"it?
:6"
I "

fr; '

"d“ we I
1"”,
”A”

g.- » ‘rr'f,

.v.s .
A.‘ -\'-

5‘

I!
° )hj'l: I
*4 ‘4 , ‘ ‘
.Y. {I"I(£;"nv ’
71.6 r

' '(.'.“
7:.Ilf

"3 .1-\‘-\
aim"

\ s x
‘3- “‘23.; :éu'v} :\~\~r&. '9“
‘4‘

‘c 5’: v.

{mi .
‘I

w ' ."r.
' 3404. ;. ‘ _ - , .
' ‘KLA l ['1'2".

' _ I”"'r1"":" 7i);
. I: 1 .
f» .

'.'.‘1'/‘ ‘

Q .‘

'1‘} 2' AKA“.
. a
”'15:;
.1 :I

:1-,» IJ,’ .
l,’ /

"All '

" ‘ l
,1‘ 11,1:
$5.14. 33.3: .7
"1: ring? '
"0.!

. I“
Iv“); .v, f?
ff} ' f

vii-My, ”w- 1-2.; .
1,11: ,._

 

('11!qu (\{NR‘
1.1 23-“ _ i. ‘24; x ‘ . ‘ «+3 3-2.1-
C“ 1.“, in. :$' AFEN‘Xg-g
ty.‘ ‘5 “ w; $93“ tsjiﬁﬁ
v. “aw ..

WV“?

12% .
‘ v '1 ‘ -t\'- ‘
“'2: . .éféw mﬁa}. .
\1K: "'3'“. . a r
A, l'.“'\ ‘ 1 ‘TS"

15/111 ..,
.Y'v'f—u'

Yuk];

“t
I
"1 111/”: 121"” ,1’1
.' a ( n I]: 'r".,( ”“2; ‘.
1w, 1 1‘ .
”6"?»'u)’}';r:/'?71I"Jlf/- 1;,"

1-191

I.” n

. -'I.1‘:{/{z/:/"/r€4”
1., gym-J - h ./ _
“6"“: 1; ~.
7".”5-2'21-1:
,2 131:.- .__ ‘1",{1‘5'ﬁ’ﬁé'
1”! - 11-5";ng
'3’ a)? Mt: u _
J " . n I

”:39: "‘0'ij ~ " r $1"
1.2;. ﬁr: : . 1’sz £1.
”liq/:1 ){{r 13).?!

I72" ‘I

'1 3:19]
’3? 21.731,
f'f’ v.2 .
.131?!) 11,1? .. ..'
’7? ‘2}: 47 ’13:, 7 2:. ,1. I in _
~11: m . 1
-r W vv '- "HI,” I
f ’ r
(7.7 r

ﬁ’Y } ' '
‘ ‘ ' '0 v
1"» ”31’304 I
M: 1,.)
‘1 r0 .
. h,
;"WH' ; l
in” ﬁr
:15:
r I Ff.’ .
. 11%;;

. .9391

:1‘4'; 554' :3;
”_{(.:;”p:" ‘v- "’
. I
' ”Cy-914‘?»-
-,. 7.:

I' ' "' ’gtx’t...) .
9‘3 ”11‘1"" 11,151,11113. ‘
Io 4.1131994“ ”.éJ-I, I?“
- {'Aé'jzf'fg'E-‘J { :1 .;. ”J

I:

4

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

A Fundamental Study of K C0
Catalysis of Hydrogasification of Carbon
and the Role of Oxygen in the Catalysis

presented by

Hossein Zoheidi

has been accepted towards fulﬁllment
of the requirements for

PhoDo degree in CHE

 

 

‘ L
Major ofessor

Date Q/17/g7

MS U is an Aﬂ'irmative Action/Equal Opportunity Institution 0- 12771

 

 

 

* RETURNING MATERIALS:
IVIESI_] Place in book drop to
remove this checkout from
w your record. FINES will

 

' be charged if book is
returned after the date
stamped below.

 

 

 

A FUNDAMENTAL STUDY OF K2003
CATALYSIS OF HYDROGASIFICATION OF CARBON
AND THE ROLE OF OXYGEN IN THE CATALYSIS

By

Hossein Zoheidi

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering

1987

ABSTRACT

A FUNDAMENTAL STUDY OF K2CO3-CATALYSIS
OF HYDROGASIFICATION OF CARBON AND
THE ROLE OF OXYGEN IN THE CATALYSIS

By

Hossein Zoheidi

The existing and projected shortages of natural gas in the United States have
stimulated extensive research on production of synthetic natural gas (SNG) from coal,
the largest fossil fuel resource in this country. Potassium carbonate has been identiﬁed
as the most economically attractive catalyst for gasilication. In this process, coal is
reacted with steam to produce a mixture of CO and H2. The product hydrogen can
further react with either CO or coal to produce methane. However, understanding the
mechanism is complicated by the reported strong retardation effect of product hydrogen
on the kinetics of both uncatalyzed and the catalyzed steam gasiﬁcation. This
retardation effect and the catalyzed reaction of product hydrogen with carbon are

relatively unexplored and are not fully understood.

Recently, studies have suggested that the catalyzed steam gasiﬁcation rates are
directly proportional to coverage of carbon surface by oxygen complexes, and that
reaction of carbon with hydrogen is enhanced by the presence of oxygen, implying that
interaction of K2003 and hydrogen with these surface complexes may be the key to

understanding the mechanism of the retardation effect and hydrogasiﬁcation.

In this work, therefore, fundamental experimental kinetic studies of K2003-
catalyzed and non-catalyzed hydrogasiﬁcation of carbon blacks at 700-865’ C and 500
psi hydrogen pressure have been conducted to gain more insight into the role of K2003
and oxygen complexes. In these studies, the effects of different catalyst loadings on

hydrogasiﬁcation rate, potassium loss at high temperature, the nature of oxygen

complexes in carbon black, and effects of oxygen chemisorption at 400° and 800°C
and /or degassing at 1000° C on non-catalyzed and the catalyzed hydrogasiﬂcation have

been investigated.

From the effect of various oxygen surface groups on rate of hydrogasiﬁcation, it is
postulated that oxygen groups thermally desorb from the carbon surface and create
‘nascent’ active sites which act as active sites for hydrogasiﬁcation, hence enhancing the
reactivity of the carbon. K2003 catalytic activity and potassium losses are inﬂuenced
by the nature of oxygen surface groups. The formation of stable and active
potassium-containing species via interaction of K2003 and oxygen complexes leads to
reduced potassium losses and enhanced hydrogasiﬁcation rates. The interaction of
K2003 with carbonyl groups forms both stable and active species. Interaction with
basic oxygen complexes, which are relatively inactive in uncatalyzed hydrogasiﬁcation,
form active but unstable intermediate species. Interaction with acidic complexes is not
well characterized from the experiments. Degassing removes non-basic groups and
thermally anneals the ‘nascent’ active sites, resulting in reduced gasiﬁcation rate and

increased potassium loss.

