A systematic study of the effect of chemical promoters on the
precipitated Fe-based Fischer-Tropsch Synthesis catalyst
Wonga Mpho Hexana
A thesis submitted to the Faculty of Science, University of the Witwatersrand,
Johannesburg, in fulfillment of the requirements for the Degree of Doctor of Philosophy.
Johannesburg, 2009
i
Declaration
I declare that this thesis is my own, unaided work. It is being submitted for the Degree of
Doctor of Philosophy in the University of the Witwatersrand, Johannesburg. It has not
been submitted before for any degree or examination in any other University.
Signature of candidate
????????day of ???????????..2009
ii
Abstract
In recent years, research interest on improving the catalytic properties of precipitated Fe-
based Fischer-Tropsch synthesis (FTS) catalysts has grown immensely. In particular the
effect of promoters on these type of catalysts has attracted much attention. Classical
promoters such as copper, potassium and silica are nowadays employed for preparing
commercially used FTS catalysts. The promoted catalysts are the catalysts of choice and
have been shown to possess chemical properties that improve the catalytic properties of
Fe-based Fischer- Tropsch synthesis catalysts.
In this thesis we attempted to systematically study effects caused by these promoters as
well as the effect of indium as a promoter for Fe. Silica which is often used as a structural
promoter and often forms a large portion of the catalyst was added in significantly small
amounts and its chemical promotional ability was investigated. It was found that
increasing the loading of silica affected the carburization and the reduction properties of
the precipitated Fe-based Fischer Tropsch synthesis catalyst. This had an effect on the
Fe/Cu and Fe/K2O contacts when the loading was increased. With regard to this study, it
was found that silica decreased the activity of the catalyst and shifted the hydrocarbon
selectivity to low weight hydrocarbons. All catalysts used in these studies were
characterized using N2 physisorption, TPR, DRIFTS, XPS and XRD.
Indium was also evaluated as a chemical promoter to a precipitated Fe-based Fischer-
Tropsch synthesis catalyst since it is believed that it may induce similar chemical effects
as that found for copper. Indeed, it was observed that indium does possess some similar
chemical properties to that of copper and also that it affected the precipitated Fe-based
Fischer-Tropsch synthesis catalyst in similar ways to copper. It was also instead realized
that indium acted as a poorer promoter than Cu for the Fe-based Fischer-Tropsch
synthesis catalyst and this was attributed to indium having a low melting point than
copper. It was also found that indium acted as a poorer promoter when it was added as a
co-promoter to an iron catalyst that contained potassium and silica. It was found that
indium lowered the activity of the Fischer-Tropsch synthesis reaction as well as the
iii
Water Gas Shift reaction. This was related to a decrease in surface area of the catalyst
after the addition of indium. The selectivity was shifted to the production of heavy weight
hydrocarbons due to the Fe/K2O contact being promoted. Characterisation techniques
such as N2 physisorption, TPR and DRIFTS were employed to elucidate the findings.
iv
Acknowledgements
I am thankful to the following people and institutions:
1. My supervisor, Professor Neil Coville, for intellectual input and mental stimulation
2. School of chemistry, for providing research facilities
3. The Catomat research group, for a great research environment and invaluable
assistance during the tenure of the project
4. Aberdeen University, for being a home away from home
5. Dr Dave Morgan, for XPS measurements and his immaculate XPS knowledge
6. RS (London) for funding the UK visit
7. My family for moral support
8. University of the Witwatersrand
9. Sasol
10. Canon Collins Education Trust
v
Presentations and publications arising from this study
Poster Presentations
1. CATSA Conference 2005, Midrand, Johannesburg, South Africa
Characterisation of iron-based Fischer Tropsch catalysts using Raman spectroscopy,
XRD and TPR
2. CATSA Conference 2006, Mossel Bay, South Africa
Characterisation of an Fe based Fischer Tropsch synthesis catalyst using DRIFTS
3. SpectroCat: Vibrational Spectroscopy for Catalysis 2007, Caen, France
?In Situ? high pressure DRIFTS vs transmission IR
4. International Catalysis Congress 2008, Seoul, South Korea
In-situ characterization of an iron based Fischer Tropsch catalyst using DRIFTS at high
pressure
Oral presentations
1. University of the Witwatersrand Postgraduate Symposium 2008, Johannesburg, South
Africa
Developing a catalyst for the production of petrol!
2. CATSA Conference 2008, Parys, South Africa
Can indium be used as a promoter for an iron-based Fischer-Tropsch synthesis catalyst?
vi
Publications
1. W. M. Hexana, N.J. Coville, ?Effect of SiO2 on a promoted and unpromoted Fe-based
Fischer-Tropsch synthesis catalyst? To be submitted
2. W. M. Hexana, J. A. Anderson, N.J. Coville, ?Effect of Cu on a promoted and unpromoted
Fe-based Fischer-Tropsch synthesis catalyst? To be submitted
3. W. M. Hexana, N.J. Coville, ?Evaluating indium as a chemical promoter in Fe-based
Fischer-Tropsch synthesis? To be submitted
4. W. M. Hexana, N.J. Coville, ?Effect of In on a promoted and unpromoted Fe-based
Fischer-Tropsch synthesis catalyst? To be submitted.
vii
Contents
Declaration........................................................................................................................ i
Abstract ............................................................................................................................ ii
Acknowledgements........................................................................................................ iiv
Presentations and publications arising from this study.................................................... v
Table of Contents........................................................................................................... vii
List of Tables ............................................................................................................... xiiii
List of Figures ................................................................................................................ xv
Abbreviations and acronyms......................................................................................... xxi
Table of Contents
CHAPTER 1 INTRODUCTION ..................................................................................... 1
CHAPTER 2 FISCHER-TROPSCH SYNTHESIS (FTS): LITERATURE REVIEW ... 3
2.1 Introduction.............................................................................................................. 3
2.2 History of the Fischer-Tropsch process ..................................................................... 3
2.3 Utilization of the FT process ................................................................................... 4
2.4 Three main steps in the FT process............................................................................ 7
2.4.1 Synthesis gas production......................................................................................... 7
2.4.2 FT synthesis: Process conditions ............................................................................ 8
2.4.3 Product Upgrading and Separation ......................................................................... 8
2.5 FT reactors ............................................................................................................... 9
viii
2.6 The chemistry behind the FT process ...................................................................... 15
2.6.1 Reactions............................................................................................................... 15
2.6.2 Mechanism and product selectivity ...................................................................... 16
2.7 Fischer-Tropsch catalysts...................................................................................... 23
2.7.1 Use of Fe catalysts in FTS .................................................................................... 23
2.8 Promoters................................................................................................................ 26
2.8.1 Structural promoters.............................................................................................. 26
2.8.2 Chemical Promoters.............................................................................................. 29
References ..................................................................................................................... 34
CHAPTER 3 EXPERIMENTAL................................................................................... 42
3.1 Catalyst preparation .............................................................................................. 42
3.2 Catalyst characterization ...................................................................................... 43
3.2.1 X-Ray Fluorescence (XRF) spectroscopy ............................................................ 43
3.2.2 N2 Physisorption ................................................................................................... 43
3.2.3 Temperature programmed reduction (TPR).......................................................... 44
3.2.4 Diffuse Reflectance Infrared Fourier Transform (DRIFT) Spectroscopy ............ 46
3.2.5 X-Ray Diffraction (XRD) measurements ............................................................. 47
3.2.5 X-Ray Photoelectron Spectroscopy (XPS) ........................................................... 51
3.3 Catalytic evaluation ............................................................................................... 53
3.2.1. FTS reactor studies .............................................................................................. 53
References ..................................................................................................................... 68
ix
CHAPTER 4 OPTIMISATION OF LOADING OF COPPER AND POTASSIUM
PROMOTERS IN A PRECIPITATED FE-BASED FISCHER-TROPSCH SYNTHESIS
CATALYST................................................................................................................... 69
4.1 Introduction............................................................................................................ 69
4.2 Experimental .......................................................................................................... 69
4.3 Results and discussion ........................................................................................... 70
4.3.1 Optimising the weight loading of copper.............................................................. 70
4.3.2 Optimising the weight loading of potassium ........................................................ 82
4.4 Conclusion .............................................................................................................. 90
References ..................................................................................................................... 91
CHAPTER 5 EFFECT OF SiO2 ON AN UNPROMOTED Fe-BASED FISCHER-
TROPSCH SYNTHESIS CATALYST ......................................................................... 93
5.1 Introduction............................................................................................................ 93
5.2 Experimental .......................................................................................................... 95
5.3 Results and discussion ........................................................................................... 95
5.3.1 Textural and structural properties of the catalysts ................................................ 95
5.3.2 Reduction and carburization behaviour of the catalysts ....................................... 96
5.3.3 Surface analysis of the catalysts ......................................................................... 101
5.3.4 Adsorption properties of the catalysts................................................................. 107
5.3.5 FTS performances............................................................................................... 110
5.4 Conclusion ............................................................................................................ 114
References ................................................................................................................... 115
x
CHAPTER 6 EFFECT OF SiO2 ON A PROMOTED Fe-BASED FISCHER-TROPSCH
SYNTHESIS CATALYST .......................................................................................... 118
6.1 Introduction.......................................................................................................... 118
6.2 Experimental ........................................................................................................ 119
6.3 Results and discussion ......................................................................................... 119
6.3.1 Catalyst characterization..................................................................................... 119
6.3.2 FTS reactor studies ............................................................................................. 129
6.4 Conclusion ............................................................................................................ 134
References ................................................................................................................... 135
CHAPTER 7 EVALUATING INDIUM AS A CHEMICAL PROMOTER IN Fe-BASED
FISCHER TROPSCH SYNTHESIS ........................................................................... 137
7.1 Introduction.......................................................................................................... 137
7.2 Motivation to compare indium to copper as a chemical promoter ........................ 139
7.3 Experimental ........................................................................................................ 141
7.4 Results and discussion ......................................................................................... 141
7.4.1 N2 physisorption results ...................................................................................... 141
7.4.2 Hydrogen Temperature Programmed Reduction (H2 TPR)................................ 142
7.4.3 X-ray Diffraction (XRD) .................................................................................... 144
7.4.4 CO adsorption measurements using DRIFTS..................................................... 145
7.4.5 In situ FTS performances using DRIFTS ........................................................... 152
7.5 Conclusion ............................................................................................................ 157
References ................................................................................................................... 158
xi
CHAPTER 8 CHEMICAL PROMOTION OF A MULTI-PROMOTED Fe-BASED
FISCHER TROPSCH SYNTHESIS CATALYST BY INDIUM .............................. 160
8.1 Introduction.......................................................................................................... 160
8.2 Experimental ........................................................................................................ 160
8.3 Results and discussion ......................................................................................... 161
8.3.1 N2 physisorption results ...................................................................................... 161
8.3.2 Temperature Programmed Reduction (TPR) ...................................................... 162
8.3.3 DRIFTS............................................................................................................... 165
8.3.4 FTS performances............................................................................................... 168
8.4 Conclusion ............................................................................................................ 175
References ................................................................................................................... 176
CHAPTER 9 GENERAL CONCLUSIONS................................................................ 177
xii
List of Tables
Chapter 3
Table 3.1 Characteristics of the GCs employed 57
Chapter 4
Table 4.1 The theoretical and XRF determined Cu loadings 70
Table 4.2 XPS data for spectra given in Fig 4.1 72
Table 4.3 Reduction temperatures for the H2 TPR profiles show in Fig.4.2 74
Table 4.4 %Fe reducibility as a function of Cu loading for all Cu loaded catalysts 76
Table 4.5 The calculated crystallite size of Fe2O3 as a function of Cu loading 77
Table 4.6 Position of IR absorbtion band as a function of Cu loading 79
Table 4.7 Calculated ratios of CH2/CH3 bands for all catalysts 81
Table 4.8 Reduction temperatures for the H2 TPR profiles show in Fig. 4.5 83
Table 4.9 %Fe reducibility as a function of K2O loading for all K2O loaded catalysts 84
Table 4.10 The calculated crystallite size of Fe2O3 as a function of K2O loading 85
Table 4.11 Peak shifts of peak at wavenumber region 2012-2015 cm-1 as a function of K2O
loading
87
Table 4.12 Estimation of the CH2/CH3 ratio as a function of K2O loading 88
xiii
Chapter 5
Table 5.1 The composition and textural properties of the calcined catalysts 95
Table 5.2 The calculated crystallite size of Fe2O3 as a function of SiO2 loading 96
Table 5.3 Peak maxima of H2 TPR profiles 98
Table 5.4 Reduction temperatures for peaks of CO TPR profiles 100
Table 5.5 Areas for peaks in Figure 5.2 100
Table 5.6 Fe (2p) and Si (2p) peak areas for all catalysts 105
Table 5.7 Band maxima in the wavenumber region 2012-2015 cm-1 as a function of SiO2
loading
108
Table 5.8 Estimation of the CH2/CH3 ratio as a function of SiO2 loading 112
Chapter 6
Table 6.1 The composition and textural properties of the catalysts 119
Reduction temperatures for peak 1 and peak 2 as well as their areas 123
Table 6.3 Reaction performances of all catalysts at steady state conditions 132
xiv
Chapter 7
Table 7.1 The composition and textural properties of the catalysts as prepared 141
Table 7.2 Comparing the reducibility of Fe-based catalysts using H2 TPR 144
Table 7.3 Fe2O3 crystallite size determined using Rietveld refinement 145
Table 7.4 Calculated ratios of CH2/CH3 for all catalysts 155
Chapter 8
Table 8.1 The composition and N2 physisorption results of the catalysts 161
Table 8.2 H2 Reduction temperatures for all the catalysts in Figure 8.1 163
Table 8.3 CO reduction temperatures for all the catalysts in Figure 8.2 164
Table 8.4 FTS reaction performances for all the catalysts 172
xv
List of Figures
Chapter 2
Figure 2.1 Franz Fischer at work in 1918 4
Figure 2.2 Multitubular fixed bed reactor with internal cooling 10
Figure 2.3 Slurry bubble column reactor (or slurry bed reactor) with internal cooling tubes or
three-phase fluidised (ebulating) bed reactors
11
Figure 2.4 Fluidised fixed bed (FFB) reactor with internal cooling 13
Figure 2.5 Circulating fluidised bed (CFB) reactor with circulating solids, gas recycle and
cooling in the gas/solid recirculating loop
14
Figure 2.6 A representation of the stepwise mechanism for hydrocarbon chain growth and
chain termination
18
Figure 2.7 Typical plot of calculated selectivities (% carbon atom basis) of carbon number
product cuts as a function of the probability chain growth
21
Chapter 3
Figure 3.1 The TRISTAR 3000 analyzer 44
Figure 3.2 Experimental set-up for TPR measurements 45
Figure 3.3 DRIFTS cell with ZnSe windows 46
Figure 3.4 Gas manifold for the introduction of gases into the DRIFTS cell 47
Figure 3.5 The Bruker D8 X-Ray diffractometer 48
xvi
Figure 3.6 Diffraction pattern obtained after the XRD measurement of Fe2O3 49
Figure 3.7 Fitting of the experimental diffraction pattern (a) blue line represents the
experimental pattern and red line is the fitted curve (b) difference curve produced after fitting
the experimental diffraction pattern
50
Figure 3.8 The AXIS UltraDLD XPS instrument 51
Figure 3.9 The stainless steel bar showing the mounted catalysts ready for XPS analysis 52
Figure 3.10 The fixed bed reactor made from a ?? Swagelok stainless steel pipe.
A = Sketch portrait; B = Digital portrait
54
Figure 3.11 The hot trap placed in a heating jacket, both situated below the reactor 55
Figure 3.12 Traps, pressure regulator, needle valve and gas line after reactor 56
Figure 3.13 GC on the left fitted with an FID detector and the one on the right fitted with a
TCD detector
58
Figure 3.14 Schematic representation of the reactor setup 59
Figure 3.15 A trace for the calibration gas using the TCD GC 61
Figure 3.16 A trace for the calibration gas product using the FID GC 61
Figure 3.17 A trace showing the calibration of the TCD GC using syngas 62
Figure 3.18 FTS products detected by the TCD GC 62
Figure 3.19 FTS products detected by the FID GC 63
Chapter 4
Figure 4.1. (a) Cu(LLM) and (b) Cu(2p) spectra for all catalysts 71
Figure 4.2 H2 TPR profiles of all the catalysts 73
Figure. 4.3 TPR profile of Fe2O3 reduced using the reduction method employed for carrying 75
xvii
out an FTS reaction
Figure 4.4 Comparing CO absorption spectra of Cu promoted catalysts to the unpromoted Fe
catalyst
78
Figure 4.5 H2 TPR profiles of all the catalysts 82
Figure 4.6 Comparing CO absorption spectra of K2O promoted catalysts to the unpromoted
Fe catalyst
86
Chapter 5
Figure 5.1 H2 TPR profiles of the catalysts 97
Figure 5.2 CO TPR profiles of catalysts 99
Figure 5.3 Survey spectrum showing elements on the surface of the 5SiO2/100Fe calcined
catalyst
101
Figure 5.4 Narrow region spectrum of Fe (2p) peak for all catalysts 103
Figure 5.5 Narrow region spectrum of Si (2p) peak for all catalysts 104
Figure 5.6 Oxygen core level spectra for a) 5SiO2/100Fe b) 10SiO2/100Fe c) 20 SiO2/100Fe
and d) 25SiO2/100Fe
106
Figure 5.7 CO adsorption spectra of all the catalysts (P = 2 bar, T = 25 ?C, CO flow rate = 12
ml/min)
107
Figure 5.8 DRIFTS spectra of all catalysts showing the C-H region after 5 hours of the FTS
reaction (Reduction conditions: H2/CO = 2, P = 2 Bar, T = 350 ?C, H2/CO flow rate = 12
ml/min, t = 1 h; FTS reaction conditions: H2/CO = 2, P = 10 Bar, T = 275 ?C, H2/CO flow rate
= 12 ml/min, t = 5h)
110
xviii
Figure 5.9 CO2 produced as a function of the SiO2 content 113
Chapter 6
Figure 6.1 H2 TPR profiles for all catalysts 122
Figure 6.2 CO adsorption on all the catalysts; Conditions: CO reduction for 1 hour (Flow rate
= 12 ml/min, T = 350oC, P = 2 bar), CO adsorption for 30 min (CO Flow rate = 12 ml/min, T
= 25oC, P = 2 bar)
124
Figure 6.3 Comparing the intensity of peak at 2014 cm-1 for all the catalysts 126
Figure 6.4 DRIFTS spectra showcasing FTS reactions for all catalysts;
Conditions: Reduction for 1 hour (H2/CO = 2/1, Flow rate = 12 ml/min, T = 350 oC, P = 2
bar), FTS reaction for 5 hours (H2/CO = 2/1, Flow rate = 12 ml/min, T = 275oC, P = 10 bar)
127
Figure 6.5 DRIFTS spectra comparing FTS reactions for SiO2 loaded catalysts to a non-
loaded SiO2 catalyst; Conditions: Reduction for 1 hour (H2/CO = 2/1, Flow rate = 12 ml/min,
T = 350 oC, P = 2 bar), FTS reaction for 5 hours (H2/CO = 2/1, Flow rate = 12 ml/min, T =
275oC, P = 10 bar)
128
Figures 6.6 The carbon monoxide conversion with time on stream for all catalysts 129
Figure 6.7 Comparing CO conversion for all catalysts at steady state conditions 130
Figures 6.8 The hydrogen conversion with time on stream for all catalysts 131
Figure 6.9 Comparing H2 conversion for all catalysts at steady state conditions 131
xix
Chapter 7
Figure 7.1 Elements that show knight?s move relationships 138
Figure 7.2 Carbon nanotubes synthesized using the Fe-Ni/CaCO3 catalyst 139
Figure 7.3 Carbon nanotubes and coils synthesized using the Fe-Ni-Cu/CaCO3 catalyst 139
Figure 7.4 Carbon nanotubes and coils synthesized using the Fe-Ni-In/CaCO3 catalyst 140
Figure 7.5 Percentage composition of coils and tubes produced for the copper and indium
promoted catalysts
140
Figure 7.6 H2 TPR profiles of Cu promoted catalysts 143
Figure 7.7 H2 TPR profiles of indium promoted catalysts 143
Figure 7.8 Comparison of CO adsorption on the unpromoted iron catalyst and copper
promoted iron catalysts; Conditions: CO reduction for 1 hour (Flow rate = 12 ml/min, T =
350oC, P = 2 bar), CO adsorption for 30 min (CO Flow rate = 12 ml/min, T = 25oC, P = 2 bar)
147
Figure 7.9 Thermal desorption of CO on the unpromoted iron catalyst 148
Figure 7.10 CO adsorption on the indium promoted iron catalysts 149
Figure 7.11 Intensity of peak at 2013 cm-1 for CO adsorption on the copper promoted and the
indium promoted iron catalysts
149
Figure 7.12 Intensity of peak at 2033 cm-1 for CO adsorption on the copper promoted iron
catalysts and the indium promoted iron catalysts
150
Figure 7.13 CO adsorption on the indium promoted iron catalysts showing the adsorbed CO
species at 2024 and 2042 cm-1
151
Figure 7.14 Comparison of the FTS reaction over the unpromoted iron catalyst (100Fe) and
copper promoted catalysts; P = 10 bar, T = 275 ?C, H2/CO = 2, H2/CO flow rate = 12 ml/min,
153
xx
Time = 5 h)
Figure 7.15 Comparison of the FTS reaction over unpromoted iron catalyst (100Fe) and
indium promoted catalysts; P = 10 bar, T = 275 ?C, H2/CO = 2, H2/CO flow rate = 12 ml/min,
Time = 5 h)
154
Chapter 8
Figure 8.1 H2 TPR profiles for all the catalysts 162
Figure 8.2 CO TPR profiles for all the catalysts 163
Figure 8.3 CO adsorption on all the catalysts; Conditions: CO reduction for 1 hour (Flow rate
= 12 ml/min, T = 350oC, P = 2 bar, CO adsorption for 30 min (CO Flow rate = 12 ml/min, T
= 25oC, P = 2 Bar)
165
Figure 8.4 Comparison of the intensities of 2014 cm-1 peak 166
Figure 8.5 DRIFTS spectra for the FTS reaction of all catalysts
Reaction conditions: H2/CO = 2/1, Flow rate = 12 ml/min, T = 275oC, P = 10 bar, t = 5 h
167
Figure 8.6 CO conversion with time on stream 168
Figure 8.7 H2 conversion with time on stream 169
Figure 8.8 Comparison of the CO conversion for all catalysts at steady state 170
Figure 8.9 Comparing of the H2 conversion for all catalysts at steady state 170
Figure 8.10 H2 TPR profile of In2O3 175
xxi
Abbreviations and acronyms
African Oxygen AFROX
Anderson, Schultz and Flory ASF
Approximately ca.
Badishe Anilin und Soda Fabrik BASF
Brunauer, Emmett and Teller BET
Calcium carbonate CaCO3
Carbon dioxide CO2
Carbon monoxide CO
Carbon number n
Chain Growth Probability ?
Chain propagation rp
Chain termination rt
Cobalt Co
Copper Cu
Copper oxide CuO
Cubic centimeter cm3
Degrees Celsius ?C
Fischer Tropsch FT
Fischer-Tropsch Synthesis FTS
Flame ionization detector FID
Gas chromatograph GC
Gram g
xxii
GC integrated area of component c cA
High-temperature Fischer Tropsch HTFT
Hydrogen H2
Indium In
Iron Fe
Liquid Petroleum Gas LPG
Low-temperature Fischer Tropsch LTFT
Methane CH4
Mole fraction Xi
Moles of carbon in the feed in C,N
Nanometer nm
Nickel Ni
Nitrogen N2
Percentage %
Product selectivity for hydrocarbons Si
United States U.S.
Rates of reaction for FTS FTSr
Rates of reaction for water gas shift WGS WGSr
Ruthenium Ru
Sigma ?
Square meter m2
Temperature Programmed Reduction TPR
Thermal conductivity detector TCD
xxiii
Total feed flow rate Fin
Total reactor exit stream outF
Tin Sn
Ultra High Purity UHP
Water gas shift WGS
Weight Percentage wt. %
X-ray Diffraction XRD
X-ray Fluorescence XRF
X-ray Photoelectron Spectroscopy XPS
Zinc Zn
1
Chapter 1
Introduction
This chapter is written to introduce the work that was performed in this thesis and to give
a comprehensive breakdown of the thesis content. The thesis is composed of nine
chapters including Chapter 1. The aim of this work as illustrated by the title was to
systematically study the effect that promoters have on a precipitated Fe-based Fischer-
Tropsch synthesis (FTS) catalyst. A catalyst promoter is classified as a chemical
substance that enhances the chemical or physical properties of a catalyst.
Over many decades various promoters have been evaluated in catalysis and extensive
publications on their effects have been published. In this thesis we undertook an approach
to studying the effects caused by copper, potassium, silica and indium on a precipitated
Fe-based Fischer- Tropsch synthesis catalyst. A literature review on the well established
Fischer-Tropsch synthesis process is first presented in Chapter 2. This describes pertinent
issues such as the reaction pathways, the catalysts often used in this process, etc. Effects
caused by copper, potassium and silica are well known and they are also discussed in
Chapter 2. In this chapter an evaluation of prior work on FTS catalyst promoters is given.
It is clear from this evaluation that little has been reported on the role of individual
promoters and their relationship to the use of multiple catalyst promoters.
Chapter 3 presents all the experimental techniques and procedures that were used to carry
out the studies. The actual work performed to evaluate the effects caused by these
promoters on the Fe-based Fischer-Tropsch synthesis catalyst is subsequently presented
(Chapter 4 to Chapter 8). Chapter 4 describes the effect of Cu and K2O on the Fe-based
Fischer-Tropsch synthesis catalyst. Since the effects of these two promoters are well
known and established this study was performed to optimize the weight loadings of these
two promoters and provide reference data for the mixed promoter studies.
2
Chapter 5 deals with the effect of silica and in particular the silica content on the
unpromoted Fe-based Fischer-Tropsch synthesis catalyst. The results presented in this
Chapter are correlated with the silica content on the Fe-based Fischer-Tropsch synthesis
catalyst.
Chapter 6 illustrates the effect of silica content on a K2O and Cu promoted Fe-based
Fischer-Tropsch synthesis catalyst. The results presented in this chapter are attributed
solely to the effect of silica in the presence of Cu and K2O. The aim of this chapter was to
assess the inter-promotional effects of the promoters and what effect they have on the
catalyst. It is to be noted that the optimum weight loadings obtained in Chapter 4 were
used to prepare the catalysts evaluated in Chapter 6.
Chapter 7 presents the effect that indium has on the precipitated Fe-based Fischer
Tropsch synthesis catalyst. The effects caused by indium on the catalyst are compared to
those of Cu.
Chapter 8 describes the effect caused by indium on the precipitated Fe-based Fischer-
Tropsch synthesis catalyst in the presence of potassium and silica. The aim of this study
was to assess the effect of indium on the potassium and silica promoters as well as the
overall effect that they have on the Fe-based Fischer Tropsch synthesis catalyst.
The general conclusions on the effect of copper, potassium, silica and indium are
presented in Chapter 9. This chapter sums up all the effects caused by the above
mentioned promoters and conclusions reached from each chapter are thus placed
perspective.
3
Chapter 2
Fischer-Tropsch Synthesis (FTS): Literature review
2.1 Introduction
The Fischer-Tropsch Synthesis (or Fischer-Tropsch process) is a catalyzed chemical
reaction in which synthesis gas (syngas), a mixture of carbon monoxide (CO) and
hydrogen (H2), is converted into gaseous, liquid and solid hydrocarbons [1-12] and an
appreciable amount of oxygenates [13-18]. The principal purpose of this process is to
produce a synthetic petroleum substitute, typically from coal, natural gas or biomass, for
use as synthetic lubrication oil or as synthetic fuel. This synthetic fuel runs trucks, cars,
and some aircraft engines. The process has also been employed to produce higher value
specialty chemicals, via 1-alkenes made from syngas derived from natural gas (methane)
or coal [19].
