsewhere.
4
  The properties of supercritical water such as 
density, viscosity and hydrogen bonding are quite different from those of steam or liquid 
water
4,5,6
. In the supercritical region, the dielectric constant of water is much lower.  
Further, the number of hydrogen bonds is much smaller and their strength is considerably 
weaker. As a result, SCW behaves as an organic solvent and exhibits extraordinary 
solubility towards organic compounds containing large nonpolar groups and most 
permanent gases.
6-9
 Another advantage of SCW reforming is that the H
2
 is produced at a 
high pressure, which can be stored directly, thus avoiding the large energy expenditures 
associated with its compression.  The process becomes economical as the compression 
work is reduced owing to the low compressibility of liquid feed compared to that of 
 50
gaseous H
2
.  For heterogeneous catalytic reactions in supercritical fluids, the high 
diffusivity of supercritical fluids can greatly reduce mass-transfer limitations and extract 
the coke precursors from the catalyst surface to prevent coking. Hydrocarbons are 
completely soluble in supercritical water, which minimizes the formation of char or slag, 
which may otherwise lead to catalyst deactivation.  Research carried out on thermal and 
catalytic gasification of biomass or its model compounds in supercritical water revealed 
that there is lower tar or char formation as compared to conventional reforming.
8-16
.  
Osada et al. 
10 
gasified lignin and cellulose in a batch reactor at 400?C in supercritical 
water with a ruthenium catalyst with gas yields of 30% and 70%, respectively with 
methane as a major gas product and no solid product was formed. 
Among the available raw materials, ethanol is an attractive option for hydrogen 
production because it is less toxic than methanol and it can be produced renewably from 
biomass with little net addition of carbon dioxide to the atmosphere.  Ethanol is already 
being produced from corn for use as a fuel or fuel additive in automobiles. But if it were 
used instead to produce hydrogen for a fuel cell, the whole process would become more 
energy efficient. Theoretically, a bushel of corn would yield three times as much power if 
its energy were channeled into hydrogen fuel cells rather than burned along with 
gasoline, because ethanol in car engines burns with 20% efficiency, whereas ethanol 
reformed to hydrogen for a fuel cell has more than 60% efficiency.
17
  The development of 
a process to produce H
2
 directly at a very high pressure is attractive since it avoids 
compression expenses for storage.  Hydrogen can be produced from ethanol and 
methanol feed stocks through steam reforming, partial oxidation or a combination of the 
two.
18-21
 Steam reforming of ethanol over Ni, Co, Ni/Cu and noble metals has been 
 51
extensively used.
21
 The main problem with the ethanol reforming is that besides the 
formation of H
2
, CO, CH
4
, the gaseous fuel product contains high levels of CO, which is 
poisonous to the Pt anode of the PEM fuel cell. Among the noble metals, Ru has been 
reported as the most promising catalyst having high activity in steam reforming of 
hydrocarbons and biomass gasification
6,10,22-24
.  Liguras et al.
24
 studied the steam 
reforming of ethanol over Ru/Al
2
O
3
 catalysts and reported high yield of hydrogen in 
subcritical water.  At high Ru loading (5 wt%) the catalyst was stable and had reasonable 
activity and selectivity.  There have been some studies to investigate the thermal H
2
 
production in supercritical water from a variety of organic feedstock such as methane, 
methanol, ethanol, glucose, and glycerol.
6-16
 Most of the hydrocarbon reforming studies 
in supercritical water have been carried out without a catalyst (except for the catalytic 
effect of the reactor walls that provide very small surface area) and some studies are 
reported over a catalyst.
6,9 ,25-27
  It has been reported that methanol reforming in 
supercritical water results in a hydrogen-rich product stream that has low concentrations 
of both carbon monoxide and methane.
25
  Watanabe investigated the chemistry of 
carbohydrate in the presence of ZrO
2
 as a catalyst.
9
  Taylor et al
26
 also conducted 
experiments to investigate the reforming of different hydrocarbons in supercritical water 
at 550?700?C and 27.6 MPa in a tubular Inconel 625 reactor. The results indicated that 
methanol can be completely converted to a product stream that is low in methane and 
near the equilibrium composition of hydrogen, carbon monoxide, and carbon dioxide. On 
the other hand ethanol and ethylene glycol resulted in less hydrogen yield and high 
concentrations of methane and carbon monoxide.  The yield of hydrogen produced in 
these experiments was significantly less and high yield of methane and CO has been 
 52
reported in the product stream.  Further, Boukis et al have studied methanol reforming in 
supercritical water catalyzed by an Inconel 625 reactor wall at both the lab scale and pilot 
plant scale.
28-36
 In conventional reforming processes, including partial oxidation and 
autothermal reforming, higher pressures disturb the equilibrium conversion. However, the 
equilibrium limitations in SCW reforming can be avoided if the reaction time is limited to 
a very short period.
4
   
The present work focuses on hydrogen production from ethanol in supercritical 
water over Ru/Al
2
O
3
 catalyst with minimal CO and methane in the product stream. In our 
recent study
6
, significantly high yield of hydrogen was obtained during the reforming of 
glucose in supercritical water over this catalyst. Effect of reactor temperature, residence 
time and concentration of ethanol in water on yield of hydrogen has been studied.  
Finally, a reaction mechanism has been proposed and the kinetics for the catalyst is 
determined. 
3.3  Experimental Section 
All the chemicals used were of high purity (99.9% pure) and of analytical grade.  
The commercial 5 wt% Ru/Al
2
O
3 
was purchased from Aldrich. The catalyst had the 
following characteristics: Crystalline structure: amorphous, total BET surface area: 100 
m
2
/g, specific pore volume: 0.30 mL/g; density: 0.95 g/cm
3
. Deionized ultra filtered 
water and HPLC grade ethanol were used as received from Fisher Scientific. The 
schematic diagram of the experimental apparatus is shown in Figure 3.1. The details of 
the experimental procedure are discussed elsewhere,
6
 and is briefly discussed here.  
Experiments were conducted in supercritical water in a fixed bed tubular reactor (0.5 m 
long, 0.25? OD and 0.12? ID) made of Inconel 600 (Microgroup) having a composition 
 53
of 73% Ni, 18% Cr, and 9% Fe, which was placed inside a temperature controlled 
furnace (Thermolyne 21100). 
  
 
Ethanol 
T
P 
HPLC pumps 
Water-cooled 
heat exchanger 
 
GC with 
TCD 
6-port injection 
valve for online 
sampling 
Gas flow 
meter 
Tubular reactor 
Tube furnace with 
temperature controller 
T 
Back 
pressure 
regulator
Gas- Liquid 
separator 
 
 
Data acquisition 
system 
Data analysis with 
software 
 
TOC analyzer 
for liquid phase 
Gas flow 
totalization 
Water preheater 
Mix pt. 
 
Water  
Figure  3.1: Schematic diagram of experimental apparatus 
 
A known mass of the catalyst, 1.92 g, was packed in the reactor. Ethanol from the 
feed tank was pumped to the reactor using an HPLC pump (Waters 590).  The feed tank 
was covered on top to avoid the loss of ethanol by evaporation.  The reactor temperature 
at the exit of the furnace was measured by using a type-K thermocouple with a tee 
arrangement.  Both ends of the tube furnace were covered properly to avoid heat loss and 
achieve uniform temperature. The gas mixture exiting the reactor was cooled using a 
water-cooled double pipe heat exchanger made of SS 316 tubing. Pressure was measured 
by a pressure gauge.  The pressure was let down to the ambient by means of a back 
pressure regulator (Straval).  The gas-liquid mixture was separated in a glass phase-
 54
separator having gas-tight valves to prevent the escape of gases.  The flow rate of the 
gases was measured using a gas flow meter (Omega FMA-1600).  A six-port injection 
valve (Valco) having a 100 ?L sample loop was used for the online sample injection. The 
gas composition was measured using a gas chromatograph (SRI 8610C) equipped with a 
TCD and 60/80 Carboxen-1000 carbon molecular sieve column (Supelco) having 
dimensions of 15? x 1/8?.  Nitrogen was used as the carrier gas. The total organic carbon 
(TOC) content of the liquid was analyzed using a TOC analyzer (Shimazdu TOC-V
CSN
).  
Characterization of the catalyst was performed before and after use by SEM (JEOL 
7000F) and X-ray diffraction (Rigaku diffractometer equipped with Cu
K?1
 radiation 
source, graphite monochrometer, and miniflex goniometer.  The diffractometer was ran at 
40 kV voltage and 40 mA current, and scanned at 5?/min with 0.05? step size.   
The total running time for each experiment was approximately three hours to 
collect necessary data.  All measurements were taken at least in triplicate to ensure 
accuracy.  Preliminary experiments were also carried out in an empty bed reactor with 
ethanol feed stock in supercritical water to study the influence of reactor wall. The 
accuracy of the run was checked by calculating the overall carbon balance for the system. 
The experiments with an error in the overall carbon balance greater than 5% were either 
repeated to check for carbon deposition or rejected.  Scattering in the data of the totalized 
gas flow rate measured by the flow meter was less than 1%.  The error in the dry gas 
composition obtained by the GC analysis was typically less than 2%.  The overall error in 
the calculation of the gas yields due to the errors introduced by the individual analysis 
techniques and experimental error was found to be less than 5%. 
3.4  Results and Discussion 
 55
The desired reaction for the reforming of ethanol is the complete conversion to 
hydrogen and carbon dioxide as 
 
C
2
H
5
OH + 3H
2
O ? 6H
2
 +2CO
2   
               (?H?
298 
 = 174 kJ/mol )                          (3.1) 
 
3.4.1  Effect of Ru/Al
2
O
3
 Catalyst.  To study the effect of the Ru/Al
2
O
3
 catalyst on 
hydrogen yield for reforming of ethanol in supercritical water, catalytic experiments were 
compared to the empty reactor experiments under identical conditions.  The gas yield is 
defined as the moles of product gases divided by the moles of ethanol fed to the reactor.  
The use of catalyst increases the hydrogen yield significantly compared to a process 
without Ru/Al
2
O
3
 catalyst. The main products of the catalytic reforming of ethanol were 
hydrogen and carbon dioxide, with small amounts of methane, carbon monoxide, and 
ethylene. Experimental conditions, product gas composition, and hydrogen yield are 
reported in Table 1 with and without Ru/Al
2
O
3
 catalyst.  The hydrogen yield increased 
from 3 to 4.5 (moles of hydrogen formed /moles of ethanol fed) upon addition of the 
catalyst. The hydrogen yield obtained in the SCW in the present study is significantly 
higher than the hydrogen yield reported for stream reforming of ethanol in subcritical 
water.
18,24
  There was also a significant reduction in carbon monoxide and methane yields 
in the presence of the catalyst.  A typical value of total organic carbon in the liquid 
product was approximately 120 ppm in most of the experiments, indicating conversion to 
gaseous products greater than 99%. Only two experiments conducted at 600 ?C showed 
higher levels of organic carbon in the liquid effluent. Although the detailed properties of 
the catalyst, such as catalyst life and strong metal-support interaction have not been 
 56
reported for Ru/Al
2
O
3
 catalyst, the probable reason for higher gasification performance is 
that the intermediate agents formed during ethanol decomposition such as dimethyl ether 
and acetaldehyde were also gasified in presence of supercritical water. The generation of 
H
2
 by the steam reforming of ethanol in sub-critical water leads to the formation of 
significant amounts carbon, which limits the yield of H
2
, and the reaction product 
contains higher hydrocarbons such as acetaldehyde, diethyl ether, ethane and ethylene in 
addition to the desired hydrogen. However in supercritical water, water becomes a strong 
oxidant, and oxygen in water can be transferred to the carbon atoms of the ethanol.  As a 
result of the high density, carbon is preferentially oxidized into CO
2
 but also low 
concentrations of CO were formed. In the present investigation, the reaction process can 
be operated continuously with the same catalyst. To study the effect of time on stream on 
catalyst performance, some experiments were repeated. There was no appreciable 
productivity change indicating that the catalyst activity does not decrease appreciably 
during the run. The gasification of hydrocarbons in supercritical water proceeds via 
several complex reactions such as ethanol decomposition, steam reforming, water gas 
shift, and methanation. The product distribution strongly depends upon the relative extent 
of these reactions.  It is assumed that during the reaction, ethanol undergoes 
dehydrogenation on the metal surface to give adsorbed intermediates before the cleavage 
of C-C or C-O bonds. Subsequent cleavage of C-C bonds leads to formation of CO and 
H
2
. CO reacts with water to form CO
2 
and H
2
 by the water gas shift reaction. Carbon 
monoxide was present only in trace amounts, probably because the Ru catalyst promotes 
the water gas shift reaction to form carbon dioxide and hydrogen from carbon monoxide 
and water. 
 57
Table 3.1:  Details of Experimental Conditions 
     
Product Gas Composition 
(mol%)  
T 
(?C) 
P 
(bar) 
Feed 
Conc. 
(wt%) 
Residence 
time (s) 
W/F
Ao
 (g-
cat*s/mol 
EtOH) H
2
 CO CH
4
 CO
2
 
H
2
 
Yield 
800 221 10 4 empty bed 62.8 3.5 10.8 22.9 3.0 
          
600 221 10 5 142449 63.1 5.6 12.0 19.3 2.1 
600 221 10 10 252433 57.5 0.5 15.3 26.8 2.6 
700 221 10 5 181515 56.5 0.2 15.6 27.6 2.4 
700 221 10 2 71563 55.6 0.5 16.1 27.8 2.5 
700 221 10 1 37520 57.7 0.6 15.0 26.7 2.8 
750 221 10 2 78297 62.0 1.1 11.9 24.9 3.3 
750 221 10 4 164234 63.4 0.8 10.2 25.6 3.9 
750 221 10 6 264062 67.2 0.5 8.8 23.5 3.3 
800 221 10 6 251264 68.8 0.9 8.3 21.9 3.5 
800 221 10 4 165517 70.4 1.0 7.5 21.1 4.5 
800 221 10 2 84223 65.7 1.0 10.3 23.0 3.6 
800 221 10 1 41845 64.8 1.2 11.6 22.4 3.7 
800 221 10 0.5 20923 63.0 1.0 12.6 23.4 3.3 
800 221 5 4 354399 73.0 0.7 2.8 23.5 5.3 
800 221 15 4 112060 64.6 1.2 11.6 22.7 3.1 
800 221 20 4 83815 59.5 1.3 14.9 24.3 2.4 
700 243 10 2 62412 56.1 0.5 15.9 27.4 2.7 
700 276 10 2 54219 53.2 0.7 18.0 28.0 2.6 
 
SEM was used to characterize the catalyst.  Figures 3.2 and 3.3 show SEM 
images of the fresh and used catalyst, respectively.  From these images it can be seen that 
the support has undergone a structural change.  The fresh catalyst consists of more-or-less 
round particles, characteristic of amorphous materials, whereas the used catalyst is 
clustered in aggregates of sharp-edged crystals.  Looking at the higher magnification 
images, it can be seen that particle size has also increased.  X-ray diffraction was used to 
further characterize the catalyst, and XRD spectra can be seen in Figure 3.4.  The spectra 
of the used catalyst show sharp narrow peaks characteristic of a crystalline compound.  
 58
Also it is evident from the width of the peaks that particle size has increased, as peak 
width is inversely proportional to particle size.  The composition of the new crystalline 
phase was not determined. 
 
 
 
 
 
 
 
 
 
 
 
 
(a) 
 59
 
 
(b) 
Figure  3.2:  SEM images of fresh Ru/Al
2
O
3
 catalyst. (a) 5500x magnification (b) 
33,000x magnification 
 
 
 
 60
(a) 
 
(b) 
Figure  3.3:  SEM images of used Ru/Al
2
O
3
 catalyst.  (a) 6000x magnification (b) 
30,000x magnification 
 
 61
0
200
400
600
800
1000
1200
10 20 30 40 50 60 70 80 90
2? (degrees)
Intensity (a.u.)
Fresh Ru/Al
2
O
3
Catalyst
Used Ru/Al
2
O
3
Catalyst
 
Figure  3.4:  X-ray diffraction spectra for fresh and spent Ru/Al
2
O
3
 catalyst 
 
3.4.2  Effect of Residence Time.  The residence time in the catalyst bed was studied by 
changing the inlet flow rate. The product yields (moles of product/ mole of ethanol fed) 
are given in Figures 3.5 - 3.7, at three different temperatures. No ethanol was seen in the 
liquid effluent indicating that it reacted completely in supercritical water.  At 800 ?C, the 
hydrogen yield varied from 2 to 4.5 as the residence time changed from 1 to 4 seconds 
and decreased thereafter.  In contrast, at the lower temperature of 600 ?C, high levels of 
organic carbon were seen in the liquid product at low residence time and higher hydrogen 
yields were obtained at relatively high residence time. The ethylene yield was also higher 
at the low temperature. Specifically at high temperatures, the low residence time can 
prevent the methanation reactions from reaching equilibrium.  This makes it possible to 
 62
significantly limit the extent of methanation and H
2
 loss by operating at a low residence 
time.  
 
0
1
2
3
4
5
6
510
Residence time in Ru/Al
2
O
3
 bed (s)
Product G
as Yield (m
ol gas/mol EtO
H
)
H2
CO2
CH4
CO
 
Figure  3.5: Effect of residence time on product gas yields (T, 600 ?C; P, 221 bar; feed 
conc., 10 wt% EtOH; 1.922g Ru/Al
2
O
3 
W/F
Ao
, 5 s:14,200g cat*s/g-mol EtOH, 10 s: 
25,200 cat*s/g-mol EtOH) 
 
 63
0
1
2
3
4
5
6
0123456
Residence time in Ru/Al
2
O
3
 bed (s)
Product
 Gas Yield (
m
ol gas/m
ol Et
OH
)
H2
CO2
CH4
CO
Figure  3.6: Effect of residence time on product yields (T, 700?C; P, 221 bar; feed conc., 
10 wt% EtOH; 1.922g Ru/Al
2
O
3
) 
 
0
1
2
3
4
5
6
01234567
Residence time in Ru/Al
2
O
3
 bed (s)
Product
 Gas Yield (
m
ol gas/m
ol Et
OH
)
H2
CO2
CH4
CO
 
Figure  3.7: Effect of residence time on product yields (T,  800?C; P, 221 bar; feed conc., 
10 wt% EtOH; 1.922g Ru/Al
2
O
3
) 
 
 64
3.4.3  Effect of Pressure.  The effect of pressure on the hydrogen yield was studied at 
700 ?C, 10 wt% EtOH in the feed with a residence time of 2 s in the catalytic bed.  As the 
pressure was raised from 221 bar to 276 bar, hydrogen yield was nearly constant, whereas 
CH
4
, CO, and CO
2
 yields increased from 0.73 to 0.88, 0.02 to 0.04, and 1.20 to 1.37, 
respectively.  This trend is illustrated in Figure 3.8. These results suggest that 
methanation reaction of CO and CO
2
 are slightly favored at higher pressures. Similar 
effect of pressure was also reported by Gadhe and Gupta
4
 during the SCW reforming of 
methanol.  
0
1
2
3
4
5
6
200 225 250 275 300
Pressure (bar)
Product
 Gas Yield (
m
ol gas/m
ol Et
OH
)
H2
CO2
CH4
CO
 
Figure  3.8:   Effect of reactor pressure (T, 700?C; feed conc., 2 s in Ru/Al
2
O
3 
bed; 
10wt% EtOH; 1.922 g Ru/Al
2
O
3
) 
 
3.4.4  Effect of Temperature.  Effect of temperature on H
2
 yield was studied by varying 
the reactor temperature from 600 to 800 ?C.  As shown in Figure 3.9, only a small 
amount of hydrogen is formed at low temperatures indicating that direct reformation 
reaction of ethanol in SCW is favored at high temperatures (>700 ?C). Two experiments 
 65
at 600 ?C showed incomplete conversion of ethanol.  The 5 and 10 s residence time 
experiments, detailed in Table 3.1, had conversions of 47.0 % and 98.4 %, respectively.  
With an increase in the temperature, the hydrogen and carbon dioxide yields increase, 
while the methane yield decreases.  The yield of carbon monoxide also decreased as 
temperature increased and it was significantly lower than that of other species.   
0
1
2
3
4
5
6
650 700 750 800 850
Temperature (?C)
Pr
oduc
t
 Ga
s
 Yi
e
l
d (
m
ol
 ga
s
/
mol
 Et
OH)
H2
CO2
CH4
CO
 
Figure  3.9:  Effect of reactor temperature (P, 221 bar; 2 s residence time in Ru/Al
2
O
3 
bed; feed conc., 10wt% EtOH; 1.922 g Ru/Al
2
O
3
) 
 
The high water excess leads to a preference for the formation of hydrogen and 
carbon dioxide instead of carbon monoxide.  The formed intermediate carbon monoxide 
reacts with water via the water?gas shift reaction to hydrogen and carbon dioxide.  The 
low carbon monoxide yield indicates that the water gas shift reaction approaches 
completion. The results indicate that water gas shift reaction is favorable at high 
temperature: 
 66
CO+ H
2
O ? CO
2
 + H
2
                                              (3.2) 
 
Due to this, the yield of H
2
 increased significantly with the increase in temperature in 
SCW.  It is also reported that the water gas shift reaction can take place non-catalytically 
in the supercritical reforming reactor.
37
 Accordingly, the methane yield was low 
indicating the occurrence of methane reforming reaction.  
 
