structure-activity correlations of cu/zno catalysts ... · biodata of the author 130 list of tables...
TRANSCRIPT
Thesis Submitted to the School of Graduate Studies, Universiti Putra Malaysia, in Fulfillment of the Requirements of the Degree of Master of Science
STRUCTURE-ACTIVITY CORRELATIONS OF Cu/ZnO CATALYSTS DERIVED FROM DIFFERENT PRECURSORS
AS FUNCTION OF AGING TIME
ERNEE NORYANA BINTI MUHAMAD
July 2005
Chairman : Associate Professor Dr. Irmawati Ramli, Ph.D. Faculty : Science Abstract
Structural modifications of Cu/ZnO catalysts for methanol steam reforming (MSR) have been investigated comparatively as function of precipitate aging in catalysts preparation process. Freshly precipitated Cu,Zn-hydroxycarbonate precursors (HC) and Cu,Zn-hydroxynitrate pre-cursors (HN) were aged in their mother liquor for 0 and 120 min followed by washing, drying, calcination and reduction. The characteristics of the precursors were determined by means of TG/MS, XRD, and SEM. Generally, the more pronounced aging effect was observed for HC precur-sor (reference catalyst) while no significant effect of aging was observed for HN precursors.
In order to determined the microstructural changes as function of aging, the bulk structure of the Cu/ZnO catalysts was investigated by in-
situ XRD, in-situ XAS, 63Cu NMR, and HRTEM. The observed increase in the activity of the catalysts prepared by HC aging coincides with a decrease in copper crystallite size (respectively an increase in Cu surface area) and an increase in the microstrain in the copper clusters presuma-bly because of the improved interface between Cu and ZnO in comparison to HN prepared catalysts. Aging of the HN precursors results in large, separated and less strained Cu and ZnO particle with an inferior catalytic activity compared to the well known Cu,Zn hydroxycarbonate prepara-tion route.
An increase in catalytic activity of HN and HC was observed significantly after temporary oxygen addition was done to the feed mixture.
The higher catalytic activity does not correlate with an increase in copper surface area, microstrain or oxygen in copper cluster (Cu-EXAFS), but due to slight changes of the catalyst in the medium range order of Cu and ZnO in XAS analysis. Furthermore, the HRTEM and 63Cu NMR inves-tigation revealed that the copper particle get more sintered which resulted to a less interfacial contact of Cu to ZnO was observed after the O2 pulse. Based on these comparison investigations, a structural model of the active catalysts as function of aging was proposed for HN preparation route.
Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)
Structure-Activity Correlations of Cu/ZnO Catalysts Derived from Different Precursors as Function of Aging Time, E.N.B.Muhamad July 2005
Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)
2
Abstrak tesis yang dikemukakan kepada Senat Universiti Putra Malaysia sebagai memenuhi keperluan untuk ijazah Master Sains
PERKAITAN ANTARA STRUKTUR-AKTIVITI MANGKIN Cu/ZnO YANG DIHASILKAN DARIPADA
PREKURSOR YANG BERBEZA SEBAGAI FUNGSI MASA PENUAAN
Oleh
ERNEE NORYANA BINTI MUHAMAD
Julai 2005
Pengerusi : Profesor Madya. Dr. Irmawati Ramli, Ph.D. Fakulti : Sains
Kajian ke atas modifikasi struktur mangkin Cu/ZnO bagi proses pembentukan semula stim metanol (MSR)
telah dilakukan secara perbandingan sebagai fungsi masa penuaan. Mendakan prekursor Cu,Zn-hidroksikarbonat
(HC) dan Cu,Zn-hidroksinitrat (HN) telah dimatangkan di dalam cecair bahan tindakbalas selama 0 dan 120 minit
dan diikuti dengan proses pembasuhan, pengeringan, pengkalsinan dan penurunan. Teknik TG/MS, XRD dan SEM
telah digunakan untuk mencirikan prekursor-prekursor tersebut. Secara keseluruhannya, kesan penuaan yang ketara
telah diperhatikan ke atas prekursor HC manakala tiada perubahan yang sangat ketara dapat diperhatikan dari pre-
kursor HN.
Bagi menentukan kesan modifikasi mikrostruktur sebagai fungsi masa penuaan, struktur pukal mangkin
Cu/ZnO telah dikaji secara in-situ XRD, in-situ XAS, 63Cu NMR dan HRTEM. Hasil kajian menunjukkan peningka-
tan di dalam aktiviti mangkin yang dihasilkan dari prekursor HC adalah sejajar dengan faktor penurunan saiz kristal
kuprum (relatif kepada peningkatan luas permukaan kuprum) dan peningkatan daya mikro regangan di dalam kekisi
kuprum yang disebabkan oleh peningkatan antara muka Cu dan ZnO. Manakala penuaan prekursor HN menghasil-
kan prekursor yang bersaiz besar, terpisah dan kurang daya regangan antara partikel Cu dan ZnO. Ini menyebabkan
kadar aktiviti yang lebih rendah bagi mangkin Cu/ZnO yang dihasilkan melalui proses penuaan prekursor HN ber-
banding HC.
Peningkatan yang ketara di dalam aktiviti mangkin Cu/ZnO bagi HC dan HN dapat diperhatikan selepas
penambahan sementara oksigen ke dalam bahan suapan. Peningkatan aktiviti yang tinggi ini didapati tidak berkaitan
dengan peningkatan luas permukaan kuprum, daya mikro regangan atau oksigen yang terdapat di dalam kekisi
Structure-Activity Correlations of Cu/ZnO Catalysts Derived from Different Precursors as Function of Aging Time, E.N.B.Muhamad July 2005
Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)
3
kuprum (Cu-EXAFS), tetapi adalah disebabkan oleh sedikit perubahan yang berlaku di antara jarak pertengahan Cu
dan ZnO seperti yang diperhatikan di dalam analysis XAS. Tambahan pula, kajian HRTEM dan 63Cu NMR menun-
jukan partikel kuprum menjadi semakin besar yang menyebabkan kurang interaksi antara muka Cu dan ZnO selepas
penambahan sementara oksigen. Oleh itu, berdasarkan kajian ini, satu model struktur bahan mangkin aktif yang di-
hasilkan dari proses penuaan HN telah dicadangkan.
This thesis submitted to the Senate of Universiti Putra Malaysia has been accepted as fulfillment of the requirements for the degree of Master of Science. The members of the Supervisory Committee are as follows: Irmawati Ramli, Ph.D. Associate Professor Department of Chemistry Faculty of Science Universiti Putra Malaysia (Chairman) Abdul Halim Abdullah, Ph.D. Associate Professor Department of Chemistry Faculty of Science Universiti Putra Malaysia (Member)
Taufiq Yap Yun Hin, Ph.D. Associate Professor Department of Chemistry Faculty of Science Universiti Putra Malaysia (Member) Sharifah Bee Abd Hamid, Ph.D. Associate Professor Department of Chemistry Faculty of Science University of Malaya (Member)
1
TABLE OF CONTENTS
Page DEDICATION ii ABSTRACT iii ABSTRAK v ACKNOWLEDGEMENTS vii APPROVAL viii DECLARATION x TABLE OF CONTENTS xi LIST OF TABLES xiii LIST OF FIGURES xiv LIST OF ABBREVIATIONS xvii CHAPTER 1 INTRODUCTION 1.1 1 1.2 5 1.3
Cu-ZnO Catalysts The Hydrogen Production from Methanol Methanol Steam Reforming 7
1.3.1 Mechanism of Methanol Steam Reforming (MSR) 12 1.3.1.1 No Formation of CO in the Reaction Route 12 1.3.1.2 The Decomposition of Methanol and Water Gas Shift (WGS) Reac-
tion 14
1.3.1.3 The Steam Reforming of Methanol and Decomposition of Methanol
14
1.3.1.4 The Steam Reforming of Methanol, Decomposition of Methanol and Water Gas Shift (WGS) Reaction
15
1.3.1.5 Steam Reforming of Methanol and Reverse Water Gas Shift (WGS) Reaction
16
2 LITERATURE REVIEW 2.1 Catalysts Preparation 17 2.1.1 Precipitation and Co-precipitation 17 2.2 Preparation Parameters 20 2.2.1 pH and Temperature 21 2.2.2 Raw Materials 24 2.2.3 Aging 26 2.4 Objectives of Research 31
3 METHODOLOGY 3.1 Materials and Gases 33 3.2 Preparation of Sample 34 3.2.1 Experimental Set-up 34 3.2.2 Catalyst Preparation 35 3.2.3 Samples Treatment Conditions 36 3.2.3.1 Thermal Analysis 36 3.2.3.2 Dynamic Analysis (In-situ method) and Methanol Steam Reforming
(MSR) Process. 37
3.3 Ex-situ Sample Characterizations 39 3.3.1 X-ray Diffraction (XRD) Analysis 39 3.3.2 Combined Thermal Treatment and Mass Spectroscopy Analysis (TG/DSC-MS) 39 3.3.3 X-ray Fluorescence (XRF) Analysis 39 3.3.4 Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray (EDX)
Analysis 40
3.4 In-situ Sample Characterization 40 3.4.1 Transmission Electron Microscopy (TEM) 40 3.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy 41 3.4.3 X-ray Diffraction Analysis (In-situ XRD) 41 3.4.3.1 Bragg-Brentano Diffractometer 42 3.4.3.2 Theta-Theta Diffractometer 42
2
3.4.4 X-ray Absorption Spectroscopy (XAS) 44 3.4.5 N2O decomposition / Reactive Frontal Chromatography (RFC) 45 3.5 Catalytic Activity 45 3.5.1 Methanol Steam Reforming Reaction / O2 pulse 45
4 RESULTS 4.1 Introduction 48 4.2 Precursor Formation 49 4.3 Characterization of Calcined Precursors 57 4.4 Reduction Behavior of CuO/ZnO 67 4.5 Activity (MSR) and Microstructural Investigation of Cu/ZnO Catalysts 87 4.6 Effect of Oxygen Pulse on the Microstructural of Cu/ZnO Catalysts 93
5 DISCUSSIONS 5.1 Precursor Formation and Decomposition 109 5.2 Microstructural and Activity Correlations 111 5.3 The Oxygen Addition (O2 pulse) Investigation 113
6 CONCLUSIONS AND RECOMMENDATIONS 115 BIBLIOGRAPHY 118 APPENDICES 128 BIODATA OF THE AUTHOR 130
LIST OF TABLES
Table Page
1.1 Summaries of some study of methanol steam reforming by related year [22] 10 2.1 Some industrially relevant catalysts and supports obtained by precipitation or co-precipitation
techniques [47] 19
2.2 Phases identified in the precipitate (reference there in 63) 26 4.1 MS data of main mass fragment during calcination of 0HNp and 120HNp 60 4.2 Total mass loss during calcination at different temperature from 300 K to 603 K and from 300
to 873 K in 21 vol % O2/He (6 K/min) 63
4.3 Total mass loss of 0HNc and 120HNc during temperature programmed reduction (TPR) under different reduction potential of 2 vol % H2/He and 15 vol % H2/He (300 K to 523 K, 6 K/min)
72
4.4 The results from NMR fitting procedure of reduced Cu/ZnO catalysts 75 4.5 Comparison in catalytic activity produced by differently prepared precursor under methanol
steam reforming (MSR) condition at 523 K 89
4.6 H2 and CO2 production rate (µmol/g*s) calculated from XAS experiment under methanol steam reforming condition
92
4.7 The comparison results from NMR fitting procedure of 0HN and 120HN samples before and after O2 pulse
96
4.8 Increase in H2 production rate (per gram catalysts) after adding O2 to the feed for about 20 min at 523 K (respect to the initial activity before O2 pulse)
102
LIST OF FIGURES
Figure Page
1.1 Four different routes used for methanol conversion process [22] 6 2.1 Parameters affecting the properties of the precipitate and main properties influenced [47] 21 2.2 Scheme of the preparation procedure for Cu/ZnO catalysts via hydroxycarbonate precursors
[37] 24
2.3 Reaction scheme for the precipitation, aging and subsequent stages in the preparation of 2:1 Cu/Zn catalysts [53]
28
2.4 Propose model of strain induce as function of aging time [18] 30 3.1 Scheme of the experimental set up used for the precipitation 35 3.2 Different stages (A - K) of sample treatments for in-situ experiments (i.e. XRD, XAS, NMR,
RFC and HRTEM) and methanol steam reforming process 37
3.3 Schematic setup of the in-situ XRD-cell in Bragg Brentano geometry 42 3.4 STOE theta-theta diffractometer (in-situ cell and sample holder) 43 3.5 Schematic diagram of the methanol steam reforming set-up 46 4.1 pH curves of precipitation of Cu/Zn nitrate with ammonia hydroxide (HN) and sodium car-
bonate (HC) as function of aging time at constant pH of 7.0 49
4.2 X-ray diffractograms of the dried precursor for different aging time of 0 and 120 minutes of HN (left); and matching diffractograms of different crystallite copper nitrate hydroxide (A-D) phases (right) for 0HNp
51
4.3 3D structure images of (A) gerhardite-monoclinic [ICDD: 15-14], (B) gerhardite-orthorombic [ICDD: 14-687] and (C) gerhardite-monoclinic [ICDD: 45-594]. Source: Inorganic Crystal Structure Database ICSD (Cyan: Cu; Red: O; Blue: N; White: H)
53
4.4 Peak broadening of FWHM due to aging of Cu/Zn hydroxynitrate of different aging precur-sors of 0HNp and 120HNp (refer to the most intense peak of gerhardite phase at 12.8 ° 2θ)
54
4.5 SEM morphology and EDX elemental analysis of unaged precursor (0HNp) under magnifica-tion of (a) 15 000x and (b) 50 000x
55
4.6 SEM morphology and EDX elemental analysis of aged precursor (120HNp) under magnifica-tion of (a) 15 000x and (b) 50 000x
56
4.7 TG/DSC signal during the calcination of 0HNp and 120HNp from 300 K to 603 K (6 K/min) under 21 vol % O2/He
58
4.8 Mass loss (TG) and the main mass fraction (MS) decomposed during the calcination of 0HNp in 21 vol % O2/He (300 K to 603 K, 6 K/min)
59
4.9 Mass loss (TG) and the main mass fraction (MS) decomposed during the calcination of 120HNp in 21 vol % O2/He (300 K to 603 K, 6 K/min)
59
4.10 TG curve (mass loss) of 0HNp and 120HNp at different calcination temperature of 300 K to 603 K and from 300 K to 873 K
63
4.11 X-ray diffractograms of CuO/ZnO under different calcination temperature from 300 K to 603 K (above) and from 300 K to 873 K (below) under 21 vol % H2/He at ramping of 6 K/min
65
4.12 Typical SEM images of CuO/ZnO of 0HNc under magnification of (A) 15 000x, (B) 50 000x, and CuO/ZnO of 120HNc under magnification of (C) 15 000x and (D) 50 000x
66
4.13 The mass loss (TG curve) and DSC signal during the reduction of 0HNc under 2 vol % H2/He (300 K to 523 K; 6 K/min) in comparison with mass fraction (m/z 1, m/z 18, and m/z 44)
67
4.14 The mass loss (TG curve) and DSC signal during the reduction of 120HNc under 2 vol % H2/He (300 K to 523 K; 6 K/min) in comparison with mass fraction (m/z 1, m/z 18, and m/z 44)
68
4.15 TG curve (mass loss) during the reduction of CuO/ZnO under reduction atmosphere of 2 vol % H2/He from 300 K to 523 K (6 K/min), shows the shift in the onset of the reduction from 479 K (0HNc) to 475 K (120HNc)
69
4.16 Onset of the reduction of 0HNc and 120HNc determined by ‘abstract concentration’ of CuO from XANES analysis
71
4.17 Diffractograms of 0HNr and 120HNr (above) with respective Cu [ICDD-PDF-2:4-836] and ZnO [ICDD-PDF-2:36-1451] peaks from 30 ° to 80 ° 2θ (below)
73
4.18 63Cu NMR spectra of differently aged Cu/ZnO catalysts (0HN and 120HN) measured at 4.2 K (Irel = intensity)
74
4.19 NMR peak profile fitting by asymmetric Pseudo-Voigt function for (a) 0HNr and (b) 120HNr implemented from WinXAS 3.1. (Red line is the fitting profile and blue line is the measured NMR pea
75
4.20 Transmission Electron micrographs (HRTEM) of big ZnO plate (A) and electron diffraction pattern of ZnO plate (B) of 0HNr
77
4.21 HRTEM image of 0HNr shows big Cu particle with some lattice fringes (left) 77
3
4.22 HRTEM images of 120HNr after reduction in 2 vol % H2/He under magnification of 300,000 x (right)
77
4.23 In-situ X-ray diffractograms of Cu/ZnO catalysts ( Cu phase; ZnO phase) for 0HNr and 120HNr measured during the reduction at 523 K (above) and after the reduction at 323 K (below)
78
4.24 Copper crystallite size calculated from Scherrer formula of Cu reflexes [111, 200, 220, 311] measured at 523 K during reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
80
4.25 Copper particle size calculated from Scherrer formula of Cu reflexes [111, 200, 220, 311] measured at 323 K (under He atmosphere) after reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
80
4.26 Microstrain of Cu [111, 200, 220, 311] reflexes measured at 523 K during reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
81
4.27 Microstrain of Cu [111, 200, 220, 311] reflexes measured at 323 K (under He flow) after re-duction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
82
4.28 Experimental Fourier-transformed Cu K-edge χ(k) (magnitude) of copper measured at 523 K in 2 vol % H2/He for the activated Cu/ZnO catalysts obtained from 0HNr, 120HNr, 0HCr and 120HCr of the corresponding precursors
83
4.29 Cu Debye-Waller factor of the multiple scattering path obtained by the refinement of theoreti-cal Cu EXAFS to the experimental data of four Cu/ZnO catalysts (0HN, 120HN, 0HC and 120HC) measured at 523 K
85
4.30 Strain indicator from Debye-Waller factor for multiple scattering path at long range order (MS10 - 5.1121 Å)
86
4.31 Fourier transformed experimental Zn K edge χ(k) of the 0HNr, 120HNr, 0HCr and 120HCr at 523 K (2 vol % H2/He)
87
4.32 H2 production rate and CH3OH conversion as function of aging for differently prepared pre-cursor (i.e. 0HNr, 120HNr, 0HCr, 120HCr, and malachite) catalysts
88
4.33 The H2 production (vol %/s) of activity of Cu/ZnO catalysts under Methanol Steam Reform-ing condition carried out in in-situ XAS cell
92
4.34 The H2 production (vol %) during the MSR reaction before (i.e. 0HNr and 120HNr) and after additional O2 pulse (i.e. 0HNo and 120HNo) measured in packed bed reactor at 523 K
94
4.35 Increase in methanol conversion after O2 pulse for 0HNo, 120HNo and 120HCo tested in tubular stainless steel reactor under working condition at 523 K
95
4.36 63Cu NMR measurements of sample (A) 0HN and (B) 120HN before and after the addition of O2 to the feed gas
96
4.37 63Cu NMR spectra of Cu/ZnO catalysts prepared by differently aged HN samples (0HNo and 120HNo) after the oxygen pulse
96
4.38 X-ray diffractograms of 120HN at different stages of treatment (a) calcined (CuO/ZnO) (b) reduced (Cu/ZnO) (c) after temporarily O2 addition and (d) re-reduction in feed at 523 K under working condition
97
4.39 Comparison of Cu [111] crystallite size calculated from Scherrer formula before and after adding O2 to the feed measured at 523 K
99
4.40 Comparison of Cu [111] microstrain before and after adding O2 to the feed measured at 523 K 99 4.41 TEM image of 0HNo after re-reduction in feed under magnification of 13,000 x (left) 101 4.42 TEM image of 120HN after re-reduction in feed under magnification of 9,600 x (right) 101 4.43 Increase in the H2 production (vol %/s) per time on stream after the O2 pulse for 0HN,
120HN, 0HC and 120HC measured by XAS cell at 523 K 102
4.44 Experimental Fourier transformed Cu K-edge χ(k) of differently aged HN and HC samples obtained at 523 K during methanol steam reforming reaction before (1 Feed) and after (2 Feed) O2 pulse
104
4.45 Experimental Fourier transformed Zn K-edge χ(k) of differently aged HN and HC samples obtained at 523 K during methanol steam reforming reaction before (1 Feed) and after (2 Feed) O2 pulse
105
4.46 Fourier transformed experimental Zn K edge χ(k) and imaginary part of the reduced Cu/ZnO derived from hydroxynitrate (HN), hydroxycarbonate (HC) and hydrozincite (HZ) sample recorded at 523 K
106
6.1 Schematic model of HN precipitate as function of aging 116
4
LIST OF ABBREVIATIONS
CH3OH Methanol
Cu/ZnO Copper zinc oxide
DSC Differential Scanning Calorimetric
DWF Debye-Waller factor
EDX Energy Dispersive X-ray
EXAFS Extended X-ray Absorption Fine Structure
FWHM Full width at half maximum
H2O Water
HC Hydroxycarbonate
HN Hydroxynitrate
HRTEM High Resolution Transmission Electron Microscopy
ICDD-PDS International Centre for Diffraction Data – Powder Diffraction Standard
ICSD Inorganic Crystal Structure Database
IUPAC International Union of Pure and Applied Chemistry
MS Mass Spectrometer
MSR Methanol Steam Reforming
NMR Nuclear Magnetic Resonance
RDF Radial Distribution Function
RFC Reactive Frontal Chromatography
SEM Scanning Electron Microscopy
TEM Transmission Electron Microscopy
TG Thermogravimetric
XAS X-ray Absorption Spectroscopy
XRD X-ray Diffraction
XRF X-ray Fluorescence
5
CHAPTER 1
INTRODUCTION
In general, this chapter represent some background of the well known Cu-based catalysts es-
pecially Cu/ZnO which is related to this study. Through the end of this chapter, the application of the
catalysts in hydrogen production particularly in Methanol Steam Reforming (MSR) reaction is de-
scribed.
1.1 Cu-ZnO Catalyst
Copper catalysts are widely used for a variety of selective hydrogenation and dehydrogenation
process and it has been known at least since the 1920’s. For instance, Cu/ZnO catalyst formulation is
well known for low-pressure methanol synthesis and low-temperature water-gas shift reaction (WGS).
Recently a lot of studies discussed its application for the production of hydrogen from methanol by
steam reforming and/or partial oxidation reaction especially for fuel cell application. Cu/ZnO catalysts
also have been used in hydrogenation of carbon monoxide, carbon dioxide, unsaturated hydrocarbons
and certain reactions of amines.
The performance of these catalysts is sensitive to the preparation methods, to the choice of ox-
ide phase used in them and the presence of small amounts of dopants such as alkali and alkaline earth
compounds as well as of Group VIII metal. Most of the published studies reported the use of simple
copper/zinc binary system as the precursors of these catalysts rather than three or four component in
one system. The implication of more components in one system makes the system much more compli-
cated to understand and because of that, most of the extensive publications on these catalysts have con-
centrated on the simple copper/zinc oxide binary system. The incorporation of zinc oxide into the
copper catalyst is of primary importance in making and maintaining a good dispersion of copper metal
crystallites and also prevents the copper particles from sintering [1]. Moreover, the high activity of this
particular system is believed to result from a strong interaction of the two phases (Cu/ZnO) leading to a
specific quality of the active copper material which is a subject under discussion.
6
Regarding that issues, there are some controversies respective to the roles of Cu and ZnO that
makes this system interesting for investigations (i.e. the effects of structural and chemical promotion).
In fact, this is widely documented in several reviews [1-5] which brought into evidences that controver-
sial issues are yet lively. Although the process (e.g. methanol synthesis) involving copper-based cata-
lysts are well established industrially, debates still exist as to:
i. The influence of the preparation method
ii. The role of the reduced copper species on the surface of catalysts
iii. The identification of the active sites
iv. Role of ZnO and Al2O3 in the catalytic process ( as Al2O3 is normally added to
Cu/ZnO catalyst for industry)
Cu-Zn-Al oxide catalysts have attracted great interest in the last decade after first paper pub-
lished by Klier. Klier [2] suggested that Cu is incorporated in the ZnO phase on interstitial and substi-
tutional sites, assuming three possible valence states Cu0, Cu+ and Cu2+. Klier proposals were made
within the framework of bulk defect equilibria based on scanning transmission electron microscopy
(STM), X-ray data and optical spectra [3]. He found that the defect structure and therefore the bulk of
the catalyst determine the catalytic activity.