2.2 History of the Fischer-Tropsch process
In 1897, Losanitsch and Jovitschitsch reported the conversion of syngas to liquid
products on electric discharge [29]. Not long after that in 1902, Sabatier and Senderens
showed that methane could be produced from CO and H2 mixtures using a nickel catalyst
[20]. This captured the interest of many catalysis researchers and provided a platform for
rigorous and intense research into this type of work. In 1913 Badische Anilin and Soda
Fabrik (BASF) were awarded a patent for showcasing the catalytic production of higher
hydrocarbons and oxygenated compounds from syngas under high pressures [21]. A
decade later in 1923, two German researchers Franz Fischer (Fig. 2.1) and Hans Tropsch,
working at the Kaiser Wilhelm Institute reported on related studies. Their work involved
the reaction of syngas over alkalised iron and many other catalysts to produce a mixture
of hydrocarbons and oxygenated compounds [22]. This was the start of what was to be
later known as the Fischer-Tropsch Synthesis.
4
Figure 2.1 Franz Fischer at work in 1918 [23]
Since the invention of the original process many refinements and adjustments have been
made, and the term "Fischer-Tropsch" now applies to a wide variety of similar processes
(Fischer-Tropsch reaction or Fischer-Tropsch chemistry). The bulk of the refinements
have been reported and a useful website for the location of publications relating to the
research and development of the Fischer-Tropsch Synthesis can be accessed at
http://www.fischer-tropsch.org [24]
2.3 Utilization of the FT process
The application of FTS at an industrial level started in Germany (rightfully so) since the
process emanated from this country. By 1938, nine plants with a combined production
capacity of about 660 x 103 t per year were in operation [25]. Even though the nine FT
plants in Germany ceased to operate after World War II, the fear of an impending
shortage of petroleum kept the interest in the FT process alive. An FT plant with a
capacity of 360 x 103 t per year was built and operated in Brownsville, TX, during the
1950s. This plant was based on syngas produced from methane but a sharp increase in the
price of methane caused the plant to be shut down [26, 27].
5
Then in 1955, Sasol, now a world-leader in the commercial production of liquid fuels and
chemicals from coal and natural gas, started Sasol I in Sasolburg, South Africa. Due to
the oil crises of the mid 1970s and the success of Sasol I, Sasol constructed two much
larger coal-based FT plants which came on line in 1980 (Sasol II) and 1982 (Sasol III)
respectively. The combined capacity of these three Sasol plants was about 6000 x 103 t
per year [27].
Some commercial ventures in FTS by Shell international in Malaysia for the production
of waxes and the Mossgas project in South Africa were subsequently initiated. Based on
methane, the Mossgas plant in South Africa and the Shell plant at Bintuli, Malaysia,
came on stream in 1992 and 1993, respectively [12]. The Mossgas plant which converts
natural gas to FT products uses a high temperature process and an iron catalyst. This
plant is still running and is now under the auspices of PetroSA. The Shell commissioned
plant in Bintuli, Malaysia uses the Shell Middle Distillate Synthesis process (SMDS),
which is essentially, an enhanced FT synthesis.
In the last few years the interest for FTS has significantly grown due to the increase in oil
price as well as the high demand for energy. Recent commercial ventures include the
development of a Gas-To-Liquid (GTL) plant, Oryx GTL, in a joint venture of Sasol with
Qatar Petroleum at Ras Laffan in Qatar. Sasol is also developing a GTL plant at Escravos
in Nigeria. Currently, Syntroleum Corporation (a United States company) is building a 10
000 barrels per day (bpd) specialty chemicals and lube oil plant located in Northwestern
Australia, also using the GTL process [28, 29].
Rentech (a small US-based company) is currently focusing on converting nitrogen-
fertiliser plants from using a natural gas feedstock to using coal or coke, and producing
liquid hydrocarbons as a by-product. In September 2005, Pennsylvania governor Edward
Rendell announced a venture with Waste Management and Processors Inc. - using
technology licensed from Shell and Sasol - to build an FT plant that will convert so-
6
called waste coal (leftovers from the mining process) into low-sulfur diesel fuel at a site
outside of Mahanoy City, northwest of Philadelphia [30].
The state of Pennsylvania has committed to buy a significant percentage of the plant's
output and together with the U.S. Dept. of Energy, has offered over $140 million in tax
incentives. Other coal-producing states are exploring similar plans. Governor Brian
Schweitzer of Montana has proposed developing a plant that would use the FT process to
turn his state's coal reserves into fuel in order to help alleviate the United States'
dependence on foreign oil [30]
With demand for energy expected to grow 5 % a year to 2020 (according to the Carbon
Sequestration Leadership Forum: www.cslforum.org/china.htm), China has been looking
at exploiting its abundant coal reserves to meet its energy requirements. Pre-feasibility
studies focusing on exploring the potential of developing two Coal-To-Liquid (CTL)
plants, using Sasol?s low temperature Fischer-Tropsch technology, each with a capacity
of about 80000 barrels per day were concluded in November 2005 [27].
In October 2006, Finnish paper and pulp manufacturer UPM announced its plans to
produce biodiesel by the Fischer-Tropsch process alongside manufacturing processes at
its European paper and pulp plants, using waste biomass from the paper and pulp
manufacturing processes as the source material [30].
In August 2007, Louisiana State University announced they had received funding from
the US Department of Energy and Conoco Phillips for development of new
nanotechnologies for catalysis of coal syngas to ethanol conversion. Conoco-Phillips is
currently building a gas-to-liquids pilot plant in Bartlesville, Oklahoma to produce diesel,
naptha, and waxes from natural gas via FT catalysis [30]. The above reports show that the
FT process is an established technology and already well applied on a large scale in some
industrial sectors.
The commercial FT process itself involves three main steps, namely: syngas production,
FT synthesis and product upgrading. These three main steps will be described next.
7
2.4 Three main steps in the FT process
2.4.1 Synthesis gas production
The initial reactants (syngas) used in the Fischer-Tropsch process are hydrogen gas (H2)
and carbon monoxide (CO). These chemicals are usually produced by one of two
methods:
1. The partial combustion of a hydrocarbon:
CnH(2n+2) + ? nO2 ? (n+1)H2 + nCO
When n=1 (methane), the equation becomes 2CH4 + O2 ? 4H2 + 2CO
2. The gasification of coal, biomass, or natural gas:
CHx + H2O ? (1+0.5x)H2 + CO
The value of "x" depends on the type of fuel. For example, natural gas has a greater
hydrogen content (x=2 to x=4) than coal (x<2). The energy needed for this endothermic
reaction is usually provided by the (exothermic) combustion of oxygen and the
hydrocarbon source.
Given its availability methane is preferred to coal for syngas production.
When using natural gas as the feedstock, many authors [31-36] have recommended
autothermal reforming or autothermal reforming in combination with steam reforming as
the best option for syngas generation. This is primarily attributed to the resulting H2/CO
ratio and the fact that there is a more favourable economy of scale for air separation units
than for tubular reactors (steam methane reforming - SMR).
8
2.4.2 FT synthesis: Process conditions
Generally, the Fischer-Tropsch process is operated in the temperature range of 180-
350?C. Higher temperatures lead to faster reactions and higher conversion rates, but also
tend to favor methane production. As a result the temperature is usually maintained at the
low to middle part of the range. Increasing the pressure leads to higher conversion rates
and also favours formation of long-chain alkanes both of which are desirable. Typical
pressures are in the range of one to several tens of bars. Even higher pressures would be
more favourable, but the benefits may not justify the additional costs of high-pressure
equipment [30].
2.4.3 Product Upgrading and Separation
Conventional refinery processes can be used for the upgrading of Fischer-Tropsch
liquid and wax products. A number of possible processes for FT products are: wax
hydrocracking, distillate hydrotreating, catalytic reforming, naphta hydrotreating,
alkylation and isomerisation [37]. Fuels produced by the FT synthesis are of a high
quality due to their very low aromaticity and zero sulfur content. The product stream
consists of various fuel types: LPG, gasoline, diesel fuel, jet fuel, etc. The diesel fraction
has a high cetane number resulting in superior combustion properties and reduced
emissions [37]. New and stringent regulations may promote replacement or blending of
conventional fuels by sulfur and aromatic free FT products [38, 39]. Also, other products
besides fuels can be manufactured with Fischer-Tropsch catalysts in combination with
upgrading processes, for example, ethene, propene, ?-olefins, alcohols, ketones, solvents,
specialty waxes, and so forth. These valuable by-products of the FT process have higher
added values, resulting in an economically more attractive process economy.
9
2.5 FT reactors
The FTS is operated in two modes. The high-temperature (300 - 350 ?C) mode with iron-
based catalysts is used for the production of gasoline and linear low molecular mass
olefins [41]. The low-temperature (200 - 240 ?C) mode with either iron or cobalt catalysts
is used for the production of high molecular mass linear waxes [42]. Efficient and rapid
removal of heat from the highly exothermic FT reaction from the catalyst particles is
essential [43-45]. If this is not adequately performed, overheating results and this
adversely affects the performance of the catalyst. Therefore ?state of the art? reactors are
needed to circumvent such problems, inevitably making reactor design a pivotal part of
the FT technology. The main types of FT reactors which have been developed since 1950
are illustrated below [37, 40, 46, 166].
10
Product Outlet
Wax
Steam
Feed inlet
Feed water inlet
Tube bundle
Fixed bed
Figure 2.2 Multitubular fixed bed reactor with internal cooling [166]
11
Figure 2.3 Slurry bubble column reactor (or slurry bed reactor) with internal cooling
tubes [166]
Multitubular fixed bed reactors (Fig. 2.2) are usually employed for the low temperature
FT operation in producing wax. The gas flows through the bed in the downward direction
and the wax produced trickles down and out of the catalyst bed. In the slurry bed reactor
(Fig. 2.3), the gas flow itself provides the agitation power required to keep the catalyst
bed in suspension.
12
The slurry bed reactor presents many advantages over the multitubular fixed bed
reactor. It is cheaper to construct (only 25% of the cost of the Multitubular fixed bed
reactor) and also requires less amount of catalyst. This catalyst can easily be removed or
added on-line. It is also more isothermal thereby enabling it to be operated at higher
temperatures which results in higher conversions. On the other hand the fixed bed is
simple to operate and allows for easy separation of the catalyst from wax.
Among the disadvantages of the fixed bed reactors are: a high pressure drop over the
reactor, a high temperature gradient (compared to other reactors) and tedious replacement
of the used catalyst [45, 46].
The third type of reactor is the fluidized bed reactor. There are two types of fluidised bed
reactors; the fluidised fixed bed (FFB) reactor (Fig. 2.4) and the circulating fluidised bed
(CFB) reactor (Fig. 2.5). In the FFB reactors, there are two phases of fluidised catalyst. In
the CFB reactor, the catalyst flows down the standpipe in a dense phase while it is
transported up the reaction zone in a lean phase. The heat of reaction is removed from the
reactor by cooling coils that generate steam. To avoid the inlet gas going up the standpipe
the pressure over the standpipe, must be higher than in the reaction zone [41].
13
Figure 2.4 Fluidised fixed bed (FFB) reactor with internal cooling [166]
14
Figure 2.5 Circulating fluidised bed (CFB) reactor with circulating solids, gas recycle
and cooling in the gas/solid recirculating loop [166]
Dry [41] compared the FFB reactor to the CFB reactor. He noted that for the same
production capacity, the FFB is smaller than the CFB, it is less costly to construct (cost is
40% lower), simpler to operate (more gas can be fed by either increasing the volumetric
flow rate or by increasing operating pressure) and easier to build.
In the FFB the whole catalyst charge participates in the reaction at any moment, whereas
in the CFB only a portion of it does since a portion of the catalyst is in the recirculation
15
loop and so not in contact with the reactant gas. The main disadvantage of the two
fluidized bed reactors is that should any poison enter the reactor the entire catalyst bed is
poisoned whereas in the fixed bed, the poison is adsorbed on the top layer of the catalyst
leaving the rest of the bed intact.
2.6 The chemistry behind the FT process
2.6.1 Reactions
The FTS has long been recognised as a polymerisation reaction [1]. It involves a variety
of competing chemical reactions, which lead to a series of desirable products and
undesirable byproducts. The most important reactions are those resulting in the formation
of alkanes (paraffins). These can be described by chemical equations of the form:
(2n+1)H2 + nCO ? CnH(2n+2) + nH2O
where 'n' is a positive integer. The simplest of these (n=1), results in formation of
methane, which is generally considered to be an unwanted byproduct (particularly when
methane is the primary feedstock used to produce the synthesis gas). Process conditions
and catalyst composition are usually chosen, so as to favor higher order reactions (n>1)
and thus minimize methane formation. Most of the alkanes produced tend to be straight-
chain, although some branched alkanes are also formed.
In addition to alkane formation, competing reactions result in the formation of alkenes
(olefins), as well as alcohols and other oxygenated hydrocarbons [47]. Usually, only
relatively small quantities of these non-alkane products are formed, although catalysts
favouring some of these products have been developed. An overview of the reactions
involved is illustrated in the equations below:
16
Methane formation:
CO + 3H2 ? CH4 + H2O
Alkene (olefin) formation:
nCO + 2nH2 ? CnH2n + nH2O
Alcohol formation:
nCO + 2nH2 ? CnH2n+1OH + (n-1)H2O
Acid formation:
(n+1)CO + (2n)H2 ? CnH2n+1COOH + (n-1)H2O
Another important reaction is the water gas shift reaction (WGS):
H2O + CO ? H2 + CO2
Although this reaction results in formation of unwanted CO2, it can be used to shift the
H2/CO ratio of the incoming syngas. This is especially important for syngas derived from
coal, which tends to have a ratio of ~0.7 compared to the ideal ratio of ~2.
Another way in which CO2 can be produced in the FTS is via the Boudouard reaction:
2CO ? C(s) + CO2
Carbon is also produced from this reaction and can be deposited on the catalyst surface
leading to catalyst deactivation. Thus, depending on a number of factors e.g. H2/CO ratio,
catalyst type, reactor type and reaction conditions used, one or the other of reactions can
predominate in the synthesis [48].
2.6.2 Mechanism and product selectivity
The FTS process produces a wide range of products - due to this - a detailed mechanism
accounting for formation of all FTS products is yet to be achieved or reported. The detail
of the mechanism has been a bone of contention for many years and an extraordinarily
?hard nut? to crack. Of all the different mechanisms proposed, most of them still remain
17
within the original four classes put forward over the decades, namely; the surface carbide,
enolic intermediate, CO-insertion and alkoxy intermediate mechanisms [49, 50]
Nonetheless there is general consensus that a stepwise growth mechanism is involved.
Thus the very wide range of products formed is as a result of sequential steps taking place
on the catalyst surface. These sequential steps largely resemble those of a polymerization
reaction and can be summarized as follows:
a) reactant adsorption
b) chain initiation
c) chain growth
d) chain termination
e) product desorption
f) re-adsorption and further reaction
Consequently a mechanism of chain growth and termination has been proposed [47] and
it is illustrated in Fig. 2.6.
18
Figure 2.6 A representation of the stepwise mechanism for hydrocarbon chain growth
and chain termination [47]
The CH2 units (Fig. 2.6) formed by the hydrogenation of CO are taken as the
?monomers? in this stepwise polymerization process. At each stage of growth the
adsorbed hydrocarbon species has the option of desorbing or being hydrogenated to form
the primary FT products or of adding another monomer to continue the chain growth.
Maitlis and co-workers have used the ideas of organometallic chemistry and
homogeneous catalysis derived from model systems, combined with the results of
experiments using 13CH2=13CH2-X (X = H, Br, etc) compounds as probes to propose the
19
?alkenyl mechanism? for the F-T reaction. In this mechanism chain growth is initiated by
a vinyl + methylene coupling and it proceeds via coupling of these two groups and
terminates via hydrogenation of the alkenyl to yield the 1-alkene [50]. This mechanism
can explain the formation of branched products (for example, by allyl isomerisation).
Labelling probe studies also suggest that oxygenates such as ethanol arise from CO but
not via methylenes in F-T reactions [50].
Other mechanisms reported describe molecules such as CO and CHOH as possible
?monomers? that add onto the growing chain. For instance CO insertion onto the growing
chain is believed to be the way that the alcohols, acids and aldehydes are formed [51-53].
The probability of chain growth (?) is assumed to be independent of the chain length. A
product distribution model known as the Anderson-Schulz-Flory (ASF) model [54, 55] is
usually used to obtain the relationship between the weight fraction of formed
hydrocarbons and the chain growth probability.
This model is described by the following equation:
Wn/n = (1-?)2?n-1
where Wn is the weight fraction of hydrocarbon molecules containing n carbon atoms and
? (alpha) is the chain growth probability or the probability that a molecule will continue
reacting to form a longer chain. In general, ? is largely determined by the catalyst and the
specific process conditions. Examination of the above equation reveals that methane will
always be the largest single product, however by increasing ? so that it is close to one, the
total amount of methane formed can be minimized compared to the sum of all of the
various long-chain products.
To easily illustrate the above point, the Anderson-Schulz-Flory (ASF) is usually
linearised into the following equation:
log (Wn/n) = n log (?) + log ((1- ?)/?)2
20
This equation is used to determine the ? value from experimental data. A plot of log
(Wn/n) versus carbon number (n) is linear and the chain growth probability is obtained
from its slope as log (?) or from the intercept as log ((1-?)/?)2 at n = 1.
Alpha (?) can also be defined in terms of the rate of chain propagation (rp) and chain
termination (rt) as:
Calculated product selectivities versus probability of chain growth are illustrated
in Fig. 2.7 [56]. This plot shows that only the light (? ? 0) or heavy (? ? ?)
products can have a high selectivity. All other products go through a maximum
yield. The product distribution is influenced by operating conditions (temperature,
pressure, feed gas composition, space velocity) and catalyst type and promoters. In other
words the alpha value (?) for product distribution ranges between 0 and 1 with the higher
value indicating a greater selectivity towards waxy products and a lower value
corresponding to gaseous products.
21
Figure 2.7 Typical plot of calculated selectivities (% carbon atom basis) of carbon
number product cuts as a function of the probability chain growth [56]
However, the FT product distributions reported in the literature [57, 58] do not always
obey the simple ASF kinetic model. Some of the deviations usually observed include:
a) A high methane selectivity. It is proposed that this is as result of methane being able to
form by more than one pathway [59].
b) A low yield of ethane, ethene and in some cases propane relative to the predicted ASF
distribution. It is suggested that this could be due to the re-insertion of the very reactive
olefins back into the growing chain.
22
c) Some negative [60] and positive [25, 55, 61-63] deviations especially when the carbon
number is greater than 8 have also been reported. Various mechanisms
accounting for chain-length related phenomena have been proposed. These include a
vapor-liquid equilibrium phenomena, diffusion enhanced olefin readsorption model [62],
different physisorption strength of the olefins [63] and the two-active-site model [55, 61].
Shi and Davis [64] have accounted for chain-length related phenomena by proposing that
the apparent products of the FTS reaction is a mixture of freshly produced FTS products
and the products left in the reactor. They concluded that in order to obtain correct product
distribution in a FTS reaction, it is necessary to find a way to evaluate or eliminate the
contribution from the products left in the reactor.
d) Further the ?-olefin to paraffin ratio decreases exponentially and the chain growth
parameter, ?, is not constant with increasing chain length.
23
2.7 Fischer-Tropsch catalysts
A variety of catalysts can be used for the Fischer-Tropsch process, but the most common
are the transition metals (group 8-10 metals) since they can dissociatively adsorb H2 and
CO. Fe, Ni, Co and Ru are the only metals that have the required FT activity for
commercial application [12]. Ni has been reported to produce too much methane under
FT conditions [11, 65-67]. On the other hand Ru has been found to be less selective to
methane and more selective to the C5+ hydrocarbon fraction than other metals [66]. Ru is
the most expensive of these four metals and its availability in the world is insufficient for
large scale application. For these reasons Fe and Co are viable catalysts for industrial
applications.
Historically, Fe has been the catalyst of choice in industrial applications due to its low
cost. It is also more suitable for low-hydrogen-content synthesis gases such as those
derived from coal due to its promotion of the water-gas-shift reaction.
More recently emphasis in industry has been placed on using Co in industrial reactors. In
this thesis, research focusing on the use of Fe has been pursued and thus only studies
done using this catalyst will be described.
2.7.1 Use of Fe catalysts in FTS
Fe catalysts have been extensively used for the F-T synthesis and scores of literature
studies can be found that describe work in this area. Extensive reviews and government
reports have also been written [6, 9, 68-70], highlighting the versatility of the Fe catalyst
in the FTS.
As explained earlier, the FTS has two temperature regimes, namely the High
Temperature Fischer Tropsch (HTFT) and Low Temperature Fischer Tropsch (LTFT)
regimes. The Fe catalyst has found use in both these regimes. The HTFT process with
iron-based catalysts, which is operated at temperatures between 300 and 350oC is used
24
for the production of low molecular mass olefins, gasoline (primarily) and diesel fuel
range liquids. The (LTFT) process is operated in the 200-250oC temperature range is used
for the production of high molecular weight waxes [71].
Thus Fe catalysts are very flexible and this is as result of much effort put into better
understanding their chemical properties [72-80]. The three key properties that have
always been studied for improving them is their lifetime, activity and product selectivity.
Optimizing these properties for desired commercial application has been the focus of Fe-
based FTS catalyst research and development. Each one of these properties can be
affected by a variety of parameters which include:
a) Catalyst preparation
Fe catalysts can be prepared by various methods varying from precipitation [81-84] to
impregnation methods. Precipitation is normally the preferred preparation method for Fe
catalysts employed commercially [9, 37].
b) Catalyst activation
The activation procedure used for Fe FTS catalysts has a great influence on their activity
and selectivity [85-87]. Precipitated catalysts are usually activated using carbon
monoxide, hydrogen or synthesis gas [85]. Activation alters the catalyst composition to
what is thought to be a mixture of iron oxides (Fe2O3, Fe3O4), various iron carbides
(FeXC, 2 ? x ? 3) and iron metal (?-Fe) [88-95].
Over many years, reseachers have sought to find the active phase of the Fe catalyst
during the FTS. Historically various studies suggested that magnetite (Fe3O4) was the
active phase [96?101], while other workers have linked the formation of magnetite to
catalyst deactivation [102]. The starting iron oxide or the reduced iron (?-Fe) is known to
transform into iron carbides during reaction. Hence, there are numerous studies that
propose Fe carbides to be the active phase for F?T synthesis [103?109]. Presently there is
25
a resounding backing for Fe carbides to be the active species in FTS and overwhelming
evidence has been presented to back this assertion [110-112].
c) Use of promoters
One way of controlling the product selectivity in an FT reaction is to introduce
promoters into the catalyst. A promoter is considered to be the component of the catalyst
that does not take part in a catalytic reaction but changes the properties of the catalyst.
Promoter chemistry usually contributes in two major ways to catalysis. Firstly a promoter
can improve a catalyst?s structural features by enhancing its surface area while
maintaining its stability in a catalytic reaction. This type of a promoter is often referred to
as a structural promoter. A structural promoter can also act as a barrier or spacer between
active metal crystallites thereby inhibiting sintering or crystallite growth [113].
The second way in which promoters affect catalysts is electronic in nature. This occurs as
a result of a change in the electronic environment of the catalyst surface. This can lead to
enhanced reactant gas?active site interactions which can lead to bonding destabilization
of the reactant gas. This type of promotion is chemical in nature, and the promoter is
referred to as a ?chemical promoter? [50].
Promoters may also serve one or more of the following purposes; they may
(i) supply a catalytic effect not possessed by the catalytic metal alone,
(ii) facilitate catalyst preparation, conditioning, or regeneration,
(iii) inhibit catalyst poisoning, and/or
(iv) improve the physical nature of the support.
The two types of promoters usually employed for improving the Fe catalysts in FTS are
chemical promoters and structural promoters, and these will shortly be discussed.
26
2.8 Promoters
2.8.1 Structural promoters
Typical structural promoters used in F-T catalysis include SiO2 [47, 113, 114], TiO2 [47,
115-118, 119], Al2O3 [47, 120-122], MnO [123], Nb2O5 [124], ZrO2 [115,125-128],
CeO2 [115,15], Cr2O3 [9], ZnO [9], MgO [129] or a mixture of supports such as
MgO/SiO2 [130] and TiO2/SiO2 [131], zeolites [132-134] and molecular sieves [135],
activated carbon [136], carbon nanotubes [84, 137] and nanofibers [138]. SiO2 has been
shown to be a superior structural promoter for precipitated Fe based FTS catalysts [9,
139, 140]. Hence only work relating to SiO2 will be discussed here.
2.8.1.2 Role of SiO2 as a promoter for Fe FT catalysts
SiO2 has been extensively investigated as a catalyst support in the FT reaction. In this
role the support material is used in a large amount relative to the amount of catalyst used.
When smaller amounts are used (typically from 1% - 20%) the materials are called
binders or promoters. In many instances the interaction between the
support/binder/promoter and the catalyst involve the same type of interactions. However,
the concentration effect can have serious implications for the physical properties of the
mixture.
SiO2 can hence either be used as a support or be used in small quantities as a structural
promoter for the Fe-based Fischer Tropsch catalysts. In both cases it is often added to Fe-
based Fischer Tropsch catalysts to maintain surface area [26, 70].
27
2.8.1.3 Low promoter concentrations of SiO2
A major problem in using an iron catalyst without addition of a structural promoter is the
formation of catalyst fines accompanying the physical breakage of the catalysts. The
addition of a binder to a precipitated iron catalyst is beneficial to the formation and
stabilization of small crystallites of the active phase and provides a robust skeletal
structure in the catalyst. This structure is needed to keep the catalyst from breaking down
(a process referred to as attrition) during the processes of activation and FTS reaction
[141-144].
A study by Jothimurugesan et al. [145] on the effect of two binders - a silica-based
system and a silica-kaolin-clay-phosphate-based system - on a doubly promoted Fischer-
Tropsch (FT) synthesis iron catalyst (100Fe/5Cu/4.2K) has revealed that 12 wt.% binder
silica gives the highest attrition resistance when the binder silica content is varied from 0
to 20 wt.%.
SiO2 also has an effect on both catalytic activity and selectivity [1, 9, 144]. Work
highlighting the ability of SiO2 to induce chemical effects on catalytic properties has also
been observed [26, 146, 147]. M?ssbauer spectroscopy studies of precipitated Fe-based
Fischer-Tropsch catalysts (100Fe/5Cu/4.2K/xSiO2), where x = 0, 8, 16, 24, 25, 40 or 100)
have shown that reduction of the oxide precursor in CO gives rise to chi-carbide Fe5C2
whose amount decreases with an increase of SiO2 content [148].
From the work discussed above, it is clear that loading levels of SiO2 onto Fe-based FT
catalysts play a huge role in affecting their chemical properties.
28
2.8.1.4 SiO2 as a support
Apart from SiO2 being used as a structural promoter, it can also be used as a support for
Fe-based Fischer Tropsch catalysts. Especially in recent years, silica has been chosen as
the principal support for the preparation of iron-based catalysts with high attrition
resistance using popular spray-drying technologies. The ability of the silica support to
prevent sintering of the Fe phases has also been observed [149].
Many other advantages of the supported catalysts, such as improved catalyst stability,
decreased deactivation rate, and improved selectivity, have also been identified. The
lower activity of supported Fe catalysts has been attributed to the effect of metal-support
interactions that affect the reducibility of the iron phase [147, 150].
Such metal support interactions and structural properties in highly dispersed catalysts are
frequently mentioned in the literature. Wielers et al. [151] studied the reduction
behaviour of silica-supported iron catalysts and revealed that reduction of the Fe/SiO2
catalyst proceeds via an iron (II) silicate phase. Cagnoli et al. [152] and Bukur et al.
[144], respectively, investigated the influence of the support on the activity and
selectivity of alumina or silica supported catalysts in FTS reaction and their results were
attributed to the interactions between the metal and supports.
Lund and Dumesic [153] studied interactions in silica-supported magnetite catalysts by
spectroscopy and suggested a model in which Si4+ substitutes for Fe3+ in the tetrahedral
sites near the surface of magnetite. In the work of Yeun et al. [154], they suggested that
on a 1 wt% Fe/SiO2 catalyst, Fe2+ strongly interacted with silica during reduction. Jun et
al. [155] studied FTS over SiO2 supported iron-based catalysts from biomass-derived
syngas. They found that the addition of SiO2 leads to the poor dispersion of iron oxide.