CH
4
 + H
2
O ?   CO + 3 H
2
                                       (3.3) 
 
No measurable quantity of carbon on the catalyst was observed during the course of the 
experiments up to an ethanol concentration of 10 wt%. This is probably due to the high 
water-to-carbon ratios and relatively higher gasification activity of the catalyst.  
3.4.5  Effect of the Ethanol Concentration. The water-to-ethanol ratio is an important 
parameter for the economy of the process, which depends on the concentration of ethanol 
in the feed. To study the effect of the ethanol concentration on the H
2
 yield, experiments 
were conducted by varying mass percent of ethanol from 5 to 20 wt%.  The results, 
shown in Figure 3.10, show that there is a continuous decrease in the H
2
 and CO
2
 yields 
coupled with an increase in the CH
4
 yield as the ethanol concentration in the feed was 
increased.  
 67
0
1
2
3
4
5
6
0 5 10 15 20 25
Feed concentration (wt% EtOH)
Gas Y
i
el
d (m
ol
 gas/
m
o
l
 E
t
OH fed)
H2
CO2
CH4
CO
 
 
Figure  3.10:  Effect of feed concentration of product gas yields (T, 800?C; P, 221 bar; 
residence time 4 s in Ru/Al
2
O
3 
bed, feed conc., 10 wt% EtOH; 1.922 g Ru/Al
2
O
3
) 
 
It is well known that the methanation reaction is favored at high feed 
concentrations.  The high water-to-ethanol ratio shifts the equilibrium of reactions 
backward, leading to a decrease in the CH
4
 yield. Alternately, the higher amounts of 
water shifts the equilibrium of the water gas shift reaction to the right to produce more H
2
 
and CO
2
.  Probably at higher ethanol concentrations the active sites of the catalyst 
become saturated with adsorbed ethanol or derivative molecules. These adsorbed 
molecules prevent other next feed molecules from reaching the surface and reacting to 
form products. Dissociation of CO and H
2
 takes place to form the intermediates, which 
combine in steps to form CH
4
 (i.e., CO methanation reaction). At low water-to-ethanol 
ratios, the reaction shown below is suppressed due to low availability of water. 
C
2
H
5
OH + H
2
O  ?          4H
2
 + 2CO                                                (3.4) 
 68
 
Another reaction which is largely responsible for formation of lighter hydrocarbon such 
as methane, ethylene, and ethane is decomposition and dehydration of ethanol.  
 
C
2
H
5
OH   ?    CH
4
+CO+H
2
                                                                (3.5) 
 
The pyrolysis of lighter hydrocarbons such as CH
4
 and C
2
H
6
 occurs if the reaction is not 
properly controlled or if the residual molecules are not further reacted via an alternative 
chemical reaction. The repeated pyrolytic fragmentations of these may ultimately form 
coke via a variety of cyclization reactions. At high concentrations of ethanol ( > 10 wt%) 
in SCW, there was some formation of  carbon  inside the reactor. Hence, low 
concentration of ethanol (<10 wt%) is favorable for improved hydrogen yield and for 
suppressing the CO concentration in the product stream. 
3.4.6  Reaction Mechanism and Kinetics.  Ideally, one mole of ethanol fed to the 
reactor would produce 6 moles of H
2 
and 2 moles of CO
2
. However, the presence of other 
species in the product stream indicates that reforming of ethanol in SCW occurs via 
several pathways.  The following are the main possible reactions describing the overall 
ethanol reforming reaction in presence of Ru/Al
2
O
3
 catalyst.  
 
C
2
H
5
OH  + H
2
O       ?   CO
2
 + CH
4
 + 2H
2  
 (reforming )                           (3.6) 
 
CO + H
2
O               ?      CO
2
 + H
2
  (water gas shift)                                  (3.7)       
 
CH
4
 + H
2
O        ?    CO + 3 H
2         
   (reforming)                                      (3.8) 
 69
Since the above reactions together represent the overall steam reforming of ethanol, it is 
important to consider all the three reactions simultaneously while developing the kinetic 
model. Additional reactions such as ethanol dehydrogenation to acetaldehyde and ethanol 
dehydration to ethylene may also contribute.
25,26
  The elementary steps of ethanol SCW 
reforming reaction reveal that reforming occurs via formation of CH
3
CHO, C
2
H
4
, C
2
H
6
 
and CH
4
 as the reaction intermediates.  The initial step in the SCW reforming of ethanol 
is dehydrogenation, which is followed by the decomposition of CH
3
CHO to CH
4
 and CO, 
and finally CO is converted in to H
2
 by the water gas shift reaction due to the presence of 
excess SCW and the active ruthenium catalyst.  The formation of small amounts of 
ethylene during the reaction suggests the dehydration of ethanol.
38
 However, in the 
presence of SCW, all these intermediates such as CH
4
, C
2
H
4
, C
2
H
6
 and CH
3
CHO undergo 
reforming to produce CO
2
 and H
2
. The results show that H
2
 yield was higher at high 
temperature which happens at the expense of methane, ethylene and acetaldehyde. The 
mechanism of ethanol reforming is not available in presence of SCW. Based on 
theoretical calculations, Takahashi et al.
39
 reported the direct participation of water 
molecules through a multi-center transition state including an ethanol molecule and 
several water molecules bridging the oxygen atom of ethanol and the ?-hydrogen.  They 
proposed a transition state consisting of an ethanol molecule and two water molecules. 
The transfer of hydrogen atoms among three molecules produces a hydrogen molecule, 
two water molecules and resulted in an oxidized molecule, acetaldehyde. Therefore, It 
can be deduced based on the experimental results that there is a formation of intermediate 
products (ABS) such as acetaldehyde, because of reaction between adsorbed ethanol and 
 70
SCW in gas phase.
6,18,38,39
.  These adsorbed intermediates further react with water in the 
supercritical phase to give CO
2
 and H
2
.  The overall reaction steps may be rewritten as
6
: 
 
A + S    AS                                                              (3.9) ????
?11
,kk
AS+ B  ABS                                                                           (3.10) ???
2
k
ABS  Intermediates  CO
2
 + H
2
                                        (3.11) ???
3
k
???
4
k
 
Where A and B denote the ethanol and water species and S represents an active site.  
Equation 3.9 represents the reversible adsorption of ethanol on the catalyst surface, and 
Equation 3.10 represents the reaction of adsorbed ethanol molecule with water to form a 
complex molecule ABS. The formation of adsorbed intermediates by decomposition of 
ABS was considered as the rate controlling step. Assuming the steady state hypothesis for 
the intermediate complex ABS and AS, the dependence of the rate (r) on reactant 
concentration (c) can be expressed as: 
)(
3
21
211
21
k
cckk
ckckk
cckk
r
BA
BA
BA
+++
=
?
                                                   (3.12) 
 
Since in our experiments, water concentration is significantly high, c
B
 is assumed 
constant and the above expression is simplified as: 
A
AR
bc
ck
r
+
=
1
                                                                                            (3.13) 
 
where, k
R
 and b are lumped parameters defined as 
 71
 
Bo
Bo
R
bck
ckk
k
+
=
?1
21
                                                                                   (3.14)  
and 
Bo
Bo
ckk
k
ckk
k
b
21
3
21
1
)(
+
+
=
?
                                                                          (3.15) 
 
The above equation was solved using a nonlinear regression technique on rate of 
hydrogen production at different ethanol partial pressures and reaction temperature.  The 
values of rate constant k
R
 as calculated are reported in Table 3.2 with 95% confidence 
interval.  From the temperature dependence of the rate constants, the activation energy of 
the reaction was found to be 65.3 kJ mol
-1
.  This value of activation energy is 
significantly lower than the values reported (96 kJ mol
-1
) over Ru/Al
2
O
3
 catalyst,
18
 and 
(82.7 kJ mol
-1
)
 
over Co/Al
2
O
3 
catalyst
21
 for reforming of ethanol in subcritical water.  The 
lower value reported here may be attributed to the nature of the reforming media, 
supercritical water.  Low viscosity coupled with high diffusivity can overcome mass 
transfer limitations present in atmospheric steam reforming, resulting in faster rate of 
reaction and lower activation energy. 
 
 
 
 
 
 
 72
Table 3.2:  Values of rate constants at different temperatures 
T (?C) 
k
R 
(?mol/g cat-s-bar) 
 
 
                          800 
 
5.2?1.2 
 
  
750  
 
 
4.5?0.9 
 
700 
 
 
2.5?0.6 
 
 
3.5  Conclusions 
Supercritical reforming of ethanol over Ru/Al
2
O
3
 catalyst is effective for the 
production of high pressure hydrogen with low methane and carbon monoxide.  Full 
conversion of ethanol to gaseous products is seen above 700?C.  The hydrogen yield 
increases with increasing temperatures, and is unaffected by pressure changes in the 
supercritical region studied.  The methanation reaction can be reduced by keeping low 
residence times and a high water-to-ethanol ratio in the feed.  The product composition is 
affected by the reactor temperature, residence time and water-to-ethanol ratio in the feed 
and contained mainly hydrogen, carbon dioxide, methane, and carbon monoxide.  The 
complete conversion of ethanol is possible during supercritical water reforming over 
Ru/Al
2
O
3 
catalyst for producing high hydrogen yield with low concentrations of methane 
and carbon monoxide in product.  Carbon formation was negligible for ethanol 
concentrations below 10 wt%.  The activation energy for the reaction was found to be 
65.3 kJ mol
-1
.   
 
 
 
 73
3.6  Acknowledgment 
This research was partially supported by the U.S. Department of Energy through the 
Office of Fossil Energy (FE), National Energy Technology Laboratory (NETL), under 
DOE Contract DE-FC26-05424.56 via the Consortium of Fossil Fuel Science. 
 
 
 74
3.7  References 
1.        Torres, J.A.; Llorca,  J.; Casanovas, A.;  Dom?nguez, M.; Salvado J., Monta D., 
Journal of Power Sources, 2006, 169,( 1), 2007,  158-166. 
2.        Freni, S.; Calogero, G.; Cavallaro, S.  J. Power Sources 2000, 87 (1-2), 28-38. 
3.        Patel, S.; Pant K.K.  J. Power Sources 2006, 159 (1), 139-143. 
4.        Gadhe, J.B.; Gupta, R.B.; Ind. Eng. Chem. Res. 2005, 44, 4577-4585. 
5.        Pinkwart, K.; Bayha, T.; Lutter, W.; Krausa, M. J. Power Sources 2004, 136 (2), 
211-214. 
6.        Byrd, A. J.; Pant, K. K.; Gupta, R.B.  Ind. Eng. Chem. Res. 2007, 46, 3574-3579.  
 
7.        Antal, M. J., Jr.; Allen, S. G.; Schulman, D.; Xu, X.; Divilio, R. J. Ind. Eng. Chem. 
Res. 2000, 39 (11), 4040-4053. 
8.        Yoshida, T.; Oshima, Y.  Ind. Eng. Chem. Res. 2004, 43 (15), 4097-4104. 
9.        Watanabe, M.; Inomata H.; Arai, K.; Biomass and Bioenergy 2002, 22, 405-410. 
10.      Osada, M.; Sato,T.; Watanabe, M.; Adschiri,T.; Arai, K.  Energy & Fuels 2004, 
18, 327-333. 
 
11.      Williams P.; Onwundili J. Energy and Fuels, 2006, 20 (3), 1259-1265. 
12.      Cortright, R. D.; Davda, R. R.; Dumesic, J. A.  Nature 2002, 418, 964-967. 
 
13.      Penninger J. M. L.; Rep, M. Int. J. of Hydrogen Energy 2006, 31, 1597-1606. 
14.      Kruse, A.; Henningsen, T.; Sinag, A.; Pfeiffer, J.  Ind. Eng. Chem. Res. 2003, 42 
(16), 3711-3717. 
15.      Hao, X. H.; Guo, L. J.; Mao, X.; Zhang, X. M.; Chen, X. J.  Int. J. Hydrogen 
Energy 2003, 28 (1), 55-64. 
16.      Yu, D.; Aihara, M.; Antal, M. J., Jr.  Energy and Fuels 1993, 7 (5), 574-577.   
17.      www.sciencedaily.com/releases/2004/02/040214081412.htm, May 2007. 
18.      Vaidya, P.V.; Rodrigues, A.E.  Ind. Eng. Chem. Res.  2006, 45, 6614-6618. 
 
19.      Patel, S.; Pant, K.K. Chemical Eng. Sci. 2007, 62, 5425-5435 
 
20.      Xu, J.; Froment, G. F.   AIChE J. 1989, 35 (1), 88-96. 
 75
21.      Sahoo, D.R; Vajpai, S.; Patel, S.; Pant, K.K. Chemical Eng. Journal 2007, 125(3) 
139-147. 
 
22.      Fierro, V.; Akdim, O.; Mirodatos, C. Green Chem. 2003, 5, 20-24. 
 
23.      Berman, A.; Karn, R.K.; Epstein, M. Applied Cat. A 2005, 282, 73-83. 
 
24.      Liguras, D.K.; Kondarides, D.I.; Verykios, X.E. Applied Cat. B 2003, 43 345-354.  
 
25.     Arita, T.; Nakahara, K.; Nagami, K.; Kajimoto, O. Tetrahedron Lett. 2003, 44 
(5), 1083-1086. 
26.      Taylor, J. D.; Herdman, C.M.; Wu, B.C.; Wally, K.; Rice, S.F.  Int. J. Hydrogen 
Energy 2003, 28, 1171-1178. 
 
27.     Xu, X.; Matsumura, Y.; Stenberg, J.; Antal, M. J., Jr.  Ind. Eng. Chem. Res. 1996, 
35 (8), 2522-2530. 
28.     Boukis,?N.;?Diem,?V.;?Galla,?U.;?Dinjus,E.?Combust.?Sci.?and?Tech.?2006,?178,?
467?485. 
29.      Boukis,?N.;?Galla,?U.;?D?Jesus,?P.;?M?ller,H.;?Dinjus,E.?14th?European?
Conference?on?Biomass?for?Energy,?Industry?and?Climate?Protection?17?
21?October?2005,?Palais?des?Congr?s,?Paris,?France,?964?967.?
 
30.     Diem,?V.;?Boukis,?N.;?Hauer,?E.;?Franz,?G.;?Dinjus,?E.?7th?Italian?Conf.?on?
Supercritical?Fluids?and?their?Applications?Trieste,?Italy,?June?13?16,?2004.?
31.?????Boukis,?N.;?Galla,?U.;?Diem,?V.;?Dinjus,?E.?Science?in?Thermal?and?Chemical?
Biomass?Conversion,?30?August?to?2?September?2004,?Vancouver?Island,?BC,?
Canada,?2,?975?990. 
32.     Diem,?V.;?Boukis,?N.;?Hauer,?E.;?Dinjus,E.?Chemical?Engineering?Transactions?
2004,?4,?99?104.?
33.?????Boukis,?N.;?Galla,?U.;?Diem,?V.;?D?Jesus,?P.;?Dinjus,?E.;?2nd?World?Conference?
and?Technology?Exhibition?on?Biomass?for?Energy,?Industry?and?Climate?Protection?10??
14?May?2004,?Rome,?ETA?Florenz,?738?741. 
34.     Diem,?V.;?Boukis,?N.;?Habicht,?W.;?Dinjus,?E.?Proceedings?of?the?6th?
International?Symposium?on?Supercritical?Fluids?28?30?April?2003,?Versailles,?France,?
Tome?2.?
 
 76
35.     Diem,?V.;?Boukis,?N.;?Habicht,?W.;?Hauer,?E.;?Dinjus,?E.?The?Sixth?Italian?
Conference?on?Chemical?and?Process?Engineering,?8?11?June?2003,?Pisa,?Italy,?AIDIC?
Conference?Series,?Vol.?6,?101?108.??
 
36.     Boukis,?N.;?Diem,?V.;?Habicht,?W.;?Dinjus,?E.?Ind.?Eng.?Chem.?Res.?2003,?42,?
728?735. 
37.     Lee, S.; Lanterman, H.B.; Wenzel, J.E.; Picou, J.; Factor, M.; Leavitt, L. Preprints-
Am. Chem. Soc. Div. Petr. Chem 2006, 51, 487-490. 
38.    Xu X.; DeAlmedia, C.P.; Antal, M. Jr.; Ind. Eng. Chem. Res. 1991, 30, 1478-
1475. 
39.     Takahashi, H.; Hisaoki, S.; Nitta, T. Chem. Phys. Lett. 2002, 363, 80-86. 
 77
4.  Hydrogen Production from Glycerol by Reforming in Supercritical Water over 
Ru/Al
2
O
3
 Catalyst 
 
4.1  Abstract 
Supercritical water is a promising medium for the reforming of hydrocarbons and 
alcohols for the production of hydrogen at high pressures in a short reaction time.  Water 
serves both as a dense solvent as well as a reactant.  In this work, hydrogen is produced 
from glycerol by supercritical water reforming over a Ru/Al
2
O
3 
catalyst with low 
methane and carbon monoxide formation.  Experiments were conducted in a tubular 
fixed-bed flow reactor over a temperature range of 700-800?C, feed concentrations up to 
40 wt% glycerol, all at short reaction time of less than 5 seconds.  Glycerol was 
completely gasified to hydrogen, carbon dioxide, and methane along with small amounts 
of carbon monoxide. At dilute feed concentrations, near theoretical yield of 7 mole of 
hydrogen/mol of glycerol was obtained, which decreases with an increase in the feed 
concentration.  Based on a kinetic model for glycerol reforming, an activation energy of 
55.9 kJ/mol was observed. 
4.2  Introduction 
Growing energy demands in a time of dwindling fossil fuel supplies has attracted 
much attention to hydrogen as an energy carrier.  Biomass too has received significant 
attention as an alternate energy source because it is renewable and essentially a carbon 
dioxide neutral since CO
2
 generated during the fuel use is subsequently fixed by growing 
plants during photosynthesis.  Glycerol, HOCH
2
-CHOH-CH
2
OH, is obtained as a 
byproduct from biodiesel manufacturing by transesterification of vegetable oils. 
 78
Approximately 1 kg of glycerol is produced for every 9 kg of biodiesel.  In only year 
2004 to 2006, annual US biodiesel production has increased from 25 to 250 million 
gallons.
1
  Johnston and Holloway
2
 project a US biodiesel production potential of almost 
850 million gallons per year. With increasing production of biodiesel, an excess of 
glycerol is expected to flood the market, and therefore it is essential to find new uses.  
Currently, glycerol is used in many applications including personal care, food, oral care, 
tobacco, polymers, and pharmaceuticals.  Another use is the production of 1,2-
propanediol and 1,3-propanediol by hydrogenation of glycerol, replacing the petroleum 
feedstock used presently.
3
  Another emerging use is to utilize glycerol for the production 
of hydrogen.  The topic of glycerol reforming for hydrogen production has received some 
attention in the literature.
4-10
 Shabaker and Dumesic
6
 produced hydrogen from biomass-
derived oxygenated hydrocarbons including glycerol in an aqueous phase reforming 
process. Czernik and co-workers
7
 have produced hydrogen via steam reforming of crude 
glycerol using a commercial nickel based naphtha reforming catalyst.  Hirai et al.
8
 have 
reported the performance of noble metal based catalysts for glycerol reforming and found 
ruthenium to be the most active of the Group 8-10 metals. 
Recently, Adhikari
11
 et al. performed a thermodynamic analysis for hydrogen 
production by steam reforming of glycerol.  Their study revealed that the best conditions 
for producing hydrogen is at a temperature > 625 
o
C and a molar ratio of water to 
glycerol of 9:1. Under these conditions methane production is minimized and carbon 
formation is thermodynamically inhibited. Although excess water in the feed increases 
the hydrogen production, a significant amount of unreacted water remains in the reactor 
effluent. Corma et al.
12
 have studied the biomass derived feeds, glycerol and sorbitol, 
 79
using fluidized catalytic cracking and reported a wide range of products including 
hydrogen, paraffins, olefins, aromatics, and coke. 
Supercritical water (SCW) reforming of hydrocarbons and biomass materials has 
been paid more attention recently.
13-18
  Kersten et al.
19
 have reported gasification results 
for glycerol and other model compounds in a variety of catalytic and non-catalytic 
reactors in SCW and found that without addition of a catalyst, only very dilute 
concentrations of model biomass feeds could be completely gasified.  SCW has 
properties entirely different from those of liquid water or steam. The dielectric constant 
of SCW is much lower, and the number of hydrogen bonds is much lower and their 
strength is weaker. Therefore SCW has high miscibility with many organic solvents and 
gases.  The density of SCW is higher than that of steam resulting in a higher space time 
yield, and higher thermal conductivity and specific heat, which are helpful in carrying out 
the endothermic reforming reactions. Transport properties, too, are unique in that SCW 
has both low viscosity and high diffusivity.  The formation of char and tar is also 
minimized because of the solubility of hydrocarbons in SCW. Importantly, hydrogen 
produced from SCW reforming is produced at high pressure, which can be stored 
directly, thus avoiding large expenses associated with compression. 
Reforming of glycerol for hydrogen production can be summarized by following 
reactions.   
First, the steam reforming of glycerol: 
C
3
H
8
O
3 
 3 CO + 4 H
2
                                            (4.1) ????
OH
2
Followed by the water-gas shift reaction: 
CO + H
2
O   CO
2
 + H
2
                                               (4.2) ?
 80
The desired overall reaction is then summarized as: 
C
3
H
8
O
3
 + 3 H
2
O ?  7 H
2
 + 3CO
2
                                         (4.3) 
Some hydrogen is also lost via the methanation of CO and CO
2
: 
CO + 3 H
2
 ?  CH
4
 + H
2
O                                               (4.4) 
 
CO
2
 + 4 H
2
 ?  CH
4
 + 2H
2
O                                            (4.5) 
 
 As a result, the product stream is a mixture of above gases. Furthermore, the yield of 
hydrogen depends on several process variables, such as system pressure, temperature, and 
water-to-glycerol feed ratio.  
Most of the above studies were done in batch mode, in which the 
biomass/water/catalyst is loaded in a small steel tube reactor and then sealed and placed 
in an oven. After the reaction, the mixture is quenched and analyzed. Typical reaction 
time varied from minutes to hours. In our recent work on methanol and glucose 
reforming, it was observed that high reaction time leads to the secondary reaction of 
methane formation.
13-14
 To limit the methane formation reaction time needs to be limited 
to the order of seconds. The aim of this study is to examine hydrogen production from 
glycerol in SCW in a continuous reactor with a short reaction time (of the order of 
seconds). Ru/Al
2
O
3
 is selected as a catalyst. The effects of the process variables such as 
temperature, contact time, and water to glycerol ratio on hydrogen yield are investigated.  
4.3  Experimental Section 
4.3.1  Materials.  Glycerol (99.5% purity) was obtained from Fisher Scientific.  The 
commercial 5 wt% Ru/Al
2
O
3 
catalyst was purchased unreduced from Aldrich, and was 
 81
subsequently reduced under hydrogen flow at 500 ?C for 6 hours. The catalyst had the 
following characteristics: total BET surface area: 100 m
2
/g; crystal structure: amorphous; 
specific pore volume: 0.30 mL/g; density: 0.95 g/cm
3
. Deionized water was obtained 
from a Millipore Direct-Q 5 water purification system. 
4.3.2  Experimental Procedure.  The schematic diagram of the experimental apparatus 
is shown in Figure 4.1. The details of the experimental procedure are discussed 
elsewhere
14
, and are briefly provided here.  Experiments were conducted in supercritical 
water in a fixed bed tubular flow reactor (0.5 m long, 0.25? OD and 0.12? ID) made of 
Inconel 600 (Microgroup) having a composition of 73% Ni, 18% Cr, and 9% Fe, which 
was placed inside a temperature controlled furnace (Thermolyne 21100) covered on each 
end to reduce heat loss and ensure temperature uniformity.  The reactor was packed with 
2 g catalyst. The details of experimental conditions are given in Table 4.1. 
A glycerol-water solution from a feed reservoir was pumped to a tee using a 
HPLC pump (Waters 590).  Due to the hygroscopic nature of glycerol, batches of pure 
glycerol were diluted with water to 66.7 wt% and kept covered to prevent absorption of 
additional water from the atmosphere.  The glycerol solution then mixes with a stream of 
supercritical water, also by HPLC pump (Alltech 301), heated to the reaction 
temperature.  The reactor temperature at the exit of the furnace was measured using a 
type-K thermocouple with a tee arrangement.  The gas mixture exiting the reactor was  
cooled using a water-cooled double pipe heat exchanger made of SS 316 tubing.  The 
pressure was let down to the ambient by means of a back pressure regulator (Straval).   
 