The formation of Cu+ has also been reported by several authors [6-8]. In particular, Fujitani et
al. [8] in their study on the interaction between support and metal catalyst suggested that the active
component is not only Cu+ but also Cu0. Thus the support may play the role to control the Cu+/Cu0 ratio
on which the catalytic activity depends. Other pronounces support effect was found by Bartley and
Burch [9] when different copper catalysts are tested for the methanol synthesis from both CO/H2 and
CO2/H2 mixtures. In particular, Burch et al. [9,10] and Spencer [11] have proposed that the role of ZnO
is to act as a reservoir for hydrogen and to promote the hydrogen spill-over.
In other point of view, the morphology effect proposed by Yoshihara and Campbell [12],
Ovesen et al. [13], Hadden et al. [14], and Topsøe and Topsøe [15], in which the morphology of copper
7
particles on a ZnO support is responsible for the effect of ZnO upon the methanol synthesis is also a
controversial issue. The view, so far described, is further complicated by the fact that, depending on the
experimental condition and on the catalysts preparation history, the formation of a Cu-Zn alloy may
also occur [16,17]. As a result, this makes the system more complicated to understand, and hence point
a great interest for further investigation.
Recent work carried out by Kniep et al. [18,19] and Günter et al. [20,21] show that the metha-
nol synthesis and methanol steam reforming activity for binary Cu/ZnO catalysts can be related to the
microstrain in copper particles. Extensive in-situ XRD analysis, for determining the microstructural
strain in both Cu and ZnO, clearly indicates that the specific Cu surface area of Cu/ZnO samples alone
cannot unequivocally account for the observed methanol production rates of the systems. Structural
defects of Cu resulting from presence of ZnO in Cu metal, incomplete reduction or epitaxial orientation
to ZnO are believed to cause strain which modifies the Cu surface area and, thus, influence the catalytic
activity.
In contrast from the idea that ZnO also plays an important role in the catalytic activity of
Cu/ZnO catalysts (i.e. methanol synthesis), a contradictory appeared while Chinchen et al. [1,5] in their
study reported that the methanol synthesis reactions occur exclusively on the surface of metallic copper
and ZnO act as carrier to prevent sintering of the copper particles. Therefore, ZnO has no special role
towards copper in the synthesis of methanol.
In summary, the complexity in understanding the synergetic effect between copper and zinc
oxide, the active states of copper and the effects of ZnO are still subjected to some debates, hence a
great interest in research area. The origin for all of these issues are the knowledge-based of the relation-
ships between catalytic activity, surface structure and bulk structure in order to come to rational cata-
lysts design. Therefore, a better understanding of the precursor phases is needed since precursor
structure plays a unique role in determining the interdispersion and the activity of the final catalysts.
Different types of mechanism and conditions have been used in the preparation of catalyst precursor
resulting in the formation of various types of crystalline phases.
8
1.2 The Hydrogen Production From Methanol
Hydrogen (H2) is used in vast quantities in the chemical industry for production of various
bulk, fine and specialty chemicals, in food processing, for fuel production in refineries, in the steel in-
dustry and also directly as a fuel. The largest portion of hydrogen in the world is manufactured at am-
monia production units and consumed on site in the process. Other large consumers are the processes
for methanol and hydrogen peroxide production.
Hydrogen can be produced from both fossil and renewable sources. The largest quantities are
manufactured from natural gas. However, in the future it can be produced by electrolysis of water using
solar energy. In this case it can clearly be viewed as a sustainable source of energy. Hydrogen is the
cleanest fuel available and ideally produces only water during combustion, which makes it an interest-
ing alternative to decrease the anthropogenic emission of carbon dioxide (CO2). The most important
driving force for using hydrogen in automotive applications is the potential of obtaining low emissions
of hazardous compounds. Hydrogen can be used in internal combustion engines or in fuel cell engines.
However, storing hydrogen on board a vehicle poses many concerns regarding safety and han-
dling and can affect customer acceptance in a negative way. Hydrogen can be stored as a compressed
gas at high pressures, as liquid at cryogenic temperature, in metal hydrides or chemically in a hydro-
gen-containing compound (i.e. methanol, ethanol, gasoline, diesel etc.). The final method is superior to
the other three in terms of energy density [22].
Methanol is found to be a favorable candidate since it is the third largest chemical commodity
after ethylene and ammonia with a production capacity in excess of 25 million tons. Methanol is cho-
sen since it is miscible with water, can be activated at low temperatures (473 – 573 K) and high ratio of
hydrogen to carbon (4:1) [23]. Methanol also has an advantage because there are no needs for desulfu-
rization and pre-reforming. Another advantage of using methanol as fuel that should also be taken into
account is that there are many sources available for the production of methanol such as natural gas, oil,
coal and biomass.
9
In principle, there are four routes available for catalytic hydrogen generation from methanol as
shown in Figure 1.1.
Q > 0 (endothermic)
Decomposition
1. 2 H2 + CO CH3OH
3 H2 + CO2
CH3OH H2O
Steam Reforming
2.
Q > 0 (endothermic)
2 H2 + CO2CH3OH
O2
Q < 0 (exothermic)
Partial Oxidation 3.
Combined Reforming at Auto-thermal conditions
2.5 H2 + CO2
CH3OH O2
H2O
4.
Q = 0 (adiabatic condition)
* Q = Heat
Figure 1.1: Four different routes used for methanol conversion process [22]
All four reactions can be carried out at moderate temperatures (473 - 573 K) over transition
metal catalysts, such as copper and palladium. On Cu/Zn-based oxide, these catalysts have received
considerable attention, due to their use in the low-pressure methanol synthesis process.
10
In this study, the catalysts under investigation are mainly used for methanol steam reforming
(MSR) process. The discussion of the MSR reaction over the Cu-based catalysts carried out by other
researchers will be delivered next.
1.3 Methanol Steam Reforming
The production of hydrogen from methanol steam reforming process (MSR) has been exten-
sively studied over supported metallic copper, nickel, palladium, rhodium, and platinum [24]. Kobaya-
shi et al. tried a variety of copper containing mixed oxides, such as Cu-Cr, Cu-Mn, Cu-Zn, Cu-Fe, Cu-
Al, Cu-Ti, Cu-Si, Cu-Ca, Cu-Sn, Cu-Co and Cu-Ni oxides for steam reforming of methanol. It was
shown that mixed oxides were active and selective for the reforming reaction, whereas single oxides
which constituted the mixed oxides were much less active [25].
An extensive study by Takezawa and Iwaza [26] on the behavior of Group 9-10 metals sup-
ported on SiO2 for the steam reforming of methanol revealed that the reaction took place via a different
mechanism compared to those over copper-based catalysts [25]. The difference in the catalytic behav-
ior was ascribed to the difference in reactivity of formaldehyde species. They suggested that formic
acid is formed either directly from formaldehyde or via a methyl formate intermediate before it was
decomposed to H2 and CO2.
However, the traditional Cu/ZnO-based well-known catalyst exhibits a high activity for the
methanol conversion and high selectivity for hydrogen production. In 1999, Breen and Ross [27] per-
formed a study to investigate the influence of zirconia addition to copper-based catalysts. Various types
of catalysts were tested, i.e. Cu-Zn, Cu-Zr, Cu-Zn-Zr, Cu-Zn-Al and Cu-Zn-Zr-Al materials with vary-
ing mass compositions with molar steam to methanol ratio of 1.3:1. The authors found that an increase
in the copper content of Cu/ZrO2 catalysts led to an increase in the conversion and the selectivity of
methanol. This comparison is with the exception of Cu-ZnO catalysts (copper content above 80 wt %)
because the Cu/ZnO catalyst was better than the Cu/ZrO2 catalysts.
Generally, recent studies [18,28-29] focusing on the performance of copper-based catalysts for
methanol steam reforming in flow rector system showed that the major products are H2 and CO2 with
11
minor quantities of carbon monoxide (CO) also being produced. It has been reported, that under favor-
able conditions, the products stream contains up to 75 % H2 and 25 % CO2 on dry basis. Therefore, the
high H2 production capacity is the main argument for choosing the steam reforming process over par-
tial oxidation, which theoretically produces a maximum H2 content of 67 % when pure oxygen is used
and about 40 % H2 if air is used as oxidant [22]. Table 1.1 summaries some related studies on metha-
nol steam reforming and their discovery.
12
Table 1.1: Summaries of some study of methanol steam reforming by related year [22]
Year Authors Catalysts Reaction conditions By-product Description
1982 Takezawa et al. Cu/SiO2 catalyst with a variety of copper loadings
Conventional flow system; H2O/CH3OH = 1:1 in N2 (96 ml/min); 220 °C; atmospheric pressure
HCOOH, CO, CH3OCHO
The effect of copper loading is studied and how it affects the selectivity and conversion for the reforming of methanol.
1985 Takahashi et al. 10 wt %Cu/SiO2, 1 wt % Pt/SiO2
Conventional flow system; 150 – 250 °C; atmospheric pressure
HCOOH, CO, CH3OCHO, HCHO
A study, which describes the different reaction paths for the reforming of methanol over palla-dium and copper-based catalysts.
1993 Iwasa et al. Pd on various supports, Pd con-tent at 1 wt % for all combinations
Conventional flow reactor H2O/CH3OH = 1:1; 0.23 -3.29 s residence time; 200 - 300 °C
CO A detailed study of catalytic effects on numer-ous supports including SiO2, Al2O3, La2O3, Nd2O3 and Nb2O5.
1994 Miyao et al. Cu/Zn/Al alloysof various mass combination
Conventional flow reactor; H2O/CH3OH = 1:1 (0.47 s resi-dence time); 220 °C; atmospheric pressure
CO An investigation of the catalytic behavior of various alloys for the steam reforming reaction.
1994 Iwasa et al. Pd/ZnO, Pd/ZrO2, Pd/SiO2 and Pd. 10 wt % Pd for all combinations
Conventional flow reactor 220 °C; atmospheric pressure; H2O/CH3OH = 1:1 (0.47 s resi-dence time); Re-reduction: 4 % H2/He
HCOOH, CO, CH3OCHO
A detailed study on the effects of the support for palladium-based catalysts.
1997 Shen et al. Various Cu/ZnO catalysts
40 mg catalyst; H2O/CH3OH = 1:1 in N2 (100 ml/min); 140-290 °C; atmospheric pressure; Prereduc-tion: 3 % H2/He; 210 °C; 1 hr
CO A study on the catalytic surface structure of copper-based catalysts and the effects on the selectivity.
1997 Takezawa andIwasa
Supported cop-per and Group 9 and 10 metals
Conventional flow reactor; H2O/CH3OH = 1:1; 0.47 s resi-dence time; atmospheric pressure
HCOOH, CO, CH3OCHO
A study on the difference of the catalytic func-tions of Group 9 and 10 metals and copper. The study also includes dehydrogenation of metha-
* continue
10
13
(Ni, Pt, Pd, Rh) nol. 1999 Breen and Ross Multiple combi-
nations of Cu, Zn, Al, Zr, La and Y
Conventional flow system; 0.1 g catalyst; H2O/CH3OH = 1.3:1 (38.6 ml/min); 140-345 °C; at-mospheric pressure; Prereduction : 5 % H2/N2; 240 °C; 4 hrs
HCOOH, CO, CH3OCHO
A detailed investigation on the performance of copper containing and a suggested reaction route.
2000 Lindstörm and Pet-tersson
Cu/Cr, Cu/Zn and Cu/Zr on Al2O3
Conventional flow reactor; H2O/CH3OH = 1:1; 180- 300 °C; atmospheric pressure; Prereduc-tion: 10 % H2/N2; 180-250 °C; 1 hr
CO A study on the performance of various copper-based catalysts supported on alumina.
2001 Lindstörm and Pet-tersson
Al2O3 -supported Cu/Cr, Cu/Zn, Cu/Zr, Cu/Cr/Zn, Cu/Cr,Zr
Conventional flow reactor; H2O/CH3OH = 1:1; Space veloc-ity: 17 000 h-1; 180-300 °C; at-mospheric pressure; Prereduction: 10 % H2/N2; 180-240 °C; 2 hrs
CO An investigation concerning the performance of binary and ternary alumina-supported copper-based catalysts.
11
14
1.3.1 Mechanism of Methanol Steam Reforming (MSR) 1.3.1 Mechanism of Methanol Steam Reforming (MSR)
There have been some controversies in the literature concerning the mechanisms for produc-
tion of hydrogen and carbon dioxide by steam reforming of methanol. The study of the mechanism of
the formation of CO as a by-product has received high attention. There are several schemes suggested
in the literature:
There have been some controversies in the literature concerning the mechanisms for produc-
tion of hydrogen and carbon dioxide by steam reforming of methanol. The study of the mechanism of
the formation of CO as a by-product has received high attention. There are several schemes suggested
in the literature:
PCO2 PH2 Kp = PCO PH2O
1.3.1.1 No Formation of CO in the Reaction Route 1.3.1.1 No Formation of CO in the Reaction Route
Some researchers [26-27,30] suggested the MSR process proceed via methyl formate
(HCOOCH3) formation, in which no CO takes part in the reaction, as shown in the below reaction
scheme.
Some researchers [26-27,30] suggested the MSR process proceed via methyl formate
(HCOOCH
According to the study performed by Takahashi et al. [30], the water gas-shift (WGS) reaction
was found to be blocked in the presence of methanol (CH3OH) on Cu/SiO2. Another argument for ex-
cluding WGS in the reaction scheme is that the equilibrium constant, , deter-
mined in the experiment greatly
According to the study performed by Takahashi et al. [30], the water gas-shift (WGS) reaction
was found to be blocked in the presence of methanol (CH
3) formation, in which no CO takes part in the reaction, as shown in the below reaction
scheme.
3OH) on Cu/SiO2. Another argument for ex-
cluding WGS in the reaction scheme is that the equilibrium constant, , deter-
mined in the experiment greatly
- H2
- H2
HCOOH H2 + CO2
H2O HCOOCH3 CH3OH + HCOOH
H2O - H2 + CH3OH HCHO + CH3OH H2 + CO2
- H2
exceeded those obtained for the WGS reaction. A detailed study of the reaction scheme on the Cu/SiO2
catalysts has been performed by Takezawa et al. [26]. They found that formaldehyde (HCHO) and
methyl formate (HCOOCH3) are involved in the reaction. By introducing HCHO to the feed of metha-
nol-water mixture, the complete conversion of HCHO to CO2 and H2 was observed. The reaction of
HCHO and water occurred more rapidly as compared to the steam reforming of methanol. Based on
exceeded those obtained for the WGS reaction. A detailed study of the reaction scheme on the Cu/SiO2
catalysts has been performed by Takezawa et al. [26]. They found that formaldehyde (HCHO) and
methyl formate (HCOOCH3) are involved in the reaction. By introducing HCHO to the feed of metha-
nol-water mixture, the complete conversion of HCHO to CO2 and H2 was observed. The reaction of
HCHO and water occurred more rapidly as compared to the steam reforming of methanol. Based on
12
these results they concluded that the production of hydrogen and carbon dioxide over copper based
catalysts includes the formation of HCHO and formic acid (HCOOH) as intermediate products can be
describes as follows:
H2O - H2CH3OH HCOOH H2 + CO2
Furthermore, the rate of formati
presence of CH3OH was determ
presence of CH3OH at the tempe
was estimated to be 20 times hi
formation of HCOOCH3 from t
drogenation of CH3OH to HCO
per based catalysts occurs throug
1.3.1.2 The Decompositio
Methanol decompositio
catalysts have been known as ac
473 to 573 K, producing mostly
shift reaction as shown in the rea
These schemes were su
a commercial low-temperature C
found that CO concentration wa
- H2CH3OH
CH3OH 2 H
CO + H2O
HCHO
+ H2on of HCOOCH3 from the reaction of HCHO in the absence and in the
ined. The rate of HCOOCH3 formation was found to be higher in the
rature from 350 to 450 K. The reaction rate in the presence of CH3OH
gher (at 393 K) than in the absence of CH3OH. This indicates that the
he mixture of HCHO and CH3OH is much more rapid than the dehy-
OCH3. They concluded that the formation of methyl formate over cop-
h a pathway:
n of Methanol and Water Gas Shift (WGS) Reaction
n is the reverse of methanol synthesis from CO and H2. Cu/ZnO-based
tive catalysts for the methanol decomposition at temperatures range of
CO and H2. In some cases, CO2 is probably formed in the water-gas
ction below:
ggested by Santacesaria et al. [31] who studied the MSR kinetics over
u/ZnO/Al2O3 shift catalyst in a continuous stirred-tank reactor. They
s negligible in the product. Based on this result they assumed that CO
HCOOCH3 + H2
CH3OH HCOH
2 + CO ∆H r = 128 kJ mol-1
2 H2 + CO2 ∆H r = - 41.2 kJ mol-1
13
is produced from decomposition of methanol and followed by water gas-shift reaction, where the de-
composition reaction was found to be the rate-determining step. According to this scheme, CO is just
an intermediate product.
1.3.1.3 The Steam Reforming of Methanol and Decomposition of Methanol
A semi-empirical model of the kinetics of the catalyst steam reforming of methanol over
CuO/ZnO/Al2O3 catalyst has been developed by Amphlett et al. [32], by using the reaction schemes of
irreversible reaction of MSR and decomposition reaction as shown below:
Amphlett and his co-workers found that the water gas-shift reaction could be neglected with-
out substantial loss in accuracy. The proposed rate equations for both reactions can be written as:
rCH3OH = - k1 CCH3OH - k2
rH2O = - k1 CCH3OH
rCO2 = k1 CCH3OH
rCO = k2
rH2 = 3 k1 CCH3OH + 2 k2
The reaction rate of methanol and water consumption is depending only on the concentration
of methanol and not on water concentration. Furthermore, the reaction rate of CO formation is a zero-
order rate which means that the formation of CO is not affected by the concentration of methanol or the
concentration of water.
1.3.1.4 The Steam Reforming of Methanol, Decomposition of Methanol and Water Gas
Shift (WGS) Reaction
The following scheme of MSR process which includes MSR, WGS and decomposition is pro-
posed by Peppley et al. [32-33].
14
CH3OH + H2O 3 H2 + CO2 ∆H r = 50 kJ mol-1
CH3OH 2 H2 + CO
CH3OH + H2O 3 H2 + CO2
CH3OH 2 H2 + CO
CO + H2O H2 + CO2
They studied the reaction network for MSR over a Cu/ZnO/Al2O3 catalyst. They claimed that
in order to fully understand the reaction network, all three reactions must be included in the model.
They found that there are two types of catalyst sites that are responsible for the catalyst activity and
selectivity, one for the MSR and WGS reactions and another for the decomposition reaction.
1.3.1.5 Steam Reforming of Methanol and Reverse Water Gas Shift (WGS) Reaction
A kinetic study of methanol steam reforming on a commercial CuO/ZnO/Al2O3 catalyst has
been performed by Purnama et al. [29]. The experimental results of CO partial pressure as function of
contact time at different temperatures showed very clearly that CO was formed as a consecutive prod-
uct. The reaction scheme used is the direct formation of CO2 and H2 by steam reforming reaction and
formation of CO as consecutive product by reverse WGS reaction. A simulation employing this scheme
is able to fit the experimental data well over a wide temperature range (503 – 573 K).
In the work of Breen and co-workers [27], they observed that CO is formed at high methanol
conversions and long contact time. No CO was formed at all at low contact time. This indicates that
CO is a secondary product, formed by reverse WGS reaction. This result agrees well with the work of
Agrell et al. [34]. They found that the level of CO decrease with decreasing contact time.
CH3OH + H2O 3 H2 + CO2
CO2 + H2 H2O + CO
CHAPTER 2
LITERATURE REVIEW
In a research field, it is important to the researcher to have some idea and review of the previ-
ous study related to their research. Thus, in this chapter the importance of the catalysts preparation is
15
represented with some relevant arguments as it is one of the main parts in this study. At the end of this
chapter, the objective of the research is described.
2.1 Catalysts Preparation
The wide number of variables involved in catalysts preparation and the method chosen for
preparing the catalysts has been known to play an important role determining the final properties of the
catalysts. Thus, it is important to choose the best method/technique to synthesize the catalysts and all
parameters involved in the catalysts preparation should be considered to make sure the desired catalysts
is obtained.
2.1.1 Precipitation and Co-precipitation
A variety of methods have been implemented for preparation of Cu/ZnO-based catalysts such
as impregnation [35], co-precipitation in aqueous solution [36-40] or other solvents [41-43], hydro-
thermal synthesis [44], deposition-precipitation and chemical vapor deposition (CVD) [45] as well as
mechanical processes such as milling [46]. However, precipitation is usually more demanding than
other preparation techniques, due to the necessity of product separation after precipitation and the large
volumes of salt-containing solutions generated in precipitation processes. The main advantage of pre-
cipitation process is the possibility of creating very pure materials and the flexibility of the process with
respect to final product quality.
One of the well-known precipitation techniques for mixture of more than one component is
co-precipitation. According to IUPAC nomenclature (2nd edition 1997), co-precipitation is the simulta-
neous precipitation of a normally soluble component with a macro-component from the same solution
by the formation of mixed crystals, by adsorption, occlusion or mechanical entrapment. However, in
catalyst preparation technology, the term is usually used in a more general sense in that the requirement
of one species being soluble is dropped. In many cases, both components to be precipitated are essen-
tially insoluble under precipitation conditions, although their solubility products might differ substan-
16
tially. So, the term co-precipitation generally referred to simultaneous precipitation of more than one
component.
For preparation of active supported metal catalysts, the metal components are generally re-
quired to be dispersed in the precursors of the catalysts at molecular level. Therefore, co-precipitation
is very suitable for the generation of a homogeneous distribution of catalyst components or for the crea-
tion of precursors with a definite stoichiometry, which can be easily converted to the active catalyst.
Such systems prepared by co-precipitation include Ni/Al2O3, Cu/Al2O3, Cu/ZnO and Sn-Sb oxides.
Table 2.1 gives an overview of industrially used precipitated catalysts and supports.
Table 2.1: Some industrially relevant catalysts and supports obtained by precipitation or co-precipitation techniques [47]
Material Use Important examples
Al2O3 (mostly γ, in special cases α or η)
catalyst, support
Claus process, dehydration of alcohols to alkenes and ethers. Hydrotreating catalysts, three-way catalyst
SiO2 Support Noble metal/SiO2 for hydrogenation reactions, Ni/SiO2 for hydrogenation reactions, V2O5/SiO2 for sulfuric acid production
Al2O3/SiO2 Catalyst Acid-catalyzed reactions such as isomerizations Fe2O3 catalyst, catalyst
component Fischer-Tropsch reactions, major component of catalyst for ethylbenzene reaction to styrene
TiO2 support, catalyst, catalyst component
Major component of DeNOx catalyst
ZrO2 Catalyst Acid catalyst after sulfate modification Cu/ZnO Catalyst Methanol synthesis (VO)2P2O7 Catalyst Selective oxidation – for instance butane to maleic
anhydride Cu-Cr oxides Catalyst Combustion reactions, hydrogenations AlPO4 support, catalyst Polymerization, acid-catalyzed reactions Sn-Sb oxide Catalyst Selective oxidation – for instance isobutene to
methacrolein Bi molybdates Catalyst Selective oxidation – for instance propene to acrolein
(mostly supported)
Even though, co-precipitation is call ‘classical route’ for preparation of Cu/ZnO catalysts, it is
found to be the best way to produce Cu/ZnO catalyst that exhibit high performance for methanol syn-
thesis and the water-gas shift reactions [48-52]. Thus, Song et al. [44] investigated several preparation
methods of Cu/Zn/Al catalysts and found the best catalyst was prepared by co-precipitation technique,
which shows high activity for methanol conversion (99 – 100 %) and hydrogen production (71 – 76 %)
in steam reforming and oxidative steam reforming. Furthermore, the first commercial catalyst for the
17
ICI low pressure methanol process were manufactured by continuous precipitation from mixed metal
nitrate (Cu, Zn and Al) and sodium carbonate [53].