Therefore, SiO2 is usually used as a support for FTS catalysts to obtain the desired
physical strength and make it attrition resistant. However the addition of SiO2 as a
29
support also leads to the corresponding poor reducibility due to the strong metal-support
interaction.
In general, the effects of SiO2 on Fe-based FTS catalysts can be summarized as follows:
(i) Changes the catalyst stability and selectivity.
(ii) Decreases the deactivation rate of the catalyst.
(iii) Maintains the surface area and thereby has a chemical effect on catalyst properties.
(iv) Prevents sintering of Fe phases.
(v) Affects the reducibility of the iron phase, especially the transformation of magnetite
to metallic iron. This is attributed to the strong interaction between the metal and the
silica support. This may lead to SiO2 indirectly weakening the surface basicity and
severely suppressing the carburization and CO adsorption of the catalyst.
(vi) Due to the lower surface basicity of the catalyst incorporated with SiO2, a higher
selectivity to light hydrocarbons and methane is observed and a decreased selectivity to
olefins and heavy products is obtained.
2.8.2 Chemical Promoters
These types of promoters affect the electronic nature of the catalyst and their presence
may result in a change in the activity and selectivity of the metal catalyst. Fe catalysts are
significantly affected by the presence of chemical promoters. For all Fe catalysts used in
the FT reaction the promotion with the optimum amount of alkali metal is vital for
satisfactory FT activity as well as the required selectivity.
Potassium [156, 157] is the preferred alkali metal commercially used in FT reaction and
has been known to increase wax and alkene yields while decreasing the production of
undesirable methane [9]. Potassium also has been implicated in increasing FTS and
water-gas shift activity [158].
However, the use of potassium as a chemical promoter may be hampered by its readiness
to form an alkali compound with common catalyst supports, or structural promoters such
30
as alumina or silica. Also, high potassium loadings may cover too large of a fraction of
the surface of the iron catalyst, resulting in a limited promotion effect or even a decrease
in FTS conversions.
Work done by O?Brein et al. [157], showed that a high potassium loading is required
when the FTS reaction temperature is decreased because it becomes harder to dissociate
the C-O bond. They found that the optimum potassium promotion was 4-5 atomic%
relative to iron.
On the contrary, Dry [41] reported that catalytic activity decreases for the low
temperature FTS as the potassium loading is increased but the opposite effect is observed
when performing a high temperature FTS reaction. He reported that at 200 ?C, the
relative catalyst activity decreased when the relative K2O content is increased from 0 to
2.6 %. At 330 ?C, the catalytic activities first increased and stabilised at a certain level as
the relative K2O content increased above 3 %. Furthermore, Davis and co-workers [68]
have found that potassium loading to give a K/Fe atomic ratio of greater than 5 failed to
further enhance the CO conversion.
It is apparent that when potassium is added in moderation to Fe-based FTS catalysts, it
enhances its characteristics. This is because when potassium containing catalysts are
heated the potassium migrates to the top (the surface) of the catalyst [57] and has a direct
influence on the active catalyst sites. Therefore if a high loading of potassium is used,
this may be detrimental to the catalyst as more of it will move to the surface and block
some of the catalyst active sites leading to lower FTS activity.
Therefore the FTS activity either increases [158] or passes through a maximum as a
function of potassium loading [9, 159], and potassium either has no effect on the activity
for FTS [157] or suppresses it [57, 159].
31
From the findings above, it appears as if the optimum positive effects of potassium on Fe-
based FTS catalysts are obtained at low loadings not greater than 1 - 5 atomic % relative
to Fe. This is obviously dependent on the reaction conditions that are employed.
Most researchers that have studied the influence of potassium on the Fe-based FTS
catalysts have come to a general consensus that potassium has the following effects on
Fe-based FTS catalysts:
(i) Influences FTS activity. The FTS activity either increases or passes through a
maximum as a function of potassium loading
(ii) Increases the activity of the WGS reaction
(iii) Potassium and other alkali metals decrease the sticking probability of the CO and H2
molecules over the iron surface and increase their probability of dissociation
(iv) Potassium leads to higher olefin-to-paraffin ratio and decreases the methane
selectivity
(v) Produces longer hydrocarbon chains. This is favorable for gasoline production
because the yield of liquid hydrocarbons increases.
(vi) Increases its heat of adsorption of CO, rate of carbon deposition and rates of
hydrocarbon chain growth
Another promoter that is commercially employed for the Fe-based FTS catalyst is copper.
Copper has traditionally been added in precipitated iron catalysts to facilitate reduction of
iron oxide to metallic iron during hydrogen activation [157]. Copper has been shown to
minimize sintering of iron catalysts when activating with hydrogen by lowering the
reduction temperature [9].
It has also been found that copper promotion appears to favour the formation of iron
carbides [160]. It may be possible that copper increases the activity of iron catalysts by
increasing the number of active sites that are formed. Assuming that the active site(s) is a
zero valence surface species, copper may serve as a means of preventing oxidation of the
active metallic iron or iron carbide.
32
Wachs et al. [161], Anderson [25] and O?Brein et al. [157] observed that copper had no
effect on product selectivity. However, Bukur et al. [158] have reported that
incorporating Cu into iron-based catalysts results in an increase in the average molecular
weight of hydrocarbon products.
Copper also appears to influence the WGS reaction and carburization. As demonstrated
earlier, the water-gas-shift reaction produces H2 from the reaction of H2O and CO (CO +
H2O ? CO2 + H2) and this enables the Cu-promoted Fe-based catalyst to be used for
syngas with low levels of H2 for the FTS reaction. This is also consistent with copper
being used in commercial low-temperature water-gas shift catalysts. Bukur et al. [158]
have also found a high water-gas shift activity for their Cu-promoted Fe-based catalyst
when performing FT synthesis reactions at 260oC. A higher carbon dioxide amount was
obtained for the catalyst with copper which indicated that copper is a promoter for the
water-gas shift reaction.
Dry [9] has stated that the precipitated iron catalyst developed by Ruhrchemie and used
in the fixed-bed reactors at SASOL contains about 5% wt. Cu. Work carried out by
Linder and Papp [162] using XPS and ISS (Ion Scattering Spectroscopy) have shown that
the degree of reduction of the Fe catalyst is strongly influenced by the amount of Cu that
is added to the catalyst. They observed that the highest amount of Fe0 (which gets
converted to the active iron carbide phase during the FT reaction) in the surface of their
Cu containing samples was obtained when 1 atomic% of Cu was added
When adding Cu above 1 %, they noticed a slight decrease in the metallic Fe (Fe0)
content of the surface. They speculated that the decrease may be due to a decrease in
dispersion of Fe and that this was as a result of sintering at higher Cu contents to bigger
Fe agglomerates leading to a lower relative amount of Fe0 measured on the surface. They
went on to conclude that all their Cu containing samples should have a higher activity in
the FT synthesis than the unpromoted Fe oxide catalyst and that those with ? 1 atomic%
of Cu should have the highest activity.
33
Meanwhile work performed by O?Brein et al. [157] has shown that reduction of iron
oxide using hydrogen is accelerated with increasing levels of copper promotion (2.6-5.0
atomic % relative to iron). They argued that the acceleration of the iron oxide reduction
(with increasing levels of copper promotion) is in agreement with more nucleation sites
being available with an increasing amount of copper.
From the findings given above, it is clear that the effect of copper loading is very much
dependent on the experimental conditions used. Also, not much work has been reported
on obtaining the optimum loading of Cu on Fe-based FTS catalysts.
The effects of Cu on Fe-based FTS catalysts can be summarized as follows:
(i) Aids in the reduction of iron oxide
(ii) Has an influence on the FTS and WGS activities
(iii) Plays a small role in FTS product selectivity
2.8.2.1 Use of indium as a chemical promoter for Fe-based catalysts
Indium as a chemical promoter for Fe-based catalysts has been reported in the literature
[163]. To the best of our knowledge it has not been used in the Fischer-Tropsch
Synthesis. It has mainly been employed in reactions such as the selective catalytic
reduction (SCR) of NOx using hydrocarbons [163-165].
34
References
[1] R.B. Anderson, The Fischer-Tropsch Synthesis, Academic Press, Orlando, 1984
[2] T. Bromfield, The effect of low-level sulfide addition on the performance of
precipitated-iron Fischer-Tropsch catalysts, PhD Thesis, University of the
Witwatersrand, Johannesburg, 1997
[3] G.C. Bond, Catalysis by Metals, Academic Press, London, 1962
[4] L. Guczi, Stud. Surf. Sci. Catal. 64 Series: New Trends in CO Activation, Elsevier,
Amsterdam, (1991)
[5] H.H. Storch, N. Golumbic, R.B. Anderson, The Fischer-Tropsch and related
syntheses: including a summary of theoretical and applied contact catalysis,
John Wiley, New York (1951)
[6] B.H. Davis, Final Report, Technology development for iron Fischer-Tropsch
catalysis, Contract No. DE-AC22-94PC94055?13 (1999)
[7] M.E. Dry, Appl. Catal. A: Gen. 138 (1996) 319
[8] J.G. Price, An investigation into novel bimetallic catalysts for use in the Fischer-
Tropsch reaction, PhD Thesis, University of the Witwatersrand, Johannesburg, 1994
[9] M.E. Dry, The Fischer-Tropsch Synthesis, in Catalysis Science and Technology, J. R.
Anderson, M. Boudart (Eds.) 1, Springer-Verlag, New York (1981) 159
[10] E. Iglesia, Stud. Surf. Sci. Catal. 107 (1997) 153
[11] M.E. Dry, Appl. Catal. A: Gen. 276 (2004) 1
[12] M.E. Dry, Catal. Today 71 (2002) 227
[13] G. Blyholder, D. Shihabi, W.V. Wyatt, R. Bartlett, J. Catal. 43 (1976) 43
[14] G. Henrici-Oliv?, S. Oliv?, J. Mol. Catal. 3 (1977/78) 443
[15] J.T. Kummer, P.H. Emmett, J. Am. Chem. Soc. 75 (1953) 5177
[16] J.T. Kummer, H. H. Podgurski, W.B. Spencer, P.H. Emmett
J. Am. Chem. Soc., 73 (1951) 564
[17] M. J. Overett, R. O. Hill, J. R. Moss, Coord. Chem. Rev.
206?207 (2000) 581
[18] R. George, J.M. Andersen, J.R. Moss, J. Organomet. Chem. 505 (1995) 131
[19] P.M. Maitlis, J. Organomet. Chem. 689 (2004) 436
35
[20] P. Sabatier, J.B. Senderens, C.R. Acad. Sci. 134 (1902) 514
[21] BASF, German Patent 293 (1913) 787
[22] H. Pichler, Adv. Catal. 4 (1952) 271
[23] http://www.fischer-tropsch.org
[24] K. Khoabane, Synthesis of nanostructured silica for use as a support for iron
Fischer-Tropsch catalysts, PhD Thesis, University of the Witwatersrand, Johannesburg,
2007
[25] R.B. Anderson, in Catalysis, P.H. Emmet (Ed.), 4, Von Norstand?Reinhold, New
Jersey, 1956
[26] M.E. Dry, in Applied Industrial Catalysis, B.E. Leach (Ed.) 2, Academic Press, New
York, 1983
[27] K. Jalama, Fischer Tropsch Synthesis over supported cobalt catalysts: Effect of
ethanol addition, precursors and gold doping, PhD Thesis, University of the
Witwatersrand, Johannesburg, 2007
[28] P.L. Spath, D.C. Dayton, Preliminary Screening ? Technical and Economic
Assessment of Synthesis Gas to Fuels and Chemicals with Emphasis on the Potential for
Biomass-Derived Syngas, NREL/TP-510-34929 (2003) [available at www.fischer-
tropsch.org]
[29] http://www1.eere.energy.gov/biomass/printable_versions/catalytic_conversion
[30] http://en.wikipedia.org/wiki/Fischer-Tropsch_process
[31] D.K. Rider, Energy: Hydrocarbon Fuels and Chemical Resources, John Wiley, New
York, 1981
[32] B. Jager, M.E. Dry, T. Shingles, A.P. Steynberg, Catal. Lett. 7 (1990) 293
[33] P.D.F. Vernon, M.L.H. Green, A.K. Cheetham, A.T. Ashcroft, Catal. Lett. 6 (1990)
181
[34] J. Abbott, B. Crewdson, Oil & Gas J. 100 (2002) 70
[35] V.R. Choudhary, K.C. Mondal, A.S. Mamman, J. Catal. 233 (2005) 36
[36] P.M. Biesheuvel, G.J. AIChE J. 49 (2003) 1827
[37] G.P. Van der Laan, Kinetics, Selectivity and Scale Up of the Fischer-Tropsch
Synthesis, PhD thesis, University of Groningen, The Netherlands, 1999
[38] J.M. Fox, III, Catal. Rev. Sci. Eng. 35 (1993) 169
36
[39] J.H. Gregor, Catal. Lett. 7 (1990) 317
[40]. S.C. Saxena, M. Rosen, D.N. Smith, J.A. Ruether, Chem. Eng. Comm., 40
(1986) 97
[41] M.E. Dry, Catal. Today 6 (1990) 183
[42] A.A. Adesina, Appl. Catal. A: Gen. 138 (1996) 345
[43] B.H. Davis, Catal. Today 71 (2002) 249
[44] B. Jager, Stud. Surf. Sci. Catal. 107 (1997) 219
[45] R.L. Espinoza, A. P. Steynberg, B. Jager, A. C. Vosloo, Appl. Catal. A: Gen. 186
(1999) 13
[46] B. Jager, R. L. Espinoza, Catal. Today 23 (1995) 17
[47] M.E. Dry, in Studies in Surface Science and Catalysis: Fischer-Tropsch
Technology, A. Steynberg, M. Dry, (eds.), 152, Elsevier, New York, 2004
[48] B.H. Weil, J.C. Lane, The Technology of the Fischer-Tropsch Process, Constable
& Co. LTD, London, 1949
[49] D.B. Bukur, X. Lang, Ind. Eng. Chem. Res. 38 (1999) 3270
[50] E.M. Mokoena, Synthesis and Use of Silica Materials as Supports for the
Fischer-Tropsch Reaction, PhD Thesis, University of the Witwatersrand,
Johannesburg, 2005
[51] K.G. Anderson, J. G. Ekerdt, J. Catal. 95 (1985) 602
[52] H. Pichler, H. Schultz, Chem. Ing. Techn. 42 (1970) 1102
[53] J.W. Thomas, W.J. Thomas, Principles and Practice of Heterogeneous Catalysis,
VCH, New York, 1996
[54] C.N. Satterfield, G.A. Huff, J.P. Longwell, Ind. Eng. Chem. Process.
Dev., 21,3 (1982) 466
[55] R.J. Madon, W.F. Taylor, J. Catal. 69 (1981) 32
[56] C.N. Mbileni, Applications of mesostructured carbonaceous materials as supports
for Fischer-Tropsch metal catalyst, PhD Thesis, University of the Witwatersrand,
Johannesburg, 2006
[57] R.A Dictor, A.T.J. Bell, J. Catal. 97 (1986) 121
[58] R.J. Madon, S.C. Reyes, E. Iglesia, J. Phys. Chem. 95 (1991) 7795
[59] V. Ponec, Coal Science 3 (1984) 1
37
[60] C.N. Satterfield, G.A. Huff, J. Catal. 73 (1982) 187
[61] G.A. Huff, C.N. Satterfield, J. Catal. 85 (1984) 370
[62] E. Iglesia, S.C. Reyes, R.J. Madon, J. Catal. 129 (1991) 238
[63] E.W. Kuipers, C. Scheper, J.H. Wilson, I.H. Vinkenburg, H. Oosterbeek, J.
Catal. 158 (1996) 288
[64] B. Shi, B.H. Davis, Appl. Catal. A: Gen. 277 (2004) 61
[65] G.Y. Chai and J.L. Falconer, J. Catal. 93 (1985) 152
[66] M.A. Vannice, J. Catal. 37 (1975) 449
[67] M.A. Vannice, J. Catal. 50 (1977) 228
[68] B.H. Davis, Final Report, Technology Development for Iron Fischer-Tropsch
Catalysts, Contract No. DE-AC22-91PC90056, 1999
[69] M. Jonardanarao, Ind. Eng. Chem. Res. 29 (1990) 1735
[70] V.U.S. Rao, G.J. Stiegel, G.J. Cinquegrane, R.D. Strvastava, Fuel Proc.
Tech. 30 (1992) 83
[71] M.E. Dry, Appl. Catal. A: Gen. 189 (1999) 185
[72] C. Bartholomew, Appl. Catal. A: Gen. 212 (2001) 17
[73] N. Sirimanothan, H. H. Hamdeh, Y. Zhang, B. H. Davis, Catal. Lett.
82 (2002) 181
[74] D.B. Bukur, W.-P. Ma, V. Carreto-Vazquez, Topics Catal. 32 (2005) 135
[75] H. Schulz, T. Riedel, G. Schaub, Topics. Catal. 32 ( 2005) 117
[76] S.A. Eliason, C.H. Bartholomew, Appl. Catal. A: Gen. 186 (1999) 229
[77] L.D. Mansker, Y. Jin, D. B. Bukur , A. K. Datye, Appl. Catal. A: Gen. 186 (1999)
277
[78] R.J. O'Brien, L. Xu, R.L. Spicer, S. Bao, D.R. Milburn, B.H. Davis Catal. Today 36
(1997) 325
[79] D.B. Bukur, K. Okabe, M.P. Rosynek, C. Li, D. Wang, K.R.P.M. Rao, G.P.
Huffman, J. Catal. 155 (1995) 353
[80] D.B. Bukur, M. Koranne, X. Lang, K.R.P.M. Rao, G.P. Huffman, Appl. Catal. A:
Gen. 126 (1995) 85
[81] D.J. Duvenhage, N.J. Coville, Appl. Catal. A: Gen. 298 (2006) 211
[82] T.C. Bromfield, N.J. Coville, Appl. Catal. A: Gen. 186 (1999) 245
38
[83] T.C. Bromfield, N.J. Coville, Appl. Surf. Sci., 119 (1997) 19
[84] M.C. Bahome, L.L. Jewell, D. Hilderbrandt, D. Glasser, N.J. Coville,
Appl. Catal. A: Gen. 287 (2005) 60
[85] B.H. Davis, E. Iglesia, Technology Development for Iron and Cobalt Fischer-
Tropsch Catalysis, Final Report, Contract No. DE-FC26-98FT40308, University of
Kenturky Research Foundation, Lexington, 2002
[86] D.B. Bukur, L. Nowicki, X. Lang, Energy and Fuels 9 (1995) 620
[87] R.J. O'Brien, L. Xu, R.L. Spicer, B.H. Davis, Energy and Fuels 10 (1996) 921
[88] D.J. Dwyer, G.A. Somorjai, J. Catal. 52 (1978) 291
[89] J.A. Amelse, J.B. Butt, L.J. Schwartz, J. Phys. Chem. 82 (1978) 558
[90] G.B. Raupp, W.N. Delgass, J. Catal., 58 (1979) 348
[91] H. Jung, W.J. Thomson, J. Catal. 134 (1992) 654
[92] D.J. Dwyer, J.H. Hardenbergh, J. Catal. 87 (1984) 66
[93] K.R.P.M . Rao, F.E. Huggins, V. Mahajan, G.P. Huffman, V.U.S. Rao, B.L. Bhatt,
D.B. Bukur, B.H. Davis, R.J. O?Brien, Top. Catal. 2 (1995) 71
[94] E.S. Lox, G.B. Marin, E. de Grave, P. Bussi?re, Appl. Catal. 40 (1988) 197
[95] H.-B. Zhang, G.L Schrader, J. Catal. 95 (1985) 325
[96] F.B. Blanchard, J.P. Raymond, B. Pommier, S.J. Teichner, J. Mol. Catal. 17 (1982)
171
[97] J.P. Reymond, P. Meriadeau, S.J. Teichner. J. Catal. 75 (1982) 39
[98] F. Blanchard, J.P. Reymond, B. Pommier, S.J. Teichner, J. Mol. Catal. 17 (1982)
171
[99] S. Soled, E. Iglesia, R.A. Fiato, Catal. Lett. 7 (1990) 271
[100] C.S. Kuivila, P.C. Stair, J.B. Butt, J. Catal. 118 (1989) 299
[101] J.B. Butt, Catal. Lett. 7 (1990) 61
[102] D.J. Duvenhage, R.L. Espinoza, N.J. Coville, Stud. Surf. Sci. Catal. 88 (1994) 351
[103] G.B. Raupp, W.N. Delgass, J. Catal. 58 (1979) 348
[104] J.A. Amelse, L.H. Schwartz, J.B. Butt, J. Catal. 72 (1981) 95
[105] M.D. Shroff, D.S. Kalakkad, K.E. Coulter, S.D. Kohler, M.S. Harrington, N.B.
Jackson, A.G. Sault, A.K. Datye, J. Catal. 156 (1995) 185
39
[106] J.W. Niemantsverdriet, A.M. van der Kraan, W.L. van Dijk, H.S. van der Baan, J.
Phys. Chem. 84 (1980) 3363
[107] J.W. Niemantsverdriet, A.M. van der Kraan, J. Catal. 72 (1981) 375
[108] D.S. Kalakkad, M.D. Shroff, S.D. K?hler, N.B. Jackson, A.K. Datye, Appl. Catal.
A: Gen. 133 (1995) 335
[109] D.B. Bukur, L. Nowicki, R.K.. Manne, X. Lang, J. Catal. 155 (1995) 366
[110] J.T. Kummer, T.W. Dewitt , EH. Emmet, J. Am. Chem. Soc. 70 (1948) 3632
[111] D.M. Stockwell, D. Bianchi, C.O. Bennet, J. Catal. 113 (1988) 13
[112] J.A. Amelse, J.B. Butt, L.H. Schwartz, J. Phys. Chem. 82 (1978) 558
[113] M.E. Dry, J.A.K. du Plessis, G.M. Leuteritz, J. Catal. 6 (1966) 194
[113] A. Khodakov, A. Griboval-Constant, R. Bechara, V.L. Zholobenko, J. Catal.
206 (2002) 230
[114] G.H. Haddad, B. Chenand, J.G. Goodwin, J. Catal. 160 (1996) 43
[115] M. Kraun, M. Baerns, Appl. Catal. A. Gen. 186 (1999) 189
[116] R.C. Reuel, C.H. Bartholomew, J. Catal. 85 (1984) 63
[117] E. Iglesia, S.L. Soled, R.A. Fiato, G.H. Via, J. Catal. 143 (1993) 345
[118] E. Iglesia, S.L. Soled, R.A. Fiato, J. Catal. 137 (1992) 212
[119] E. Iglesia, Appl. Catal. 161 (1997) 59
[120] R. Bechara, D. Balloy, D. Vanhove, Appl. Catal. A: Gen. 207 (2001) 343
[121] D. Schanke, A.M. Hillmen, E. Bergene, K. Kinnari, E. Rytter, E. ?dnanes, A.
Holmen, Energy and Fuels 10 (1996) 867
[122] A. Holmen, D. Schanke, S. Vada, E.A. Blekkan, A.M. Hillmen, A. Hoff, J.
Catal. 156 (1995) 85
[123] F. Morales, E. de Smit, F.M.F. de Groot, T. Visser, B.M. Weckhuysen, J. Catal.
246 (2007) 91
[124] F.B. Noronha, A. Frydman, D.A.G. Aranda, C. Perez, R.R. Soares, B. Morawek,
D. Castner, C.T. Campbell, R. Frety, M. Schmal, Catal. Today 28 (1996) 147
[125] L.A. Bruce, M. Hoang, A.E. Hughes, T.W. Turney, Appl. Catal. A: Gen. 100
(1993) 51
[126] J. van der Loosdrecht, M. van der Haar, A.M. van der Kraan, A.J. van Dillen,
J.W. Geus, Appl. Catal. 150 (1997) 365
40
[127] M.K. Niemel?, A.O.I. Krause, T. Vaara, J. Lathinen, Top. Catal. 2 (1995) 45
[128] M.K. Niemel?, A.O.I. Krause, T. Vaara, J.J. Kiviaho, M.K.O. Reikainen,
Appl. Catal. A. Gen. 147 (1996) 325
[129] G.S. Sewell, E. van Steen, C.T. O.Connor, Catal. Lett. 37 (1996) 255
[130] R.J. O.Brien, L. Xu, S. Bao, A. Raje, B.H. Davis, Appl. Catal. A: Gen. 196
(2000) 173
[131] B. Jongsomjit, T. Wongsalee, P. Praserthdam, Mater. Chem. Phys. 97 (2006) 343
[132] J. Eilers, S.A. Posthuma, S.T. Sie, Catal. Lett. 37 (1997) 253
[133] S. Bessel, Appl. Catal. A: Gen. 96 (1993) 253
[134] V.U.S. Rao, R.J. Gormley, Catal. Today 6 (1990) 207
[135] D. Yin, W. Li, W. Yang, H. Xiang, Y. Sun, B. Zhong, S. Peng, Microporous and
Mesoporous Mat. 47 (2001) 15
[136] H.-J. Jung, P.L. Walker Jr., M.A. Vannice, J. Catal. 75 (1982) 416
[137] E. van Steen, F.F. Prinsloo, Catal. Today 71 (2002) 327
[138] G.L. Bezemer, J.H. Bitter, H.P.C.E. Kuipers, H. Oosterbeek, J.E. Holewijn, X.
Xu, F. Kaptein, A.J. van Dillen, K.P. de Jong, J. Am. Chem. Soc. 128 (2006)
3956
[139] Y. Yang, H. Xiang, L. Tian, H. Wang, C. Zhang, Z. Tao, Y.
Xu, B. Zhong, Y. Li, Appl. Catal. A: Gen. 284 (2005) 105
[140] H. Dlamini, T. Motjope, G. Joorst, G. ter Stege, M. Mdleleni, Catal. Lett.
78 (2002) 201
[141] R. Zhao, J.G. Goodwin Jr., K. Jothimurugesan, S.K. Gangwal, J.J. Spivey, Ind.
Eng. Chem. Res. 40 (2001) 1065
[142] K. Sudsakorn, J.G. Goodwin Jr. K. Jothimurugesan, A.A. Adeyiga, Ind. Eng.
Chem. Res. 40 (2001) 4778
[143] H.N. Pham, A. Viergutz, R. Gormley, A.K. Datye, Powder Technol. 110 (2000)
196
[144] D.B. Bukur, X. Lang, D. Mukesh, W.H. Zimmerman, M.P. Rosynek, C. Li, Ind.
Eng. Chem. Res. 29 (1990) 1588
[145] K. Jothimurugasen, J.G. Goodwin, S.K. Gangwal, J.J. Spivey, Catal. Today 58
(2000) 335
41
[146] J.L. Rankins, C.H. Bartholomew, J. Catal. 100 (1986) 533
[147] S.L. Soled, E. Iglesia, S. Miseco, B.A. DeRites, R.A. Fiato, Top. Catal. 2 (1995)
193
[148] K.R.P.M Rao, F.E. Huggins, V. Mahajan, G.P. Huffman, D.B. Bukur, V.U.S. Rao,
Hyperfine Interactions 93 (1994) 1751
[149] Y. Jin, A.K. Datye, J. Catal. 196 (2000) 8
[150] M.A. Vannice, J. Catal. 37 (1975) 462
[151] A.F.H. Wielers, A.J.H.M. Kock, C.E.A. Hop, J.W. Geus, A.M. van der Kraan, J.
Catal. 117 (1989) 1
[152] M.V. Cagnoli, S.G. Marchetti, N.G. Gallegos, M. Alvarez, R.C. Mercader, A.A.
Yeramin, J. Catal. 123 (1990) 21
[153] C.R.F. Lund, J.A. Dumensic, J. Phys. Chem. 85 (1981) 3175
[154] S. Yuen, Y. Chen, J.E. Dumesic, J. Phys. Chem. 86 (1982) 3022
[155] K.W. Jun, H.S. Roh, K.S. Kim, J.S. Ryu, K.W. Lee, Appl. Catal. A: Gen. 259
(2004) 221
[156] M. Luo, R.J. O?Brien, S. Bao, B.H. Davis, Appl. Catal. A: Gen. 239 (2003) 111
[157] R. J. O'Brien, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal.
