Oxidation of methanol to formaldehyde

over Ag catalyst
A study of reaction pathways and Ag catalyst
restructuring during MTF

Kamilla Arnesen

Chemical Engineering and Biotechnology

Submission date: June 2018
Supervisor: Hilde Johnsen Venvik, IKP

Norwegian University of Science and Technology

Department of Chemical Engineering
ii
Abstract

Formaldehyde is recognised as an important industrial chemical, being a versatile organic

molecule. Due to high reactivity, it is used as a building block to synthesise other chemicals
and industrial products. Production of formaldehyde is mainly based on two reactions,
dehydrogenation and partial oxidation of methanol. Silver can be used as a catalyst and
the technology is recognised by Dynea AS as the process method with most potential for
competitive development, based on market and environmental considerations
In this master thesis silver catalyst morphology, activity and interaction between silver
and various components relevant to the formaldehyde synthesis have been studied. The
catalytic morphology was investigated for dependency on different atmospheres, exposure
time and process variables. Catalytic activity was measured for hydrogen, carbon monox-
ide and, to an extent, methanol oxidation reactions, in addition to focusing on structure
dependencies. The catalyst was characterised with XRD and SEM.
Polycrystalline silver catalyst was treated with various atmospheres, such as nitrogen, hy-
drogen, air and methanol/water, at elevated temperatures. It was also used as the catalyst
in oxidation reactions, performed with varying temperatures, to investigate the levels of
reactant conversion.
It was found that treating the silver with different atmospheres resulted in great changes
of the silver surface and morphology. Structural changes in the silver was observed when
oxygen was present. Silver particles treated in air showed increasing restructuring with
time, eliminating angular grain structure, production of pinholes and promoting irregular
and smooth facet formation. Using the silver as a catalyst for hydrogen, carbon monoxide
and methanol oxidation resulted in the same observations. Equal morphological changes
were observed on the mesoscale, for silver used as catalyst in hydrogen and carbon monox-
ide oxidation. The catalyst display high catalytic activity for hydrogen oxidation, increas-
ing with time. A declining trend was evident for catalytic activity for carbon monoxide
oxidation, suggesting the restructured silver promotes hydrogen oxidation, as opposed to
carbon monoxide oxidation. In general, oxygen display an important role promoting and
activating reactions and influence the silver catalyst morphology. This is especially critical
for silver catalyst activation and performance during process start-up in the formaldehyde
production.

iii
iv
Sammendrag

Formaldehyd er et allsidig organisk molekyl og er ansett som et viktig industrielt kjemikalie.

Grunnet dens høye reaktivitet blir det brukt som grunnstein for å syntetiseres andre kjemikalier
og industrielle produkter. Produksjonen av formaldehyd er i all hovedsak produsert ved
hjelp av to reaksjoner, dehydrogenisering og delvis oksidasjon av metanol. Sølv blir gjerne
bukt som katalysator for reaksjonene og er identifisert av Dynea As som den teknologien
med størst potensial for konkurransedyktig utvikling basert på økonomisk og miljømessige
betrakninger.
Denne masteroppgaven fokuseres det på sølvkatalysatorens morfologi, aktivitet og in-
teraksjonen mellom sølvet og komponenter relevant for produksjonen av formaldehyd.
Katalysatorens morfologi var undersøkt i sammenheng med ulike atmosfærer, eksponer-
ingstid og prosessvariabler. Katalytisk aktivitet ble målt for hydrogen, karbonmonoksid,
og til delvis, for oksidasjon av metanol. Det ble i tillegg fokusert på reaksjonens struktu-
ravhengighet. Katalysatoren ble karakterisert ved hjelp av XRD og SEM.
Polykrystalinsk sølvkatalysator ble utsatt for ulike atmosfærer som, nitrogen, hydrogen,
luft og metanol/vann-blanding, ved høye temperaturer. Sølvet ble også brukt som katalysator
for oksidajonsreaksjoner som ble utført ved varierende temperarurer for å undersøke om-
setningsgrad.
Resultatene viste at behandling av sølvet ved ulike atmosfærer gjorde store utslag for
sølvets morfologi. Strukturelle endringer i sølvet ble observert ved tilstedeværelse av
oksygen. Behandling med luft over lengre tid førte til økende restrukturering. Skarp
kornstruktur forsvant, hull og uregelmessige falsetter ble dannet. De samme morfologiske
forandringene ble observert i sølv som ble brukt som katalysator for hydrogen, karbon-
monoksid og metanol oksidasjon. Katalysatoren viste høy katalytisk aktivitet for oksi-
dasjon av hydrogen som økte med eksponeringstiden. En nedadgående trend ble sett for
den katalytiske aktiviteten for oksidasjon av karbonmonoksid. Dette tyder på at sølvets re-
strukturering fremmer oksidasjon av hydrogen, i motsetning til karbonmonoksid. Generelt
viser oksygen en viktig egenskap ved å promotere og aktivere reaksjoner, samt påvirke
sølvkatalysatorens morfologi. Dette er spesielt viktig for aktivering av sølvkatalysatoren
og dens prestasjon ved oppstart av formaldehydproduksjon.

v
vi
Table of Contents

Preface i

Abstract iii

Sammendrag v

List of Tables x

List of Figures xiv

Abbreviations xv

1 Introduction 1

2 Theory 3
2.1 Formaldehyde Production by Silver Catalyst . . . . . . . . . . . . . . . . 3
2.1.1 Interaction of Oxygen with Silver . . . . . . . . . . . . . . . . . 5
2.1.2 Oxidation of Hydrogen and Carbon Monoxide . . . . . . . . . . 6
2.2 Silver Catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Catalyst Characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.1 X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.2 Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . 11
2.4 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 Catalytic Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3 Experimental 17
3.1 Health, Safety and Environment . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Catalyst Characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.1 X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.2 Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . 18
3.3 Catalyst Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

vii
 3.3.1 Experimental set-up . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3.2 Calcination Experiments . . . . . . . . . . . . . . . . . . . . . . 22
3.3.3 Oxidation Experiments . . . . . . . . . . . . . . . . . . . . . . . 22
3.3.4 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4 Results 27
4.1 Catalyst Characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.1 X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.2 Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . 27
4.2 Catalyst Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2.1 Hydrogen oxidation . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2.2 Carbon monoxide oxidation . . . . . . . . . . . . . . . . . . . . 39
4.2.3 Methanol oxidation . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Discussion 45
5.1 Catalyst Characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.1.1 X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.1.2 Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . 46
5.2 Catalyst Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.2.1 Hydrogen Oxidation . . . . . . . . . . . . . . . . . . . . . . . . 48
5.2.2 Carbon Monoxide Oxidation . . . . . . . . . . . . . . . . . . . . 49
5.2.3 Methanol Oxidation . . . . . . . . . . . . . . . . . . . . . . . . 51

6 Conclusion 53

7 Further work 55

Bibliography I

Appendix V

A Methanol to formaldehyde test set-up VII

B Feed Analysis XIII

C Calculations XV

D Oxidation Experiments XVII

D.1 Hydrogen Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . XVII
D.2 Carbon Monoxide Oxidation . . . . . . . . . . . . . . . . . . . . . . . . XXI

viii
List of Tables

3.1 Variables and parameters used during the experiments on the Ag catalyst. 18
3.2 Composition of the calibration gases used to calibrate the GC. . . . . . . 21
3.3 Flow set points for hydrogen (top) and methanol/water (bottom) calcina-
tion experiments. The total flow was set to 250 Nml/min. . . . . . . . . . 22
3.4 H2 and CO oxidation experimental overview showing the temperature
range, time on stream and which catalyst bed was used. . . . . . . . . . . 24
3.5 Flow set points for H2 and CO oxidation experiments performed in the
MTF set-up. Experiments were performed with 2 % H2 , O2 /H2 :3 ratio
and a total flow of 250 Nml/min. . . . . . . . . . . . . . . . . . . . . . . 24
3.6 Flow set points for methanol oxidation experiment performed in the MTF
set-up. Experiments were performed with 8 mol% methanol, MeOH/O2 :2.314
ratio and a total flow of 500 Nml/min. . . . . . . . . . . . . . . . . . . . 24

4.1 Calculated crystallite size for a set of silver samples exposed to various
atmospheres, using the software Diffrac.Eva, K-value = 0.89 and FWHM. 28
4.2 EDX analysis of silver catalyst used in methanol oxidation experiment.
The spectrum relate to a position in Figure 4.10. . . . . . . . . . . . . . 36
4.3 Maximum hydrogen conversion for the hydrogen oxidation experiments at
the set point temperature of 700 ◦ C. . . . . . . . . . . . . . . . . . . . . 38
4.4 Maximum CO conversion for the carbon monoxide oxidation experiments
at the set point temperature of 700 ◦ C. . . . . . . . . . . . . . . . . . . . 41
4.5 Average carbon mass balance, methanol conversion and catalyst bed tem-
perature for the methanol oxidation experiment. . . . . . . . . . . . . . . 44

B.1 Feed analysis for a hydrogen oxidation experiment with 2 mol% H2 and
excess oxygen (6 mol%). . . . . . . . . . . . . . . . . . . . . . . . . . . XIII
B.2 Feed analysis for a carbon monoxide oxidation experiment with 2 mol%
CO and excess oxygen (6 mol%). . . . . . . . . . . . . . . . . . . . . . . XIV
B.3 Feed analysis for methanol oxidation experiment with MeOH/O2 :2.314
ratio. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIV

ix
 D.1 Temperatures trough the reactor at 20.5 cm axial position compared to the
temperature set point for the carbon monoxide oxidation experiment cata-
lyst bed 2, experiment 3, cooling cycle following the three day experiment.
The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . . . . . . . XVII
D.2 Temperatures trough the reactor at 20.5 cm axial position compared to the
temperature set point for the carbon monoxide oxidation experiment cata-
lyst bed 2, experiment 3, cooling cycle following the three day experiment.
The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . . . . . . . XXI

x
List of Figures

2.1 Illustration of the methanol to formaldehyde process using silver catalyst

at Dynea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Illustration of the interactions of various oxygen species with silver cata-
lyst at 923 K. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 Illustration of a fcc unit cell (left) and basal plane structures (100), (111)
and (110) (right). Stippled lines in figure to the right indicate the second
layer atom locations and a is the lattice constant determining the distance
between atoms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Illustration of on-top, bridge and hollow adsorption sites on fcc structure. 8
2.5 Sketch of a heterogeneous solid surface showing various surface sites. . . 9
2.6 An illustration showing X-ray reflection related to Bragg’s law in Equation
2.10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.7 Left: Sketch showing the interactions between the primary electron beam
and the sample in a electron microscopy setup. Right: Schematic set up of
an scanning electron microscope with x-ray detector. . . . . . . . . . . . 11
2.8 Illustration showing the general layout of an gas chromatograph. 1 = car-
rier gas, 2 = reduction valve, 3 = injector, 4 = oven, 5 = column, 6 =
detector, 7 = result processing. . . . . . . . . . . . . . . . . . . . . . . . 13
2.9 A sketch showing the working principle of a thermal conductivity detector
used for gas chromatography. . . . . . . . . . . . . . . . . . . . . . . . . 14

3.1 Image of the calcination set-up showing the eurotherm, valve system and
oven with reactor to the left and a close-up of the oven with the reactor to
the right. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Process flow diagram of the MTF test set-up per January 2018. . . . . . . 20
3.3 Catalyst bed for calcination experiments in the MTF test set-up. From the
left; quartz sinter, quartz wool and silver particles. . . . . . . . . . . . . . 22
3.4 Graph showing the temperature program used in oxidation experiments in
the MTF set-up. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

xi
 4.1 X-Ray Diffraction pattern for the silver catalyst exposed to various condi-
tions. Black = Fresh Ag, Red = N2 calcination (24h), Blue = CH3 OH/H2 O,
Green = H2 oxidation, Brown = CO oxidation . . . . . . . . . . . . . . . 28
4.2 SEM images of unused particular silver catalyst at different magnifications
((a): x500, (b): x1.00k) using secondary electron detector. . . . . . . . . 29
4.3 SEM images of silver catalyst exposed to nitrogen for 5 and 24 hours, at
different magnifications ((a): 5h, x300, (b): 24h, x1.00k) using secondary
electron detector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.4 SEM images of silver catalyst exposed to hydrogen (1 mol%) for 5 and 24
hours, at similar magnifications ((a): 5h, x1.00k, (b): 24h, x1.00k) using
secondary electron detector. . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.5 SEM images of silver catalyst exposed to air for 5, 24 and 72 hours, at
various magnifications ((a): 5h, x1.00k, (b): 72h, x1.00k, (c): 24h, x500k,
(d): 24h, x3.00k) using secondary electron detector. . . . . . . . . . . . . 31
4.6 SEM images of silver catalyst exposed to a methanol and water vapour at-
mosphere for 24 hours, at different magnifications ((a): x500, (b): x1.00k)
using secondary electron detector. . . . . . . . . . . . . . . . . . . . . . 32
4.7 SEM images of silver catalyst used in hydrogen oxidation experiments,
at various magnifications ((a): Catalyst bed 1, x650, (b): Catalyst bed 1,
x1.00k, (c): Catalyst bed 2, x500k, (d): Catalyst bed 2, x3.00k) using
secondary electron detector. . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.8 SEM images of silver catalyst used in carbon monoxide oxidation experi-
ments, at various magnifications ((a): Catalyst bed 1, x500, (b): Catalyst
bed 1, x1.00k, (c): Catalyst bed 2, x1.00k, (d): Catalyst bed 2, x3.00k)
using secondary electron detector. . . . . . . . . . . . . . . . . . . . . . 34
4.9 SEM images of silver catalyst used in methanol oxidation experiment, at
different magnifications ((a): x500, (b): x1.00k) using secondary electron
detector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.10 Left: SEM image from EDX analysis, of silver catalyst used in methanol
oxidation experiment at x1.50k magnification using X-ray detection. Right:
Image of the catalyst bed used for methanol oxidation after ended experi-
ment, showing dark spot on the top of the bed. . . . . . . . . . . . . . . . 35
4.11 Graphs showing the hydrogen conversion plotted against catalyst bed tem-
perature in the hydrogen oxidation experiments (H2 /O2 :1/3). (a): Reactor
heating, (b): Reactor cooling. . . . . . . . . . . . . . . . . . . . . . . . . 37
4.12 Graph showing the H2 conversion plotted against time on stream for the
hydrogen oxidation experiment (H2 /O2 :1/3) which lasted for three days
with the oven temperature kept stable at 700 ◦ C. . . . . . . . . . . . . . . 38
4.13 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in a hydrogen oxidation experiment
Catalyst bed 1, Experiment 1, for reactor heating (a) and cooling (b) cycle.
The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . . . . . . . 39
4.14 Error in hydrogen mass balance plotted against hydrogen conversion in the
hydrogen oxidation experiment using Catalyst bed 2, Experiment 1. . . . 40

xii
4.15 Graphs showing the CO conversion plotted against catalyst bed temper-
ature in the carbon monoxide oxidation experiments (CO/O2 :1/3). (a):
Reactor heating, (b): Reactor cooling. . . . . . . . . . . . . . . . . . . . 41
4.16 Graph showing the CO conversion plotted against time on stream for the
carbon monoxide oxidation experiment (CO/O2 :1/3) which lasted for three
days with the oven temperature kept stable at 700 ◦ C. . . . . . . . . . . . 41
4.17 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in a carbon monoxide oxidation ex-
periment Catalyst bed 1, Experiment 1, for reactor heating (a) and cooling
(b) cycle. The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . 42
4.18 Error in carbon mass balance plotted against CO conversion in the carbon
monoxide oxidation experiment using catalyst bed 1. . . . . . . . . . . . 43
4.19 Graph showing the conversion of methanol and oxygen, and selectivity
towards the carbon products; formaldehyde, carbon monoxide and car-
bon dioxide in the methanol oxidation experiment (CH3 OH/O2 :2.314 and
heating set points 200 - 600 ◦ C). . . . . . . . . . . . . . . . . . . . . . . 44
4.20 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the methanol oxidation experi-
ment, for furnace set points 200 - 600 ◦ C. The catalyst bed is located at
18.5 - 20 cm, marked by vertical dotted lines. . . . . . . . . . . . . . . . 44