 82
 
 
Diluted 
Glycerol 
T
P 
HPLC pumps 
Water-cooled 
heat exchanger 
 
GC with 
TCD 
6-port injection 
valve for online 
sampling 
Gas flow 
meter 
Tubular reactor 
Tube furnace with 
temperature controller 
T 
Back 
pressure 
regulator
Gas- Liquid 
separator 
 
 
Data acquisition 
system 
Data analysis with 
software 
 
TOC analyzer 
for liquid phase 
Gas flow 
totalization 
Water preheater 
Mix pt. 
 
Water  
Figure  4.1:  Schematic of SCW hydrogen production apparatus 
 
Table 4.1:  Details of experimental conditions 
Reactor Reactor Feed Residence W/F
Ao
     H
2
 Yield 
Temp.
a
 Pressure Conc. Time g cat-s/ 
Product Gas Composition 
(mol%) mol gas/ 
?C bar wt% s 
?mol 
glycerol H
2
 CO CH4 CO
2
 
mol glycerol 
fed 
800 241 5 4 0.353 45.3 0.0 19.6 35.2 2.0 
800 241 5 2 0.221 49.4 0.0 15.9 34.7 2.8 
800 241 5 1 0.137 70.0 1.1 3.7 25.2 6.5 
800 241 15 1 0.050 57.9 0.6 10.7 30.9 4.1 
800 241 20 1 0.035 53.5 2.2 11.5 32.9 3.9 
800 241 30 1 0.023 47.2 3.2 15.6 34.0 2.9 
800 241 35 1 0.022 46.5 3.8 16.8 32.9 2.6 
800 241 40 1 0.019 42.2 4.3 18.9 34.6 2.2 
750 241 2.5 1 0.243 66.9 0.1 3.9 29.1 5.8 
750 241 5 1 0.137 66.2 1.2 3.6 29.0 6.1 
750 241 15 1 0.045 56.9 2.0 9.6 31.5 4.4 
750 241 30 1 0.023 46.5 3.4 15.8 34.4 2.6 
700 241 5 1 0.134 63.5 0.7 6.7 29.1 5.1 
a
 Uncertainty ?3?C 
 83
The gas-liquid mixture was separated in a glass phase-separator having gas tight 
valves to prevent the escape of gases.  The exit gas flow rate was measured using a gas 
flow meter (Omega FMA-1600).  The gas composition was determined using a gas 
chromatograph (SRI 8610C) equipped with a thermal conductivity detector (TCD) and a 
60/80 Carboxen-1000 carbon molecular sieve column (Supelco 15?x1/8?) using nitrogen 
as the carrier gas. Sample injection to the gas chromatograph was done online by means 
of a six-port injection valve (Valco) having a 100 ?L sample loop. The total organic 
carbon (TOC) content of the liquid effluent was measured using a TOC analyzer 
(Shimazdu TOC-V
CSN
). 
All measurements were taken at least in triplicate to ensure accuracy, which were 
further checked by calculating the overall carbon balance for the system. Scattering in the 
data of the totalized gas flow rate measured by the flow meter was less than 1%.  The 
error in the dry gas composition obtained by the GC analysis was typically less than 2%.  
The overall error in the calculation of gas yields due to the errors introduced by the 
individual analysis techniques and experimental error was found to be less than 5%.  
Experimental results in which the carbon exiting the system was calculated to differ by 
more than 10% of the known amount entering were discarded. 
4.4  Results and Discussion 
Production of hydrogen and other compounds at different temperatures, water to 
glycerol feed ratios, and residence times have been analyzed. Supercritical reforming of 
glycerol produced a stream rich in H
2
 and CO
2
 with small amounts of CH
4
 and CO.  Over 
the temperature, residence times, and water to glycerol mass ratio ranges analyzed, the 
conversion of glycerol was always greater than 99%, and it can be considered that the 
 84
conversion was complete.   Our previous study
20
 showed that changing pressure has 
minor effect in the supercritical region, thus only one pressure was studied. 
Thermodynamic equilibrium calculations were done by minimizing Gibbs free energy 
using Peng-Robinson equation of state in CHEMCAD 5.2.0. The calculated results are 
shown as solid lines in Figures 4.2-4.4. The simulation did not predict coke formation for 
any of the experimental conditions in this paper. 
4.4.1  Effect of Reaction Time.  Residence time in the catalyst bed was controlled by the 
inlet flowrate of reactants.  The product gas yields are shown as a function of residence 
time in Figure 4.2.  The residence times of 1, 2 and 4 seconds correspond to space time 
(mass of catalyst in bed/molar flow rate of glycerol, W/F
Ao
) values of 0.136, 0.221, and 
0.353 g cat-s/?mol glycerol, respectively.  The shortest residence time gave high 
hydrogen yield, however at larger residence times the hydrogen yield drops sharply with 
a decline in CO
2 
yields as well.  Considering this along with the increase in methane 
suggests that Equation 4.5, the methanation of carbon dioxide, becomes important at 
longer residence times. Hence, it is desirable to keep residence time short in order to 
maximize the hydrogen yield.   
 85
0
1
2
3
4
5
6
7
12345
Residence Time (s)
Product Gas Yield (
m
ol
 gas/mol glycerol 
fed
)
H2
CO2
CH4
CO
H2 equil
CO2 equil
CH4 equil
CO equil
Hydrogen
 
Figure  4.2:  Effect of residence time on product gas yields. T: 800?C, P: 241 bar, feed 
conc.:5 wt% glycerol, 2.0 g Ru/Al
2
O
3
 catalyst 
 
4.4.2  Effect of Temperature.  The influence of temperature was studied by varying the 
furnace temperature from 700 to 800?C.  Hydrogen yield increases with temperature as 
shown in Figure 4.3.  It can also be seen that the methane yield is slightly higher at 
700?C.  The same trend is present in the calculated equilibrium concentration with 
slightly more methane present at lower temperatures.  It should be noted that continued 
operation at 700 ?C for feed concentrations greater than 5 wt% glycerol resulted in 
plugging of the reactor, however this problem was solved by operating at 800 
o
C.  At 
lower temperatures, the reaction rates for reactions leading to coke formation are higher 
than the rates of the reforming and carbon gasification reactions.
21
  At temperatures 
below 800?, the experimental hydrogen yield is lower than the equilibrium values, but is 
accompanied by higher than predicted methane yield. 
 86
0
1
2
3
4
5
6
7
650 700 750 800 850
Temperature (?C)
P
r
od
uct Gas Y
i
el
d ( m
o
l
 gas/
m
o
l
 gl
y
cerol
 fed)
H2
CO2
CH4
CO
H2 equil
CO2 equil
CH4 equil
CO equil
Hydrogen
 
 
Figure  4.3:  Effect of temperature on product gas yields. P: 241 bar, 1 s residence time, 
feed conc.: 5 wt%, 2.0 g Ru/Al
2
O
3
  
 
4.4.3  Effect of Glycerol Feed Concentration.  Feed concentration is an important 
economic consideration, it is not desirable to heat and pump more water through the 
system than is necessary.  To study the effect of feed concentration, the glycerol 
concentration was varied from 5 ? 40 wt% glycerol.  Figure 4.4 shows that increasing the 
feed concentration was coupled with a decrease in the yield of hydrogen and an 
accompanying increase in the methane yield.  This can be explained by considering that 
less water is present at the higher concentrations, and it is known that at low 
steam/carbon ratios CO is more likely to produce methane by consuming hydrogen.
22
  
The carbon balance showed that complete conversion of glycerol to gaseous products was 
realized even for the highest feed concentrations tested.  It can be seen that the 
experimental gas yields closely mirror the equilibrium concentrations calculated from 
 87
simulation, indicating that the reaction is near its thermodynamic equilibrium.  For higher 
concentrations, experimental carbon monoxide yields are smaller than predicted at 
equilibrium.  Paired with higher than predicted hydrogen and carbon dioxide yields, this 
indicates the water-gas shift reaction going near completion. 
0
1
2
3
4
5
6
7
1020304050
Feed Concentration (wt% glycerol)
Pr
oduct
 Gas Yi
el
d (
m
ol
 g
as/
mol
 gl
ycerol
 
f
ed)
H2
CO2
CH4
CO
H2 equil
CO2 equil
CH4 equil
CO equil
Hydrogen
 
Figure  4.4:  Effect of glycerol feed concentration on product gas yields.  T: 800?C, P: 
241 bar, 1 s residence time, 2.0 g Ru/Al
2
O
3
 catalyst 
 
4.4.4  Reaction Kinetics.  The reforming of oxygenated hydrocarbons entails a complex 
network of reactions. Adsorption and decomposition of oxygenated compounds 
containing hydroxyl groups such as methanol and ethylene glycol has been studied 
extensively on noble metal surfaces.  Oxygenated compounds containing hydroxyl 
groups have been shown to adsorb to the catalytic Ru surface predominantly through one 
or more oxygen atoms.  First the reactant undergoes dehydrogenation on the catalyst 
surface, followed by subsequent cleavage of C-C or C-O bonds.  Cleavage of C-C bonds 
leads to synthesis gas which is subjected to the water-gas shift reaction and possible 
 88
methanation, while cleavage of C-O bonds gives organic acids and alcohols.
3
   The 
ruthenium catalyst is known to have a high activity for C-C bond scission.
23
  Very low 
levels of organic carbon in the liquid effluent suggests that any intermediate alcohol or 
organic acids formed from C-O bond breaking were further reacted to gaseous products.  
Hence, the following kinetic model is proposed for the reforming of glycerol in 
supercritical water: 
 
A + S    AS                                                              (4.6) ????
?11
,kk
AS+ B  ABS                                                                           (4.7) ???
2
k
ABS  Intermediates  CO
2
 + H
2
                                        (4.8) ???
3
k
???
4
k
 
where, A and B represent glycerol and water, respectively, and S represents an active site 
on the catalyst surface.  Equation 4.6 describes the reversible adsorption of glycerol on 
the catalyst, and Equation 4.7 represents the reaction of the adsorbed glycerol with water 
to form an adsorbed complex molecule, ABS.  Assuming steady state of adsorbed 
intermediates and that decomposition of ABS to form intermediate products is the rate-
limiting step, the dependence of rate (r) on partial pressures (p) can be expressed as: 
 
)(
3
21
211
21
k
ppkk
pkPkk
ppkk
r
BA
BA
BA
+++
=
?
                                            (4.9) 
Considering that water is in excess, Equation 4.9 may be simplified to: 
A
AR
bp
pk
r
+
=
1
                                                               (4.10) 
 89
where k
R
 and b are lumped parameters defined as: 
Bo
Bo
R
bpk
pkk
k
+
=
?1
21
                                                        (4.11) 
and 
Bo
Bo
pkk
k
pkk
k
b
21
3
21
1
)(
+
+
=
?
                                                         (4.12) 
Equation 4.10 was solved using a regression on rate of hydrogen production for several 
values of partial pressures of glycerol at 750 and 800?C.  Calculated values of k
R
 are 1.9 x 
10
-5
 and 2.6 x 10
-5
 ?mol/g cat-s-bar at 750?C and 800?C, respectively.  Based on the 
temperature dependence of the rate constant, activation energy E
a
 for the reaction was 
found to be 55.9 kJ/mol.  Glycerol reforming experiments were continued to evaluate the 
catalyst activity loss due to carbon formation. There was negligible change in reaction 
rate and hydrogen product yield as same catalyst was used for all the runs without any 
regeneration.  Dispersion of Ru was not measured, but it is expected that it should 
decrease under the operating conditions used.  Kellner and Bell
24
 have reported 
previously that decreased dispersion can result in increased methane formation.  Our 
recent study
18
 found changes to the catalyst support after exposure to supercritical water.  
SEM images and XRD spectra showed that crystallinity increased in our previous study 
using this amorphous catalyst support, as shown in Figure 4.5. The fresh catalyst consists 
of more-or-less round particles, characteristic of amorphous materials, whereas the used 
catalyst is clustered in aggregates of sharp-edged crystals.  It is expected that the catalyst 
support underwent similar morphological changes in the present study. 
 90
 
(a) 
 
 
(b) 
Figure  4.5:  SEM images of fresh (a) and used (b) Ru/Al
2
O
3
 catalysts. 
 
 91
4.5  Conclusions 
Reforming of glycerol in supercritical water over Ru/Al
2
O
3
 catalyst is an effective 
means of high pressure hydrogen production from a biomass-derived source.  Although 
near-theoretical hydrogen yields were obtained for dilute glycerol concentrations at 
800?C, it was also found to be possible to completely gasify feed containing upto 40 wt% 
glycerol, but with increased methane formation.  Hydrogen yields were found to increase 
directly with temperature.  Methane formation can be reduced by operating at low 
residence times.  Values of the rate constant based upon a simplified kinetic model are 
reported at 750 and 800?C, and based upon those an activation energy of 55.9 kJ/mol was 
observed. 
 
4.6  Acknowledgment 
This research was partially supported by Alabama Center for Paper and Bioresource 
Engineering, and the U.S. Department of Energy through the Office of Fossil Energy 
(FE), National Energy Technology Laboratory (NETL), under DOE Contract DE-FC26-
05424.56. 
 
 
 
 
 
 
 
 92
4.7  References 
1.       < http://www.biodiesel.org/pdf_files/fuelfactsheets/backgrounder.PDF> accessed 
29 February 2008. 
 
2.        Johnston, M and Holloway, T. A Global Comparison of National Biodiesel 
Production Potentials. Env Sci and Tech 2007; 41: 7967-73. 
 
3.        Montassier, C, Menezo, JC, Hoang, LC, Renaud, C, Barbier, J. Aqueous polyol 
conversions on ruthenium and on sulfur-modified ruthenium. J. Molecular Catalysis 
1991; 70: 99-110 
 
4.        Cortright RD, Davda RR, Dumesic JA. Hydrogen from catalytic reforming of 
biomass-derived hydrocarbons in liquid water. Nature 2002;418:964?6. 
 
5.        Davda RR, Shabker JW, Huber GW, Cortright RD, Dumesic JA. A review of 
catalytic issues and process conditions for renewable hydrogen and alkanes by aqueous 
phase reforming of oxygenated hydrocarbons over supported metal catalysts. App Cat B: 
Environ 2005;56: 171?86. 
 
6.        Shabaker JW, Dumesic JA. Kinetics of aqueous-phase reforming of oxygenated 
hydrocarbons: Pt/Al2O3 and Sn-modified Ni catalyst. Ind Eng Chem Res 2004; 43(12): 
3105? 12. 
 
7.        Czernik S, French R, Feik C, Chornet E. Hydrogen by catalytic steam reforming of 
liquid byproducts from biomass thermoconversion process. Ind Eng Chem Res 2002;41: 
4209?15. 
 
8.    Hirai, T, Ikenaga N, Miyake T, Suzuki T. Production of Hydrogen by Steam 
Reforming of Glycerin on Ruthenium Catalyst. Energy and Fuels 2005;19: 1761-62. 
 
9.        Zhang B, Tang X, Li, Y, Xu Y, Shen W. Hydrogen production from steam 
reforming of ethanol and glycerol over ceria-supported metal catalysts. Int J of Hyd 
Energy 2007; 32: 2367-73 
 
10.      Simonetti D A, Kunkes E L, Dumesic J A. Gas-phase conversion of glycerol to 
synthesis gas over carbon-supported platinum and platinum-rhenium catalysts. Journal of 
Catalysis 2007; 247:  298-306. 
 
11.      Adhikari S, Fernando S, Haryanto A. A Comparative Thermodynamic and 
Experimental Analysis on Hydrogen Production by Steam Reforming of Glycerin. 
Energy and Fuels 2007; 21: 2306-10 
 
 
 93
12.      Corma A, Huber G, Sauvanaud L, O'Connor, P. Processing biomass-derived 
oxygenates in the oil refinery: Catalytic cracking (FCC) reaction pathways and role of 
catalyst. Journal of Catalysis 2007; 247: 307-27.  
 
13.      Gadhe JB, Gupta RB. Hydrogen Production by Methanol Reforming in 
Supercritical Water: Suppression of Methane Formation. Ind Eng Chem Res 2005; 44:  
4577-85 
 
14.       Byrd AJ, Pant KK, Gupta RB. Hydrogen Production from Glucose Using 
Ru/Al2O3 Catalyst in Supercritical Water. Ind Eng Chem Res 2007;46: 3574-79. 
 
15.       Penninger JML, Maass GJJ, Rep M.   Compressed hydrogen-rich fuel gas 
(CHFG) from wet biomass by reforming in supercritical water. Int J Hyd Energy 
2007;32: 1472-76.   
 
16.       Penninger JML, Rep M. Reforming of aqueous wood pyrolysis condensate in 
supercritical water. Int J Hyd Energy 2006;31: 1597-1606. 
 
17.       Lu Y, Guo L, Zhang X, Yan Q.   Thermodynamic modeling and analysis of 
biomass gasification for hydrogen production in  supercritical  water. Chem Eng J 
2007;131: 233-44.   
 
18.       Lu YJ, Guo LJ, Ji CM, Zhang XM, Hao XH, Yan QH. Hydrogen production by 
biomass gasification in supercritical water: A parametric study. Int J Hyd Energy 
2006;31: 822-31.  
 
19.       Kersten SRA, Potic B, Prins W, Van Swaaij WPM. Gasification of Model 
Compounds and Wood in Hot Compressed Water. Ind Eng Chem Res 2006;45: 4169-77. 
 
20.       Byrd, AJ, Pant KK, Gupta RB. Hydrogen Production from Ethanol by Reforming 
in Supercritical Water using Ru/Al
2
O
3
 Catalyst.  Energy and Fuels 2007; 21: 3541-47 
 
21.       Coll, R, Salvado, J, Farriol, X, Montane, D. Steam reforming model compounds 
of biomass gasification tars: conversion at different operating conditions and tendency 
towards coke formation.  Fuel Processing Tech 2001; 74: 19-31. 
 
22.      Xue, E, O?Keeffe, M, Ross, JRH. Water-gas shift conversion using a feed with a 
low steam to carbon monoxide ratio and containing sulfur. Cat Today 1996; 30:107-18. 
 
23.      Sinfelt, John H.; Yates, David J.C. Catalytic hydrogenolysis of ethane over the 
noble metals of Group VIII J. Catalysis 1967;8: 82-90. 
 