Today 36 (1997) 325
[158] D.B. Bukur, D. Mukesh, S.A. Patel, Ind. Eng. Chem. Res. 29 (1990) 194
[159] D.G. Miller, M. Moskovits. J. Phys. Chem. 92 (1988) 6081
[160] K.R.P.M. Rao, F.E. Huggins, G.P. Huffman, R.J. Gormly, R.J. O?Brein, B.H.
Davis, Energy and Fuels 10 (1996) 546
[161] I.E. Wachs, D.J. Dwyer, E. Iglesia, Appl. Catal. 12 (1984) 201
[162] U. Lindner, H. Papp, Appl. Surf. Sci. 32 (1988) 75
[163] X. Wang, T. Zhang, X. Sun, W. Guan, D. Liang, L. Lin, Appl. Catal. B:
Environmental 24 (2000) 169
[164] R. Serra, M. J. Vecchietti, E. Mir?, A. Boix, Catal. Today 133?135 (2008) 480
[165]X. Wang, X. Zhao, J. Shen, X. Sun, T. Zhang, L. Lin, Phys. Chem. Chem. Phys. 4
(2002) 2846
[166] M.C. Bahome, Synthesis and use of carbon nanotubes as a support for the Fischer-
Tropsch Synthesis, PhD Thesis, University of the Witwatersrand, Johannesburg, 2007
42
Chapter 3
Experimental
3.1 Catalyst preparation
All the catalysts were prepared using the co-precipitation method [1, 2]. Fe (NO3)3.9H2O,
Cu(NO3)2.3H2O, In(NO3)3 and KNO3 were used as precursors, while the ammonia
solution (25% NH3) was used as a precipitating agent (All were purchased from MERCK
Chemicals (PTY) LTD). SiO2 (purchased from Davisil with BET surface area = 303 m2/g
and pore volume = 1.05 cm3/g) in form of a white powder was also employed as part of
the reagents.
The nitrate precursors were dissolved in distilled water. This was followed by stirring
using an overhead stirrer. While stirring, SiO2 was added (when needed for the
preparation of catalysts that contained SiO2). The ammonia solution was then added
dropwise to produce a brown slurry. The resultant slurry (precipitate) was stirred for 15
minutes and the final pH (pH = 8-9) was recorded. The slurry was dried at 120oC
overnight to give a brown solid. The dried slurry was then calcined for 4 hours at 350oC.
The calcined catalyst was ready for analysis using N2 physisorption, XRF, XPS, XRD,
TPR, DRIFTS and FTS reactor studies.
43
3.2 Catalyst characterization
3.2.1 X-Ray Fluorescence (XRF) spectroscopy
The XRF experiments were carried out using a PW2404 wavelength dispersive XRF
spectrometer from Panalytical. A Rh target tube was used to generate the X-Rays with
K? = 24.9 and K? = 22. The samples were mixed with polyvinyl glue (Mowiol) and were
pressed to pellets using 10 MPa pressure prior to analysis.
3.2.2 N2 Physisorption
N2 physisorption was employed for surface-area determination and pore volume
measurements of the calcined catalysts. It is noted that the surface areas could change
significantly following various pretreatments and could be different from those
determined after calcination. For consistency and comparison purposes, surface areas
reported in this thesis were determined on only calcined samples. The samples were
degassed using N2 at 150 ?C for 2 hours before measurement. N2 adsorption-desorption
isotherms at N2 boiling point (-196 ?C) were measured on a Micromeritics TRISTAR
3000 analyzer (Fig. 3.1). The surface areas were determined by the Brunauer-Emmett-
Teller (BET) method.
44
Figure 3.1 The TRISTAR 3000 analyzer
3.2.3 Temperature programmed reduction (TPR)
Temperature programmed reduction (TPR) was used to assess the reducibility of the
catalysts. The home-build apparatus used (Fig. 3.2) was the same as that used by
Duvenhage [3], Mokoena [4] and Bahome [5].
45
Figure 3.2 Experimental set-up for TPR measurements
The catalyst sample was first weighed before being loaded into a U-shaped quartz tube.
Typical mass values weighed were ca. 20 mg. A glass wool plug was inserted into the U-
tube before the catalyst was added. This was to circumvent any of the catalyst material
being carried into the reactor outlet. The ends of the U-tube were then attached to the gas-
inlet and outlet points of the apparatus.
The flow rate of the affluent gas stream was kept at 50 ml/min and a thermal conductivity
detector (TCD) was used to monitor the concentration variation of the gas stream. The
TCD output was calibrated based upon 100% reducibility of Ag2O powder.
For CO TPR measurements the temperature was ramped from room temperature to
800?C under a flow of 10% CO balanced in Helium. For the H2 TPR measurements the
temperature was ramped from room temperature to 900 ?C under a flow of 5% H2
balanced in Argon. The temperature of the sample was monitored by a thermocouple
placed in the catalyst bed.
46
3.2.4 Diffuse Reflectance Infrared Fourier Transform (DRIFT) Spectroscopy
A Bruker Tensor 27 infrared spectrometer fitted with a Harrick Praying Mantis Diffuse
Reflectance accessory was employed. Typically 50 mg of the catalyst was loaded into the
DRIFTS cell fitted with ZnSe windows (Fig. 3.3).
Figure 3.3 DRIFTS cell with ZnSe windows
The cell was equipped with a heating system that allowed operation under different
temperatures and pressures. Spectra were collected at a resolution of 4 cm-1 and an
average of 64 scans were employed during the measurements. Gases were led into the
cell using a homemade gas manifold (Fig. 3.4).
47
Figure 3.4 Gas manifold for the introduction of gases into the DRIFTS cell
3.2.5 X-Ray Diffraction (XRD) measurements
Powder samples were loaded on a sample holder and their diffraction patterns were
recorded from 5 to 90o 2? on a Brucker D8 X-Ray diffractometer using monochromatized
Cu K? radiation (40 kV, 40 mA). The Bruker D8 X-Ray diffractometer employed for
carrying out the measurements is shown below (Fig. 3.5).
48
Figure 3.5 The Bruker D8 X-Ray diffractometer
Typically a diffraction pattern as illustrated in Fig. 3.6 was obtained after XRD analysis.
From this pattern the iron oxide phase was identified using the reported diffraction
patterns in the Diffracplus evaluation package with the aid of the EVA (V11.0) software
package. It is to be noted that this was done for all the catalysts employed in this thesis
and the only phase identified after calcination was the hematite (Fe2O3) phase.
49
Figure 3.6 Diffraction pattern obtained after the XRD measurement of Fe2O3
Rietveld refinement was also employed to estimate the average crystallite size of Fe2O3.
During Rietveld refinement the diffraction peaks were fitted using mathematical
functions (Gaussian, Lorentzian and Pearson functions). The idea was to try and
minimize the differences between the fitted curve and the experimental diffraction
pattern. A fit was deemed excellent, if the difference curve between the observed and
calculated curves was minimized and revealed as a straight line as illustrated in Fig. 3.7b.
The average crystallite size was estimated using the fit and mathematical equations
within the EVA software package.
1 0 2 0 3 0 4 0 5 0 6 0 7 0
Li
n (
C
ou
nt
s)
2 T h e t a
50
Figure 3.7 Fitting of the experimental diffraction pattern (a) blue line represents the
experimental pattern and red line is the fitted curve (b) difference curve produced after
fitting the experimental diffraction pattern
a
b
2 Theta
51
3.2.6 X-Ray Photoelectron Spectroscopy (XPS)
The surface analysis of all the catalysts was performed using the XPS instrument based at
the University of Cardiff in Wales, United Kingdom. The AXIS UltraDLD manufactured
by KRATOS Analytical (A Shimadzu Group Company) was employed and the set-up is
illustrated below (Fig. 3.8).
Figure 3.8 The AXIS UltraDLD XPS instrument
The samples were placed on a stainless steel bar as depicted in Fig. 3.9 below and were
transferred into the analysis chamber of the XPS instrument.
52
Figure 3.9 The stainless steel bar showing the mounted catalysts ready for XPS analysis
53
3.3 Catalytic evaluation
3.3.1 FTS reactor studies
3.3.1.1 Gases
All gases used were supplied by AFROX (African Oxygen) Ltd. The gas used for catalyst
reduction prior to the FT synthesis was an Ultra High Purity (UHP) grade Carbon
monoxide gas (99.97 % purity) and only this gas was used for all the reduction reactions.
Gas cylinders containing H2/CO/N2 mixtures (60.2 %/29.6 %/10.2 % v/v) were used to
supply the reactant gas stream to the catalyst. N2 was used as an internal standard in order
to ensure accurate mass balances.
3.3.1.2 Catalyst reactor setup
The fixed bed reactor system is shown in Fig. 3.10 and was used for all the FT reactions.
It consisted of a ?? Swagelok stainless steel pipe and this served as the reactor. The
reactor was placed into a heating jacket to maintain a constant temperature profile across
it and inside the reactor a 1/4? Swagelok stainless steel pipe was placed. Quartz wool was
placed on top of this 1/4? pipe so that the catalyst bed could rest on it. The 1/4? pipe had
an opening at the bottom for the ejection of liquid and wax products into the traps.
54
Figure 3.10 The fixed bed reactor made from a ?? Swagelok stainless steel pipe.
A = Sketch portrait; B = Digital portrait
Reactor
Heating jacket
B
A
55
All gas lines after the reactor were kept at 150 oC as shown in Fig. 3.12 and a hot trap
(Fig. 3.11) placed immediately after the reactor was held at this temperature in order to
collect wax. A second trap kept at ambient temperature was used to collect the oil and
water mixture. The flow rate was controlled using a needle valve and measured by a
means of a bubble meter.
Figure 3.11 The hot trap placed in a heating jacket, both situated below the reactor
Reactor
Heating
jacket
56
Figure 3.12 Traps, pressure regulator, needle valve and gas line after reactor
Both the collected wax and liquid products were decanted and were analysed using an
offline gas chromatograph (G.C.). The gaseous stream which was not collected in the hot
and cold traps was analysed online using two GCs and both of them are depicted in Fig.
3.13. Table 3.1 below illustrates the instrumental characteristics for the GCs used [6] and
an overview of the detailed schematic representation of the reactor setup is depicted in
Fig 3.14.
Hot trap
Cold trapPressure
regulator
Needle
valve
Gas line
57
Table 3.1 Characteristics of the GCs employed
Online GC
Make PYE Unicam (Series 204)
Column type
Packed, stainless steel, 2m x 2.2mm, O.D = 1/8"
Stationary
phase
Carbosieve S-II, 60-80 mesh
Detector
Thermal conductivity detector (TCD)
Online GC
Make Hewlett Packard 5890
Column type
Packed, stainless steel, 1.5 m x 2.2 mm, O.D = 1/8"
Stationary
phase
ZB-5, 80/100 mesh
Detector
Flame ionization detector
(FID)
Offline GC
Make Varian 3700
Column type
30 m x 5 ?FT, O.D.= 0.53 mm
Stationary
phase
ZB-1
Detector
Flame ionization detector
(FID)
58
Figure 3.13 GC on the left fitted with an FID detector and the one on the right fitted with
a TCD detector
59
Figure 3.14 Schematic representation of the reactor setup
3.3.1.3 Activity measurement of catalysts
Catalyst (0.1 g) was added to the reactor and reduced in situ at 350 ?C for 20-24 hours
under a stream of CO (2 bar pressure, 12 ml/min). After reduction, the temperature was
decreased to room temperature. Synthesis gas was introduced and the pressure was
gradually increased to 10 bar. The temperature was then ramped to 200?C for 40 minutes
and thereafter, ramped from 200 ?C to 275 ?C for a period of 1 hour. The FTS reaction
was then carried out at 275?C for a period of 140 hours
To
React
P
P
H2
H2
C
N2 TC FICold Trap
Integrat
FI : Flame ionization
P : Pressure
PI : Pressure indicator
TC : Thermal conductivity
TI : Temperature indicator
P
Key
vent
tor
R
R
O
D D
Hot Trap
Computer with
Clarity software
S ft
PR Pressure
TC Thermocouple
NV : N edle valve
D
R
PR
TC
Bubble meter
NV
FID : Flam Ionization Detector
60
3.3.1.4. Product analysis
The analysis of the product spectrum was divided into two parts. The first part being the
online analysis of the gaseous product stream using two GCs. The second part being that
of the analysis of the liquid (oil and water) and wax products using an offline GC.
For the online analysis, the two GCs employed were respectively equipped with a thermal
conductivity detector (TCD) and a flame ionization detector (FID). The TCD was used to
analyze H2, N2, CO, CH4, CO2 whereas the FID was mainly employed for the analysis of
hydrocarbons, primarily C1-C8.
Prior to the gas product analyses, the two online gas chromatographs (GC) were
calibrated using a gas mixture of a known concentration. The gas mixtures employed
were 20.6 % H2/20.3 % N2/20.3 % CO/19.1 % CH4/19.7 % CO2 (v/v) and 2.5 %
CH4/0.20% C2H4/0.50 % C2H6/10 % CO/5 % CO2/81 % Ar. Syngas (10 % N2/29.6 %
CO/60.2 % H2 ) was also used as a calibration gas for the estimation of the number of
moles of reactants entering the reactor (feed stream) prior to the FT reaction. Typical
traces produced from the calibration and reaction analyses were recorded and plotted
using a DataApex Chromatograph software package known as Clarity (v. 2.5). These
plots are illustrated in Figures 3.15, 3.16, 3.17, 3.18 and 3.19.
61
Figure 3.15 A trace for the calibration gas using the TCD GC
Figure 3.16 A trace for the calibration gas product using the FID GC
CH4
C2H4 C2H6
H2
N2
CO
CH4
CO2
62
Figure 3.17 A trace showing the calibration of the TCD GC using syngas
Figure 3.18 FTS products detected by the TCD GC
H2
N2 CO
H2
CH4 CO2
N2 CO
63
Figure 3.19 FTS products detected by the FID GC
C1
C2-
C8C7
C6
C2=
C3- C4-
C5-
C3=
C4=
C5=
64
3.3.1.5 Mass balance calculations
The calculations used to determine the mass balance are similar to those used by
Duvenhage [3], Mokoena [4], Bahome [5], Phadi [6], and Price [7]. The mass balance
was performed on carbon and oxygen. Mass balance data of 95% to 105% was accepted
as adequate.
The analysis of feed and products in the two gas chromatographs was recorded and
plotted using the Clarity software as explained in the previous section. The areas of the
components were converted to molar composition by calculation.
The reaction steady state was typically reached 24 hours after the beginning of the
reaction. Once this period was reached, the mass balance period was initiated and was
recorded till the end of the experiment. The liquid and the wax products were then
collected separately from the cold trap and hot trap successively and weighed. They were
then analysed using the offline GC. It is to be noted that the oil was separated from water
before analysis. The actual offline analysis involved syringe injection (0.02 ?l) of liquid
(oil) and wax products into the GC.
The outlet flow stream was measured on a daily basis using a bubbler at ambient pressure
and temperature. The feed inlet flow rate to the reactor was determined using N2 gas
contained in the syngas cylinder. The equation used to determine the feed flow rate is
given below:
out
outN
in,N
in Fx
,X
XF
2
2 ??
???
?= (3.1)
where Fin is the total feed flow rate in mol/s, in,N 2X and out ,N2X are mole fractions of
nitrogen in the feed (Syngas) and reactor exit streams respectively and outF is the total
reactor exit stream in mol/s.
65
The number of moles of carbon in the feed stream in the total mass balance period was
calculated by:
in CO,inin c, X t..F N = (3.2)
where in C,N is the moles of carbon in the feed, Fin is the total feed flow rate in mol/s, t is
the total mass balance time and in CO,X is the mole fraction of CO in the feed gas.
Calibration of the components was carried out with a premixed gas of known
composition containing CH4, C2H6, C2H4, CO, CO2, and Ar. The moles product of each
of the component present in the calibration gas was calculated using the following
equation:
t.F . X .
A
A
N outcal c,
cal c,
c
out c, = (3.3)
where cA is the GC integrated area of component c, cal c,A is the area of the component c
in the calibration gas and cal c,X is the mole fraction of the component c in the calibration
gas.
The hydrocarbon product areas were corrected for C2H4 (olefins) and C2H6 (paraffins) by
using the response factors based on those presented by Bahome [5] and Phadi [6]. The
mole fractions of hydrocarbons i HC,X were calculated using the equation below:
cal ,C
cal ,C
i HC,i
i HC, 2
2
X .
A
A . RF
X = (3.4)
where iRF is the response factor for carbon number i, i HC,A is the integrated GC area for
a hydrocarbon with carbon number i, cal ,C2A and cal ,C2X refer to peak area and mole
fraction of the C2 hydrocarbon in the calibration gas [3, 4].
66
The mass response factors for the hydrocarbon with carbon number greater than 15 were
assumed to be one. The mass fractions of these hydrocarbons (i > 15) were thus
determined directly from the GC integrated areas using the following equation:
?= i HC,
i HC,
i A
A
m (3.5)
The product selectivity for hydrocarbons Si was calculated for component xi as follows:
%100x
x
xcomponent mass
S
i
i
i ???
?
???
?= ? (3.6)
The olefin to paraffin ratio x2 was given as:
2
2
2 n xhydrocarbo totalMass
olefin x Mass xratioParaffin Olefin to = (3.7)
Carbon and oxygen mass balances were determined using the information obtained from
the above analysis and calculations:
in CO,
COin vapour CO,solidin CO,out CO,in CO, 2
N -N -N -N N
x100balance Mole %
N
?= (3.8)
The % CO conversion was calculated as:
100x
CO
n contractio Gas x COCO
in
outin ??
???
? ? (3.9)
where the gas contraction was determined from the
out2
in2
N
N calibration
67
The individual rates of reaction for FTS ( FTSr ) and water gas shift WGS ( WGSr ) were
calculated from experimentally obtained quantities as:
2COWGS
rr = (3.10)
2COCOFTS
rrr ?= (3.11)
where rCO2 is the rate of carbon dioxide formation and rCO is rate of carbon monoxide
conversion.
68
References
1. M. Bowker, The Basis and Applications of Heterogeneous Catalysis, Oxford
University Press, New York (1998)
2. A. Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692
3. D.J. Duvenhage, The Preparation, Characterization and Evaluation of Titania
Supported Fe:Co Bimetallic Catalysts for the Hydrogenation of CO, PhD Thesis,
University of the Witwatersrand, Johannesburg (1994)
4. E.M. Mokoena, Synthesis and use of silica materials as support for the Fischer-
Tropsch reaction, PhD Thesis, University of the Witwatersrand, Johannesburg (2005)
5. M.C. Bahome, Synthesis and use of carbon nanotubes as a support for the Fischer-
Tropsch Synthesis, PhD Thesis, University of the Witwatersrand, Johannesburg (2007)
6. T.T. Phadi, Titanates and titania coated titanates as supports in the Fischer-Tropsch
synthesis, MSc Dissertation, University of the Witwatersrand, Johannesburg (2008)
7. J.G. Price, An investigation into novel bimetallic catalysts for use in the Fischer-
Tropsch reaction, PhD Thesis, University of the Witwatersrand, Johannesburg (1994)
69
Chapter 4
Optimisation of the weight loading of copper and potassium promoters in a
precipitated Fe-based Fischer-Tropsch synthesis catalyst
4.1 Introduction
Copper and potassium are classic chemical promoters often used in the iron-based
Fischer-Tropsch synthesis catalyst. They are typical promoters used to prepare FTS
catalysts employed in industrial catalysts. The effects caused by copper and potassium
are well documented and most of these have been described in Chapter 2. The aim of this
study was to optimize the weight loadings of these two promoters and the studies were
carried out systematically. The weight loading range investigated was 1 ? 5 wt. %
4.2 Experimental
All the catalysts were prepared using the precipitation method as outlined in Chapter 3.
The catalysts were characterized using XRF, XPS, TPR, XRD and DRIFTS techniques.
A comprehensive discussion of how these characterisation experiments were carried out
is also given in Chapter 3.
70
4.3 Results and discussion
The optimum weight loading of copper in Fe FTS catalysts will be presented first. The
results obtained on studying potassium will be presented after the copper results.
4.3.1 Optimising the weight loading of copper
4.3.1.1 XRF
The intended weight loadings of Cu and those determined using XRF are displayed in
Table 4.1
Table 4.1 The theoretical and XRF determined Cu loadings
Catalyst composition Theoretical value of Cu
(wt. %)
XRF determined
(wt. %)
1Cu/100Fe 1 1.2
2Cu/100Fe 2 2.2
3Cu/100Fe 3 3.5
4Cu/100Fe 4 4.8
5Cu/100Fe 5 6.0
It is noticed that the intended weight loadings are similar to those determined using XRF.
The maximum error that exists between the theoretical values and the XRF determined
values is 1 wt. %. The catalysts were then characterised using the techniques mentioned
above. All the comparisons were done relative to a catalyst containing only Fe (100Fe).
In other words 100Fe was used as the benchmark catalyst.
71
4.3.1.2 XPS
XPS spectra of the copper loaded catalysts are shown in Fig. 4.1 where the Cu(2p)
spectra and copper Auger spectra are presented. The Cu(2p) spectra show several peaks,
the most intense of which is centered at 934.5 eV, and is assigned to the Cu(2p3/2)
photoelectron line of CuO. The other peaks are ascribed to the satellites of Cu(II) [1-2].
The copper Auger spectra were recorded to aid exact determination of the copper
oxidation state within each sample. From the Cu(LLM) spectra (Fig 4.1a), it is clear that
the copper exists as CuO rather than Cu2O or metallic copper, due to the presence of a
peak having a binding energy at 569.3 eV. For copper metal, the LLM peak would have
shifted to a lower binding energy (ca. 568 eV) [3].
It is noticeable that as the copper loading is increased, the peak at 934.5 eV increases in
intensity. This is an indication that the copper content on the surface is increasing.
Evidence of this assertion is illustrated in Table 4.2 and it is observed that as the atomic
% of Fe (Fe 2p peak) decreases the atomic % of Cu (Cu 2p peak) increases.
Figure 4.1 (a) Cu(LLM) and (b) Cu(2p) spectra for all catalysts
5Cu/100Fe
569.3 eV 934.5 eV(a) (b)
4Cu/100Fe
3Cu/100Fe
2Cu/100Fe
1Cu/100Fe
72
Table 4.2 XPS data for spectra given in Fig 4.1
Catalyst composition
(parts by weight)
Peak identity Binding energy
(eV)
Peak area Atomic %
Fe 2p 711.4 24662 43.7 1Cu/100Fe
Cu 2p 935.4 587 0.58
Fe 2p 711.9 18595 42.0 2Cu/100Fe
Cu 2p 935.4 576 0.72
Fe 2p 711.4 20683 40.8 3Cu/100Fe
Cu 2p 935.4 849 0.93
Fe 2p 711.4 20445 42.7 4Cu/100Fe
Cu 2p 933.9 1156 1.34
Fe 2p 711.4 19159 39.5 5Cu/100Fe
Cu 2p 934.4 1668 1.91
73
4.3.1.3 H2 TPR
The H2 TPR results are shown in Fig. 4.2, Tables 4.3 and 4.4. All TPR profiles show 2
distinct reductions peaks with the last three profiles showing an extra small peak before
the first peak. This peak is ascribed to the reduction of CuO to Cu. The occurrence of this
peak has also been reported in the literature [4, 5]. It has also been reported that the H2
reduction of Fe2O3 occurs via 2 main steps: Fe2O3 ? Fe3O4 ? Fe. These two elementary
reactions are assigned to the first and second peaks in the H2 TPR profiles, respectively
[6-8]. It is noticeable that the addition of Cu shifts the reduction peaks to lower
temperatures. This is also a well known effect and has been widely published [4, 5]. As
the Cu loading is increased the reduction temperature of the first peak is lowered,
demonstrating a linear relationship between copper loading and the iron oxide reduction
temperature. An increase in the copper content generally increases the reduction
temperature of the second peak. The peak is only decreased to a lower temperature for the
catalyst loaded with 5 wt. % of Cu. From these results it is seen that the 5 wt. % loading
of Cu greatly improves the reduction of the iron oxide phase more than the other
loadings.
0 200 400 600 800 1000
100Fe/5Cu
100Fe/4Cu
100Fe/3Cu
100Fe/2Cu
100Fe/1Cu
100FeH
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
Temperature (oC)
Figure 4.2 H2 TPR profiles of all the catalysts
74
Table 4.3 Reduction temperatures for the H2 TPR profiles show in Fig.4.2
Reduction temperature (?C) Catalyst
composition
(parts by weight)
Peak 1 Peak 2 Peak 3
100Fe - 529 798
1Cu/100Fe - 430 815
2Cu/100Fe - 379 815
3Cu/100Fe 336 404 823
4Cu/100Fe 332 396 828
5Cu/100Fe 285 345 789
When carrying out an FTS reaction, reduction is normally carried out for longer than the
time employed when performing an H2 TPR experiment. An isothermal temperature is
employed instead of the changing temperature as is the case for a TPR experiment. TPR
is often employed for relative comparisons and does represent the reduction process used
for the FTS reaction.
In order to be able to determine the amount of Fe reduced prior to reaction, the reduction
conditions normally employed for an FTS reaction were employed in a TPR reaction. The
100Fe catalyst was used for these experiments and the experimental procedure involved
heating the catalyst from room temperature to 350 ?C under the flow of H2 and then
holding the temperature at 350 ?C for 24 hours. The TPR profile obtained is shown in
Fig. 4.3.
75
0 500 1000 1500 2000 2500
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
Time (min)
Figure 4.3 TPR profile of Fe2O3 reduced using the reduction method employed when
carrying out an FTS reaction
It is noticeable from Fig. 4.3 that only one peak is present and this represents the
transformation of Fe2O3 to Fe3O4. However, it is likely that some Fe2O3 could be reduced
to Fe3O4 and then to Fe rapidly for small crystallites when Cu is present. This could also
be possible in the absence of Cu as is with the current situation. Fig. 4.3 also shows that
holding the temperature at 350 ?C for 24 hours does not increase the reducibility of Fe.
As a result only the first reduction peak during TPR, accounts for the reduction of Fe
under the standard reduction procedure used. The %Fe reducibility shown in Table 4.4,
therefore, was calculated only from the first TPR peak shown in Fig. 4.2 and represents
the degree of reducibility of the catalyst prior to reaction. Another way of determining
%Fe reducibility could be to reduce the catalyst at 350 oC for 24 hrs, followed by cooling
76
to room temperature (all under H2) and then performing an H2 TPR experiment thereafter.
But for the purpose of our studies the former method was employed.
Table 4.4 %Fe reducibility as a function of Cu loading for all Cu loaded catalysts
aCatalyst composition
(parts by weight)
%Fe reducibilityb
100Fe 15
1Cu/100Fe 30
2Cu/100Fe 33
3Cu/100Fe 38
4Cu/100Fe 39
5Cu/100Fe 46
aParts by weight
bMaximum error = ? 5%
From Table 4.4 it is observed that varying the loading amount of Cu has some effect on
the %Fe reducibility. The catalyst loaded with 5 wt. % Cu gives the highest %Fe
reducibility.
Dry [9] has stated that the precipitated iron catalyst developed by Ruhrchemie and used
in the fixed-bed reactors at Sasol contains about 5 wt. % Cu. Meanwhile work performed
by O?Brein et al. [10] has shown that reduction of iron oxide using hydrogen is
accelerated with increasing levels of copper promotion (2.6-5.0 atomic % relative to
iron). They argued that the acceleration of the iron oxide reduction (with increasing levels
of copper promotion) is in agreement with more nucleation sites being available with an
increasing amount of copper.
77
Based on these findings, the 5 wt. % loading of Cu can be nominated as the best loading
amount for enhancing the reduction properties of the Fe based catalyst.
4.3.1.4 XRD
The XRD experiments were performed to determine the crystallite size of Fe2O3.
The actual crystallite size was determined using Rietveld refinement and the crystallite
sizes determined are given in Table 4.5.
Table 4.5 The calculated crystallite size of Fe2O3 as a function of Cu loading
aCatalyst
composition
Fe2O3 crystallite
size (nm)
100Fe 33
1Cu/100Fe 44
2Cu/100Fe 46
3Cu/100Fe 48
4Cu/100Fe 46
5Cu/100Fe 45
aParts by weight
It is noticeable that the introduction of even 1 wt. % Cu increases the crystallite size of
Fe2O3. The crystallite size increases irrespective of the loading amount of Cu added and
the crystallite sizes for all the Cu loaded catalysts are comparable to one another.