A.1 Valve system controlling GC injections. . . . . . . . . . . . . . . . . . . VIII

A.2 Reactor details for experiments in the MTF test set-up with particular silver.VIII
A.3 Calibration curve for nitrogen MFC, line 1, low area. . . . . . . . . . . . IX
A.4 Calibration curve for nitrogen MFC, line 1, high area. . . . . . . . . . . . IX
A.5 Calibration curve for nitrogen MFC, line 2 . . . . . . . . . . . . . . . . . X
A.6 Calibration curve for air MFC. . . . . . . . . . . . . . . . . . . . . . . . X
A.7 Calibration curve for carbon monoxide MFC. . . . . . . . . . . . . . . . X
A.8 Calibration curve for hydrogen MFC. . . . . . . . . . . . . . . . . . . . XI
A.9 Calibration curve for LFC feeding methanol. . . . . . . . . . . . . . . . . XI
A.10 Calibration curve for LFC feeding a mix of methanol and water. . . . . . XI

D.1 Graph showing the conversion plotted against set point temperature in the
blank hydrogen oxidation experiment. . . . . . . . . . . . . . . . . . . . XVIII
D.2 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the blank hydrogen oxidation
experiment, for reactor heating. The catalyst bed would be located at 19 -
21 cm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XVIII
D.3 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the hydrogen oxidation experi-
ment catalyst bed 2, experiment 1, for reactor heating (a) and cooling (b)
cycle. The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . . . XIX
D.4 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the hydrogen oxidation experi-
ment catalyst bed 2, experiment 2, for reactor heating (a) and cooling (b)
cycle. The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . . . . . XIX

xiii
 D.5 Error in hydrogen mass balance plotted against hydrogen conversion in the
hydrogen oxidation experiment using catalyst bed 1, experiment 1. . . . . XX
D.6 Error in hydrogen mass balance plotted against hydrogen conversion in the
hydrogen oxidation experiment using catalyst bed 2, experiment 2. . . . . XX
D.7 Error in hydrogen mass balance plotted against hydrogen conversion in
the hydrogen oxidation experiment using catalyst bed 2, experiment 3,
duration of 3 days. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXI
D.8 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the blank carbon monoxide oxi-
dation experiment. The catalyst bed would be located at 19 - 21 cm. . . . XXII
D.9 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the carbon monoxide oxidation
experiment catalyst bed 2, experiment 1, for reactor heating (a) and cool-
ing (b) cycle. The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . XXII
D.10 Graph showing the temperature profile trough the reactor plotted against
the axial position of the thermocouple in the carbon monoxide oxidation
experiment catalyst bed 2, experiment 2, for reactor heating (a) and cool-
ing (b) cycle. The catalyst bed is located at 19 - 21 cm. . . . . . . . . . . XXIII
D.11 Error in carbon mass balance plotted against CO conversion in the carbon
monoxide oxidation experiment using catalyst bed 2, experiment 1. . . . . XXIII
D.12 Error in carbon mass balance plotted against CO conversion in the carbon
monoxide oxidation experiment using catalyst bed 2, experiment 2. . . . . XXIV
D.13 Error in carbon mass balance plotted against CO conversion in the car-
bon monoxide oxidation experiment using catalyst bed 2, experiment 3,
duration of 3 days. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXIV

xiv
Abbreviations

BSE = Backscattered Electrons

CEM = Controlled Evaporation and Mixing
CO = Carbon Monoxide
EDX = Energy Dispersed X-ray
FCC = Face Centred Cubic
GC = Gas Chromatoraph
H2 = Hydrogen
HSE = Health, Safety and Environment
LFC = Liquid Flow Controller
MeOH = Methanol
MFC = Mass Flow Controller
MTF = Methanol to Formaldehyde
NTNU = Norwegian University of Science and Technology
SE = Secondary Electrons
SEM = Scanning Electron Microscopy
TCD = Thermal Conductive Detector
TOF = Turnover Frequency
TOS = Time on Stream
TPD = Temperature Programmed Desorption
XRD = X-Ray Diffraction

xv
xvi
Chapter 1
Introduction

Formaldehyde is recognised as an important industrial chemical with a yearly production

of 29 million tons (2010) [1]. It is a versatile organic molecule and due to high reactivity,
it is used as a building block to synthesise other chemicals and industrial products. Some
products are resins for wood panel industry, polyacetal and other acetylenic chemicals
making it important for construction, automotive, and the pharmaceutical industry [2].
Formaldehyde production has a long history and was knowingly first synthesised by hy-
drolysing methylene acetate in 1859. Production went from small to industrial scale after
1923 as a result of the high-pressure synthesis of methanol by BASF [3]. Since then, for-
malin production, formaldehyde dissolved in water, has expanded by focusing on process
and reaction efficiency [2].
Today, two process technologies dominate and account for the majority of formaldehyde
production. Initially, a gauze made of silver wires were used as a catalyst, which later have
been replaced with a shallow catalyst bed made of silver crystals where methanol is feed
in excess with oxygen and H2 O at 600 - 680 ◦ C. Continued development have resulted in
the use of iron oxide-molybdenum oxide catalyst. This process operates with excess air
and a near complete conversion of methanol at 270 - 400 ◦ C [2].
Based on market and environmental considerations, the silver catalyst technology has been
recognised by Dynea AS as the process method with most potential for competitive devel-
opment [4].
As a way to develop the formaldehyde process, several aspects are under investigation,
one being to increase the understanding behind the reactions and catalytic system. Side re-
actions, gas phase reactions, changes in catalyst morphology and activity are among some
of the challenges. This project aims to focus on reaction pathways, the silver catalyst and
contribute to the investigation of the interaction between silver and various components
present in the formaldehyde synthesis.

1
Chapter 1. Introduction

2
Chapter 2
Theory

2.1 Formaldehyde Production by Silver Catalyst

Producing formaldehyde from methanol accounts for about a third of the worlds methanol
consumption. The synthesis is mainly based on two reactions shown in Equations 2.1 and
2.2. The former, dehydrogenation, is endothermic, while the latter, partial oxidation, is
exothermic [5].

CH3OH CH2O + H2 ∆H 84 kJ/mol (2.1)

1
CH3OH + O CH2O + H2O ∆H − 159 kJ/mol (2.2)
2 2
The two reactions can be carried out simultaneously over a silver catalyst bed at atmo-
spheric pressure and temperatures of 600 - 680 ◦ C, where the exothermic partial oxidation
reaction supplies heat to the dehydrogenation. Several side reactions also takes place pro-
ducing carbon dioxide, methyl formate and formic acid. Carbon dioxide can be produced
by complete oxidation of methanol shown in Equations 2.3 and 2.6.
Typically methanol, air and water are mixed and evaporated in a vaporiser and heated
to reaction temperature before being fed to the reactor. The reaction occur on a silver
catalyst bed (10 - 50 mm). To avoid product decomposition, the product gas is cooled
directly after the reactor in a waste heat boiler and fed to an absorber. Water is mixed in to
stabilise the product and yield formalin (37 wt% formaldehyde in water). Formaldehyde
is an unstable chemical and can polymerise easily. Methanol (3 - 12 wt%) is added to the
formalin solution as an inhibitor for this purpose. It is also important to minimise formic
acid content to increase product quality. The formaldehyde synthesis is run at adiabatic
conditions, making temperature control important. Methanol is recycled and fed in excess

3
Chapter 2. Theory

Figure 2.1: Illustration of the methanol to formaldehyde process using silver catalyst at Dynea.

for this purpose, in addition to avoid operating in the explosive methanol - air area [5]. This
process can yield a near complete methanol conversion with about 90 - 92 % selectivity
towards formaldehyde [6] and is illustrated in Figure 2.1.
Silver catalyst is operational from a few months to a year before regeneration is needed,
which usually is done by electrolysis. A common reason for decreasing catalytic per-
formance is sintering, which happen at high reaction temperatures, and poisoning, such
as iron. Water is generally added to the feed at a methanol/water 60/40 weight ratio. It
is often called water ballast process and is shown to increase the catalyst lifetime by dis-
tributing the heat across the catalyst bed and lower the risk of sintering and coke poisoning.
Coke can be produced by methanol decomposing to carbon, water and hydrogen as shown
in Equation 2.4. As the risk of overheating is lowered with adding water to the feed, the
oxygen concentration can be increased, improving the conversion and yield [3].

3
CH3OH + O CO2 + 2 H2O ∆H − 674 kJ/mol (2.3)
2 2

CH3OH C + H2O + H2 ∆H − 31 kJ/mol (2.4)

Qian et al. have performed a study comparing formaldehyde synthesis with and without
water in the feed. The goal was to investigate the effect of water on the catalyst, formalde-
hyde selectivity and study the effect it had on the production of byproducts. The research
concluded that adding water to the reaction feed increased the methanol conversion and
gave a better selectivity towards formaldehyde [7]. In addition, it reduced the formation of
the byproduct formic acid, which is unwanted as it is a corrosive acid and will contribute
to polymerisation of formaldehyde [7, 2]. Formaldehyde is an unstable chemical and can
decompose at high temperatures to H2 or CO, oxidise to CO2 and H2 O or formic acid

4
 2.1 Formaldehyde Production by Silver Catalyst

in the presence of oxygen, as shown in Equations 2.5 to 2.7. The complete oxidation of
formaldehyde and formation of formic acid are strongly exothermic reactions which may
produce local hot spots and sintering of the catalyst [7].

CH2O CO + H2 ∆H 12 kJ/mol (2.5)

CH2O + O2 CO2 + H2O ∆H − 514 kJ/mol (2.6)

1
CH2O + O HCOOH ∆H − 314 kJ/mol (2.7)
2 2

In addition to increase the catalyst lifetime, water can also influence catalyst activity [3].
Bao et al. investigated the effect of water on bonds created on a silver (111) surface
discovering correlations between promotion of surface restructuring and the formation of
oxygen species active in the formaldehyde synthesis, promoting formaldehyde selectivity
[8].
Oxygen in the feed is primarily used by the exothermic reactions in Equation 2.2 and to
some extent in Equations 2.3 and 2.6. By adjusting air in the feed, the reactor and cata-
lyst bed temperature vary, controlling to what extent the endothermic reactions occur, in
Equations 2.1 and 2.5. The reaction in Equation 2.5 is not a desired reaction and shows the
importance of cooling the product stream after the reactor [2]. Waterhouse et al. concluded
with an ideal methanol/oxygen feed ratio of 2.25 when studying industrial formaldehyde
manufacturing with a focus on examining the role of oxygen species in methanol oxidation
on silver [9].

2.1.1 Interaction of Oxygen with Silver

A complete reaction mechanism for the formaldehyde synthesis is still not known and
several theories exist. The two main theories will be mentioned further.
Wachs and Madix performed experiments using isotopic labelling showing absorbed atomic
oxygen must be present to promote methoxy, formaldehyde and formate species and cre-
ated a subsequent reaction mechanism [10]. Andreasen et al. used this model as a basis
when developing a microkinetic model for oxidation of methanol to formaldehyde and
oxidation of formaldehyde to carbon dioxide [11]. They argued only one type of atomic
oxygen, Oα , is active in the formaldehyde synthesis and based the model on a consecutive
reaction path and kinetic calculations.
Researchers have also studied the interactions between oxygen and silver with regards to
the formaldehyde synthesis to explain the reaction mechanism [9, 12, 13]. Three species
of atomic oxygen have been identified as Oα , Oβ and Oγ . The oxygen species exhibit
different properties and behaviour with regards to silver interactions and formaldehyde
formation, as illustrated in Figure 2.2. Oα is agreed to be chemisorbed atomic oxygen

5
Chapter 2. Theory

Figure 2.2: Illustration of the interactions of various oxygen species with silver catalyst at 923 K.
Figure is obtained from [13]

species creating bridge bonds (Ag-O-Ag) on the silver surface on the low index planes
of silver. This oxygen is related to partial oxidation of methanol to formaldehyde and
oxidation of methanol and formaldehyde to carbon dioxide and dominates at temperatures
lower then 350 ◦ C [12]. At elevated temperatures oxygen may dissolve in the silver bulk
and grain boundaries to create Oβ , which is associated with surface restructuring of the
silver [9]. The last oxygen species, Oγ , is a result of volume diffusion of Oβ through the
silver bulk at high temperatures, 600 ◦ C, creating strong terminal bonds (Ag=O), which
catalyse dehydrogenation of methanol to formaldehyde and hydrogen [7]. Bulk dissolved
oxygen reacting with hydrogen creating water is also discussed to be the source of hole
formation in the catalyst surface [14].
Even though oxygen is not directly involved in all reactions creating formaldehyde it is
an important factor in the conversion of methanol and formaldehyde yield, as discovered
by Qian et al. During a 300 days MTF experiment the catalyst was temporarily exposed
to an oxygen-free atmosphere from steady-state. When the normal oxygen feed returend
an decrease in methanol conversion and a shift in selectivity towards carbon dioxide and
formic acid was observed, associating this behaviour with an increased amount of the the
weakly bound atomic Oα on the silver surface [7].

2.1.2 Oxidation of Hydrogen and Carbon Monoxide

According to the industrial review by Millar and Collins, there is still much to understand
and learn to improve the industrial production of formaldehyde. Some of the concluding
remarks concerning improvements were about process optimisation and eliminating sil-
ver poisons to expand the catalyst lifetime [3]. This would include obtaining a greater
understanding of the reaction pathways and production of byproducts.
Two of these products are water and carbon dioxide. Production of water and carbon

6
 2.2 Silver Catalyst

dioxide are included in several reactions previously shown, and are suggested signs of
reactant and product decomposition. The amount of water can also indicate which of
the reaction pathways dominate and produce formaldehyde (dehydrogenation or partial
oxidation of methanol). For the production of formaldehyde it is naturally most beneficial
to maximise the production of the desired product and minimise byproducts. It is also of
interest to understand the source and the amount being produced of byproducts to indicate
to what extent methanol combustion and formaldehyde decomposition occur during the
reaction [3].
Deactivation of silver, as mentioned, is caused by poisons and sintering. A suspected
cause of sintering is the presence of water vapour at elevated temperatures and as a result
of deactivation the increased selectivity towards the byproducts carbon dioxide and formic
acid occur [6, 14]. In addition to the role of oxygen acting as a promoter of silver surface
reconstruction, a study by Millar et al. suggest the production of water and carbon dioxide
by oxidation can be the cause of pinhole formation in the silver surface [14].
Qu et al. performed experiments concluding with oxygen pretreated silver exhibit good
activity towards the oxidation of carbon monoxide at room temperature, suggesting carbon
dioxide could be a product from carbon monoxide oxidation [15]. The activity at higher
temperatures are not certain. Work by Van Santen state that recombination of adsorbed
oxygen and CO to CO2 occurs on most transition metal surfaces, but have a maximum
rate as a function of temperature and CO surface coverage. The reaction rate increase with
temperature to the point where the adsorption equilibrium of CO is shifted towards the
gas phase. Experiments have also been performed with single surface Pt crystal, which
stabilise the structure by shifting from (100) to (111) structure, in a vacuum. The absorp-
tion of CO activate the surface towards creating a more reactive (100) surface, promoting
CO2 formation. Desorption of the product promote surface restructuring toward the stable
(111) structure and a oscillatory behaviour [16].
If the silver surface is active for oxidation of hydrogen and carbon monoxide, these reac-
tions are competing for consumption of oxygen. The exothermic oxidation reactions are
shown in Equations 2.8 and 2.9.