Dedication

I would like to dedicate this work to my parents, Hassan and Forough
Zoheidi, who provided me with love and support and gave me all I have, my
brothers Mohammad Ali and Hammeed and my sisters, Afsauneh, Elauheh and
Parvauneh. They made this work possible with their continued support and

encouragement.

ii

Acknowledgments

I would like to take this opportunity to express my gratitude to Dr.
Dennis J. Miller, the chairman of my doctoral committee, for his guidance and
support throughout this work, and Prof. Charles A. Petty for invaluable and
constructive comments and suggestions during the preparation of this
document. My thanks to Prof. Martin C. Hawley and Prof. Thomas J.
Pinnavaia for kindly accepting the unpleasant duty of serving on my doctoral

committee.

I wish to thank all of my good and valuable friends, in particular, Dr.
Alireza Abbasabadi, Dr. Nikrouz Faroughi, Dr. Hamid Ghods, Dr. Touran
Cheraghi, Mehrnam Sharif, Partash Moallemian, Sun (Natasha) Yong Hong,
Javad Kalantar, and Farzin Khorasanizadeh, who bestowed me with nothing but
their encouragement, understanding, patience and forgiveness. Thank you all

for tolerating me during the hard period of consummating this work.

I would also like to thank the Department of Chemical Engineering at
Michigan State University for supporting me ﬁnancially through most of this

work.

Finally, I would like to Thank Dr. Klaus Otto at Ford Motor Company
for carrying out the XPS analysis and Mr. Xiao—Wen Qian at the Department of
Physics at Michigan State University for providing the XRD analysis. Their

help is greatly appreciated.

iii

Table of Contents

List of Tables ...................................................................................... ix
List of Tables ...................................................................................... x
Nomenclature ..................................................................................... xii
Chapter 1: Introduction 69’ Background ........................................... 1
Chapter 2: Experimental Apparatus Design ................................... 10
2.1 Design Objectives ............................................................................. 10

2.2 Reactor System Design .................................................................... 11
2.2.1 Reactor Design ....................................................................... 13

2.2.1.1 Heating System ........................................................... 13

2.2.1.2 Material of Construction ............................................. 14

2.2.1.3 Reactor Wall Thickness .............................................. 15

2.2.1.4 Flange Design .............................................................. 17

2.2.1.5 Cooling System & Thermal Stresses ............................ 18

2.2.1.5.1 Thermal Stresses ............................................... 19

2.2.1.5.2 Cooling Coils .................................................... 26

2.2.1.6 Reactor Quartz Lining ................................................. 27

2.3 v-Ray Detection System .............................................. i ..................... 28
2.3.1 Theoretical Derivations .......................................................... 29

2.3.1.1 Mass-Radiation Relation ............................................. 29

2.3.1.2 Radiation Measurement .............................................. 31

2.3.1.2.] Problems ........................................................... 31

2.3.1.2.2 ﬂ-Rays .............................................................. 32

2.3.1.2.3 'y-Rays .............................................................. 32

2.3.1.2.4 Interaction and Absorption of ’7—Rays .............. 33

iv

2.3.1.2.5 Attenuation Coefﬁcients .................................... 34

2.3.1.2.5.l Photoelectric Effect ................................. 34

2.3.1.252 Compton Effect ...................................... 35

2.3.1.2.5.3 Pair Production ...................................... 35

2.3.1.2.6 _Absorption ........................................................ 37

2.3.1.2.7 Geometric Factor .............................................. 39

2.3.1.2.8 Sample Activation ............................................ 39

2.3.2 Potassium Content Measurements ......................................... 41
2.3.3 Biological Effects of 'y-Rays ................................................... 41

2.4 Gas Collection System ..................................................................... 42
2.4.1 Solenoid Valves & Electronics ................................................ 42
2.4.2 Cold Traps ............................................................................ 43
2.4.3 Silica Gels .............................................................................. 43

2.5 Gas Analyzing & Calibration System .............................................. 44
2.5.1 Gas Chromatograph & Recorder ........................................... 44
2.5.2 Calibration Setup .................................................................. 45

2.6 Gas Purifying System ...................................................................... 45
2.7 Evacuation System .......................................................................... 46
2.8 Flow & Pressure Control ................................................................. 46
2.8.1 Flowmeters .................................................................. _ .......... 4 6
2.8.2 Pressure Regulators ............................................................... 47

2.9 Rupture Disc .................................................................................... 47
2.10 Sample Holders .............................................................................. 48
2.11 Temperature Probes ....................................................................... 50
Chapter 3: Characterization of Carbon ........................................... 51
3.1 Coconut Charcoal . ........................................................................... 52
3.1.1 BET Surface Area .................................................................. 52
3.1.2 Ash Content .......................................................................... 53

3.1.3 Impregnation with K2003 ...................................................... 53

3.1.4 Reactivity .............................................................................. 54

3.2 Graphite ........................................................................................... 54
3.2.1 BET Surface Area .................................................................. 54
3.2.2 Impregnation with K2003 ...................................................... 55
3.2.3 Reactivity .............................................................................. 55

3.3 Carbon Black ................................................................................... 56
3.3.1 Ulimate Analysis & Ash Content .......................................... 56
3.3.2 Impregnation with K2003 ...................................................... 57
3.3.3 BET Surface Area .................................................................. 58
3.3.4 Reactivity .............................................................................. 61
3.3.5 XRD Analysis ........................................................................ 62
3.3.6 Microstructure of Carbon Black ............................................ 62
3.3.7 XPS Analysis ......................................................................... 65
3.3.8 Oxygen Surface Complexes .................................................... 65
3.3.8.1 Procedure .................................................................... 67

3.3.8.1.1 Combustion ...................................................... 67

3.3.8.1.2 pH Measurement ............................................... 68

3.3.8.2 Results ........................................................................ 68

3.3.8.3 Discussion .......................................................... , .......... 7 2
Chapter 4: Uncatalyzed Hydrogaszﬁcation ...................................... 76
4.1 Background ...................................................................................... 78
4.1.1 Hydrogasiﬁcation Products .................................................... 78
4.1.2 Thermodynamic Equilibrium ................................................. 79
4.1.3 Kinetics .................................................................................. 83
4.1.4 Mechanism ............................................................................. 85

4.2 Intrinsic Nature of Reaction Rates ................................................... 87
4.2.1 Mass Transfer Resistances ..................................................... 87

vi

4.2.2 Temperature Variation of Carbon Bed .................................. 96

4.3 Hydrogasiﬁcation of Carbon Black .................................................. 97
4.3.1 Procedure ............................................................................... 97
4.3.2 Results .............................................. 99
4.3.3 Discussion .............................................................................. 105

Chapter 5: Catalyzed Hydrogasiﬁcation .......................................... 109

5.1 Background ....................................................................................... 110
5.1.1 Thermodynamic Equilibrium ................................................. 110
5.1.2 Kinetics & Mechanism ........................................................... 111
5.1.3 Carbon-K2003 Interaction ..................................................... 112
5.1.4 Effect of Carbon Impurities on Reactivity ............................. 117

5.2 Intrinsical Nature of Reaction Rates ................................................ 119
5.2.1 Mass Transfer Resistances ...................................................... 119

5.3 Hydrogasiﬁcation of Coconut Charcoal ............................................ 126

5.4 Hydrogasiﬁcation of Graphite .......................................................... 128

5.5 Hydrogasiﬁcation of Carbon Black .. ................................................ 129
5.5.1 Procedure ............................................................................... 129
5.5.2 Results ................................................................................... 130