24.   Kellner, C and Bell, A. Effects of dispersion on the activity and selectivity of 
alumina-supported ruthenium catalysts for carbon monoxide hydrogenation. J Catal 1982; 
75:251-61. 
 94
5.  Hydrogen Production from Catalytic Gasification of Switchgrass Biocrude in 
Supercritical Water 
 
 
5.1 Abstract 
 Biomass can be liquefied to produce biocrude for the ease of transportation and 
processing.  Biocrude contains oxygenated hydrocarbons of varying molecular structure 
and molecular weights, including lignin derived products, sugars and their decomposition 
products.  In this work several catalysts were screened for hydrogen production by 
gasification of switchgrass biocrude in supercritical water at 600?C and 250 bar.  Nickel, 
cobalt, and ruthenium catalysts were prepared and tested on titania, zirconia, and 
magnesium aluminum spinel supports.   Magnesium aluminum spinel was seen to be an 
inappropriate support as reactors quickly plugged.  Ni/ZrO
2
 gave 0.98 mol H
2
/mol C, the 
highest hydrogen yield of all tested catalysts; however, over time, increase in pressure 
drop lead to reactor plugging with all zirconia supported catalysts.  Titania supported 
catalysts gave lower conversions, however they did not plug during the course of the 
study.   Charring of all catalysts was seen to occur at the entrance of the reactor as the 
biocrude was heated.  All support materials suffered significant surface area loss due to 
sintering.   
 
5.2 Introduction 
 Hydrogen has been considered as a potential energy carrier, however 95% of all 
hydrogen production currently relies on fossil fuel feedstocks via the thermo-catalytic 
and gasification processes of methane, napthas, and coal.
1
  Concerns over carbon 
 95
emissions and depletion of fossil fuel reserves have led many researchers to investigate 
carbon neutral renewable biomass feedstocks for hydrogen production.
2-4
  A recent 
review comparing the economic feasibility of several hydrogen production strategies 
found hydrogen production from the thermochemical processing of biomass to have a 
lower cost of hydrogen per kilogram than current wind, solar, or nuclear technologies.
5
 
Traditional gasification technologies require several additional steps not necessary in the 
supercritical water gasification scheme presented here.  For example, the energy intensive 
grinding and drying of biomass prior to gasification is not required with sub- or 
supercritical water (SCW) technologies, and the amount of tarry material is greatly 
reduced, owing to SCW?s ability to solubilize nonpolar tar precursor compounds.
6
   
Further, hydrogen is obtained directly at high pressure, eliminating the need for 
compression of the gas before storage. 
In this work we use a feedstock of aqueous oxygenated hydrocarbons termed 
biocrude, where switchgrass was partially solubilized in sub-critical water. In the sub-
critical region water has an increased ionic character and the preferred reaction pathway 
includes the formation of phenols and furfurals.
7
   Details of the biocrude production 
have been published by  Kumar and Gupta.
8
   In the gasification of biocrude the 
maximum theoretical hydrogen yield would be attained by suppressing methanation and 
having the water-gas shift go to near completion, giving approximately 2 mol H
2
/ mol C.  
Examples of the desired reactions are shown in Equations 1-3 for glucose, furfural, and 
catechol, respectively:  
C
6
H
12
O
6
 + 6 H
2
O ? 12 H
2
 + 6 CO
2
             Yield: 2 H
2
/C          (5.1) 
C
5
H
4
O
2
 + 8 H
2
O ? 10 H
2
 + 5 CO
2
             Yield: 2 H
2
/C          (5.2) 
 96
C
6
H
6
O
2
 + 10 H
2
O ? 13 H
2
 + 6 CO
2
          Yield: 2.17 H
2
/C        (5.3) 
This approach of using solubilized biomass instead of whole biomass has received 
some attention in the literature.  Matsumura et al. liquefied cellulose and cabbage at 
150?C under autogeneous pressure and gasified the aqueous product with partial 
oxidation at 400?C and 25 MPa in a batch system over a commercial Ni catalyst and 
several mixed oxides from the lanthanide series.  The oxides did not effectively catalyze 
gasification or oxidation of tars, whereas the nickel catalyst gave a 68% gasification 
efficiency and was active in both the water-gas shift (WGS) reaction and methanation.
13
  
Elliott et al. have taken a similar approach, liquefying several biomass feedstocks then 
gasifying them at 350?C and 20 MPa over Ru/C catalyst, obtaining a product gas 
composed mostly of methane and carbon dioxide.  They studied continuous systems, both 
on the bench scale as well as a scaled up mobile reactor system.  They encountered 
problems with reactor plugging, as well as catalyst deactivation from trace elements 
present in the biomass feed.
14
  Recently another approach has been taken by studying the 
aqueous phase reforming of hydrolyzed wheat straw in a batch reactor and the results 
showed Raney nickel to be more active than supported noble metal catalysts.
15
 
Osada et al. have previously found Ru/TiO
2
 catalysts to be active and stable in the 
gasification of lignin in a batch reactor at 400?C and 371 bar.  They achieved complete 
gasification in 120 minutes, obtaining a gas rich in CO
2
 and CH
4
.  Both Ru/C and Ru/?-
Al
2
O
3
 were seen to lose surface area due to gasification or phase transformations of the 
support, respectively.
9
  Yu and Savage reported that after 100 h in supercritical water 
oxidation service at 380?C and 250 atm a mixed rutile/anatase support had completely 
transformed to rutile.
10
  Elliott et al. found that commercial titania tablets lost their 
 97
mechanical strength after water exposure at 350?C and 200 bar and also report all anatase 
titania being transformed to the rutile form, while monoclinic zirconia was reported to be 
stable.
11, 12
  
The formation of tar and char materials from biomass components and model 
compounds has been investigated by several researchers.  For example, Sato et al. studied 
the noncatalyzed alkylation of phenol by alcohols and aldehydes in SCW.
16-18
  Saisu et al. 
later expanded on this work to include the repolymerization of lignin decomposition 
products by cross-linking with formaldehyde, as well as between the decomposition 
products themselves.
19
  Chuntanapum and Matsumura have recently clarified the role of 
5-HMF in tarry materials formation, finding polyaromatic char formation occurring only 
in the subcritical condition, resulting from polymerization of 5-HMF and it aromatic 
degradation products.
20, 21
 
In our previous work we have gasified model and biomass derived compounds.
22-
24
  Here we will extend the field of knowledge on the catalytic gasification of real 
liquefied biomass for hydrogen production, studying nickel, cobalt, and ruthenium 
catalysts on titania, zirconia, and magnesium aluminum spinel supports.  A focus of this 
work is to understand the activity of the catalyst and stability of the support materials in 
supercritical water. 
 
5.3 Experimental 
5.3.1 Catalyst preparation 
TiO
2
 and ZrO
2
 supports were supplied by St. Gobain Norpro as pellets.  MgAl
2
O
4
 
was prepared by coprecipitation from aluminum nitrate nonahydrate (98%, Sigma 
 98
Aldrich) and magnesium nitrate hexahydrate (99%, Sigma Aldrich) following the method 
of Bocanegra et al.
25
  All supports were crushed and sieved to particle sizes between 150 
? 600 ?m.  Metals were loaded onto the supports by an incipient wetness technique.  
Precursor materials for metals were nickel nitrate hexahydrate (98%, Alfa Aesar), cobalt 
nitrate hexahydrate (98%, Alfa Aesar), and ruthenium (III) nitrosyl nitrate (1.5% Ru, 
Strem Chemicals).  Following impregnation, catalyst samples were dried at 110?C 
overnight, followed by calcination in air at 500?C for 4 h.  Prior to each experiment, a 
fresh 4 g portion of catalyst was reduced in situ at 500?C for 2 h in a stream of 5% H
2
 in 
helium flowing at 0.2 Nl/min.  All nickel and cobalt catalysts were prepared to have 10 
wt% metals loading, while the ruthenium catalysts had a 1.5 wt% metals loading.  A 
smaller metal loading was used with ruthenium catalysts due to concern over the high 
cost of ruthenium. 
 
5.3.2 Feedstock preparation 
Switchgrass was treated in subcritical water at 300?C and 100 bar for 30 minutes 
with no additional catalyst in a batch reactor to give an aqueous solution of oxygenated 
hydrocarbons termed biocrude.  Biocrude is not extremely stable and some carbon 
precipitated during refrigerated storage, as has been noted by other researchers.
21
  
Precipitated particles were removed from the biocrude prior to a run by vacuum filtration 
through a 2.5 ?m filter paper.  Depending on the age of the biocrude it contained 1.0-1.3 
wt% carbon, and an approximate composition is given in Table 5.1.  For details of the 
experimental procedure and characterization of solid and liquid products, see Kumar and 
Gupta.
8
   
 99
 
Table 5.1:  Oxygenated Hydrocarbons Present in Biocrude 
 
Components identified by HPLC wt.% 
     5 & 6 Carbon Sugars 15 
     5-HMF 4 
     Furfural 6 
     Organic Acids (lactic, formic, acetic) 17 
Other components identified by GC-MS 57 
     2-furancarboxaldehyde  
     1,2-benzenediol  
     4-hydroxyvanillin  
     2,3-dihydrobenzofuran  
     2-methoxy-4-vinylphenol  
     1,4-benzenediol  
     2-methylphenol  
     2,6-dimethoxybenzaldehyde  
     homovanillyl alcohol  
     4-hydroxy-3,5-dimethoxy-benzaldehyde  
 
 
5.3.3 Experimental Procedure 
 A schematic of the experimental apparatus used for hydrogen production is shown 
in Figure 5.1.  After the catalyst reduction period, the reactor and apparatus were 
pressurized with water while the temperature of the furnace was raised such that a 
thermocouple at the midpoint of the 6.4 mm OD x 3.2 mm ID x 50 cm  Inconel 600 
reactor read 600?C.  After one hour, the water feed to the reactor was replaced with 
biocrude fed directly to the reactor with no preheating at 0.6 ml/min, resulting in a weight 
hourly space velocity (WHSV) of 9 h
-1
. Thirty minutes were allowed to elapse between 
the introduction of the biocrude feed and the start of product sampling.  The reactor 
effluent was quenched by a double pipe heat exchanger with water on the shell side.  
Pressure control in the system was accomplished by two back pressure regulators in 
series, the first set at 250 bar and the second at 100 bar. This arrangement reduced 
pressure fluctuations in the reactor caused by two phase flow through the back pressure 
 100
regulator.  Product gases were separated from the liquid effluent in a glass phase 
separator before online sampling by a GC equipped with TCD (SRI 8610C).  The carbon 
content of the liquid effluent was measured by a TOC analyzer (Shimadzu TOC-V
CSN
), 
and gas flow was measured by an inverted beaker type flowmeter.  Gas composition and 
flow were measured at least five times over the course of a run.  Error in the gas yields is 
reported as the standard deviation of the measured gas flow, all other experimentally 
measured quantities having very low variability. 
 
 
Figure 5.1:  Schematic of experimental apparatus 
 
X-ray spectra were collected on a Rigaku diffractometer equipped with a Cu K?
1
 
radiation source, graphite monochromator, and miniflex goniometer. The diffractometer 
was run at 40 kV voltage and 40 mA current, and scanned at 5?/min with 0.05? step size.  
Specific surface areas and pore volume data were determined by nitrogen physisorption 
Tube furnace with 
temperature controller 
Gas- Liquid 
Separator 
Rupture Disc 
 
Biocrude  
 
HPLC pump 
Water-cooled 
heat exchanger 
 
GC with 
TCD 
6-port injection 
valve for online 
sampling 
Packed Bed Reactor 
P
Back pressure regulators 
 
 
Liquid Effluent 
to TOC 
Inverted Beaker 
Flowmeter 
 
P TP 
 101
at 77 K after outgassing for 3 h at 300?C on a Quantachrome NOVA 2200e.  Scanning 
electron microscopy (SEM) was performed on a Zeiss EVO 50.  The samples were 
scattered onto two sided adhesive carbon tape on an aluminum stub, followed by a sputter 
coat of gold using an Electron Microscopy Services EMS 550X sputterer.  
Thermogravimetric data was collected on a TA Instruments TGA Q5000 under air flow 
of 120 mL/min. 
 
5.4 Results  
The major product gases formed were hydrogen, methane, and carbon dioxide, 
with small amounts of carbon monoxide, ethane, and ethylene.  Gas yields for titania and 
zirconia supported catalysts are shown in Figure 5.2, while gasification efficiency, carbon 
remaining in the aqueous effluent and closure of the carbon balance are given in Table 
5.2.  Gas yields are reported as moles of gas formed per mole of carbon reacted, as 
measured by the difference between the total organic carbon content of the biocrude feed 
and reactor effluent.  The gasification efficiency (GE) is defined as moles of carbon in the 
product gases divided by the moles of carbon in the feed.  Residual aqueous carbon is 
defined as carbon concentration of the liquid effluent divided by the feed carbon 
concentration.  The carbon balance is defined as the carbon exiting the system in the 
aqueous and gaseous phases divided by the input carbon flowrate.  Over the course of a 
typical 2 h run, the pressure drop across the reactor for zirconia supported catalysts was 
seen to steadily increase from approximately 5 bar at the start of the run to 100 bar when 
the experiment would be aborted.  This large increase in pressure drop was not an issue 
for titania supported catalysts in the timeframe of experiments conducted.    No results 
 102
are presented for gas yields from MgAl
2
O
4
 supported catalysts, as these each plugged and 
overpressurized the reactor before any measurements could be taken, typically within 
twenty minutes of operation.  To ensure that the plugging of the reactor was not related to 
sintering during the initial water pressurization, this period was extended to three hours.  
Again no increase in pressure drop was seen until biocrude feed was introduced.  Based 
on these results we find MgAl
2
O
4
 to be an inappropriate support for gasification of 
biocrude.  However, the stability of MgAl
2
O
4
 in SCW has received no attention in the 
literature and will be discussed herein. 
Ruthenium catalysts gave greatly differing results depending on the support.  The 
second highest hydrogen yield was seen from Ru/TiO
2
, however with Ru/ZrO
2
 the 
product gas contains very little hydrogen, consisting almost entirely of methane and 
carbon dioxide.  The poor closure of the carbon balance along with the lack of carbon in 
the liquid effluent stream point to extensive char formation over this catalyst formulation.  
Incomplete conversion of carbon in the liquid phase was seen for all of the titania 
supported catalysts, as can be seen from comparing the residual aqueous carbon in Table 
2.  Zirconia itself is known to have catalytic activity in the gasification of biomass
26, 27
, 
and as such does not simply act as an inert support. 
 
Table 5.2:  Gasification Efficiencies, Residual Aqueous Carbon and Carbon Balance 
Closure 
  Ru/TiO
2
Ru/ZrO
2
Ni/TiO
2
Ni/ZrO
2
 Co/TiO
2
 Co/ZrO
2
Gasification Efficiency (%) 78 67 74 96 83 102 
Residual Aqueous Carbon (%) 12 2 16 3 17 4 
Carbon Balance (%) 90 70 90 99 100 106 
 
 
 103
0.0
0.2
0.4
0.6
0.8
1.0
1.2
56
G
a
s 
Y
i
el
d 
(
m
ol
 gas
/
m
o
l
 C
 r
e
a
c
t
e
d)
0.00
0.05
0.10
0.15
0.20
0.25
G
a
s 
Y
i
el
d 
(
m
ol
 gas
/
m
o
l
 C
 r
e
a
c
t
e
d)
1234
CH
4
 
Ru/TiO2
Ru/ZrO2
Ni/TiO2
Ni/ZrO2
Co/TiO2
Co/ZrO2
H
2
 CO
2
 CO C
2
H
4
C
2
H
6
 
 
 
Figure 5.2:  Product gas yields for TiO
2
 and ZrO
2
 supported catalysts 
 
The highest observed hydrogen yield of 0.98 mol H
2
 /mol C was obtained with 
Ni/ZrO
2
, as was the lowest methane yield.  A lower hydrogen yield of 0.67 was obtained 
with Ni/TiO
2
, however the yield of methane was only slightly higher.  Small amounts of 
ethane and ethylene were also detected in the product gas over the Ni/TiO
2
 catalyst. 
The third highest hydrogen yield of 0.68 was achieved with the Co/ZrO
2
 catalyst.  
Co/TiO
2
 was the only catalyst to give a significant yield of carbon monoxide, not 
surprising since cobalt has a low activity for the water-gas shift reaction.
28
  Over 
Co/ZrO
2
, however, little CO was detected.  The increased activity of the water-gas shift 
may be due to strong metal support interactions between cobalt and zirconia, as has been 
observed for gold on zirconia.
29
   Yields of hydrogen, methane, carbon dioxide, and 
ethane were higher for Co/ZrO
2
 than Co/TiO
2
, as was the overall gasification efficiency. 
 
 
 
 104
5.4.1 Crystalline structure 
The stability of the crystalline structure of the catalyst supports was studied by 
XRD before and after exposure to SCW.  Figure 5.3 shows X-ray spectra for zirconia 
supported catalysts.  The diffraction patterns show that the monoclinic crystal structure of 
the zirconia was unchanged in the hydrothermal conditions, although in each case the 
peaks are more intense and have a smaller full width at half maximum after exposure to 
SCW, indicating a more uniform crystallinity after SCW exposure.  After calcination a 
diffraction peak at 43.4? for NiO is present in the Ni/ZrO
2
 sample, while the used catalyst 
instead shows a peak for Ni at 44.8?.  No diffraction peaks are visible for cobalt, 
ruthenium, or their oxides. 
20 25 30 35 40 45 50 55 60 65
2 ? (?)
Intensity (a.u.)
 
f
e
d
c
b
a
 
Figure 5.3:  XRD spectra of ZrO
2
 supported catalysts  before and after SCW exposure.  
(a) fresh and (b) used Ni/ZrO
2
; (c) fresh and (d) used Ru/ZrO
2
; (e) fresh and (f) used 
 Co/ZrO
2 
 
 105
The X-ray spectra of titania supported catalysts are shown in Figure 5.4.  The 
anatase structure is still the dominant phase after being on stream; however another 
unidentified pattern not corresponding to rutile or brookite is apparent after SCW 
exposure for Ni/TiO
2
 and Co/TiO
2
.  NiO is seen in the calcined Ni/TiO
2
 sample, while no 
peaks corresponding to Ni or NiO are seen in the used sample. 
20 25 30 35 40 45 50 55 60
2 ? (?)
Intensity (a.u.)
 
f
e
d
c
b 
a
Figure 5.4:  XRD spectra of TiO
2
 supported catalysts  before and after SCW exposure.  
(a) fresh and (b) used Ni/TiO
2
; (c) fresh and (d) used Ru/TiO
2
; (e) fresh and (f) used 
 Co/TiO
2
 
 
Figure 5.5 shows X-ray spectra of the freshly calcined and used catalysts 
supported on MgAl
2
O
4
.  Reflections from the MgAl
2
O
4
 structure dominate all the spectra 
both before and after SCW exposure with the peaks becoming narrower and more intense 
after SCW exposure, indicating a greater degree of crystallinity.  Each MgAl
2
O
4
 
supported catalyst also shows an additional phase of ?-Al
2
O
3
 after SCW exposure, 
characterized by the peak at about 43.3?.  Additional unidentified peaks at 2? = 12.1?, 
28.5?, 35.7?, and 54.8? are seen in the calcined Ru/MgAl
2
O
4
 sample after which are not 
 106
present after exposure to supercritical water.  These may be some artifact from the 
preparation method. 
10 15 20 25 30 35 40 45 50 55 60 65 70
2 ? (?)
Intensity (a.u.)
 
?-Al
2
O
3
 
e 
f
d
c 
b
T
a
Figure 5.5:  XRD spectra of MgAl
2
O
4
 supported catalysts  before and after SCW  
exposure.  (a) fresh and (b) used Ni/ MgAl
2
O
4
; (c) fresh and (d) used Ru/ MgAl
2
O
4
; (e)  
fresh and (f) used Co/ MgAl
2
O
4
 
 
5.4.2 Surface area 
Surface area, average pore radius, and pore volume were measured for the 
supports before impregnation, after new catalyst calcination, and of used catalyst from 
near the entrance and exit of the reactor.  These results are given in Table 5.3.  The 
surface area of each catalyst was decreased after metals impregnation and calcination.  
The ruthenium catalysts, which had a 1.5 wt% metals loading, showed a smaller loss of 
surface area than the nickel or cobalt catalysts, both of which had a 10 wt% metals 
loading.  A similar trend is seen in the pore volume, where each ruthenium catalyst had a 
small decrease in pore volume and a larger decrease for the nickel and cobalt catalysts.  
 107
This indicates that metals blocked access to some of the smallest pores, decreasing the 
total accessible volume and correspondingly the available surface area. 
Table 5.3:  Surface Area, Pore Radii and Volume for Calcined and Used Catalysts 
  
Surface Area (m
2
/g) 
 
Average Pore radius (?) 
 