1
H2 + O H2 O ∆H − 243 kJ/mol (2.8)
2 2

1
CO + O CO2 ∆H − 283 kJ/mol (2.9)
2 2

2.2 Silver Catalyst

In heterogeneous catalysis the common interface is that between solid and gas. The chem-
ical reaction occur on said surface on active catalytic sites. The surface can exhibit local
variations resulting in individual site reactivity. These sites can activate the reactant differ-
ently and based on the geometry of the site, the interaction can result in different adsorbed

7
Chapter 2. Theory

Figure 2.3: Illustration of a fcc unit cell (left) and basal plane structures (100), (111) and (110)
(right). Stippled lines in figure to the right indicate the second layer atom locations and a is the
lattice constant determining the distance between atoms. The figures are obtained from [18].

Figure 2.4: Illustration of on-top, bridge and hollow adsorption sites on fcc structure. The figure is
obtained from [18].

states on the metal surface. These states are important because it is the interaction between
molecule and active site that decide which bonds are broken and created making product
selectivity the result of molecule activation. Metallic surfaces also effect the formation of
molecular states in the absorbate making the basal plane structure and orbital interactions
a factor which also determine reactivity and product formation [17].
Silver is a frequently used catalyst with a face-centred cubic crystal structure (fcc), which
is one of the closest packed spheres with only 26 % empty space. For cubic structures
the main basal planes are (100), (110) and (111), as defined by Miller indices. The fcc
structure and illustration of the basal planes are shown in Figure 2.3, where the broken lines
indicate the second layer of atoms and a is the lattice constant determining the distance
between atoms [18] .
Metal surface atoms change their bonding geometry to adapt to not having a complete
set of nearest amtomic neighbours, resulting in different properties than atoms in the bulk
material. The number of unsaturated bonds and increased energy level effect the activity
of a site on the catalyst resulting in different adsorption sites. Some of the most common
sites are illustrated in Figure 2.4, showing the geometry of on-top, bridge and hollow sites
on the fcc structure. Complex surfaces also contain defects, such as steps, kinks, terraces,
adatoms and vacancies, which are illustrated in Figure 2.5. Due to the different energy
levels some bonds are more likely to be broken at step and kink sites then at terraces [19].
In polycrystalline materials, composed of small crystals or grains, the grain boundary
create an interface similar to the surface with increased surface energy and reactivity. An

8
 2.2 Silver Catalyst

Figure 2.5: Sketch of a heterogeneous solid surface showing various surface sites. The figure is
inspired by [17].

example of a grain boundary in polycrystalline silver is illustrated in Figure 2.2 [20].

When a surface is created, bonds may be broken, resulting in increased surface free energy.
Some surface processes and phenomenons are therefore a result of minimising the surface
free energy, such as creating a dense packed surface, maximising the amount of nearest
neighbours and keeping impurities at the surface [18]. The minimum energy configuration
for silver is the closed packed (111) structure. Facet formation require atom mobility to
reconstruct both surface and bulk material and occur more easily at temperatures above
345 ◦ , which is the Tammann temperature for silver. The Tammann temperature is defined
as the temperature where the atoms has gained sufficient thermal energy to clearly notice
increased bulk mobility and activity [13].

With regards to the formaldehyde process it is known that the silver morphology effects
the reactivity and methanol selectivity towards formaldehyde, especially during the start-
up period of the process. Electrolytic silver vary in shape, size, surface area, exposed
plains and defects depending on supplier. These variations effect the time until steady-
state production is reached and the silver structure has reformed [6]. The changes can be
both thermally or reaction induced, where the former result in a decrease in surface free
energy. Morphological changes occurring as a result of reactions, between reactants or
interactions with the catalyst, produce a structure different then what would occur in an
inert atmosphere.

Oxygen is believed to induce a facet formation in silver to form thermodynamic favourable

step or kink sites both by surface absorbed oxygen (Oα ) and oxygen incorporated in the
bulk (Oβ ). This result in a steady-state morphology for the silver in the formaldehyde
syntheses as an equilibrium between thermal and oxygen induced influences [12].

Another observed morphological change in the silver surface from formaldehyde synthe-
sis is the formation of so-called pinhole defects [6, 14]. Millar et al. preformed in-situ
observations of pinhole formation at reaction temperatures above 600 ◦ C. The holes were
found in areas with surface defects, such as edges, and are believed to be the result of near-
surface explosions created by sub-surface hydroxy recombination [14]. Waterhouse et al.
suggest the hole formation may be due to the hydrostatic pressure created in the reaction
between bulk-dissolved oxygen and hydrogen [6].

9
Chapter 2. Theory

Figure 2.6: An illustration showing X-ray reflection related to Bragg’s law in Equation 2.10. The
figure is obtained from [18].

2.3 Catalyst Characterisation

Characterization of a catalyst involves the use of a series of techniques to identify physical,

chemical and mechanical properties. It is an important part with the goal to gain a better
understanding of how the catalyst work and how it can be improved and tailored to a
specific need [18].

2.3.1 X-Ray Diffraction

X-ray diffraction (XRD) is a structural analysis technique and is often used to identify
crystalline phases in the bulk and particle size in a crystalline material. Emitted X-ray
photons are scattered by atoms in the sample and a diffraction pattern is created by the
scattering which, due to the lattice sample parameters are unique for a particular com-
pound. Diffraction patterns and lattice spacing are related by Bragg’s law in Equation 2.10
[18],

nλ = 2dsinθ; n = 1, 2, ... (2.10)

where n is an integer, λ is the X-ray wavelength, d is the distance between two lattice
planes and θ is the angle between emitted and reflected X-ray radiation normal to the
reflecting plane, as shown in Figure 2.6.
The technique have some drawbacks such as the XRD pattern is effected by the crystal
size and amount of crystalline phase present in the sample and particles that are too small
or amorphous can not be detected. Some of the limitations of the method present it self as
a change in peak shape and intensity, but this information can be used to estimate particle
size via the Scherrer formula in Equation 2.11. This formula relates crystal size and peak
width,

Kλ
L= (2.11)
βcosθ

10
 2.3 Catalyst Characterisation

Figure 2.7: Left: Sketch showing the interactions between the primary electron beam and the sample
in a electron microscopy setup. Right: Schematic set up of an scanning electron microscope with
x-ray detector. The figure is inspired by [21].

where L is particle dimension, K is a constant, λ is the X-ray wavelength, β is the peak

width (often at half peak height) and θ is the angle between emitted and reflected X-ray
radiation normal to the reflecting plane [21].

2.3.2 Scanning Electron Microscopy

Electron microscopy is a frequently used technique investigating the field of surface sci-
ence. It provides a way of inspecting the surface of a material by using electrons to create
images with magnifications up till one million. A high energy electron beam is used and
by studying the effect it exert on the sample a number of qualities can be investigated, such
as illustrated to the left in Figure 2.7. The interactions can provide information related to
morphology, crystallography and chemical composition of the material [21].
In a scanning electron microscope (SEM) the electron beam is focused by a set of lenses
and scanning coils, to analyse a small part of the sample, which is set in a vacuum. Figure
2.7 (right side) shows a typical SEM set up. The energy level of the primary electron
beam is usually in the range of 0.2 - 40.0 keV. Interactions between electron beam and the
surface result in diagrams due to detection of backscattered electrons (BSE) or secondary
electrons (SE) emission. In SEM the image contrast is a result of surface topography and
sample composition. Backscattered electrons carry information about the heavy elements
in a sample producing a contrast between elements. The secondary electrons are lower
in energy then BSE, but give a better resolution. Surfaces facing the detector, and edges
appear brighter then the surfaces facing away, crating a three dimensional effect [21].
Emitted X-ray radiation is characteristic for an element and can be used to identify chem-
ical composition. This technique is called energy-dispersed x-ray analysis (EDX). To-

11
Chapter 2. Theory

gether, these methods can create a good representation of a material surface.

SEM is a non destructive method and it can easily be used to investigate composition
and changes in a catalyst surface. A disadvantage with SEM compared to other electron
microscopy methods, is the low resolution (5 - 10 nm). Therefor, it is mainly used to
characterise particles larger then 10 nm. Also, it may not provide a presentable image
of a large surface, as only specific areas are scanned. The material must also be electron
conducting, an issue that can be dealt with by coating the material, with, for example,
carbon or gold [22].

2.4 Gas Chromatography

Chromatography is a frequently used analysis technique based on separation to identify

molecular composition and quantities. It is based on the principle where samples will sep-
arate between two phases, one stationary and one mobile phase, due interactions between
sample and the phases. The mobile phase is often a gas or liquid, defining the difference
between gas and liquid chromatography. Several factors are important to achieve a suc-
cessful separation of substances. Some of the conditions are the nature of the stationary
and mobile phase, temperature and pressure [23].
A typical gas chromatograph (GC) set-up consists of several parts, where each has an im-
portant role. Figure 2.8 illustrate a typical set-up, highlighting the main parts. A carrier
gas, the mobile phase, is often nitrogen, helium or hydrogen because it acts as an inert in
the system and will transport the sample from the injection system trough the column to
the detector. The sample is injected onto the column as a gas. When the sample and mobile
phase reach the column the separation begins. Components can be separated by properties
as weight, boiling point, volatility and polarity depending on the packing material, sta-
tionary phase, in the column. Last, the elements are detected and creates a signal based
on mass or concentration. The signal made from detection is transformed to crate peaks
in a chromatogram. Further result processing can be done on a computer [23]. An ideal
detection system should result in sharp, symmetric, base separated peaks. The following
section will explain more about the column and detector system.
Column
A solute migrating trough a column will, depending on its properties, interact with the
stationary phase and absorb temporarily onto the liquid surface. This absorption cause the
elements to travel through the column with a different velocity then the carrier gas. As a
goal, the columns are tailored to be highly efficient, resulting in successfully separating
solutes. Open-tubular capillary columns are the most common choice for GC analysis.
The walls of the column can be coated with a film of liquid polymer (wall-coated open-
tubular), where the side branches of the polymer are tailored to get suitable properties.
Another option is coating the walls with a porous solid layer (porous-layer open-tubular)
with a large surface area. Capillary columns often have a inner diameter of 0.2 - 0.7 mm
and can be 10 - 100 m long. The capillary columns generally have a high efficiency due
to the coating method and the column length, resulting in a high theoretical plate height,

12
 2.4 Gas Chromatography

Figure 2.8: Illustration showing the general layout of an gas chromatograph. 1 = carrier gas, 2 =
reduction valve, 3 = injector, 4 = oven, 5 = column, 6 = detector, 7 = result processing. The figure is
obtained from [23].

a parameter for achieving efficient separation. It is important that the stationary phase is
thermally stable and does not react with the samples [24, 25].

Detector

There are many different techniques to choose from when selecting a detection method.
Flame ionisation, thermal conductivity detection and mass spectrometer are some of the
frequent choices. The detection principle is either based on mass or concentration.

The thermal conductivity detector (TCD), which is used in this thesis experimental work,
detects concentration levels. The concentration is proportional with the resulting signal
and the method is therefor sensitive to varying flow rates. TCD is a universal, non-
destructive detector and is mostly used for detection of organic and in-organic compounds,
more specifically light and inert gasses.

As all compounds have the ability to absorb heat in some capacity, the detector is set up to
measure the heat difference between two chambers. The carrier gas acts as a reference and
flows through one heated chamber with a filament. An equilibrium will develop between
the gas and the filament and this acts as a baseline against the second chamber. Here, gas
exiting the column will flow trough. When a solute pass through the chamber some of the
heat is absorbed and the difference in temperature can be detected as a change in current
compared to the reference. This signal generate a peak in the chromatogram. Figure 2.9
show a sketch of a TCD system [24].

To achieve a sutible base line, carrier gas with a high thermal conductivity is used, as
helium and hydrogen. A drawback with the system is the detectors low sensitivity, com-
pared to other detectors, making it not ideal for detecting low concentrations. Eventually,
the filaments burn out and need replacement. The lifetime is dependent on the system
temperature and the corrosively and reactivity of the samples [23].

13
Chapter 2. Theory

Figure 2.9: A sketch showing the working principle of a thermal conductivity detector used for gas
chromatography. The figure is obtained from [24].

2.5 Catalytic Activity

Catalytic properties are measured to define catalyst performance. Activity measurements

can be accomplished by measuring the reaction rate, conversion, selectivity, turnover fre-
quency and stability. Some of the properties are specific to a set of process conditions and
can not be generalised [18].
Turnover frequency (TOF) is a value that indicates the activity per active site on a catalyst.
It is defined as number of molecules converted, or reacted, per active site, given in s−1 ,
as described in Equation 2.12. TOF is often a difficult catalyst property to specify as it
is effected by the methods used for preparation, characterisation and it requires specific
knowledge of the surface area under reaction conditions [26, 27].

Total number of reacted molecules

TOF = (2.12)
Total number of sites · time

Conversion is also a term used to define activity. It is often used together with selectivity,
as a fitting catalyst will produce the desired product. The definitions of conversion and
selectivity are presented in Equations 2.13 and 2.14. Measuring conversion at a set space
velocity or temperature for a fixed conversion are practical ways of measuring activity and
is often used when catalyst comparisons are of interest [26].

Amount of reagent consumed

Conversion = (2.13)
Amount supplied

Moles of desired product formed

Selectivity = (2.14)
Moles of desired product stoichiometrically possible

14
 2.5 Catalytic Activity

In this study conversion, selectivity and stability are used to measure activity.

15
Chapter 2. Theory

16
Chapter 3
Experimental

3.1 Health, Safety and Environment

Experimental work was performed at NTNU laboratories, in accordance with the existing
rules and regulations regarding health, safety and environment (HSE), including working
after normal work hours routines. The main risks associated with the work are using ex-
plosive and toxic gases at high temperature when operating the methanol to formaldehyde
set-up. The risk is acceptable, as the set-up is enclosed with proper ventilation system,
safety valves and controllers. Routines were followed to ensure a safe work environment,
such as leak testing, using hydrogen and carbon monoxide detectors and correct protective
gear. A risk assessment for the experimental work was prepared as a part of the require-
ments of the master thesis and is accessible in full, at the university risk assessment portal
with ID-number 22597.

3.2 Catalyst Characterisation

3.2.1 X-Ray Diffraction

X-ray diffraction was performed using a diffractometer of the model D8 DaVinci Ad-
vanced with the software Diffrac.Measuerment Centre v6.5.0. The instrument use a Cu
anode with Kα radiation, wavelength of 1.54Å. The samples were analysed using a method
for crystalline samples measuring at 2θ angles from 35 to 105 degrees with fixed V6 slit.
The fixed slit illuminates 6 mm of the sample at all angles. To ensure an homogeneous
sample with a level surface, the silver particles were ground with a mortar and pestle be-
fore added to a sample holder (Si 10 mm cavity). Further result processing was performed
using Diffrac.Eva v4.2.1.10 software for identifying metal phases and crystal size.

17
Chapter 3. Experimental

3.2.2 Scanning Electron Microscopy

Scanning electron microscope was used to analyse the silver samples after each exper-
iment, in addition to fresh catalyst. The instrument used was of the model Hitachi S-
3400N. A secondary electron detector was used to investigate the silver morphology. For
this analysis an acceleration voltage of 15 keV, with 40 % probe current at about 5 mm
working distance and the images were captured at a speed of 40 seconds. Silver samples
were mounted on a sample holder using carbon tape.
EDX analysis was performed on specific samples using an X-ray detector. The accelera-
tion voltage varied based on which element was of interest. For silver detection 7 keV and
15 keV were used and the probe current and working distance were increased to 50 % and
10 mm.