5.5.2.1 Catalyzed Hydrogasiﬁcation ........................................ 130
5.5.2.1.1 Fresh Carbon Black .......................................... 130

5.5.2.1.2 Treated Carbon Black ...................................... 133

5.5.2.2 Catalyst Loss .............................................................. 138
5.5.2.2.1 Fresh Carbon .................................................... 138

5.5.2.2.2 Treated Carbon Black ...................................... 141

5.5.3 Discussion .............................................................................. 144
5.5.3.1 Hydrogasiﬁcation of Fresh Carbon Black .................... 144
5.5.3.2 Potassium Catalyst Loss ............................................. 151
5.5.3.3 Effect of Treatments on Catalysis ............................... 152

vii

Chapter 6: Conclusions 89’ Recommendation ................................. 160

6.1 Conclusions ....................................................................................... 161
6.1.1 Uncatalyzed Hydrogasiﬁcation ............................................... 161

6.1.2 Alkali-Catalyzed Hydrogasiﬁcation ........................................ 162

6.2 Recommendation ............................................................................... 163
Appendix A: Reactor Design Calculations ..................................... 165
Flange Design ...................................................................................... 165
Thermal Stresses .................................................................................. 169
Temperature Proﬁle .................................................................. 169

Shear Strain .............................................................................. 171

Cooling Coils ............................................................................ 172

’y-Ray Detection System ...................................................................... 177

'ikRay Absorption ..................................................................... 177

Sample Activity ........................................................................ 179

Sample Activation .................................................................... 180

Rupture Disc Sizing ................................................ 182

List of References ............................................................................. 186

viii

2.1
2.2
2.3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8

List of Tables

Maximum overall normal stresses in the reactor wall. ..................................... 24
Attenuation coefficient of Lead and Haynes-25 components. ........................... 36
Attenuation coeﬂicient A for various shielding media. ...................................... 37
Elemental analysis of ash in Raven 16. ........................................................... 57
Ultimate analysis of Raven l6. ....................................................................... 57
Surface concentration (atomic %) of non-metalic elements in Raven 16. ......... 67
ApH of 0.1 M KC1 solution for fresh, degassed and gasified carbons. ........... 69
ApH of 0.1 M KC] solution for carbons oxidized at 400 ° C. ............................ 69
ApH of 0.1 M KC1 solution for carbons oxidized at 800 ° C. ............................ 69
ApH of KC] solution after combustion at 400 ' C and hydrogasiﬁcation. ........ 72
ApH of KC] solution after combustion at 800 ° C and hydrogasification. ........ 72

ix

2.1
2.2
2.3
2.4
3.1
3.2
3.3
4.1

4.2
4.3

4.4
4.5

4.6

4.7

4.8

5.1

5.2

5.3

5.4

5.5

5.6

5.7

5.8

5.9

List of Figures

Schematic of the complete differential micro-reactor system. .................................
Temperature zones of the differential micro-reactor. ..............................................
'1-Ray detection conﬁguration ................................................................................
Sample holders used for hydrogasiﬁcation studies.

Ng-BET surface area of Raven 16 at various conversions. .....................................
ESCA scan of fresh and heat treated Raven 16. ....................................................
Nature of oxygen complexes in bulk Raven 16. .....................................................

Mass transfer resistances during hydrogasiﬁcation of Raven 16 at 865 ' C and
500 psig. ...............................................................................................................

Arrhenius plot of hydrogasiﬁcation of Raven 16 at 500 psig H2. .........................

Effect of Raven 16 bed conﬁguration on hydrogasiﬁcation rate at 865 ° C and
500 psig H2. ...........................................................................................................

Raw hydrogasiﬁcation rate data of Raven 16 at 865 ° C and 500 psig H2.

Eﬂect of heatrtreatment on hydrogasification rate of Raven 16 at 865 ' C and
500 psig H2. ..........................................................................................................

Effect of combustion at 400 ° C on hydrogasiﬁcation rate of Raven 16 at 865 ° C
and 500 psig H2. ...................................................................................................

Effect of combustion at 800 ° C on hydrogasiﬁcation rate of Raven 16 at 865 ° C
and 500 psig H2. ...................................................................................................

Effect of combustion on hydrogasiﬁcation rate of degassed Raven 16 (1000 ' C)
at 865 ° C and 500 psig H2. ...................................................................................

Absolute methane evolution rate from Raven 16 at 865 ° C and 500 psig
H2. .......................................................................................................................

Mass transfer resistances during hydrogasiﬁcation of catalyzed Raven 16 bed at
865 ' C and 500 psig H2. ........................................................................................

Reproducibility of hydrogasiﬁcation rates of 10 wt%—K2CO3 samples at 865 ° C
and 500 psig H2. ....................................................................................................

Arrhenius plot for hydrogasiﬁcation of Raven 16 at 22% conversion under
500 psig H2. ..........................................................................................................

Absolute methane formation rate from coconut charcoal at 865 ' C and 500
psig H2. ................................................................................................................

Absolute methane evolution rate from Raven 16 at 865 ° C and 500 psig
H2. .......................................................................................................................

Catalyzed hydrogasiﬁcation rate at 865 ' C and 500 psig H2 versus
conversion. ............................................................................................................

Effect of oxidation at 400 ° C and catalyst addition on hydrogasiﬁcation rate
at 865 ' C and 500 psig H2 of degassed carbon. .....................................................

Effect of oxidation at 400 ' C and catalyst addition on gasiﬁcation rate at
865 ° C in H2 of 47% gasiﬁed carbon. ....................................................................

20
38

101

102

103

123

124

125

134

136

5.10

5.11

5.12

5.13

5.14

5.15

5.16

5.17

5.18

Effect of oxidation at 400 ° C and catalyst addition on gasiﬁcation rate of 65%
gasiﬁed carbon at 865 ° C in H2. ............................................................................

Potassium loss from reaction zone during hydrogasification at 865 ' C and 500
psig H2 measured by q-ray detector. .....................................................................

Potassium loss from KgCoa-loaded Raven 16 during hydrogasiﬁcation at
865 ' C and 500 psig H2 .........................................................................................

Effect of high temperature treatment on potassium content of carbon samples
at various levels of conversion (865 ° C). ...............................................................

Effect of oxidation at 400 ’ C on potassium content of carbon samples at
various levels of conversion (865 ° C). ....................................................................

Speciﬁc methane evolution rate of Raven 16 at 865 ° C and 500 psig versus
carbon conversion. ................................................................................................

Hydrogasiﬁcation rates relative to uncatalyzed rates versus carbon conversion
at 865 ° C and 500 psig H2. ...................................................................................

Effect of heat treatment on reactivity of impregnated Raven 16 at 865 ° C and
500 psig H2. ..........................................................................................................

Relative gasiﬁcation rates at 865 ’ C of heat-treated K2C03-loaded
carbons. ................................................................................................................

mi

137

139

140

143

150

3':

a

bppbb

Q

‘

gyms

[‘13
.r

"' 32' R R? 9b.?“

3"
:E
.5

k4.