Pore Volume (cc/g) 
 
  
New 
Calcined 
Used 
Entrance 
Used 
Exit 
New 
Calcined 
Used 
Entrance 
Used 
Exit 
New 
Calcined 
Used 
Entrance 
Used 
Exit 
ZrO
2
 50.4 - - 102.0 - - 0.256 - - 
Ru/ZrO
2
51.1 51.7 28.4 98.0 89.1 159.1 0.250 0.230 0.226 
Ni/ZrO
2
 43.0 56.3 24.3 96.3 65.1 161.6 0.207 0.184 0.196 
Co/ZrO
2
42.0 71.8 20.7 92.9 71.8 184.0 0.195 0.228 0.190 
               
TiO
2
 39.3 - - 79.4 - - 0.156 - - 
Ru/TiO
2
37.0 36.1 16.9 83.8 73.8 120.1 0.155 0.133 0.101 
Ni/TiO
2
 31.7 55.1 9.1 83.4 54.1 117.2 0.132 0.135 0.053 
Co/TiO
2
31.5 44.2 11.0 76.7 62.4 107.6 0.121 0.138 0.059 
               
MgAl
2
O
4
 113.8 - - 78.0 - - 0.443 - - 
Ru/MgAl
2
O
4
 108.3 127.2 55.2 79.4 68.8 152.2 0.430 0.437 0.420 
Ni/MgAl
2
O
4
 87.0 113.5 45.0 86.7 62.1 169.2 0.377 0.352 0.385 
Co/MgAl
2
O
4
 86.3 124.2 46.8 84.9 58.2 105.7 0.366 0.362 0.247 
 
 
Surface area of the used catalysts varied whether the catalyst sample is taken from 
near the entrance or the exit of the reactor.  Samples taken from the reactor entrance show 
a marked increase in surface area, while those from near the exit show a large decrease.  
As the biocrude is heated in the entrance zone of the reactor it forms char particles on the 
catalyst surface.  These carbon deposits are responsible for the apparent increase in 
surface area of the used catalyst.  At the downstream end of the reactor the temperature is 
higher than the entrance zone.  Little charring takes place in this zone and the surface 
area decrease is due to sintering in the hydrothermal environment.  The pore radius data 
for the three supports suggests this.  Smaller pore sizes are seen at the entrance as they 
become plugged, while the collapse of small pores at the downstream end results in larger 
average pore sizes. 
 108
5.4.3 Particle size and morphology 
 Representative SEM images of Ni/ZrO
2
 taken from the entrance and exit of the 
catalytic bed are shown in Figures 5.6 and 5.7, respectively.  The surface of the catalyst 
sample taken from the entrance of the reactor is covered and obstructed by char materials.  
The sample taken from the end of the reactor lacks much of this additional material and is 
charred to a lesser degree.  A similar situation occurs with samples taken from the 
entrance and exit of the Ni/TiO
2
 bed, shown in Figures 5.8 and 5.9, respectively.  The 
surface of the Ni/TiO
2
 catalyst from the entrance has a uniform covering of char particles.  
Larger agglomerates are seen at the downstream end, although the covering is not as 
thorough. 
 
Figure 5.6:  SEM of used Ni/ZrO
2
 catalyst taken from the reactor entrance 
 109
 
Figure 5.7: SEM of used Ni/ZrO
2
 catalyst taken from the reactor exit 
 
 
Figure 5.8: SEM of used Ni/TiO
2
 catalyst taken from the reactor entrance 
 110
 
Figure 5.9: SEM of used Ni/TiO
2
 catalyst taken from the reactor exit 
 
5.4.4 Thermo-gravimetric Analysis 
 Used nickel catalysts were selected for further analysis by TGA.  Samples of the 
catalyst were taken from the entrance and exit of the catalytic bed and heated at 10?C/min 
to 600?C in an air atmosphere.  Differing weight losses were seen from catalyst samples 
at the entrance and exit of the reactor for each support, and are given in Table 5.4. 
Weight loss associated with the burning of carbon deposits occurred over the range 200-
400?C for Ni/TiO
2
 and Ni/MgAl
2
O
4
, with the maximum rate of weight loss at 300?C.  
Ni/ZrO
2
 lost weight over the range 300-500?C with the maximum rate of weight loss at 
440?C.  The large weight loss from MgAl
2
O
4
 after only 20 minutes on stream indicates 
the strong tendency for charring over this support, although no quantifiable amount was 
observed from the sample taken from the downstream end of the reactor. After a 2 hour 
 111
period on stream, Ni/ZrO
2
 was seen to have over twice the amount of carbon as Ni/TiO
2
.  
A small amount of weight loss was detected from the end of the bed of Ni/TiO
2
, whereas 
none was observed from Ni/ZrO
2
. 
Table 5.4:  Weight Loss (%) of Carbonaceous Material During TGA 
 
  Ni/TiO
2
 Ni/ZrO
2
 Ni/MgAl
2
O
4
Entrance 1.8 3.9 6.1 
Exit 0.06 ND ND 
ND: None Detected  
 
5.5  Discussion 
To understand the charring behavior of the complex biocrude feedstock we 
examine the literature on model biomass compounds.  In the supercritical water 
gasification of glucose, a fast heating rate is known to increase gasification efficiency
30
 as 
well as enhance hydrogen yield
31
.  Obviously with a slow heating rate the reactants will 
spend more time in a subcritical water environment before the critical temperature is 
reached.  In the subcritical region furfurals and phenolic compounds can react to form 
higher molecular weight compounds, while direct gasification is preferred above the 
critical point.
7
  The intermediate compound 5-hydroxymethylfurfural has been observed 
to form high molecular weight char in subcritical water at 25 MPa and 350?C, but is 
completely gasified in supercritical water at 450?C.  In our system carbon deposition in 
the catalyst bed occurred mainly near the entrance of the reactor in each instance.  This 
was suggested by the decrease in pore size seen in catalyst samples taken from the reactor 
entrance and confirmed by weight loss from TGA.  The biocrude was fed to the reactor at 
room temperature without preheating, and it follows that the carbon deposition occurred 
while the reactants were being heated.   
 112
It was not expected that MgAl
2
O
4
 supported catalysts would char and plug as 
quickly as they did.  It is generally thought that surfaces with higher acidity tend to coke 
more quickly in steam reforming reactions.  Accordingly, basic MgO is often added to 
Al
2
O
3
 catalyst supports.  The reason for the rapid charring of the MgAl
2
O
4
 supported 
catalysts is at this point unclear, but perhaps this points to the need for the study of the 
surfaces of metal oxides in the presence of supercritical water.  
Crystallographically, the monoclinic zirconia support was stable in supercritical 
water, although BET analysis of a portion of each catalyst taken from the downstream 
end of the reactor lost approximately half of its surface area through sintering.  As 
discussed above, there are varying reports on the stability of the tetragonal phase of 
titania in the high temperature and pressure hydrothermal environment in the literature.  
In this work the structure of the commercial anatase titania used was found to be stable at 
250 bar and 600?C.  We note that in previous studies, titania supports containing a mix of 
anatase and rutile transform to exclusively to rutile after exposure to hot compressed 
water, while pure anatase is stable under the hydrothermal conditions examined.  In the 
case of the mixed phase titania supports it is likely that the presence of the rutile phase 
acts as a seed allowing anatase to transform at a lower temperature.  The MgAl
2
O
4
 
supported catalysts also sintered in the hydrothermal environment, each losing 
approximately half of their surface area after a short exposure to SCW.  An ?-Al
2
O
3
 
phase was seen to appear after exposure to SCW, however the origin of this phase is 
unclear from X-ray analysis.  The alumina could have originated from the MgAl
2
O
4
 
itself, however it is also possible that some amorphous alumina, which is not detectable 
 113
by XRD, existed following the preparation of the support and crystallized during the 
exposure to supercritical water.  
5.6 Conclusions 
Liquefied switchgrass was gasified over nickel, cobalt, and ruthenium catalysts 
supported on TiO
2
, ZrO
2
, and MgAl
2
O
4
 at 600?C and 250 bar with a WHSV of 9 h
-1
.  
Catalysts supported on MgAl
2
O
4
 charred immediately and were found to be an 
inappropriate support for biocrude reforming.  A given metal supported on ZrO
2
 gave a 
higher conversion of biocrude than those supported on TiO
2
, although charring at the face 
of the catalyst bed led to reactor plugging within a few hours for ZrO
2
 supported 
catalysts.  Char formation occurred to the smallest degree over TiO
2
 supports.  The 
highest hydrogen yield was obtained with Ni/ZrO
2
, while the lowest was with Ru/ZrO
2
, 
which gave a product gas composed of mostly methane and CO
2
.  The lowest gasification 
efficiency was seen with Ru/ZrO
2
.due to extensive char formation.  Co/TiO
2
 was the only 
catalyst to give a significant CO yield.  The anatase titania and monoclinic zirconia were 
crystallographically stable, but both lost significant amounts of surface area through 
hydrothermal sintering.  Following SCW exposure MgAl
2
O
4
 also sintered and was found 
to partially transform to ?-Al
2
O
3
. 
 
 114
5.7 References 
1. Balat, H.; Kirtay, E.; Hydrogen from biomass: Present scenario and future 
prospects. International Journal of Hydrogen Energy 2010, 35, 7416-7426. 
 
2. Matsumura, Y.; Minowa, T.; Potic, B.; Kersten, S. R. A.; Prins, W.; van Swaaij, 
W. P. M.; van de Beld, B.; Elliott, D. C.; Neuenschwander, G. G.; Kruse, A.; Antal, M. 
J., Jr., Biomass gasification in near- and super-critical water: Status and prospects. 
Biomass and Bioenergy 2005, 29, 269-292. 
 
3. Guo, Y.; Wang, S. Z.; Xu, D. H.; Gong, Y. M.; Ma, H. H.; Tang, X. Y., Review 
of catalytic supercritical water gasification for hydrogen production from biomass. 
Renewable and Sustainable Energy Reviews 2010, 14, 334-343. 
 
4. Gupta, R. B., Hydrogen Fuel: Production, Transport and Storage. CRC Press: 
Boca Raton, FL, 2008. 
 
5. Bartels, J. R.; Pate, M. B.; Olson, N. K., An economic survey of hydrogen 
production from conventional and alternative energy sources. International Journal of 
Hydrogen Energy 2010, 35, 8371-8384. 
 
6. Akiya, N.; Savage, P. E., Roles of Water for Chemical Reactions in High-
Temperature Water. Chemical Reviews 2002, 102, 2725-2750. 
 
7. Kruse, A.; Gawlik, A., Biomass Conversion in Water at 330-410 ?C and 30-50 
MPa. Identification of Key Compounds for Indicating Different Chemical Reaction 
Pathways. Industrial Engineering Chemistry Research 2003, 42, 267-279. 
 
8. Kumar, S.; Gupta, R. B., Biocrude Production from Switchgrass Using Subcritical 
Water. Energy & Fuels 2009, 23, 5151-5159. 
 
9. Osada, M.; Sato, T.; Arai, K.; Shirai, M., Stability of Supported Ruthenium 
Catalysts for Lignin Gasification in Supercritical Water. Energy & Fuels 2006, 20, 2337-
2343. 
 
10. Yu, J.; Savage, P. E., Catalyst activity, stability, and transformations during 
oxidation in supercritical water. Applied Catalysis B: Environmental 2001, 31, 123-132. 
 
11. Elliott, D. C.; Sealock, J., Jr.; Baker, E. G., Chemical Processing in High-pressure 
Aqueous Environments. 2. Development of Catalysts for Gasification. Industrial 
Engineering Chemistry Research 1993, 32, 1542-1548. 
 
12. Elliott, D. C.; Hart, T. R.; Neuenschwander, G. G., Chemical Processing in High-
Pressure Aqueous Environments. 8. Improved Catalysts for Hydrothermal Gasification. 
Industrial Engineering Chemistry Research 2006, 45, 3776-3781. 
 
 115
13. Matsumura, Y.; Kato, A.; Sasaki, H.; Yoshida, T. In Gasification of liquefied 
biomass in supercritical water using partial oxidation, Progress in Thermochemical 
Biomass Conversion, Tyrol, Austria, 2000; Bridgwater, A. V., Ed. Blackwell Science 
Ltd.: Tyrol, Austria, 2000; pp 237-251. 
 
14. Elliott, D. C.; Neuenschwander, G. G.; Hart, T. R.; Butner, R. S.; Zacher, A. H.; 
Engelhard, M. H.; Young, J. S.; McCready, D. E., Chemical Processing in High-Pressure 
Aqueous Environments. 7. Process Development for Catalytic Gasification of Wet 
Biomass Feedstocks. Industrial Engineering Chemistry Research 2004, 43, 1999-2004. 
 
15. Meryemoglu, B.; Hesenov, A.; Irmak, S.; Atanur, O. M.; Erbatur, O., Aqueous 
phase reforming of biomass using various types of supported precious metal and raney-
nickel catalysts for hydrogen production. International Journal of Hydrogen Energy 
2010, 35, 12580-12587. 
 
16. Sato, T.; Sekiguchi, G.; Adschiri, T.; Arai, K., Non-catalytic and selective 
alkylation of phenol with propan-2-ol in supercritical water. Chemical Communications 
2001, 17, (1566-1567). 
 
17. Sato, T.; Sekiguchi, G.; Adschiri, T.; Arai, K., Ortho-Selective Alkylation of 
Phenol with 2-Propanol without Catalyst in Supercritical Water. Industrial Engineering 
Chemistry Research 2002, 41, 3064-3070. 
 
18. Sato, T.; Sekiguchi, G.; Adschiri, T.; Smith, R. L., Jr; Arai, K., Regioselectivity of 
phenol alkylation in supercritical water. Green Chemistry 2002, 4, 449-451. 
 
19. Saisu, M.; Sato, T.; Watanabe, M.; Adschiri, T.; Arai, K., Conversion of Lignin 
with Supercritical Water-Phenol Mixtures. Energy & Fuels 2003, 17, 922-928. 
 
20. Chuntanapum, A.; Yont, T. L.-K.; Miyake, S.; Matsumura, Y., Behavior of 5-
HMF in Subcritical and Supercritical Water. Industrial Engineering Chemistry Research 
2008, 47, 2956-2962. 
 
21. Chuntanapum, A.; Matsumura, Y., Formation of Tarry Material from 5-HMF in 
Subcritical and Supercritical Water. Industrial Engineering Chemistry Research 2009, 
48, 9837-9846. 
 
22. Byrd, A. J.; Pant, K. K.; Gupta, R. B., Hydrogen Production from Glucose Using 
Ru/Al2O3 Catalyst in Supercritical Water. Industrial Engineering Chemistry Research 
2007, 46, 3574-3579. 
 
23. Byrd, A. J.; Pant, K. K.; Gupta, R. B., Hydrogen Production from Ethanol by 
Reforming in Supercritical Water Using Ru/Al2O3 Catalyst. Energy & Fuels 2007, 21, 
3541-3547. 
 
 116
24. Byrd, A. J.; Pant, K. K.; Gupta, R. B., Hydrogen production from glycerol by 
reforming in supercritical water over Ru/Al2O3 catalyst. Fuel 2008, 87, 2956-2960. 
 
25. Bocanegra, S. A.; Ballarini, A. D.; Scelza, O. A.; de Miguel, S. R., The influence 
of the synthesis routes of MgAl2O4 on its properties and bavior as support of 
dehydrogenation catalysts. Materials Chemistry and Physics 2008, 111, 534-541. 
 
26. Watanabe, M.; Inomata, H.; Arai, K., Catalytic hydrogen generation from 
biomass (glucose and cellulose) with ZrO2 in supercritical water. Biomass and Bioenergy 
2002, 22, 405-410. 
 
27. Watanabe, M.; Inomata, H.; Osada, M.; Sato, T.; Adschiri, T.; Arai, K., Catalytic 
effects of NaOH and ZrO2 for partial oxidative gasification of n-hexadecane and lignin in 
supercritical water. Fuel 2003, 82, 545-552. 
 
28. Dalai, A. K.; Davis, B. H., Fischer?Tropsch synthesis: A review of water effects 
on the performances of unsupported and supported Co catalysts. Applied Catalysis A: 
General 2008, 348, 1-15. 
 
29. Menegazzo, F.; Pinna, F.; Signoretto, M.; Trevisan, V.; Boccuzzi, F.; Chiorino, 
A.; Manzoli, M., Highly Dispersed Gold on Zirconia: Characterization and Activity in 
Low-Temperature Water Gas Shift Tests. ChemSusChem 2008, 1, 320-326. 
 
30. Sinag, A.; Kruse, A.; Rathert, J., Influence of the Heating Rate and the Type of 
Catalyst on the Formation of Key Intermediates and on the Generation of Gases 
During Hydropyrolysis of Glucose in Supercritical Water in a Batch Reactor. Industrial 
Engineering Chemistry Research 2004, 43, 502-508. 
 
31. Watanabe, M.; Aizawa, Y.; Iida, T.; Levy, C.; Aida, T. M.; Inomata, H., Glucose 
reactions within the heating period and the effect of heating rate on the reactions in hot 
compressed water. Carbohydrate Research 2005, 340, 1931-1939. 
 
 117
 
6.  Stability of Cerium-modified ?-Alumina Catalyst Support in Supercritical Water 
 
 
6.1 Abstract  
Supercritical water (above 374.1?C and 220.6 bar) is emerging as a promising 
medium to carry out a variety of catalytic reactions, including reforming to produce 
hydrogen.  However, when using a heterogeneous catalyst the support material can 
undergo transformations in the hydrothermal environment.  In this work the stability of ?-
Al
2
O
3
 modified with 1-10 wt% Ce in supercritical water is examined, specifically in the 
temperature range of 500 ? 700?C at 246 bar.  Transformations of the ?-phase were 
slowed but not prevented.  Based on X-ray analysis, the transformation of ?-Al
2
O
3
 
proceeded through the ? phase toward the stable ? phase.  Reduced cerium species were 
seen to be oxidized in the supercritical water environment, and low Ce-loading supports 
maintained the highest BET surface areas.  The stabilization was greatest at 700?C, where 
Ce-modified aluminas retained significantly higher specific surface areas than 
unmodified alumina.  
6.2.  Introduction 
Supercritical water (T
c 
= 374.1?C, P
c 
= 220.6 bar) is an emerging reaction medium 
for hydrogen production due to its desirable thermophysical properties such as high 
diffusivity, low viscosity, and ability to solubilize polar molecules as well as gases.  As a 
dense reaction medium it allows for small reactor volumes, and gasification and 
reforming reactions may proceed homogeneously.  Employment of a heterogeneous 
catalyst can further increase throughput by allowing even shorter reaction times as well 
 118
as lowering operating temperatures.  Aluminum oxide is commonly used as a catalyst 
support owing to its high specific surface area, acidic nature of the surface, and low 
cost.
1, 2
  A number of meta-stable phases (?, ?, ?, ?, ?, ?) collectively known as transition 
aluminas can be prepared by calcining hydrated alumina precursors such as boehmite (?-
AlO(OH)), gibbsite (?-Al(OH)
3
), or bayerite  (?-Al(OH)
3
) at a suitable temperature.  This 
paper examines the stability of ?-Al
2
O
3
, as it generally contains the highest specific 
surface area of the transition aluminas.  Under atmospheric pressure all of the transition 
aluminas transform to the thermodynamically stable alpha phase (mineralogically known 
as corundum) at approximately 1100?C, although the exact transition temperature is 
dependent on impurities and particle size.
3
  ?-Al
2
O
3
 transforms with increasing 
temperature to corundum by the sequence ? ? ? ? ? ? ?.  Dehydration sequences for 
several alumina hydrates under air at atmospheric pressure are illustrated in Figure 6.1.  
While the surface area decreases with each step, the final transition to the alpha phase is 
associated with the greatest decrease.
4
 
From a crystallographic point of view, ?-Al
2
O
3
 can be described as a defect spinel 
structure with a face-centered cubic (fcc) sublattice of O atoms and a sublattice of Al
3+
 
cations distributed randomly between octahedral and tetrahedral interstitial sites.  During 
the transformation to the higher transition aluminas there is a tendency for Al
3+
 to migrate 
from tetrahedral sites where they are stabilized by four anions to octahedral sites which 
are stabilized by six anions.  The final transformation to the ? phase involves a complete 
recrystallization resulting in a hexagonally close packed (hcp) oxygen sublattice with 
cations at 2/3 of the octahedral sites.
5
  Burtin et al.
6
 describe the entire transition from 
 119
boehmite to corundum by a progressive dehydration due to loss of hydroxyl groups.  This 
transition can be described with the following formula in Kr?ger-Vink notation:  
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Temperature (?C)
Boehmite Gamma Delta Alpha
Gibbsite Chi Kappa Alpha
Bayerite Eta Theta
Theta
Alpha
a
a
b
b
 
Figure 6.1:  Dehydration sequence of selected alumina hydrates in air at ambient 
pressure.  Path a favored by d
p
 < 10 ?m; Path b favored by alkalinity and d
p
 > 100 ?m.  
Open areas indicate transitional zones. Adapted from reference [3] 
 
                                                                             (6.1) 
/2-1OO/2-3O
''
M2Al
)(V)OH()(OV)Al(
???
?????
where represents an aluminum cation in a trivalent site, a cationic vacancy at a 
divalent site, an oxygen anion at an oxygen site, a hydroxyl substituted onto an 
oxygen site, and  an oxygen vacancy.  Thus for v = 2 the formula reduces to that of 
boehmite, and for v = 0 the formula for corundum is obtained. Dehydration through 
dehydroxylation gives rise to the appearance of an oxygen vacancy, described in Kr?ger-
Vink notation as: 
?
Al
Al
''
M
V
?
O
O
?
O
OH
??
O
V
2( ) ? H
2
O + O
O
 +                                             (6.2) 
?
O
OH
??
O
V
 120
Dehydroxylation proceeds as anionic and cationic vacancies annihilate each other, 
resulting in the transformation to corundum.  The transformation to the alpha phase is 
initiated at points of contact between particles
1, 7
 where necking regions contain large 
numbers of anion vacancies from the formation of Al-O-Al bonds during 
dehydroxylation according to Equation 6.2.  The growth of the neck regions gives rise to 
sintering and thus the drastic loss of surface area associated with transformation to the 
alpha phase.  Sintering by this mechanism is depicted in Figure 6.2. 
 