3.3 Catalyst Analysis

A series of experiments were performed to investigate catalytic activity and morphological
changes. Polycrystalline silver particles (0.25 - 0.50 mm) were used for all experiments,
produced by electrolysis at K.A. Rasmussen. To resemble industrial process conditions for
formaldehyde production the temperature of 650 ◦ C and interval between 200 to 700 ◦ C
were chosen for all experiments. Atmosphere and experiment length a varying parameter
during the testing. An overview of the experiments and parameters are shown in Table 3.1.

Table 3.1: Variables and parameters used during the experiments on the Ag catalyst.

Atmosphere TOS [hours] Temperature [◦ C] Activity

Nitrogen 5α , 24 650 -
Air 5α , 24α , 72 650 -
Hydrogen 5α , 24α 650 -
Methanol/Water 24 650 -
H2 oxidation - 200 - 700 Yes
CO oxidation - 200 - 700 Yes
Methanol oxidation - 200 - 600 Yes
α Performed during the Specialization Project

Experiments without catalyst were also performed to investigate gas phase activity. All
experiments were performed in a calcination oven set-up and in the methanol to formalde-
hyde (MTF) test set-up, to be presented in the following sections.

18
 3.3 Catalyst Analysis

Figure 3.1: Image of the calcination set-up showing the eurotherm, valve system and oven with
reactor to the left and a close-up of the oven with the reactor to the right.

3.3.1 Experimental set-up

Calcination set-up
The calcination set-up consists of gas cylinders, where the gas is feed through stainless
steel pipe lines with Swagelock fittings, to a glass reactor placed in an isolated oven. The
gas flow is regulated using a rotameter (0 - 120 ml/min) and is ventilated via a water
seal. A N-type thermocouple and Eurotherm were used to measure and regulate the oven
temperature and can be used to set a temperature program fit for the experiment. The
set-up is shown in Figure 3.1.
Methanol to formaldehyde test set-up
The methanol to formaldehyde (MTF) experimental set-up is mainly used to perform cat-
alytic activity testing related to the industrial MTF process. A process flow diagram for
the MTF set-up is presented in Figure 3.2.
The main components in the set-up consist a set of pressure regulators, mass flow con-
trollers (MFC), liquid tanks, vaporizer and a reactor, set in an isolated oven. Gas is feed
through four high pressure stainless steel pipe lines (1/4”), with Swagelock fittings. Two
lines are used for nitrogen, at different volumetric flow rates (4 - 200 Nml/min and 30
- 1500 Nml/min), one line is used for synthetic air (10 - 500 Nml/min) and the last is
for CO or H2 gas (1 - 100 Nml/min). Two stainless steel containers are used to hold
methanol/water and formaldehyde solution and are pressurised with helium gas. All the
pipe lines are equipped with pressure regulators and valves used for manual control. The
system is operated at ambient pressure and one mechanical safety valve is placed before
the reactor to prevent pressure build up. Gas and liquid are introduced to the system by

19
Chapter 3. Experimental

the mass and liquid flow controllers (Bronkhorst), which is digitally controlled by the soft-
ware, LabVIEW v.13. An evaporator (CEM - Controlled Evaporation and Mixing) is used
to mix and vaporise the feed before entering the fixed bed reactor, which is made of quartz
glass. Details regarding the reactor design is shown in Appendix A, Figure A.2. A bypass
line is also available to avoid the CEM and reactor.

Figure 3.2: Process flow diagram of the MTF test set-up per January 2018.

20
 3.3 Catalyst Analysis

A gas chromatograph is used to analyse the product and feed stream (Agilent Technolo-
gies 7890A) and is operated by the program GC ChemStation. An automatic valve sys-
tem, which is operated by pressurised air, control the sample injection. The GC uses
two columns, molsieve 5A and CP-sil 5 CB columns, to separate the product gas and use
thermal conductive detectors (TCD) for detection. The PLOT molsieve column is used to
detect light weigh gases, such as hydrogen, oxygen, methane, nitrogen and carbon monox-
ide. The WCOT CP-sil column separate compounds based on boiling point and is used for
methane, carbon dioxide, methanol, formaldehyde, formic acid and water separation. An
illustration of the automated injection valve system is shown in Appendix A, Figure A.1.

Pipelines from the evaporator to the GC are heated by electric heating bands and isolated
to prevent condensation and thermocouples (K-type) are used to monitor the temperature
with the software EasyVIEW. One thermocouple is placed inside the reactor and is used
to measure the temperature profile in the reactor by moving the thermocouple vertically.

Prior to the experiment leak tests are performed with nitrogen and leak test spray, followed
by hydrogen or carbon monoxide (2 mol%) and using a gas detector.

Mass and liquid flow controllers, in addition to the GC, were calibrated during the special-
ization project and master thesis project period.

Calibration

Mass flow controllers were calibrated using a bobble flow column, where a flow was set
in LabVIEW and the bubble travel time was measured corresponding to a fixed column
volume. The liquid flow controller was calibrated by weighing samples of methanol/water
and pure methanol produced by the controller for a set time period. Further, the GC was
calibrated by analysing calibration gas with compositions shown in Table 3.2. Parameters,
retention time, concentration and area, were adjusted based on the calibration gas analyses.
The GC was also calibrated for formaldehyde, by manual injection of a standard solution
(27.6 mol%).

Calibration curves for the flow controllers are shown in Appendix A, Figures A.3 to A.10.

Table 3.2: Composition of the calibration gases used to calibrate the GC.

Concentration
Component
[mol%]
Hydrogen 5
Methane 1
Carbon dioxide 1.5
Carbon monoxide 0.5
Nitrogen Balance
Methanol 0.5
Nitrogen Balance

21
Chapter 3. Experimental

3.3.2 Calcination Experiments

Calcination experiments were performed to investigate the effect of varying atmospheres

had on the silver morphology.
Experiments with nitrogen (5.0) and air (synthetic) were performed in the calcination set-
up. Silver catalyst was added to a calcination reactor (∼ 1 g), using a quartz sinter as
support, and installed in an isolated oven. A gas line was connected to and from the
reactor and flow was set using a rotameter. The total flow for the experiments were set to
100 ml/min. An Eurotherm was programmed to hold the oven temperature at 650 ◦ C at a
given time period and set the heating and cooling rate at 10 ◦ C/min.
Experiments performed with H2 (1 %) in nitrogen and a mix of methanol (8 mol%) and
water (H2 O/CH3 OH:1.34), were carried out in the MTF set-up. The feed compositions
are shown in Table 3.3. The temperature program is similar as for the previously men-
tioned experiments, but increased to 700 ◦ C to compensate for heat loss. Total flow for
the experiments were set to 250 Nml/min. The reactor was assembled with a silver bed
containing a quartz sinter, a layer of quartz wool as support and silver particles, as shown
in Figure 3.3.

Table 3.3: Flow set points for hydrogen (top) and methanol/water (bottom) calcination experiments.
The total flow was set to 250 Nml/min.

Flow
Component
[Nml/min]
Hydrogen 2.5
Nitrogen 247.5
Methanol 20
Water 23.6
Nitrogen 206.4

Figure 3.3: Catalyst bed for calcination experiments in the MTF test set-up. From the left; quartz
sinter, quartz wool and silver particles.

3.3.3 Oxidation Experiments

Oxidation experiments were performed with hydrogen, carbon monoxide and methanol to
investigate catalytic activity and how reactions influence the silver morphology. Experi-
ments were performed with both a heating and cooling sequence, between 200 and 700 ◦ C

22
 3.3 Catalyst Analysis

with a 45 minute hold at every 100 ◦ C, as shown in Figure 3.4. The temperature profile
trough the reactor was measured at every 100 ◦ C. An Eurotherm was used to set the tem-
perature program which was programmed using the software iTools Engineering Studio.
Product gas was continuously analysed by the GC.

Figure 3.4: Graph showing the temperature program used in oxidation experiments in the MTF
set-up.

Blank experiments, without catalyst, were performed to investigate gas phase reactions
and confirm the effect of the catalyst for H2 and CO experiments. The catalyst bed was of
identical design as described in Figure 3.3. For experiments using the same catalyst bed
the reactor was flushed with nitrogen between runs.
Previous to the experiments a feed analysis was performed by GC analyses to be used in
calculations. GC analyses were continued until stable values were achieved.
Hydrogen and Carbon Monoxide Oxidation
Hydrogen and carbon monoxide oxidation experiments were performed at conditions shown
in Table 3.4. Experiments were performed with catalyst (∼ 250 mg), H2 or CO (2 mol%),
in excess oxygen (6 mol%), which was supplied as air. Two different silver beds were used
to investigate the reproducibility of the experiment (referred to as catalyst bed 1 and 2), in
addition to see if the prolonged used of the silver effected the morphology. The catalytic
activity was also tested over time, for three days, followed by a cooling sequence. In Table
3.5 the flow set points are shown. Oxidation experiments were performed equally with two
exceptions, both occurred during the hydrogen oxidation experiments. The experiment us-
ing catalyst bed 1 were performed 1 mol% H2 and a temperature program with 30 minute
hold each 50 ◦ C. The experiment using catalyst bed 2, experiment 2 were performed with
a temperature program with 30 min hold each 100 ◦ C.
Methanol Oxidation
Only one methanol oxidation experiment with catalyst was performed due to equipment
malfunction. The experiment was performed with a similar temperature program as de-
scribed in Figure 3.4. About 450 mg catalyst was used with 8 % methanol and a methanol

23
Chapter 3. Experimental

Table 3.4: H2 and CO oxidation experimental overview showing the temperature range, time on
stream and which catalyst bed was used.

Temperature TOS
Experiment Catalyst bed
[◦ C] [hours]
H2 /CO oxidation Blank 200 - 700 6
Catalyst bed 1.1∗ 200 - 700 - 200 12
Catalyst bed 2.1 200 - 700 - 200 12
Catalyst bed 2.2◦ 200 - 700 - 200 12
Catalyst bed 2.3 700 - 200 72
∗ Performed with 1 mol% H2 30 min hold each 50 ◦ C for H2 ox.
◦ Performed with 30 min hold each 100 ◦ C for H2 ox.

Table 3.5: Flow set points for H2 and CO oxidation experiments performed in the MTF set-up.
Experiments were performed with 2 % H2 , O2 /H2 :3 ratio and a total flow of 250 Nml/min.

Flow Concentration
Component
[Nml/min] [mol%]
Nitrogen 230.0 92.0
O2 15.0 6.0
H2 /CO 5.0 2.0

to oxygen ration of 2.314, as provided by Dynea AS. Oxygen was supplied as air. In Table
3.6 the flow set points are shown. The experiment produced activity data until 700 ◦ C
was reached. For this type of experiment H2 O is usually a part of the reactant stream. In
this case, it was of interest to investigate if, and what effect was shown when H2 O was
removed from the feed, in addition to activity data.

3.3.4 Calculations

From the continuous GC analyses the information could be further analysed using Mi-
crosoft Excel. Conversion, defined as converted reactant, was calculated for the relevant
experiments. Nitrogen is an inert in the reaction system and was used as an internal stan-
dard in the calculations. To validate the calculations a mass balance over the system can

Table 3.6: Flow set points for methanol oxidation experiment performed in the MTF set-up. Ex-
periments were performed with 8 mol% methanol, MeOH/O2 :2.314 ratio and a total flow of 500
Nml/min.

Flow Concentration
Component
[Nml/min] [mol%]
Nitrogen 442.7 88.5
O2 17.3 3.5
Methanol 40.0 8.0

24
 3.3 Catalyst Analysis

be performed [28]. Examples on the calculations are shown in Appendix C.

Conversion
Conversion of a reactant can be calculated by Equation 3.1. The flow of any reactant
can be defined by Equation 3.2, where FT OT is the total flow and yi is the component
mole fraction, provided by the GC analysis for a given component, i. yi is generated by
the components thermal conductivity, transformed via an electric signal to a peak in the
chromatogram, with a specific area and response time.

Freactant,in − Freactant,out
Xreactant = (3.1)
Freactant,in

Freactant = yi · FTot (3.2)

Further, the reactant H2 is used as an example. Combining Equations 3.1 and 3.2, the
conversion of hydrogen can be expressed by Equation 3.3.

yH2 ,in FTot,in − yH2 ,out FTot,out

XH2 = (3.3)
yH2 ,in FTot,in

Next, with the assumption that all converted reactant produce the desired product and the
total flow into the system is not assumed to be equal to the total flow out of the system,
the total flow can be expressed using the inert, N2 . Finally, by combining Equations 3.3 to
3.5, the conversion can be expressed by Equation 3.6.

FN2 ,in
FTot,in = (3.4)
yN2 ,in

FN2 ,out
FTot,out = (3.5)
yN2 ,out

yH2 ,out yN2 ,in

XH2 = 1 − (3.6)
yH2 ,in yN2 ,out

Conversion can also be calculated by the amount of reacted oxygen or by the amount of
formed product.
Selectivity
As a measurement on how much desired product is produced based on consumed reactant
in the reaction selectivity is used. In this case, the carbon selectivity can be calculated as
a function of the total amount of converted methanol or as a function of the total amount
of produced carbon products. Selectivity can also be calculated based on hydrogen, but is
not done in the thesis due to the instability of water GC analys.

25
Chapter 3. Experimental

Equation C.3 show how to calculate the selectivity of CH2 O based on the total amount of
converted methanol. The carbon product method is shown in Equation 3.8 for component,
i.

FTot,out yCH2 O,out

SCH2 O = (3.7)
FTot,in yCH3 OH,in − FTot,out yCH3 OH,out

yC,i,out
SC,i = P (3.8)
yC,product,out

In this thesis Equation C.3 is used, due to uncertainty of the specific amount of carbon
products.
Mass balance
A mass balance over the system is calculated as a validation of the experimental set-up
and methods of detection. In general, an error with the limit of 5 % is accepted due to
systematic errors like gas leakages and simplifications.
As an example, by assuming all hydrogen either reacts to form water or leave the system
as un-reacted hydrogen the following component balance can be expressed in Equation
3.9.

FH2 ,in = FH2 ,out + FH2 O,out (3.9)

When applying the correlations in Equations 3.1 and 3.2, the error in the hydrogen balance
can be calculated by Equation 3.10,
P P
FTot,out nyH,out − FTot,in nyH,in
EH = P (3.10)
FTot,in nyH,in

where, n, is a stoichiometric factor. The same balance can be carried out for the compo-
nents, oxygen and carbon.

26
Chapter 4
Results

4.1 Catalyst Characterisation

4.1.1 X-Ray Diffraction

The X-Ray diffractrograms for the silver samples exposed to various atmospheres are pre-
sented in Figure 4.1, where Fresh Ag is indicated by black, N2 calcination (24 hours) by
red, MeOH/H2 O calcination by blue, H2 oxidation by green and CO oxidation by brown.
The intensity of the different phases are indicated by peak height. The crystalline face-
centred cubic silver and the phases were identified by comparison in the Diffrac.Eva soft-
ware database. The peaks exhibit similar characteristics and no great change is evident. In
the diffractogram representing N2 calcination, the (200) and (220) phases have increased
intensities compared to the fresh Ag. Peaks representing the (111) and (311) phases in the
silver used for CO oxidation also have increased in intensities.
Crystallite size for the samples were calculated using the Scherrer formula (see Chapter
2.3.1) in the Diffrac.Eva software, with a K-value of 0.89 and using full width at half
maximim (FWHM). The values are presented in Table 4.1 and the crystallite size has not
been effected noticeably by the various experimental treatment. Catalyst used in oxidation
experiments show a slight decrease in size compared to the fresh silver catalyst.