78
4.3.1.5 CO adsorption measurements using DRIFTS
CO adsorption on the copper loaded catalysts was also performed and the results are
illustrated in Fig 4.4 and Table 4.6. It is noticeable that adsorption of CO on Fe produces
two CO bands at around 2013 and 2033 cm-1. These two bands show that CO is adsorbed
on Fe in a linear fashion [11, 12]. The CO adsorption measurements indicate that the
presence of Cu causes a red shift of the 2013.7 cm-1 peak to 2011.7 cm-1. This red shift by
Cu indicates that Cu enhances the backdonation ability of Fe. This means that in the
presence of Cu, the ability of Fe to transfer electrons via backdonation into the
antibonding orbitals (2?*) of carbon is enhanced [13, 14]. This strengthens the Fe-C bond
making it possible to increase the hydrocarbon chain during the FTS reaction. This
postulation is further confirmed by the number increase of the calculated CH2/CH3 ratio
(see below). In fact for all the copper loaded catalysts the 2013 cm-1 band shifts to lower
wavenumbers. The results are illustrated in Table 4.6.
Figure 4.4 Comparing CO absorption spectra of Cu promoted catalysts to the
unpromoted Fe catalyst
2070 2060 2050 2040 2030 2020 2010 2000
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
K
ub
el
ka
-M
un
k
un
its
2033
2013
5Cu/100Fe
4Cu/100Fe
3Cu/100Fe
2Cu/100Fe
1Cu/100Fe
100Fe
Wavenumber (cm-1)
79
Table 4.6 Position of IR absorption band as a function of Cu loading
Catalyst Peak wavenumber (cm-1)
100Fe 2014
1Cu/100Fe 2012
2Cu/100Fe 2012
3Cu/100Fe 2012
4Cu/100Fe 2012
5Cu/100Fe 2012
80
4.3.1.6 In situ CO hydrogenation using DRIFTS
In situ CO hydrogenation reactions were performed using the DRIFTS reactor. Only the
part of the spectrum that monitors the production of C-H species (2750 ? 3100 cm-1) was
assessed. This C-H species gives an indication of the hydrocarbon molecules produced
during the reaction.
To estimate the average carbon chain length of the hydrocarbon molecules produced after
5 hours of reaction, the ratio of CH2/CH3 species was calculated using the following
formula:
1
)( 2
?
speciesCHArea ?? /
2
)( 3
?
speciesCHArea ??
where
1) Area of -CH2- species is the area of the peak at 2925-2930 cm-1 representing the
asymmetric stretch of CH2 species
2) Area of ?CH3 species is the area of the peak at 2955-2960 cm-1 representing the
asymmetric stretch of CH3 species
3) ?1 is the molar extinction coefficient of the CH2 species (75 mole-1.l.cm-1) [15]
4) ?2 is the molar extinction coefficient of the CH3 species (70 mole-1.l.cm-1) [15]
The calculated ratios are given in Table 4.7. It is evident that addition of copper leads to
an increase in the average chain length of the hydrocarbons.
81
Table 4.7 Calculated ratios of CH2/CH3 bands for all catalysts
Catalyst Ratio of CH2/CH3
100Fe 1
100Fe/1Cu 7
100Fe/2Cu 6
100Fe/3Cu 7
100Fe/4Cu 6
100Fe/5Cu 6
As stipulated in the previous section the addition of Cu leads to an increase in the
CH2/CH3 ratio indicating an increase in the chain length of the hydrocarbon molecules.
Further when the loading of Cu is varied, the CH2/CH3 ratio is not drastically changed,
indicating that changing the Cu loading does not significantly change the FTS product
spectrum. This could mean that the FTS product selectivity is not affected by varying the
loading of Cu. This again could mean Cu plays a small role in changing the FTS product
selectivity. Zhang et al. [16] have also reported that copper plays only a small role in FTS
product selectivity.
It was also important to make sure that deductions made on the results presented above
were real and did not necessarily constitute a scenario of a one point deduction. To verify
this postulation, two unpromoted 100Fe samples were prepared and DRIFTS experiments
(i.e. calculation of CH2/CH3) were done on both samples. The results obtained came out
to be similar (not presented here). Clearly, confirming that deductions made above were
true. Unfortunately the XRD experiment was performed only on one sample (as already
discussed above). But it is logical to conclude that if the DRIFTS results came out to be
the same for both prepared samples surely the XRD results of the two samples would
give similar results.
82
4.3.2 Optimising the weight loading of potassium
4.3.2.1 H2 TPR
The H2 TPR results are presented in Fig. 4.5, Tables 4.8 and 4.9. All the catalysts show
the two dominant peaks for the reduction of Fe2O3. These peaks represent the two step
reduction of Fe2O3 into metallic Fe as explained earlier. It is noticeable that catalysts
loaded with a loading amount of ? 2 wt. % K2O show 3 peaks. The first peak in all these
profiles could reflect the reduction of easily reducible iron oxide crystallites. This still
reflects the transformation of Fe2O3 into Fe3O4.
0 200 400 600 800 1000
100Fe/3K
2
O
100Fe/2K
2
O
100Fe
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
100Fe/5K
2
O
100Fe/4K
2
O
100Fe/1K
2
O
Temperature (oC)
Figure 4.5 H2 TPR profiles of all the catalysts
83
Table 4.8 Reduction temperatures for the H2 TPR profiles show in Fig. 4.5
Reduction temperature (?C) Catalyst
composition
(parts by weight)
Peak 1 Peak 2 Peak 3
100Fe - 537 802
1K2O/100Fe - 422 721
2K2O /100Fe 358 469 747
3K2O /100Fe 426 537 785
4K2O/100Fe 417 520 764
5K2O/100Fe 486 597 832
From Table 4.8 it is observed that K2O loading up to 2 wt. % lowers the reduction
temperature of peak 2, thereafter the reduction temperatures are shifted to higher
temperatures.
The %Fe reducibility was also determined in the same manner as it was done for the
Cu/Fe loaded catalysts. It is clear that as the loading of potassium is increased the %Fe
reducibility is decreased. This could be attributed to potassium suppressing the ability of
Fe to adsorb H2 [17, 19]. The catalyst loaded with 2 wt. % K2O gives the highest %Fe
reducibility.
84
Table 4.9 %Fe reducibility as a function of K2O loading for all K2O loaded catalysts
aCatalyst
composition
%Fe
reducibilityb
100Fe 15
1K2O/100Fe 64
2K2O/100Fe 70
3K2O/100Fe 55
4K2O/100Fe 54
5K2O/100Fe 41
aParts by weight
bMaximum error = ? 5%
85
4.3.2.2 XRD
XRD was employed to determine the crystallite size of Fe2O3. The crystallite size was
determined in the same way as it was done for the Cu loaded catalysts. Table 4.9 below
illustrates the calculated crystallite size of Fe2O3 for all the K2O loaded catalysts.
Table 4.10 The calculated crystallite size of Fe2O3 as a function of K2O loading
aCatalyst
composition
Fe2O3 crystallite
size (nm)
100Fe 33
1K2O/100Fe 51
2K2O/100Fe 52
3K2O/100Fe 52
4K2O/100Fe 52
5K2O/100Fe 47
aParts by weight
It is seen that the presence of K2O increases the crystallite size of Fe2O3 but increasing
the K2O loading from 1-4 wt. % does not significantly change the crystallite size of
Fe2O3.
86
4.3.2.3 CO adsorption measurements using DRIFTS
The CO adsorption results are presented in Fig. 4.6 and Table 4.10. Again it is noticeable
that all the catalysts have two bands showcasing CO adsorbed linearly on Fe. The
adsorbed CO bands on the benchmark catalyst are around 2013 and 2033 cm-1 and no
other peaks could be identified.
2070 2060 2050 2040 2030 2020 2010 2000
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
100Fe/5K
2
O
100Fe/4K
2
O
100Fe/3K
2
O
100Fe/2K
2
O
100Fe/1K
2
O
100FeK
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 4.6 Comparing CO absorption spectra of K2O promoted catalysts to the
unpromoted Fe catalyst
These bands are shifted to lower wavenumbers for all the K2O loaded catalysts as
illustrated in Table 4.10. This red shift illustrates that the Fe-C bond is strengthened and
that potassium enhances CO adsorption on Fe. It is well known that potassium increases
the CO adsorption ability of Fe [18, 19].
87
Table 4.11 Peak shifts of peak at wavenumber region 2012-2015 cm-1 as a function of
K2O loading
Catalyst Peak wavenumber (cm-1)
Fe 2014
1K2O/100Fe 2012
2K2O/100Fe 2008
3K2O/100Fe 2006
4K2O/100Fe 2008
5K2O/100Fe 2012
The catalyst loaded with 3 wt. % K2O shifts the 2013 cm-1 peak to the lowest
wavenumber (2006.0 cm-1). This peak is then shifted to ca. 2007.9 cm-1 at a loading of 4
wt. % K2O and to 2011.7 cm-1 for the 5 wt. % loading. It is clear that weight loadings
above 3 wt. % do not significantly enhance the adsorption of CO.
88
4.3.2.4 In situ CO hydrogenation using DRIFTS
The CH2/CH3 ratio was also determined for all the catalysts. The ratio was determined as
explained in section 4.2.1.6. All the ?in situ? CO hydrogenation reactions were performed
as explained in Chapter 3.
Table 4.12 Estimation of the CH2/CH3 ratio as a function of K2O loading
Catalyst Ratio of CH2/CH3
100Fe 1
1K2O/100Fe 2
2K2O/100Fe 8
3K2O/100Fe 8
4K2O/100Fe 7
5K2O/100Fe 7
Increasing the loading of K2O increases the CH2/CH3 ratio. It is also noticeable that 2 and
3 wt. % loadings of K2O give the highest CH2/CH3 ratio. These results are consistent with
the CO adsorption results. It can also be noted that an increased CH2/CH3 ratio on
addition of K2O promotion reflects that the hydrocarbon chain length is increased. This is
consistent with literature reports since it is well known that K2O promotes chain growth
and shifts selectivity to longer chained hydrocarbons [19-23].
It is apparent that when potassium is added in moderation to Fe-based FTS catalysts, it
enhances its characteristics (e.g. FTS activity is enhanced and selectivity to methane
lowered). This is because when potassium containing catalysts are heated potassium
moves to the top (the surface) of the catalyst [24] and has a direct influence on the active
sites of the catalyst. Therefore if a high loading of potassium is used, this may be
detrimental to the catalyst as more of it will move to the surface and block some of the
catalyst active sites leading to a lower FTS activity.
89
Therefore the FTS activity either increases [22, 25] or passes through a maximum as a
function of potassium loading [23], and potassium either has no effect on the activity for
FTS [10] or suppresses it [23, 24].
Not much systematic work has been carried out on the effect of the level of potassium
loading on Fe-based FTS catalysts. Work done by O?Brein et al. [10] in 1997, showed
that a high potassium loading is required when the FTS reaction temperature is decreased
because it becomes harder to dissociate the C-O bond. They found that the optimum
potassium promotion was 4-5 atomic % relative to iron. They further found that
potassium promotion increased wax selectivity.
In these studies we have noticed that 2 ? 3 wt. % loading amount of K2O significantly
enhanced the chemical properties of the precipitated Fe-based FTS catalyst. It improved
the reduction properties (TPR results) and the CO adsorption ability of the catalyst. It
also increased the CH2/CH3 ratio which can be used as a qualitative way of measuring the
average hydrocarbon chain length.
90
4.4 Conclusion
The aim of this study was to optimize the weight loadings of Cu and K2O. The weight
loading range investigated was 1-5 wt. % for both promoters. Various characterization
techniques were used to assess the effects caused by all the promoter loadings. It was
found that the 5 wt. % loading of Cu was the optimum loading amount for the copper
loaded catalysts, because this loading significantly enhanced the reduction properties of
the precipitated Fe-based FTS catalyst.
For the K2O loading, the wt. % loading range of 2-3 wt. % K2O significantly improved
the reduction properties as well as the CO adsorption ability of the precipitated Fe-based
FTS catalyst and due to the 2 wt. % K2O loading giving the best %Fe reducibility, this
loading was chosen as the optimal loading of K2O.
As a result the 5 wt. % loading of Cu and 2 wt. % loading of K2O were used to prepare
all Cu/K2O containing catalysts used in this thesis. The work showcasing the use of these
optimum loading amounts of Cu and K2O is illustrated in Chapter 6 and Chapter 8.
91
References
[1] F.M. Capece, V. Dicastro, C. Furlani, G. Mattogno, C. Fragale, M. Gargano, M.
Rossi, J. Electron Spec. Relat. Phenom. 27 (1982) 119.
[2] B. Peplinske, W.E.S. Unger, I. Grohmann Appl. Surf. Sci. 62 (1992)
[3] Z. Wang, Q. Liu, J. Yu, T. Wu, G. Wang, Appl. Catal. A: Gen. 239 (2003) 87
[4] D.B. Bukur, K. Okabe, M.P. Rosynek, C. P Li, D. J Wang, K. R. P. M. Rao,
G. P. Huffman, J. Catal. 155 (1995) 353
[5] Y. Jin, A. K. Datye, J. Catal. 196 (2000) 8
[6] I.S.C Hughes, J.O.H. Newman, G.C. Bond, Appl. Catal. 30 (1987) 303
[7] C.N. Mbileni, Applications of mesostructured carbonaceous materials as supports for
Fischer-Tropsch metal catalyst, PhD Thesis, University of the Witwatersrand,
Johannesburg, 2006
[8] M. Luo, R.J. O?Brein, S. Bao, B.H. Davis, Appl. Catal. A: Gen. 239 (2003) 111
[9] M.E. Dry, The Fischer?Tropsch synthesis, in: J.R. Anderson, M. Boudart (Eds.),
Catalysis Science and Technology 1, Springer-Verlag, New York, 1981, 159
[10] R. J. O?Brein, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal. Today
36 (1997) 325
[11] M.J. Heal, E.C. Leisegang, R.G. Torrington, J. Catal. 51 (1978) 314
[12] G. Bian, A. Oonuki, Y. Kobayashi, N. Koizumi, M. Yamada, Appl. Catal. A: Gen.
219 (2001) 13
[13] F. Morales, E. de Smit, F.M.F. de Groot, T. Visser, B.M. Weckhuysen, J. Catal. 246
(2007) 91
[14] G. Blyholder, L.D. Neff, J. Chem. Phys. 66 (1962) 1464
[15] K. Nakanishi, Infrared Absorption spectroscopy ? PRACTICAL - , Nankodo
Company Limited, Japan, 1964
[16] C. Zhang, Y. Yang, Z. Tao, T. Li, H. Wan, H. Xiang, Y. Li, Acta Physico-Chimica
Sinica 22 (2006) 1310
[17] G. Zhao, C. Zhang S. Qin, H. Xiang, Y. Li, J. Mol. Catal. A: Chemical 286 (2008)
137
92
[18] H.J. Wan, B.S. Wu, Z.C. Tao, T.Z. Li, X. An, H.W. Xiang, Y.W. Li, J. Mol. Catal.
A: Chemical 260 (2006) 255
[19] N. Lohitharn, J.G. Goodwin Jr., J. Catal. 260 (2008) 7
[20] R.B. Anderson, The Fischer-Tropsch Synthesis, Academic Press, Orlando, 1984
[21] Y.Yang, H.W. Xiang, Y.Y. Xu, L. Bai, Y.W. Li, Appl. Catal. A Gen. 266 (2004) 181
[22] D.B. Bukur, D. Mukesh, S.A. Patel, Ind. Eng. Chem. Res. 29 (1990) 194
[23] D.G. Miller, M. Moskovits, J. Phys. Chem. 92 (1988) 6081
[24] R.A. Dictor, A.T. Bell. J. Catal. 97 (1986) 121
[25] R.B. Anderson, B. Sekigman, J.F. Schulz, M.A. Elliot, Ind. Eng. Chem. 44 (1952)
391
93
Chapter 5
Effect of SiO2 content on an unpromoted Fe-based Fischer-Tropsch
synthesis catalyst
5.1 Introduction
Many studies have been performed employing SiO2 as part of the ingredient for preparing
precipitated iron-based catalysts for the FTS. In most instances it has been employed as a
support (structure promoter) or as a major component of the catalyst [1].
The major reason for using it as a support or structural promoter is that it lowers the
deactivation rate of the catalyst especially in slurry phase reactors [2]. Deactivation of
catalysts particularly those without a binder, support or structural promoter occurs via
attrition. Attrition is the breakage of the catalyst leading to the formation of very small
particles. These small particles are readily lost as ?fines?.
Previous studies with supported iron catalysts have been reported [3-10], but there are
very few studies that have looked at the effect of SiO2 as a chemical promoter. SiO2 has
been shown to possess chemical promotional abilities [11].
Recent work by Zhang et al. [11] has highlighted that the introduction of 20 wt. % SiO2
into a precipitated iron-based FTS catalyst results in improved light hydrocarbon
selectivity. A deleterious effect was observed on introduction of 20 wt. % SiO2 as this
high loading lowered the FTS activity. They attributed this effect to the strong interaction
that exists between iron and silica (iron-silica interaction) thus rendering some of the iron
to be inactive for FTS. Work reported by other researchers [12-16] also confirms the
latter postulation made by Zhang et al., but most of this work was performed on multi-
component systems.
94
Surprisingly not much work has been carried out to evaluate the effect of SiO2 as a sole
chemical promoter for the precipitated Fe-based FTS catalysts. Employing SiO2 in a
multi-component system makes it difficult to establish its chemical effect. Our aim then
was to evaluate the effect of SiO2 as a chemical promoter for Fe-based FTS catalyst in the
absence of other promoters. We were interested in studying the effect of the SiO2 content
(5 wt. % to 25 wt. %) on a precipitated iron-based Fischer-Tropsch catalyst.
95
5.2 Experimental
All the catalysts were prepared in the same manner as outlined in Chapter 3. The catalysts
were characterised using N2 physisorption, XRD, XPS, TPR and DRIFTS. A
comprehensive discussion on how the characterisation experiments were performed is
also outlined in Chapter 3.
5.3 Results and discussion
5.3.1 Textural and structural properties of the catalysts
The textural properties of the catalysts were determined using N2 physisorption as
illustrated in Table 5.1. The structural properties were examined using XRD.
Table 5.1 The composition and textural properties of the calcined catalysts
Catalyst composition (parts
by weight)
BET Surface areab
(m2/g)
Pore volumeb
(cm3/g)
100Fe 22.8 0.086
100Fe/5SiO2 72.7 0.13
100Fe/10SiO2 131 0.16
100Fe/20SiO2 196 0.22
100Fe/25SiO2 229 0.24
bMaximum error = ? 2%
As expected the addition of SiO2 increased the surface area of the precipitated catalyst.
The surface area as well as the pore volume increased with increasing SiO2 content. This
is consistent with work carried out by other researchers [17, 18].
96
Several authors [14, 19] have gone on to mention that SiO2 provides a more dispersed
rigid matrix, which helps to prevent the catalyst from a fast pore collapse and stabilizes
the small iron oxide crystallites from sintering. In other words, SiO2 favours a high
dispersion of Fe2O3. This means that increasing the SiO2 content favours a formation of a
porous structure and the high dispersion of Fe2O3. Consequently, the average crystallite
size of Fe2O3 was decreased as shown by XRD results (Table 5.2). The crystallite size
was calculated from Rietveld refinement of the XRD patterns of the catalysts.
Table 5.2 The calculated crystallite size of Fe2O3 as a function of SiO2 loading
aCatalyst composition Fe2O3 crystallite size
(nm)
100Fe 32.8
5SiO2/100Fe 30.5
10SiO2/100Fe 23.4
20SiO2/100Fe 20.5
25SiO2/100Fe 12.5
aParts by weight
5.3.2 Reduction and carburization behaviour of the catalysts
H2 and CO TPR techniques were employed to investigate the effect of the silica content
on the reduction and carburization behaviour of the catalysts. The TPR profiles of the H2
absorption and the corresponding quantitative results are presented in Fig. 5.1 and Table
5.3.
97
0 200 400 600 800 1000
100Fe
25SiO
2
/100Fe
20SiO
2
/100Fe
10SiO
2
/100Fe
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
Temperature (oC)
5SiO
2
/100Fe
Figure 5.1 H2 TPR profiles of the catalysts
As shown in Fig. 5.1, the expected two stage reduction process of Fe2O3 occurs but as the
loading of SiO2 is increased the reduction peaks are shifted to higher temperatures. In fact
it becomes difficult to distinguish between the two reduction peaks when SiO2 loading is
increased above 10 wt. %. This may be attributed to increased Fe-SiO2 interactions [20].
It has also been reported that adding SiO2 to iron-based FTS catalyst restrains its
reduction [11] and this is clearly indicated by Fig. 5.1 and Table 5.3.
98
Table 5.3 Peak maxima of H2 TPR profiles
Peak Maximum
(?C)
aCatalyst
composition
Peak 1 Peak 2
100Fe 537 793
5SiO2/100Fe 561 845
10SiO2/100Fe 585 845
20SiO2/100Fe 630 -
25SiO2/100Fe 633 -
aParts by weight
CO TPR results are illustrated by Fig. 5.2. It is known that CO also reduces Fe2O3 via a
two step process into iron carbides [21-23].
3Fe2O3 + CO ? 2Fe3O4 + CO2 (5.1)
5Fe3O4 + 32CO ? 3Fe5C2 + 26CO2 (5.2)
The catalyst with no SiO2 shows the two peaks associated with FeOx reduction to FeCx.
As soon as SiO2 is added, three peaks emerge. The first peak is attributed to the reduction
of easily reducible iron oxide crystallites. The second peak is ascribed to the reduction of
iron oxide via Eq. 5.1 and the third peak represents the carburization of the crystallites
(Eq. 5.2). It is noticeable that as the SiO2 content is increased the first peak increases in
intensity and broadens (Fig 5.2 and Table 5.5). This would indicate an increase in the
number of easily reducible iron oxide crystallites. This is possible since increasing the
SiO2 content increases the surface area (N2 physisorption results) which in turn improves
the dispersion of the crystallites.
99
0 200 400 600 800
25SiO
2
/100Fe
20SiO
2
/100Fe
10SiO
2
/100Fe
5SiO
2
/100Fe
100Fe
Temperature (oC)
C
O
C
on
su
m
pt
io
n
(m
m
ol
C
O
/m
ol
F
e)
Figure 5.2 CO TPR profiles of catalysts
It is also noticed that the first peak (for SiO2 loaded catalysts) shifts to higher
temperatures as the SiO2 loading is increased. This is attributed to increasing Fe-SiO2
interactions [17]. It is also evident that increasing the SiO2 loading shifts all three peaks
to higher temperatures (Table 5.4) and decreases the areas of all three peaks (Table 5.5).
Again this effect is attributed to the increased Fe-SiO2 interaction. It is therefore
concluded that both the Fe-SiO2 interaction and the iron dispersion affect the reduction
and carburization behaviour of the catalyst [11, 18].
100
Table 5.4 Reduction temperatures for peaks of CO TPR profiles
Peak Maximum
(?C)
aCatalyst
composition
Peak 1 Peak 2 Peak 3
100Fe - 266 536
5SiO2/100Fe 106 285 567
10SiO2/100Fe 121 304 609
20SiO2/100Fe 133 322 640
25SiO2/100Fe 186 338 640
aParts by weight
Table 5.5 Areas for peaks in Figure 5.2
Peak Area
aCatalyst
composition
Peak 1 Peak 2 Peak 3
100Fe - 3732 5813
5SiO2/100Fe 487 2868 2868
10SiO2/100Fe 1521 2289 2033
20SiO2/100Fe 1994 1954 1868
25SiO2/100Fe 3378 1339 1327
101
5.3.3 Surface analysis of the catalysts
The surfaces of the catalysts were studied using XPS. The regions that were looked at
were the core levels of oxygen, iron and silicon, as well as C(1s) spectral regions. Survey
spectra were collected for each sample and an example of this is shown in Fig. 5.3.
Figure 5.3 Survey spectrum showing elements on the surface of the 5SiO2/100Fe
calcined catalyst
In all the survey spectra, a peak having a binding energy of 284.7 eV (Fig. 5.3) was
identified. This is assigned to adventitious carbon [37-39]. This peak was used as a
reference peak for the analysis. It was also used for correction of charge compensation on
the surfaces of all catalysts analysed.
102
Common to all spectra, the carbon peak at 284.7 eV exhibited a shoulder at 288.2 eV.
The former peak is attributed to adventitious carbon, whilst the latter is ascribed to
adsorbed carbonate from reaction with CO2 in the atmosphere. A peak at 529.8 eV with a
shoulder at ca. 531 eV was identified in all the survey spectra and is clearly illustrated in
Fig. 5.3. This peak is attributed to oxygen [26]. The shoulder at ca. 531 eV is
characteristic of a more O?- like oxygen state and is probably attributable to adsorbed
surface hydroxyl groups.
The Fe (2p) spectra region (Fig. 5.4) showed a peak with binding energy in the region of
711 eV, characteristic of Fe2O3 [25], in addition, the characteristic satellite for Fe(III) is
clearly visible at ca. 719.5 eV.
103
Figure 5.4 Narrow region spectrum of Fe (2p) peak for all catalysts
5 SiO2/100Fe
10 SiO2/100Fe
20 SiO2/100Fe
25 SiO2/100Fe
104
Fig. 5.5 shows the narrow region spectrum of the Si (2p) peak. It is observed that as the
weight loading of SiO2 is increased the Si (2p) peak at ~ 109.3 eV increases in size. To
get a qualitative measure of the increase in size of the Si (2p) peak. The ratio of Si (2p)
peak/Fe (2p) peak was calculated using the areas of the peaks. The results are shown in
Table 5.6. It is noticeable from Table 5.6 that as the SiO2 content is increased the ratio of
the Si (2p) peak/Fe(2p) peak increases. This is indicative of the fact that more SiO2 goes
to the surface as its loading amount is increased. This would explain why it was
extremely difficult to reduce and carburize the Fe especially at weight loadings above 10
wt. % .
Figure 5.5 Narrow region spectrum of Si (2p) peak for all catalysts
10 SiO2/100Fe
20 SiO2/100Fe
25 SiO2/100Fe
Si 2p
5 SiO2/100Fe
105
Table 5.6 Fe (2p) and Si (2p) peak areas for all catalysts
Area of peak aCatalyst composition
Fe (2p) Si (2p)
Ratio of Si (2p)
peak/ Fe (2p) peak
5SiO2/100Fe 20723 474 0.023
10SiO2/100Fe 19177 625 0.033
20SiO2/100Fe 16818 871 0.052
25SiO2/100Fe 14191 917 0.065
The oxygen core level spectra (Fig. 5.6), shows a notable increase in the component at
ca. 532 eV as the SiO2 content of the catalyst increases. This high O (1s) binding energy
is characteristic of SiOx (x=1.8 to 2) [27-30].
106
Figure 5.6 Oxygen core level spectra for a) 5SiO2/100Fe b) 10SiO2/100Fe c) 20
SiO2/100Fe and d) 25SiO2/100Fe
Increasing SiOx Content
d
c
b
a
107
It is clear that as the SiO2 loading is increased. There is a notable increase in the amount
of SiO2 that goes to the surface of the catalyst. This leads to increased Fe-SiO2
interactions resulting in the reduction and carburization of the catalyst to be suppressed as
confirmed by the H2 and CO TPR results.
5.3.4 Adsorption properties of the catalysts
The adsorption properties of the catalysts were studied using CO as a probe molecule.
The adsorbed CO species were monitored using DRIFTS. The spectra obtained for
differently SiO2 loaded samples after 30 minutes of adsorption are shown in Fig. 5.7.
2040 2020 2000 1980 1960
0.0
0.4
0.8
1.2
1.6
20
13
100Fe
25SiO
2
/100Fe
20SiO
2
/100Fe
10SiO
2
/100Fe
5SiO
2
/100Fe
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 5.7 CO adsorption spectra of all the catalysts (P = 2 bar, T = 25 ?C, CO flow rate
= 12 ml/min)
108
A band at 2013 cm-1 could be identified for all the catalysts. It is noticeable that as the
SiO2 loading is increased the intensity of this band is decreased (Fig 5.7). This indicates
that fewer iron surface sites are available for binding CO, most likely due to masking by
SiO2 species. This could be assumed to be the case since XPS results showed that more of
the SiO2 goes to the surface as its loading amount is increased.