Nomenclature

total heat transfer surface area in cooling coils, ft2
ﬂange ring design parameter, see Eq. A-l

ﬂange ring design parameter, see Eq. A-l

internal cross sectional area of copper cooling coils, in2
cross sectional area of reactor wall, in2

cross sectional area of rupture disc, in2

atomic weight

constant

constant

flow constant of gases through rupture disc

molar concentration of reacting component A, mole/cm3
speciﬁc heat capacity of cooling water, Btu /]b,,,- ° F
speciﬁc heat capacity of the i th component of a reaction mixture, ca] / g mole- ° K
gaseous binary diffusivity of Hg—CH4 system, cm2/ sec
Knudsen diffusion coefficient, cmz/sec

effective diffusivity, cm2/sec

inside diameter of reactor vessel, in

diameter of 7-ray detector crystal. in

outside diameter of reactor vessel, in

inside diameter of cooling coils, in

Young modulus of elasticity, psi

intrinsic activation energy, kcal/g-mole

apparent activation energy, kcal/g-mole

7-ray energy, MeV

fraction of potassium lost from a sample

fraction of surface covered by detector crystal

fraction of '1-ray radiation passing through shielding layers
shearing modulus of elasticity, psi

heat transfer coefficient, Btu / ftz-h- ' F

constant, set to 0.62 by ASME Code when rupture disc is the sole relieving device
equilibrium constant

ratio of r, to r0

ratio of speciﬁc heats

rate constant of uncatalyzed hydrogasiﬁcation, sec’1

rate constant of a ﬁrst order reaction, sec—l

rate constant of a. second order reaction, cm3/gmole-sec
constant

constant

constant

constant

distance from cooling coils to furnace inlet, in .
X-ray width of microcrystallites of carbon black, A
X-ray height of microcrystallites of carbon black, A
length of cooling coils, ft

distance between sample and '7-ray detector crystal, in
total mass of potassium at any time, g

total initial mass of potassium, g

mass ﬂow rate of cooling water, lbm/ h

mass ﬂux of cooling water inside cooling coils, lbm/inQ-h

A

assess

Q
a
A

b-

3
a

:3: '* tempo innate

msga

an.“ O
ssssca fﬁﬁyﬁs

a5=
ab

Avogadro’s number, 6.023)<1023 atoms/g-mole
number of atoms of “K at time t

number of atoms of 42K at time t

initial number of atoms of “K

initial number of atoms of 42K

number of major metal components in Haynes-25
effective load on ﬂange, psi

reactor vessel total internal pressure, psi

exhaust pressure of rupture disc, psi

hydrogen partial pressure, psi

methane partial pressure, psi

total axial heat ﬂow in reactor wall, Btu / h

volumetric ﬂow rate of gases through rupture disc, SCFM

axial heat ﬂux in reactor wall, Btu / (inz-h)

universal gas constant, ems-atm/mole- ’ K

rate of nuclear disintegration of 42K nuclei in a sample, d/s

measured rate of nuclear disintegration of 42K nuclei by '7-ray detector, d/s
rate of evolution of methane, gmole CH4/cm3-sec

Reynolds number, dimensionless

hydraulic radius, cm

rate of reaction, gmole/cms-sec

radial distance from the central axis of reactor, in
mean radius of pores in a carbon black particle, cm
inside radius, in

outside radius, in

inside radius of ﬂage ring, in

radius of bolt circle, in

total surface area of carbon black particles, cm2/ g
inverse of Poisson’s ratio p

temperature of reactor wall, 'F

temperature of furnace, ° F

room temperature, ' F

inlet temperature of cooling water, ° F

outlet temperature of cooling water, ° F

reaction temperature, °K

time elapsed after the end of neutron bombardment, h
period of neutron bombardment of potassium atoms, h
ﬂange ring design parameter which is 1 at weakest spot of ﬂang ring
volume fraction of ith elemental component of Haynes-25
volume of a control element, in3

molecular weight, g/ mole

hydrogen molecular weight, 2.016 g/ mole

methane molecular weight, 16.043 g/ mole

potassium molecular weight, 39.102 g/ mole

fraction of carbon gasiﬁed

thickness of '7-ray shielding material, cm

fraction of hydrogen converted to methane at equilibrium
ﬂange ring deﬂection under stress, in

methane mole fraction

atomic number
distance along the central axis of reactor, in

xiii

Greek

0 = thermal expansion coeﬂicient, in/ (in - ° F)
I‘ = Gamma constant of '7-ray radiation
A0? = 214C“,
i
AH = heat of reaction, kcal/ g mole

A1125 0: heat of reaction at 25°C, kcal/g mole

ApH = change in pH of a solution induced by carbon black oxygen surface groups
AT = heat transfer driving force in cooling coils, F

ATc = temperature rise of cooling water in cooling coils, ' F

AT,“ = log-mean temperature difference, 'F

Ar -——-— radial thickness of a control shell, in

A2 = axial length of a control shell, in

e = In 2/ty5 proportionality constant of radiation and mass, h'1
(C = porosity of a carbon black particle

(p = porosity of carbon bed

g = thickness of the reactor wall, in

g, = thickness of ﬂange ring, in

ft = thickness of closed end of the reactor vessel, in

n = cross section of “K atom for neutron absorption, 1.5X10"24 cm2
IC = thermal conductivity of reactor wall, Btu-in/ftz-h- ' F

Kc = thermal conducitvity of cooling water, Btu/ft—h- ° F

A = attenuation coefficient of q-ray shielding material, cm-l

). = attenuation coeﬂicient of pure elements, cm"l

p = Poisson’s ratio

p, = viscosity, lbm/in—h

u = absorption coefficient of photoelectric effect, cm-l

u, = stoichiometric coefficient of the i th component of a reaction mixture
E = compton scattering absorption coefficient, cm'1

p = density, g/m]

am“ = maximum overall normal stress, psi

0, = radial stress, psi

0”,," = tensile rupture stress, psi

0, == longitudinal stress, psi

0, = circumferential or tangential stress, psi

1', = tortuosity factor of carbon bed

r, = circumferential or tangential shear stress, psi

11 = longitudinal shear stress, psi

1,, = radial shear stress induced from z-direction, psi

<l> = neutron ﬂux, neutrons/cm2-s

d = Thiele modulus

x = mass fraction of total potassium as “K

xiv

Chapter 1

Introduction 8’ Background

In the United States, it has been realized for the last two decades that the
natural gas resources (6. g. methane) are being depleted at a rate faster than it is
possible to replace by nature, and the estimated reserves of some 2500 trillion
cubic feet (in 1974) could easily be depleted by the turn of the century. These
existing and projected shortages have stimulated extensive research to ﬁnd viable
alternatives and to develop the technology for commercial production of
synthetic high BTU gases. Imported liquid natural gas can ease shortages, but it

is much more expensive than natural gas (5-10 times) [1].

In this general effort, one of the most logical alternatives is to convert
coal, the largest fossil fuel resource in this country, to gas and oil. Gasiﬁcation

of coal with steam is such a process, and today the technology for commercial

production of synthetic natural gas (SNG) exists, albeit possibly not in the most
economical or elegant way. In this process, coal is reacted with steam to produce

a mixture of CO and H2 (synthesis gas)

0 + H20 -+ H2 + 00 (1-1)

which can subsequently undergo nickel-catalyzed methanation reaction

co + 3H2 -> CH4 + H20 (1-2)

or, alternatively, the product hydrogen gas can directly react with carbon to

produce methane

c + 2H2 —. CH4 (1-3)

The products of these reactions can further react with steam, coal or other

reaction products in secondary reactions:

co + H20 -» H2 + (:02 (1-4)

002 "l" 4H2 —* CH4 + 2H20 (1'5)

200 —> C + C02 (1-6)

Reactions 1-2, 1-3 and 1-5 produce the highly desirable methane and all are

exothermic [1,2].