4.1.2 Scanning Electron Microscopy

Experiments using silver catalyst, as listed in Chapter 3, Table 3.1, were performed to in-
vestigate if the catalyst morphology was effected when exposed to various atmospheres at
industrially relevant temperatures and conditions. Scanning electron microscope (Hitachi
S-3400N) was used to analyse the silver and the scans will be presented in the following

27
Chapter 4. Results

Figure 4.1: X-Ray Diffraction pattern for the silver catalyst exposed to various conditions. Black =
Fresh Ag, Red = N2 calcination (24h), Blue = CH3 OH/H2 O, Green = H2 oxidation, Brown = CO
oxidation

Table 4.1: Calculated crystallite size for a set of silver samples exposed to various atmospheres,
using the software Diffrac.Eva, K-value = 0.89 and FWHM.

Crystallite size
Experiment
[nm]
Fresh Ag 37.5
N2 , 24h 37.9
CH3 OH/H2 O 37.9
H2 oxidation 33.1
CO oxidation 34.9

28
 4.1 Catalyst Characterisation

section. Image editing tools have been used to adjust, such as brightness and contrast,
to improve the image quality. Unfortunately, this will not guarantee for sufficient image
quality in print.

Fresh Silver Catalyst

A sample of unused silver catalyst was analysed to be used as a referance for comparison.
Figure 4.2 show the silver at two different magnifications. At low magnification, (a), a
clear faceted structure is visible and at higher magnifications, (b), a terrace structure and
surface defects in the metal are shown. The silver is produced by electrolytic refining,
promoting a structure with a large surface area.

(a) (b)

Figure 4.2: SEM images of unused particular silver catalyst at different magnifications ((a): x500,
(b): x1.00k) using secondary electron detector.

Nitrogen Atmosphere

The silver was exposed to nitrogen at 650 ◦ C for 5 and 24 hours in the calcination set-
up. From the images in Figure 4.3, structural changes in the silver are evident. In the
high magnification images, (b), the terrace structures are less pronounced, but still visible.
Widespread holes are also visible in the structures, but are not consistent throughout the
sample. The experiment length does not seem to significantly influence the hole size or
silver morphology.

Hydrogen Atmosphere

Figure 4.4 show SEM images from hydrogen (1 mol%) calcination in the MTF set-up, for
5 and 24 hours with temperature set point at 650 ◦ C. Due to a heat gradient in the set-up
the actual temperature in the catalyst bed was 600 ◦ C. Compared to the unused catalyst
sample, facet and terrace structures seem to be intensified in the hydrogen atmosphere and
minor hole formations are present. Lengthened exposure time does not seem to have a
great impact on the silver morphology.

29
Chapter 4. Results

(a) (b)

Figure 4.3: SEM images of silver catalyst exposed to nitrogen for 5 and 24 hours, at different
magnifications ((a): 5h, x300, (b): 24h, x1.00k) using secondary electron detector.

(a) (b)

Figure 4.4: SEM images of silver catalyst exposed to hydrogen (1 mol%) for 5 and 24 hours, at
similar magnifications ((a): 5h, x1.00k, (b): 24h, x1.00k) using secondary electron detector.

30
 4.1 Catalyst Characterisation

(a) (b)

(c) (d)

Figure 4.5: SEM images of silver catalyst exposed to air for 5, 24 and 72 hours, at various mag-
nifications ((a): 5h, x1.00k, (b): 72h, x1.00k, (c): 24h, x500k, (d): 24h, x3.00k) using secondary
electron detector.

Air Atmosphere

The silver was exposed to synthetic air at 650 ◦ C for 5, 24 and 72 hours in the calcination
set-up. From the images in Figure 4.5 great changes in topology for all samples are evident.
The morphological changes are more extensive as the exposure time increase, resulting in
larger and an increased number of holes at 72 hours (Figure 4.5 (b)). The angular grain
structure is eliminated and at higher magnification (Figure 4.5 (d)) well-defined facets are
visible. The facet structure and hole formation has only been clearly visible in calcination
experiments where oxygen has been present.

Methanol and Water Atmosphere

The silver was exposed to a CH3 OH/H2 O atmosphere for 24 hours at 650 ◦ C in the MTF
set-up. From the images in Figure 4.6 no great structural changes in the silver are evident
as similar to the fresh silver and nitrogen calcination experiment. During the experiment
GC analysis was performed and trace amounts of CH2 O and H2 were detected.

31
Chapter 4. Results

(a) (b)

Figure 4.6: SEM images of silver catalyst exposed to a methanol and water vapour atmosphere for
24 hours, at different magnifications ((a): x500, (b): x1.00k) using secondary electron detector.

Hydrogen Oxidation

Silver particles were used as a catalyst in hydrogen oxidation experiments producing water.
Two catalyst beds were used, one for a single experiment, noted as ”Catalyst bed 1”, where
the temperature was increased step-wise to set point temperature of 700 ◦ C and cooled
back to room temperature. This experiment was performed over the course of two days,
and the reactor was exposed to a nitrogen atmosphere over night at high temperature. The
second bed, ”Catalyst bed 2”, was used in two heating/cooling experiments, in addition to
a three day period at set point temperature of 700 ◦ C. Between the experiments the reactor
was exposed to a nitrogen atmosphere over a longer period of time.

In Figure 4.7 the SEM images for both the silver catalyst beds are shown and great changes
in topology for all samples are evident. The angular grain structure is eliminated and at
higher magnification (Figure 4.7 (d)) well-defined facets are visible, particularly in Cata-
lyst bed 2. Images of Catalyst bed 1 show more frequent and larges holes then Catalyst
bed 2, where the structure is smoother and holes are smaller.

Carbon Monoxide Oxidation

Silver particles were used as a catalyst in oxidation of carbon monoxide experiments pro-
ducing carbon dioxide. Two catalyst beds were used, one for a single experiment, de-
scribed as ”Catalyst bed 1”, where the temperature was increased step-wise to set point
temperature of 700 ◦ C and cooled back to room temperature. During the experiment the
laboratory experienced a power shortage, lasting about 45 min. For that time the tem-
perature slowly declined in the reactor and oven, but was increased back to correct tem-
perature when the power returned. The temperature was keep stable for one hour before
continuing the temperature program. The second bed, ”Catalyst bed 2”, was used in two
heating/cooling experiments, in addition to a three day period at set point temperature of
700 ◦ C. Between the experiments the reactor was exposed to a nitrogen atmosphere over
a longer period of time.

32
 4.1 Catalyst Characterisation

(a) (b)

(c) (d)

Figure 4.7: SEM images of silver catalyst used in hydrogen oxidation experiments, at various mag-
nifications ((a): Catalyst bed 1, x650, (b): Catalyst bed 1, x1.00k, (c): Catalyst bed 2, x500k, (d):
Catalyst bed 2, x3.00k) using secondary electron detector.

33
Chapter 4. Results

(a) (b)

(c) (d)

Figure 4.8: SEM images of silver catalyst used in carbon monoxide oxidation experiments, at var-
ious magnifications ((a): Catalyst bed 1, x500, (b): Catalyst bed 1, x1.00k, (c): Catalyst bed 2,
x1.00k, (d): Catalyst bed 2, x3.00k) using secondary electron detector.

In Figure 4.8 the SEM images for both the silver catalyst beds are shown and great changes
in topology for all samples are evident. The angular grain structure is eliminated and at
higher magnification (Figure 4.8 (c) and (d)) well-defined facets are visible, particularly in
Catalyst bed 2. A greater degree of smoothing also seem to be evident in Catalyst bed 2.

Methanol Oxidation

Silver particles were used as a catalyst in a oxidation of methanol experiment. Experi-

mental equipment broke down half way into the experiment, resulting in only 5 hours on
stream for the catalyst and being the only methanol oxidation experiment performed. The
catalyst bed was analysed by SEM and the result is shown in Figure 4.9. Dark spots were
noticed visually on top of the catalyst bed after use (Figure 4.10 (right)) and both bulk and
top of the catalyst bed were analysed. This was further investigated by EDX analysis.

In Figure 4.9 different morphologies are shown. Bulk analysis does not revile a great
structural change, as shown in, a, to the left. Clear terrace structure is present, similar as

34
 4.1 Catalyst Characterisation

(a) (b)

Figure 4.9: SEM images of silver catalyst used in methanol oxidation experiment, at different mag-
nifications ((a): x500, (b): x1.00k) using secondary electron detector.

Figure 4.10: Left: SEM image from EDX analysis, of silver catalyst used in methanol oxidation
experiment at x1.50k magnification using X-ray detection. Right: Image of the catalyst bed used for
methanol oxidation after ended experiment, showing dark spot on the top of the bed.

to fresh catalyst in Figure 4.2. Analysis of silver from the top of the bed clearly shows a
large number of holes produced on the surface.

EDX analysis was performed on silver from the top of the catalyst bed. Figure 4.10 (left)
show the scan used for the analysis and marks the locations of the EDX analysis. The
silver show no noticeable structural changes. Table 4.2 present the EDX results and the
elements detected on the surface of the silver particle. In the ”cloud-like” formation to
the right in the image, silver, carbon and oxygen are the elements detected. Silicon is also
detected, but only as trace amounts.

35
Chapter 4. Results

Table 4.2: EDX analysis of silver catalyst used in methanol oxidation experiment. The spectrum
relate to a position in Figure 4.10.

Spectrum 21 Wt% Spectrum 22 Wt%

Ag 100 Ag 90.8
C 9.2
Spectrum 23 Spectrum 24
Ag 52.0 Ag 100
C 46.0
O 2.0
Si 0.1

4.2 Catalyst Activity

Catalyst activity experiments were performed as described in Section 3.3.3. In the follow-
ing section results from the analysis will be presented.
Each experimental run produced flow and temperature logs, chromatograms from the GC
and temperature measurements from the thermocouple placed in the reactor. Experimental
data was compared based on the logged time stamps to correlate temperatures and GC data.
Also, to ensure correct and stable temperature for the data points. The GC chromatograms
provided data used to calculate conversion and selectivity of hydrogen, carbon monoxide
and methanol in the respective oxidation experiments. All calculations are based on Equa-
tions 3.6, C.3 and 3.10, presented in Section 3.3.4 and calculation examples are presented
in Appendix C.
Temperature logs and temperature profiles over the catalyst bed, were used to match av-
erage conversion data and stable reactor temperature, every 100 ◦ C between 200 ◦ C and
700 ◦ C, for both reactor heating and cooling sequences. The conversion was then plotted
against the appropriate catalyst bed temperature.
For each experiment the accuracy was calculated by a mass balance over the reaction. This
was done by using the mass balance Equation 3.10, in Section 3.3.4 and GC data.
Typical feed analysis for the various experiments are shown in Appendix B, Tables B.1 to
B.3.

4.2.1 Hydrogen oxidation

Hydrogen oxidation experiments were performed in the MTF set-up to investigate the
catalytic activity of silver for the reaction presented in Section 2.1, Equation 2.8. The ex-
periments were performed with polycrystalline silver catalyst, in addition to one reference
test without catalyst. For experiments with catalyst the reactor was set up as shown in
Section 3.3, Figure 3.3. Temperature profiles over the catalyst bed were measured for each

36
 4.2 Catalyst Activity

(a) (b)

Figure 4.11: Graphs showing the hydrogen conversion plotted against catalyst bed temperature in
the hydrogen oxidation experiments (H2 /O2 :1/3). (a): Reactor heating, (b): Reactor cooling.

experiment to observe the temperature gradient and to control where the exothermic reac-
tion is taking place. Temperature profiles and other information related to the experiment
are presented in Appendix D.
The experiments were performed with hydrogen (2 mol%) and excess oxygen (6 mol%) in
nitrogen and a temperature program as shown in Section 3.3.3, Figure 3.4. All, except for
the experiment using Catalyst bed 1, where 1 mol% hydrogen was used, and the temper-
ature was keep at a 30 minute hold every 50 ◦ C. The temperature profiles were measured
at every 100 ◦ C. Feed analysis to be used in conversion calculations, and an example of a
typical analysis for carbon monoxide oxidation is shown in Appendix B, Table B.1.
Initially, a blank experiment was performed at equal conditions as the experiments with
catalyst, without the cooling sequence. The experiment resulted in 11 % conversion of hy-
drogen at the maximum set point temperature of 700 ◦ C. The graph is shown in Appendix
D, Figure D.1 and show a distinct increase in conversion at 400 ◦ C. The temperature pro-
file, in Appendix D, Figure D.2 show the warmest location in the reactor at 20 cm, 613
◦
C.
Experiments with catalyst were performed as presented in Section 3.3.3, Table 3.4, were
two different catalyst beds were used, noted as Catalyst bed 1 and 2. Figure 4.11 present
the hydrogen conversion plotted against the catalyst bed temperature for the reactor heating
(a) and cooling cycle (b).
Comparing the two graphs in Figure 4.11, the conversion increase with increasing temper-
atures. During the heating cycle Catalyst bed 1 and 2, Experiment 1, show the same trend
(red squares and black circles), even though the conversion is higher for Catalyst bed 1.
Both graphs show a flat segment between 300 and 400 ◦ , before the conversion continues
to increase, to 85.1 and 77.7 %. The conversion of Catalyst bed 2, Experiment 2 (green tri-
angles), progress similarly to the graphs in the cooling cycle, where the conversion change
evenly with the temperature. In the cooling cycle the results are similar, except from the

37
Chapter 4. Results

Figure 4.12: Graph showing the H2 conversion plotted against time on stream for the hydrogen
oxidation experiment (H2 /O2 :1/3) which lasted for three days with the oven temperature kept stable
at 700 ◦ C.

Table 4.3: Maximum hydrogen conversion for the hydrogen oxidation experiments at the set point
temperature of 700 ◦ C.

H2 conversion
Experiment
at 700 ◦ C [%]
Catalyst bed 1.1 85.1
Catalyst bed 2.1 77.7
Catalyst bed 2.2 92.9
Catalyst bed 2.3 93.0

experiment with bed 1 (red squares), which again, have a higher conversion than the other
experiments. The third experiment performed with Bed 2 (yellow stars) in the cooling
cycle follows the three day experiment, and the maximum H2 conversion is equal to the
second experiment using the same Bed of 93 %. Table 4.3 show the maximum conversion
for all the experiments at the set point temperature of 700 ◦ C.
In Figure 4.12 the H2 conversion is plotted against time on stream for the three day ex-
periment at a constant temperature set point of 700 ◦ C, is shown. Over the course of the
experiment the hydrogen conversion is stable, fluctuating between 91.7 % and 93.0 %
conversion.
The temperature profile was measured at every 100 ◦ C during all the experiments and
an example is shown in Figure 4.13, for Catalyst bed 1, Experiment 1. The maximum
temperature measured over the catalyst bed is 640 ◦ C, with the oven set point temperature
at 700 ◦ C. The profiles for the heating cycle are shown in (a) and reactor cooling cycle in
(b), and at all temperature set points above 200 ◦ C, the warmest spot is located between 19
- 21 cm, where the catalyst bed is located. The remaining temperature profiles are shown

38
 4.2 Catalyst Activity

(a) (b)

Figure 4.13: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in a hydrogen oxidation experiment Catalyst bed 1, Experiment 1, for
reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

in Appendix D, Figures D.3 and D.4 and Table D.1.

Mass balance calculations for hydrogen were performed for all experiments to evaluate the
accuracy of the measurements made by the gas chromatograph. An example is shown in
Figure 4.14 for the experiment performed using Catalyst bed 2, Experiment 1, where the
error is plotted against the conversion of hydrogen. The mass balance error decrease with
increasing conversing of hydrogen. The remaining mass balances are shown in Appendix
D, Figures D.5 to D.7.