Figure 6.2:  Model of alumina surface dehydroxylation resulting in loss of surface area.  
Adapted from reference [4] 
 
Several groups have explored phase transitions of alumina under increased 
pressure.  For example, Ito et al. observed that the transition to ?-Al
2
O
3
 is accelerated by 
an increase in pressure, resulting in a decrease in the transition temperature.
8
  Using hot 
isostatic pressure under an argon atmosphere, they observed that nucleation was induced 
by stress concentration at particle contact points.  Panasyuk et al. observed that upon 
exposure to hydrothermal environment at 450?C and 10 MPa, gibbsite transforms first to 
boehmite, then to corundum; here an intermediate state between boehmite and corundum 
was referred but was not characterized. 
9, 10
 Indeed, the accelerated transformations in the 
 121
hydrothermal environment have given rise to novel syntheses of both transition
11
 and 
alpha aluminas
12
, as well as other complex oxides
13
. 
The suitability of alumina as a catalyst support in sub- and supercritical water has 
received some attention in the literature, and our group has also noted the instability of 
alumina in our previous work on hydrogen production in supercritical water.
14
  Elliott et 
al. reported that after exposure to subcritical water at 350?C and 20 MPa, ?- and ?-Al
2
O
3
 
were completely hydrolyzed to give boehmite, while ?-Al
2
O
3
 formed a mix of boehmite 
and ?-Al
2
O
3
.  Corundum was the only polymorph of alumina found to be stable under 
conditions investigated.
15
  Osada et al. evaluated ruthenium catalysts on several supports 
for their activity in lignin gasification at 400?C and 37 MPa.  The Ru/Al
2
O
3
 catalyst lost 
most of its activity after one reaction period of 180 minutes in their batch system.  This 
was attributed to leaching of Ru metal during a phase transition from ?- to ?-Al
2
O
3
.  In 
the supercritical water oxidation (SCWO) of phenol in a continuous packed bed reactor, 
Yu and Savage
16
 saw that a CuO/?-Al
2
O
3
 catalyst was unstable under the reaction 
conditions of 380?C and 25 MPa, finding that after exposure to SCW the dominant 
aluminum-containing species was boehmite. 
A number of researchers have investigated the thermal stabilization of the 
transition aluminas by inclusion of small amounts of other materials such as Ba
17
, Ca
18
, 
Ce
17, 19, 20
, Ga
18
, In
18
, La
17, 18, 21
, Mg
17, 18
, Pr
17
, SrO
22
 , Th
18
, Zr
18, 23
, and mixed oxide Ce-
Zr
24, 25
 systems.   These studies show that most of the stabilizing agents suppress the 
phase transition to the alpha phase to some degree, while In
3+
, Ga
3+
, and Mg
2+
 accelerate 
the transition
6
.  Oudet et al. 
26-28
 proposed that thermal stabilization of noble-metal 
automotive-exhaust catalysts supported on alumina by elements from the lanthanide 
 122
series (Ln = La, Pr, Nd) is related to the presence of LnAlO
3
 microdomains in the 
corundum nucleation sites.  The model was later extended to include Ce, as all of the 
mixed oxides are isostructural with identical lattice parameters.  For example, there is a 
structural coherence between the CeAlO
3
 and Al
2
O
3
 phases due to continuity of the 
anionic lattices of both compounds.
29
 
The fundamental studies have focused on the thermal stability under a variety of 
atmospheres (reducing/oxidizing, presence of moisture), but at ambient pressure.  As 
discussed above, exposure to a high pressure environment can accelerate phase 
transformations and enable them to take place at lower temperatures.  This work 
investigates the stability of Ce-modified alumina in a supercritical water environment at 
both high temperature and pressure.  Cerium was chosen over other stabilizing agents 
because in addition to its known ability to suppress phase transformations in alumina, it 
also gives other benefits to catalyst formulations including water-gas shift activity, 
increased oxygen storage capacity, as well as aiding in retaining high dispersion of 
supported metals.
30, 31
   
 
6.3.  Experimental 
6.3.1  Sample Preparation 
?-Al
2
O
3
 having a specific surface area of 256 m
2
 g
-1
 and a pore volume of 0.62 
cm
3
 g
-1
 was obtained from Alfa Aesar as 1/8? extruded pellets.  The pellets were crushed 
and sieved to particle sizes ranging from 150-600 ?m.  Cerium (III) nitrate hexahydrate 
(99.9% Ce, metals basis) was obtained from Strem Chemicals.  Deionized water was used 
in all catalyst preparations and hydrothermal experiments. 
 123
  Cerium loading of the alumina carrier was performed by an incipient wetness 
impregnation technique where an appropriate amount of cerium nitrate was dissolved in 
water to give a solution whose volume was equivalent to the pore volume of the support. 
The samples were then dried overnight at 110?C and subsequently calcined in air at 
500?C for 6 h, followed by reduction at 750 or 950?C in a stream of 5%H
2
 in He flowing 
at 0.2 SLPM for 6 h.  Throughout the article, the nomenclature to designate cerium 
loading is written as xCeAl, where x is the loading of cerium in weight percent, defined 
as grams of elemental cerium in 100 grams of alumina plus elemental cerium. 
 
6.3.2 Experimental Procedure 
 
A schematic of the experimental apparatus is shown in Figure 6.3.  In each 
experiment, approximately 0.5 g of alumina or Ce-modified alumina was placed into a 50 
cm Inconel 600 tube (OD: 6.35 mm, ID: 3.05 mm) with pressed steel frits at both ends 
having pore sizes of 2 ?m.  Additionally, there was a quantity of smaller steel frits 
randomly packed into the last 5 cm of the tube to serve as an inert reactor packing and to 
ensure that the sample did not experience any uneven heating effects from heat losses 
through the fittings at the end of the tube.  
 124
V-1 
 
T
HPLC 
Pumps 
Water-Cooled 
Heat Exchanger 
Fixed Bed 
Tube Furnace with 
Temperature 
Controller
P 
Back 
Pressure 
Regulator 
 
Water  
 Water  Preheater 
N
2 
Collect 
Aqueous 
Sample 
V-3 
Pressure 
Relief 
Valve 
Drain 
V-5 
V-4 
V-2 
P 
Figure 6.3:  Experimental apparatus 
 
The reactor was placed inside of a Thermolyne 21100 tubular furnace with the ends 
covered in ceramic insulation and additional insulation was placed around the pressure 
fittings protruding from the furnace to reduce heat losses.  The furnace was then brought 
to the desired temperature and held for ten minutes before pressurization with water.  
Water was supplied at 2 mL min
-1
 by an HPLC pump (Alltech 301) and pumped through 
a 2 m section of 1/16? tubing serving as a preheater before entering the reactor.   The 
time at which the desired temperature and pressure were achieved was taken as t = 0 in 
each experiment, and the reactor was then isolated by closing valves V-3 and V-5.  A 
fixed pressure of 246 bar was used in all experiments.  After the allotted time had passed, 
nitrogen gas slightly above the pressure of the hydrothermal system was applied and 
needle valve V-4 was carefully opened to collect water from the system.  In this way the 
sample was not exposed to subcritical water at the end of the run.  High pressure nitrogen 
was supplied by means of a 1.5 L high pressure vessel (HiP TOC31-10-P) equipped with 
a close fitting piston.  Prior to the start of an experiment, nitrogen gas was supplied to one 
 125
side of the piston at cylinder pressure by a high pressure regulator.  The high pressure 
side of the piston was then isolated and water pumped to the other side to increase 
pressure to the desired level.  The time to drain water from the hydrothermal system was 
5 minutes, after which nitrogen gas was bled out for another ten minutes before fully 
opening valve V-4.  Upon the increase in nitrogen flow, the furnace was quickly cooled 
by blowing compressed air into the furnace while the remaining gas from the piston 
passed through the bed.  The total time to depressurize was approximately 30-40 minutes, 
during which the sample was exposed to nitrogen at high temperature and pressure. 
 Samples were characterized by X-ray Diffraction (XRD), Brunauer-Emmett-
Teller (BET) method, and inductively coupled plasma atomic emission spectroscopy 
(ICP-AES).  X-ray spectra were collected over the range 2? = 10?- 80?on a Rigaku 
diffractometer equipped with a Cu
K?1
 radiation source, graphite monochrometer, and 
miniflex goniometer.  The diffractometer was run at 40 kV voltage and 40 mA current, 
and scanned at 5?/min with 0.05? step size.  X-ray spectra obtained were compared 
against JCPDS/ICDD cards 21-1307 (AlO(OH)), 04-0878 (?-Al
2
O
3
), 46-1131 (?-Al
2
O
3
), 
35-0121 (?-Al
2
O
3
), 10-173 (?-Al
2
O
3
), 34-394 (CeO
2
), and 18-0315 (CeAlO
3
).
32
  Specific 
surface areas were determined by N
2
 physisorption using a 5-point BET method at 77 K 
on an Autosorb-1 instrument (Quantachrome).  Following the depressurization of the 
apparatus, water that was in contact with the Ce-modified alumina samples was checked 
for presence of cerium species by ICP-AES (Varian Vista-MPX ICP AES).   
6.4.  Results and Discussion 
6.4.1 XRD 
 126
X-ray spectra of the freshly calcined CeAl samples and ?-Al
2
O
3
 are shown in Fig 6.4.  
Diffraction peaks appear at 2? = 29.2?, 48.4?, and 57.1? corresponding to CeO
2
 for 4 wt% 
(i.e., 4CeAl) and greater Ce loading, while no additional diffraction peaks were seen on 
the lower loading samples.  CeO
2
 on the lower loading samples is expected to be 
undetectable by XRD due to small crystallite size and/or weak signal.
31
  After a reducing 
treatment at 750?C, diffraction peaks from CeO
2
 in the 4CeAl sample were absent but 
still persisted on 6CeAl samples.  Following reduction at 950?C, however, 6CeAl and 
10CeAl showed no diffraction peaks from CeO
2
.  No additional diffraction peaks 
assignable to CeAlO
3
 or Ce
2
O
3 
were seen in the X-ray spectra of any samples following 
reduction. 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
te
n
s
it
y
 (
a
.u
.)
f
a
b
c
d
e
 
Figure 6.4:.  XRD spectra of (a) ?-Al2O3 as received; freshly calcined (b) 1CeAl, (c) 
2CeAl, (d) 4CeAl, (e) 6CeAl, (f) 10CeAl.  ? corresponds to CeO
2
. 
 
 127
XRD spectra of unmodified alumina before and after exposure to supercritical water 
between 500-700?C at 246 bar for three hours are shown in Fig. 6.5.  After exposure to 
SCW at 500?C the dominant phase present is boehmite, indicating extensive hydrolysis 
of the support.  Weaker reflections are also present corresponding to the alpha and kappa 
phases.  After SCW exposure at 600?C boehmite is not present.  Instead, the alpha phase 
shows the strongest reflections, while kappa is also visible.  Broad diffraction peaks 
appearing at 32?-34? and 45-47? are not easily assignable to a particular phase, as ?-, ?-, 
and ?-Al
2
O
3
 all contain multiple reflections in this region with very similar positions. All 
peaks present on the spectrum after SCW exposure at 700? can be indexed as a well-
crystallized alpha phase. 
 
 
 
 
 
 
 
 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
te
n
s
ity
 (a
.u
.)
a
b
c
Figure 6.5:  XRD spectra of unmodified Al
2
O
3
 after 3h exposure to supercritical water at 
(a) 500?C, (b) 600?C, and (c) 700?C.  Labels correspond to (?) ?-Al
2
O
3
, (?) ?-Al
2
O
3
, 
and (?) Boehmite 
 128
 
Examining the X-ray spectra from Ce-modified alumina after exposure to 500?C 
SCW in Fig. 6.6, it is apparent that the intensity of the reflections associated with 
boehmite are greatly reduced in all three spectra compared with plain ?-Al
2
O
3
 after SCW 
exposure.  The degree of recrystallization to the ? or ? phase varies with Ce loading.  The 
2CeAl sample shows intense reflections from the alpha phase which are much smaller in 
the higher loading CeAl samples.  The most intense reflection from the alpha phase 
((113) at 43.65?) is only discernable as a weak shoulder to the intense peak from the 
kappa phase at 43.1? in the 6CeAl sample, but is distinct in the highest Ce loading 
sample. 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
t
e
n
s
ity
 (
a
.u
.)
a
b
c
 
Figure 6.6:  XRD spectra of CeAl samples after exposure to SCW at 500?C (a) 2CeAl (b) 
6CeAl (c) 10CeAl. Labels correspond to (?) ?-Al
2
O
3
, (?) ?-Al
2
O
3
, (?) Boehmite, (?) 
CeO
2
 
 129
 
     As was the case with plain ?-Al
2
O
3
, Ce-modified aluminas exposed to SCW at 600?C 
do not show any evidence of boehmite formation, as seen in Fig. 6.7.  The dominant 
alumina phases present are alpha and kappa.  Again the tendency towards ? or ? phase 
dominance after 3 h is related to the cerium loading.  As the loading of cerium was 
increased from 1 to 4 wt%, the intensity of the ?-(113) peak decreases while the 
neighboring peak for ?-Al
2
O
3
 increases.  The samples with the highest cerium loading 
show only a broad shoulder at the position of the ?-(113) peak. 
 
 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
t
e
n
s
ity
 (
a
.u
.)
a
b
c
 
Figure 6.7:  XRD spectra of CeAl samples after exposure to SCW at 600?C (a) 1CeAl (b) 
2CeAl (c) 4CeAl (d) 6CeAl (e) 10CeAl. Labels correspond to (?) ?-Al
2
O
3
, (?) ?-Al
2
O
3
, 
(?) CeO
2
 
 
 130
X-ray spectra from samples exposed to SCW at 700?C are shown in Fig 6.8.  The 
2, 4, and 6 wt% Ce-modified alumina samples all exhibit mixed alpha and kappa phases, 
in contrast to plain alumina that was completely converted to the alpha phase under 
identical conditions.  The intensity of reflections from the alpha phase overshadows the 
kappa phase for the 2 and 4 wt% samples, however the 6 wt% Ce sample shows higher 
intensity diffraction peaks for the kappa phase. 
 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
t
e
n
s
it
y (
a
.
u
.
)
c 
b 
a 
 
 
 
Figure 6.8:  XRD spectra of CeAl samples after exposure to SCW at 700?C (a) 2CeAl (b) 
4CeAl (c) 6CeAl. Labels correspond to (?) ?-Al
2
O
3
, (?) ?-Al
2
O
3
, (?) CeO
2
 
 
 131
The presence of diffraction peaks assignable to CeO
2
 after exposure to SCW 
indicates the oxidation of CeAlO
3
 to some degree.  The reappearance of diffraction peaks 
for ceria after exposure to SCW occurred at all temperatures explored in the present 
study.  Further, there is evidence for sintering of ceria after exposure to SCW.  Whereas 
freshly calcined 2CeAl showed no diffraction peaks for CeO
2
, small reflections at the 
positions expected for ceria can be seen post-SCW treatment, indicating that ceria 
domains below the size detection limits for XRD have coalesced to observable sizes. 
 Additional experiments were performed with time extended to 6 h as well as with 
6CeAl reduced at 950?C for hours.  Figures 6.9a and 6.9c show the X-ray spectra of 
6CeAl samples reduced at 950? and 750?C, respectively, and exposed to SCW for three 
hours.  In both cases the kappa phase is the only clearly visible phase of alumina, 
although the (113) reflection from the alpha phase is present as a shoulder in both cases.  
Reflections from the sample reduced at 750?C have a higher intensity.  Spectra shown in 
Figures 6.9b and 6.9d were obtained after exposure to SCW for six hours.  Examining 
Fig. 6.9b, it can be seen that for the sample reduced at 950?C that the kappa phase 
reflections have increased in intensity, while (113) reflection from the alpha phase still 
remains as a shoulder.  This contrasts with Fig. 6.9d, reduced at 750?C, where reflections 
from the alpha phase are clearly identifiable.  There are several differences in the 
reflections from ceria in these samples as well.  The most intense peak associated with 
ceria is at 29?, corresponding to the (111) plane.  This peak has a greater intensity in both 
samples reduced at 750?.  The diffraction peak from the ceria (111) plane is narrower and 
more intense in Fig. 6.9d than the freshly calcined sample, indicating a greater degree of 
crystallinity and the sintering of ceria.  This effect is not noticeable in the 6CeAl sample 
 132
reduced at 950?C.  Additionally, in Fig. 6.9d a weak reflection for boehmite is present at 
14.8? which may be due brief period of exposure to subcritical water during the course of 
the experiment caused by a leak in the system. 
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
2 ? (?)
In
t
e
n
s
ity
 (a
.u
.)
b
c
d
a
 
Figure 6.9:  XRD spectra of 6CeAl samples after exposure to SCW at 600?C.  (a) 
Reduced at 950?C, SCW exposure 3h (b) Reduced at 950?C, SCW exposure 6 h (c) 
Reduced at 750?C, SCW exposure 3h (d) Reduced at 750?C, SCW exposure 6 h.  Labels 
correspond to (?) ?-Al
2
O
3
, (?) ?-Al
2
O
3
, (?) CeO
2
, (?) Boehmite 
 
6.4.2  BET Surface Area 
 Surface areas measured by N
2
 physisorption for as received ?-Al
2
O
3
, CeAl 
samples following calcination and reduction (indicated by C&R), and after exposure to 
SCW at 500-700?C, 246 bar for 3h are shown in Table 6.1.  Generally, all CeAl samples  
 133
 
experienced some loss of surface area after calcination and reduction, with losses 
increasing with higher Ce loading.   At 500? and 600?C, samples loaded with 1-2% Ce 
retained both the highest specific surface area as well as the largest percentage of original 
surface area.  After exposure to SCW at 700?C all of the Ce-modified aluminas reduced 
at 750?C retained approximately 30-40 m
2
 g
-1
, significantly more than the 2 m
2
 g
-1
 
exhibited by the unmodified sample under the same conditions. Comparing the specific 
surface areas of 6CeAl samples reduced at 750 and 950?C, the samples reduced at 950?C 
have a lower initial surface area, however each retained a higher overall surface area as 
well as a higher percentage of the specific area before exposure to SCW.  
 
Table 6.1:  BET Surface Area of Samples 
 
    SCW Exposure Temperature 
Ce loading (wt%) C&R
a
 700?C 600?C 500?C 
10 164.4 37.4 30.1 39.1 
6 215.5 42.3 57.2 50.6 
 6
b
 143.3 56.9 70.0 91.3 
4 172.7 37.1 56.0 90.2 
2 187.0 42.0 63.7 101 
1 197.0 27.7 105 71.4 
0 256.5
c
 2.4 97.2 49.1 
              
a
  Surface area after calcination at 500?C and reduction at 750?C 
              
b
  Reduction performed at 950?C 
              
c
  Surface area of alumina as received 
 
6.4.3 ICP-AES 
Inductively coupled plasma atomic emission spectroscopy was used to detect cerium 
species present in the water after depressurizing the experimental apparatus.  
 134
Concentrations of cerium corresponding to ~1% of the original loading were seen after 
exposure to SCW at 500? and 600?C in the effluent from 4CeAl samples calcined at 
500?C and receiving no reductive treatment.  Analysis of a 4CeAl sample which was 
calcined at 800?C and exposed to identical SCW treatments did not show the presence of 
detectable Ce in the effluent water.  No cerium was detected in samples which were 
subjected to reductive treatments at 750?C. 
 
6.5 Conclusions 
 The stability of cerium-modified ?-Al
2
O
3
 was evaluated in supercritical water.  
XRD and BET studies showed that the inclusion of cerium slowed phase transformations 
in the supercritical water environment; however, the ?-phase was transformed in all cases 
and was accompanied by the loss of a large percentage of surface area.  Ce-modified 
alumina samples showed reduced boehmite formation at 500?C, while at higher 
temperatures the transformation towards the thermodynamically stable ?-Al
2
O
3
 was 
slowed.  Formation of ?-Al
2
O
3
 was observed in the transformation sequence towards 
corundum, and reduced cerium species were oxidized after exposure to SCW.  Low 
cerium loadings (1-2 wt%) maintained the highest BET surface areas at 500-600?C, while 
at 700?C all loadings retained approximately 30-40 m
2
 g
-1
, compared to 2 m
2
 g
-1
 for 
unmodified alumina. 
 
 
 135
6.6 References 
1. Johnson, M. F. L., Surface area stability of aluminas. Journal of Catalysis 1990, 
123, 245-249. 
 
2. Jun-Cheng, L.; Lan, X.; Feng, X.; Zhan-Wen, W.; Fei, W., Effect of hydrothermal 
treatment on the acidity distribution of ?-Al2O3 support. Applied Surface Science 2006, 
253, 766-770. 
 
3. Gitzen, W. H., Alumina as a Ceramic Material. The American Ceramic Society, 
Inc.: Columbus, Ohio, 1970. 
 
4. Arai, H.; Machida, M., Thermal stabilization of catalyst supports and their 
application to high-temperature catalytic combustion. Applied Catalysis A: General 1996, 
138, 161-176. 
 
5. Levin, I.; Brandon, D., Metastable Alumina Polymorphs: Crystal Structures and 
Transition Sequences. Journal of the American Ceramic Society 1998, 81, 1995-2012. 
 
6. Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M., Influence of surface area and 
additives on the thermal stability of transition alumina catalyst supports.  II.  Kinetic 
model and interpretation. Applied Catalysis 1987, 34, 239-254. 
 
7. Tucker, D. S., Gamma-to-alpha transformation in spherical aluminum oxide 
powders. Journal of the American Ceramic Society 1985, 68, C163-164. 
 
8. Ito, S.; Umehara, N.; Takata, H.; Fujii, T., Phase transition of [gamma]-Al2O3 
under hot isostatic pressure. Solid State Ionics 2004, 172, (1-4), 403-406. 
 