Therefore, the gasiﬁer effluent gas stream is a mixture of H2, H20, CO,
C02 and CH4, and its composition depends on the pressure and temperature
inside the gasiﬁer and on the catalyst used. The optimum reaction conditions
are thus the conditions under which methane formation is maximized; i. e., lower
temperatures are favorable because of the exothermicity of Reactions 1—2, 1-3
and 5. In fact, it was the search for gasiﬁcation conditions at lower
temperatures (to favor methane formation) that led Exxon researchers and other
investigators to explore the alkali-metal catalysis of steam gasiﬁcation of coal,

ﬁrst realized more than a century ago [3].

As far as the catalytic activity is concerned, most transition metals of
groups IIIB-VIIIB, such as nickel, are superior catalysts compared to alkali
carbonates [4-7]. However, these metal catalysts are expensive and are easily
poisoned by sulfur which is invariably present in coal. In contrast, alkali
carbonates are not as effective, but are not poisoned severely by sulfur and are
relatively cheaper and thus more attractive [8-10]. Among carbonates, catalytic
activity increases with alkalinity; but, again the determining factor for
commercial feasibility is cost effectiveness, or abundance in nature. In terms of
abundance of the elements in earth’s crust, sodium ranks sixth, and potassium
at seventh is 90 times as plentiful as carbon [8]. Yet, potassium has more than
enough activity relative to sodium to justify its somewhat higher cost. On the
other hand, rubidium and cesium, though more active than potassium, are far
too costly to be used commercially. Hence, potassium carbonate seems to be the

most promising catalyst for any steam gasiﬁcation of coal on a commercial scale.

To maximize gasiﬁer throughput and to use potassium carbonate
effectively it is essential to understand kinetics and mechanisms of various
reactions taking place inside the gasiﬁer and the inﬂuence of the catalyst on
these reactions. A knowledge of mechanism and kinetics allows selection of
optimum gasiﬁcation conditions such as temperature, pressure, catalyst loadings,
etc., so that more economical processes and gasiﬁers can be designed. For
instance, if the mechanism of catalyzed interaction of hydrogen with carbon is
known, perhaps the strong retardation effect of hydrogen on reaction of steam
with carbon [10-14] can be diminished and, at the same time, the rate of
Reaction 1-3 can be enhanced signiﬁcantly. This way, SNG (methane) could be
produced at much higher rates and the process would be more economically

attractive.

Catalysis of steam and carbon dioxide gasiﬁcation by alkali carbonates, in
particular KQCO3, has been studied extensively during the last few decades and
several different mechanisms have been proposed. McKee et al. [15-18] have
proposed a redox cycle which involves the decomposition and reformation of the
carbonate on the surface of the carbon with the alkali oxide and hydroxide as
intermediates in CO2 and steam gasiﬁcation, respectively. Veraa & Bell [19] and
Huttinger & Minges [20] have supported this mechanism for steam gasiﬁcation.
Mims et al. [21,22,13] have believed that the formation of C-O—K complexes on
the surface of the carbon is an essential step in the catalysis and thus determines
the activity of the catalyst. Yuh & Wolf [23] have proposed a reduction-
oxidation cycle involving these groups. Wood et al. [24,25] have postulated the
formation of a non-stoichiometric oxide with excess alkali metal which acts as a
center for extraction of oxygen from gaseous reactants and extraction of
electrons from the carbon matrix, thus facilitating the reaction of adsorbed

oxygen atoms [26]. Huttinger & Minges [14] have used this mechanism to

explain their steam gasiﬁcation results, but have gone one step further and have
suggested KOH as the active intermediate. Recently, Saber et al. [27] have
concluded that presence of oxygen on the carbon surface is essential for the
catalyst to remain active, implying that oxygen might be involved in the
catalysis process. Wen [28] has outlined a mechanism involving formation of
potassium-intercalation compounds of carbon. \Nigmans et al. [29,30] have
suggested that all three catalysis processes are simultaneously competing and

dominate at diﬁerent stages of the gasiﬁcation.

Applications of FTIR [31], in situ FTIR [32,33], in situ e.s.r. [34,35] and
EXAFS [36] have provided further support for each of the these proposed
mechanisms. However, in spite of all these studies, and many similar ones
employing a wide range of different techniques, the nature of active sites and the

sequence of events are not yet fully understood.

Catalysis of hydrogasiﬁcation is relatively unexplored compared to steam
gasiﬁcation. Much less attention has been given to this reaction and much less is
known about the catalysis process, in spite of the fact that hydrogen is a
product of steam gasiﬁcation (Reaction 1-1) and is involved in an important step
(Reaction 1-3) in production of SNG from coal. A priori, the above mentioned
redox cycles cannot be accepted as reasonable mechanisms for hydrogasiﬁcation,
because the oxidizing gases (002 or H2O) which account for the oxidation step of
these mechanisms are absent. This is further supported by the observation that
hydrogen gas strongly inhibits the catalysis of steam gasiﬁcation by alkali
carbonates [10,14]. Thus, in the reducing environment of high temperature
hydrogen gas, the driving force for maintaining these cycles operating is

diminished if not completely eliminated.

Wood et al. [37] and Gardner ct al. [38] have already shown that
potassium carbonate can not only catalyze steam gasiﬁcation but also catalyze
reaction of hydrogen with carbon appreciably. The catalytic activity of K9003
in hydrogasiﬁcation thus may arise via another reaction pathway or process.
For uncatalyzed hydrogasiﬁcation, Cao & Back [39] have shown that presence of
oxygen in the gas stream enhances hydrogenation rate of pure carbon
signiﬁcantly and Blackwood [40] has related methane formation rate in pure
hydrogen to oxygen content of a coconut char. However, the mechanism of this
enhancement is not known. In addition, there has been no study on the

inﬂuence of surface oxygen complexes on K2003 catalysis of hydrogasiﬁcation.

The importance of oxygen in catalysis of steam gasiﬁcation has only been
realized very recently [27 ,41,42]. Hoshimoto et al. [41] have indicated that the
rate of steam gasiﬁcation is directly proportional to the amount of oxygen on
carbon surface. In contrast, Sams et al. [42] have demonstrated that gasiﬁcation
rate is a function of potassium to carbon ratio only, and that heat treatment of
carbon is immaterial. Saber et al. [27] have reconciled these ﬁndings by
illustrating that the amount of potassium retained on carbon surface is indeed
proportional to the amount of oxygen on carbon surface. Yet, no study has so
far been done to examine the nature of oxygen groups in conjunction with their

impact on catalysis process in steam gasiﬁcation or hydrogasiﬁcation.