4.2.2 Carbon monoxide oxidation

Carbon monoxide oxidation experiments were performed in the MTF set-up to investigate
the catalytic activity for the reaction presented in Section 2.1, Equation 2.9. The exper-
iments were performed with polycrystalline silver catalyst, in addition to one reference
test without catalyst. For experiments with catalyst the reactor was set up as shown in
Section 3.3, Figure 3.3. Temperature profiles over the catalyst bed were measured for each
experiment to observe the temperature gradient and to control where the exothermic reac-
tion is taking place. Temperature profiles and other additional information related to the
experiment are presented in Appendix D.
The experiments were performed with carbon monoxide (2 mol%) and excess oxygen (6
mol%) in nitrogen and a temperature program as shown in Section 3.3.3, Figure 3.4. The
temperature profiles were measured at every 100 ◦ C. Feed analysis to be used in conversion
calculations, and an example of a typical analysis for carbon monoxide oxidation is shown
in Appendix B, Table B.2.
Initially, a blank experiment was performed at equal conditions as the experiments with
catalyst, without the cooling sequence. No carbon dioxide was detected at the maximum

39
Chapter 4. Results

Figure 4.14: Error in hydrogen mass balance plotted against hydrogen conversion in the hydrogen
oxidation experiment using Catalyst bed 2, Experiment 1.

set point temperature of 700 ◦ C, which was 615 ◦ C for the warmest spot in the reactor.
The temperature profile was measured during the experiment and is shown in Appendix
D, Figure D.8 and shows the warmest spot in the reactor to be at 20 cm. This is where the
catalyst bed would be located, at 19 - 21 cm.

Experiments with catalyst were performed as presented in Section 3.3.3, Table 3.4, were
two different catalyst beds were used, noted as Catalyst bed 1 and 2. Figure 4.15 present
the CO conversion plotted against the catalyst bed temperature for the reactor heating (a)
and cooling cycle (b).

Comparing the two graphs in Figure 4.15, the conversion increase with increasing tem-
peratures. When heating the reactor the conversion is low until the temperature exceed
500 ◦ C, resulting in a sharp increase, unlike the cooling sequence when the conversion
decrease evenly with decreasing temperature. The first experiment for Bed 1 and 2 are
similar (red squares and black circles), except for some irregularities, both experiments
end with a maximum CO conversion of 68.6 and 69.3 %. For Bed 2, first and second ex-
periment (black circles and green triangles), the maximum conversion decrees from 69.6
% to 46.3 %. The third experiment performed with Bed 2 (yellow stars) is the cooling
cycle following the three day experiment, and the maximum CO conversion has decreased
further. Table 4.4 show the maximum conversion for all the experiments at the set point
temperature of 700 ◦ C.

In Figure 4.16 the CO conversion is plotted against time on stream for the three day ex-
periment at a stable temperature set point of 700 ◦ C. Except for an area with scattered
data points, the graph follow a decreasing trend, starting with a conversion of 42.1 % and
ending at 9.6 %.

The temperature profile was measured at every 100 ◦ C during all the experiments and an
example is shown in Figure 4.17, for Catalyst bed 1, Experiment 1. The maximum temper-
ature measured over the catalyst bed is 635 ◦ C, with the set point temperature of 700 ◦ C.

40
 4.2 Catalyst Activity

(a) (b)

Figure 4.15: Graphs showing the CO conversion plotted against catalyst bed temperature in the
carbon monoxide oxidation experiments (CO/O2 :1/3). (a): Reactor heating, (b): Reactor cooling.

Figure 4.16: Graph showing the CO conversion plotted against time on stream for the carbon
monoxide oxidation experiment (CO/O2 :1/3) which lasted for three days with the oven tempera-
ture kept stable at 700 ◦ C.

Table 4.4: Maximum CO conversion for the carbon monoxide oxidation experiments at the set point
temperature of 700 ◦ C.

CO conversion
Experiment
at 700 ◦ C [%]
Catalyst bed 1.1 68.6
Catalyst bed 2.1 69.3
Catalyst bed 2.2 41.3
Catalyst bed 2.3 9.1

41
Chapter 4. Results

(a) (b)

Figure 4.17: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in a carbon monoxide oxidation experiment Catalyst bed 1, Experiment
1, for reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

The heating cycle is shown in (a) and reactor cooling cycle in (b), and at all temperature
set points above 200 ◦ C, the warmest spot is located between 19 - 21 cm, where the cata-
lyst bed is located. The remaining temperature profiles are shown in Appendix D, Figures
D.9 and D.10 and Table D.2.
Mass balance calculations for carbon were performed for all experiments to evaluate the
accuracy of the measurements made by the gas chromatograph. An example is shown in
Figure 4.18 for the experiment performed using Catalyst bed 1, where the error is plotted
against the conversion of carbon monoxide. The mass balance error increase with increas-
ing conversing of carbon. The remaining mass balances are shown in Appendix D, Figures
D.11 to D.13.

4.2.3 Methanol oxidation

A methanol oxidation experiment was performed in the MTF set-up to investigate the
catalytic activity of polycrystalline silver catalyst. The reactor was set up as shown in
Section 3.3, Figure 3.3. Temperature profiles over the catalyst bed were measured for
the experiment to observe the temperature gradient and to control where the exothermic
reactions are taking place.
The experiment was performed with excess methanol (8.0 mol%) and oxygen (3.5 mol%)
in nitrogen and a temperature program as shown in Section 3.3.3, Figure 3.4. The tem-
perature profiles were measured at every 100 ◦ C. The feed analysis used in conversion
calculations is shown in Appendix B, Table B.3. Due to experimental equipment malfunc-
tion, only one experiment was performed with a reactor heating cycle at temperatures from
200 - 600 ◦ C
Figure 4.19 show the reactant conversions and carbon selectivity plotted against the cata-

42
 4.2 Catalyst Activity

Figure 4.18: Error in carbon mass balance plotted against CO conversion in the carbon monoxide
oxidation experiment using catalyst bed 1.

lyst bed temperature. Both reactant conversions, methanol (blue circles) and oxygen (red
triangles) increased rapidly at the temperature set point of 300 ◦ C, before stabilising at
their respective conversions of about 65 % and 95 %. The selectivities of the carbon prod-
ucts, formaldehyde (yellow stars), carbon monoxide (green diamonds) and carbon dioxide
(grey squares) are also shown. Formaldehyde and carbon dioxide selectivity increase with
the same rate as the reactants conversion, but formaldehyde continues to increase, where
carbon dioxide decrease. According to GC analysis, formic acid was also a product of
the reaction, but is not shown here, as the GC was not recently calibrated for formic acid
making it difficult to give exact amounts of this product. The trend show an initially high
production of the acid, but decreasing with increasing temperature. At the reactor set point
temperature of 500 ◦ C, formic acid was no longer detected.
The temperature profile was measured at every 100 ◦ C during the experiment and is shown
in Figure 4.20. The maximum temperature measured over the catalyst bed is 560 ◦ C, with
the set point temperature of 600 ◦ C. The warmest spot is at 18.5 - 20 cm, where the catalyst
bed is located. At the oven set points of 300 ◦ C and 400 ◦ C, the temperature over the
catalyst bed exceeded the oven temperature by 70 ◦ C and 25 ◦ C, respectively, indicating
highly exothermic reactions taking place.
Table 4.5 show the average carbon mass balance, methanol conversion and catalyst bed
temperatures for the experiment. The balance is negative, which indicate more carbon
leaving the system is detected then carbon in the feed. The error initially increase with
increasing temperature and conversion levels, before it decrease at higher temperatures.
Values for formic acid are not included, and may account for the trend in the balance.

43
Chapter 4. Results

Figure 4.19: Graph showing the conversion of methanol and oxygen, and selectivity towards the
carbon products; formaldehyde, carbon monoxide and carbon dioxide in the methanol oxidation
experiment (CH3 OH/O2 :2.314 and heating set points 200 - 600 ◦ C).

Figure 4.20: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in the methanol oxidation experiment, for furnace set points 200 - 600
◦
C. The catalyst bed is located at 18.5 - 20 cm, marked by vertical dotted lines.

Table 4.5: Average carbon mass balance, methanol conversion and catalyst bed temperature for the
methanol oxidation experiment.

Catalyst Bed Temperature CH3OH Conversion Carbon Balance Error

[◦ C] [%] [%]
191 1.2 -0.3
369 63.0 -9.9
425 64.0 -10.5
490 64.6 -7.1
560 68.6 -6.3

44
Chapter 5
Discussion

5.1 Catalyst Characterisation

The main purpose of using X-ray diffraction and scanning electron microscopy character-
isation was first, to create a base for comparison and learn more about the structure of the
silver catalyst. After performing experiments, the techniques were used to provide infor-
mation of the effect said experiments had on the structure. This information will also be
discussed in supplement to the results in its entirety.

5.1.1 X-Ray Diffraction

X-ray diffraction analysis was used on a selection of the silver catalyst samples to investi-
gate the crystalline phases in the silver. The silver was identified as having a face-centred
cubic structure by the Diffrac.Eva software database and the narrow, sharp peaks, in Fig-
ure 4.1, indicate high crystalline samples [18]. Comparing the samples, some shift in peak
intensity is visible. This indicate a phase restructuring in the bulk of the silver. However,
for verification, the experiment and analysis should be repeated.
The crystal size was calculated using Diffrac.Eva software and the Scherrer formula, in
Equation 2.11. The Scherrer constant, K, is a factor of particle shape. This constant is
usually 0.89, but can vary from 0.89 to 1.0, influencing the crystal size. No great change
in crystal size was detected, except for a slight decrease in the silver used in oxidation
experiments, as presented in Table 4.1. Normally, a catalyst exposed to high temperatures
over a longer period of time experience particle growth in the form of sintering. According
to Nagy et al., the opposite can occur when silver is heating in the presence of oxygen.
Diffused γ-oxygen can substitute for silver lattice atoms and will eventually cause failure
in the structure which result in crystal size reduction [29]. Still, an important factor is that

45
Chapter 5. Discussion

transport and diffusion mechanism in the catalyst is not known and the results can not be
verified without further testing.
An important factor in this chase, the XRD method is mainly for gaining information of
bulk and not surface structure, indicating these results are best used as indications.

5.1.2 Scanning Electron Microscopy

Micrographs of all silver catalyst samples were analysed to investigate morphological

changes in fresh and exposed catalyst.
When using SEM, only a random selection of scan sites are selected, scanned one at a
time. Therefor, if the sites reflect the entire sample is uncertain. The technique is also
limited as one can only scan the surface, not different angles or inside the particles and
the resolution is only valid to a certain magnification. To verify a representative surface,
selected samples were analysed on several occasions.
The silver particles show a great variety in the visible surface changes by SEM, as a re-
sult of the various atmospheres and reaction conditions compared to the fresh silver, as
presented in Section 4.1.2. Comparing the fresh silver, in Figure 4.2, to nitrogen and
hydrogen calcinated silver, Figures 4.3 and 4.4, show the greatest similarities. Some topo-
graphic changes are evident in the nitrogen and hydrogen treated samples, such as minor
hole formation. Nitrogen is present in the system as an inert, not adsorbing or reacting
on the surface, making it likely that the changes are a result of thermal energy and min-
imisation of surface energy. Hydrogen is reported to diffuse into the silver bulk, but when
comparing the silver structures, neither time or components present in the gas feed seems
to have an effect on the silver morphology in SEM scale, suggesting hydrogen alone also
acts as an inert.
Waterhouse et al. used Raman spectroscopy and temperature programmed desorption
(TPD) to characterise chemisorbed oxygen species on silver. Results from analysis of the
as-received silver catalyst showed presence of subsurface oxygen, identified as β-oxygen
[13]. High temperature and trace amounts of subsurface oxygen, or feed gas impurities,
seems likely to be the cause of the surface changes in the silver treated with nitrogen and
hydrogen.
The silver particles exhibiting a pronounced restructuring of the surface on the mesoscopic
scale, are the silver treated with oxygen, both in air and as catalyst for the oxidation ex-
periments. Exposure to synthetic air resulted in massive changes in morphology for all
samples, shown in Figure 4.5, though more extensive with time. The micrographs are con-
sistent with results obtained by Waterhouse et al. where polycrystalline silver catalyst was
exposed to oxygen for 5 hours at 923 K [13].
Facet formation shown in the SEM images, are in accordance with the increase in inter
and surface diffusion occurring in the presence of oxygen and temperatures above the
Tammann temperature (345 ◦ C for silver) [13]. Present oxygen is believed to chemisorbe
onto the silver surface, lowering the surface free energy, promoting densely packed, (111)

46
 5.1 Catalyst Characterisation

facet termination, and simultaneously incorporate into the silver bulk causing the silver
lattice to expand [13].

Silver used as a catalyst for the hydrogen and carbon monoxide oxidation exhibit similar
restructuring, independent of the bed has been used for one or multiple experiments, as
show in Figures 4.7 and 4.8. A smooth surface with irregular facetted areas and holes are
visible. As described by Millar et al. the formation of pinholes are based in the hypothesis
that the holes are created by recombination of sub-surface hydroxy species in the silver,
especially, close to surface defects [14]. Waterhouse et al. argue the pinholes are formed
by reaction between bulk-dissolved β-oxygen and hydrogen [6]. In situ observations were
performed of polycrystalline silver, by Millar et al., in steam atmosphere and as a catalyst
for a mix of methanol and oxygen. For both reactions, the hole formation rate and size
increased with temperature. After, the silver was used in TPD analysis where hydrogen,
water and carbon dioxide were detected.

This may indicate that the formation of carbon dioxide also can be the source of pinhole
formation, as the SEM micrographs in Figure 4.8 show. The morphology of the holes pro-
duced by hydrogen and carbon monoxide oxidation are similar, but if they are structurally
equal is unknown. During the first experiment using carbon monoxide the lab experienced
a power shortage, how much, and if the catalyst structure was effected is not know, but
when comparing with the second bed, the effect seems minimal. Some of the structural
changes can also be due to the nitrogen atmosphere used in between experiments for the
catalyst bed used for multiple experiments.

The third oxidation experiment, using methanol, showed two different morphological
states during SEM analysis, as presented in Figure 4.9. The random sample from the
catalyst bed bulk shows little signs of use. This is similar to images from fresh catalyst
and the hydrogen calcination, with clear angels and terrace structures. The was unlike the
sample from the top of the bed. Here, hole formation is clearly visible, but still lacking
the smoothing and grain annealing visible from other oxidation experiments. This may
be explained by the short duration of the experiment due to equipment malfunction. The
catalyst was only exposed to temperatures close to 600 ◦ C for a short period of time,
as the complete experiment lasted for 5 hours. During the experiment the temperatures
rose quickly, as shown in the temperature profile in Figure 4.20, promoting exothermic
reactions. This may have lead to local hot-spot, normally evened out by the steam in the
reactant feed [7]. The reactions on these sites could have contributed to the variations
in morphology, producing the holes in the silver surface. As observed by Millar et al.,
the pinhole production first occurred at temperatures lower then for the steam atmosphere
(360 and 440 ◦ C, respectively). The pinholes where also of a greater concentration with
methanol present [14].

Last, the experiment performed with a mix of methanol and steam produced no great struc-
ture change present in the catalyst (Figure 4.6). This is in contradictory to the, previously
mentioned, experiments performed by Millar et al. [14] and coincides with the theory of
Waterhouse et al., where the Oβ , is key to pinhole formation [6]. More experiments should
be performed and repeated SEM analysis should be performed before a conclusion can be
drawn.