2.2 Preparation Parameters
Basically all process parameters, some of which are fixed and some of which are variable, in-
fluence the quality of the final product of the precipitation. It is therefore necessary to optimize the
parameters in order to produce the desired material. Figure 2.1 summarizes the parameters which can
be varied in precipitation processes and the properties which are mainly influenced by these parame-
ters. Since this study is a continuous investigation from previous study done by Kniep et al. [18-19]
and Bems et al. [37] where the pH and temperature was chosen as control parameters, the reasons for
controlling both parameters are justified in the next section based on previous study reported. As addi-
tion, the references on the effect of different precipitating agent (raw material) used in this study and
the aging effect will followed thereafter.
18
Figure 2.1: Parameters affecting the properties of the precipitate and main properties influenced [47]
2.2.1 pH and Temperature
Study on pH and temperature effect on the precipitation was carried out extensively by Li and
Inui [36]. They quoted that different precursor phases were obtained by three different groups of au-
thors although the precipitates were prepared in the same composition and the same precipitating agent
but different pH values. These results clearly show that the pH values have a significant influence on
19
the composition of the precursors hence the activity of the final catalysts. As an addition, the effect of
the temperature in the precipitation was also investigated in their study on copper/zinc aluminium
mixed oxide catalysts for methanol synthesis.
As to investigate the effect of pH and temperature on the precursor phase, a series of precur-
sors with the same agents (i.e. copper/zinc/aluminium nitrate, sodium carbonate) were prepared by co-
precipitation technique at different pH levels (but for each precursor, they were constant). From the
resulting results, Li and Inui concluded that a pH of 7 and the temperature of 343 K was an optimum
condition for the initial co-precipitation procedure. When the precipitation was conducted at pH = 7.0,
the precipitates mainly consisted of the malachite-like phase and the catalysts therefore were more ac-
tive. Precipitation at pH ≤ 6 favored the formation of hydroxynitrate, which led to less active catalysts.
Comparing the activity produced by different precipitation temperature, they observed although the
composition of the precursors obtained at pH = 7.0 and temperature lower than 323 K was the same as
those obtained at pH = 7.0 and T = 343 K, the activity of the catalysts for methanol synthesis was much
lower.
They also concluded that although the effects of both the pH value and the temperature on the
activity of catalyst can be ascribed to the different interdispersion of copper and zinc, it seems that the
pH value exerts its effect through altering the phase composition of the precursors, while the tempera-
ture does so through the precipitation kinetics of the precursors with same compositions. The catalyst
particle size is also affected by the pH; under acidic and alkaline conditions large particles are pro-
duced, but under neutral conditions they are smaller. Spencer stated that the best catalyst is obtained by
co-precipitation at around pH of 7, and this is achieved by a precipitation procedure in which the acidic
and alkaline solutions are mixed continuously [53].
In principle, besides those effect mentioned above each step in the preparation of Cu/ZnO
catalyst for instances precipitation rates, concentration of the solution during the precipitation affects
the microstructural characteristics of the catalyst. However, this issue did not end up here; all of the
processes take place during and/or after the precipitation such as aging, washing, drying, calcining [54-
55] and reducing [55] contributed to the final results of the catalysts. Figure 2.2 illustrates the typical
20
preparation route of the catalyst via the hydroxycarbonate route. Its shows multistep procedure where
each treatment subsequently adds to the structural characteristics of the catalysts finally obtained.
drying
washing
aging
Active catalyst
Controlled precipitation
processing
calcination oxides
precursor
filter cake
precipitate
Cu+2, Zn+2, + CO3
2-
(CuxZn1-x)2(OH)2CO3rosasite
(CuxZn1-x)5(OH)6(CO3)2aurichalcite
Structural evolution & complexity Figure 2.2: Scheme of the preparation procedure for Cu/ZnO catalysts via hydroxycarbonate pre-
cursors [37]
21
2.2.2 Raw Materials
One of the significant points to consider in the preparation of catalysts is the type or choice of
raw material. Normally metal nitrates salt and ammonia or sodium carbonate as precipitating agent
were chosen because it can easily be decomposed to volatile product during heat treatment or calcina-
tion. The choice of raw materials and the composition of the solution are important as they determine
the composition of the final product. Generally, Cu/ZnO catalysts were prepared from mixed (or a mix-
ture of) metal nitrate solution (Cu/Zn molar ratio 2:1) by co-precipitation at 333 K under constant pH
(pH 7.0) [51-52]. Different precipitating agents have been employed for the common precipitation
route, i.e. sodium carbonate [36-39], sodium aluminate [40], ammonia solution [57], ammonium car-
bonate [58] and oxalic acid [41-42,59]. Most commonly, carbonate solution was used as precipitant to
prepare Cu/ZnO catalyst for methanol synthesis and other reactions [17-21,60-62].
Attempt of Sengupta et al. [58] by using ammonium carbonate and ammonium hydroxide as
precipitating agent contributed some ideas on the influence of the precursor phase on the performance
of the catalyst. They claimed that the use of ammonium bicarbonate lead to mostly aurichalcite phase
that is active for water-gas shift reaction in comparison to gerhardite-like phase which was obtained
from ammonium hydroxide. Although, the precipitate precursors did not influence the phase composi-
tion of the calcined or the reduced Cu/ZnO catalysts, their composition determine the interdispersion
and morphology of the final catalyst. Thus, this influences the performance of catalyst in their reaction
study.
Several crystalline phases, known to form in the precipitates, are termed according to their
mineral analogues. Among them are malachite [Cu2(OH)2CO3] which is capable of substituting up to
30 % of Cu with Zn; zincian-malachite [(Cu,Zn)2(OH)2CO3],also known as rosasite, aurichalcite
[(Cu,Zn)5(OH)6(CO3)2], hydrozincite [Zn(OH)6(CO3)2] and gerhardite [Cu2(NO3)(OH)3] [37]. The pre-
cursor produced, either monophasic and/or mixed phases, are according to the catalyst composition (i.e.
ratio between the metal nitrate solutions) and the preparation conditions [37,54]. The identification of
phases in the precipitate has been identified by X-ray diffraction (XRD) and the most important ones
are shown in Table 2.2.
22
Rasmussen et al. [52] in his studies, stated that the maximum activity for the methanol synthe-
sis was observed in an intermediate compositional range (~ 70 mol % Cu) which is industrially interest
whereas, Cu and ZnO alone exhibit only negligible activity. The 2:1 Cu/Zn ratio (67/33 % CuO-ZnO)
also exhibited highly active catalyst in its reduced state [2]. Moreover, many successful commercial
catalysts have rather high copper-to-zinc ratios. For example, ICI: 58 at. -% copper, 27 at. -% zinc, 16
at. -% aluminium, 1 at. -% carbon; Topsøe: 52 - 60 at. -% copper, 21 - 22 at. -% zinc, 15 – 28 at. -%
aluminium, 6 – 8 at. -% carbon [4]. This difference cannot be ascribed to the presence of alumina. As a
structural promoter, which enhances the Cu dispersion and catalyst stability, alumina exerts little effect
on the activity of a reaction.
Table 2.2: Phases identified in the precipitate (reference there in 63)
Chemical formula Name
Cu2(OH)2CO3 Malachite Cu2(OH)3NO3 Gerhardite Zn5(OH)6(CO3)2 Hydrozincite Zn4(OH)6CO3.H2O Zinc hydroxycarbonate Al(OH)3 Aluminium hydroxide (CuxZn1-x)5(OH)6(CO3)2 Copper hydrozincite (x < 0.1) (CuxZn1-x)5(OH)6(CO3)2 Aurichalcite (0.27 < x < 0.45) (Cu1-xZnx)2(OH)2CO3 Zinc malachite (x < 0.3) (Cu1-xZnx)2(OH)2CO3 Rosasite (0.33 < x < 0.5) (CuxZn1-x)16Al2(OH)6CO3.4H2O (Cu, Zn, Al) hydroxycarbonate (CuxZn1-x)16Al2(OH)6CO3.4H2O Roderite
2.2.3 Aging
In this few years, aging effect received great interest in research area. Aging (of precipitate) is
defined as the time-dependent change of those properties of a precipitate, i.e. loss of water, growth of
crystals, recrystallization, decrease of the specific surface, loss of co-precipitated substances, which
generally improve the filtering properties. The process of aging is very often promoted by maintaining
the precipitate and precipitation medium together at elevated temperatures for a period of time. The
terms chemical, physical and thermal aging may be used in cases in which some of the (usually com-
bined) effects named above are to be emphasized specifically (IUPAC 2nd edition 1997).
23
Aging of a precipitate is a process which can have a significant effect on the catalytic activity
of the resulting materials. Figure 2.3 shows the reaction scheme for the precipitation, aging and subse-
quent stages in the preparation of 2:1 Cu/Zn catalysts. During the process of aging the initial precipitate
remains in contact with the precipitating reagents, and the particle can partially redissolve or increase in
size due to Ostwald ripening. According IUPAC Compendium of Chemical Terminology, definition of
Ostwald ripening is the growth of larger crystals from those of smaller size which have a higher solu-
bility than the larger ones. In other word, many small crystal forms in a system initially but slowly dis-
appear except for a few that grow larger, at the expense of the small crystals. The smaller crystals act as
‘nutrients’ for the bigger crystals. As the larger crystals grow, the area around them is depleted of
smaller crystals. In addition, the composition is often reported to change [55].
It has also been known that, in the production of commercial copper zinc oxide catalysts, a pe-
riod of aging is essential. The strong influence of aging of the precursor on the catalytic activity of the
resulting catalyst has been described in the literature. Work by Stone’s group [48] on a simple Cu-ZnO
system, partly overlapping and extending work by Porta’s group [59], showed that aging process led to
both smaller and better distributed Cu and ZnO crystallites in the final catalyst. The catalyst precursor
obtained from a mixed nitrate solution (Cu/Zn molar ratio 2:1) by precipitation at 333 K and pH 7.0
consisted of small crystallites of zincian malachite (Cu/Zn ≈ 85:15), [(Cu/Zn)2CO3(OH)2], and rela-
tively large platelets of aurichalcite [(Cu/Zn)5(CO3)2(OH)6]. Both these hydroxy-carbonates disap-
peared on aging, to give a more finely divided, copper-enriched (Cu/Zn = 2:1) malachite phase. This
monophasic precursor was shown to yield much more active catalysts than the unaged, biphasic precur-
sor.
24
2:1 Cu/Zn nitrate (aq. sol) Na2CO3 (aq. sol)
precipitation
blue zincian georgeite 2:1 Cu/Zn
aging CO2
blue-green low-zincian malachite 85:15 Cu/Zn
+ high-zincian aurichalcite 60:40 Cu/Zn
aging
blue-green high-zincian malachite 2:1 Cu/Zn
drying, calcinations, reduction
Cu/ZnO catalyst
Figure 2.3: Reaction scheme for the precipitation, aging and subsequent stages in the preparation of
2:1 Cu/Zn catalysts [53]
Aging of the precipitated material before filtering and washing was found to be a key step, es-
sential to obtain catalysts of high activity and stability. Whittle and his co-workers [61] studied the per-
formance of Cu/ZnO catalyst for CO oxidation and found that the phase formation changes by
increasing the aging time (0 - 300 min). On aging, the precursor was found to be more crystalline and
formed a characteristic mixture of aurichalcite and rosasite phase. The reflections (XRD diffractogram)
are very broad, particularly at longer aging times, which indicate relatively small crystallites are pre-
sent. They also suggested that there are three morphologies present, observed by using TEM and
HRTEM during the post 30 min aging process: (a) needles (b) platelets and (c) nanoparticles. The in-
crement in the aging time led to smaller crystallite size, therefore higher surface area catalyst is ob-
tained.
25
On the other hand, there is also a report stated that the microstructural characteristics of
Cu/ZnO catalysts can be tailored directly during the preparation, for instance by an appropriate aging
procedure. In the investigation towards rational design of nanostructured copper-zinc oxide catalysts by
Kniep et al. [18], they revealed that there is unexplained linear correlation between copper surface area
and activity as have been reported earlier. They proposed that by increasing aging of precipitate led to
more active catalyst than non-aged one based on the microstructural changes of the active Cu/ZnO
catalyst. The in-situ XRD analysis carried under methanol steam reforming condition shows the Cu
[111] diffraction line reveals a decrease in copper particle size from 110 Å to 70 Å for 0 min and 120
min aging time, respectively.
In addition, an increase in the microstrain in the copper particles with increasing aging time,
from initially 0.09 % after 0 min aging to 0.26 % after 120 min aging, was observed [19]. The NMR
spectra of the catalysts obtained for different aging period (0, 15, 30 and 120 min) also exhibited broad
signal with an asymmetric line profile corresponding to small and strained Cu nanoparticles for the 30
and 120 min samples. From their investigation, it is clearly indicated that the increase of activity in
copper catalysts correlated not only to the decreasing particle size, but also to the increasing mi-
crostrain in the copper nanoparticles. It is also evidence that the structure of the catalyst affects the ac-
tivity more than variations in the chemical composition. Figure 2.4 represent schematic model of strain
induced as function of aging time.
aged unaged precursors precursors
Strain Zn
Cu Cu Cu
ZnO ZnO Interface
ZnO
Larger nanopar-ticles less strai-
ned
Smaller nano-particles more
strained
less more active active
Figure 2.4: Propose model of strain induce as function of aging time [18]
26
However, it should be noted that the optimal aging time will be highly dependent on a wide
range of factors, including temperature, ionic strength, stirring rate, pH and the atmosphere above the
solution. It was found that the aging temperature was an important factor in determining the interdis-
persion and activity of the final catalysts [36]. A study on copper zinc oxide catalysts for carbon mon-
oxide oxidation under different atmosphere shows the atmosphere used during co-precipitation has a
marked influence on the activity of the copper zinc oxide catalysts. Regardless of aging time, the gen-
eral trend in terms of carbon monoxide conversion is air > hydrogen ≈ nitrogen >carbon dioxide. In-
creasing the aging time up to 180 min increased the carbon monoxide conversion over all the copper
zinc oxide catalysts [60]. The catalysts also have surface areas of similar magnitude regardless of the
precipitation atmosphere. However, there is a general increase in surface area as a result of increased
aging time.
In principle, each step in the preparation of Cu/ZnO catalysts via precipitation either pre- or
post precipitation (i.e. aging, washing, drying, calcining, reducing) affects the microstructural charac-
teristics of the final catalysts. However, the corresponding structure-activity relationships are still re-
mained unclear.
2.4 Objectives of Research
Current research in heterogeneous catalysis is devoted to establish structure-activity relation-
ships of real catalysts. Those insights are required in order to arrive at a rational approach in catalysis
which eventually enables a rational catalysts design for more efficient process. The key future for this
issue is better understanding of the system to unravel the ‘chemistry’ between the precursor phases and
the active catalyst.
Therefore a comparative study between well-known catalysts for methanol steam reforming
(MSR) process was investigated by means of the influence of the precipitation parameter. This work
focused on the preparation of Cu/ZnO catalyst prepared by precipitation method with ammonium hy-
droxide. The effect of aging on the catalyst precursor and activity was also investigated as compared to
27
hydroxycarbonate precursor. Several characterization techniques (i.e. XRD, TG/DSC-MS, SEM-EDX,
HRTEM and XAS) were applied to study the correlation between precursor structure and catalyst per-
formance in MSR.
This work was done related to the study of hydroxycarbonate (HC) precursor conducted by
Benjamin L. Kniep (Ph.D student) and other related preceding Ph.D works [18-20,37] in Fritz Haber
Institute (FHI) of Max Plank Gesellschaft (MPG), Berlin Germany with the objectives of;
1. To synthesize Cu/ZnO catalysts through copper hydroxynitrate (HN) route as comparison
to the conventional copper hydroxycarbonate (HC) preparation route.
2. To compare the characteristics of the obtained HN samples with the prepared HC sam-
ples.
3. To observe the catalytic activity of the Cu/ZnO catalysts derived from HN and HC sam-
ples in methanol steam reforming (MSR) reaction.
CHAPTER 3
METHODOLOGY
This chapter consists of the experimental part of this study. The details on the materials, tech-
nique and the instruments used to produce and characterize the catalysts are described under this meth-
odology part.
3.1 Materials and Gases
28
The chemicals used are as follows:
1. Copper (II) nitrate pentahemihydrate, Cu(NO3)2.2.5H2O (98 % - Riedel-de Häen)
2. Zinc nitrate hexahydrate, Zn(NO3)2.6H2O (99 % - Fluka Chimica)
3. Ammonia solution, NH3 solution 25 % (25 % - Merck)
4. Barium sulfate, BaSO4 (97.5 – 100.5 % - Merck)
5. Boron nitride, BN (99.5 % - Alfa Aesar)
6. Methanol p.a.
The gases used are as follows:
1. Hydrogen ( H2 5.0 (99.999 %) - Westfalen)
2. Helium (He 5.0 (99.999 %) - Westfalen)
3. Oxygen (O2 5.0 (99.999 %) - Westfalen)
4. Nitrous oxide, N2O (1 % in He - Westfalen)
3.2 Preparation of the Sample
3.2.1 Experimental Set-up
In the constant pH method, a computer-controlled installation has been used as shown in Fig-
ure 3.1 in order to achieve the highest possible control over the precipitation process. The set-up con-
sists of a thermostated reactor of 2000 ml volume equipped with a wing stirrer and a lid on the top.
Ports are provided with fittings for the pH electrode (single junction, Schott BlueLine 13, temperature
effect compensated by pH-meter, voltage range –1.9 to 1.9 V), thermocouple (Pt-100 resistance ther-
mometer), double holder plastic cap for the metal nitrate and reagent solution. An additional port was
used for measuring the UV-Visible spectrums.
An online computer `Versa Module Eurocard bus´ (VME bus) provides via analog I/O inter-
faces an accurate control over the reaction in the tank. The interfaces are connected to the sensors using
buffer amplifiers and to the pump supply (magnetically coupled toothed wheel pumps from Ismatec). A
real time operating system (VxWorks) allows an isochronous handling of the process parameters. Proc-
29
ess control is realized with a software implemented Proportional Integral Differential (PID) algorithm
which allows processing with a temporal resolution of 100 ms.
NH3 solution Mi n xture of Cu/Z
nitrate solution
pH-electrode
Figure 3.1: Scheme of the experimental set up used for the precipitation
3.2.2 Catalysts Preparation
The copper zinc hydroxynitrate precursors for the Cu/ZnO catalysts were prepared by co-
precipitation method with a molar ratio of copper to zinc of 70:30. The precipitation was carried out in
a reaction container filled with 400 ml bidistilled water, thermostated at 338 K by simultaneous mixing
of an aqueous solution of metal nitrates (Cu(NO3)2.2.5H2O and Zn(NO3)2.6H2O) with ammonia hy-
30
droxide (1.2 M) at constant pH of 7. The solution of the metal nitrates was added at a constant rate of
25 ml/min. The addition of the precipitating reagent was controlled by computer as shown in Figure
3.1.
The resulting precipitates were allowed to age in the mother liquor for 0 min and 120 min un-
der continuous stirring (150 rpm) at the precipitation temperature. After filtering the precipitates were
washed five times with bidistilled water (170 ml – each time) whereby the suspension was stirred for
15 min at ambient temperature. The filter cake was then dried in an oven at 393 K for 20 hours to ob-
tain the dried precursor. About 2 g of the dried precursor was placed in a porcelain crucible and under-
went calcination in a muffle furnace (static air) at 603 K for 180 min with a heating ramp of 6 K/min.
3.2.3 Samples Treatment Conditions
A series of preliminary treatment was conducted on the samples to further characterize by
means of static and dynamic approach. In section 3.2.3.1, the precursors were underwent a series of
thermal analysis (i.e. calcination and reduction) and later was characterized by ex-situ techniques.
Where as the calcined samples which underwent a series of treatment in section 3.2.3.2, were analyzed
dynamically by means of in-situ XRD, XAS, TEM, NMR and RFC method.
3.2.3.1 Thermal Analysis
(a) Calcination conditions:
1. 21 vol % O2/He ; 6K/min ; 300 to 603 K ; TG/DSC/MS detector.
2. 21 vol % O2/He ; 6K/min ; 300 to 873 K ; TG/DSC/MS detector.
(b) Calcination/Reduction conditions:
1. Calcination in 21 vol % O2/He ; 6K/min ; 300 to 603 K ; followed by reduction in 2 vol % H2/He
; 6K/min ; 300 to 523 K ; TG/DSC/MS detector.
31
2. Calcination in 21 vol % O2/He ; 6K/min ; 300 to 603 K ; followed by reduction in 15 vol %
H2/He ; 6K/min ; 300 to 523 K ; TG/DSC/MS detector.
3.2.3.2 Dynamic Analysis (In-situ method) and Methanol Steam Reforming (MSR) Process
K
J
I H
A
G
F
E
D
C
B
RT/He RT / He RT/He
6 K/min 6 K/min
2 vol % H2/He Feed I Feed II O2
pulse 523 K
First half - Reduction Second half – MSR reaction
Part II
Part I
300 K
Figure 3.2: Different stages (A - K) of sample treatments for in-situ experiments (i.e. XRD,
XAS, NMR, RFC and HRTEM) and methanol steam reforming process
Description of the Treatment Conditions:
A - Steady stage at 300 K under 100 vol % He flow of (40 ml/min) for approximately 10 min until the
MS signal stable.
B - Ramping up (6 K/min) from 300 to 523 K under 2 vol % H2/He flow (40 ml/min)
C - Holding for 30 min under 2 vol % H2/He at 523 K (reduction atmosphere)
D - Cooling down (quench) under reduce atmosphere until approximately 353 K and then changing
the atmosphere to fully He (100 vol % He – 40 ml/min) until the temperature reach 300 K.
32
E - Cleaning stage with fully He flow (40 ml/min) at 300 K for 10 min.
F - Ramping up (6 K/min) from 300 to 523 K under Methanol Steam Reforming (MSR) condition
with a total flow of 44 ml/min;
6 ml/min CH3OH
Called as `Feed´ 34 ml/min H2O
4 ml/min He (additional He)
G - First MSR reaction (Feed I) with a total flow of 44 ml/min at 523 K for 20 min.
H - ‘Oxygen pulse’ – 10 vol % O2 (4 ml/min) was purged into the Feed by replacing the additional He
which is switched off during the O2 pulse. The O2 pulse was carried out for 20 min at 523 K.
I - Second MSR reaction (Feed II) – The O2 flow was switched off and the 4 ml/min additional He was
switched on again. The reaction was carried out for 20 min at 523 K.
J - Cooling down (quench) under MSR atmosphere from 523 to 300 K.
K - Cleaning stage under fully He flow 10 min at 300 K.
3.3 Ex-situ Sample Characterization
3.3.1 X-ray Diffraction (XRD) Analysis
The phase composition of the precursors has been determined by X-ray diffraction using a
STOE STADI-P focusing monochromatic transmission diffractometer employing CuKα radiation,
equipped with a Ge (111) monochromator, and a position sensitive detector (linear PSD) at ambient
temperature. The samples were scanned in the 2θ range from 2 to 100 ° with a scanning rate of
0.01°/step. The phase analysis was done with the STOE WinXPow software package (version 1.06,
Stoe Darmstadt, Germany).
3.3.2 Combined Thermal Treatment and Mass Spectroscopy Analysis (TG/DSC-MS)
Several series of thermal analyses were conducted on a NETZSCH STA 449 thermobalance
under constant gas flow of 100 ml/min as described in section 3.3.3.1 earlier. Approximately 10 mg of
33
samples were heated up in an Al2O3 crucible with a heating ramp of 6 K/min. The gases evolved in the
thermal analyses were monitored with a quadrupole mass spectrometer (Pfeiffer QMS 200 Omnistar,
Balzers).
3.3.3 X-ray Fluorescence (XRF) Analysis
The XRF elemental analysis was performed on SEA2010 Bench Top X-ray Fluorescence
Analyzer. The X-rays were generated from rhodium as target metal at 50 kV and a Si(Li) semiconduc-
tor was used as the detector. Approximately, 500 mg sample were placed in the sample holder (mylar
container) and covered with mylar film surface under vacuum condition.