The nature of oxygen containing groups on carbon surfaces has been the
center of intensive studies for over six decades due to their important and vital
impact on rubber reinforcing properties of carbon blacks. Numerous chemical
techniques using different reagents have been devised to study the chemical
behavior and hence the nature of these groups [43-48]. These studies have
shown that a great part of the oxygen surface groups are chemically inert toward

the reagents [46-48]. Aromatic ring-type ether groups have been proposed to

account for these inert groups [45,49]. The reactive part of the surface groups
show acidic, basic or inert characteristics depending on the conditions under
which they are formed [47] (acidic or basic in the sense that they neutralize
dilute hydroxide solutions, such as NaOH or KOH, or dilute acid solutions such
as HC], respectively). For example, if carbon black is freed from all surface
compounds by heating in vacuo followed by cooling in the absence of oxygen, its
surface becomes covered with basic surface oxides upon exposure to oxygen [47].
It has also been found that oxidation at temperatures of 300-500° C establishes
acidic groups on graphite [50], charcoal [51] and carbon blacks [52,53]. The
nature of these acidic groups has been the source of great controversy. Diﬁerent
functional groups such as carboxy] and phenolic hydroxyl groups [47,48], a
rather indeﬁnite type of “CO2-complex” [52], and lactone groups [40,49,54,55]
have been proposed as the acidic surface complexes. These groups all decompose
and give off CO2 upon heating the carbon to higher temperatures. These studies
have also shown that the amount of base neutralized by these groups is directly
proportional to the amount of CO2 evolved upon thermal decomposition. The
optimum condition for the formation of these acidic groups has been suggested

to be 400 ° C [43,51].

The oxygen surface groups formed at temperatures of 750°C or higher
are predominantly basic in nature [56,57]. Pyrone—like structures [48] as well as
benzpyran structures called chromenes [58] have been speculated to be
responsible for the basicity of these surfaces; both thermally decompose to yield
CO. It has been suggested that carbonyl groups also decompose to yield CO,
but they are neutral in character [50]. The chromene structures are known to
oxidize to lactones [58]. The chromene structures are also thought to be
responsible for the low reactivity of residual chars toward hydrogen (to produce

methane) at temperatures as high as 870 ° C [40]. These residual chars had lost

their highly active sites associated with lactone groups and temperatures above

1200 ° C were used to decompose their chromene groups completely.

Oxygen surface groups also play a key role in the wettability of the
carbon surface. A well-cleaned graphitic carbon surface is essentially
hydrophobic; the presence of oxygen complexes on some fraction of the surface
renders that surface hydrophilic [59]. The wetting by water is due to the
formation of hydrogen bonds both between the oxygen complexes on the surface

and between the adsorbed molecules themselves [60].

As evident from the above literature survey (additional and more detailed
and in depth review of the litrature will be given in the appropriate sections
when seems necessary), various studies have concentrated on either just
identifying the nature of oxygen surface complexes under isolated conditions or,
at most, the inﬂuence of mere presence of oxygen on potassium retention. No
study has yet been done to study the inﬂuence of ‘diﬁerent’ oxygen complexes on
potassium carbonate catalysis of either steam gasiﬁcation or hydrogasiﬁcation.
Such a study is of vital importance since a knowledge of nature of the most
active oxygen complex(es) would eventually provide more insight into the steps

involved in the catalysis process and its mechanism.

The work presented here is an attempt to ﬁll this gap. The study of the
inﬂuence of diﬂerent oxygen surface groups on alkali catalysis of
hydrogasiﬁcation (Reaction 1-3) of carbon blacks is the focus of the work
presented here. Within this context, the inﬂuence of different oxygen groups on
hydrogasiﬁcation reaction in the absence of potassium carbonate is also reported
and a mechanism is postulated. These fundamental studies of alkali
carbonate-catalyzed hydrogasiﬁcation of carbons are necessary to enhance our
knowledge of the catalysis process and the nature of the active sites.

Examination of similarities and differences in the behavior of the catalyst in

oxidizing and reducing environments will reveal more information about the
steps involved in the catalysis process in oxidizing gasiﬁcation as well as in

hydrogen.

Chapter 2

Experimental Apparatus Design

2.1 Design Objectives

The design objectives are to design a complete high temperature/high
pressure reactor system suited for gas-solid reaction kinetic studies. The heart of
this system is a differential micro-reactor that can operate at 1000° C and 1000
psi pressure simultaneously and can handle severe oxidizing and reducing
atmospheres of high temperature air and hydrogen gas, respectively. The
operating temperature of 1000° C is chosen because this reactor is primarily to
be used for gasiﬁcation studies of carbons with hydrogen and, as will be
discussed in Chapters 4 and 5, the previous studies have shown that
temperatures above 700° C are needed to obtain reasonable reaction rates even

for reactive chars and an additional temperature range of loo-150°C are

10

ll

necessary to evaluate kinetic parameters such as activation energies. For
graphite samples, which are the least reactive of carbons, even higher
temperatures are necessary. Besides, degassing of carbon samples that is
essential for studying of surface groups on carbons is traditionally done at

temperatures higher than 950 ° C.

Gasiﬁcation rates of carbons with hydrogen are much slower than those of
other gas reactants such as steam or carbon dioxide and, therefore, in addition
to using higher temperatures much higher pressures are also used. Pressures as
high as 1000 psi are reported in the literature, but not simultaneously with
1000°C. Thus a set of design objectives of 1000°C and 1000 psi seems

reasonable and practical.

2.2 Reactor System Design

A preliminary design outlay of this complete reactor system reveals that
this system must consist of a differential micro—reactor, a gas collection system, a
gas-analyzing system, a reactant-gas purifying system and a gas ﬂow control
scheme. A 'y-ray detection unit is also needed to monitor the amount of

potassium in KQCO3-loaded carbon samples during gasiﬁcation experiments.

Such a system was designed, constructed and tested. As more
experiments were conducted some unpredicted problems (such as presence of air
inside the reactor even after prolonged purging of the system and interference of
metallurgical carbon from the reactor wall with hydrogenation rate measurement
of graphite) were faced and a new evacuation system and a quartz lining for the
reactor were added to this system to resolve these problems. At occasions some

minor revisions were also implemented and what is presented in Figure 2.1 and

12

.639? acaoaecéemﬁ €55.8ch 329:8 e5 .0 oSsEeﬁm ﬁn ohm—ME

z. “(a
we 9 5.5.5

               

wmta>n
swap—(U

  

as: OJOU
2332..
ml... #355 2

ﬁlm) mac?

 

was: 0.50

     

mi? 5335’ Oh

 

 

Emhm>m
ZOEuﬁOU
v.10

0 :5 2.3.5:
I...”—
So 553

I I -
I I
wE!(m
ugaogzcmxh “

-
'
J

25¢) ><>>.m $3.5:

 

 

 

 

 

 

 

 

 

 

 

  

 

35:6 / I ’4. l.
26 532: W
.H. " -Vﬁzzﬁﬁﬁ/ﬁ/zﬁxﬁx-
A: 39523.. \\ I . _¢_ _[ 35:2 :5»
a.
5:3 02:08

 

195230: —K M_
.33)

13

is described hereafter is the ﬁnal scheme of the system that was successfully used

for gasiﬁcation experiments.

2.2.1 Reactor Design

 

Due to the high operating pressure of 1000 psi the reactor to be designed
is basically a high pressure cylindrical vessel closed at one end and a ﬂange and a
cover disk bolted together at the other end for enclosing the sample and isolating
the reaction environment from the surroundings. The design and description of

individual components of this reactor unit are discussed below.