It is also evident from Fig. 5.7 that the addition of SiO2 shifts the band maxima to lower
wavenumbers and this shift is independent of the SiO2 loading (Table 5.5).
Table 5.7 Band maxima in the wavenumber region 2012-2015 cm-1 as a function of SiO2
loading
Catalyst Peak maxima
(cm-1)
100Fe 2014
5SiO2/100Fe 2012
10SiO2/100Fe 2012
15SiO2/100Fe 2012
20SiO2/100Fe 2012
25SiO2/100Fe 2012
It is possible that the introduction of SiO2 increases the dispersion of the Fe2O3
crystallites (as postulated earlier). The smaller particles are more easily reduced into the
iron metal phase resulting in an increased d-electron density. This in turn results in a
strong Fe-C bond.
The generally accepted explanation of Fe-C bond formation of the Fe-C-O system is
based on a molecular orbital model of the CO molecule. A C ? Fe ? bond would be
ineffective, being opposed by electrostatic repulsions arising from electron transfer.
An Fe ? C ? bond would be similarly ineffective. However, taken together, the two
109
electron transfer act in opposite directions and tend to cancel each other; Relatively
strong synergetic bond processes results. The Fe ? C component arises from the
availability of d-electron density from the metal which occupies an empty CO
antibonding orbital. Thus as the Fe-C bond strength increases, through greater availability
of d-electron density, the C-O bond strength diminishes as shown by lower C-O vibration
frequencies [31].
These results may also suggest that the silica species tightly interact with the surface iron
species and promote iron oxide reduction to form the fine metallic iron clusters as
suggested earlier with the H2-TPR results.
110
5.3.5 FTS performance
The FTS performance of the catalysts was monitored ?in-situ? using DRIFTS. Figure 5.8
below shows the C-H region of the DRIFTS spectra for all catalysts after 5 hours of
reaction. Infrared absorption bands located at 2855 cm-1 and 2927 cm-1 can be assigned to
symmetric and asymmetric stretching vibrations of CH2 groups, while the 2960 cm-1 band
is usually assumed to arise from the asymmetric stretch of CH3 groups [32]. The peak at
3016 cm-1 is assigned to gaseous CH4 [33].
Figure 5.8 DRIFTS spectra of all catalysts showing the C-H region after 5 hours of the
FTS reaction (Reduction conditions: H2/CO = 2, P = 2 Bar, T = 350 ?C, H2/CO flow rate
= 12 ml/min, t = 1 h; FTS reaction conditions: H2/CO = 2, P = 10 Bar, T = 275 ?C, H2/CO
flow rate = 12 ml/min, t = 5h)
3100 3000 2900 2800
3016
2960
2855
2927
25SiO
2
/100Fe
20SiO
2
/100Fe
10SiO
2
/100Fe
5SiO
2
/100Fe
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
111
It is noticeable that as the SiO2 loading is increased the intensity of C-H peaks is lowered
indicating a decline in the FTS activity. These results are in agreement with H2 and CO
TPR results. The decrease in activity is attributed to the Fe-SiO2 interaction as described
earlier. This observation has been reported previously in literature [11].
It is also well known that iron carbides are the main active phases required for FTS
reactions [34, 35]. CO TPR results suggest that increasing SiO2 loading suppresses
carburization of the precipitated Fe-based catalyst. This may mean that during syngas
reduction the carburization is suppressed leading to a lower content of iron carbides. This
in turn, results in lower FTS activity for catalysts with high SiO2 loadings.
We also estimated the average chain length of the products produced in the FTS reaction.
The estimation of the average chain distribution was done by calculating the CH2/CH3
ratio using the equation illustrated below:
1
)( 2
?
speciesCHArea ?? /
2
)( 3
?
speciesCHArea ??
where
1) Area of -CH2- species is the peak at 2925-2930 cm-1 which represents the
asymmetric stretch of CH2 species
2) Area of ?CH3 species is the peak at 2955-2960 cm-1 which represents the
asymmetric stretch of CH3 species
3) ?1 is the molar extinction coefficient of CH2 species (75 mole-1.l.cm-1) [36]
4) ?2 is the molar extinction coefficient of CH3 species (70 mole-1.l.cm-1) [36]
It is evident that as the SiO2 loading is increased the CH2/CH3 ratio is decreased (Table
5.8). This equates to formation of light weight (short chain) hydrocarbons. This is in
excellent agreement with work reported by other researchers who have performed full
FTS reactor studies [2, 11, 18].
112
Table 5.8 Estimation of the CH2/CH3 ratio as a function of SiO2 loading
Catalyst Ratio of CH2/CH3
100Fe 1
5SiO2/100Fe 4
10SiO2/100Fe 4
15SiO2/100Fe 3
20SiO2/100Fe 1
25SiO2/100Fe -
Work by Zhang et al. [11] suggests that adding SiO2 to a precipitated Fe-based FTS
catalyst results in enhanced high selectivity towards low weight hydrocarbons and light
olefins. They attribute this to the Fe-SiO2 interaction and they believe it inhibits chain
growth and secondary hydrogenation reactions.
The precipitated Fe-based FTS catalyst is also known to be reactive for the Water-Gas
Shift (WGS) reaction (Eq. 5.4), where H2O produced from the FTS reaction (Eq. 5.3)
reacts with CO to produce H2 and CO2.
FTS reaction:
CO + 2H2 ? -CH2- + H2O (5.3)
WGS reaction:
CO + H2O ? CO2 + H2 (5.4)
The 2250-2450 cm-1 region of the DRIFTS spectrum can be assigned to gaseous CO2. We
notice that as the SiO2 loading is increased these peaks diminish (Fig. 5.9).
113
2450 2400 2350 2300 2250
0
1
2
3
4
5
6
7
8
2363
2335
25SiO
2
/100Fe
20SiO
2
/100Fe
10SiO
2
/100Fe
5SiO
2
/100Fe
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 5.9 CO2 produced as a function of the SiO2 content
It is also important to note that CO2 can also be produced from the Boudouard reaction
(Eq. 5.5), together with the carbon that deposits on the surface of the catalyst.
Boudouard reaction:
2CO ? CO2 + C (5.5)
Even if the CO2 was produced via this reaction clearly the addition of SiO2 reduces the
formation of CO2. Therefore one would assume that reaction 5.5 is also not enhanced
when the SiO2 content is increased.
114
5.4 Conclusion
The effect of the SiO2 content on a precipitated Fe-based Fischer Tropsch synthesis
catalyst was investigated by comparing the textural, structural, reduction, carburization
properties as well as the FTS performances. Increasing the SiO2 content increased the
surface area of the catalyst which improved the dispersion of the iron oxide crystallites,
resulting in a decrease in the average size of the iron oxide crystallites.
But a decrease in the average size of the iron oxide crystallites strengthened the Fe-SiO2
interaction, XPS surface analysis confirmed that as the SiO2 loading was increased, more
of the SiO2 stayed on the surface allowing the Fe-SiO2 interaction to be enhanced.
This resulted in the reduction and carburization ability of the catalyst to be suppressed.
This affected the FTS performance of the catalyst and lowered its activity. In general, the
reduction and carburization behaviour reflects the activation capability of catalyst. Hence
the effect of the Fe-SiO2 interaction on the reduction/carburization behaviour, directly
affects the FTS performance of the precipitated Fe-based catalyst. It therefore appears
that low loadings of SiO2 are required to induce chemical activation.
115
References
[1] V.U.S. Rao, G.J. Steigel, G.J. Cinquegrane, R.D. Srivastava, Fuel Proc. Technol. 30
(1992) 83
[2] D.B. Bukur, C. Sivaraj, Appl. Catal. A: Gen. 231 (2002) 201
[3] M.E. Dry, in Catalysis Science and Technology 1, J.R. Anderson, M. Boudart (Eds.),
Springer, New York, 1981, p. 160
[4] R.J. O?Brien, L. Xu, S. Bao, A. Raje, B.H. Davis, Appl. Catal. 196 (2000) 173
[5] L. Guczi, Catal. Rev. Sci. Eng. 23 (1981) 329
[6] H.J. Jung, P.L. Walker, M.A. Vannice, J. Catal. 75 (1982) 416.
[7] R.B. Anderson, The Fischer?Tropsch Synthesis, Academic Press, New York, 1984
[8] C.H. Bartholomew, Stud. Surf. Sci. Catal. 64 (1990) 159
[9] G.B. McVicker, M.A. Vannice, J. Catal. 63 (1980) 25
[10] J.L. Rankins, C.H. Bartholomew, J. Catal. 100 (1986) 533
[11] C.H. Zhang, H.J. Wan, Y. Yang, H.W. Xiang, Y.W. Li, Catal. Comm. 7 (2006) 733
[12] Y. Yang, H.W. Xiang, L. Tian, H. Wang, C.H. Zhang, Z.C. Tao, Y.Y. Xu, B. Zhong,
Y.W. Li, Appl. Catal. A: Gen. 284 (2005) 105
[13] N. Egiebor, W.C. Cooper, Can. J. Chem. Eng. 63 (1985) 81
[14] D.B. Bukur, X. Lang, D. Mukesh, W.H. Zimmerman, M.P. Rosynek, C. Li, Ind.
Eng. Chem. Res. 29 (1990) 1588
[15] P.K. Basu, S.B. Basu, S.K. Mitra, S.S. Dasandhi, S.S. Bhattcherjee, P. Samuel, Stud.
Surf. Sci. Catal. 113 (1998)
[16] J.J. Jothimurugesan, S.K. Spiey, S.K. Gangwal, J.G. Goodwin, Stud. Surf. Sci. Catal.
119 (1998) 215
[17] W. Hou, B. Wu, X. An, T. Li, Z. Tao, H. Zheng, H. Xiang, Y. Li, Catal. Lett. 119
(2007) 353
[18] W. Hou, B. Wu, Y. Yang, Q. Hao, L. Tian, H. Xiang, Y. Li, Fuel Process. Technol.
89 (2008) 284
[19] K.W. Jun, H.S. Roh, K.S. Kim, J.S. Ryu, K.W. Lee, Appl. Catal. A: Gen. 259 (2004)
221
116
[20] H. Dlamini, T. Motjope, G. Joorst, G.T. Stege, M. Mdleleni, Catal. Lett. 78 (2002)
201
[21] C.H. Zhang, Y. Yang, B.T. Teng, T.Z. Li, H.Y. Zheng, H.W. Xiang, Y.W. Li
J. Catal. 237 (2006) 405
[22] C.N. Mbileni, Applications of mesostructured carbonaceous materials as supports
for Fischer-Tropsch metal catalyst, PhD Thesis, University of the Witwatersrand,
Johannesburg, 2006
[23] M. Luo, R.J. O?Brein, S. Bao, B.H. Davis, Appl. Catal. A: Gen. 239 (2003) 111
[24] G. Zhao, C. Zhang, S. Qin, H. Xiang, Y. Li, J. Mol. Catal. A: Chemical 286 (2008)
137
[25] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Muilenberg, Handbook
of X-Ray Photoelectron Spectroscopy, Perkin-Elmer Corporation, Physical Electronics
Division, Eden Prairie, Minn. (1979) 55344
[26] J.C. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, PCCP 2 (2000) 1319
[27] G. Hollinger, Appl. Surf. Sci. 8 (1981) 318
[28] J.R. Pitts, T.M. Thomas, A.W. Czanderna, M. Passler, Appl. Surf. Sci. 26 (1986) 107
[29] J. Finster, E.-D. Klinkenberg, J. Heeg, Vacuum 41 (1990) 1586
[30] A. Carnera, P. Mazzoldi, A. Boscolo-Boscoletto, F. Caccavale, R. Bertoncello, G.
Granozzi, J. Non-cryst. Solids 125 (1990) 293
[31] M. J. Heal, E. C. Leisegang, R. G. Torrington, J. Catal. 51 (1978) 314
[32] D. Schanke, G.R. Fredriksen, E.A. Blekkan, A. Holmen, Catal. Today 9 (1991) 69
[33] G. Bian, A. Oonuki, Y. Kobayashi, N. Koizumi, M. Yamada, Appl. Catal. A: Gen.
219 (2001) 13
[34] L.D. Mansker, Y. Jin, D.B. Buker, A.K. Datye, Appl. Catal. A 186 (1999) 277
[35] S. Li, G.D. Meitzner, E. Iglesia, J. Phys. Chem. B 105 (2001) 5743
[36] K. Nakanishi, Infrared Absorption spectroscopy ? PRACTICAL - , Nankodo
Company Limited, Japan, 1964
[37] Y. Lei, K. Ng, L. Weng, C. Chan, L. Li, Surf. Interface Anal. 35 (2003) 852
[38] D.T. Clark, A. Dilks, H.R. Thomas, J. Polym. Sci., Polym. Chem. Ed. 16 (1978)
1461
[39] D.T. Clark, W.J. Feast, D. Kilcast, W.K.R.Musgrave, J. Polym. Sci.,
117
Polym. Chem. Ed. 11 (1973) 389
118
Chapter 6
Effect of SiO2 content on a promoted precipitated iron-based Fischer-
Tropsch synthesis catalyst
6.1 Introduction
There is an increasing interest in studying the structural, electronic and chemical
properties of precipitated iron-based Fischer-Tropsch synthesis catalysts. In particular,
the Ruhrchemie catalyst (Cu/K2O/SiO2/Fe) has received much attention due to its
excellent catalytic performance.
SiO2 is often incorporated into this catalyst as a structural promoter [1-4]. However, its
addition suppresses the reduction as well as the activity of the catalyst due to the
variations in surface structure and interaction between iron and silica [5]. To circumvent
this from happening, chemical promoters such as K2O and Cu [6-8] are often added.
These chemical promoters are thought to facilitate the reduction of the catalyst as well as
the adsorption and dissociation of CO [9-17].
It is also well known that the intimate contacts between iron and chemical promoters
result in an important influence on the catalyst activity and selectivity [8, 15, 18]. On an
iron-based catalyst incorporated with SiO2, the existence of Fe-SiO2 interaction has been
extensively discussed in the literature [5, 19]. Nevertheless, little attention has been
focused on the effect of SiO2 on the interaction between iron and chemical promoters,
especially for multi-component catalysts.
Also most of the studies reported have focused on the effect of SiO2 as a structural
promoter and not as a chemical promoter [23, 26]. The present study was undertaken to
investigate the effect of SiO2 content on the interaction of Fe/Cu and Fe/K2O, as well as
FTS performances.
119
6.2 Experimental
All the catalysts employed in this Chapter were prepared using the precipitation method
as explained in Chapter 3. Several characterisation techniques such as N2 physisorption,
H2 TPR and ?in situ? DRIFTS were employed to characterise the iron-promoters contacts
and to illustrate the function of SiO2 in the catalyst. The way in which the
characterisation experiments were carried out is comprehensively discussed in Chapter 3.
These results will be discussed first and will be followed by the results from the FTS
reactor studies.
6.3 Results and discussion
6.3.1 Catalyst characterization
6.3.1.1 N2 physisorption
The textural properties of the catalysts were determined using N2 physisorption as
illustrated in Table 6.1. The effect of SiO2 content on the interaction of Cu and K2O on
Fe was studied. The weight loading of SiO2 was varied from 1 wt. % to 10 wt. % and the
amounts of Cu and K2O were kept constant (Table 6.1).
Table 6.1 The composition and textural properties of the catalysts
Catalyst (parts by weight) Surface area (m2/g)a Pore volume (cm3/g)
1SiO2/2K2O/5Cu/100Fe 65.2 0.14
3SiO2/2K2O/5Cu/100Fe 99.0 0.16
5SiO2/2K2O/5Cu/100Fe 130 0.17
10SiO2/2K2O/5Cu/100Fe 142 0.17
a Maximum error = ? 2 %
120
It is noticed that as the weight loading of SiO2 is increased the total surface area of the
catalyst increases. This is expected and is in agreement with work carried out by
Hayakawa [20] and Hou et al. [24]. This explains the reason why SiO2 is always
incorporated into precipitated Fe-based FTS catalysts; the SiO2 enhances the surface area
of the active Fe crystallites. This surface area provided by SiO2 also allows the Fe
crystallites to be well dispersed and not to easily come together or sinter. Sintering
produces larger Fe crystallites which are deemed less active for the FTS reaction and also
lead to deactivation of the catalyst [21, 26].
6.3.1.2 H2 TPR
The reduction behaviour of the Fe catalysts as determined using H2-TPR are shown in Fig
6.1. All Fe catalysts show two distinct peaks at 280-290 ?C and 700-760 ?C, which are
assigned to the reduction of Fe2O3 ? Fe3O4 and Fe3O4 ? Fe reactions respectively [22,
23].
Increasing the SiO2 loading does not alter the reduction temperature of the first peak; in
fact it is only when 10 wt. % SiO2 loading is added that we see a temperature shift.
Increasing SiO2 loading increases the total surface area, leading to an improved
dispersion of the Fe oxide crystallites.
However this has no effect on the reduction behaviour of the catalyst. The Fe3O4 ? Fe
reduction step is represented by the second peak. This peak shifts from 760 ? C when
SiO2 loading is 1 wt.% to 702 ?C for 10 wt.% SiO2 loading. Here the increased SiO2
content, unexpectedly leads to an increase in reducibility.
Hou et al. [24] have found that increasing the SiO2 content for a xSiO2/4.2K/5Cu/100Fe
catalyst (x = 15 to 40, catalyst composition based on parts by weight) facilitates the
dispersion of Fe2O3 and CuO and decreases the crystallite size of Fe2O3, leading to more
Fe2O3 being exposed to the surface of the catalyst. Other researchers have also published
similar findings [25, 26].
121
It is also known that Cu improves the reduction of the iron oxide to metallic iron via a
?spillover phenomenon? [17]. It may be that the dispersion of copper species caused by
SiO2 allows CuO to be easily reduced to metallic Cu allowing the crystallites formed to
provide H2 dissociation sites [13, 14, 16], which in turn lead to reactive hydrogen species
that are able to reduce Fe oxides at lower temperatures [17].
It is also noticeable that as the SiO2 loading is increased, the total area under the two
peaks gets smaller. In fact the area under the two peaks represents the amount of H2 being
consumed during the reduction reaction. To quantitatively get a better sense of the
amount of H2 being consumed. The area of each peak was measured for each TPR
profile. The area under peak 1 represents the Fe2O3 ? Fe3O4 reduction step and the area
under peak 2 illustrates the Fe3O4 ? Fe reduction step.
The results of these calculations are displayed in Table 6.2. It is noticed that as the SiO2
loading is increased the area of peak 1 increases while the area of peak 2 decreases.
These results imply that SiO2 favours the reduction of Fe2O3 ? Fe3O4 but restrains the
reduction of Fe3O4 ? Fe. The results may also mean that as the SiO2 loading is increased
it becomes extremely difficult to completely reduce the Fe2O3 phase to the metallic Fe
phase or the increment of SiO2 loading results in the amount of available iron oxide for
reduction to be decreased. It is to be noted that increasing the SiO2 content may lead to
increased Fe-SiO2 interaction and this could have also played a role in restraining the
Fe3O4 ? Fe reaction.
Overall the reduction of the Fe catalyst becomes difficult as shown by a decrease in the
total area of peaks with increasing SiO2 content. This would make sense and may mean
that more of the SiO2 is located at the surface masking some of the iron oxide species,
making it difficult for reactive H2 species to interact with them. Certainly XPS surface
analysis experiments would be required to back this assertion.
122
0 200 400 600 800 1000
0
40
80
120
Temperature (oC)
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
10SiO
2
/2K
2
O/5Cu/100Fe
5SiO
2
/2K
2
O/5Cu/100Fe
3SiO
2
/2K
2
O/5Cu/100Fe
1SiO
2
/2K
2
O/5Cu/100Fe
Figure 6.1 H2 TPR profiles for all catalysts
123
Table 6.2 Reduction temperatures for peak 1 and peak 2 as well as their areas
Total area = Area of Peak 1 + Area of Peak 2
Reduction
Temperature [oC]
Areas Catalyst composition
Peak 1 Peak 2 Peak 1 Peak 2 Total Area
1SiO2/5Cu/2K2O/100Fe 288 760 10300 8905 19205
3SiO2/5Cu/2K2O/100Fe 288 735 10967 8001 18968
5SiO2/5Cu/2K2O/100Fe 289 718 11211 7201 18412
10SiO2/5Cu/2K2O/100Fe 281 702 11271 6210 17481
124
6.3.1.3 ?In situ? CO adsorption using DRIFTS
CO adsorption measurements were carried out as outlined in Chapter 3. The results are
illustrated in Figures 6.2 and 6.3. Fig 6.2 compares CO adsorption on all the catalysts
studied. It is evident that two CO bands were identified for all the catalysts. These two
bands (2014 and 2034 cm-1) can be assigned to CO linearly bound to a Fe (0) species
[27].
2040 2020 2000 1980
0.00
0.04
0.08
0.12
0.16
0.20
2034 2014
10SiO
2
/2K
2
O/5Cu/100Fe
5SiO
2
/2K
2
O/5Cu/100Fe
3SiO
2
/2K
2
O/5Cu/100Fe
1SiO
2
/2K
2
O/5Cu/100Fe
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 6.2 CO adsorption on all the catalysts; Conditions: CO reduction for 1 hour (Flow
rate = 12 ml/min, T = 350oC, P = 2 bar), CO adsorption for 30 min (CO Flow rate = 12
ml/min, T = 25oC, P = 2 bar)
This Fe0 species may be Fe carbides or metallic Fe [30]. This is possible since Fe carbides
can directly form from iron oxides or via the carburization of metallic iron [23].
The reason why it is difficult to differentiate between metallic Fe species and Fe carbide
species is because the adsorption features of probe molecules on iron carbides are quite
125
similar to those on metallic iron particles. This is greatly supported by the adsorption
features of CO on H2 reduced iron samples performed in our lab (not reported herein).
Also Bian et al [30] have shown that the adsorption of CO on the iron carbide phase
produces adsorption bands with only a small shift in wavenumber from that on metallic
iron.
So the CO species at 2014 and 2034 cm-1 may represent the adsorption of CO on Fe-
carbides. This has been reported before by Bian et al. [30]. Unfortunately bridged CO
species could not be identified. They normally appear in the wavenumber region 1800 ?
2000 cm-1) [28, 29].
The intensity of the peak at 2014 cm-1 was compared for all the catalysts. This is
illustrated in Fig 6.3. It was found that as the SiO2 loading is increased above 3 wt. %, the
intensity of this peak decreases. This means that increasing the SiO2 loading above 3 wt.
% results in the decrease of Fe0 type species. Obviously this would mean a smaller
number of Fe2O3 crystallites were present at the surface, when they were exposed to CO
reduction. This result ties in well with postulation made earlier on (for the H2-TPR
results) that SiO2 interacts with the iron oxide crystallites, leading to a decreased
reduction of the iron oxide species.
126
A B C D
0.00
0.04
0.08
0.12
0.16
A = 1SiO
2
/2K
2
O/5Cu/100Fe
B = 3SiO
2
/2K
2
O/5Cu/100Fe
C = 5SiO
2
/2K
2
O/5Cu/100Fe
D = 10SiO
2
/2K
2
O/5Cu/100Fe
Catalysts
In
te
ns
ity
o
f p
ea
k
at
2
01
2-
20
15
c
m
-1
(K
ub
el
ka
-M
un
k
un
its
)
Figure 6.3 Comparing the intensity of peak at 2014 cm-1 for all the catalysts
It is also well known that K2O increases the extent of CO adsorption for iron based
catalysts [13, 31] and this is due to Fe coming into contact with K2O [26]. A decrease in
intensity of the peak at 2014 cm-1 as the SiO2 loading is increased, suggests that the
incorporation of SiO2 into the catalyst overwhelms the effect of the Fe/K2O contact. In
summary ? a clear relationship between SiO2 and CO adsorption exists and that is,
increasing the SiO2 loading suppresses the adsorption of CO.
127
6.3.1.4 ?In situ? FTS using DRIFTS
Fischer Tropsch synthesis reactions were also performed using the DRIFTS reactor. They
were monitored in-situ. Only the wavenumber region 2800-3100 cm-1 which showcases
the production of C-H type species will be presented here. Hydrocarbons produced from
the FTS reaction can be monitored using this part of the DRIFTS spectrum. The spectra
for all catalysts after 5 hours of reaction are illustrated in Fig. 6.4.
Figure 6.4 DRIFTS spectra showcasing FTS reactions for all catalysts;
Conditions: Reduction for 1 hour (H2/CO = 2/1, Flow rate = 12 ml/min, T = 350 oC, P = 2
bar), FTS reaction for 5 hours (H2/CO = 2/1, Flow rate = 12 ml/min, T = 275oC, P = 10
bar)
3100 3000 2900 2800 2700
0
2
4
6
8
10
10SiO
2
/2K
2
O/5Cu/100Fe
5SiO
2
/2K
2
O/5Cu/100Fe
3SiO
2
/2K
2
O/5Cu/100Fe
1SiO
2
/2K
2
O/5Cu/100FeK
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
128
When the loading of SiO2 is increased the intensities of C-H peaks are lowered. This
invariably means that SiO2 lowers the production of the C-H species. It also suggests that
the Fe-SiO2 interaction is more expected than Fe/Cu and Fe/K2O interactions. This
postulation is further confirmed in Fig. 6.5 where spectra showcasing SiO2 containing
catalysts are compared to a catalyst only containing Cu and K2O. It is clear from Fig. 6.5
that SiO2 does inhibit the Fe/Cu and Fe/K2O interactions. The intensity of the C-H
species are drastically lowered immediately after the addition of 1 wt. % SiO2 and
becomes worse with higher loadings of SiO2.
3100 3000 2900 2800 2700
0
40
80
120
160
5Cu/2K
2
O/100Fe
10SiO
2
/5Cu/2K
2
O/100Fe
5SiO
2
/5Cu/2K
2
O/100Fe
3SiO
2
/5Cu/2K
2
O/100Fe
1SiO
2
/5Cu/2K
2
O/100Fe
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 6.5 DRIFTS spectra comparing FTS reactions for SiO2 loaded catalysts to a non-
loaded SiO2 catalyst; Conditions: Reduction for 1 hour (H2/CO = 2/1, Flow rate = 12
ml/min, T = 350 oC, P = 2 bar), FTS reaction for 5 hours (H2/CO = 2/1, Flow rate = 12
ml/min, T = 275oC, P = 10 bar)
129
6.3.2 FTS reactor studies
The carbon monoxide (CO) and hydrogen (H2) conversions with time on stream for all
catalysts are displayed in Fig. 6.6, Fig 6.8 and Table 6.2. It can be seen from Figure 6.6
and 6.8 that increasing the SiO2 content has an effect on both CO and H2 conversions. To
get a sense of this effect, CO and H2 conversion values at steady state conditions
(constant CO and H2 conversion) were plotted for all the catalysts. These plots are
illustrated by Figures 6.7 and 6.9. Fig. 6.7 shows the CO conversion for all catalysts and
it is noticeable that CO conversion goes through a maximum at 3 wt. % SiO2 loading.
The same trend is observed with the H2 conversion plot (Fig 6.9). At a fixed set of
process conditions, the CO conversion can be used as an indication of FTS activity [24].
So it is clear that 3 wt. % SiO2 leads to the maximum activity when incorporated into an
Fe-based catalyst. Even before steady state is reached the 3 wt. % SiO2 loaded catalyst
has the highest activity (Fig. 6.6).
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
0 20 40 60 80 100 120 140 160
Time on stream (h)
C
O
c
on
ve
rs
io
n
(%
)
1SiO2/2K2O/5Cu/100Fe
3SiO2/2K2O/5Cu/100Fe
5SiO2/2K2O/5Cu/100Fe
10SiO2/2K2O/5Cu/100Fe
Figures 6.6 The carbon monoxide conversion with time on stream for all catalysts
130
0
10
20
30
40
50
60
CO conversion (%)
A B C D
Catalyst
A = 1SiO2/2K2O/5Cu/100Fe
B= 3SiO2/2K2O/5Cu/100Fe
C = 5SiO2/2K2O/5Cu/100Fe
D = 10SiO2/2K2O/5Cu/100Fe
Figure 6.7 Comparing CO conversion for all catalysts at steady state conditions
It is also interesting to note that the 3 wt. % SiO2 loaded catalyst reaches stability fairly
quickly and appears to be more stable on stream for a long time when compared to the
other catalysts. Another interesting observation on the stability of all the catalysts is that
the 3 wt. % and 5 wt. % SiO2 loaded catalysts have similar maximum activities before
reaching steady state. But the activity of 1 wt. % SiO2 loaded catalysts continues to
decrease with time on stream whereas the activity of the 10 wt. % SiO2 remains stable
and tends to increase with time on stream. This clearly illustrates that increasing the
loading amount of SiO2 improves the catalyst?s stability.