3.3.4 Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray (EDX) Analysis
The surface morphology of the samples was observed with a scanning electron microscope Hitachi –
S4000. For all the samples, the acceleration voltage was set at 15 kV with 10 mm working distance.
The SEM micrographs were recorded at various magnifications (15 000x, 50 000x, and 30 000x) com-
plemented by an Energy dispersive X-ray system (EDX-DX4) for elemental composition analysis.
3.4 In-situ Sample Characterization
3.4.1 Transmission Electron Microscopy (TEM)
Transmission Electron Microscopy (TEM) of the samples was conducted after different stages
of reaction (Part I and II) as described earlier in Figure 3.2. The purpose of measuring TEM after Part I
and Part II is to observe the microstructural changes (if any) after several treatments under different
atmosphere. The TEM analysis was performed on a Philips CM 200 FEG TEM with an electron beam
of 200 kV. The fine resulting samples were pipetted and deposited on holey-carbon films supported on
34
gold grids and transferred to a vacuum transfer holder. The preparation of all TEM samples was done
in a glove box to prevent any contact with air.
3.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy
For NMR investigations, various samples were collected from a fixed bed reactor at room
temperature after different set of reaction steps (Part I and II) as depicted in Figure 3.2. The samples
were transferred into Schlenk flasks in the glove box to prevent any contact with air. Ex-situ Cu NMR
spectra of the Cu/ZnO catalysts were measured with a Bruker MSL 300 Spectrometer at 79.618 MHz
for 63Cu and 85.288 MHz for 65Cu at 4.2 K in an Oxford cryostat. Spin-echo experiments (90°-tau-
180°) were performed with a 90° pulse of 5.5 µs, a ‘recycling delay’ of 2 s, and a tau value of 25 µs.
The 63Cu-NMR spectra shown are calibrated against CuBr (s) at – 381 ppm.
3.4.3 X-ray Diffraction Analysis (In-situ XRD)
For in-situ XRD analysis, the XRD diffractograms were recorded during the reduction stage
(C), after the reduction (E), under Feed 1 (G), during the O2 pulse (H), under Feed II (I) and at final
stage under fully He flow (K) as described in detail at section 3.2.3.2. Two XRD cell was used for the
in-situ XRD analysis as described below.
3.4.3.1 Bragg-Brentano Diffractometer
The in-situ X-ray powder diffraction experiments were conducted using a STOE Bragg-
Brentano diffractometer (CuKα radiation, secondary monochromator, scintillation counter). The in-situ
cell consisted of a Bühler HDK S1 high diffraction chamber mounted onto the goniometer, Bronkhorst
mass flow controllers were employed to control the gas flow. Figure 3.3 shows the schematic diagram
of the Bregg-Brentano set up used. Typically, 25 mg of the catalyst sample were ground and sieved
(200 mesh) and then applied to the sample holder (steel band) in a slurry with acetone.
35
Figure 3.3: Schematic setup of the in-situ XRD-cell in Bragg Brentano geometry
3.4.3.2 Theta-Theta Diffractometer
Combined in situ XRD/MS experiments were performed on a STOE theta/theta diffractometer
with secondary Si [111] monochromator (Cu Kα1 and Kα2 radiation). The diffractometer was
equipped with a XRK900 high temperature cell for in situ measurements. Approximately, 70 mg of the
samples were placed on top of the sample holder (Figure 3.4) and the gas entered the cell over three
inlets, one from the front side, the second from the backside and the third from the bottom of the sam-
ple holder. The bottom of the sample holder was made from a sieve that allows the gas to flow through
the sample. The bottom with its holes was covered with glass microfibre filter to prevent the sample
from blocking the gas flow. This construction provides the entire sample to be in a close contact with
the chosen gas atmosphere. All measurements were conducted under atmospheric pressure in flowing
atmosphere depending on the treatment by means of reduction, MSR or O2 pulse. The gas phase com-
position at the cell outlet was analyzed on line with a mass spectrometer (Quadrupole, Omnistar Pfeif-
fer).
36
Figure 3.4: STOE theta-theta diffractometer (in-situ cell and sample holder)
For these experiments (both in Bragg Brentano and Theta-theta cell), the diffractogram were
recorded at 523 K under methanol steam reforming conditions in a 2θ range of 28.0 ° to 93.0 ° with a
step width of 0.04° 2θ (counting time: 2s/dp). Prior to methanol steam reforming, the catalyst was re-
duced under 2 vol % H2/He (T = 313 to 523 K, heating ramp of 6 K/min) in the same XRD cell. The
XRD diffractogram of the catalysts were recorded during reduction (523 K) and after reduction (323 K
– as the minimum temperature of the detector) using the same scan conditions as for methanol steam
reforming. The microstructure analysis was done by using Powdercell (Version 2.3) software package,
WinXas 3.1, FINDiT and Diamond 2 software.
3.4.4 X-ray Absorption Spectroscopy (XAS)
In-situ XAS experiments were performed in a flow reactor in transmission geometry at the Cu
K-edge (E = 8.979 keV) and the Zn K-edge (E = 9.659 keV) at the Hamburger Synchrotron Radiation
Laboratory (HASYLAB, Beamline E4 and X1). The measurements was carried out in a flow reactor
containing a pellet with a mixture of Cu/ZnO catalyst diluted in boron nitride (ratio Cu/ZnO:BN =
1:10) pressed for 1 min at 1.5 ton. Reduction of the catalysts is performed in 2 vol % H2 in He (1bar,
37
total flow 40 ml/min) in a temperature range from 303 K to 523 K with a heating ramp of 6 K/min. The
software WinXAS 3.1 [64] was used to analyze the XAS data.
After reduction, the extended X–ray absorption fine structure (EXAFS) was monitored in
steam reforming gas mixture at 523 K before and after an additional oxidation-re-reduction cycle. In
the EXAFS data analysis copper foil (respectively Zn foil) as reference was used for energy calibration,
two polynomials for background subtraction and, the atomic background in the EXAFS region was
determined by cubic splines. The radial distribution function was obtained by Fourier transformation of
the k3 weighted experimental spectrum χ(k) in a k range of 2.3 - 10.6 Å-1 for copper and 1.9 - 9.2 Å-1
for Zn. The multiple and single scattering paths of Cu EXAFS spectra were calculated using FEFF 8
software [65] in an R range of 1.5 Å to 6.0 Å (5 % lower limit with respect to the strongest scattering
path). Simulation of the experimental EXAFS was carried out in R space by refining the single scatter-
ing shell distances (R), the Debye Waller factors of the single scattering paths (σ2), and one E0 shift for
all scattering paths. The coordination numbers (CN) were kept constant. In addition, refinement of a
Debye Waller factor of a particular scattering path was used as indicator for structural disorder (strain).
3.4.5 N2O decomposition / Reactive Frontal Chromatography (RFC)
The specific copper surface area was quantified by surface titration with N2O (RFC : N2O +
2Cus N2 + Cus-O-Cus ) [66] after reduction and activation in methanol and water and after tempo-
rary oxygen addition at 523 K (Figure 3.2). Prior to the RFC measurements, the sample was purged for
30 min in He with a flow of 50 ml/min to diminish the amount of adsorbed molecules. The measure-
ments were performed at 313 K with a mixture of 0.5 % N2O/He at a flow of 15 ml/min. The catalysts
were diluted with boron nitride (BN) with ratio of BN to sample of 3:1 to provide a sample bed of
about 15 mm height with a thermocouple directly positioned in the powder bed. The materials were
placed onto a quartz frit in a quartz tube reactor. The gases evolved were measured with a calibrated
quadrupole mass spectrometer (Omnistar, Pfeiffer Vacuum) for quantitative product analysis.
3.5 Catalytic Activity
38
3.5.2 Methanol Steam Reforming Reaction / O2 pulse
Measurements of catalytic activity of methanol steam reforming were performed in a reactor
which consists of separate saturators for methanol (CH3OH) and water (H2O), mass flow controllers for
the measured gases and equipped with a calibrated quadrupole mass spectrometer for quantitative
product analysis. For this set up, two different temperature controllers were used. One is used to heat
the reactor with a heating element and the other one is used for measuring the temperature inside the
catalyst bed. The schematic diagram of the set up is presented in Figure 3.5.
CH3OH
Gases source
He sour-ce
Reactor
saturator
Mass flow controller
Thermo 2
Gas channel
Mass spectrometer
Four wheel panel
PC
H2O
Thermo 1
Figure 3.5: Schematic diagram of the methanol steam reforming set-up
The reaction study performed in this work can be divided into four major segments (i.e. reduc-
tion, Feed I, O2 pulse and Feed II) as previously described in Figure 3.2. First, the catalyst is reduced in
2 vol % H2/He by ramping the temperature at 6 K/min from 300 K up to 523 K and hold at that tem-
perature for 30 min. Later, it was cooled to 300 K in reducing atmosphere followed by purging the He
through the system as a cleaning stage for almost 10 min. Then, the flow was switched to feed gas
(Feed I) and the temperature was increased to 523 K at a linear heating rate of 6 K/min. The composi-
tion of the feed gas used was 2 vol % CH3OH and 2 vol % H2O diluted in He with a total flow of 44
ml/min. In addition to the MSR reaction, the system was purged with 10 vol % O2/Feed as an interme-
39
diate treatment to the catalyst which is called as oxygen pulse (O2 pulse). Immediately, after the O2
pulse was finished the second MSR reaction took over (Feed II). The temperature was kept isothermal
at 523 K for 20 min each for Feed 1, oxygen pulse and Feed II, respectively. The activity measure-
ments were conducted at 523 K under atmospheric pressure in a quartz tube reactor using typically 10
mg sample diluted in 90 mg of boron nitride. Further details on the reaction conditions are as described
in section 3.2.3.2.
CHAPTER 4
RESULTS
The results of the entire experiments conducted on hydroxynitrate (HN) samples, starting from
the precursor formation throughout the catalytic activity of the Cu/ZnO catalysts, are presented in this
chapter. Those results will be compared with the results obtained by other researchers’ whom worked
on the formation of hydroxycarbonate (HC) precursors up to the performance of the Cu/ZnO catalysts
in the Methanol Steam Reforming (MSR) reaction.
4.1 Introduction
A comparative study of the precipitate prepared via precipitation method at constant pH was
investigated as function of different aging time of 0 min and 120 min whereupon all parameters were
kept constant. The samples are denoted according to the scheme t-Xy, where t is the aging time, X is
the precipitating agent that was used in preparing the sample (i.e. ammonia hydroxide (HN) and so-
dium carbonate (HC)) and y is the stage of the samples (i.e. p = precursor, c = calcined/oxide, r = re-
duced and o = after oxygen pulse).
Precipitation of copper and zinc nitrate with a ratio 70:30 with ammonia hydroxide was con-
ducted in this study. Whereas the samples from precipitation of copper and zinc nitrate with sodium
40
carbonate was already prepared by former PhD students and was used in this study as a comparison
[18-19,37]. Besides typical methods (i.e. XRD, TG/DSC-MS, TEM, NMR) used to characterize the
precursor and its oxide, the in-situ technique such as in-situ XRD and XAS were applied to investigate
the microstructural properties of the catalysts. Finally, the activity of the prepared catalyst (HN sam-
ples) in comparison with HC samples for MSR reaction will be investigated and discussed.
4.2 Precursor Formation
Figure 4.1: pH curves of precipitation of Cu/Zn nitrate with ammonia hydroxide (HN) and sodium
carbonate (HC) as function of aging time at constant pH of 7.0
The pH curves of the two different precipitation processes are plotted in Figure 4.1. The pre-
cipitation of Cu/Zn nitrate with ammonia hydroxide shows no change in pH evolution as function of
aging which is contradictory to the precipitation of Cu/Zn nitrate with sodium carbonate. In the case of
hydroxycarbonate (HC) precursor, strong aging effect was found on the precursor phase as depicted in
Figure 4.1. The drop point in pH curve approximately after 22 min aging of the precipitate shows the
transformation point in crystallite phase from initially amorphous precursor (0 min) to crystalline phase
of rosasite and aurichalcite (120 min). This transformation is detected in the XRD diffractogram for HC
sample as reported by Bems et al. [37]. Whereas, no significant changes in pH signal of HN shows
there is no phase transformation happened during the aging process.
0 20 40 60 80 100 120 1406.0
6.5
7.0
7.5
8.0
precipitation process
120 min agingaging started
crystallization point
pH
Time [min]
Cu/Zn hydroxynitrate (HN) Cu/Zn hydroxycarbonate (HC)
41
42
10 20 30 40 50 60 2-th
eta [deg]
Dried precursor
0HNp
120HNp
Inte
nsity
[a.u
]
10 20 30 40 50 600
1000
2000
3000
4000
5000
6000
copper nitrate hydroxide/GerharditeCu2(OH)3NO3orthorhombic
Inte
nsity
0HNp 14-687
10 20 30 40 50 600
1000
2000
3000
4000
5000
6000
copper nitrate hydroxideCu2(OH)3NO3monoclinic
Inte
nsity
0HNp 15-14
10 20 30 40 50 600
1000
2000
3000
4000
5000
6000
copper nitrate hydroxideCu2(NO3)(OH)3monoclinic
Inte
nsity
0HNp 45-594
10 20 30 40 50 600
1000
2000
3000
4000
5000
6000
copper nitrate hydroxideCu2NO2(OH)3monoclinic
Inte
nsity
2-theta [deg]
0HNp 46-1165
D
C
B
A
Figure 4.2: X-ray diffractograms of the dried precursor for different aging time of 0 min and 120
min of HN (left); and matching diffractograms of different crystallite copper nitrate hy-droxide (A-D) phases for 0HNp (right)
After washing and drying were done, the dried HN precursors were analyzed by XRD for
phase identification. The resulting diffractogram of the precursors were matched with several phases
that reasonably could be formed by taking into account all the cations (Zn+2, NH4+) and anions (CO3
-2)
present, especially copper (Cu+2), hydroxide (OH-) and nitrate (NO3-) as the main composition contain-
ing in the precursor phase. From the XRD diffractogram, there is no significant change of the precursor
phase as function of aging. Both samples exhibit similar diffractogram as depicted in Figure 4.2.
As has been reported by Sengupta et al. [58], the precipitation of Cu/Zn nitrate with ammonia
hydroxide led to gerhardite phase. Thus, this idea was used to find significant phases match with the
resulting diffractogram of 0HNp and 120HNp. By using International Centre for Diffraction Data
(ICDD) database provided in STOE software and considering all the main composition containing in
the precursor phase, the diffractogram is considerably match with copper nitrate hydroxide
(Cu2(OH)3NO3) [ICDD-PDF-2: 14-687] or known as gerhardite evident by strong 002-reflection at
12.8 ° 2θ. Other reflexes correspond to the gerhardite phase are 25.5 °, 35.5 °, 41.8 ° and 59.6 ° 2θ.
However, some of the reflexes correspond to gerhardite phase for instance at 37.5 ° and 47.2 ° 2θ are
missing and some unexplained peaks appear at 31.5 °, 32.3 ° and 33.1 ° 2θ . Thus, the precursor phase
which resembling gerhardite is called gerhardite-like phase.
For the HN precursor, the orthorhombic copper nitrate hydroxide was chosen as the reference
for the precursor phase due to many reflexes match with the 2θ value within the whole range of 2 ° to
60 ° 2θ. Furthermore, the reference orthorhombic copper nitrate hydroxide is obtained in much favor-
able condition compared to other references as stated in ICDD-PDF-2 databases. Thus as comparison,
several matches diffractogram for the precursor as shown previously in Figure 4.2 (A-D) was plotted to
shows that the precursor phase is relatively matched with the orthorhombic copper nitrate hydroxide
[ICDD-PDF-2: 14-687] phase compared to others. However, the difference in peak intensity relative to
the ratio of the reflexes is due to the occupation factor and texture effect.
43
There are several copper nitrate hydroxide phases in the ICDD database which have same
stoichiometry as gerhardite [ICDD-PDF-2: 14-687] but different in crystallite system. All the phases
reveal a layer structure arrangement as shown in Figure 4.3 with a ratio 3:1 of OH- to NO3-.
C
A B
Figure 4.3: 3D structure images of (A) gerhardite-monoclinic [ICDD: 15-14], (B) gerhardite-
orthorombic [ICDD: 14-687] and (C) gerhardite-monoclinic [ICDD: 45-594]. Source: Inorganic Crystal Structure Database ICSD (Cyan: Cu; Red: O; Blue: N; White: H)
On the other hand, effort also was put to find any significant ammonia nitrate, NH4NO3 phases
which is considerably could be formed from precipitation of nitrate salt with ammonia hydroxide. Un-
fortunately, there is no NH4NO3 phases match with the precursor diffractogram. Even if that so, it can
not be detected in the XRD due to the limitation error (5 %) of the instrument.
44
It is also found out that the aging of the HN precursor in its mother liquor does not affect any
phase transformation or formation of other different phases except for particle growth. This was ob-
served from the normalization of the most intense peak at 12.8 ° 2θ of the full width at half maximum
(FWHM) of 0HNp and 120HNp precursor as depicted in Figure 4.4. The peak broadening of 0HNp is
wider which relatively led to smaller particle size compared to 120 min aging sample (120HNp).
10 11 12 13 14 15
0.0
0.2
0.4
0.6
0.8
1.0
Nor
m. I
nt [a
.u]
2-Theta [deg]
0HNp 120HNp
Figure 4.4: Peak broadening of FWHM due to aging of Cu/Zn hydroxynitrate of different aging
precursors of 0HNp and 120HNp (refer to the most intense peak of gerhardite phase at 12.8 ° 2θ)
The morphology of the precursor obtained at different aging time was studied by using Scan-
ning Electron Microscopy (SEM) in cooperation with elemental composition analysis by using Energy
Dispersive X-ray (EDX). The overview morphology of the secondary particle measured from SEM
acquired at (a) 15 000x and (b) 50 000x magnification for unaged (0 min) and aged (120 min) HN pre-
cursors are presented in Figure 4.5 and 4.6, respectively.
45
Element Wt % At % CuK 71.97 72.54 ZnK 28.03 27.46
Element Wt % At % CuK 71.97 72.54 ZnK 28.03 27.46
(b) 50 000x (a) 15 000x
Figure 4.5: SEM morphology and EDX elemental analysis of unaged precursor (0HNp) under magnification of (a) 15 000x and (b) 50 000x
15 000x46
Figure 4.6: SEM morphology and EDX elemental analysis of aged precursor (120HNp) under
magnification of (a) 15 000x and (b) 50 000x
Element Wt % At % CuK 67,62 68,24 ZnK 32,38 31,76
Element Wt % At % CuK 67,62 68,24 ZnK 32,38 31,76
Element Wt % At % CuK 67,52 68,14 ZnK 32,48 31,86
Element Wt % At % CuK 67,52 68,14 ZnK 32,48 31,86
(b) 50 000x (a) 15 000x
The SEM images of 0 min aging HN precursor (Figure 4.5) shows mostly needle-like shape
particle was formed inhomogeneously and big agglomerate particle (secondary particle) clearly could
be seen under magnification of 15 000x (Figure 4.5a). At higher magnification of 50 000x, sharp small
edges of the needle-like particle is observable (Figure 4.5b). The ratio between copper and zinc is ap-
proximately 72:28 atomic % with more copper to zinc ratio was detected from EDX analysis.
Prolonging the precursor aging time up to 120 min in its mother liquor caused slight changes
in morphology as shown in Figure 4.6. The secondary particles formed are more homogeneous in dis-
tribution if compared to unaged precursor (Figure 4.6a). By comparing the morphology of the secon-
dary particles of aged (120HNp) and unaged precursor (0HNp), obviously the needle-like shape
particle getting bigger by increasing the aging time. The elemental analysis show the precursor contains
mainly copper and zinc as expected, with an atomic ratio of copper to zinc of approximately 68:32 in
47
percent. The slight difference in copper ratio for longer aging time is probably due to ion exchanged
and re-dissolving copper ion in mother liquor during aging time [75].
4.3 Characterization of Calcined Precursors
Along the line of preparing active catalyst from the precursor, thermal decomposition is ana-
lyzed in detail as it models the calcinations step in the catalyst synthesis. The calcined precursor was
then further investigated in order to get more information about the phase formation and the morphol-
ogy changes during the thermal treatment. Combined technique TG/MS was applied during the calcina-
tions of 0HN and 120HN from 300 K to 603 K in 21 vol % O2/He. Qualitative analysis of the mass loss
curve reveals that all of the impurities decomposed in one single step which also could be seen in the
DSC signal (Figure 4.7). The sharp exothermic DSC signals exhibit that the decomposition of the
0HNp and 120HNp started approximately at 453 K and 458 K, respectively.
350 400 450 500 550 600
65
70
75
80
85
90
95
100
Mas
s lo
ss [%
]
Temperature [K]
TG 0HNp TG 120HNp
-1.0
-0.5
0.0
0.5
1.0
1.5
DSC
(uV
/mg)
DSC 0HNp DSC 120HNp
Figure 4.7: TG/DSC signal during the calcination of 0HNp and 120HNp from 300 K to 603 K (6 K/min) under 21 vol % O2/He
Figure 4.8 and 4.9 illustrated the mass loss from thermal analysis (TG) and main mass fraction
(MS) decomposed during the calcination of 0HNp and 120HNp, respectively. From the figures, both
48
samples exhibit similar decomposition profiles. During calcination in synthetic air, huge water signal
(m/z 17 and m/z 18) and NOx (m/z 30 and m/z 46) were formed. While small traces of ammonia, NH3
(m/z 15) and N2O (m/z 44) were also detectable. Looking closely to the TG curve of 0HNp and
120HNp (Figure 4.7), the decomposition of 120HNp is slightly shifted to the higher temperature com-
pared to the 0HNp. This small difference approximately ~ 5 K, is probably due to the decomposition of
slightly bigger particle of the 120HNp sample coincides with a slightly slower response in the MS sig-
nal as revealed in Figure 4.9.
350 400 450 500 550 600
1E-12
1E-11
1E-10
Ion
curr
ent [
A]
Temperature [K]
m/z 15 m/z 17 m/z 18 m/z 30 m/z 44 m/z 46
65
70
75
80
85
90
95
100
Mas
s lo
ss [%
]
Figure 4.8: Mass loss (TG) and the main mass fraction (MS) decomposed during the calcination
of 0HNp in 21 vol % O2/He (300 K to 603 K, 6 K/min)
49
350 400 450 500 550 600
1E-12
1E-11
1E-10
Ion
Cur
rent
[A]
Temperature [K]
m/z 15 m/z 17 m/z 18 m/z 30 m/z 44 m/z 46
65
70
75
80
85
90
95
100
Mas
s lo
ss [%
]
Figure 4.9: Mass loss (TG) and the main mass fraction (MS) decomposed during the calcination
of 120HNp in 21 vol % O2/He (300 K to 603 K, 6 K/min)
The similar features of the TG curves and DSC signals obtained from the decomposition of
differently aged HN precursors give an indication that the precursor consists of qualitatively one crys-
talline phase as evidenced earlier in XRD analysis. However, looking closely to the MS profile (Figure
4.8 and 4.9), the decomposition of both precursors actually proceeds via three stages shown by 0HNp
(T = 467, 497, 504 K) and 120HNp (T = 478, 500, 506 K). These data are tabulated in Table 4.1.
Table 4.1: MS data of main mass fragment during calcination of 0HNp and 120HNp
Maximum Decomposition Tem-perature (K)
Mass number (m/z)
Key Frag-ments
Probable Par-ent Molecule
(s)
Additional mass number (m/z)
0HNp 120HNp
15 NH+ NH3 14,16,17 497 500 17 OH+ H2O 16, 18 467, 497 478, 500 18 H2O+ H2O 16,17,19,20 467, 497 478, 500 30 NO+ NO2 46,16,14 504 506 44 N2O+ N2O 30,14,16 467, 497 478, 500 46 NO2
+ NO2 30,16,14 504 506
Surprisingly, both samples exhibit the nitrous oxide (N2O) evolution (m/z 44) which is rea-
sonably could not be formed easily, unless there is significant substances such as ammonium nitrate
(NH4NO3) present in the sample. De Soete [66] reported that the decomposition of ammonium nitrate
50
can occurred at 513 K produces nitrous oxide and water. However, the reaction must be carefully con-
trolled to prevent explosive decomposition of the nitrous oxide.