9. Panasyuk, G. P.; Danchevskaya, M. N.; Belan, V. N.; Voroshilov, I. L.; Ivakin, Y. 
D., Phenomenology of corundum crystal formation in supercritical water fluid. Journal of 
Physics: Condensed Matter 2004, 16, S1215-S1221. 
 
10. Panasyuk, G. P.; Belan, V. N.; Voroshilov, I. L.; Shabalin, D. G., Aluminum 
hydroxide transformations during thermal and vapor heat treatments Inorganic Materials 
2008, 44, (1), 45-50. 
 
11. Noguchi, T.; Matsui, K.; Islam, N. M.; Hakuta, Y.; Hayashi, H., Rapid synthesis 
of ?-Al2O3 nanoparticles in supercritical water by continuous hydrothermal flow reaction 
system. The Journal of Supercritical Fluids 2008, 46, (2), 129-136. 
 
12. Liao, J.; Huang, L.; Meng, Y.; Wu, B.; Zhang, L., Phase Transition Actibity 
Characteristics of Nanosized AlOOH in the Hydrothermal Synthesis of nanosized alpha-
Al2O3. Key Engineering Materials 2008, 368-372, 675-678. 
 
 136
13. Danchevskaya, M. N.; Ivakin, Y. D.; Torbin, S. N.; Muravieva, G. P.; 
Ovchinnikova, O. G., Thermovaporous synthesis of complicated oxides. Journal of 
Material Science 2006, 41, 1385-1390. 
 
14. Byrd, A. J.; Pant, K. K.; Gupta, R. B., Hydrogen production from glycerol by 
reforming in supercritical water over Ru/Al2O3 catalyst. Fuel 2008, 87, 2956-2960. 
 
15. Elliott, D. C.; Sealock, L. J.; Baker, E. G., Chemical Procesing in High-Pressure 
Aqueous Environments. 2. Development of Catalysts for Gasification. Industrial 
Engineering Chemistry Research 1993, 32, 1542-1548. 
 
16. Yu, J.; Savage, P. E., Catalyst activity, stability, and transformations during 
oxidation in supercritical water. Applied Catalysis B: Environmental 2001, 31, 123-132. 
 
17. Rossignol, S.; Kappenstein, C., Effect of doping elements on the thermal stability 
of transition aluminas. International Journal of Inorganic Materials 2000, 3, 51-58. 
 
18. Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M., Influence of surface area and 
additives on the thermal stability of transition alumina catalyst supports. I: Kinetic Data. 
Applied Catalysis 1987, 31, 225-238. 
 
19. Piras, A.; Trovarelli, A.; Dolcetti, G., Remarkable stabilization of transition 
alumina operated by ceria under reducing and redox conditions. Applied Catalysis B: 
Environmental 2000, 28, (2), L77-L81. 
 
20. Piras, A.; Colussi, S.; Trovarelli, A.; Sergo, V.; Llorca, J.; Psaro, R.; Sordelli, L., 
Structural and Morphological Inverstigation of Ceria-Promoted Al2O3 under Severe 
Reducing/Oxidizing Conditions. Journal of Physical Chemistry B 2005, 109, 11110-
11118. 
 
21. Schaper, H.; Doesburg, E. B. M.; Van Reijan, L. L., The influence of lanthanum 
oxide on the thermal stability of gamma alumina catalyst supports. Applied Catalysis 
1983, 7, (211-220). 
 
22. Bloch, B.; Ravi, B. G.; Chaim, R., Stabilization of transition alumina and grain 
growth inhibition in ultrafine Al O ?5 wt.% SrO alloy. Materials Letters 2000, 42, 61-65. 
23. Horiuchi, T.; Teshima, Y.; Osake, T.; Sugiyama, T.; Suzuki, K.; Mori, T., 
Improvement of thermal stability of alumina by addition of zirconia. Catalysis Letters 
1999, 62, 107-111. 
 
24. Wu, X.; Yang, B.; Duan, W., Effect of Ce-Zr mixed oxides on the thermal 
stability of transition aluminas at elevated temperature. Journal of Alloys and Compounds 
2004, 376, 241-245. 
 
 137
25. Wei, Z.; Li, H.; Zhang, X.; Yan, S.; Lv, Z.; Chen, Y.; Gong, M., Preparation and 
property investigation of CeO2-ZrO2-Al2O3 oxygen-storage compounds. Journal of 
Alloys and Compounds 2008, 455, (1-2), 322-326. 
 
26. Oudet, F.; Bordes, E.; Courtine, P.; Lambert, C.; Guerlet, J. P., Structural 
consideration with respect to the thermal stability of a new platinum supported 
lanthanum-alumina catalyst. Studies in Surface Science and Catalysis 1987, 30, 313-321. 
 
27. Oudet, F.; Courtine, P.; Vejux, A., Thermal stabilization of transition alumina by 
structural coherence with lanthanide aluminum oxide (LnAlO3, Ln = lanthanum, 
praseodymium, neodymium). Journal of Catalysis 1988, 114, 112-120. 
 
28. Oudet, F.; Vejux, A.; Courtine, P., Evolution during thermal treatment of pure and 
lanthanum-doped Pt/Al/sub 2/O/sub 3/ and Pt-Rh/Al/sub 2/O/sub 3/ automotive exhaust 
catalysts. Transmission electron microscopy studies on model samples. Applied Catalysis 
1989, Medium: X; Size: Pages: 79-86. 
 
29. Humbert, S.; Colin, A.; Monceaux, L.; Oudet, F.; Courtine, P., Simultaneous 
atmosphere and temperature cycling of three-way automotive exhaust catalysts. Studies 
in Surface Science and Catalysis 1995, 96, 829-839. 
 
30. Yao, M. H.; Baird, R. J.; Kunz, F. W.; Hoost, T. E., An XRD and TEM 
Investigation of the Structure of Alumina-Supported Ceria-Zirconia. Journal of Catalysis 
1997, 166, (1), 67-74. 
 
31. Damyanova, S.; Perez, C. A.; Schmal, M.; Bueno, J. M. C., Characterization of 
ceria-coated alumina carrier. Applied Catalysis A: General 2002, 234, (1-2), 271-282. 
 
32. JCPDS-ICDD, Powder Diffraction File. 2001. 
 
 
 
 138
7.  Stability of Binary Aluminum, Titanium, and Zirconium Oxides for Catalyst 
Supports in Sub- and Supercritical Water 
7.1  Abstract 
Hot compressed water is an attractive reaction medium due to its desirable tunable 
thermophysical properties; however, the severe conditions of water?s critical point 
(374.1?C, 22.1 MPa) leads to extensive sintering and phase transformations of traditional 
support materials.  In this work, binary oxides of aluminum, titanium, and zirconium with 
1:1 mole ratios of the component metals were synthesized by a coprecipitation method. 
Their stability in sub- and supercritical water was evaluated at 25 MPa over a temperature 
range of 350 ? 650 ?C for a period of three hours by XRD and BET studies.  The 
compound ZrTiO
4
 was crystallographically stable at all conditions.  It maintained its 
surface area in subcritical water, although it sintered and lost much of its pore volume in 
supercritical water.  ZrO
2
/Al
2
O
3
 maintained high surface area up to 450?C, but sintered 
above this temperature as a result of phase transformation of both ZrO
2
 and Al
2
O
3
.  The 
TiO
2
/Al
2
O
3
 mixed oxide, while having the highest initial surface area, sintered 
extensively following all hydrothermal treatments.  Alumina in the TiO
2
/Al
2
O
3
 system 
hydrolyzed in subcritical water and transformed to corundum in supercritical water, while 
anatase titania was transformed to rutile only at 650?C. 
7.2  Introduction 
Supercritical water is an emerging reaction medium due to its desirable physical 
and chemical properties.  The severity of water?s critical point (371.4?C, 22.1 MPa) often 
leads to sintering and phase transformation of the high surface area metastable metal 
oxide support materials traditionally used in heterogeneous catalysis.  For example, 
 139
transition aluminas are hydrolyzed in subcritical water and transform to corundum at 
higher temperatures.
1, 2
  Pure anatase titania has been reported to be stable in supercritical 
water over the range of 400-600?C,
3, 4
 while mixed rutile-anatase titania has been 
observed to transform completely to the rutile phase following exposure to hot 
compressed water.
2, 5
  Owing to its solubility in supercritical water, silica is inappropriate 
as a support.
6
  Activated carbons have been used as support materials in heterogeneously 
catalyzed reactions, however varying results have been reported.  Carbon supported 
catalysts have been sometimes found to be stable in hot compressed water,
7
 but others 
report that the carbon is slowly gasified.
2, 3, 8
   
Titania can exist in three polymorphs: the metastable tetragonal anatase and stable 
tetragonal rutile phase or the orthorhombic brookite phases.  At atmospheric pressure the 
transformation temperature from anatase to rutile is approximately 800?C, however 
depending on preparation conditions, precursor materials, particle size, and impurities, 
the transformation may take place at temperatures ranging from 400-1200?C.
9, 10
  For 
pure zirconia, the monoclinic form is the stable polymorph at ambient temperature, while 
tetragonal zirconia forms at above 1170?C.
11
  Alumina exhibits a number of metastable 
phases (?, ?, ?, ?, ?, ?) collectively known as transition aluminas, each sharing defect 
spinel structures.  At approximately 1100?C, each of the transition aluminas will 
transform to ?-Al
2
O
3
, which is the only thermodynamically stable phase of alumina.
12
 
 Phase transition temperatures of catalyst supports can be greatly reduced in the 
high pressure supercritical water environment.  Ito et al. observed that the temperature for 
phase transformation of ?-Al
2
O
3
 to the stable ? phase was reduced to 720?C under 50 
MPa and took place at only 500?C under 200 MPa hot isostatic pressure.  Nucleation of 
 140
the ? phase is induced by stress concentration at the points of contact between particles 
and is accelerated by the presence of water.
13
 
The use of a binary oxide can significantly increase the effective operating 
temperature over a single oxide.  A 1:1.3 molar ratio of Al
2
O3:TiO
2
 has been seen to 
increase the anatase to rutile transition temperature by 200?C for supported vanadia 
catalysts.
14
  TiO
2
/ZrO
2
 catalysts have been found to retain reasonably high surface areas 
after high temperature calcinations.
15
  Mixed oxides may also be desirable as they can 
exhibit a larger number of acid sites than the pure component oxides.
16
  In this work we 
examine the stability of binary oxides of aluminum, titanium, and zirconium in sub- and 
supercritical water to evaluate their potential as support materials for heterogeneous 
catalytic reactions.  Nomenclature used to represent mixed oxides in this paper will be 
TiAl for TiO
2
 + Al
2
O
3
, ZrAl for ZrO
2
 + Al
2
O
3
, and ZrTi for ZrTiO
4
. 
 
7.3  Experimental 
7.3.1  Sample Preparation 
The binary oxides were prepared by a coprecipitation method from chloride salts.  
AlCl
3
 (99% purity), ZrCl
4
 (99.5% purity), and TiCl
4
 (99.9% purity) were obtained from 
Sigma Aldrich.  The synthesis of ZrTi was performed as follows:  ZrCl
4
 and TiCl
4
 were 
added dropwise to a volume of water at 60?C with stirring to give a total metal 
concentration of 0.5 M.  NH
4
OH (29.6%) was then slowly added to the solution until the 
pH rose from 1 to 10.  The resulting hydroxide gel was allowed to age for 1h before 
undergoing vacuum filtration.  The gel was rinsed with deionized water until the filtrate 
was free of the chloride anion, as evidenced by a lack of precipitation of AgCl upon 
 141
addition of AgNO
3
 to the filtrate.  The gel was then dried at 110?C overnight, followed 
by a 3 hour calcination at 750?C.  Syntheses of ZrAl and TiAl were analogous to the 
preparation of ZrTi. 
7.3.2  Experimental Procedure 
All hydrothermal treatments were performed in an Inconel 600 tube (OD: 6.35 
mm, ID: 3.05 mm) with pressed steel frits at both ends having pore sizes of 2 ?m.  
Approximately 1.6 g of a binary oxide was loaded into the reactor tube to fill it halfway, 
allowing a 0.5 g sample to be removed from the middle portion of the reactor after 
hydrothermal treatment.  This ensured that the sample taken for analysis was not exposed 
to uneven heating near the ends of the reactor.  A schematic of the experimental 
apparatus is shown in Figure 7.1.  
T
HPLC 
Pumps 
Water-Cooled 
Heat Exchanger 
Fixed Bed 
Tube Furnace with 
Temperature 
Controller 
P 
Back 
Pressure 
Regulator 
 
Water  
 Water  Preheater 
N
2 
Collect 
Aqueous 
Sample 
P 
Pressure 
Relief 
Valve 
V-2 
V-1 
V-3 
V-4 
V-5 
 
 
Figure 7.1:  Experimental apparatus for hydrothermal treatment of binary oxides 
 
The reactor was placed inside of a Thermolyne 21100 tubular furnace with the 
ends covered in ceramic insulation with additional insulation placed around the pressure 
 142
fittings protruding from the furnace to reduce heat losses.  The furnace was then brought 
to the desired temperature and held for ten minutes before pressurization with water.  A 
fixed pressure of 25 MPa was used in all experiments.  Water was supplied at 2 mL min
-1
 
by an HPLC pump (Alltech 301) and pumped through a 2 m section of 1/16? tubing 
serving as a preheater before entering the reactor.   The time at which the desired 
temperature and pressure were achieved was taken as t = 0 in each experiment, and the 
reactor was then isolated by closing valves V-3 and V-5.  After 3 h had elapsed, nitrogen 
gas slightly above the pressure of the hydrothermal system was applied and needle valve 
V-4 was slightly opened to collect water from the system.  This prevented many of the 
samples from being exposed to subcritical water at the end of the run and also dried the 
samples as the system depressurized.  High pressure nitrogen was supplied by means of a 
1.5 L high pressure vessel (HiP TOC31-10-P) equipped with a close fitting piston.  Prior 
to the start of an experiment, nitrogen gas was supplied to one side of the piston at 
cylinder pressure by a high pressure regulator.  The high pressure side of the piston was 
then isolated and water was pumped to the other side to increase pressure to the desired 
level.  The time to drain water from the system was 5 minutes, after which nitrogen gas 
was bled out for another ten minutes before fully opening valve V-4.  Upon the increase 
in nitrogen flow, the furnace was quickly cooled by blowing compressed air into the 
furnace while the remaining gas from the piston passed through the bed.  The total time 
to depressurize was approximately 30-40 minutes, during which the sample was exposed 
to nitrogen at high temperature and pressure. 
 
 
 143
7.3.3  Characterization 
 X-ray spectra were collected on a Discover D8 diffractometer equipped with a Cu 
K?
1
 radiation source. The diffractometer was run at 40 kV voltage and 40mA current, 
and scanned at 6?/min with 0.01? step size.  Specific surface areas and pore volume data 
were determined by nitrogen physisorption at 77 K after vacuum outgassing for 3 h at 
300?C on a Quantachrome NOVA 2200e.   
 
7.4  Results and Discussion 
7.4.1  Crystalline structure 
X-ray diffraction spectra of the ZrTi samples after calcination and subsequent 
hydrothermal treatment are shown in Fig. 7.2.  Freshly calcined ZrTi shows diffraction 
pattern from a single ZrTiO
4
 phase, which presents an orthorhombic structure.
17
  There is 
no evidence of TiO
2
 or ZrO
2
, as expected after calcination above 700?C.
15
  Mechanical 
stresses from cold pressing ZrTiO
4
 at 147 MPa have been seen to induce partial 
decomposition of ZrTiO
4
 to monoclinic zirconia at temperatures as low as 300?C;
18
 
however, the crystalline structure of ZrTi remained unchanged after exposure to sub- and 
supercritical water.   
 144
20 25 30 35 40 45 50 55 60 65 70
2 ? (?)
In
t
e
n
s
it
y
 (
a
.
u
.
)
 
e 
d 
c 
b 
a 
Figure 7.2: XRD spectra of ZrTi (a) after calcination, and after hydrothermal treatment at 
25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.  All peaks correspond to 
ZrTiO
4
 
 
XRD spectra from TiAl samples are shown in Fig. 7.3.  Calcined TiAl shows only 
a crystalline anatase TiO
2
 phase.  The lack of any diffraction peaks from alumina 
indicates that all alumina present in the calcined sample is amorphous.  Following 
hydrothermal treatment below 650?C only the anatase polymorph of titania is present, 
although as the temperature of the exposure was increased, the peaks become narrower 
and more intense, indicating increased crystallinity of the tetragonal phase.  At 650?C, 
transformation from the anatase to rutile is evidenced by the appearance of the peak at 
27.5?.  The amorphous state of alumina was changed after each hydrothermal treatment 
in both sub- and supercritical water.  In subcritical water at 350 ?C the peak at 23.16 ?C is 
indicative of the formation of AlO(OH) (boehmite) from the hydrolysis of alumina.  In all 
cases after exposure to supercritical water the only phase of alumina visible from XRD is 
?-Al
2
O
3
. 
 145
20 30 40 50 60 70
2 ? (?)
In
te
n
s
it
y
 (a
.
u
.)
e 
d 
c 
b 
a 
 
Figure 7.3: XRD spectra of TiAl (a) after calcination, and after hydrothermal treatment at 
25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.  Labels correspond to (?) 
AlO(OH), (?) ?-Al
2
O
3
, ({) anatase TiO
2
, (z) rutile TiO
2
 
 
The XRD spectra of calcined ZrAl also shows no evidence of a crystalline  
Al
2
O
3
 phase.  The only phase detectable by XRD after calcination is tetragonal ZrO
2
.  
Tetragonal zirconia is normally formed above 1170?C and reverts to the monoclinic form 
at lower temperatures.
11
  The presence of alumina, however, stabilizes the tetragonal 
phase.
19
  Depending on preparation conditions and ratios of the two metals, the exclusive 
formation of tetragonal zirconia in the presence of alumina may be achieved, 
20, 21
 
however it is also possible to obtain a mixture of tetragonal, monoclinic, and cubic 
zirconia.
22
  After exposure to subcritical water ZrAl showed a weak reflection at around 
14.5? (not shown), evidencing the hydrolysis of a small amount of alumina to form 
boehmite.  No changes in the alumina are seen at 450?C, however at 550?C and above ?-
 146
Al
2
O
3
 can clearly be seen by the peak at about 43.4?.  After each hydrothermal treatment 
there is some recrystallization of zirconia to the monoclinic form, although only a small 
amount is transformed up to 550?C.  Following hydrothermal treatment at 650?C, 
however, reflections for the monoclinic phase are much more intense than at lower 
temperatures.  In the TiAl and ZrAl samples, no evidence for a transition through the ?-
phase of alumina before the ?-phase is observed, as has been for CeO
2
 coated ?-Al
2
O
3
.
23
 
20 30 40 50 60 70
2 ? (?)
In
t
e
n
s
ity
 (
a
.u
.)
 
e 
d 
b 
c 
T
a 
Figure 7.4: XRD spectra of ZrAl (a) after calcination, and after hydrothermal treatment at 
25 MPa and (b) 350?C, (c) 450?C, (d) 550?C, and (e) 650?C.  Labels correspond to (?) t-
ZrO
2
, (?)m-ZrO
2
, (?)AlO(OH), (?) ?-Al
2
O
3
 
 
7.4.2  Surface area and pore characterization 
Results from the characterization of the surface area and pores of the supports are 
presented in Table 1.  After hydrothermal treatment at 350?C, ZrAl was mostly 
unchanged, with only a slight decrease in surface area.  A noticeable increase in surface 
area was seen after hydrothermal treatment at 450?C; pore volume was unchanged, 
although the average pore radius slightly decreased from 38 ? to 29 ?.  The large 
 147
increase in pore size after exposure to supercritical water at 550?C is likely due to the 
collapse of smaller pores while extensive sintering at 650?C, indicated by the large 
decrease in pore volume, decreases the average pore size.  
 
Table 7.1:  Surface Area, Average Pore Size, and Pore Volume of Binary Oxides 
After Calcination and Subsequent Hydrothermal Treatment at 25 MPa 
 
ZrAl  ZrTi  TiAl 
 
Surfac
e Area 
(m
2
 g
-1
) 
Averag
e Pore 
Radius 
(?) 
Pore 
Volum
e 
(cc/g) 
 
Surfac
e Area 
(m
2
 g
-1
) 
Averag
e Pore 
Radius 
(?) 
Pore 
Volum
e 
(cc/g) 
 
Surfac
e Area 
(m
2
 g
-1
) 
Averag
e Pore 
Radius 
(?) 
Pore 
Volume 
(cc/g) 
Calcine
d 81.7 38 0.15  30.8 172 0.26  118.5 61 0.36 
350?C 80.5 37 0.15  31.9 171 0.27  55.3 110 0.30 
450?C 105.3 29 0.15  21.9 105 0.11  21.6 85 0.09 
550?C 19.9 132 0.13  17.4 51 0.04  10.0 50 0.02 
650?C 8.3 41 0.02  19.5 61 0.06  6.0 44 0.01 
 
ZrTi had the lowest initial surface area of the binary oxides prepared.  Like ZrAl, 
it was largely unaffected after exposure to subcritical water at 350?C and 25 MPa.  At 
450?C there was a decrease in surface area and pore volume with more extensive 
sintering at 550? and 650?C.  Of the three binary oxides tested, however, ZrTi retained 
the largest surface area and pore volume after hydrothermal treatment at the most severe 
condition. 
The TiAl sample, while having the largest surface area of the three binary oxides 
following calcination, sintered extensively after all hydrothermal treatments.  Whereas 
the other two oxides were stable at 350?C, TiAl lost over 50% of its surface area through 
sintering after exposure to subcritical water at 350?C.  The degree of sintering increased 
with further increases in temperature. 
 148
 
7.5  Conclusions 
The hydrothermal stability of binary oxides of titanium, zirconium, and aluminum 
was evaluated in sub- and supercritical water at 25 MPa over the temperature range 350?-
650?C.  The TiO
2
/Al
2
O
3
 sample had the highest initial surface area but sintered 
extensively after all hydrothermal treatments, with alumina in the samples being 
hydrolyzed in subcritical water and transformed to corundum in supercritical water.  
Anatase titania was transformed to rutile only at 650?C.  ZrO
2
/Al
2
O
3
 maintained high 
surface area up to 450?C, but sintered above this temperature as a result of phase 
transformation of both ZrO
2
 and Al
2
O
3
.  The compound ZrTiO
4
 was crystallographically 
stable at all conditions.  In subcritical water ZrTiO
4
 maintained its pore structure and 
surface area. In supercritical water its surface area was reduced by approximately 30%, 
but at the most severe condition it retained the largest specific surface area and pore 
volume of the three compounds evaluated.   
 