131
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
0 20 40 60 80 100 120 140 160
Time on stream (h)
H
2
co
nv
er
si
on
(
%
)
1SiO2/2K2O/5Cu/100Fe
3SiO2/2K2O/5Cu/100Fe
5SiO2/2K2O/5Cu/100Fe
10SiO2/2K2O/5Cu/100Fe
Figures 6.8 The hydrogen conversion with time on stream for all catalysts
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
H
2
co
n
ve
rs
io
n
(
%
)
A B C D
Catalyst
A =1SiO2/2K2O/5Cu/100Fe
B = 3SiO2/2K2O/5Cu/100Fe
C = 5SiO2/2K2O/5Cu/100Fe
D = 10SiO2/2K2O/5Cu/100Fe
Figure 6.9 Comparing H2 conversion for all catalysts at steady state conditions
132
Table 6.3 Reaction performances of all catalysts at steady state conditions
Catalysta 1SiO2/100Fe 3SiO2/100Fe 5SiO2/100Fe 10SiO2/100Fe
CO conversion (%) 28.9 60.0 51.6 37.1
H2 conversion (%) 18.5 30.2 29.5 24.8
Rate CO (mol/s) -7.3 x 10-7 -1.9 x 10-6 -1.1 x 10-6 -1.0 x 10-6
Rate CO2 (WGS)
(mol/s)
1.95 x 10-7 7.0 x 10-7 3.34 x 10-7 3.58 x 10-7
Rate FT 5.39 x 10-7 1.16 x 10-6 7.95 x 10-7 6.67 x 10-7
Activity
(?mol/sec.gFe)
7.34 18.6 11.3 10.2
? 0.69 0.64 0.68 0.61
C2 olefin %b 39.7 45.2 44.1 34.3
Selectivity
C1 19.3 19.8 20.4 23.3
C2-C4 36.9 43.1 40.7 42.7
C5-C11 38.2 31.2 33.4 28.0
C12+ 4.40 4.56 4.46 4.49
CO2 2.88 10.3 4.92 5.28
a All catalysts contained 2K2O and 5Cu
b C2= /(C2 + C2= ) [olefin to total C2 hydrocarbon weight ratio]
Data consists of ? 5% experimental error
Catalyst mass: 0.1 g
Reduction: CO, flow rate = 12 ml/min, t = 20-24 h, T = 350 oC, P = 2 bar
Reaction conditions: H2/CO = 2, flow rate = 12 ml/min, t = 140 h, T = 275 ?C, P = 10 bar
133
The effect of silica content on the product selectivity in the FTS reaction is shown in
Table 6.3. It can be seen that SiO2 loading greater than 1 wt. % increases selectivity to
C2-C4 hydrocarbons. It can also be noticed that the highest loading of SiO2 (10 wt. %)
gives the highest the methane selectivity and the lowest selectivity to C5-C11
hydrocarbons and C2 olefins. The ? value (0.61) for this catalyst is also the lowest. All of
these results imply that chain growth is restrained whereas the hydrogenation reaction is
enhanced. This could be attributed to SiO2 retarding the Fe/K2O interaction, since
potassium is known to promote the chain propagation reaction and olefin selectivity [8,
12, 25, 32, 37].
It has also been reported that potassium enhances the dissociative adsorption of CO and
suppresses H2 adsorption. Because it is an alkali promoter, it increases the basicity of the
iron surface leading to increased CO adsorption [26, 33]. These chemical effects lead to
the promotion of chain growth and olefin selectivity.
Previous reports have also suggested that K2O can interact with SiO2 and this may lead to
the promotional effect of potassium on FTS activity and selectivity to be decreased [8,
33, 34, 35, 36]. It is logical to think that a K2O-SiO2 interaction could suppress the
promotional effect of potassium since SiO2 is acidic in nature [33]. This means that the
interaction of SiO2 with K2O could decrease the basicity of the iron surface, leading to
the dissociative adsorption of CO to be suppressed, thereby retarding the chain growth
reaction [20, 26]. This will result in a lower coverage of carbon species on the Fe surface
whereas the H2 present will enhance chain termination rates and the production of light
paraffins due to olefins being hydrogenated [26].
CO adsorption results presented earlier demonstrated that as the SiO2 content is increased
the adsorption ability of the Fe surface is decreased. This may serve as evidence that as
the K2O-SiO2 interaction increases the Fe-K2O interaction diminishes.
134
6.4 Conclusion
The effect of SiO2 content on an unpromoted precipitated iron-based catalyst was studied
in the previous chapter. Interesting observations were noted and were all attributed to the
presence of SiO2. In this chapter the effect of SiO2 content on the promoted precipitated
iron-based catalyst was studied. Incorporation of SiO2 to the promoted precipitated iron-
based catalyst was found to have a significant influence into the reduction and adsorption
behaviours, as well as the catalytic activity of the catalyst. The changes in catalytic
activity could primarily be attributed to the effects of SiO2 on the Fe/Cu and Fe/K2O
interactions, which led to different degrees of H2 reduction and CO adsorption and further
significantly affected the FTS performances of the catalyst.
SiO2 stabilized the iron oxide crystallites by providing adequate surface area. This
facilitated the high dispersion of Fe2O3 and CuO and enhanced the contact between Fe2O3
and CuO. The enhanced Fe/Cu contact promoted the reduction of Fe2O3 to Fe3O4,
whereas the transformation of Fe3O4 to Fe was suppressed. Furthermore, due to the K2O-
SiO2 interaction, the catalyst loaded with 10 wt. % SiO2 (highest SiO2 loading) had a
weak contact between Fe and K2O, which reduced the surface basicity of the catalyst and
severely suppressed the CO adsorption.
In the FTS reaction, the FTS activity went through a maximum at 3 wt. % loading of
SiO2 and further increments of SiO2 loading decreased the catalyst activity. The SiO2
content also affected the hydrocarbon selectivity. At the highest SiO2 loading, the product
distribution shifted to light hydrocarbons and the C5-C11 hydrocarbons and C2 olefins
selectivity were suppressed.
135
References
[1] M.A. Vannice, R.L. Garten, J. Catal. 66 (1980) 242
[2] H.J. Jung, P.L. Walker, M.A. Vannice, J. Catal. 75 (1982) 416
[3] H.N. Pham, A. Viergutz, R. Gormley, A.K. Datye, Powder Technol. 110
(2000) 196
[4] R. Zhao, J.G. Goodwin, K. Jothimurugesan Jr., S.K. Gangwal, J.J. Spivey,
Ind. Eng. Chem. Res. 40 (2001) 1065
[5] C.H. Zhang, H.J. Wan, Y. Yang, H.W. Xiang, Y.W. Li, Catal. Comm. 7 (2006) 733
[6] A.F.H. Wielers, A.J.H.M. Kock, C.E.A. Hop, J.W. Geus, A.M. van der Kraan, J.
Catal. 117 (1989) 1
[7] M.V. Cagnoli, S.G. Marchetti, N.G. Gallegos, M. Alvarez, R.C. Mercader, A.A.
Yeramian, J. Catal. 123 (1990) 21
[8] D.B. Bukur, X. Liang, D. Mukesh, W.H. Zimmerman, M.P. Rosynek, C. Li, Ind. Eng.
Chem. Res. 29 (1990) 1588
[9] M.E. Dry, G.J. Oosthuizen. J. Catal. 11 (1968) 18
[10] M.E. Dry, T. Shingles, L. Boshoff, G.J. Oosthuizen. J. Catal. 15 (1969) 190
[11] J. Benziger, R. Madix, Surf. Sci. 94 (1980) 119
[12] D.B. Bukur, D.S. Mukesh, S.A. Patel, Ind. Eng. Chem. Res. 29 (1990) 194
[13] R. J. O?Brein, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal. Today
36 (1997) 325
[14] S. Li, A. Li, S. Krishnamoorthy, E. Iglesia Catal. Lett. 77 (4) (2001) 197
[15] Y. Jin, A.K. Datye, J. Catal. 196 (2000) 8
[16] R.J. O?Brein, B.H. Davis, Catal. Lett. 94 (2004) 1
[17] I.E. Wachs, D.J. Dwyer, E. Iglesia, Appl. Catal. 12 (1995) 35
[18] Y. Yang, H.W. Xiang, Y.Y. Xu, L. Bai, Y.W. Li, Appl. Catal. A: Gen. 266
(2004) 181
[19] W. Hou, B. Wu, X. An, T. Li, Z. Tao, H. Zheng, H. Xiang, Y. Li, Catal. Lett. 119
(2007) 353
[20] H. Hayakawa, H. Tanaka, K. Fujimoto, Appl. Catal. A: Gen. 328 (2007) 117
136
[21] B.H. Davis, E. Iglesia, Technology Development for Iron and Cobalt Fischer-
Tropsch Catalysis, Final Report, Contract No. DE-FC26-98FT40308, University of
Kenturky Research Foundation, Lexington, 2002
[22] R. Brown, M.E. Cooper, D.A. Whan, Appl. Catal. 3 (1982) 177
[23] D.B. Bukur, C. Sivaraj, Appl. Catal. A Gen. 231 (2002) 201
[24] W. Hou, B. Wu, Y. Yang, Q. Hao, L. Tian, H. Xiang, Y. Li, Fuel Process. Technol.
89 (2008) 284
[25] Y. Yang, H.W. Xiang, L. Tian, H. Wang, C.H. Zhang, Z.C. Tao, Y.Y. Xu, B. Zhong,
Y.W. Li, Appl. Catal. A: Gen. 284 (2005) 105
[26] H.J. Wan, B.S. Wu, Z.C. Tao, T.Z. Li, X. An, H.W. Xiang, Y.W. Li, J. Mol. Catal.
A: Chemical 260 (2006) 255
[27] A.F.H. Wielers, A.J.H.M. Kock, C.E.C.A. Hop, J.W. Geus, A.M. Van der Kraan, J.
Catal. 117 (1989) 1
[28] M.J. Heal, E.C. Leisegang, R. Torrington, J. Catal. 51 (1978) 314
[29] G. Blyholder, L.D. Neff, J. Phys. Chem. 66 (1962) 1464
[30] G. Bian, A. Oonuki, Y. Kobayashi, N. Koizumi, M. Yamada, Appl. Catal. A: Gen.
219 (2001) 13
[31] D.G. Miller, M. Moskovits, J. Phys. Chem. 92 (1998) 6081
[32] S. M?rup, H. Tops?e, Appl. Phys. 11 (1976) 63
[33] G. Zhao, C. Zhang S. Qin, H. Xiang, Y. Li, J. Mol. Catal. A: Chemical 286 (2008)
137
[34] M.E. Dry, Catal. Today 71 (2002) 227
[35] M.E. Dry, in: J.R. Anderson, M. Boudart (Eds.), Catalysis Science and Technology,
Vol. 1, Springer, New York, 1981, p. 160.
[36] M.E. Dry, Catal. Lett. 7 (1990) 241
[37] T.J. Donnelly, C.N. Satterfield, Appl. Catal. 52 (1989) 93
137
Chapter 7
Evaluating indium as a chemical promoter in Fe-based Fischer-Tropsch
synthesis
7.1 Introduction
Although the impact of copper as a promoter on the activity of Fe-based catalysts has
been extensively studied, amazingly, the effects of few other potential promoters have not
been significantly investigated or compared directly to the effects caused by copper when
carrying out a systematic study. In this study the impact of adding indium as a promoter
(which is suspected to have similar chemical properties to copper) on the catalytic
properties of precipitated bulk Fe-based catalysts was investigated using the same
preparation method and reaction conditions.
Copper is normally added to Fe-based Fischer Tropsch catalysts as a chemical promoter.
It is added to enhance hematite reducibility [1]. When copper oxide is reduced to metallic
Cu, the crystallites formed provide H2 dissociation sites [1-4], which in turn lead to
reactive hydrogen species that are able to reduce Fe oxides at lower temperatures. This
phenomenon is often referred to as H2 spillover [5]. Increased loading of copper onto Fe
FTS catalysts increases the FTS rate as well as the water gas shift (WGS) reaction [3].
Copper also has a positive effect on product selectivity over a wide range of conversions
[2].
Work carried out by several researchers over the years has highlighted relationships that
exist between various elements of the periodic table. A less known relationship is the
?Knight?s Move relationship? [6]. It takes its name from the knight?s move in the game
of chess, referring to a move of one step in any direction followed by two steps in a
direction at right angles to the first movement (Fig. 7.1).
138
Cu Zn Ga
Ag Cd In Sn Sb
Au Hg TI Pb Bi Po
114
Figure 7.1 Elements that show knight?s move relationships (after E.R. Scerri [6])
Fig. 7.1 suggests that elements like Zn and Sn should show a ?Knight?s Move
relationship? and have similar properties [7]. For example both are used for plating steel
such as in the case of food cans [6]. Not only do layers of both metals successfully delay
the onset of corrosion in the iron, but they are also non-poisonous.
With this in mind we decided to examine the use of indium as a catalyst promoter in the
FTS reaction and compare the results with those of copper, since both copper and indium
are in a position to each other to exhibit the ?Knight?s Move relationship?. Indeed not
much work has been carried to evaluate indium?s effects as a potential promoter and to
the best of our knowledge no work has been published to compare the promotional
effects of indium to copper for Fe-based FTS catalysts.
139
7.2 Motivation to compare indium to copper as a chemical promoter
Recent work in our laboratory to synthesise carbon nanotubes from acetylene, has shown
that indium can exhibit similar chemical properties to copper [8]. Employing a CaCO3
supported Fe-Ni catalyst results in the synthesis of nanotubes as shown in Fig. 7.2.
Figure 7.2 Carbon nanotubes synthesized using the Fe-Ni/CaCO3 catalyst
But adding copper to a CaCO3 supported Fe-Ni catalyst, results in the formation of tubes
as well as coils (Fig. 7.3). To our amazement, the same effect was observed when indium
was added to the Fe-Ni/CaCO3 (Fig. 7.4).
Figure 7.3 Carbon nanotubes and coils synthesized using the Fe-Ni-Cu/CaCO3 catalyst
140
Figure 7.4 Carbon nanotubes and coils synthesized using the Fe-Ni-In/CaCO3 catalyst
Furthermore the ratio of the tubes to coils produced for the copper and indium promoted
catalysts were found to be comparable (Fig. 7.5).
0
10
20
30
40
50
60
70
Indium Copper
P
er
ce
n
ta
ge
C
o
m
p
os
iti
on
Coils
Tubes
Figure 7.5 Percentage composition of coils and tubes produced for the copper and
indium promoted catalysts
141
Therefore with this possible link that exists between the chemical properties of copper
and indium, we decided to carry out a study to compare the promotional properties of the
two elements on the precipitated Fe-based FTS catalyst.
7.3 Experimental
Five catalysts were prepared using the precipitation method as explained in Chapter 3.
They include two copper-promoted catalysts and two indium promoted catalysts. The
weight loading for the promoters was 1 and 3 wt. %. The fifth catalyst prepared was the
unpromoted catalyst (100Fe) which was used as the benchmark catalyst. It is important to
note that this catalyst is not the same as the one used in Chapters 5. The characterization
results and a comparison of the properties of the catalysts are given below.
7.4 Results and discussion
7.4.1 N2 physisorption results
Table 7.1 The composition and textural properties of the catalysts
Catalyst compositiona
(parts by weight)
BET surfaceb
area (m2/g)
Pore volumeb
(cm3/g)
100Fe 18.6 0.073
100Fe/1Cu 19.4 0.077
100Fe/3Cu 19.9 0.077
100Fe/1In 25.8 0.12
100Fe/3In 26.5 0.12
aWeight loadings verified using XRF, Maximum error = ? 5%
bMaximum error = ? 2%
142
Copper appears not to alter the surface area as well as the pore volume of Fe. Indium
increases the surface area and pore volume of the Fe-based catalyst.
7.4.2 Hydrogen Temperature Programmed Reduction (H2 TPR)
A comparison of TPR results for all catalysts is shown in Table 7.2 and Figures 7.6 and
7.7. All promoted catalysts were compared directly to the unpromoted Fe catalyst
(Fe2O3). It is reasonable to assume that only the Fe2O3 is detected after calcination based
on the similar TPR profiles obtained for all the Fe catalysts prepared in this study. XRD
work carried out (not reported herein) has also shown Fe2O3 to be the predominant Fe
phase after calcination.
All TPR profiles show 2 distinct reduction peaks. It has been suggested that the H2
reduction of Fe2O3 occurs via 2 main steps: Fe2O3 ? Fe3O4 ? Fe. These 2 elementary
reactions have been assigned to the first and second peaks in the H2 TPR profiles,
respectively (Fig 7.6) [9-11].
143
Figure 7.6 H2 TPR profiles of Cu promoted catalysts
Figure 7.7 H2 TPR profiles of indium promoted catalysts
0 200 400 600 800 1000
100Fe/3Cu
100Fe/1Cu
100Fe
Temperature (oC)
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/g
c
at
.)
0 200 400 600 800 1000
100Fe/3In
100Fe/1In
100Fe
Temperature (oC)
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/g
c
at
.)
144
It is noticeable that copper lowers the reduction peaks of iron oxide to metallic iron. This
same effect has been observed by other authors [4, 12, 13]. A similar effect is noticed
with the indium promoted catalysts. But a striking effect with these results is that, the H2
reduction peaks of the indium promoted catalysts are smaller than those of the copper
promoted catalysts. To get a sense of the size of the peaks, we calculated the moles of H2
consumed by the catalysts. Thereafter, the Fe reducibility was calculated in the same way
as reported in literature [15].
Table 7.2 Comparing the reducibility of Fe-based catalysts using H2 TPR
H2 TPR Catalyst
Peak temperature
(?C)
H2 Consumption
(mmol/molFe)
Fe reducibility
(%)
100Fe 536 72 15
100Fe/1Cu 436 217 46
100Fe/3Cu 411 189 38
100Fe/1In 451 9.8 2
100Fe/3In 422 9.8 2
From Table 7.2, it is evident that the percentage amount of iron reduced in the indium
promoted catalysts is relatively low compared to the copper promoted catalysts. This
quantifies the small size of the reduction peaks and suggests that indium decreases the
%Fe reducibility.
7.4.3 X-Ray diffraction (XRD)
The XRD technique was employed to determine the crystallite size of Fe2O3. The
objective was to see how both copper and indium affect the crystallite size of iron oxide.
Table 7.3 depicts the crystallite size determined using Rietvelt refinement (as explained
in Chapter 3).
145
Table 7.3 Fe2O3 crystallite size determined using Rietveld refinement
Catalyst Crystallite size (nm)
100Fe 32.8
100Fe/1Cu 43.6
100Fe/3Cu 48.6
100Fe/1In 41.6
100Fe/3In 40.9
The addition of copper and indium increases the crystallite size of Fe2O3. An explanation
for this effect could be that during the preparation of the catalysts, the introduction of
promoters (precursors) modifies the precipitation behaviour of ions in the solution. This
leads to the net repulsive effect of the Fe3+ ions to be neutralized, hence making the Fe3+
particles to come together. This is the same analogy that can be used to explain the
destabilization of colloidal systems [14].
7.4.4 CO adsorption measurements using DRIFTS
A study of adsorbed CO provides information about the extent of, and number of types of
adsorbed CO on the Fe. It can also act as a probe molecule with which to study the metal
on which it is adsorbed. Carbon monoxide is an ideal probe molecule for the
characterisation of Fischer-Tropsch catalysts using DRIFTS. The CO is able to accept
electron density from metal surface sites, resulting in formation of metal?carbonyl
complexes that can readily be monitored by the CO stretching frequency.
The complexes are characterized by IR absorption bands at 2100?1800 cm?1. The shift
from the vibrational energy of gas-phase CO (2143 cm?1) can be explained in terms of
simple molecular orbital (MO) theory. The 5? orbital of the CO molecule forms a ? bond
with an empty orbital on the metal, and for electron-rich surfaces, back-donation from the
metal d-orbitals into the antibonding ?*-orbitals of the CO molecule occurs, weakening
the C?O bond. The result is a red shift of the CO stretching frequency (compared with
146
?free? CO gas) and the appearance of bands caused by CO linearly and bridged bonded to
the metal surface. The precise position of these bands can provide valuable information
about the electron density of the metal sites [16-18].
When CO adsorption was performed on the unpromoted iron catalyst (100Fe), two peaks
were obtained at 2033 and 2013 cm-1, showcasing CO linearly adsorbed on Fe0 (Fig. 7.8).
This is consistent with work carried out by other researchers [19-21].
These peaks decreased in intensity during thermal desorption until they were completely
desorbed at 300?C (Fig. 7.9). The introduction of Cu to Fe produced a red shift of the
peak at 2013 cm-1 to ca. 2011 cm-1 (Fig. 7.8) highlighting the increased backdonation
ability of the d-orbitals of Fe. These results are in agreement with the H2 TPR results for
the Cu promoted catalysts. No other peaks were observed, although it is possible that
some were hidden by the gaseous CO peaks (2173 and 2115 cm-1), since a pressure of 2
bar CO was employed in the experiments.
147
2060 2040 2020 2000 1980
0.4
0.8
1.2
1.6
2.0 100Fe/3Cu
100Fe/1Cu
100Fe
Wavenumber (cm-1)
20
13
20
33
K
ub
el
ka
-M
un
k
un
its
Figure 7.8 Comparison of CO adsorption on the unpromoted iron catalyst and copper
promoted iron catalysts; Conditions: CO reduction for 1 hour (Flow rate = 12 ml/min, T
= 350oC, P = 2 bar), CO adsorption for 30 min (CO Flow rate = 12 ml/min, T = 25oC, P =
2 bar)
148
2300 2200 2100 2000 1900 1800 1700
0
5
10
15
20
CO desorption at 300oC
CO desorption at 200oC
CO desorption at 100oC
CO desorption at 50oC
CO adsorption at 25oC
2013
20
33
21
1521
73
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 7.9 Thermal desorption of CO on the unpromoted iron catalyst
Fig. 7.10 shows the adsorption of CO on indium promoted Fe catalysts. It is evident that
the intensity of species at 2033 and 2013 cm-1 has decreased with indium addition to Fe.
A comparison of the intensities of these peaks for copper promoted and indium promoted
catalysts is shown in Figures 7.11 and 7.12. This is to give a sense of CO adsorption
ability of these catalysts. It is clear from both figures that the intensity of the adsorbed
CO bands of the Cu promoted catalysts is at least seven times more than the intensity of
the indium promoted catalysts.
It is also noticed that the 2013 cm-1 peak intensity of the 3Cu case is lower than that of
the 1 Cu case. This could be due to the copper particles covering some Fe active sites
available for CO adsorption as a result of the higher loading of Cu. We do acknowledge
149
that at this point this is a mere speculation and confirmatory evidence in a form of XPS
data would be required to back this assertion
2060 2040 2020 2000 1980
0.4
0.8
1.2
1.6
2.0
Wavenumber (cm-1)
K
ub
el
ka
-M
un
k
un
its
100Fe/3In
100Fe/1In
100Fe
20
33
20
13
Figure 7.10 CO adsorption on the indium promoted iron catalysts
Figure 7.11 Intensity of peak at 2013 cm-1 for CO adsorption on the copper promoted
and the indium promoted iron catalysts
A B C
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8 A= 100Fe
B = 100Fe/1Cu
C = 100Fe/3Cu
Catalysts
In
te
ns
ity
o
f p
ea
k
at
2
01
3
cm
-1
(K
ub
el
ka
-M
un
k
un
its
)
A B C
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Catalysts
A= 100Fe
B = 100Fe/1In
C = 100Fe/3In
150
Figure 7.12 Intensity of peak at 2033 cm-1 for CO adsorption on the copper promoted
iron catalysts and the indium promoted iron catalysts
It is also noted that the CO adsorption spectra of the indium promoted catalysts reveal
CO adsorption peaks at 2024 and 2042 cm-1 (Fig. 7.13). This might indicate the presence
of a different type of iron species or simply a blue shift of the 2013 and 2033 cm-1 peaks.
If the emergence of these peaks is more likely as a result of the blue shift of the 2013 and
2033 cm-1 peaks, then this might suggest that indium inhibits the backdonation ability of
iron. This suggesting that indium acts as a poorer promoter for the Fe-based FTS catalyst.
It therefore appears that indium lowers the CO adsorption ability of the Fe catalyst.
A B C
0.0
0.2
0.4
0.6
0.8
1.0
1.2
In
te
ns
ity
o
f
pe
ak
a
t 2
03
3
cm
-1
(K
ub
el
ka
-M
un
k
un
its
)
Catalysts
A= 100Fe
B = 100Fe/1Cu
C = 100Fe/3Cu
A B C
0.0
0.2
0.4
0.6
0.8
1.0
Catalysts
A= 100Fe
B = 100Fe/1In
C = 100Fe/3In
151
Figure 7.13 CO adsorption on the indium promoted iron catalysts showing the adsorbed
CO species at 2024 and 2042 cm-1
2060 2050 2040 2030 2020 2010
0.00
0.04
0.08
0.12
0.16
20
42
20
24
100Fe/3In
100Fe/1In
Wavenumber (cm-1)
K
ub
el
ka
-M
un
k
un
its
152
7.4.5 In situ FTS performances using DRIFTS
After performing the FTS reaction for the copper and indium promoted catalysts for 5
hours, spectra were recorded. These are shown in Fig 7.14 and Fig 7.15. All the spectra
are compared to that of the spectrum showing FTS reaction performed over the
unpromoted Fe catalyst.
The peak at 3016 cm-1 is associated with gaseous methane (CH4), whereas peaks at 2957
cm-1 and 2929 cm-1 represent the asymmetric CH stretching vibration of the methyl
species (-CH3) and the asymmetric CH stretching of methylene species (-CH2-)
respectively. The peak at 2878 cm-1 is assigned to a symmetric CH stretching vibration of
the methyl species (-CH3) and the one at 2854 cm-1 is assigned to the symmetric CH
stretching vibration of the methylene species (-CH2-) [19, 29].
153
Figure 7.14 Comparison of the FTS reaction over the unpromoted iron catalyst (100Fe)
and the copper promoted catalysts; P = 10 bar, T = 275 ?C, H2/CO = 2, H2/CO flow rate =
12 ml/min, Time = 5 h)
3100 3050 3000 2950 2900 2850 2800
0
2
4
6
8
10
12
28
55
29
28
29
58
30
16
100Fe/3Cu
100Fe/1Cu
100Fe
Wavenumber (cm-1)
K
ub
el
ka
-M
un
k
un
its
154
3100 3050 3000 2950 2900 2850 2800
0
2
4
6
8
10
100Fe/3In
100Fe/1In
100Fe
28
55
29
28
29
58
30
16
K
ub
el
ka
-M
un
k
un
its
Wavenumber (cm-1)
Figure 7.15 Comparison of the FTS reaction over unpromoted iron catalyst (100Fe) and
indium promoted catalysts; P = 10 bar, T = 275 ?C, H2/CO = 2, H2/CO flow rate = 12
ml/min, Time = 5 h)
It is evident that the addition of indium to the Fe catalyst lowers the intensity of the C-H
peaks obtained after 5 hours of reaction whereas copper has a negligible effect on the
intensity of the C-H peaks. This tells us that the addition of indium to the Fe catalyst
hampers its activity. From this observation it appears that indium has a deleterious effect
on the activity of the Fe catalyst.
To estimate the average carbon chain length of the hydrocarbon molecules produced after
5 hours of reaction, the ratio of CH2/CH3 species was calculated using the following
formula:
155
1
)( 2
?
speciesCHArea ?? /
2
)( 3
?
speciesCHArea ??
where
1) Area of -CH2- species is the area of the peak at 2925-2930 cm-1 representing the
asymmetric stretch of CH2 species
2) Area of ?CH3 species is the area of the peak at 2955-2960 cm-1 representing the
asymmetric stretch of CH3 species
3) ?1 is the molar extinction coefficient of the CH2 species (75 mole-1.l.cm-1) [22]
4) ?2 is the molar extinction coefficient of the CH3 species (70 mole-1.l.cm-1) [22]
The calculated ratios are given in Table 7.4. It is evident that both indium and copper lead
to an increase in the average chain length of the hydrocarbons. This suggests that both
copper and indium can induce similar effects properties to the Fe FTS catalyst.