2 NH4NO3 2 N2O + 4 H2O ∆H = -124 kJ/mol [66]
In both cases, there are two maximum peaks of mass fraction (m/z 44) which is referred to the
evolution of H2O at 467 K and 478 K for 0HNp and 120HNp, respectively. This was then followed by
the evolution of N2O occurred at peak maxima 497 K for 0HNp and 500 K for 120HNp. These obser-
vations agreed well with the above reaction. Thus base on these results, assumption was made on the
presence of other solid state phase of NH4NO3 in the precursors. However, as mentioned before due to
the limitation of the XRD analysis, there was no ammonium nitrate phase could be detected. Again,
these results support of the previous discussion that the precursor is not purely gerhardite but resemble
gerhardite phase.
Additional information can be gathered by calculating the total weight loss during calcination.
Supposing that all gerhardite decomposed to cupric oxide, along with ammonia and nitrate (as detected
from the MS signal), the experimental value is assumed to be much higher than the theoretical ones.
However, unexpectedly higher weight loss in theoretical value was obtained from the decomposition of
the HN precursors. This observation supports that the precursor phase resembles gerhardite-like phase.
There could be a new additional solid state phase (e.g. NH4x(Cu0.7,Zn0.32 - x/2)(OH)3NO3), which does
not completely decomposed during the calcination of HN precursors but instead incorporated in the
structure. The incorporated species for instance NH4+ will lead to deviation in stoichiometry ratio of
anions (OH-, NO3-) and cations (Cu2+, Zn2+, NH4
+) of the gerhardite phase. Thus, influence the decom-
position of the precursor and the initial structure of gerhardite.
Besides the instrument detection limit, the new additional solid state phase (if is there so)
which may match with the unexplained peak is not be able to be confirmed due to lack of present XRD
database. Thus the real precursor phase is still uncertain. Nevertheless, taking into account that almost
90 % of the peaks in the XRD diffractogram matched with gerhardite phase, and according to Sengupta
et al. [58] that precipitation of Cu/Zn nitrate with ammonia hydroxide led to gerhardite phase, we will
51
keep refering our precursor as gerhardite-like phase assisted with all the supportive data explained be-
fore.
A further, effort was also put to find any related indicator to support the postulate which has
been made to the precursor phase. As stated in the precursor formation part (section 4.2), the referred
precursor phase (gerhardite) or known as copper nitrate hydroxide (Cu2(OH)3NO3) has a ratio of 3:1 of
OH- to NO3-. Thus from the MS signal given, the integration for the normalized area of m/z 18 (H2O)
and m/z 30 (NO) was done qualitatively. The results show that the ratio of OH- to NO3- was found to be
3:1 for 0HNp and 2.9:1 for 120HNp respectively. Highly similar value of hydroxide and nitrate ratio of
HN precursor with the gerhardite phase fortifies the proposition that has been made for precursor
phase.
Besides calcining the HN samples in 21 vol % O2/He from 300 K to 603 K, additional thermal
treatment was done at relatively higher temperature (300 K to 873 K, 21 vol % O2/He, 6 K/min) to
compare the thermal behavior of HN precursors with HC precursors. Figure 4.10 revealed the mass loss
(TG curve) obtained from both calcination temperatures. From the figure, there is no further decompo-
sition detected after 523 K for HN precursors. This feature shows that the chosen calcination tempera-
ture which ended at 603 K is enough to decompose all the impurities contain in the HN precursor.
Experimentally, the decomposition process of HN precursors finished approximately within 7 min for
0HNp and 8 min for 120HNp after the decomposition started. Table 4.2 presented the total mass loss
calculated from the theoretical and experimental thermal decomposition of HN precursors.
52
Figure 4.10: TG curve (mass loss) of 0HNp and 120HNp at different calcination temperature of
300 K to 603 K and from 300 K to 873 K Table 4.2: Total mass loss during calcination at different temperature from 300 K to 603 K and
from 300 to 873 K in 21 vol % O2/He (6 K/min)
400 500 600 700 80063
72
81
90
99
Mas
s lo
ss [%
]
Temperature [K]
0HNp_603 K 120HNp_603 K 0HNp_873 K 120HNp_873 K
0HNp
120HNp
Mass of catalyst used Calcination
Temperature (K)
Sample name Before (mg) After (mg)
Weight loss ex-perimental per
mg (%)
Theoretical value
(wt %) 0HNp 8.601 5.791 32.67 33.71
603 120HNp 8.964 6.048 32.53 33.26
0HNp 9.617 6.470 32.72 33.71
873 120HNp 14.87 10.110 32.01 33.26
However, this was not happen in the case of hydroxycarbonate samples. As reported by Bems
et al. [37] the calcination of HC precursor from 300 K to 873 K revealed further decomposition at tem-
perature ranges from 713 – 753 K. They found that the high temperature decomposition step is exclu-
sively due to the release of CO2. According to this high temperature resistance of the carbonate species
it will now be referred to as HT-CO3. As these decomposition temperatures are above the calcination
temperatures (603 K) of the precursors, the species causing this feature will persist in activated cata-
lytic systems. The presence of HT-CO3 in HC precursor will results in a modified microstructures not
only of the oxides but also of the reduced catalysts [37]. From this thermal treatment study, it is clearly
53
shows that calcination has a great factor to influence the properties of the produced CuO/ZnO. Still,
however the initial precursor formation is one of the key factor determining the final properties of the
catalysts.
To determine the percentage of copper and zinc in CuO/ZnO of HN samples, the XRF (X-ray
Fluorescence Analysis) analysis was done. The elemental analysis allows the determination of all the
elements and traces (up to limitation of 1 ppm) contained in the samples. The XRF analysis shows that
calcined 0HNc samples contain 73.16 atomic % of Cu and 26.84 atomic % of Zn. The copper contain-
ing in 120HNc oxide is 72.44 atomic % and zinc is 27.56 atomic % respectively.
The oxide form of Cu/Zn derived from HN precursors was send to the XRD for phase identifi-
cation. The resulting diffractogram were found matched with CuO and ZnO of the ICDD database as
depicted in Figure 4.11. This reckons that calcination temperature from 300 K to 603 K is sufficient to
transform the precursor phase into CuO/ZnO. While, an increase of the calcination temperature up to
873 K will lead to highly crystalline and bigger particle size as shown by narrower XRD peaks. The
bigger particle size could be formed as a result from sintering due to high temperature effect. Thus, the
chosen calcination temperature from 300 K to 603 K is efficient to decompose all the impurities and at
the same time able to avoid sintering effect.
54
30 40 50 60 70 80 90 100
Inte
nsity
[a.u
]
CuO [5-661]ZnO [36-1451]
120HNc_603 K
0HNc_603 K
30 40 50 60 70 80 90 100
2-theta [deg]
0HNc_873 K
120HNc_873 K
Figure 4.11: X-ray diffractograms of CuO/ZnO under different calcination temperature from 300
K to 603 K (above) and from 300 K to 873 K (below) under 21 vol % H2/He at ramp-ing of 6 K/min
The morphology of the secondary particle of the calcined oxides was studied by using SEM
under different magnification as shown in Figure 4.12. The 0HNc oxide sample shows a mixture of
inhomogeneously small and big particle of CuO/ZnO (Figure 4.12B). Typically define rod shape parti-
cles with a Cu:Zn ratio of 69:31 atomic % was detected in 0HNc. For the 120 min aging sample, the
oxide formed is more homogeneous in size distribution with Cu:Zn ratio of 67:33 atomic %. Mostly
rectangular and hexagonal shaped particles with a smooth edges could be seen under magnification of
50 000x for 120HNc (Figure 4.12D).
55
(a) 15 000x
A B
(b) 50 000x
(b) 50 000x
D
C
(a) 15 000x
Figure 4.12: Typical SEM images of CuO/ZnO of 0HNc under magnification of (A) 15 000x, (B) 50 000x, and CuO/ZnO of 120HNc under magnification of (C) 15 000x and (D) 50 000x
4.4 Reduction Behavior of CuO/ZnO
The redox behavior of CuO/ZnO from differently aged HN samples was investigated by ther-
mogravimetry analysis under reducing atmosphere of 2 vol % H2 / 98 vol % He. Figure 4.13 and 4.14
shows the TG-DSC/MS signal during the reduction of 0HNc and 120HNc from 300 K to 523 K respec-
tively.
56
0 20 40 60 80
2.00E-009
4.00E-009
6.00E-009
Ion
curr
ent [
A]
Time [min]
m/z 1 m/z 18
5.00E-012
1.00E-011
1.50E-011
Ion
curr
ent [
A] (
m/z
44) m/z 44
0 20 40 60 80
84
86
88
90
92
94
96
98
100
Mas
s lo
ss [%
]
TG
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
DSC
DS
C [u
V/m
g]
0 20 40 60 80
300
350
400
450
500
550
Temp [K]
Tem
pera
ture
[K]
Figure 4.13: The mass loss (TG curve) and DSC signal during the reduction of 0HNc under 2 vol % H2/He (300 K to 523 K; 6 K/min) in comparison with mass fraction (m/z 1, m/z 18, and m/z 44)
57
0 20 40 60 80
5.00E-010
2.00E-009
3.00E-009
4.00E-009
5.00E-009
Ion
curr
ent [
A]
Time [min]
m/z 1 m/z 18
1.00E-012
1.20E-012
1.40E-012
1.60E-012
1.80E-012
2.00E-012
2.20E-012
Ion
curr
ent [
A] (
m/z
44) m/z 44
0 20 40 60 80
84
86
88
90
92
94
96
98
100
102
Mas
s lo
ss [%
]
TG
0.0
0.2
0.4
0.6
0.8
1.0
1.2
DSC
DS
C [u
V/m
g]
0 20 40 60 80
300
350
400
450
500
550
Temp [K]Te
mpe
ratu
re [K
]
Figure 4.14: The mass loss (TG curve) and DSC signal during the reduction of 120HNc under 2
vol % H2/He (300 K to 523 K; 6 K/min) in comparison with mass fraction (m/z 1, m/z 18, and m/z 44)
On the DSC signal, a huge exothermic profile is exhibited by both samples. The signal is co-
incident with the evolution of water (m/z 18). The peak evolution of H2O from the reduction of 0HNc
is rather symmetric compared to 120HNc. The asymmetric profile of H2O evolution during the reduc-
tion of 120HN is caused by a decrease in reduction kinetics due to the isothermic conditions. There is
no distinct evolution of nitrous oxide (m/z 44) could be detected on both samples. A small peak ob-
served in 0HNc can be neglected due to detection error.
The weight loss determined by TG analysis (14.95 wt % for 0HNc and 14.71 wt % for
120HNc) agreed well with the theoretical values of 14.63 wt % for 0HNc and 14.49 wt % for 120HNc,
58
respectively. A slightly lower mass loss of the 120HNc in comparison to the 0HNc is due to a decrease
in CuO content by increasing aging time as evidenced by XRF analysis. The onset temperature of re-
duction for 0HNc started at 479 K and for 120HNc is at 475 K (Figure 4.15). The reduction process of
0HNc and 120HNc was completed approximately in 25 min and 24 min accordingly.
c
c
Figure 4.15: TG curve (mass loss) d
of 2 vol % H2/He from set of the reduction from
The TPR results obtained from
compared with the works carried out by K
samples in hydrogen reduction atmosphe
462 K to 444 K for 120HCc with a differe
in the onset of reduction temperature betw
which the copper crystallite size was 110
that the small particle was found to be an
For further investigated the redu
120HNc) were analyzed by X-ray absorpt
sis of the time resolved XANES spectra
the principle component analysis (PCA).
0HN
120HNuring the reduction of CuO/ZnO under reduction atmosphere 300 K to 523 K (6 K/min), shows the shift in the on- 479 K (0HNc) to 475 K (120HNc)
the reduction of HN samples (i.e. 0HNc and 120HNc) were
niep et al. [19] on the HC samples. The TPR profiles of HC
re showed a shifted in the onset temperature of 0HCc from
nce about 18 K. Kniep et al. reported that the large difference
een these two samples were due to the particle size effect, in
Å for 0HCc and 70 Å for 120HCc. Hence, they concluded
induction factor for the reduction process.
ction behavior of HN samples, both samples (i.e. 0HNc and
ion near edge structure analysis (XANES). Qualitative analy-
of the reduced CuO in 2 vol % H2/He was analyzed by using
The PCA analysis was performed in order to obtain the num-
59
ber of chemical phases and to identify suitable references included in the set of spectra recorded during
the reduction of CuO/ZnO samples [19,28]. From the PCA analysis, three reference components were
detected from the reduction of CuO as explain by Kniep et al. [19]. Figure 4.16 depicted the onset of
the reduction for 0HNc and 120HNc calculated from the ‘abstract concentration’ resulting from the
PCA analysis. As could be seen in Figure 4.16, the onset of the reduction for 120HNc is slightly earlier
than 0HNc which is comparable with the TG analysis.
5 10 15 20 25
0.0
0.2
0.4
0.6
0.8
1.0
abst
ract
con
cent
ratio
n of
CuO
Time [min]
120HNc 0HNc
Figure 4.16: Onset of the reduction of 0HNc and 120HNc determined by ‘abstract concentration’
of CuO from XANES analysis
Another temperature programmed reduction (TPR) experiment was done on both samples, but
this time 15 vol % H2/He was used. All of the reduction conditions were kept the same except for a
higher volume of hydrogen was flowed in. The reason of using higher hydrogen concentration is to
ensure that all Cu2+ are reduced to Cu0. This may be obtained by comparing the total weight loss de-
tected from the TG curve. Results of percentage of weight loss of both samples are tabulated in Table
60
4.3, in which 0HNc recorded 15.00 wt % while 120HNc gave a value of 14.78 wt %. These data were
compared with the works carried using 2 vol % H2/He and they are comparable. Thus, it reveals that
CuO was found reduced even when using 2 vol % H2/He at 300 to 523 K.
Table 4.3: Total mass loss of 0HNc and 120HNc during temperature programmed reduction
(TPR) under different reduction potential of 2 vol % H2/He and 15 vol % H2/He (300 K to 523 K, 6 K/min)
Reduction potential
Experimental value (wt %) Sample name (CuO/ZnO)
2 vol % H2/He 15 vol % H2/He
Theoretical value (wt %)
0HNc 14.95 15.00 14.63 120HNc 14.71 14.78 14.49
Immediately after the reduction process has completed, the reduced sample were sent to XRD
analysis for phase determination. The resulting XRD diffractogram was compared with several copper
phases (i.e. Cu2O and Cu) and ZnO from the database. The diffractogram was found well matched with
Cu [ICDD-PDF-2: 4-836] and ZnO [ICDD-PDF-2: 36-1451]. There is no Cu2O phase being detected
from the XRD analysis after reduction in 2 vol % H2/He. This indicated that CuO (i.e. Cu2+) was totally
reduced to Cu0 in 2 vol % H2/He at 523 K. The reduced Cu/ZnO diffractograms with the respective
phases (i.e. Cu and ZnO) are shown in Figure 4.17. The crystallite size which corresponds to Cu [111]
reflexes is 275 Å and 236 Å for 0HNr and 120HNr, respectively calculated from Scherrer formula [64].
This finding is agreed well with the study done by Kniep et al. [19] for the HC precursor route in which
the copper crystallite size was decreased as a function of aging time.
61
30 40 50 60 70 80
[220][200]
[111]
Inte
nsity
[a.u
]
0HNr 120HNr
30 40 50 60 70 800
20
40
60
80
100
Inte
nsity
[rel
.]
2-Theta [deg]
Cu [ICDD-PDF-2:4-836] ZnO [ICDD-PDF-2:36-1451]
Figure 4.17: Diffractograms of 0HNr and 120HNr (above) with respective Cu [ICDD-PDF-2:4-
836] and ZnO [ICDD-PDF-2:36-1451] peaks from 30 ° to 80 ° 2θ (below)
Ex-situ NMR spectroscopy investigation was performed on the reduced 0HNr and 120HNr in
order to gather more information about the characteristic of the reduced samples. Similar to the X-ray
diffraction line broadening, the peak profile and width of the 63Cu NMR signal is correspond with the
crystallite size and microstrain of the copper particles [69]. Large and highly ordered copper crystallites
result in narrow and symmetric NMR signals. A decrease in copper crystallite size led to wider and
symmetric NMR signals, whereas an increase in the microstrain (disorder of the copper) results in
asymmetric signal profiles or even additional NMR signals [18]. The 63Cu NMR spectra for differently
aged Cu/ZnO are depicted in Figure 4.18.
62
1000 1500 2000 2500 3000
0.0
0.4
0.8
1.2
norm
. I re
l.
δ /ppm
0HNr 120HNr
Figure 4.18: 63Cu NMR spectra of differently aged Cu/ZnO catalysts (0HN and 120HN) measured
at 4.2 K (Irel = intensity)
Two types of fitting procedure were done to characterize the NMR peak profile of the sample.
One is called Integration method and the other one is fitting procedure with Pseudo-Voigt function. The
Integration method was done by smoothing the measured data by using Full Fourier Transform (FFT
fit) to get the maximum point, and the peak was divided into two halves. The integration was done for
both side and the ratio between left and right side was taken to determine the peak profile. On the other
hand, the asymmetric Pseudo-Voigt function was used to determine the NMR peak of reduced 0HN
and 120HN as to compare the results with the Integration method. The background was determined by
polynomial function and the FWHM was calculated. This procedure was implemented in WinXAS 3.1
software [64]. The resulting fitting procedure is presented in Table 4.4.
63
(a) (b)
0.0
5000
10000
15000
0.0
0 2000 40000
Figure 4.19: NMR peak profile fitting by asymmetric Pseudo-Voigt function for (a) 0HNr and
(b) 120HNr implemented from WinXAS 3.1. (Red line is the fitting profile and blue line is the measured NMR peak
Table 4.4: The results from NMR fitting procedure of reduced Cu/ZnO catalysts
Fitting procedures Sample name Integration method Asymmetric Pseudo-Voigt function
0HNr
Chemical shift: 2010 (δ) Ratio: 0.94461
Chemical shift: 2004 (δ) Ratio: 0.92728 FWHM: 374.5
120HNr
Chemical shift: 2006 (δ) Ratio: 0.9188
Chemical shift: 2010 (δ) Ratio: 0.91716 FWHM: 386.5
From the Table 4.4, the ratio between Integration method and Pseudo-Voigt was found nearly
the same. The peak profile of 0HNr and 120HNr shows that large and ordered Cu crystallites were
formed with a ratio lies between 0.9-1.0 which is closer to ideal copper. Nevertheless, from NMR
peaks profile of HN samples, there is some effect (disorder) of copper crystallite but it is not as pro-
nounced as could be seen on HC samples [18]. Furthermore small deviation in ratio between 0HNr and
120HNr is hardly to compare. Still, the NMR investigation is in good agreement with the XRD analy-
sis. The FWHM determined from the Pseudo-Voigt function reveal that the 0HNr have larger particle
size than the 120HNr which is qualitatively matched with calculated XRD particle size of 275 Å
(0HNr) and 236 Å (120HNr) respectively.
Y
X
0.0
2000
4000
6000
0.0
0 2000 40000
Xδ/ppm
YIn
tens
ity
Inte
nsity
δ/ppm
(a) (b)
64
Detail studies on the morphology of the activated catalysts obtained from HN precursors at
different aging time was carried out by using High Resolution Transmission Electron Microscopy
(HRTEM). Generally, large and separate Cu/ZnO particle was observed for both HN samples. Figure
4.20A revealed big ZnO plates which are known to crystallize in a hexagonal lattice formation shows a
well define crystalline form illustrated by distinct and bright spots of electron diffraction pattern (Fig-
ure 4.20B). Separate large and big Cu particle was observed in a range of ~ 50 nm with some lattice
fringes as could be seen in Figure 4.21. Relatively much smaller Cu particle overlapping from one to
another was observed for 120 min aging of HN sample. The high resolution transmission electron mi-
croscopy (HRTEM) image of 120HN shows different orientation and arrangement of lattice fringes as
revealed in Figure 4.22. The overview of the TEM images of Cu/ZnO catalysts derived from 0HNr and
120HNr shows less Cu and ZnO interaction. The high degree of segregation between Cu and ZnO par-
ticle in HN samples results in less strain and less active Cu/ZnO catalysts [20].
B
BA
Figure 4.20: Transmission Electron micrographs (HRTEM) of big ZnO plate (A) and electron
diffraction pattern of ZnO plate (B) of 0HNr
65
Figure 4.21: HRTEM image of 0HNr shows big Cu particle with some lattice fringes (left) Figure 4.22: HRTEM images of 120HNr after reduction in 2 vol % H2/He under magnification of
300,000x (right)
In-situ X-ray diffraction allows the determination of microstructural characteristics (i.e. mi-
crostrain and particle size) of the active Cu/ZnO catalysts by analyzing the line profile of the X-ray
diffraction. Thus, in-situ XRD analysis was performed on CuO/ZnO (i.e. 0HNc and 120HNc) under
reducing atmosphere to determine the microstructural of active Cu/ZnO catalysts (i.e. 0HNr and
120HNr). Figure 4.23 depicted the diffractograms taken by theta-theta diffractometer during the reduc-
tion in 2 vol % H2/He at 523 K. Followed by the diffractograms of 0HNr and 120HNr after the reduc-
tion at 323 K under helium (He) flow. (Note: The minimum temperature of theta-theta diffractometer is
323 K).
66
30 40 50 60 70 80 90
0HNr (523 K) 120HNr (523 K)
30 40 50 60 70 80 90
2-theta [deg]
Inte
nsity
[a.u
]
0HNr (323 K) 120HNr (323 K)
Figure 4.23: In-situ X-ray diffractograms of Cu/ZnO catalysts ( Cu phase; ZnO phase) for 0HNr
and 120HNr measured during the reduction at 523 K (above) and after the reduction at 323 K (below)
From the XRD diffractograms there are only Cu and ZnO phases were detected. The analysis
of the Cu reflexes [111, 200, 220, 311] concerning the copper crystallite size from the theta-theta dif-
fractometer at 523 K and 323 K is represented in Figure 4.24 and 4.25, respectively. A sum of Pseudo
Voigt functions and an appropriate background function were refined according to the Pawley method
(“full pattern refinement”) to analyze the in-situ X-ray powder diffraction diffractograms of the
Cu/ZnO catalysts. Lattice constants of Cu and ZnO, a linear zero shift in 2θ scale, four coefficient of
the background polynomial, the intensities and the Gaussian and Lorentzian part in the profile of the
individual hkl-reflexes were free running parameters in the refinement. The particle size and mi-
crostrain of the copper particles were determined from the Lorentzian and Gaussian part of the individ-
ual profile functions [19].
67
As reported by Kniep et al. [18-19], aging process strongly affects the copper crystallite size
regarding Cu [111] reflexes of HC samples. From the in-situ XRD investigation, they found that by
increasing the aging time led to decrease in copper crystallite size from initially 110 Å (0HCr) to 70 Å
(120HCr). The same effect of aging was observed for HN samples. For HN samples, the copper crys-
tallite size measured at 323 K decrease from initially 275 Å (0HNr) to 236 Å (120HNr). Thus, again it
revealed that the aging process has a great factor influence the Cu crystallite size.