47
Chapter 5. Discussion

During the methanol/steam calcination experiment, some activity indicating methanol de-
hydrogenation was observed by GC analysis (Equation 2.1). It is unknown if the reaction
occurred as gas phase or catalytic reaction, but is another indication that atomic oxygen
must be present to activate the silver surface for both structural changes and reactions,
such as methanol dehydrogenation reaction, where oxygen is not directly involved in the
reaction [9]. A general comparison of the micrographs representing silver exposed to at-
mospheres with and without oxygen in the feed, indicate that oxygen seems to be the main
cause of morphological structure changes for experiments performed in this study.
The silver catalyst experience great structural change during the start up of formaldehyde
synthesis and influence both product formation and formaldehyde selectivity. Understand-
ing what effects the structure and activity can be important in the work of understanding
and improve the silver catalyst for formaldehyde production [3, 6].

5.2 Catalyst Activity

Side reactions, gas phase reactions, changes in catalyst morphology and activity are among
some of the challenges in the formaldehyde process. Catalytic reactions were performed
to investigate reaction pathways and catalyst activity. Further, the results from oxidation
experiments will be discussed.
The experiments were performed in the MTF set-up were temperature and gas composi-
tion were the variable conditions. To have better control of the reaction temperatures, the
silver catalyst were set up with a quartz sinter and wool as support. For all the experiments
performed, the reactive system has not been at steady state, as the experiments being per-
formed are meant to represent the start up of a process using fresh catalyst. As the catalyst
restructure, the process will not reach steady state during the time scale of these investi-
gations. Temperature profiles were measured for every experiment. This was performed
manually, and will represent an uncertainty. Both placement of the thermocouple and the
time of the measurement may have influenced the values.

5.2.1 Hydrogen Oxidation

Hydrogen oxidation using hydrogen (2 mol%) and excess oxygen (6 mol%), was per-
formed using two different catalyst beds. Initially a blank experiment was performed to
investigate gas phase reactions. Results from the gas chromatograph showed detection of
H2 O, indicating that at a set point of 700 ◦ C and a temperature of 613 ◦ C measured in
the reactor, were sufficient to oxidise hydrogen to some extent (Appendix D, Figure D.1).
Temperature profile was measured at every 100 ◦ C (Appendix D, Figure ??) and show the
warmest location along the reactor at 20 cm, which was the location of the quartz sinter.
The first experiment, using catalyst bed 1, was performed under sightly different condi-
tions then the remaining experiments. A lower amount of hydrogen (1 mol%) was used
with the same amount of silver catalyst, in addition to the temperature program variations.

48
 5.2 Catalyst Activity

The difference in the ratio between reactant and catalyst might be the cause of the higher
conversion levels at intermediate temperatures, shown in Figure 4.11, both in the heating
and cooling cycle.
Comparing the first experiments for catalyst bed 1 and 2, a flat segment between 350 ◦ C
and 450 ◦ C is noticeable for both experiments. This coincide with the Tammann tem-
perature of silver and is according to Waterhouse et al. [13] where silver restructuring
is activated by increased bulk mobility for the silver atoms. The change in silver mor-
phology, is a result of inter and surface diffusion, and enable incorporation of oxygen and
hydrogen into the bulk. As the oven temperature continues to increase, the promotion
of active oxygen species occurs and hydrogen conversion rise. Seen in connection with
the SEM analysis, massive morphological changes are visible, such as pinholes and facet
termination.
The catalyst bed, used in several experiments, shows increased conversion with time, as
shown in Table 4.3. This may indicate a silver catalyst restructuring which promotes
hydrogen oxidation. The conversion was stable over a longer period of time in the third
experiment (3 days), indicating the initial restructuring is somewhat stabilised, to the point
there it no longer effect the reactivity of hydrogen (Figure 4.12). The silver samples used
for one experiment exhibit equal topology as silver used in several experiments, on the
scale provided by SEM analysis. This suggest the restructuring happens rather quickly, in
the time scale of the formaldehyde synthesis start up, which is said to last for periods up to
a week until steady-state is reached. The structure observed in this work is similar to that
shown in similar research [14, 9].
The mass balance error is larger then what is within reason for sections of the experiments,
decreasing the reliability of the results (Figure 4.14). From past experience, the GC does
not perform satisfactory water analysis. During the period of the experiments, measures
were taken to improve the analysis and the mass balance error, resulting in slightly im-
proved results. Conversion is therefor calculated based on reacted hydrogen, not produced
water.
The experiment using catalyst bed 2, second experiment, the temporary stay in temperature
was shorter then for the other experiments. This variation may have been made visible by
the increased conversion at the end of the cooling cycle, suggesting the system did not have
sufficient time to adapt to the change in temperature. Temperature profiles are similar for
all experiments, where the warmest location is measured in the area of the catalyst bed and
support the oxidation occurring as catalytic reaction and not as gas phase reaction.
The results clearly shows hydrogen oxidation to be an active reaction on the silver cata-
lyst, in addition to be a source of catalyst restructuring. Also, the reaction is capable of
competing for the oxygen in the formaldehyde synthesis.

5.2.2 Carbon Monoxide Oxidation

Oxidation of carbon monoxide (2 mol%) and excess oxygen (6 mol%), was performed
using two different catalyst beds. Initially a blank experiment was performed to investigate

49
Chapter 5. Discussion

gas phase reactions. Temperature profile at every 100 ◦ C were measured and are shown
in Appendix D, Figure D.8. The graph show the warmest spot along the reactor at 20 cm,
which was the location of the quartz sinter.
Results from the gas chromatograph showed no detection of CO2 , indicating that at a set
point of 700 ◦ C and a temperature of 615 ◦ C measured in the reactor, was not sufficient
to oxidise carbon monoxide. This is consistent with the research done by Ekholdt, where
COMSOL simulations of gas phase carbon monoxide oxidation were performed at process
conditions relevant to formaldehyde synthesis [30].
Further, experiments using silver catalyst were performed. All experiments were executed
equally, but the experiment using catalyst bed 1 experienced a power shortage half way
in the experiment. Comparing the conversion data with the other relevant experiment, it
does not seem to have much impact on the results. Experiment 1 for catalyst bed 1 and
2, show great likeness and reproducibility, ending with similar conversion at maximum
temperatures, as listed in Table 4.4.
Figure 4.15 show the conversion of carbon monoxide plotted against the catalyst bed tem-
perature, during reactor heating and cooling cycle. All experiments show the same be-
haviour during the heating cycle, with low conversion of carbon monoxide until sufficient
temperature is reached, and the conversion increase rapidly. This is evident also for the
catalyst bed used for multiple experiments. The conversion results in the cooling cycle
follow a trend, but not equal to the heating cycle. The maximum conversion of carbon
monoxide decrease as the catalyst bed is submitted to continued process conditions. This
is clearly shown in Figure 4.16, where the conversion is reduced from 41 % to 9 % over the
course of three days at approximately 630 ◦ C over the catalyst bed. The scattered points
on this graph is most likely caused by pressure variations in the GC analysis, resulting in
different amounts of gas in the injection loop. The pressure change was caused by pre-
cipitation in the water-lock located on a parallel pipeline to the GC. The precipitation was
removed at the time where the data points continue to follow the decreasing trend.
Carbon mass balance graphs show the same trend, where the error increase with increasing
carbon monoxide conversion, which indicate a uncertainty in the GC analysis for low
amounts of reactant.
The temperature profiles are similar for all experiments showing the warmest location the
catalyst bed at the position of the catalyst bed, indicating the exothermic reactions occur
on the bed and not as gas phase reactions.
These results correlate with the observations made by Qu et al., and silver is an active
catalyst for carbon monoxide oxidation [15]. A hypothesis to explain the trends shown
between heating and cooling cycles may indicate that carbon monoxide have similar influ-
ence on silver as on platinum crystals, described by Van Santen [16]. The carbon monoxide
activates the platinum surface, restructuring the phase form a stable, dense, (111) structure
to an open, more active (100) structure, which promote carbon dioxide formation.
The results also indicate that the activity is not maintained over time. This may be caused
by silver restructuring, permanently shifting towards the (111) structure, being the thermo-
dynamically favourable structure for silver. Comparing with the XRD results, an increased

50
 5.2 Catalyst Activity

intensity for the (111) structure is observed in the diffractrogram for the silver used in car-
bon monoxide oxidation, in Figure 4.1, compared to the fresh silver.
Carbon loss in the formaldehyde synthesis is mostly caused by selectivity towards the main
byproduct, carbon dioxide, and understanding how to avoid this could be of great use to
increase formaldehyde production. Results from this work indicate that after sufficient
silver catalyst restructuring, the carbon dioxide source does not seem to generate from the
reaction between carbon monoxide and oxygen. Also, the oxidation reaction would not be
a competing reaction for oxygen consumption.

5.2.3 Methanol Oxidation

Only one methanol oxidation experiment was performed due to equipment failure. The ex-
periment was performed by feeding methanol (8 mol%) in excess and oxygen (3.5 mol%).
No water was added to the reactant feed, which is common in formaldehyde synthesis.
The methanol/ oxygen ratio was between the needed values for partial and total oxidation
of methanol, Equations 2.2 and 2.3.
Figure 4.19 show the conversion of methanol and oxygen, in addition to the carbon selec-
tivity’s. Both conversion of methanol and oxygen increase rapidly at temperatures reach-
ing 370 ◦ C, and can be experienced as the ”light-off”, where the silver catalyst begins an
initialisation phase, as described by Nagy [12]. Bulk-diffusion by β-oxygen was observed
at temperatures above 300 ◦ C. The oxygen diffuse along grain bounties and surface defects
in the silver catalyst, activating restructuring.
As shown in the temperature profile in Figure 4.20, the catalyst bed temperature increase
above the oven set point in correlation with the light-off temperature. This indicate highly
exothermic reactions are taking place, such as partial and total oxidation of methanol and
formaldehyde, show in Section 2.1. This elevated temperatures may also have lead to
carbon deposition on the catalyst bed, explaining the black spots on the catalyst, as show
in Figure 4.10. The EDX analysis confirmed the presence of carbon and oxygen on the
catalyst, but can not confirm if it was a result of coke or a carbon species introduced to the
silver surface after the experiment was carried out (Table 4.2). SEM analysis show local
morphological changes on the catalyst similar to results shown for other oxidation exper-
iments, being exposed to high temperatures for a longer period of time. The comparison
may indicate active catalyst sites and local hot spots on the catalyst.
Initially, formic acid was detected by the GC analysis. The amount decreased as the tem-
perature rose, and was no longer present at 500 ◦ C. Formic acid is not shown as a part of
the results, because the GC method was not calibrated for the compound, making quantita-
tive analysis challenging. This provides an additional error in the mass balance for carbon,
hydrogen and oxygen, resulting in only the carbon balance being presented (Table 4.5).
The balance error is higher then the allowed 5 %, which is expected, but decrease with
temperature, when formic acid levels fall. Conversion and carbon selectivity is therefor
calculated on the basis of methanol reacted, and not product levels.
Research on comparing formaldehyde synthesis with and without water vapour in the re-

51
Chapter 5. Discussion

action gas was performed by Qian et al. [7], Results showed an increased conversion of
methanol and formaldehyde selectivity with the water ballast method. Light off at 300
◦
C and increased production of formic acid at low temperatures when water was excluded
from the reaction gas, correlating with the present results. Also, carbon dioxide selectivity
was higher at lower temperatures and the absence of steam.
The initial carbon dioxide production decrease with temperature. As shown in previous
results, the carbon dioxide is not likely to be a product from carbon monoxide oxidation,
as the reaction does not seem to be favoured at temperatures bellow 500 ◦ C for near fresh
silver catalyst.
The reaction did not achieve full oxygen conversion (max. 97.8 %). This is surprising,
as the oxygen was not feed in excess and participates in several reactions, including hy-
drogen oxidation presented in the current work. Bao et al. state, water and hydroxy must
be present to promote surface restructuring of the silver and formation of γ-oxygen, rel-
evant for methanol oxidation and formaldehyde selectivity at elevated temperatures [8].
Strongly bound γ-oxygen is associated with exclusively partial oxidation of methanol to
formaldehyde. The work by Bao et al., indicate that producing bound γ-oxygen from hy-
droxy reaction is more favourable then by electron donation from the silver metal. When
water is removed this will increase the barrier for γ-oxygen to be formed, also increasing
the chance of oxygen desorbing from the silver surface forming oxygen molecules. The
experiments should be completed and reproduced before conclusions can be drawn, but
the theory gives rise to the idea of repeating hydrogen and carbon monoxide oxidation
experiments with steam to investigate the effect.
Formaldehyde could react with oxygen, downstream of the reactor, but with elevated gas
velocity and sufficient cooling after the reactor, oxygen detection in the product gas might
be explained.

52
Chapter 6
Conclusion

In this master thesis, polycrystalline silver catalyst have been used to investigate catalytic
activity and the effected of different atmospheres on the morphological structure of the
catalyst relevant to the formaldehyde synthesis. The motivation for the project was to con-
tribute to the continued development and understanding of the silver catalyst technology
used in formaldehyde production.
The silver catalyst was exposed to various conditions to investigate the dependency be-
tween the conditions and catalyst structure, analysed by SEM and XRD. Conditions, such
as time dependency, temperature and atmosphere compositions were used to gather infor-
mation about the interaction between silver and various components present in the exper-
iments. Further, the effect of catalytic oxidation reactions were tested in the connection
with activity data.
The XRD diffraction confirmed the presence of highly crystalline fcc silver and a shift in
the crystalline base plane structures after use. SEM investigations indicated that the silver
particles were effected differently based on the process conditions. At 650 ◦ , thermal
energy was the main driving force causing signs of restructuring in the silver exposed
to nitrogen and hydrogen atmospheres, independent of the exposure time. This was also
the case with exposure to methanol and steam in nitrogen, where in addition, only trace
amounts of produced formaldehyde and hydrogen was detected.
Morphological changes in the silver was observed when oxygen was present. Silver cal-
cined in air showed increasing restructuring with time, eliminating angular grain structure,
production of pinholes and promoting irregular and smooth facet formation. Using the
silver as a catalyst for hydrogen, carbon monoxide and methanol oxidation resulted in the
same observations.
It was found that the silver catalyst display high catalytic activity for hydrogen oxidation,
facilitated by catalytic restructuring towards increased conversion of hydrogen. The cat-
alyst displayed an initial catalytic activity for carbon monoxide oxidation which declined

53
Chapter 6. Conclusion

over time. Equal morphological changes were observed on the mesoscale, suggesting the
restructured silver promotes hydrogen oxidation, as opposed to carbon monoxide oxida-
tion. This indicate the oxidation of hydrogen to be a competing reaction for oxygen in the
formaldehyde synthesis. Performing oxidation of methanol without water, promoted the
essential role of water in the reaction, in connection with temperature control and product
selectivity.
In general, oxygen displays an important role promoting and activating reactions and influ-
ence the silver catalyst morphology. This is especially critical for silver catalyst activation
and performance during process start-up in the formaldehyde production.

54
Chapter 7
Further work

In this work, only a few methods for silver characterisation was used. The project work
could have benefited from using more surface sensitive techniques to gain a better under-
standing of the silver surface structure and properties. Advanced characterisation methods,
such as STM, TEM, Raman, AFM and Krypton BET could be beneficial.
Operating the MTF set-up, some challenges arose. The experimental result could benefit
from changing flow and liquid controllers more suited to the feed interval, to increase feed
precision. Also, a calibration of formic acid is required. To improve the product stream
analysis, using GC combined with MS or switch to IR detection could be promising.
Further investigating the reaction pathways and silver restructuring should be investigated
to gain a better understanding of the catalytic system. Experiments, such as calcination of
silver in water, and CO, continuing methanol oxidation experiments with and without wa-
ter, and using equal conditions as for hydrogen and CO oxidation performed in the work,
would be interesting. To get a better understanding of how water effect the formaldehyde
process, the current experimental work could be repeated with added steam.
As previously mentioned, the catalyst structure varies from start-up and at steady-state
conditions. By performing experiment with pre-calcinated silver it could provide informa-
tion about how the start-up phase of the experiments is effected.