Table 7.4 Calculated ratios of CH2/CH3 bands for all catalysts
Catalyst Ratio of CH2/CH3
100Fe 1
100Fe/1Cu 7
100Fe/3Cu 7
100Fe/1In 4
100Fe/3In 5
156
From the results presented above it is clear that indium has similar properties to copper
but it is a poorer promoter than copper. In fact indium lowers the activity of the
precipitated Fe-based FTS catalyst. It is thought that indium poisons the active sites of
the catalyst by interacting with them. Our postulation is that during calcination, In(NO3)3
(which was used as the indium precursor) is transformed to In2O3 and during
pretreatment (before reaction) In2O3 is reduced to indium metal which has a low melting
point (157 ?C) [23] causing it melt during the FTS reaction and this resulted in some of
the active sites to be covered by this melted indium rendering them inactive.
157
7.5 Conclusion
The ability of indium to act as a chemical promoter for the Fe-based FTS catalyst was
evaluated. Its effect on Fe was evaluated and compared to that of copper. This was to
evaluate if both indium and copper possessed similar promotional abilities for the Fe-
based FTS catalyst. N2 physisorption, temperature programmed reduction (TPR), X-ray
diffraction (XRD) and diffuse reflectance infrared fourier transform spectroscopy
(DRIFTS) were employed to characterize the catalysts. ?In situ? Fischer Tropsch
synthesis (FTS) reactions were also performed in the DRIFTS reactor. It was found via
TPR studies that indium exhibited similar chemical properties to that of copper. Results
obtained from XRD and N2 physisorption showed indium promoted catalysts give
comparable results to those of copper promoted catalysts. It therefore appears that indium
does exhibit similar chemical properties to copper.
However indium decreased the reducibility and CO adsorption ability of the Fe catalyst.
Indium also lowered the FTS activity of the Fe-based catalyst. It is thus concluded that
indium is a poorer promoter for the iron-based FTS catalyst and acts as a poison for this
catalyst.
158
References
[1] S. Li, A. Li, S. Krishnamoorthy, E. Iglesia, Catal. Lett. 77 (2001) 197
[2] R. J. O?Brein, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal. Today
36 (1997) 325
[3] R.J. O?Brein, B.H. Davis, Catal. Lett. 94 (2004) 1
[4] Y. Jin, A.K. Datye, J. Catal. 196 (2000) 8
[5] I.E. Wachs, D.J. Dwyer, E. Iglesia, Appl. Catal. 12 (1995) 35
[6] E.R. Scerri, The Periodic Table, Its Story and Its Significance, Oxford University
Press, New York, 2007
[7] M. Laing, The Knight?s Move in the Periodic Table, Education in Chemistry, 36
(1999) 160
[8] A. Shaikjee, N.J. Coville, in preparation.
[9] D.B. Bukur, C. Sivaraj, Appl. Catal. A: Gen. 231 (2002) 201
[10] K. Jothimurugesan, J.G. Goodwin, Jr., S.K. Gangwal, J.J. Spivey, Catal. Today 58
(2000) 335
[11] I.S.C Hughes, J.O.H. Newman, G.C. Bond, Appl. Catal. 30 (1987) 303
[12] R. J. O?Brien, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal. Today
36 (1997) 325
[13] U. Lindner, H. Papp, Appl. Surf. Sci. 32 (1988) 75
[14] N. Lohitharn, J. G. Goodwin, Jr., E. Lotero, J. Catal. 255 (2007) 104
[15] http://en.wikipedia.org/wiki/Colloid
[16] F. Morales, E. de Smit, F. M.F. de Groot, T. Visser, B. M. Weckhuysen
J. Catal. 246 (2007) 91
[17] J. Ryczkowski, Catal. Today 68 (2001) 263
[18] G. Blyholder, L.D. Neff, J. Phys. Chem. 66 (1962) 1464
[19] G. Bian, A. Oonuki , Y.Kobayashi, N. Koizumi, M. Yamada, Appl. Catal. A: Gen.
219 (2001) 13
[20] E. Boellaard, A.M. van der Kraan, J.W. Geus, Appl. Catal. A: Gen. 147 (1996) 207
[21] E. Guglielminotti, F. Boccuzzi, F. Pinna, G. Strukul, J. Catal. 167 (1997) 153
159
[22] K. Nakanishi, Infrared Absorption spectroscopy ? PRACTICAL - , Nankodo
Company Limited, Japan, 1964
[23] http://www.chemicalelements.com/elements/cu.html (01 February 2009)
160
Chapter 8
Chemical promotion of a multi-promoted Fe-based Fischer-Tropsch
synthesis catalyst by indium
8.1 Introduction
In Chapter 7 it was reported that indium does exhibit some chemical properties similar to
those of copper. It was decided to employ indium as a promoter and evaluate its effect on
a multi-promoted precipitated iron catalyst. It was also noticed that at loadings of > 1 wt.
% indium had a deleterious effect on the chemical properties of the precipitated iron
catalyst. This prompted us to evaluate wt. % loadings of less than 1 wt. % for this study.
The aim of this study was thus to evaluate the effect of low loadings of indium on a
multi-promoted precipitated iron catalyst.
8.2 Experimental
The catalysts were prepared in the same manner as those reported in Chapter 6 and the
detailed experimental procedure is discussed in Chapter 3. The three catalysts prepared
contained indium, potassium, silica and iron and their compositions are shown in Table
8.1. All these catalysts were characterized using N2 physisorption, TPR and DRIFTS.
Their FTS performances were also evaluated.
161
8.3 Results and discussion
8.3.1 N2 physisorption measurements
The BET surface area measurements are given in Table 8.1. It is seen that upon addition
of indium on the promoted precipitated iron based FTS catalyst, both the surface area and
pore volume are decreased. The decrease in surface area could be due to indium filling
the pores of the SiO2.
Table 8.1 The composition and N2 physisorption results of the catalysts
Catalyst composition
(parts by weight)
BET Surface areaa
(m2/g)
Pore volume
(cm3/g)
2K2O/5SiO2/100Fe 87.3
0.15
0.01In/2K2O/5SiO2/100Fe 78.7
0.14
0.1In/2K2O/5SiO2/100Fe 64.3
0.14
aMaximum error = ? 2%
SiO2 is known to provide a high surface area [1, 2]. Thus as more promoters are added to
the catalyst, the high surface area of SiO2 diminishes.
162
8.3.2 Temperature programmed reduction (TPR)
8.3.2.1 H2 TPR
H2 TPR measurements were also performed on the catalysts. The results for all the
catalysts studied are shown in Fig. 8.1 and Table 8.2. In all the TPR profiles there are two
peaks and these two peaks illustrate the transformation of Fe2O3 to Fe via a two step
process [3-6]. The first peak represents the transformation of Fe2O3 ? Fe3O4 and the
second one shows the transformation of Fe3O4 ? Fe. It is noticed that the addition of
indium shifts the two reduction peaks to higher temperatures. This means that indium
suppresses the reduction of Fe2O3.
0 200 400 600 800 1000
0
20
40
60
80
0.1In/2K
2
O/5SiO
2
/100Fe
0.01In/2K
2
O/5SiO
2
/100Fe
2K
2
O/5SiO
2
/100Fe
Temperature (oC)
H
2 C
on
su
m
pt
io
n
(m
m
ol
H
2/m
ol
F
e)
Figure 8.1 H2 TPR profiles for all the catalysts
Table 8.2 shows that the first peak is shifted from 422 ?C to 473 ?C and the second peak
from 733 ?C to ca. 768-772 ?C. This change occurs for both the indium loaded catalysts.
It is interesting to note that this trend occurs even if the loading amount of indium is
increased tenfold (0.01 ? 0.1 wt. %). This signifies that the suppression ability of indium
is complete after addition of very small amounts of indium.
163
Table 8.2 H2 Reduction temperatures for all the catalysts in Figure 8.1
Reduction Temperature [oC] Catalyst composition
Peak 1 Peak 2
2K2O/ 5SiO2/100Fe 422 733
0.01In/2K2O/ 5SiO2/100Fe 473 772
0.1In/2K2O/ 5SiO2/100Fe 473 768
8.3.2.2 CO TPR
CO TPR measurements were also performed on the catalysts and the results are given in
Fig. 8.2 and Table 8.3. All the profiles show four peaks. The first peak is in the
temperature range 120 ? 150 ?C and may be the reduction of easily reducible iron-oxide
crystallites. Luo et al. [7] have shown that Fe2O3 occurs via a two step process as given
by Eqs. 8.1 and 8.2.
0 200 400 600 800
-8
-4
0
4
8
12
Temperature (oC)
C
O
C
on
su
m
pt
io
n
(m
m
ol
C
O
/m
ol
F
e)
0.1In/2K
2
O/5SiO
2
/100Fe
0.01In/2K
2
O/5SiO
2
/100Fe
2K
2
O/5SiO
2
/100Fe
Figure 8.2 CO TPR profiles for all the catalysts
164
3Fe2O3 + CO ? 2Fe3O4 + CO2 (8.1)
5Fe3O4 + 32CO ? 3Fe5C2 + 26CO2 (8.2)
The second peak located in the 290 ? 315 ?C range is ascribed to the reduction of Fe2O3 to
Fe3O4 as illustrated by Eq. 8.1. It is noticeable that this peak is unaffected by the addition
of a small amount of indium. However, when the indium loading is increased to 0.1 wt.
% the peak is shifted to 296 ?C. This indicates that indium promotes the reduction of
Fe2O3 to Fe3O4 and Fe2O3 is more easily reduced using CO than using H2.
Table 8.3 CO reduction temperatures for all the catalysts in Figure 8.2
Reduction Temperature [oC] Catalyst composition
Peak 1 Peak 2 Peak 3 Peak 4
2K2O/5SiO2/100Fe 146 315 622 736
0.01In/2K2O/5SiO2/100Fe 138 314 626 739
0.1In/2K2O/5SiO2/100Fe 127 296 593 699
The third peak in the temperature range 590 ? 630 ?C corresponds to the carburization of
iron oxides as illustrated by Eq. 8.2, and the fourth peak corresponds to the carburization
of the difficult to reduce iron oxide species. These difficult to reduce iron oxide species
could be present as a result of the Fe-SiO2 interaction, which many researchers have
widely reported [8-11]. Nonetheless, the addition of indium (especially 0.1 wt. %) has a
marked effect on the carburization peaks. Both the carburization peaks are shifted to
lower temperatures. This suggests that indium promotes the carburization of the iron
oxide phase. It is known that K2O promotes the dissociative adsorption of CO [12-14]
and in doing this it promotes the carburization of the iron oxide phase. This could mean
that indium plays a role in enhancing the carburization ability of K2O and of also
enhancing the Fe-K2O contact.
165
8.3.3 DRIFTS
8.3.3.1 ?In situ? CO adsorption measurements
The CO adsorption results are shown in Figures 8.3 and 8.4. CO adsorption peaks at 2014
and 2034 cm-1 were obtained and these bands represent the adsorption of CO on Fe0
species [15]. The intensity of the 2014 cm-1 band gives a qualitative measure of the
adsorption of CO and this is used to compare the CO adsorption for all the catalysts. This
is illustrated in Fig. 8.4.
2040 2020 2000 1980
0.00
0.04
0.08
0.12
0.16
0.20
0.24
2034
2014
2K
2
O/5SiO
2
/100Fe
0.01In/2K
2
O/5SiO
2
/100Fe
0.1In/2K
2
O/5SiO
2
/100Fe
Wavenumber (cm-1)
K
ub
el
ka
-M
un
k
un
its
Figure 8.3 CO adsorption on all the catalysts; Conditions: CO reduction for 1 hour (Flow
rate = 12 ml/min, T = 350oC, P = 2 bar, CO adsorption for 30 min (CO Flow rate = 12
ml/min, T = 25oC, P = 2 Bar)
166
A B C
0.00
0.04
0.08
0.12
0.16
0.20
A = 2K
2
O/5SiO2/100Fe
B = 0.01In/2K
2
O/5SiO2/100Fe
C = 0.1In/2K
2
O/5SiO2/100Fe
Catalysts
In
te
ns
ity
o
f p
ea
k
at
2
01
2-
20
15
c
m
-1
(K
ub
el
ka
-M
un
k
un
its
)
Figure 8.4 Comparison of the intensities of 2014 cm-1 peak
The 2014 cm-1 peak intensity is slightly increased when indium is added to the catalyst.
This implies that CO adsorption is slightly enhanced when indium is added. It may be
that indium improves carburization as shown by the CO TPR results and it also enhances
the Fe-K2O contact for the adsorption of CO. This would mean that K2O improves the
ability of Fe to adsorb CO. It is widely accepted that K2O promotes the CO adsorption
ability of Fe [14, 16].
8.3.3.2 ?In situ? FTS
The IR spectra showing the production of C-H peaks after 5 hours of FTS reaction
performed in a DRIFTS reactor for all the catalysts are shown in Fig. 8.5. It appears as if
indium has no marked effect on the production of C-H peaks as illustrated in Fig. 8.5.
167
3100 3000 2900 2800 2700
0.0
0.2
0.4
0.6
2926
29
59
30
16 2
05
6
Wavenumber (cm-1)
K
ub
el
ka
-M
un
k
un
its
2K
2
O/5SiO
2
/100Fe
0.1In/2K
2
O/5SiO
2
/100Fe
0.01In/2K
2
O/5SiO
2
/100Fe
Figure 8.5 DRIFTS spectra for the FTS reaction of all catalysts
Reaction conditions: H2/CO = 2/1, Flow rate = 12 ml/min, T = 275oC, P = 10 bar, t = 5 h
8.3.4 FTS performances
8.3.4.1 Catalyst activity and stability
The conversions of carbon monoxide and hydrogen over the catalysts with time on
stream are shown in Fig. 8.6, Fig. 8.7 and Table 8.4. It is noticeable that the addition of
indium to the catalyst has a marked effect on both the CO and H2 conversions. In fact
changing the loading of indium also has an effect on both the CO and H2 conversions.
168
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0 20 40 60 80 100 120 140 160
Time on stream (h)
C
O
c
on
ve
rs
io
n
(%
)
5SiO2/2K2O/100Fe
0.01In/5SiO2/2K2O/100Fe
0.1In/5SiO2/2K2O/100Fe
Figure 8.6 CO conversion with time on stream
169
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
0 20 40 60 80 100 120 140 160
Time on stream (h)
H
2
co
nv
er
si
o
n
(%
)
5SiO2/2K2O/100Fe
0.01In/5SiO2/2K2O/100Fe
0.1In/5SiO2/2K2O/100Fe
Figure 8.7 H2 conversion with time on stream
Figures 8.8 and 8.9 respectively show a comparison of CO and H2 conversion at steady
state conditions for all catalysts. It is noticeable that the addition of indium lowers the
activity of the catalyst, since both the CO and H2 conversions are decreased. The
calculated activity (Table 8.4) confirms this point. The activity decreases from 14.5 to
10.8 ?mol/sec.gFe. The activity is further decreased when the loading amount of indium
is increased from 0.01 wt. % to 0.1 wt. % (i.e. from 10.8 to 7.0 ?mol/sec.gFe). It is also
noticeable from Table 8.4 that the rate of CO conversion is lowered from -1.1 x 10-6
mol/s to -7.0 x 10-7 mol/s as the indium content is increased to 0.1 wt. %.
170
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
50.0
C
O
c
on
ve
rs
io
n
(%
)
A B C
Catalyst
A = 5SiO2/2K2O/100Fe
B = 0.01In/5SiO2/2K2O/100Fe
C = 0.1In/5SiO2/2K2O/100Fe
Figure 8.8 Comparison of the CO conversion for all catalysts at steady state
0.0
5.0
10.0
15.0
20.0
25.0
30.0
H
2
C
o
nv
er
si
o
n
(%
)
A B C
Catalyst
A= 5SiO2/2K2O/100Fe
B= 0.01In/5SiO2/2K2O/100Fe
C= 0.1In/5SiO2/2K2O/100Fe
Figure 8.9 Comparison of the H2 conversion for all catalysts at steady state conditions
171
The decrease in activity could be attributed to a decrease in surface area as confirmed by
the N2 physisorption results. It may also be that a suppression of the reduction properties
of the catalyst (H2 TPR results) is responsible for a decrease in the catalyst?s activity. It is
noted that the catalyst was activated in CO for the FTS runs. CO TPR results suggest that
indium improves the reduction/carburization of the iron oxide phase. It thus appears as if
a decrease in surface area of the catalyst is a logical explanation for the decrease in
activity.
It is also noticed that after 70 hours of reaction the activity of the indium promoted
catalysts continued to decline indicating that they had not reached stability.
172
Table 8.4 FTS reaction performances for all the catalysts
Catalyst 5SiO2/2K2O/100Fe 0.01In/5SiO2/2K2O/100Fe 0.1In/5SiO2/2K2O/100Fe
CO
conversion
(%)
49.9 38.4 26.3
H2 conversion
(%)
28.5 23.8 18.0
Rate CO
(mol/s)
-1.4 x 10-6 -1.1 x 10-6 -7.0 x 10-7
Rate CO2
(mol/s)
4.67 x 10-7 2.98 x 10-7 1.80 x 10-7
Rate FT
(mol/s)
9.79 x 10-7 7.76 x 10-7 5.22 x 10-7
Activity
(?mol/sec.gFe)
14.5 10.8 7.0
? 0.48 0.60 0.62
C2 olefin %a 44.0 40.7 36.8
Selectivity
C1 19.8 18.4 20.4
C2-C4 39.8 34.9 34.6
C5-C11 25.1 30.5 27.7
C12+ 10.2 11.7 11.9
CO2 6.89 4.40 2.66
Data has ? 5% experimental error
a C2= /(C2 + C2= ) [olefin to total C2 hydrocarbon weight ratio]
Catalyst mass: 0.1 g, Reduction: CO, flow rate = 12 ml/min, t = 20-24 h, T = 350 oC, P = 2 bar,
Reaction conditions: H2/CO = 2, flow rate = 12 ml/min, t = 140 h, T = 275 ?C, P = 10 bar
173
8.3.4.2 Product selectivity
The selectivity to FTS products produced is illustrated in Table 8.4. The introduction of
indium lowers the selectivity to low weight hydrocarbons (C2-C4 and C2 olefins), whereas
the selectivity to heavy weight hydrocarbons (C5-C11 and C12+) is increased. This trend
stays the same even when the loading of indium is increased. It may be that the addition
of indium enhances the Fe-K2O contact leading to K2O promotion to be boosted.
K2O promotion leads to enhanced CO adsorption and this in turn increases the
concentration of C atoms on the Fe surface, which promotes the chain growth reaction
over chain termination reactions [16]. CO adsorption results presented earlier did show a
slight increase in CO adsorption when indium was added, so this assertion could be true.
Indium addition also lowers CO2 selectivity. CO2 production may be a qualitative way of
evaluating a catalyst?s effect on the Water Gas Shift reaction (WGS) (Eq. 8.4). It is also
important to note that CO2 can be produced by the Boudouard reaction (Eq. 8.5).
FTS reaction: CO + H2 ? -CH2- + H2O (8.3)
WGS reaction: CO + H2O ? CO2 + H2 (8.4)
Boudouard reaction: CO + CO ? CO2 + C(s) (8.5)
Assuming that most of the CO2 produced is from the WGS reaction. This then suggests
that adding indium to the catalyst decreases the WGS activity. This could be true since
the rate of the WGS reaction is mainly controlled by the amount of H2O available. This
H2O is produced as a byproduct of the FTS reaction (Eq. 8.3). It is seen from Table 8.4
that the FTS rate is decreased when indium is added and when its loading amount is
increased. This would invariably slow down the production rate of H2O and this would in
turn decrease the formation rate of CO2. It is noticeable from Table 8.4 that the rate of
CO2 formation is also lowered from 4.67 x 10-7 to 2.98 x 10-7 mol/s when indium is
added. Increasing the loading amount of indium still results in a decrease as well. The
174
rate of the FTS reaction is also decreased further with an increment in the loading amount
of indium.
It is suggested that indium acts more as a ?poison? than a promoter for the FTS reaction.
As it was suggested in Chapter 7, it may be possible that during calcination and reduction
indium moves to the surface of the catalyst to cover some of the actives sites of the
catalyst.
During calcination In(NO3)3 is transformed into In2O3 (indium oxide) and this In2O3 is
transformed into In (indium) metal during reduction. Since indium has a very low melting
point (157?C) [17], it melts during the FTS reaction and covers some of the active sites of
the Fe-based FTS catalyst. To verify this postulate an H2 TPR experiment of indium
oxide was carried out (Fig. 8.10). In2O3 was reduced to metallic indium at ca. 250-350 ?C
and this temperature range is well within the temperature conditions employed for the
FTS reaction. The broad peak stretching from 440-950?C is ascribed to the volatilization
of indium.
The broadness of this peak may suggest that as In volatilizes, it is picked up by the TCD
detector and since a TCD detector was used which monitors the change in conductivity of
the effluent gas stream. It may be that as a component of In was volatizing, a change in
conductivity of the effluent stream occurred and the detector picked it up, consequently
resulting in a broad peak on our TPR profile.
175
0 200 400 600 800 1000
H
2 C
on
su
m
pt
io
n
(a
.u
.)
Temperature (oC)
Figure 8.10 H2 TPR profile of In2O3
8.4 Conclusion
The effect of adding indium to a multi-promoted Fe-based FTS catalyst was investigated.
The addition of indium suppressed the reduction properties of the catalyst when H2 was
employed as a reductant. When CO was employed as a reductant, the
reduction/carburization properties were improved. This improved the CO adsorption
ability of Fe and resulted in a selectivity shift to heavy weight hydrocarbons during the
FTS reaction, whereas low weight hydrocarbons were suppressed. Adding indium to the
promoted catalyst also lowered the catalyst surface area which resulted in a decrease to
the FTS activity.
176
References
[1] M.E. Dry, in Applied Industrial Catalysis, B.E. Leach (Ed.) 2, Academic Press, New
York, 1983
[2] V.U.S. Rao, G.J. Stiegel, G.J. Cinquegrane, R.D. Strvastava, Fuel Proc.
Tech. 30 (1992) 83
[3] H. Lin, Y. Chen and C. Li, Thermochimica Acta 400 (2003) 61-67
[4] D. B. Bukur, K. Okabe, M. P. Rosynek, C. P. Li, D. J. Wang, K. R. P. M. Rao
G. P. Huffman, J. Catal. 155 (1995) 353
[5] Y. Jin, A. K. Datye, J. Catal. 196 (2000) 8
[6] R. Brown, M.E. Cooper, D.A. Whan, Appl. Catal. 3 (1982) 177
[7] M. Luo, R. J. O?Brien, S. Bao, B. H. Davis, Appl. Catal. A: Gen. 239 (2003) 111
[8] C.H. Zhang, H.J. Wan, Y. Yang, H.W. Xiang, Y.W. Li, Catal. Comm. 7 (2006) 733
[9] H. Dlamini, T. Motjope, G. Joorst, G.T. Stege, M. Mdleleni, Catal. Lett. 78 (2002)
201
[10] P.K. Basu, S.B. Basu, S.K. Mitra, S.S. Dasandhi, S.S. Bhattcherjee, P. Samuel, Stud.
Surf. Sci. Catal. 113 (1998)
[11] J.J. Jothimurugesan, S.K. Spiey, S.K. Gangwal, J.G. Goodwin, Stud. Surf. Sci. Catal.
119 (1998) 215
[12] R. J. O?Brein, L. Xu, R. L. Spicer, S. Bao, D. R. Milburn, B. H. Davis, Catal. Today
36 (1997) 325
[13] D.G. Miller, M. Moskovits, J. Phys. Chem. 92 (1998) 6081
[14] G. Zhao, C. Zhang S. Qin, H. Xiang, Y. Li, J. Mol. Catal. A: Chemical 286 (2008)
137
[15] G. Bian, A. Oonuki, Y. Kobayashi, N. Koizumi, M. Yamada, Appl. Catal. A: Gen.
219 (2001) 13
[16] H.J. Wan, B.S. Wu, Z.C. Tao, T.Z. Li, X. An, H.W. Xiang, Y.W. Li, J. Mol. Catal.
A: Chemical 260 (2006) 255
[17] http://www.chemicalelements.com/elements/cu.html (01 February 2009)
177
Chapter 9
General conclusions
This study describes the effects that copper, potassium, silica and indium have on a
precipitated Fe-based Fischer Tropsch synthesis catalyst.
Chapters 1-3 were introductory chapters contained background material relevant to the
thesis. Chapter 4 described studies in which the optimum weight loadings of Cu and K2O
were investigated. For both the promoters a weight % loading range of 1-5 wt. % was
studied. It was established that the 5 wt. % loading was the optimal weight loading for the
copper promoter since it gave optimal reduction properties as well as the CO adsorption
properties of the Fe catalyst. For the K2O promoter it was concluded that 2 wt. % K2O
loading was also optimal for the Fe-based catalyst. At this loading the reduction
properties and the CO adsorption properties of the Fe-based catalyst were optimal. These
weight loadings were then employed to prepare catalysts containing Cu and K2O.
The effect that silica content had on the Fe-based catalyst was presented in Chapter 5. It
was found that increasing the SiO2 content increased the surface area of the catalyst
which improved the dispersion of the iron oxide crystallites, resulting in a decrease in the
average size of the iron oxide crystallites. But a decrease in the average size of the iron
oxide crystallites strengthened the Fe-SiO2 interaction. An XPS surface analysis of the
catalysts confirmed that as the SiO2 loading was increased, more of the SiO2 stayed on
the surface allowing the Fe-SiO2 interaction to be enhanced. This resulted in the
reduction and carburization ability of the catalyst being suppressed. This affected the FTS
performance of the catalyst and lowered its activity.
In Chapter 6 the effect of silica content on the Fe-based Fischer Tropsch synthesis
catalyst that was promoted with potassium and copper was presented. The conclusions
reached from this study were that silica stabilized the iron oxide crystallites by providing
adequate surface area. This facilitated the high dispersion of Fe2O3 and CuO and
178
enhanced the contact between Fe2O3 and CuO. The enhanced Fe/Cu contact promoted the
reduction of Fe2O3 to Fe3O4, whereas the transformation of Fe3O4 to Fe was suppressed.
Furthermore, due to the K2O-SiO2 interaction, the catalyst loaded with the highest SiO2
loading had a weak contact between Fe and K2O, which reduced the surface basicity of
the catalyst and severely suppressed the CO adsorption.
In the FTS reaction, the FTS activity went through a maximum at 3 wt. % loading of
SiO2 and further addition of SiO2 decreased the catalyst activity. The SiO2 content also
affected the hydrocarbon selectivity. At the highest SiO2 loading, the product distribution
shifted to light hydrocarbons and the C5-C11 hydrocarbons and C2 olefins selectivity were
suppressed.
In Chapter 7 the effect of indium on the Fe-based Fischer Tropsch catalyst was evaluated
and the results obtained were compared to those of the effects caused by copper on the
Fe-based Fischer Tropsch synthesis catalyst. It was found that indium exhibited similar
promotional properties to copper. However indium decreased the reducibility and CO
adsorption ability of the Fe catalyst. It also lowered the FTS activity of the Fe-based
catalyst. It was thus concluded that indium was a poorer promoter for the iron-based FTS
catalyst and acted as a poison for this catalyst.
Finally in Chapter 8 the effect of indium on an Fe-based Fischer Tropsch synthesis
catalyst promoted with potassium and silica was investigated. It was determined that
indium suppressed the reduction properties of the catalyst when H2 was employed as a
reductant, whereas when CO was employed as a reductant, the reduction/carburization
properties were improved. This improved the CO adsorption ability of Fe and resulted in
a selectivity shift to heavy weight hydrocarbons during the FTS reaction whilst low
weight hydrocarbons were suppressed. Adding indium to the promoted catalyst also
lowered the catalyst surface area which resulted in a decrease to the FTS activity.