[Å]
50
75
100
125
150
175
200
225
250
275
311220200111
Cu reflexes
Cu
crys
talli
te s
ize
0HNr (523 K) 120HNr (523 K) 0HCr (523 K) 120HCr (523 K)
Figure 4.24: Copper crystallite size calculated from Scherrer formula of Cu reflexes [111, 200,
220, 311] measured at 523 K during reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
68
[Å]
50
100
150
200
250
300
350
Cu reflexes
311220200111
Cu
crys
talli
te s
ize
0HNr (323 K) 120HNr (323 K) 0HCr (323 K) 120HCr (323 K)
Figure 4.25: Copper particle size calculated from Scherrer formula of Cu reflexes [111, 200, 220,
311] measured at 323 K (under He atmosphere) after reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
Apart from Cu crystallite size, the microstrain of copper particle can also be determined from
the XRD peak profile by using Pawley method as described earlier. Figure 4.26 and 4.27 presented the
microstrain determination of Cu reflexes [111, 200, 220, 311] at 523 K and 323 K, respectively. From
both figures, the strain effect on copper (overall trend) can be divided into two groups. As could be
seen in Figure 4.27, higher microstrain (0.33 %) was detected for 120HCr and lower (0.01 – 0.07 %)
strain effect for 0HNr, 120HNr and 0HCr regarding Cu [111] reflexes. These results clearly show that
the effect of aging on microstrain of copper is more pronounced on aged HC sample (120HCr) com-
pared to others. In the case of HN samples, aging of HN samples up to 120 min seems to have no pro-
nounced effect on the copper crystallite size and microstrain effect. Hence, may not greatly influence
the catalytic activity of both Cu/ZnO catalysts.
69
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
311220200111
Stra
in [%
]
Cu reflexes
0HNr (523 K) 120HNr (523 K) 0HCr (523 K) 120HCr (523 K)
Figure 4.26: Microstrain of Cu [111, 200, 220, 311] reflexes measured at 523 K during reduction
in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
0.000.050.100.150.200.250.300.350.400.450.500.550.600.650.70
Cu reflexes
311220200111
Stra
in [%
]
0HNr (323 K) 120HNr (323 K) 0HCr (323 K) 120HCr (323 K)
Figure 4.27: Microstrain of Cu [111, 200, 220, 311] reflexes measured at 323 K (under He flow)
after reduction in 2 vol % H2/He for 0HNr, 120HNr, 0HCr and 120HCr
In order to further analyze the local structure of the Cu/ZnO catalysts, energy absorption ex-
tended fine structure (EXAFS) was applied as an addition to the in-situ XRD experiment. The EXAFS
70
analysis in this section is restricted to a qualitative analysis of the Fourier transformed χ(k) to detect
changes in the local structure with different preparation reagent (i.e. ammonia hydroxide and sodium
carbonate). The EXAFS of the activated catalysts was measured at the Cu K-edge (E = 8.979 keV) and
Zn K-edge (E = 9.659 keV) under reduction conditions at 523 K in 2 vol % H2/He in order to analyze
microstructural changes in the short and medium range order.
[Å]
0 1 2 3 4 5 6
0.00
0.05
0.10
0.15
0.20
0.25
0HCr 0HNr 120HCr 120HNr
FT(χ
(k)*
k3 ]
R
Figure 4.28: Experimental Fourier-transformed Cu K-edge χ(k) (magnitude) of copper measured
at 523 K in 2 vol % H2/He for the activated Cu/ZnO catalysts obtained from 0HNr, 120HNr, 0HCr and 120HCr
The resulting pseudo radial distribution function (RDF) for Cu-EXAFS of all four Cu/ZnO
catalysts (i.e. 0HNr, 120HNr, 0HCr and 120HCr) measured under isothermic conditions is compared in
Figure 4.28. The peaks distances of the plots for Cu K-edge and Zn K-edge are not phase corrected for
the phase shift and hence are shifted to lower R values [21]. The radial distribution function repre-
sented the probability of finding the nearest-neighbors at specific distance, r for specific atom (i.e. Cu-
EXAFS and Zn-EXAFS). Qualitative analysis of the first peak in Figure 4.28 at R = 2.2 Å correspond
to the first Cu-Cu shell, reveal a decrease in amplitude for 0HCr and more or a less the same level of
amplitude for 0HNr, 120HNr and 120HCr.
71
The decrease in amplitude may originate from an increase in the Debye-Waller factor which
respect to the increase in the structural disorder caused by microstrain effect. This is in contrast with
the previous finding [19] where the microstrain of HC samples is increased by increasing the aging
time. The damp in amplitude also can cause by the changes in coordination number (CN) due to the
morphology change in copper nanoparticle [69]. This factor can be excluded because it is normally
works for particle size less than 40 Å which is not the case in the entire samples. The damp in ampli-
tude also can cause by slight shift in the imaginary part of the Fourier transform indicating the presence
of phase mixture (i.e. Cu-Zn alloy formation). Qualitatively fitting procedure of four Cu/ZnO sample
was done and there is no shift in the imaginary part for both HN sample and 120HC, but slightly shift
in 0HC which can caused damping in amplitude as reported by Kniep et al. [19,69] as additional Cu-Zn
alloy phases.
The Debye-Waller factors of the multiple scattering paths for 0HNr, 120HNr, 0HCr and
120HCr sample assuming the copper metal are depicted in Figure 4.29. The increase in the Debye-
Waller factors for 0HCr at short range order of Cu-Cu shell indicates the Cu-Zn alloy formation which
extendedly explains in reference 21. The FT(χ(k)) of the Cu/ZnO sample from HN samples coincide at
distances above 4 Å which correlated with the DWF calculation. Slight deviation in Debye-Waller fac-
tors for 0HCr and 120HCr at higher range order of Cu-Cu shell is difficult to explain because it in-
volves multiple scattering paths underlying the peaks. However, the structural disorder in HC samples
is more pronounced when compared to that of HN samples. This finding is nicely correlated with all
other experimental determination (i.e. in-situ XRD, NMR).
72
1 2 3 40.008
0.010
0.012
0.014
0.016
0.018
DW
F
Cu-Cu shells
0HCr 120HCr 120HNr 0HNr
Figure 4.29: Cu Debye-Waller factor of the multiple scattering path obtained by the refinement of
theoretical Cu EXAFS to the experimental data of four Cu/ZnO catalysts (0HN, 120HN, 0HC and 120HC) measured at 523 K
The theoretical refinement of the experimental Cu FT (χ(k)) at higher range order of Cu-Cu
shells was done for qualitative microstrain determination [19]. The multiple scattering path area at R =
5.1121 Å is taken as the strain indicator and the calculated DWF is represented in Figure 4.30. The
result shows that the 120HCr consist higher strain in copper particle (structural disorder) when com-
pared to that of 0HNr, 120HNr and 0HCr. This is coincides with all the experiment data obtained from
the in-situ XRD and NMR analysis. This result is considered reliable by taking into account the ratio of
the amplitude (46 %) relative to the total amplitude of Cu-Cu distance at R = 2.55 Å [64].
73
0.010
0.012
0.014
0.016
0.018
0.020
0.022
0.024
0.026
0.028
Cu/ZnO catalysts
120HCr
0HNr
120HNr
0HCr
DW
F
Figure 4.30: Strain indicator from Debye-Waller factor for multiple scattering path at long range
order (MS10 - 5.1121 Å)
The theoretical refinement of the experimental Zn FT(χ(k)) of all four Cu/ZnO samples is re-
vealed in Figure 4.31. For Zn K-edge there are three distinct peaks can be observed at ~ 1.5 Å, ~ 2.9 Å
and ~ 4 Å respectively. The first peak of Zn K-edge is attributed to the Zn-O shell and the second peak
are corresponding to Zn-Zn shell. Large deviation in the second amplitude for Zn-Zn shell is clearly
observed for sample prepared from HN and HC. This is maybe an indication of highly disorder of Zn
particle in HC sample rather than in HN sample. There is also deviation in the imaginary part which
apparently observed at higher range. However more detail in the microstructural of Zn is still in pro-
gress.
74
Figure 4.31: Fourier transformed experimental Zn K edge χ(k) of the 0HNr, 120HNr, 0HCr and 120HCr at 523 K (2 vol % H2/He)
4.5 Activity (MSR) and Microstructural Investigation of Cu/ZnO Catalysts
The methanol steam reforming (MSR) activity of all Cu/ZnO catalysts prepared from different
precursor was investigated with packed bed reactor. The calculated methanol conversion and hydrogen
production rate (per gram catalyst) of all samples are depicted in Figure 4.32. Distinct deviation is ob-
served in hydrogen production rate for Cu/ZnO catalysts prepared by sodium carbonate (i.e. 0HCr and
120HCr) compared to ammonia hydroxide (i.e. 0HNr and 120HNr). In the case of HC samples, the
differences in hydrogen production rate and methanol conversion between aged and non-aged catalysts
is apparently observed. Qualitatively, the hydrogen production rate increase about 15 % by prolonging
the aging time of HC samples.
0.0
0.05
0.0
- 0.05
0 2 4 6
R [Å]
120HC 0HC 120HN 0HN
0.0
0.05
0.0
- 0.05
0 2 4 6
R [Å]
120HC 0HC 120HN 0HN
0.0
0.05
0.0
- 0.05
0 2 4 6
R [Å]
0.0
0.05
0.0
120HCr 0HCr 120HNr 0HNr
FT (χ
(k)*
k3
- 0.05
0 2 4 6
R [Å]
75
Figure 4.3 -
On the other hand, slight increase in hydrogen production rate from 45.7 µmol/g*s
(0HNr) to 48.6 µmol/g*s (120HNr) was detected for HN samples. At lower methanol conversion, this
small deviation (~ 6 %) between non-aged and aged sample of HN is difficult to compare. However,
the same trend is observed in both cases. More active catalyst was obtained by increasing the aging
time. MSR reaction also was conducted to unsupported copper catalyst obtained from malachite
[(Cu2(OH)2CO3)] precursor as comparison to the HN samples. The resulting activity data (Table 4.5)
revealed that even though the experimentally determined crystallite size of these three catalysts (i.e.
0HNr, 120HNr, and malachite) is more or less similar, the HN samples were found to be more active
than the malachite. This indicates that not only Cu plays an important role in MSR but also ZnO. ZnO
may not only disperse copper particles and prevent them from sintering [1] but also have a considerably
influence on the microstructure of copper particle which affect the activity [19-20].
10
20
70
80
90
120HCr0HCr120HNr0HNrmalachite
CH
3OH
con
vers
ion
(%)
Cu/ZnO catalysts
CH3OH conversion
20
40
60
120
140
160
180
H2 p
rod.
rate
(µm
ol/g
*s)
H2 prod. rate
2: H2 production rate and CH3OH conversion as function of aging for differently prepared precursor (i.e. 0HNr, 120HNr, 0HCr, 120HCr, and malachite) catalysts
76
Table 4. Comparison in catalytic activity produced by differently prepared precursor under methanol steam reforming (MSR) condition at 523 K
Sample name (Å) sion rate
CO2 production rate (µmol/g*s)
5:
Cu [111] Methanol conver- H2 production
(%) (µmol/g*s) 0HNr 275 22.3 45.7 17.6 0HCr 110 52.6 75.2 139.9
1 20HNr 236 26.7 48.6 17.5 1 70 20HCr 82.2 161.1 63.9
malachite 300 13.9 29.5 10.6
Additional experiment was carried out in the three channel fixed bed reactor [29] as compari-
son to th
-situ XRD experiment allowed determining the copper crystallite size of the active catalysts
under w
(21 m2/g ≡ XRD) after 120 min aging of HC samples.
e above reaction set up. The catalyst powder was diluted in boron nitride (BN), pressed into
pellets which later crushed into smaller particles and sieved to obtain a defined particle size. The
methanol conversion at 523 K of the differently aged catalyst was determined by using a Varian GC
3800. The resulting activity (trend) for 0HNr, 120HNr and 120HCr is qualitatively matched with the
result obtained from the packed bed reactor. Under controlled condition (less error compared to the
packed bed reactor) the methanol conversion calculated from the experiment is 34.5 % (0HNr), 37.3 %
(120HNr) and 41.7 % for 120HCr respectively by diluting the catalyst with BN in a ratio of 1:5. The
same qualitative evolution of the methanol steam reforming activity as function of aging time was
found for the in-situ XRD experiments.
In
orking conditions for the Cu/ZnO catalysts obtained by precipitate aging from different precur-
sor phases. The calculated copper crystallite size is decreased (respectively higher copper surface area)
by increasing the aging time from initially 275 Å (0HNr) to 236 Å (120HNr) and from 110 Å (0HCr) to
70 Å (120HCr). Thus, the resulting copper surface area calculated from the crystallite size assuming
spherical particle (XRD estimation) was compared to copper surface area determined by N2O titration.
Excellent qualitative match was observed from the HC samples which described in details by Kniep et
al. [19]. They found that by increasing aging time leads to an increase copper surface area (respectively
a decrease in copper particle size) from initially 20 m2/g (13 m2/g ≡ XRD) after 0 min aging to 33 m2/g
77
However, reversed result was found for the HN samples. The copper surface area determined
by N2O titration give a value of 1.15 m2/g for 0 min aging (16.64 m2/g ≡ XRD) and 1.08 m2/g after
120 min
HNr = 236
Å) revea that several domain size may formed or sticking together as secondary particle as detected by
SEM. O
t which
flects the activity data. Small changes (error) in the preparation step (e.g. during weighing the sample
and dilu
aging (19.16 m2/g ≡ XRD). Small deviation in copper surface area detected by N2O titration is
not comparable because it is in the range of error (instrument limitation). The large differences in the
surface area detected by N2O titration and the XRD estimation may originate from a small degree of
agglomeration of copper particles (i.e. one particle which exposed to N2O titration may consist of sev-
eral crystallite size where each crystallites are included and calculated in XRD spherical estimation).
This assumption was driven by qualitative analysis of the morphology of the HN samples.
The HRTEM images of apparently large crystallite size (i.e. 0HNr = 275 Å and 120
l
ther possibility is support interactions which reduce the accessible Cu surface area [19]. Fur-
thermore, the N2O titration procedure is not so precise depending on the preparation and the set up
condition. For a better copper surface area determination by using N2O titration, the experiment should
be carried out several times to minimize the error. Nevertheless, there is a qualitative correlation be-
tween copper particle size determine by in-situ XRD and the NMR investigation. Both experimental
data show smaller copper crystallite size was obtained from 120 min aging time of HN sample.
A small difference in copper surface area for HN samples is assumed as critical poin
re
ting with BN) may have affected the result. This was observed in the in-situ XAS experiment
under working condition. The result shows (Table 4.6) reversed activity produced from 0HNr and
120HNr as presented in Figure 4.33.
78
Figure 4.33: The H2 production (vol %/s) of activity of Cu/ZnO catalysts under Methanol Steam
Reforming condition carried out in in-situ XAS cell
However, when comparing the two different precipitate groups, the hydroxycarbonate (0HCr
nd 120HCr) is proven to be more active than the hydroxynitrate (0HNr and 120HNr). A small error in
preparin
mol/g*s) calculated from XAS experiment under methanol steam reforming condition
Sample name Mass of catalyst H production rate CO production rate
a
g the HC samples is not affecting the total activity of the catalyst because of the distinct devia-
tion between both sample (0HCr and 120HCr).
Table 4.6: H2 and CO2 production rate (µ
(mg) 2
(µmol/g*s) 2(µmol/g*s)
0HNr 3.26 2.55 7.12 120HNr 2.51 5.58 2.56
0HCr 2.53 53.7 13.3 120HCr 2.52 79.0 19.2
.6 Effect of Oxygen Pulse on the Microstructural of Cu/ZnO Catalysts
copper-zinc catalysts
repared from HN samples was investigated due to promising feature of O2 addition to the catalysts
0 20 40 60 800.0
0.2
0.4
0.6
0.8
1.0
H2 [v
ol%
/s]
Time [min]
0HNr 120HNr 0HCr 120HCr
50 %
120HNr 0HNr
14 %
4
The effect of O2 addition during the steam reforming of methanol over
p
79
perform
ctivity of the Cu/ZnO
catalysts repared from HN samples as presented in Figure 4.34. After temporarily oxygen addition (10
vol %), t
igure 4.34: The H2 production (vol %) during the MSR reaction before (i.e. 0HNr and 120HNr) and after additional O2 pulse (i.e. 0HNo and 120HNo) measured in packed bed reac-tor at 523 K
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
ance as have been reported by several authors [28,70-72]. The effect of temporarily oxygen
addition followed by reduction in feed to the activity and microstructural of Cu/ZnO catalysts was in-
vestigated by means of TEM, NMR and XAS. Before further investigation using the in-situ technique
(i.e. in-situ XRD and XAS) one catalytic test in the packed bed reactor was done for Cu/ZnO catalysts
of HN samples. This experiment was carried out to get an idea if there is any significant change in ac-
tivity of HN samples after O2 pulse as reported by Günter [72] for HC samples.
The resulting activity shows promising effect of the O2 addition on the a
p
he hydrogen production increase about 21.7 vol % for 0HNo and 23.3 vol % for 120HNo. The
calculated hydrogen production rate normalized per gram catalysts shows an increment from initially
45.7 µmol/g*s to 55.6 µmol/g*s (0HNo) and from 48.6 µmol/g*s to 59.9 µmol/g*s (120HNo) respec-
tively after 20 min adding oxygen to the feed.
O2 pulse After O2 pulse
Before O2 pulse
H2 co
ncen
tratio
n [%
]
Cycle
0HN 120HN
300
350
400
450
500
Tem
pera
ture
[K]
Tempearture [K]
0 200 400 600 800 1000 1200 1400 16000
2
4
10
6
8
O2
conc
entra
tion
[vol
%]
O2 pulse
F
80
Additio
eel reactor for both HN samples and aged 120 min HC samples. The resulting activity from this reac-
tor set u
Figure 4. 5: sted in
The effects of O2 pulse to the Cu/ZnO catalysts prepared from HN samples were further inves-
gated by NMR analysis. Figure 4.36 presented the normalized 63Cu NMR signal of 0HNr and 120HNr
(before)
0 5 10 15 20
0.35
0.40
0.45
0.50
0.55
nal experiment under more controlled condition was conducted in tubular stainless
st
p is depicted in Figure 4.35. Upon addition of oxygen to the feed gas and followed by the re-
reduction process in MSR, an increase in catalytic activity was observed for those three samples. The
methanol conversion increase approximately 38 % for 0HN, 28 % for 120HN and 32 % for 120HC.
Generally, the addition of oxygen to the feed gas during the reaction study enhanced the catalytic activ-
ity of Cu/ZnO catalysts of HN and HC samples.
0.60
After O2 pulse
Before O2 pulse
CH
3OH
con
vers
ion
[%]
Time [min]
0HN 120HN 120HC
3 Increase in methanol conversion after O2 pulse for 0HN, 120HN and 120HC tetubular stainless steel reactor under working condition at 523 K
ti
and 0HNo and 120HNo (after) the O2 pulse. The peak profile of the NMR analysis shown in
Figure 4.36 revealed that the Cu particles become more sinter after the addition of oxygen for about 20
min on stream. Analysis of the individual NMR peak profile (as described in section 4.4) shows that
the sintering effect on Cu particle is more pronounced for 120HNo compared to that of 0HNo (Figure
4.37). The resulting fitting procedure revealed that the copper particle become more ideal after the O2
81
pulse as shown in Table 4.7. The increment in calculated ratio value (nearly 1.0) from both fitting pro-
cedure reveal that the Cu particle becomes more ordered (ideal) after O2 pulse. The narrow peak broad-
ening determined from FWHM shows the sintering effect take place after the additional O2 pulse
(respectively to bigger particle size).
igure 4.36: 63Cu NMR measurements of sample (A) 0HN and (B) 120HN before and after the
addition of oxygen to the feed gas
igure 4. d HN samples
1500 2000 2500 3000
0.0
0.2
0.4
0.6
0.8
F
1.0
F 37: 63Cu NMR spectra of Cu/ZnO catalysts prepared by differently age
(0HNo and 120HNo) after the oxygen pulse
norm
. I re
l
δ [ppm]
0HNo 120HNo
1500 2000 2500
0.0
0.5
1.0
norm
. I re
l
δ [ppm]
120HNr (before O)2 120HNo (after O2)
1500 2000 2500
0.0
0.5
1.0
norm
. I re
l
δ [ppm]
0HNr (before O)2 0HNo (after O2)A B
82
Table 4.7: The comparison results from NMR fitting procedure of 0HN and 120HN samples before and after O2 pulse
0HN 120HN
Fitting procedures / pa rameters Before After Before After
Chemical shift (δ) 2010 2010 2006 2007 Integration method Ratio 0.94461 1.02908 0.9188 0.98413
Chemical shift (δ) 2004 2094 2010 2083 Ratio 0 10.92728 .98441 0.91716 .05269
AssymePseudo-Voigt
3
tric
function FWHM 374.5 37.35 386.5 257.9
A dynamic experi procedure w out fo e samp r to in
gate the structural changes of Cu/ZnO catalysts prepared from differently aged HN and HC samples. A
program
igure d ) re-
ctogram indicate that CuO is converted to metallic copper after it was reduced in 2
ol % H2/He with no additional copper oxide phases was detectable. The addition of 10 vol % O2 to the
30 35 40 45 50 55 60
mental as carried r all of th les in orde vesti-
med in-situ XRD analysis was conducted in Bregg Brentano diffractometer (Figure 3.3) to take
the X-ray diffractogram (scan) at each stage under working condition as described earlier in section
3.2.3.2. The X-ray diffractograms of all stages (i.e. calcined precursor, reduced Cu/ZnO, O2 pulse, re-
reduction in MSR) were taken at 523 K except for the calcined sample which is at 323 K under He at-
mosphere. For instance, Figure 4.38 presented the X-ray diffractograms of 120HN obtained (a) after
calcination, (b) reduction (c) oxidation, and (d) re-reduction in feed.
(d) re-reduction
(c) O2 pulse
(b) reduced Cu/ZnO
(a) calcined precursor
Cu [111]
Cu2O [200]
Cu2O [111]
*
[101]
[100] [002]
ZnO
* CuO [-111]
CuO [200]
Inte
nsity
[a.u
]
2-theta [deg]
F 4.38: X-ray diffractograms of 120HN at different stages of treatment (a) calcine
(CuO/ZnO) (b) reduced (Cu/ZnO) (c) after temporarily O2 addition and (dreduction in feed at 523 K under working condition
The diffra
v
83
feed at 5
alitative analysis of the
ffect of O2 pulse on copper crystallite size and strain was done by using the same fitting procedure as
escribe
igure 4.39 before and
23 K for 20 minutes resulted in the formation of minor amount of Cu2O which slightly ap-
peared at 36.4 ° and 42.3 ° 2θ. However, no CuO phase was detectable (respect to the 5 % XRD limita-
tion). After temporarily O2 pulse, the sample underwent re-reduction in feed at 523 K. The XRD
diffractogram of re-reduce Cu/ZnO catalysts shows that metallic copper can be seen clearly evidenced
by the presence of Cu [111] reflexes. Slight decrease in peak width (FHWM) of Cu [111] reflexes cor-
respond to the first reduction process in H2 and after re-reduction process in feed, indicates that the Cu
crystallite size become more sintered after the oxidation-re-reduction process.
In addition to the in-situ diffractogram recorded for each samples, qu
e
d d earlier in section 4.4. The Cu crystallite size determination by means of Scherrer formula was
depicted in Figure 4.39. The general trend shows that the size of crystallite Cu increases after the O2
pulse for all of the samples. The sintering effect in HN samples is more pronounced compared to the
HC samples. The increase in the size (respective lower surface area) after addition of O2 to the feed is
deemed to cause the failure in the surface area detection by using the N2O titration after the O2 pulse.
This was observed especially for the large Cu particle as in HN samples. (Note: Additional N2O titra-
tion experiment was conducted separately in packed bed reactor for the surface area determination after
O2 pulse for 0HN and 120HN. Unfortunately no N2O titration was done for the HC samples and the
resulting copper surface area determination for the HN samples after the O2 pulse was failed).