 
 
 149
References 
1. Yu, J.; Savage, P. E., Catalyst activity, stability, and transformations during 
oxidation in supercritical water. Applied Catalysis B: Environmental 2001, 31, 123-132. 
 
2. Elliott, D. C.; Sealock, L. J.; Baker, E. G., Chemical Processing in High-pressure 
Aqueous Environments. 2. Development of Catalysts for Gasification. Industrial 
Engineering Chemistry Research 1993, 32, 1542-1548. 
 
3. Osada, M.; Sato, O.; Arai, K.; Shirai, M., Stability of Supported Ruthenium 
Catalysts for Lignin Gasification in Supercritical Water. Energy and Fuels 2006, 20, 
2337-2343. 
 
4. Byrd, A. J.; Kumar, S.; Ramsurn, H.; Kong, L.; Gupta, R. B., Hydrogen 
production from catalytic gasification of switchgrass biocrude in supercritical water. 
International Journal of Hydrogen Energy 2011. 
 
5. Elliott, D. C.; Hart, T. R.; Neuenschwander, G. G., Chemical Processing in High-
Pressure Aqueous Environments. 8. Improved Catalysts for Hydrothermal Gasification. 
Industrial Engineering Chemistry Research 2006, 45, 3776-3781. 
 
6. Kennedy, G. C., A portion of the system silica-water. Economic Geology 1950, 
45, 629-653. 
 
7. Yamaguchi, A.; Hiyoshi, N.; Sato, O.; Bando, K. K.; Osada, M.; Shirai, M., 
Hydrogen production from woody biomass over supported metal catalysts in supercritical 
water. Catalysis Today 2009, 146, 192-195. 
 
8. Matsumura, Y.; Xu, X.; Antal, M. J. J., GASIFICATION CHARACTERISTICS 
OF AN ACTIVATED CARBON IN SUPERCRITICAL WATER. Carbon 1997, 35, (6), 
819-824. 
 
9. Kumar, K.-N. P.; Keizer, K.; Burggraf, A. J., Stabilization of the porous texture of 
nanostructured titania by avoiding a phase transformation. Journal of Materials Science 
Letters 1994, 13, 59-63. 
 
10. Santos, J. G.; Ogasawara, T.; Correa, R. A., Synthesis of nanocrystalline rutile-
phase titania at low temperatures. Material Science Poland 2009, 27, 1067-1076. 
 
11. Block, S.; Jornada, J. A. H. d.; Piermarini, G. J., Pressure-Temperature Phase 
Diagram of Zirconia. Journal of the American Ceramic Society 1985, 68, 497-499. 
 
12. Gitzen, W. H., Alumina as a Ceramic Material. The American Ceramic Society, 
Inc.: Columbus, Ohio, 1970. 
 
 150
13. Ito, S.; Umehara, N.; Hidenori, T.; Fujii, T., Phase transition of g-Al2O3 under 
hot isostatic pressure. Solid State Ionics 2004, 172, 403-406. 
 
14. Reddy, B. M.; Rao, K. M.; Reddy, G. K.; Bharali, P., Characterization and 
catalytic activity of V2O5/Al2O3-TiO2 for selective oxidation of 4-methylanisole. 
Journal of Molecular Catalysis A: Chemical 2006, 253, 44-51. 
 
15. Reddy, B. M.; Chowhurdy, B., Dispersion and Thermal Stability of MoO3 on 
TiO2-ZrO2 Mixed Oxide Support. Journal of Catalysis 1998, 179, 413-419. 
 
16. Zou, H.; Lin, Y. S., Characterization and catalytic activity of V2O5/Al2O3-TiO2 
for selective oxidation of 4-methylanisole. Applied Catalysis A: General 2004, 265, 35-
42. 
 
17. Newnham, R. E., Crystal Structure of ZrTiO4. Journal of the American Ceramic 
Society 1967, 50, 216. 
 
18. Cosentino, I. C.; Muccillo, E. N. S.; Muccillo, R.; Vichi, F. M., Low-Temperature 
Sol-Gel Synthesis of Single-Phase ZrTiO4 Nanoparticles. Journal of Sol-Gel Science and 
Technology 2006, 37, 31-37. 
 
19. Reddy, B. M.; Sreekanth, P. M.; Yamada, Y.; Kobayashi, T., Surface 
characterization and catalytic activity of sulfate-, molybdate- and tungstate-promoted 
Al2O3?ZrO2 solid acid catalysts. Journal of Molecular Catalysis A: Chemical 2005, 227, 
81-89. 
 
20. Amairia, C.; Fessi, S.; Ghorbel, A., Sol gel derived Pd/Al2O3-ZrO2 as catalysts 
for methane combustion: effect of zirconium loading. Journal of Sol-Gel Science and 
Technology 2010, 54, 29-35. 
 
21. Soisuwan, S.; Panpranot, J.; Trimm, D. L.; Praserthdam, P., A study of alumina?
zirconia mixed oxides prepared by the modified Pechini method as Co catalyst supports 
in CO hydrogenation. Applied Catalysis A: General 2006, 303, 268-272. 
 
22. Dominguez, J. M.; Hernandez, J. L.; Sandoval, G., Surface and catalytic 
properties of Al2O3?ZrO2 solid solutions prepared by sol?gel methods. Applied 
Catalysis A: General 2000, 197, 119-130. 
 
23. Byrd, A. J.; Gupta, R. B., Stability of cerium-modified gamma-alumina catalyst 
support in supercritical water. Applied Catalysis A: General 2010, 381, 177-182. 
 151
8.  Overall Conclusions 
 
 This dissertation has focused on production strategies for the environmentally 
friendly energy carrier hydrogen from renewable resources by catalytic gasification and 
reforming in supercritical water.  The associated problem of instability of catalyst 
supports in the supercritical water environment has also been addressed by developing 
several new metal oxide supports.  Parametric studies have been performed for the 
catalytic reforming of the simple molecules ethanol and glycerol, as well as glucose and 
biocrude which have more complex reaction networks.  In these studies the effect of 
temperature, pressure, feedstock concentration, residence time, and catalyst has been 
investigated.   Many parallels can be drawn from the results of the glucose, glycerol and 
ethanol studies.  Hydrogen yields increase with increasing temperature, as can be 
expected from thermodynamic equilibrium, however further increases in pressure in the 
supercritical region studied had negligible effect on product yields.  In all cases hydrogen 
concentration in the product gases was highest when the feed concentration was low, 
while methane dominated with higher feed concentrations.  Hydrogen yield is maximized 
when the formation of methane is minimized, as hydrogen is consumed in the formation 
of methane.  By operating the supercritical water reactor with a short residence time it is 
possible to suppress methanation to a large extent.  General kinetic models were 
developed for hydrogen production in supercritical water and rate constants were 
reported over the temperature ranges studied.   
In the studies using glucose, glycerol, and ethanol as a feedstock a commercially 
available Ru/?-Al
2
O
3
 catalyst was used.  While Ru was a good catalyst, the ?-Al
2
O
3
 
 152
support was unstable at the temperatures and pressures associated with supercritical 
water.  Accordingly, a suite of catalysts were developed for testing in the supercritical 
water gasification of biocrude.  Further, the catalysts were characterized by XRD, BET, 
and TGA to evaluate their stability in supercritical water and tendency toward char 
formation.  Switchgrass was first liquefied in subcritical water using a batch reactor 
before catalytic gasification in a separate step.  Fixed temperature, pressure, feed 
concentration, and space time were used to evaluate Ru, Ni, and Co catalysts on anatase 
TiO
2
, monoclinic ZrO
2
, and MgAl
2
O
4
 supports to make a total of nine catalysts.  This 
was the first report of MgAl
2
O
4
 being used as a support in supercritical water.  In spite of 
the lower inherent surface acidity of the support material, it was seen to char extensively 
and rapidly upon exposure to a carbonaceous feedstock. A given metal supported on ZrO
2
 
gave a higher conversion of biocrude than those supported on TiO
2
 due to the catalytic 
activity of zirconia, although charring at the face of the catalyst bed led to reactor 
plugging within a few hours for ZrO
2
 supported catalysts.  No plugging was observed for 
TiO
2
 supported catalysts, which also had the smallest amount of char formation.  The 
highest hydrogen yield was obtained with Ni/ZrO
2
, while the lowest was with Ru/ZrO
2
, 
which gave a product gas composed of mostly methane and CO
2
.  The anatase titania and 
monoclinic zirconia were crystallographically stable, but both lost significant amounts of 
surface area through hydrothermal sintering.  Following SCW exposure MgAl
2
O
4
 also 
sintered and was found to partially transform to ?-Al
2
O
3
. 
Other support materials were also developed as potential catalyst supports for the 
severe supercritical water environment.  A series of Ce-modified ?-aluminas were 
prepared by an incipient wetness method followed by a reducing treatment which were 
 153
then exposed to supercritical water for several hours at 250 bar and 500 - 700?C.  XRD 
and BET studies showed that the inclusion of cerium slowed phase transformations in the 
supercritical water environment; however, the ?-phase was transformed in all cases and 
was accompanied by the loss of a large percentage of surface area.  Compared to plain 
alumina samples, the Ce-modified aluminas showed some resistance to being hydrolyzed 
at 500?C, while at higher temperatures the transformation towards the thermodynamically 
stable ?-Al
2
O
3
 was slowed.  The first observation was made of transition through the ?-
Al
2
O
3
 phase during the transformation sequence towards corundum.  Low cerium 
loadings (1-2 wt%) maintained the highest BET surface areas at 500-600?C, while at 
700?C all loadings retained approximately 30-40 m
2
 g
-1
, compared to 2 m
2
 g
-1
 for 
unmodified alumina. 
Binary oxides of aluminum, titanium, and zirconium were also synthesized and 
tested as potential support materials for supercritical water applications.  These oxides 
were tested at 250 bar in both sub- and supercritical water over the temperature range 350 
? 650?C.  The TiO
2
/Al
2
O
3
 sample had the highest initial surface area but sintered 
extensively after all hydrothermal treatments, with alumina in the samples being 
hydrolyzed in subcritical water and transformed to corundum in supercritical water.  
Anatase titania was transformed to rutile only at 650?C.  ZrO
2
/Al
2
O
3
 maintained high 
surface area up to 450?C, but sintered above this temperature as a result of phase 
transformation of both ZrO
2
 and Al
2
O
3
.  The compound ZrTiO
4
 was crystallographically 
stable at all conditions.  In subcritical water ZrTiO
4
 maintained its pore structure and 
surface area. In supercritical water its surface area was reduced by approximately 30%, 
 154
but at the most severe condition it retained the largest specific surface area and pore 
volume of the three compounds evaluated.   
 155
9.  Directions for Future Work 
 
There is still much work to be done to develop hydrogen production strategies 
from biomass derived feedstocks.  With small molecules like ethanol and glycerol it is 
possible to feed high concentrations to the reactor without coke formation on the catalyst; 
however, thermodynamics dictates a higher selectivity towards methane when the feed 
concentration is increased.  We have seen that methanation can be suppressed by utilizing 
a short reaction time.  Here lies an opportunity for design of more complex reactors than 
the simple tubular packed beds that were employed in this research.  Reactors having 
better heat transfer characteristics could heat reactants to the desired temperature quickly 
and prevent lowering of temperature in the reactor from the endothermic reactions taking 
place, especially as throughput is increased to further reduce residence time.  Working 
with the more complex feedstock biocrude it was also found that char formation took 
place primarily at the face of the catalyst bed where heating was taking place.  Since it is 
known that char formation is related to heating rate, the rapid heating rate that could 
potentially realized with new reactor designs would further minimize char formation and 
increase gas yields. 
 Continued efforts are needed to further develop support materials capable of 
withstanding the supercritical water environment and to find less expensive catalytic 
materials.  Some of the catalyst supports developed in this research were able to retain 
their morphological properties under limited high temperature and high pressure aqueous 
conditions, but this remains a challenge for future researchers to develop materials able of 
retaining high surface areas.  Ruthenium metal has been found to be a good catalyst in 
 156
supercritical water gasification; however its high cost makes it prohibitive for large scale 
applications.  Other researchers have found that noble metals can be replaced with 
combinations of base metals in other applications, in some cases exceeding the activity of 
the most active single element catalysts. Perhaps bimetallic base metal catalysts have a 
place in hydrogen production that has not been discovered by researchers yet. 
 Another area of study ripe with opportunity stemming from this research is the 
development of effective strategies for gas separation.  The research of this dissertation 
has not addressed separation or purification of the produced gases.  One promising 
possibility is a catalytic membrane reactor with in situ hydrogen separation.  The 
development of a catalytic membrane reactor for hydrogen separation would offer several 
advantages.  First, it would provide a high purity hydrogen stream for any application.  
Secondly, it would replace expensive processes such as pressure swing adsorption 
commonly used for hydrogen purification.  A third boon lies in the ability of a membrane 
reactor to achieve conversions beyond the dictations of thermodynamics.  As the product 
hydrogen is removed in situ, the equilibrium is shifted towards products.  A reactor of 
this type can also serve to minimize methanation.  Hydrogen produced in the reactor will 
be separated so it cannot react with other products to form methane. 
 The high pressures and temperatures found in supercritical water reactors provide 
large driving forces for porous ceramic membranes, the permeance of which scales 
directly with pressure.  Proton conducting membranes have been recently studied for use 
in solid oxide fuel cells and could be further developed for hydrogen separation.  A major 
shortcoming of both of these types of inorganic membranes is poor mechanical strength 
 157
and poor stability in a hydrothermal environment.  These materials could add 
significantly to hydrogen production technologies if these limitations are overcome. 
 158
Appendix 1 ? Determination of Gas Composition 
 An SRI 8610C gas chromatograph equipped with a 15? x 1/8? Supelco 60/80 
Carboxen 1000 and a thermal conductivity detector was used to determine gas 
composition in all supercritical water gasification and reforming experiments.  The oven 
was operated isothermally at 200?C, and nitrogen or helium supplied at 15 psi
g
 was used 
as carrier gas.  An online 150 ?Lgas sampling loop from Supelco was used to inject the 
sample.  The GC essentially has two parts: a packed column in an oven to separate the 
mixture into its components and a detector to quantify them.  The detector compares the 
thermal conductivity of the carrier gas flowing in a reference cell to another gas in the 
sample cell and registers the difference as a mV signal via a Wheatstone bridge.  As a gas 
passes through the detector the signal will quickly rise then fall back to the baseline.  
When plotted against time the signal will show a series of peaks for each component. 
 The area of each peak is proportional to it?s concentration in the mix.  A gas 
having a known composition similar to a typical product gas was used for calibration.  
Composition of a typical calibration gas is given in Table A1.  To calculate the response 
factor from the calibration gas, first the number of moles of each gas was found by 
multiplying the total moles in the sample loop by each specie?s molar concentration and  
assuming ideal gas behavior.  Pressure and temperature were precisely known from an 
electronic gas flowmeter.  Knowing the order of elution of the gases from the column 
manufacturer, a multiplication factor having units of moles per area could be calculated 
for each gas.  By multiplying the peak area obtained from a sample of unknown 
concentration by this factor, the number of moles of each component can be calculated. 
 
 159
Table A1.1:  Composition of Calibration Gas in mol% 
H2 CO CH4 CO2 C2H4 C3H6 
50.886 9.709 9.907 19.800 4.844 4.854
 160
Appendix 2 ? Calculation of Gas Flow using Viscosity Correction 
 In Chapters 2,3,4, and 5 gas flow was measured by an Omega 1607A mass 
flowmeter.  This flowmeter measures the pressure drop across a venturi tube and 
calculates flowrate for a given gas using the viscosity.  The flowmeter was operated in 
hydrogen mode as it was typically the major component of the product gas.  The gas flow 
being measured, however, is not a single component but a mixture.  In order to obtain 
accurate information from the flowmeter the viscosity of the mix must be calculated.  
Viscosity of the mix was calculated from Wilke?s empirical formula: 
?
?=
?
=
N
mix
x
x
1?
??
?
?
??
?
?                                          Eq. A.2.1 
where 
2
4/1
2/12/1
11
8
1
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
+
?
?
?
?
?
?
?
?
+=?
?
?
?
?
?
?
?
??
?
?
M
M
M
M
                     Eq. A.2.2 
 
and where M
i
 is the molecular weight of species i, ?
i
 is the viscosity of species i, and x
i
 is 
the mole fraction of species i.  Dividing the viscosity of hydrogen by the calculated 
viscosity of the mix gives a correction factor to multiply the measured flowrate by to 
obtain the flowrate of the gas mix. 
 161
Appendix 3 ? List of Publications and Presentations 
JOURNAL ARTICLES 
1.  Byrd, Adam J.; Gupta, Ram B. Stability of Binary Oxides of Aluminum, Titanium,     
        and Zirconium in Sub- and Supercritical Water.  (submitted to Industrial &  
        Engineering Chemistry Research) 
2.  Byrd, Adam J.; Kumar, S.; Kong, L.; Ramsurn, H.; Gupta, R. B. Hydrogen   
       Production from Catalytic Gasification of Switchgrass Biocrude in Supercritical  
       Water. International Journal of Hydrogen Energy (2011) available online  
        doi:10.1016/j.ijhydene.2010.12.026 
3.  Byrd, Adam J.; Gupta, Ram B. Stability of cerium-modified ?-alumina catalyst 
       support in supercritical water. Applied Catalysis, A: General (2010), 381(1-2), 
       177-182. 
4.  Byrd, Adam J.; Pant, K. K.; Gupta, Ram B. Hydrogen production from glycerol by 
       reforming in supercritical water over Ru/Al
2
O
3
 catalyst. Fuel (2008), 87(13-14),    
        2956- 2960. 
5.  Byrd, Adam J.; Pant, K. K.; Gupta, Ram B. Hydrogen Production from Ethanol by 
       Reforming in Supercritical Water Using Ru/Al
2
O
3
 Catalyst. Energy & Fuels  
       (2007), 21(6), 3541-3547. 
6.  Byrd, Adam J.; Pant, K. K.; Gupta, Ram B. Hydrogen Production from Glucose 
       Using Ru/Al
2
O
3
 Catalyst in Supercritical Water. Industrial & Engineering 
       Chemistry Research (2007), 46(11), 3574-3579.  
 
CONFERENCE PRESENTATIONS 
1.  Byrd, Adam J.; Kumar, S.; Gupta, R.B. Hydrogen Production From Liquefied 
       Switchgrass in Supercritical Water Over Ru, Ni, and Co Catalysts.  AIChE 
       National Meeting, Salt Lake City, UT, United States, 2010 
 
2.  Kumar, S.; Byrd, A.J.; Gupta, R.B. Sub- and Supercritical Water Technology for 
       Biofuels:  Swithchgrass to Ethanol, Biocrude, and Hydrogen.  31st Symposium   
       on Biotechnology for Fuels and Chemicals.  San Francisco, CA, United States. 2009. 
3.  Byrd, Adam J.; Gupta, Ram B. Stability of Ce-Doped Alumina Catalyst Support in 
       Supercritical Water.  AIChE National Meeting, Nashville, TN, United States, 2009. 
4.  Byrd, Adam J.; Kumar, S.; Gupta, R.B. Cellulose Liquefaction and Hydrogen 
       Production in Supercritical Water.  AIChE National Meeting, Salt Lake City, UT, 
       United States, 2008. 
5.  Byrd, Adam J.; Pant, K. K.; Gupta, Ram B.  Hydrogen Production from Glucose 
       Using Ru/Al
2
O
3
 Catalyst in Supercritical Water.  AIChE National Meeting, Salt 
       Lake City, UT,   United States, 2007. 
6.  Byrd, Adam J.; Gupta, Ram B. Hydrogen production by reforming in    
       supercritical water. 231st ACS National Meeting, Atlanta, GA, United States, 
       March 26-30, 2006. 
 
 
 
 162
 163
 
Posters 
 
1.  Byrd, Adam J.; Gupta, R.B.  Hydrogen Production by Reforming in Supercritical   
       Water.   Alternative Energy Solutions from Alabama's Natural Resources. Auburn,   
       AL, United States, 2006 
 
2.  Byrd, Adam J. and Gupta, Ram B.  Stability of Binary Oxides of Al, Ti, and Zr 
        in Sub- and Supercritical Water.  AIChE National Meeting, Salt Lake City, UT, 
       United States, 2010