2.2.1.1 Heating Swem

 

A heating system is needed to heat a sample under study inside the
reactor to a desired temperature up to 1000° C. This could be achieved by
either internal or external heating: internal heating being inserting some heating
elements inside the reactor and heating the sample directly. Internal heating has
three main advantages:

i ) Much faster heating capability than the external heating can be obtained.
The main mode of heat transfer from the heated surfaces to the sample inside
the reactor is by radiation at the high temperatures of interest, and the sample
can be heated fast if the hot surface of heating elements is close to the sample
and faces the sample directly. In external heating the reactor wall has to be
heated ﬁrst and only then can the sample be heated by radiation from the
reactor wall. This may take considerably longer times specially if the reactor
wall material has low thermal conductivity and is thick.

ii ) The heating elements inside the reactor can be insulated from the reactor
wall and thus the reactor wall can be kept at much lower temperatures than the

sample temperature inside the reactor. This translates into much higher

14

strength of the reactor wall material, and that [in turn results in requiring a
thinner reactor wall and lower construction costs.

iii ) The heat loss from heating elements for internal heating are much less due
to a smaller size and insulation from the reactor wall, so that heating is localized
only around the sample. The smaller size of the heating elements and their

better heating efficiency reduces construction and operating costs.

However, contamination is a big disadvantage to using internal heating.
Both the heating elements and insulating materials (refractory bricks)
contaminate the gaseous reactant stream and introduce complications and
uncertainity into interpretation of the experimental results. This is highly
undesirable, and therefore it was decided to concentrate on designing and
evaluating the external heating option. For this purpose a Lindberg single-zone
tube furnace model 54232 is chosen. This model has a 12"-long heating zone and
is suited for heating tubes with 2" OD. This model is chosen for the reasons
that will be discussed in Reactor Wall Thickness section. This implied that the
reactor wall material has to be an alloy that is suitable for operation at 1000 ° C

and 1000 psi pressure.

2.2.1.2 Material of Construction

 

The problem of selecting a suitable metal alloy for construction of the
reactor is complicated by the so called “Hydrogen Environment Embrittlement”
phenomenon (HEE) which is induced by hydrogen gas. This phenomenon has
been observed over a wide range of hydrogen pressure, 10"8 Torr to 10000 psi.
The exact mechanism of this process is not fully understood, but it has been well
agreed upon that dissociative chemisorption of hydrogen molecule and
absorption are necessary steps. The necessity of diffusion of the hydrogen atom

into the alloy lattice has provoked marked disagreement [61]. Subsequent crack

15

initiation, internally or at the surface, and crack growth degrades one or more

properties such as ductility of the alloy [62].

The high-strength structural alloys such as steels and nickel-base alloys
are particularly susceptible to HEE [61]. Super strength alloys such as stainless
steels 316, 304L, A-286 and some other alloys such as 19-9DL, N-155, IN 800 and
aluminum, magnesium and copper alloys have excellent resistance to HEE but
all lack structural strength that is needed at 1000 ° C [61,63,64]. Generally these
alloys are usable only at temperatures below 800°C. Pure titanium is also a
very good metal but is extremely expensive. The choice is then limited to some
cobalt alloys such as Haynes-25 (L—605) that have acceptable resistance to HEE
and have good structural strength at 1000° C. Haynes-25 alloy is thus selected
as the material of construction for the micro-reactor. This alloy can also handle

the oxidizing environment of air at 1000 ° C.

2.2.1.3 Reactor Wall Thickness

 

The differential micro-reactor to be designed here can be classiﬁed as a
thick-walled cylinder. Traditionally, two widely accepted formulae are used to
calculate the desired thickness of this vessel wall, Lamé’s formula and Barlow’s
formula, which is only applicable to cylinders subject to internal pressures [65].
Barlow’s formula embodies more assumptions on the side of safety than the

Lamé’s formula and thus is used as the design equation. This formula is

 

c = (2-1)

16

where a, is the circumferential or tangential stress, p is the internal pressure of
the reactor, Do is the outside diameter of the reactor and g“ is the thickness of

the reactor wall.

The rupture stress in 10000 hours at 1000° C is 2000 psi and is used as
the value for 0‘, in the above equation [64]. Internal pressure p is set at 1100 psi
to include a safety factor of 10% into the calculation. Equation 2-1 clearly
shows that the wall thickness 3' is directly proportional to the outside diameter
of the reactor D, and therefore smaller diameters are more favorable because the
reactor wall material would cost less. However, the axial hole inside the reactor
has to be big enough to provide sufficient space for convenient insertion and
removal of samples inside the reactor and thus bigger reactor diameters are
favored in this respect. Hence, a compromise has to be made for an optimal
diameter size. It was then decided to try an outside diameter size of 2" as a ﬁrst
guess for the optimal size; a wall thickness of 0.55" was obtained from Equation
2-1. This leaves a cylindrical space with 0.9" diameter for housing the sample
inside the reactor, which seems to be a reasonable and practical space for a

sample. The actual size of the drilled hole is 7/ 8".

The required thickness of the closed end of the reactor vessel is calculated

from an equation for circular ﬂat plates [66]

{h = _' P (M)

 

where g, is the head thickness, D,- is the inside diameter (7 / 8") and am“ is the
maximum, allowable stress which for this case is equal to 2000 psi (rupture stress
of Haynes-25 alloy at 1000° C in 10000 hours [64]). Substitution of these values

into Equation 2-2 yields a thickness of 0.28". The actual thickness of the

l7

fabricated head is 1" and thus the highest stress, which is at the edge is only 160

psi.

2.2.1.4 Flange Design

 

For the purposes of containing the high pressure reaction environment
and convenient insertion and removing of samples from the open end of the
reactor, three options are available; a ﬂange, a screw cap and a clamp-on hinged
cover disk. The last two options are not reliable due to leaks and inconvenience
in handling and proper positioning of samples. Diﬁ'erent types of ﬂanges such as
forged integral, loose or fusion lap-welded hubbed and ring ﬂanges, etc., can
safely be designed and operated [67]. A plain fusion welded ring ﬂange is chosen
because of its simplicity of design and lower costs. This ﬂange is made of
Haynes-25 alloy so that it can be easily welded to the Haynes-25 reactor

cylinder.

The surface contact between the ﬂange ring face and the cover disk is not
gas-tight enough to contain the high pressure reactant gas(es) and some gasket is
thus required for this purpose. The type of gasket needs to be known a priori
so that the required sealing stress can be included in ﬂange stress analysis and
thickness calculations. Metal gaskets generally require very large sealing stresses,
much higher than the hydrostatic pressure in the present case, and furthermore
are not reusable. O—ring gaskets, when mounted properly, are sufficient for
handling 1000 psi design pressure. A 1%" II) and 3/16" thick Viton O-ring is
used with the mounting grooves on the cover disc. The sealing stress for this
O-ring is 100 lb I per inch of the O-ring [68]. The stress on the O-ring should be
uniformly distributed for an optimum seal and thus larger number of bolts are
favorable. Four to six bolts, with nuts and locker washers, are reasonable for a

3%"-diameter bolt-circle. This bolt-circle diameter is the smallest diameter for

18

which the bolts and nuts can easily be put together without interference from

the reactor wall and the ﬂange welded laps.