350
400
F : Comparison of Cu [111] crystallite size calculated from Scherrer formula
after adding O2 to the feed measured at 523 K
50
100
150
200
250
300
Cu
cr s
ize
After O2 pulseBefore O2 pulse
Cu [111]
0HN 120HN 0HC 120HC
[Å]
ysta
llite
84
igure 4.40 measured
Figure 4.40 the microstrain analysis after temporarily O2 addition to the feed for all HN and
HC sam es are displayed. As could be seen, the O2 pulse treatment resulted in less strained copper
except f
o and 120HNo after underwent several treatments in feed
I, O2 pulse and re-reduction in feed II (as described in section 3.2.3.2) were depicted in Figure 4.41 and
4.42 resp
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
After O2 pulseBefore O2 pulse
Stra
in [%
]
Cu [111]
0HN 120HN 0HC 120HC
F : Comparison of Cu [111] microstrain before and after adding O2 to the feed
at 523 K
In
pl
or 120HNo. The higher value in strain detected for 0HNr is unexplained because the given
value calculated from this in-situ experiment (bregg-brentano) is mistrusted and not relevant with in-
situ theta-theta data evaluation of microstrain before the O2 pulse (i.e. reduced Cu/ZnO). (Note: The
in-situ XRD carried out in bregg-brentano diffractometer is rather difficult to control due to some ex-
perimental error during the MSR reaction).
The overview TEM images of 0HN
ectively. Dark images presented in the TEM pictures are assigned as Cu particles. The images
showed that the Cu particles merged together and formed big agglomerate. For effective HRTEM
measurements, small particle is needed. Thus, from the overview picture, the big Cu particle (~ 250
nm) doesn’t suit for the HRTEM measurement and thus no lattice fringes were observed. As revealed
previously in TEM images of reduced Cu/ZnO catalysts, the minor interfacial contact of Cu to ZnO in
85
HN samples will increase the sintering effect especially after underwent the O2 pulse treatment. This
finding is in good agreement with NMR and in-situ XRD of Cu particle of HN samples, where higher
degree of sintering was observed after the oxygen treatment. On the other hand, Energy Dispersive X-
ray (EDX) analysis of re-reduced Cu/ZnO in feed II revealed a presence of oxygen content in the sam-
ple. However, from the in-situ XRD determination of re-reduced Cu/ZnO catalysts, there is no Cu2O
and/or CuO phase were detected after the re-reduction in feed II (respective to XRD limitation). This
probably due to non-fully reduced of CuO to Cu metal after second MSR reaction in feed II or the
sample was exposed to air during relocation.
igure 4.41: TEM image of 0HNo after re-reduction in feed under magnification of 13,000 x (left)
igure 4.42: TEM image of 120HN after re-reduction in feed under magnification of 9,600 x
In addition to the microstructural investigation by in-situ XRD experiments, the X-ray absorp-
on spectroscopy (XAS) was carried out to analysis the effect of O2 pulse to the microstructural
F F
(right)
ti
changes in Cu/ZnO catalysts. The catalytic activity of HN and HC samples after adding O2 to the feed
mixture at 523 K is depicted in Figure 4.43 as function of H2 production per time on stream. The feed 1
indicate the activity before the O2 pulse and feed 2 indicate the activity after approximately 20 minutes
of oxygen addition. The positive effect of the O2 addition to the catalytic activity is clearly observed for
all four samples especially for HN. The total increment of HN samples is about ~ 400 % by means of
86
H2 production. The percent of increment in the H2 production rate normalized per gram catalysts for all
four samples are summarized in Table 4.8.
igure 4.43: Increase in the H2 production (vol %/s) per time on stream after the O2 pulse for 0HN, 120HN, 0HC and 120HC measured by XAS cell at 523 K
able 4.8: Increase in H2 production rate (per gram catalysts) after adding O2 to the feed for about 20 min at 523 K (respect to the initial activity before O2 pulse)
F
T
H2 production rate (µmol/g*s) Sample name
(mg) Before O pulse After O pulse rcent of incre-ment (%)
Mass of catalyst Pe2 2
0HN 2.55 7.12 36.9 418.3 120HN 2.51 5.58 33.0 491.4
0HC 2.53 53.7 93.6 74.3 120HC 2.52 79.0 126 59.5
The extended fine structure (EXAFS) of the experimental XAS spectra measured under reac-
tion con
are slight differences detectable after the oxidation-re-reduction process.
0 20 40 60 8 100 120 140 160 180
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0
Feed 2
Feed 1
0HN 120HN 0HC 120HC
H2 pr
oduc
tion
[vol
%/s
]
Time [min]
ditions at the Cu K-edge and Zn K-edge was investigated to analyze changes in the short range
and medium range order structure of Cu and Zn after oxidation and re-reduction at 523 K in methanol
steam reforming gas mixture. The experimental Cu K-edge EXAFS (Figure 4.44) and Zn K-edge
EXAFS (Figure 4.45) of (a) 0HN, (b) 120HN, (c) 0HC and (d) 120HC before and after the O2 pulse are
presented. In principle, there is no significant change detected for Cu EXAFS and Zn EXAFS after the
O2 pulse for all of the samples. However, by looking further to the EXAFS spectra of Cu and Zn, there
87
Qualitatively the resulting Cu EXAFS shows slight changes in Cu-Cu shells especially in the
middle range order ~ 3 Å to 4.5 Å. The changes in the amplitude at higher distances in the FT(χ(k))
after the O2 pulse is more pronounced in the HN samples compared to the HC samples. The increase in
amplitude may be caused by a decrease in structural disorder (decreasing DWF) or sintering effects
leading to changes in coordination number. However, analysis of Cu EXAFS peaks in the middle range
order (R ~ 3 Å to 4.5 Å) reveal slight changes in the imaginary part caused by distance changes in cop-
per clusters. Thus, further investigations have to follow in order to elucidate if microstructural or sur-
face structural characteristics enhance the catalytic activity after oxidation re-reduction cycles.
88
104
113 58
(a) Cu K-edge of 0HN (b) Cu K-edge of 120HN
(c) Cu K-edge of 0HC (d) Cu K-edge of 120HC
Figure 4.44: Experimental Fourier transformed Cu K-edge χ(k) of differently aged HN and HC samples obtained at 523 K during methanol steam reforming reac-tion before (1 Feed) and after (2 Feed) O2 pulse
59
(a) Zn K-edge of 0HN (b) Zn K-edge of 120HN
(c) Zn K-edge of 0HC (d) Zn K-edge of 120HC
Figure 4.45: Experimental Fourier transformed Zn K-edge χ(k) of differently aged HN and HC samples obtained at 523 K during methanol steam reforming reaction before (1 Feed) and after (2 Feed) O2 pulse
105
60
Qualitative analysis of the Zn K-edge EXAFS of reduced Cu/ZnO catalysts recorded at 523 K using refer-
ences was done in order to elucidate if the Zn RDF can be described by ZnO. Therefore, pure ZnO obtained from
hydrozincite (HZ) was used as a reference to compare the Zn K-edge EXAFS derived from HN and HC preparation
route (Figure 4.46).
0.0
0.05
0.0
-0.05
0 2 4 6
FT
( χ(k
)*k
3)
R [Å]
ZnO (HN) ZnO (HZ) ZnO (HC)
Figure 4.46: Fourier transformed experimental Zn K edge χ(k) and imaginary part of the reduced Cu/ZnO de-
rived from hydroxynitrate (HN), hydroxycarbonate (HC) and hydrozincite (HZ) sample recorded at 523 K
Qualitatively, the HN sample matches well with the HZ as could be seen in the first Zn-O shell and Zn-Zn
distances. Large deviation in the amplitude was observed for the HC sample indicating a non ideal ZnO character or
additional Zn containing solid state phases. However, simulation of the experimental HN sample assuming ideal
ZnO (single crystal data) was found to fit much better than the HC prepared catalyst, even though it is only contain
ZnO (HZ). This was observed in the damp in amplitude and slight shifts in the imaginary part of HC and HZ sam-
ples compared to HN samples which indicated less ideal ZnO. These results clearly demonstrate again that Cu/ZnO
preparation using HC precursors result in a more defect rich character of the final catalyst.
The evolution of the experimental Ft(χ(k)) of ZnO of all the samples (i.e. 0HN, 120HN, 0HC, and 120HC)
as depicted in Figure 4.45, shows qualitatively no significant change in the amplitude especially for the first Zn-O
shell as function of oxygen pulse. However, slight changes in the amplitude of the Ft(χ(k)) in the medium range
order mainly correspond to the Zn-Zn single scattering path. The slight increase in the amplitude was observed for
all samples at distances ~ 2.9 Å, but no change in the intercept of the imaginary part was observed. It is assumed
that the equidistance shift in the imaginary part of the whole Zn EXFAS for the 0HN sample after oxygen pulse does
61
not correspond to bulk structural changes in ZnO but caused by bad data quality (energy calibration, signal to noise
ratio). In principle, by adding O2 to the feed followed by re-reduction process led to more ideal ZnO for all Cu/ZnO
samples. Further investigations on the effect of the O2 pulse to the microstructural of Cu and ZnO have to follow.
CHAPTER 5
DISCUSSIONS
According to the positive effect of aging on the Cu/ZnO catalysts (i.e. microstructural and activity) particu-
larly on Cu-Zn hydroxycarbonate preparation route [18-19], comparative study of differently aged freshly precipi-
tated HN was carried out to investigate the effect of aging on the different precursor phase. The resulting results
from previous study of HC precursors [18-19,37] was used in this study as comparison to the results obtained from
the HN precursors. Thus in this study the effect of aging from the precursor phase towards the activity of the
Cu/ZnO catalysts in methanol steam reforming reaction between HN and HC was comparatively shown.
Generally, the discussion part which is stressed more to HN preparation route will be divided into three
main parts. The first part is dedicated to the precursor formation and the characteristic of the precursor phase and the
oxide form. While the reduced Cu/ZnO catalysts and the catalysts performance in methanol steam reforming (MSR)
reaction will be discussed in the second part. Adding the oxygen to the feed gas during the MSR received great in-
terest lately due to promising feature [28,72]. Thus, the effect of the oxygen pulse on the microstructural and the
activity of the catalysts also will be discussed in the third part of this chapter.
5.1 Precursor Formation and Decomposition
Aging of a precipitate is a process that can have a significant effect on the catalytic activity of the resulting
material. In the production of commercial Cu/ZnO catalysts, a period of aging is essential [22]. During the process
62
of aging the initial precipitate remains in contact with the precipitating reagents, and the particles can partially redis-
solve or increase in size due to Ostwald ripening [73]. Ostwald ripening or known as particle growing concerns only
crystal of the same composition and structure (i.e. crystals of the same phase) which is in contrast with phase transi-
tion. In addition to Ostwald ripening effect, the phase composition is often reported to change during the aging time
[61,73].
The Ostwald ripening effect was observed during the aging of HN precursor. The constant pH evolution
and the resulting XRD phase determination reveal that the precursor phase remained as gerhardite-like even after
120 min aging time. The increase in particle size was detected qualitatively from the narrowing of the most intense
gerhardite peak (FWHM).
On the other hand, the drop in pH curve during aging of HC shows that the crystallization takes place. The
initially amorphous precipitate (0 min) (i.e. georgeite; [(Cu,Zn)(CO3)3(OH)2]) was transformed into crystalline phase
of rosasite [(Cu,Zn)2CO3(OH)2] and aurichalcite [(Cu,Zn)5(CO3)2(OH)6] after approximately 22 min aging time
(Figure 4.1) [21].
The HN precursors phase seems to be composed of one single phase; resembling gerhardite with no zinc
hydroxynitrate phase was detected. This is in good agreement with the study by Sengupta et al. [58], which showed
no presence of zinc hydroxynitrate even by using as low as copper to zinc ratio of 30:70. The gerhardite phase pre-
sent in zinc-containing samples (i.e. (Cu0.7Zn0.3)2(OH)3NO3) in this study is then called gerhardite-like phase.
The uncertainty occur in determining the real precursor phase due to some unexplained peak appear in the
XRD diffractogram. One possibility is there is other phase presence in the precursor phase such as NH4NO3. This
was strengthened by decomposition of NH4+ and NOx during the calcination of HN precursors. However, there is no
NH4NO3 phase match with the unexplained peak. Even if that so, washing process was involved during the prepara-
tion which may remove the NH4+ and NO3
- excessively [51]. Consequently, it could not be detected due to lower
amount of NH4NO3 compared to the instrument detection limit. Besides that, it could be assumed that some of the
zinc ions have been incorporated in a gerhardite lattice and/or the ammonia probably ‘located’ in between the layer
structure of gerhardite. This may led to another possibility of forming a new solid state phase containing ammonia
in the structure (e.g. NH4x(Cu0.7,Zn0.3-x/2)(OH)3NO3). Hence may influence the pure gerhardite phase formation and
thus change the peak intensity relative to the ratio of the reflexes of the precursors.
63
5.2 Microstructural and Activity Correlations
The decrease in the Cu crystallite size as function of aging time from 275 Å (0HNr) to 236 Å (120HNr) is
reflected in the onset of the reduction of CuO/ZnO precursor prepared from HN samples (Figure 4.15). The slight
differences in the mass loss of the 120 min aged sample in comparison to the 0 min aged sample is caused by slight
decrease in CuO content with increasing aging time as revealed by EDX analysis. Both reduction potential in 2 vol
% H2/He and 15 vol % H2/He shows that both aged and non-aged HN samples is totally reduced to Cu/ZnO (Table
4.3) with no residual NOx detected. Furthermore, as revealed in the thermal decomposition step (calcination) of
0HNp and 120HNp there is no additional decomposition happened at higher temperature of 603 K.
The effect of precipitate aging on the microstructural changes of the final Cu/ZnO catalysts derived from
HN samples is not as pronounced as precipitate aging of HC samples. Generally, both samples revealed more or less
same characteristics except in Cu crystallite size determined by XRD. It has been proved that the increase in activity
of the copper catalysts correlates not only to the decreasing in particle size, but also with the increasing of mi-
crostrain (disorder of copper) in the copper nanoparticles [18]. The strain determination obtained by in-situ XRD
experiments and the increase in the Debye-Waller factor used as strain indicator calculated from the EXAFS refine-
ment for both catalysts revealed less strain in copper lattice and thus results to less active catalysts [19]. In addition,
small and symmetric NMR signal of 0HNr and 120HNr revealed large and more to ideal copper particle. The large
copper particle on ZnO upon calcination and reduction (TEM) minimize the Cu/ZnO interaction. Assuming that
copper microstrain originates from epitactical orientation of copper on ZnO, the reduced Cu/Zn interface resulted in
less strained copper particles. Consequently, less strain could be diverted to the copper surface (surface reaction)
which results to less active catalysts in MSR reaction compared to the activity of Cu/ZnO catalysts of HC. In addi-
tion to all the unpromising result of Cu/ZnO catalysts derived from HN preparation route is the formation of
gerhardite phase (precursor). The form and dispersion of copper which is the active component depends to a great
extent on the amount of gerhardite formed. According to Sengupta et al. [58] the presence of gerhardite causes sin-
tering copper oxide formed which make the catalysts inactive.
Unlike HN samples, the distinct deviation between aged and non-aged HC samples clearly could be seen
by means of Cu crystallite size (XRD), microstrain effect (in-situ XRD and EXAFS), specific Cu surface area (N2O
decomposition) and also in the MSR reaction. The decrease in Cu crystallite size from initially 110 Å to 70 Å and
increase in microstrain effect (Figure 4.26 and 4.27) clearly indicated that aging has tailored the microstructural of
64
the precursor which led to highly defect structure [19]. This was also observed from the NMR investigation which
results in asymmetric line broadening and even additional NMR signal by increase the aging time [18]. In summary,
by increasing the precursor aging time led to increase in the nanostructure character of the final Cu/ZnO catalysts
which accompanied by higher copper surface area is due to the smaller crystallite size evidence by TEM, XRD, RFC
and MSR investigations.
5.3 The Oxygen Addition (O2 pulse) Investigation
Addition of oxygen in feed (i.e. CH3OH and H2O) during the methanol steam reforming may result in
modification of the copper-zinc catalysts surface, causing an increase in the rate of methanol decomposition and
consequently enhanced the steam reforming activity. As reported by Huang et al. [70-71], the presence of oxygen in
the steam reforming enhances methanol conversion due to occurrence of the partial oxidation of methanol. Further-
more high hydrogen yield can be obtained although the catalysts stability is poor [24].
The positive effect of O2 addition to the feed was observed in both HN and HC samples. Surprisingly, the
total increment of approximately 400 % in H2 production was achieved from HN samples respect to the initial activ-
ity of the catalysts in the first MSR reaction (Table 4.8). According to Szizybalski et al. [28] the increased activity
of the Cu/ZrO2 catalyst after oxidative treatment can be directly correlated to the increasing amount of oxygen in the
copper metal clusters. This was proven by unfitted spectra (Cu-EXAFS) of the copper clusters by of pure copper
metal as was applied in this study. Nevertheless, this feature is not observable in the Cu/ZnO system. In addition, the
in-situ XRD analysis of the re-reduced catalysts showed the presence of only Cu0 and ZnO crystalline phases (Fig-
ure 4.38). Wang et al. [74] have been reported that highly dispersed Cu0 is an important factor for high activity of
the catalysts in hydrogen generation from methanol partial oxidation while the presence of Cu+ species inhibits the
hydrogen production from the methanol. This indicates that in order to enhance the catalytic activity of Cu/ZnO
catalysts, CuO should be fully reduced to metallic copper [24]. However, there are also other results indicating that
Cu+ species helps to increase the activity of the Cu-based catalysts, and it is suggested that both Cu0 and Cu+ species
are essential for hydrogen generation from methanol and the activity of the catalysts is dependent on the ratio of
Cu+/Cu0 in the catalysts [71]. Thus in determination of the active species involved during the reaction, further inves-
tigation and other technique should be involved.
65
After the oxidative treatment for temporarily 20 min, both series (HN and HC) shows an increase in Cu
particle size which results to less strain copper. Additional ex-situ 63Cu NMR investigation on 0HNo and 120HNo
also revealed the Cu particle get sintered after the oxygen addition and it is more pronounced for longer aging time.
As expecting, high sintering effect on HN samples after the O2 pulse is caused by higher degree of segregation be-
tween Cu and ZnO as revealed by TEM images of the reduced catalysts. The high interfacial contact of Cu to ZnO
in HC sample led to less sinter Cu particle (Figure 4.39).
The Cu EXAFS experiment shows that the slight changes in Cu-Cu shells (R ~ 3 Å to 4.5 Å) is more
pronounced on HN samples (Figure 4.45). The increase in amplitude is probably caused by a decrease in structural
disorder or sintering effect. On the other hand, large deviation in the amplitude of Zn EXAFS indicating a non ideal
ZnO character was observed for HC but not on HN samples. The criterion of ZnO in HN samples is fit more to ideal
ZnO. However, further details on the microstructural changes on Cu and ZnO after the O2 pulse is still under inves-
tigation.
CHAPTER 6
CONCLUSIONS AND RECOMMENDATION
In addition to the effect of using different precipitating agent, an appropriate aging time also play an impor-
tant role by tailoring the microstructural properties of the active catalysts. Precipitation using ammonia hydroxide
lead to the formation of monophasic metal hydroxy nitrate precursor (HN), whereas the standard method using so-
dium carbonate result in a mixture of several metal hydroxy carbonates (HC). The characteristics phase transforma-
tion (pH evolution) during aging of HC was not observed for the aging process of HN.
The lower catalytic activity of Cu/ZnO catalysts obtained from HN samples compared to the catalysts pre-
pared by HC samples correlates qualitatively with the larger copper crystallite size (determined by detailed XRD
line profile analysis) and the resulting lower copper surface area detected by N2O decomposition. Furthermore, the
positive effect of aging on the catalytic activity of HC samples is not observable for the HN samples due to insig-
nificant changes in copper crystallite size. NMR and EXAFS investigation revealed a more ideal copper bulk struc-
66
ture for the HN prepared catalysts compared to HC. In other word, the effect of aging on HN samples is not as sig-
nificant as HC samples.
Besides the fact that specific copper surface area exposed to the gas mixture mainly determines the cata-
lytic activity of Cu/ZnO catalysts, it is now have been extendedly proven that bulk structural characteristics such as
microstrain enable to enhance the catalytic activity of the copper surface by comparing the relatively ideal copper
system (i.e. HN) to ‘real catalysts’ design (i.e. HC). The higher strain amount of the catalyst prepared by 120 min
aging of HC clearly indicates that the nanostructured character of the catalysts and the resulting distinct interfacial
contact of ZnO to copper are necessary to prepare strained copper particles. Aging of the HN precursors led to big-
ger Cu particle and less strain copper as illustrated in Figure 4.47.
Cu to ZnO in-terface
Cu
Cu Cu
ZnO ZnO
ZnOCu
strain
Figure 6.1: Schematic model of HN precipitate as function of aging
Upon addition of oxygen to the feed followed by reduction, an increase in catalytic activity was observed
for both catalysts (HN and HC). From the whole experiment conducted on the Cu/ZnO catalysts especially for HN
samples (i.e. in-situ XRD, N2O decomposition, NMR, TEM) after the oxygen addition, the higher catalytic activity
does not correlated with an increase in copper surface area, microstrain or oxygen in copper cluster but slight struc-
tural changes of the catalysts in the medium range order of Cu and ZnO determined by XAS experiment.
In order to further investigate the effect of aging on precipitate precursor and correlations between micro-
structural and activity of Cu/ZnO catalysts, several other experiments should be carried out. The longer aging time
(more than 120 min) especially for HN precursors is suggested to observe if there is any significant change on the
precursor phase and thus affecting the final properties of Cu/ZnO catalyst. Since the oxygen addition has significant
67
effect in activity of the Cu/ZnO catalysts, more oxygen addition to feed cycles is suggested to investigate the cata-
lysts performance and stability. Instead of using oxygen, other oxidant for instance N2O can be employed due to
much higher methanol conversion and more active oxidizing agent compared to O2 [24]. On the other hand, as been
known it was assumed that the reaction of Cu/ZnO catalysts occurred on the copper surface. Thus, a surface method
analysis for example X-ray photoelectron spectroscopy (XPS) is proposed in this study in order to further character-
ize the samples. XPS analysis allows the verification of species present on the samples. Therefore further informa-
tion of the active species can be obtained.
BIBLIOGRAPHY
1. Chinchen, G.C., Denny, P.J., Jennings, J.R., Spencer, M.S and Waugh, K.C. 1986. Synthesis of methanol:
Part I. Catalysts and Kinetics. Applied Catalysis 36:1-65.
2. Klier, K. 1982. Methanol synthesis. Advances in Catalysis 31:243-313.
3. Klier, K. 1984. Structure and function of real catalysts. Applications of Surface Science 19: 267-297.
4. Bart, J.C.J and Sneeden, R.P.A. 1987. Copper-zinc oxide-alumina methanol catalysts revisited. Catalysis
Today 2:1-124.
5. Chinchen, G.C., Waugh, K.C and Whan, D.A. 1986. The activity and state of the copper surface in metha-
nol synthesis catalysts. Applied Catalysis 25:101-107.
6. Sheffer, G.R and King, T.S. 1989. Potassium’s promotional effect of unsupported copper catalysts for
methanol synthesis. Journal of Catalysis 115:376-387.
7. Sheffer, G.R and King, T.S. 1989. Differences in the promotional effect of the group IA elements on un-
supported copper catalysts for carbon monoxide hydrogenation. Journal of Catalysis 116: 488-497.
8. Fujitani, T., Saito, M., Kanai, Y., Watanabe, T., Nakamura, J and Uchijima, T. 1994. The role of metal ox-
ide in promoting a copper catalysts for methanol synthesis. Catalysis Letters 25:271-276.
9. Bartley, G.J.J and Burch, R. 1988. Support and morphological effects in the synthesis of methanol over
Cu/ZnO, Cu/ZrO2 and Cu/SiO2 catalysts. Applied Catalysis 43:141-153.
10. Burch, R., Golunski, S.E and Spencer, M.S. 1990. The role of copper and zinc oxide in methanol synthesis
catalysts. Journal of Chemical Society: Faraday Transactions 86(15): 2683-2691.
11. Spencer, M.S.1998. Role of ZnO in methanol synthesis on copper catalysts. Catalysis Letters 50:37-40.
68
12. Yoshihara, J and Campbell, C.T.1996. Methanol synthesis and reverse water-gas shift kinetics over
Cu(110) model catalysts:Structural sensitivity. Journal of Catalysis 161:776-782.
13. Ovesen, C.V., Calusen, B.S., Schiøtz, J., Stoltze, P., Topsøe, H and Nøorkov, J.K.1997. Kinetic implica-
tions of dynamical changes in catalyst morphology during methanol synthesis over Cu/ZnO catalysts.