55
Chapter 7. Further work

56
Bibliography

[1] I. Business Wire. (2011, May) Annual consumption growth of 5 % expected

on formaldehyde market according to merchant research & consulting, ltd.
[Online]. Available: https://www.businesswire.com/news/home/20110526005760/
en/Annual-Consumption-Growth-5-Expected-Formaldehyde-Market
[2] A. W. Franz, H. Kronemayer, D. Pfeiffer, R. D. Pilz, G. Reuss, W. Disteldorf,
A. O. Gamer, and A. Hilt. (2016, November) Ullmann’s encyclopedia of industrial
chemistry, formaldehyde. [Online]. Available: http://onlinelibrary.wiley.com/doi/10.
1002/14356007.a11 619.pub2/abstract
[3] G. Millar and M. Collins, “Industrial production of formaldehyde using polycrys-
talline silver catalyst,” Industrial and Engineering Chemistry Research, vol. 56,
no. 33, pp. 9247–9265, August 2017.
[4] C. D. NTNU. (2017) icsi - industrial catalysis science and innovation. [Online].
Available: https://www.ntnu.edu/icsi
[5] J. A. Moulijn, Chemical process technology, 2nd ed. Chichester: Wiley, 2013.
[6] G. I. Waterhouse, G. A. Bowmaker, and J. B. Metson, “Influence of catalyst mor-
phology on the performance of electrolytic silver catalysts for the partial oxidation
of methanol to formaldehyde,” Applied Catalysis A, General, vol. 266, no. 2, pp.
257–273, 2004.
[7] M. Qian, M. Liauw, and G. Emig, “Formaldehyde synthesis from methanol over
silver catalysts,” Applied Catalysis A, General, vol. 238, no. 2, pp. 211–222, 2003.
[8] X. Bao, M. Muhler, B. Pettinger, Y. Uchida, G. Lehmpfuhl, R. Schlögl, and G. Ertl,
“The effect of water on the formation of strongly bound oxygen on silver surfaces,”
Catalysis Letters, vol. 32, no. 1, pp. 171–183, March 1995.
[9] G. I. Waterhouse, G. A. Bowmaker, and J. B. Metson, “Mechanism and active sites
for the partial oxidation of methanol to formaldehyde over an electrolytic silver cat-
alyst,” Applied Catalysis A, General, vol. 265, no. 1, pp. 85–101, 2004.

I
[10] I. E. Wachs and R. J. Madix, “The oxidation of methanol on a silver (110) catalyst,”
Surface Science, vol. 76, no. 2, pp. 531–558, 1978.
[11] A. Andreasen, H. Lynggaard, C. Stegelmann, and P. Stoltze, “Simplified kinetic mod-
els of methanol oxidation on silver,” Applied Catalysis A, General, vol. 289, no. 2,
pp. 267–273, 2005.
[12] A. Nagy, G. Mestl, T. Rühle, G. Weinberg, and R. Schlögl, “The dynamic restruc-
turing of electrolytic silver during the formaldehyde synthesis reaction,” Journal of
Catalysis, vol. 179, no. 2, pp. 548–559, October 1998.
[13] G. I. Waterhouse, G. A. Bowmaker, and J. B. Metson, “Oxygen chemisorption on an
electrolytic silver catalyst: a combined tpd and raman spectroscopic study,” Applied
Surface Science, vol. 214, no. 1, pp. 36–51, 2003.
[14] G. J. Millar, M. L. Nelson, and P. J. R. Uwins, “In situ observation of structural
changes in polycrystalline silver catalysts by environmental scanning electron mi-
croscopy,” Journal of the Chemical Society, Faraday Transactions, vol. 94, no. 14,
pp. 2015–2023, 1998.
[15] Z. Qu, M. Cheng, W. Huang, and X. Bao, “Formation of subsurface oxygen species
and its high activity toward co oxidation over silver catalysts,” Journal of Catalysis,
vol. 229, no. 2, pp. 446–458, 2005.
[16] R. A. Sheldon and R. A. v. Santen, Catalytic oxidation: principles and applications
- a course of the netherlands institute for catalysis research (niok). World Scientific
Pub Co Pte, October 1995.
[17] B. Viswanathan, Catalysts and surfaces : characterization techniques. Oxford:
Alpha Science International, 2010.
[18] I. Chorkendorff, Concepts of modern catalysis and kinetics, 2nd ed. Weinheim:
Wiley-VCH, 2007.
[19] K. W. Kolasinski, Surface science : foundations of catalysis and nanoscience.
Chichester: Wiley, 2002.
[20] W. D. Callister, Materials science and engineering, 8th ed. Hoboken, N.J: Wiley,
2011.
[21] J. Niemantsverdriet, Spectroscopy in catalysis : an introduction, 3rd ed. Weinheim:
Wiley-VCH, 2007.
[22] J. M. Thomas, Handbook for Heterogeneous Catalysis. 2., completely revised and en-
larged Edition. Vol. 1–8. Edited by G. Ertl, H. Knözinger, F. Schüth, and J. Weitkamp.
Wiley-VCH, Weinheim, 2008.
[23] T. Greibrokk, Kromatografi : separasjon og deteksjon, 3rd ed. Oslo: Universitets-
forlaget, 1994.
[24] C. F. Poole, Gas chromatography, 1st ed. Amsterdam: Elsevier, 2012.

II
[25] A. Technologies. (2018) Cp-sil 5 cb capillary column. [Online].
Available: https://www.agilent.com/en/products/gas-chromatography/gc-columns/
capillary/cp-sil-5-cb
[26] J. T. Richardson, Principles of catalyst development, ser. Fundamental and applied
catalysis. New York: Plenum Press, 1989.

[27] T. E. C. Industry. (2007) ”catalyst science and technology,

http://nptel.ac.in/courses/”. [Online]. Available: http://nptel.ac.in/courses/
103103026/7
[28] H. S. Fogler, Elements of chemical reaction engineering, 5th ed., ser. Prentice-Hall
international series in the physical and chemical engineering sciences. Boston:
Prentice Hall, 2016.
[29] A. Nagy and G. Mestl, “High temperature partial oxidation reactions over silver cat-
alysts,” Applied Catalysis A, General, vol. 188, no. 1, pp. 337–353, 1999.
[30] R. J. Ekholdt, “Oxidation of methanol to formaldehyde over ag,” Master’s thesis,
Norwegian University of Science and Technology, 2017.

III
IV
Appendix

V
VI
Appendix A
Methanol to formaldehyde test
set-up

This Appendix contain details about selected equipment in the MTF set-up. Details re-
garding the reactor design is presented in Figure A.1 and the GC automatic valve injection
system is shown in Figure A.2.
All the flow controllers were calibrated either during the specialization project (TKP 4580)
or the period of the master thesis. A description of the calibration method is presented in
Chapter 3.3.1 and the resulting graphs are presented in this Appendix.

VII
 Figure A.1: Valve system controlling GC injections.

Figure A.2: Reactor details for experiments in the MTF test set-up with particular silver.

VIII
Calibration curves
Mass and liquid flow controllers were calibrated using a bobble column and gravimetric
method, respectably. Figures A.3 to A.10 show the calibration curves for both nitrogen
supply lines (low and high area), air, CO/H2 and liquid controller feeding pure methanol
and a methanol/water mix.

Figure A.3: Calibration curve for nitrogen MFC, line 1, low area.

Figure A.4: Calibration curve for nitrogen MFC, line 1, high area.

IX
 Figure A.5: Calibration curve for nitrogen MFC, line 2

Figure A.6: Calibration curve for air MFC.

Figure A.7: Calibration curve for carbon monoxide MFC.

X
 Figure A.8: Calibration curve for hydrogen MFC.

Figure A.9: Calibration curve for LFC feeding methanol.

Figure A.10: Calibration curve for LFC feeding a mix of methanol and water.

XI
XII
Appendix B
Feed Analysis

Feed analysis were performed for all activity experiments. Examples are shown in Tables
B.1 - B.3. The average values are used further in conversion, selectivity and mass balance
calculations.
Table B.1: Feed analysis for a hydrogen oxidation experiment with 2 mol% H2 and excess oxygen
(6 mol%).

Experiment Hydrogen Oxygen Nitrogen

H2 oxidation,
2.5906 6.0270 93.2632
with catalyst
2.5838 6.0408 93.2354
2.5795 6.0468 93.2700
2.5689 6.0421 93.3594
2.5633 6.0284 93.2318
2.5646 6.0487 93.3647
2.5542 6.0507 93.3455
2.5587 6.0346 93.3239
Average 2.5705 6.0399 93.2992

XIII
Table B.2: Feed analysis for a carbon monoxide oxidation experiment with 2 mol% CO and excess
oxygen (6 mol%).

Experiment Carbon Monoxide Oxygen Nitrogen

CO oxidation,
2.7275 5.9657 93.0092
with catalyst
2.7377 5.9592 92.9090
2.7560 5.9863 93.0544
2.7218 6.0124 93.0604
2.7512 6.0344 93.0518
2.7550 6.0261 93.0380
Average 2.7415 5.9973 93.0205

Table B.3: Feed analysis for methanol oxidation experiment with MeOH/O2 :2.314 ratio.

Experiment Methanol Oxygen Nitrogen

MeOH oxidation,
6.6160 3.3713 88.5235
with catalyst
6.7892 3.4003 89.4285
6.5792 3.3847 88.8120
6.3204 3.3969 89.1759
6.6511 3.4541 89.2488
6.6926 3.5819 88.8551
6.9780 3.6133 87.9802
Average 6.6609 3.4575 88.8606

XIV
Appendix C
Calculations

For all activity experiments the data was further processed using Microsoft Excel. Exam-
ples showing the calculations are shown in this Appendix. The calculations are based on
the equations shown in Chapter 3.3.4.
Calculating the conversion of, for example hydrogen, XH2 , is done using Equation C.1,

yH2 ,out yN2 ,in

XH2 = 1 − · 100% (C.1)
yH2 ,in yN2 ,out

where yxin are the element values based on the feed analysis and yxout are the element
values based on the GC analysis.

0.61 93.3
XH2 = 1 − · 100% = 76.7% (C.2)
2.57 94.58
The product selectivity is calculated based on the amount of reacted methanol, using Equa-
tion C.3 and formaldehyde as an example, in Equation C.4.

FTot,out yCH2 O,out

SCH2 O = (C.3)
FTot,in yCH3 OH,in − FTot,out yCH3 OH,out

520.7 · 2.28
SCH2 O = · 100% = 51.2% (C.4)
500.0 · 6.50 − 520.7 · 1.83
The mass balance is used to control the reliability of the GC values and calculations. The
mass balance for hydrogen, EH , is done using Equation C.5,
P P
FTot,out n · yH,out − FTot,in n · yH,in
EH = P · 100% (C.5)
FTot,in n · yH,in

XV
where FT ot,out is the total flow calculated based on the inert nitrogen entering and leaving
the system, FT ot,in is the total inlet flow, yH,out is the fraction of the elements containing
hydrogen leaving the reactor, yH,in is the inlet fraction of elements containing hydrogen
based on the feed analysis and n is the stoichiometric factor.

246.61(2.01 + 0.61) − 250.00 · 2.56

EH = · 100% = 0.37% (C.6)
250.00 · 2.56

XVI
Appendix D
Oxidation Experiments

D.1 Hydrogen Oxidation

Additional information and results from hydrogen oxidation experiments are shown in this
section.

The graph showing conversion data for the blank experiment is shown in Figure D.1.

Temperature profiles over the catalyst bed were measured for each experiment to measure
the temperature gradient and to control where the exothermic reactions are taking place.
The profiles are shown in Figures D.3 and D.4 and Table D.1.

The mass balance error for hydrogen was calculated for each hydrogen oxidation exper-
iment to assess the accuracy of the GC analysis and is shown in Figures D.5 to D.7 for
experiments using catalyst bed 1 and bed 2, experiments 2 and 3.

Table D.1: Temperatures trough the reactor at 20.5 cm axial position compared to the temperature
set point for the carbon monoxide oxidation experiment catalyst bed 2, experiment 3, cooling cycle
following the three day experiment. The catalyst bed is located at 19 - 21 cm.

Set point temperature Temperature at

[◦ C] 20.5 cm axial position
200 185
300 277
400 374
500 472
600 569
700 668

XVII
Figure D.1: Graph showing the conversion plotted against set point temperature in the blank hydro-
gen oxidation experiment.

Figure D.2: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in the blank hydrogen oxidation experiment, for reactor heating. The
catalyst bed would be located at 19 - 21 cm.

XVIII
 (a) (b)

Figure D.3: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in the hydrogen oxidation experiment catalyst bed 2, experiment 1, for
reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

Figure D.4: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in the hydrogen oxidation experiment catalyst bed 2, experiment 2, for
reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

XIX
Figure D.5: Error in hydrogen mass balance plotted against hydrogen conversion in the hydrogen
oxidation experiment using catalyst bed 1, experiment 1.

Figure D.6: Error in hydrogen mass balance plotted against hydrogen conversion in the hydrogen
oxidation experiment using catalyst bed 2, experiment 2.

XX
Figure D.7: Error in hydrogen mass balance plotted against hydrogen conversion in the hydrogen
oxidation experiment using catalyst bed 2, experiment 3, duration of 3 days.

Table D.2: Temperatures trough the reactor at 20.5 cm axial position compared to the temperature
set point for the carbon monoxide oxidation experiment catalyst bed 2, experiment 3, cooling cycle
following the three day experiment. The catalyst bed is located at 19 - 21 cm.

Set point temperature Temperature at

[◦ C] 20.5 cm axial position
200 178
300 255
400 344
500 430
600 534
700 631

D.2 Carbon Monoxide Oxidation

Additional information and results from carbon monoxide oxidation experiments are shown
in this section.
Temperature profiles over the catalyst bed were measured for each experiment to measure
the temperature gradient and to control where the exothermic reactions are taking place.
The profiles for the blank experiment and catalyst bed 2, experiments 1 - 3, are shown in
Figures D.8 to D.10 and Table D.2.
The mass balance error for carbon was calculated for each carbon monoxide oxidation
experiment to assess the accuracy of the GC analysis and is shown in Figures D.11 to D.13
for experiments using catalyst bed 2, experiments 1-3.

XXI
Figure D.8: Graph showing the temperature profile trough the reactor plotted against the axial
position of the thermocouple in the blank carbon monoxide oxidation experiment. The catalyst bed
would be located at 19 - 21 cm.

(a) (b)

Figure D.9: Graph showing the temperature profile trough the reactor plotted against the axial po-
sition of the thermocouple in the carbon monoxide oxidation experiment catalyst bed 2, experiment
1, for reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

XXII
 (a) (b)

Figure D.10: Graph showing the temperature profile trough the reactor plotted against the axial po-
sition of the thermocouple in the carbon monoxide oxidation experiment catalyst bed 2, experiment
2, for reactor heating (a) and cooling (b) cycle. The catalyst bed is located at 19 - 21 cm.

Figure D.11: Error in carbon mass balance plotted against CO conversion in the carbon monoxide
oxidation experiment using catalyst bed 2, experiment 1.

XXIII
Figure D.12: Error in carbon mass balance plotted against CO conversion in the carbon monoxide
oxidation experiment using catalyst bed 2, experiment 2.

Figure D.13: Error in carbon mass balance plotted against CO conversion in the carbon monoxide
oxidation experiment using catalyst bed 2, experiment 3, duration of 3 days.

XXIV