Journal of Catalysis 168:133-142.
14. Hadden, R.A., Sakakini, B., Tabatabaei, J and Waugh, K.C. 1997. Adsorption and reaction induced mor-
phological changes of the copper surface of a methanol synthesis catalysts. Catalysis Letters 44:145-151.
15. Topsøe, N.Y and Topsøe, H. 1999. FTIR studies of dynamic surface structural changes in Cu-based metha-
nol synthesis catalysts. Journal of Molecular Catalysis A 141:95-105.
16. Barr, T.L and Hackenberg, J.J. 1982. Studies of the low temperature oxidation of alloys by X-ray photo-
electron spectroscopy: Cu-Zn. Applications of Surface Science 10:523-545.
17. Kanai, Y., Watanabe, T., Fujitani, T., Saito, M., Nakamura, J and Uchijima, T. 1994. Evidence for the mi-
gration of ZnOx in a Cu/ZnO methanol synthesis catalysts. Catalysis Letters 27:67-78.
18. Kniep, B.L., Ressler, T., Rabis, A., Girgsdies, F., Beanitz, M., Steglich, F and Schlögl, R. 2004. Angewand-
te Chemie International Edition 43:112-115.
19. Kniep, B.L., Girgsdies, F., Ressler, T and Schlögl, R. Effect of precipitate ageing on the microstructural
characteristics of Cu/ZnO catalysts for methanol steam reforming. Journal of Catalysis : submitted.
20. Günter, M.M., Ressler, T., Bems, B., Büscher, C., Genger, T., Hinrichsen, O., Muhler, M and Schlögl, R.
2001. Implication of the microstructure of binary Cu/ZnO catalysts for their catalytic activity in methanol
synthesis. Catalysis Letters 71(1-2):37-44.
21. Günter, M.M., Ressler, T., Jentoft, R.E and Bems, B. 2001. Redox behavior of copper oxide/zinc oxide
catalysts in the steam reforming of methanol studied by in situ X-Ray Diffraction and Absorption Spectros-
copy. Journal of Catalysis 203.133-149.
22. Agrell, J., Lindström, B., Pettersson, L and Järås, S.G. 2002. Catalytic hydrogen generation from methanol.
Catalysis 16:67-132.
23. Lindstörm, B and Pettersson, L.J. 2002. Steam reforming of methanol over copper based monoliths: the
effects of zirconia doping. Journal of Power Source 106: 264-273.
24. Alejo, L., Lago, R., Pena, M.A and Fierro, J.L.G. 1997. Partial oxidation of methanol to produce hydrogen
over Cu-Zn-based catalysts. Applied Catalysis A 162:281-297.
25. Kobayashi, H., Takezawa, N and Minochi, C. 1976. Methanol reforming reaction over copper-containing
mixed oxides. Chemistry Letters 12:1347-1350.
69
26. Takezawa, N and Iwaza, N. 1997. Steam reforming and dehydrogenation of methanol: Difference in the
catalytic functions of copper and group VIII metals. Catalysis Today 36:45-56.
27. Breen, J.P and Ross, J.R.H. 1999. Methanol reforming for fuel-cell applications: development of zirconia-
containing Cu-Zn-Al catalysts. Catalysis Today 51:521-533.
28. Szizybalski, A., Girgsdies, F., Jentoft, R., Kassatkin, I., Schlögl, R., Ressler, T., Wang, Y., Schattka, J.H
and Niederberger, M. Binary Cu/ZrO2 catalyst for steam reforming of methanol: The influence of prepara-
tion on structure and activity. Journal of Catalysis : submitted.
29. Purnama, H., Ressler, T., Jentoft, R.E., Soerijanto, H., Schlögl, R and Schomäcker, R. 2003. CO forma-
tion/selectivity for steam reforming of methanol with a commercial CuO/ZnO/Al2O3 catalysts. Applied Ca-
talysis A : 259:83-94.
30. Takahashi, K., Takezawa, N and Kobayashi, H. 1982. The mechanism of steam reforming of methanol over
copper-silica catalyst. Applied Catalysis 2:363-366.
31. Santacesaria, E and Carrá, S. 1983. Kinetics of catalytic steam reforming of methanol in a CSTR reactor
Applied Catalysis 5:345-358.
32. Amphlett, J.C., Creber, K.A.M., Davis, J.M., Mann, R.F., Peppley, B.A and Stokes, D.M. 1994. Hydrogen
production by steam reforming of methanol for polymer electrolyte fuel cells. International Journal of Hy-
drogen Energy 19:131-137.
33. Peppley, B.A., Amphlett, J.C., Kearns, L. M and Mann, R.F. 1999. Methanol-steam reforming on
Cu/ZnO/Al2O3 catalysts. Part 2. A comprehensive kinetic model. Applied Catalysis 179:21-29.
34. Agrell, J., Birgersson, H and Boutonnet, M. 2002. Steam reforming of methanol over a Cu/ZnO/Al2O3 cata-
lysts: A kinetics analysis and strategies for suppression of CO formation. Journal of Power Sources
106:249-257.
35. El-Shobaky, H.G., Mokhtar, M and El-Shobaky, G.A. 1997. Partial oxidation of methanol to produce hy-
drogen over Cu-Zn- based catalysts. Applied Catalysis A 162:281-297.
36. Li, J.L and Inui, T. 1996. Characterization of precursors of methanol synthesis catalysts, cop-
per/zinc/aluminum oxides, precipitated at different pHs and temperatures. Applied Catalysis A 137:105-
117.
37. Bems, B., Schur, M., Dassenoy, A., Junkes, H., Herein, D and Schlögl, R. 2003. Relations between synthe-
sis and microstructural properties of copper/zinc hydroxycarbonates. Chemistry: A European Journal
9:2039-2052.
70
38. Shen, G.C., Fujita, S.I., Matsumoto, S and Takezawa, N. 1997. Steam reforming of methanol on binary
Cu/ZnO catalysts: Effects of precipitation condition upon precursors, surface structure and catalytic activ-
ity. Journal of Molecular Catalysis A 124:123-136.
39. Fujita, S.I., Moribe, S., Kanamori, Y., Kakudate, M and Takezawa, N. 2001. Preparation of a coprecipi-
tated Cu/ZnO catalysts for the methanol synthesis from CO2 - effects of the calcination and reduction con-
ditions on the catalytic performance. Applied Catalysis A 207:121-128.
40. Li, J.L., Zhang, X.G and Inui, T. 1996. Improvement in the catalyst activity for direct synthesis of dimethyl
ether from synthesis gas through enhancing the dispersion of Cu/ZnO/γ-Al2O3 in hybrid catalysts. Applied
Catalysis A 147:23-33.
41. Zhang, Y., Sun, Q., Deng, J., Wu, D and Chen, S. 1997. A high activity Cu/ZnO/Al2O3 catalysts for metha-
nol synthesis: Preparation and catalytic properties. Applied Catalysis A 158:105-120.
42. Ma, Y., Sun, Q., Wu, D., Fan, W.H., Zhang, Y.L and Deng, J.F. 1998. A practical approach for the prepara-
tion of high activity Cu/ZnO/ZrO2 catalyst for methanol synthesis from CO2 hydrogenation. Applied Ca-
talysis A 171:45-55.
43. Ma, Y., Sun, Q., Wu, D., Fan, W.H and Deng, J.F. 1999.A gel-oxalate co-precipitation process for prepara-
tion of Cu/ZnO/Al2O3 ultrafine catalysts for methanol synthesis from CO2 + H2 :(II) effect of various calci-
nation conditions. Applied Catalysis A 177:177-184.
44. Shen, J.P and Song, C. 2002. Influence of preparation method on performance of Cu/Zn-based catalysts for
low-temperature steam reforming and oxidative steam reforming of methanol for H2 production for fuel
cells. Catalysis Today 77:89-98.
45. Kurtz, M., Bauer, N., Büscher, C., Wilmer, H., Hinrichen, O., Becker, R. Rabe, S., Merz, K., Driess, M.,
Fischer, R.A and Muhler, M. 2004. New synthesis route to more active Cu/ZnO catalysts used for methanol
synthesis. Catalysis Letters 92:49-52.
46. Castricum, H.L., Bakker, H and Poels, E.K. 2001. Oxidative and reduction in copper/zinc oxides by me-
chanical milling. Materials Science and Engineering A 304-306:418-423.
47. Schüth, F and Unger, K. 1997. Precipitation and coprecipitation. In Handbook of Heterogeneous Catalysis
Vol. 1, ed. G. Ertl, H. Knözinger, and J. Weitkamp, pp:72-85.Weinheim: VCH Verlagsgesellschaft.
48. Waller, D., Stirling, D., Stone, F.S and Spencer, M.S. 1989. Copper-zinc oxide catalysts. Activity in rela-
tion to precursor structure and morphology. Faraday Discussion of the Chemical Society 87:107-120.
49. Porta, P., de Rossi, S., Ferraris, G., Lo Jacono, M and Minelli, G. 1988. Structural characterization of mala-
chite-like coprecipitated precursors of binary CuO-ZnO catalysts. Journal of Catalysis 109:367-377.
71
50. Herman, R.G., Klier, K., Simmons, G.W., Finn, B.P., Bulko, J.B and Kobylinski, T.P. 1979. Catalytic syn-
thesis of methanol from CO/H2: I. Phase composition, electronic properties, and activities of the
Cu/ZnO/M2O3 catalysts. Journal of Catalysis 56:407-429.
51. Himelfarb, P.B., Simmons, G.W., Klier, K and Herman, R.G. 1985. Precursors of the copper-zinc oxide
methanol synthesis catalysts. Journal of Catalysis 93:442-450.
52. Rasmussen, B.S., Nielsen, P.E.H., Vialladsen, J., Hansen, J.B and Topsøe, H. 1987. The activity of copre-
cipitation Cu, Zn catalysts for methanol synthesis. In Preparation of Catalysts IV, eds. B. Delmon, P.
Grange, P.A. Jacobs and G. Poncelet, pp 785- 794. Amsterdam: Elsivier.
53. Spencer, M.S. 1999. The role of zinc oxide in Cu/ZnO catalysts for methanol synthesis and water-gas shift
reaction. Topics in Catalysis 8:259-266.
54. Millar, G.J., Holm, I.H., Uwins, P.J.R and Dreennan, J. 1998. Characterization of precursors to methanol
synthesis catalysts Cu/ZnO system. Journal of Chemical Society: Faraday Transaction 94(4):593-600.
55. Spencer, M.S. 2000. Precursors of copper/zinc oxide catalysts. Catalysis Letters 66:255-257.
56. Sloczyński, J., Grabowski, R., Kozlowska, A., Olszewski, P.K and Stoch, J. 2003. Reduction kinetics of
CuO in CuO/ZnO/ZrO2 systems. Physical Chemistry Chemical Physics 5:4631-4640.
57. Frolich, K., Fenske, M.R and Quiggle, D. 1928. Catalysts for the formation of alcohols from carbon mon-
oxide and hydrogen. I- Decomposition of methanol by catalysts composed of copper and zinc. Industrial
and Engineering Chemistry 20(7):694-698.
58. Sengupta, G., Das, D.P., Kundu, M.L., Dutta, S., Roy, S.K., Sahay, R.N and Mishira, K.K. 1989. Study of
copper-zinc oxide catalysts, characterization of the coprecipitate and mixed oxide. Applied Catalysis
55:165-180.
59. Agrell, J., Hasselbo, K., Jansson, K., Järås, S.G and Boutonnet, M. 2001. Production of hydrogen by partial
oxidation of methanol over Cu/ZnO catalysts prepared by microemulsion technique. Applied Catalysis A
211:239-250.
60. Taylor, S.H., Hutchings, G.J and Mirzaei, A.A. 1999. Copper zinc oxide catalysts for ambient temperature
carbon monoxide oxidation. Chemical Communications 15:1373-1374.
61. Whittle, D.M., Mirzaei, A.A., Hargreaves, J.S.J., Joyner, R.W., Kiely, C.J., Taylor, S.H and Hutchings,
G.J. 2002. Co-precipitated copper zinc oxide catalysts for ambient temperature carbon monoxide oxidation:
effect of precipitate ageing on catalysts activity. Physical Chemistry Chemical Physics 4:5915-5920.
62. Stone, F.S. and Waller, D. 2003. Cu-ZnO and Cu-ZnO/Al2O3 catalysts for the reverse water-gas shift reac-
tion. The effect of the Cu/Zn ratio on precursor characteristics and on the activity of the derived catalysts.
Topics in Catalysis 22(3-4):305-318.
72
63. J.B. Hansen. 1997. Methanol Synthesis. In Handbook of Heterogeneous Catalysis Vol 4, ed. G. Ertl, H.
Knözinger, and J. Weitkamp, pp:1856-1872.Weinheim: VCH Verlagsgesellschaft.
64. Ressler, T. 1998. WinXAS: A program for X-ray Absorption Spectroscopy data analysis under MS-
Windows. Journal of Synchrotron Radiation 5:118-122.
65. Ankundinov, A.L., Ravel, B., Rehr, J.J and Conradson, S.D. 1998. Real-space multiple-scattering calcula-
tion and interpretation of x-ray-absorption near-edge structure. Physical Review B 58:7565-7576.
66. De Soete, G.G., 1990. Nitrous Oxide Formation and Destruction by Industrial NO Abatement Techniques
Including SCR. Oil & Gas Science and Technology 45(5): 663-682.
67. Chinchen, G.C., Hay, C.M., Vandervell, H.D and Waugh, K.C. 1987. The measurement of copper surface
areas by reactive frontal chromatography. Journal of Catalysis 103:79-86.
68. Clausen, B.S., Gråbæk, L., Topsøe, H., Hansen, L.B., Stoltze, P and Nørskov, J.K. 1993. A new procedure
for particle size determination by EXAFS based on molecular dynamics simulations. Journal of Catalysis
141:368-379.
69. Grunwaldt, J.D., Molenbroek, A.M., Topsøe, N.Y., Topsøe, H and Clausen, B.S. 2000. In situ investigation
of structural changes in Cu/ZnO catalysts. Journal of Catalysis 194:452-460.
70. Huang, T.J and Wang, S.W. 1986. Hydrogen production via partial oxidation of methanol over copper-zinc
catalysts. Applied Catalysis 24:287-297.
71. Huang, T.J and Chren, S.L. 1988. Kinetics of partial oxidation of methanol over a copper-zinc catalyst.
Applied Catalysis 40:43-52.
72. Günter, M.M., 2001. Structure and activity of copper/zinc oxide catalysts studied using X-ray diffraction
and Absorption Spectroscopy. Ph.D Thesis, Technical University of Berlin, Germany.
73. Boistelle, R and Astier, J.P. 1988. Crystallization mechanisms in solution. Journal of Crystal Growth
90:14-30.
74. Wang. Z., Wang, W and Lu, G. 2003. Studies on the active species and on dispersion of Cu in Cu/SiO2 and
Cu/Zn/SiO2 for hydrogen production via methanol partial oxidation. International Journal of Hydrogen
Energy 28:151-158.
APPENDICES
1. Theoretical calculation for the decomposition of HN precursor
Example: Assuming that the precursor phase consists only single monophasic gerhardite phase, thus Cu2(OH)3NO3 O
MCu2(OH)3NO3 = 120 / mol Cu MCuO = 79.55 / mol Cu So, Precursor Oxide
120 = 79.55 Let say, 100 = X
X = [100/120 x 79.55] = 66.29 %
Thus, 100 % – 66. 29 % = 33.71 % (mass loss)
d
2. Theoretical calculation for the reduction of CuO/ZnO
Example: From the XRF analysis, the 0HNp consist of 73.16 % copper and 26.84 % zinc, supposing that all CuO is reduced to Cu in 2 vol % H2/He, thus
MCuO = 79.55 MZnO = 81.39 Because of, 73.16 % CuO = 73.16/100 x 79.55 = 58.20 26.84 % ZnO = 26.84/100 x 81.39 = 21.85 Thus, total CuO/ZnO = 58.20 + 21.85 = 80.05 CuO
73.16/100 x 16
Let say, 80.05 = So, 11.71 = X
3. Calculation for the methanol conversio
Conv. = CH OH inCH3OH 3
4. Calculation for the hydrogen productio
By using, n =
Where, n = Hp = 1V =
reduced
n
n
p
2
v
Cucalcine
73
Cu
= 11.71 (O2 content in CuO)
100 % X = 11.71/80.05 x 100 = 14.63 % (mass loss)
- CH CH
3OH out x 100 % 3OH out
rate
V/RT
production bar olume of H2 produced (L/s)
74
R = 0.08314 bar/L mol K T = 303 K
Thus, n = 1 bar x ( V L/s)
0.08314 bar/L mol K x 303 K = y µmol/s (1)
From eq. (1)
The H2 production rate = y µmol/s (per gram catalyst) g catalyst
= x µmol/gs The H2 production rate = y µmol/s (per gram Cu catalyst) g Cu catalyst
= z µmol/g*s
5. Calculation for the specific Cu surface area from N2O decomposition
1. V [l/g] = (Integral N – Blank) x (flow rate) N2 2
Sample mass 2. η N2 [mol/g] = pVN2
RT 3. Assuming NCu(s) = 1.47 x 1019 atoms/m2
4. η Cu [mol/m2] = NCu(s) / NA , η N2 (max) = 1/2 η Cu(s)
5. SCu [m2/g] = η N2 [mol/g] η N2 (max) [mol/m2]
75
BIODATA OF THE AUTHOR Name: Ernee Noryana Binti Muhamad Date of Birth: 9-11-1979 Place of Birth: Ayer Hitam, Johor Marital Status: Single E-mail Address: [email protected] , [email protected] Education Background: 1992-1994 Sekolah Tinggi Kluang PMR
1996 Sekolah Tinggi Kluang SPM (Grade 1) 1997-1998 Kolej Yayasan Pelajaran Mara, Science Matriculation Kota Bharu (CGPA : 2.946 ) 1998-2002 Universiti Putra Malaysia, Serdang Bachelor of Science (Hons.) – Industrial Chemistry
(CGPA: 3.582) List of the seminar/conference/workshop attended:
1. Industrial Chemistry VII Seminar on Kimia : Pemangkin dalam Perindustrian Negara” (19th September 2001), organized by Chemistry Department, Universiti Putra Malaysia at IDEAL hall, UPM.
2. TA Conference on Advanced Technology in Material Characterization (23 & 24 May 2002), organized by
USM, UKM and Mettler-Toledo (M) Sdn Bhd at Dewan Cemerlang 1, MINES International Exhibition Convention Centre.
3. Chemistry Department Seminar “Solid State Chemistry of Ferroelectric Ceramics” by Prof A. R. West at
Multi-Purpose Room, FSAS UPM, 20 December 2002.
4. Seminar on “TEM and EELS of VPO Catalysts”, (27th December 2002) by Dr. D.S. Su at Auditorium Ground Floor, Institute of Postgraduate Studies, UM.
5. Seminar on “Tribomechanical Activation of VPO Catalysts”, (28th December 2002) by Dr. D.S. Su at Bilik
Seminar Al-Khawarizmi, Jabatan Matematik, UPM.
6. TPDRO 1100 Training, by Guee Eng Hwa (O’Connor’s Engineering Sdn. Bhd.) (2th January 2003) at COMBICAT-UPM Laboratory.
7. Seminar on Robotics by Dr. Bashir Harji (Cambridge Reactor Design Ltd.) at Auditoium, Institut Asia-
Eropah, UM, 17th January 2003
8. Seminar of Surfactant Combicat UKM, Chemistry Department UKM (24 January 2003.
9. Seminar on “The Role of Metal Oxides in Selective Oxidation Catalysts”, (19th February 2003) by Prof. Robert Schlogl (Department of Inorganic Chemistry, Fritz Haber of Max Planck Society, Germany) at Bilik Seminar Al-Khawarizmi, Jabatan Matematik, UPM.
10. Seminar on “The Role of in-situ techniques in making combinatorial heterogeneous catalysis a success”,
(20th February 2003) by Prof. Robert Schlogl (Department of Inorganic Chemistry, Fritz Haber of Max Planck Society, Germany) at Auditorium 1, IPS, UM.
76
11. Seminar on “Preparation of hetegeneous catalysis”, (21th February 2003) by Prof. Robert Schlogl (Depart-ment of Inorganic Chemistry, Fritz Haber of Max Planck Society, Germany) at Anggerik Room, IPS, UM.
12. Pameran Reka Cipta dan Penyelidikan UPM 2003, 8-10th July 2003, Pusat Kebudayaan dan Kesenian Sul-
tan Salahuddin Abdul Aziz Syah (P.K.K.S.S.A.A.S), UPM. – Poster presenter with the title of “Synthesis and characterization of nanocrystalline CuO powders” – Silver medal
13. National Symposium on Science and Technology – Strategic Research and Innovation Towards Economic
Development, 28-30th July 2003, The MINES Resort City- Poster presenter with a the title of “Synthesis and characterization of nanocrystalline CuO powders”.
14. Workshop on Electron Microscopy, (18 - 22th August 2003) organised by Institute of Bioscience, Universiti
Putra Malaysia.
15. SKAM-16/EXPERTS 2003 (Simposium Kimia Analisis Malaysia Ke-16), (9 -11th September 2003) at Damai Lagoon, Kuching, Sarawak. – oral presentation entitled “Effect of Washing on Nanocrystalline of CuO Samples”.
16. Scientific stay at Fritz Haber Institute of MPG, Berlin Germany (Department of Inorganic Chemistry –
FHI) from 8.11.2003 until 5.11.2004.
17. 4th International Conference on Inorganic Materials, 19-21 September 2004, Antwerp, Belgium. Abstract, Synthesis and characterization of nanocrystalline CuO powders. (Poster Exhibitor – represented by super-visor)
18. Exhibition of Invention, Research and Innovation UPM 2005, 16-17th March 2005, UPM. – Poster pre-
senter with the title of “Effect of Aging on the Microstructural and Activity of Cu/ZnO Catalysts Derived From Different Precursor” – Bronze medal
Papers published in proceeding of conferences and seminars:
1. Ramli, I., E. N. Muhamad,A. H. Abdullah, Y.H. Taufiq-Yap, S.B. Abdul Hamid. Synthesis and charac-terization of nanocrystalline CuO powders – “National Symposium on Science and Technology – Strategic Research and Innovation Towards Economic Development”, The MINES Resort City. 28-30th July 2003.
2. Ramli, I., E. N. Muhamad, A. H. Abdullah, Y. H. Taufiq-Yap, S. B. Abd. Hamid. Effect of Washing on
Nanocrystalline of CuO Samples. “16th Annual Malaysian Symposium on Analytical Chemistry”, Damai Lagoon, Kuching, Sarawak. 9-11th September 2003.
3. R. Irmawati, E.N. Muhamad, I. Hamadneh, A.H. Abdullah, Y.H. Taufiq-Yap, S.B. Abdul Hamid, Synthe-sis and characterization of nanocrystalline CuO powders. “Fourth International Conference on Inorganic Materials”, Antwerp, Belgium, 19-21 September 2004.
4. E. N. Muhamad, B. L. Kniep, F. Girgsdies, R. E. Jentoft, T. Ressler, Microstructural Characteristics of Cu/ZnO Catalysts as Function of Precipitate Ageing. “XXXVII. Jahrestreffen Deutscher Katalytiker in Verbindung mit dem Fachtreffen Reaktionstechnik”, Weimar, Germany, 17- 19th March 2004.
5. B. L. Kniep, E. N. Muhamad, A. Szizybalski, F. Girgsdies, R. Schlögl, T. Ressler, Structural Modifica-tions of Cu/ZnO-based Catalysts as a Function of Precipitate Ageing. “EuropaCat VII” Sofia, Bulgaria, 28th August – 1st September 2005.
Papers published in manuscript and refereed journals:
1. I. Ramli, E.N. Muhamad, A.H. Abdullah, Y.H. Taufiq-Yap and S.B. Abdul Hamid, Influence of number of washing on the characterisatic of nanocrystalline copper oxide powders, Nanotech 2004, vol. 3, 99-102, ISBN: 0-9728422-9-2