The detailed calculations for the ﬂange ring thickness are presented in the
Flange Design section of Appendix A. The O—ring is assumed to be the only
place of contact between the ﬂange ring face and the cover disc and the analysis
and design formulas [69] presented in Appendix A are true only for this
particular case. The ﬂange ring thickness is calculated to be 0.256" and for the
sake of safety a plain ring ﬂange it" thickness was constructed. A 1"-thick
stainless steel-316 cover disc was used because it is much cheaper than Haynes-25
alloy and has sufficient strength at room temperature. The calculated thickness
is 0.611". The total deﬂection of the ﬂange ring and the cover disc under the
applied stresses (internal hydrostatic pressure plus the sealing stress of the bolts
and the weight of the reactor) is 0.0011" at the bolt-circle. This value is too
small to allow the ﬂange ring and the cover disc to contact each other under
stress bending and, therefore, the assumption of no contact other than at the

site of the O-ring is not violated and the design analysis is valid.

2.2.1.5 Cooling System 3 Thermal Stresses

 

At the ﬂanged end of the reactor vessel there are two NPT ﬁttings for gas
inlet/outlet and insertion of thermocouples that are leak proof at room
temperature. The temperature rise from room temperature at this end of the
reactor vessel, caused by heat transfer from the heated zone of the vessel during
experiments, may cause these ﬁttings to leak and therefore it is essential to keep
the temperature at the ﬂanged end at room temperature. Convective air-cooling
requires a fairly long reactor. An existing 44" long Haynes-25 reactor at
Department of Chemical Engineering at University of Florida at Gainesville

provides enough length to air-cool the reactor to room temperature. Thus it was

19

decided to study the possibility of cutting down the reactor length to 22" by
water-cooling the ﬂanged end to room temperature without producing severe
thermal stresses in the reactor wall. For this purpose a cooling system was
devised which consisted of a rotary water pump, a ten gallon glass jar, a three
way valve for draining the glass jar, a cooling coil around the reactor vessel near

the ﬂange, and some it" ID copper and Tygon plastic tubing.

2.2.1.5.1 Thermal Stresses

To calculate the thermal stresses, the length over which the reactor wall
has to be cooled from the furnace temperature to room temperature has to be set
ﬁrst and only then can thermal stresses be calculated. However, in setting this
cooling length some factors have to be taken into consideration. The cooling
system cannot be mounted on the reactor vessel very close to the heated zone,
because severe and damaging thermal stresses may develop inside the reactor
wall. On the other hand, mounting the cooling coils on a long reactor vessel far
away from the heated zone calls for a long and very expensive reactor, which is
also inconvenient for insertion and removal of samples. Thus an optimum
cooling length has to be found which will make the reactor short enough and at

the same time will avoid development of severe thermal stresses.

Preliminary estimations established that at the ﬂanged-end 2" is needed
for mounting a 1/8" Swagelok tube ﬁtting for gas inlet, mounting and
supporting the reactor and welding the ﬂange, and an additional 2%" for
mounting cooling coils (6 to 8 laps). At the closed end 8%" is needed to be
inserted inside the furnace for placing samples under study in the middle of the
12" heating zone of the furnace. That leaves 9" in between the heated zone and
the cooling coils for gradual cooling of the reactor wall from furnace temperature

to room temperature (Figure 2.2).

.meGoN 2350an» 3.. was 588.7238 @368”: 355559 «é oaswmh

mcﬂm wmcme/w
\

 

8\-.-,-\.\\. \

 

 

no>oo
ova—mam \ \\

 

/V-/-/-///%
mHZ-ﬁ?

 

 

 

mn\x_.\\\\\\\\\\\\M\\\\\\\

\mnse H cccccc no HOE onEnmI/

 

Imaj

 

vﬂ- -

 

 

 

8.88.%-\§§\\\\\\\\4\\\\\\\

UCHHIO
[In

 

 

 

 

 

 

\ .

 

l mHHOU 71.5 ocoN
a 000

 

\ \

 

 

21

The criteria for determining whether or not the thermal stresses are
damaging or not is to calculate all tensile stresses and thermal shear stresses in
radial, axial and angular directions and then evaluate the maximum overall
normal stress by Mohr’s circle [72]. The maximum overall normal stress at a
point is the highest combination of coupling tensile stresses with normal stresses
generated by shear stresses. If this maximum overall normal stress exceeds the
rupture stress in 10000 b (which at 1000°C is only one tenth of the yield
strength for Haynes 25 alloy) of the material, the shear stresses are damaging;
i. e., the reactor wall will fail much sooner than expected. The tensile stresses
alone are not damaging because the wall is already designed to handle the tensile
stresses; however, when coupled with normal stresses produced by thermal shear
stresses, the result may be damaging. This may become critical if the reactor

wall is just thick enough to handle only tensile stresses.

The underlying idea for evaluating shear stresses is to ﬁnd the
temperature proﬁle of the cooling zone and then to estimate elastic shear strain
caused by temperature gradients and thermal expansion in the cooling region.

The temperature proﬁle is calculated by solving the following equation [75]

—(k—) = (2’3)

with the boundary conditions
T = room temperature Tm = 77 °F at z = 0

T = furnace temperature T, at z = L

where L is the distance from the cooling coils to the furnace heating zone, T is
the temperature of the reactor wall, T, is the furnace temperature, 2 is the axial

distance from the cooling coils and K. is the thermal conductivity of the reactor

22

wall material. The important assumption made here is to ignore convective heat
transfer to ambient air or gases inside the reactor; i.e., there is no radial
temperature gradient and heat is transferred only in the axial direction. This
temperature proﬁle yields the maximum heat duty of the cooling coils, but
underestimates the thermal shear stresses in the hot regions of the wall because
the reactor wall is cooled faster than what is predicted from this equation. The
details of computing this temperature proﬁle are presented in the Thermal

Stresses section in Appendix A (Equations A-3 to A—9).

The next step is to relate elastic shear strain dr/ dz to the above

temperature proﬁle. This is done using the following equation

ﬂailiﬂ—T
dz dT'dz

(2-4)
where r is the radial distance of a point from the axis of the cylindrical reactor
vessel. The derivative dr/dT can be estimated from thermal expansion

coeﬂ'icient a of the reactor wall and the details can be found in Thermal Stresses

section in Appendix A (Equations A-10 to A—l4).

Once the elastic shear strain is known, the corresponding radial shear

stress rz, can be calculated [66]

Tzr = GE"; (2'5)

where G is the shear modulus of elasticity of the reactor wall. The ﬁnal

estimation from Appendix A for 7,, is

23

__ raG 3561

T2,. — 3 z
1+8.35—I:-

It can be seen clearly that 72,. is a function of both axial position 2 and radial

(2-5)

position r for a given cooling length L. It can also be seen that the points of
maximum shear stress are the points at 2:0 and r=ro; i. e., the points next to
the cooling coils on the outside of the reactor. At these points G is also a
maximum. However, this does not mean that these points are the weakest
points of the reactor wall, because the yield and the rupture strengths of the

reactor wall at these points are also at a maximum.

Thermal shear stresses in axial and angular directions (1', and n,
respectively) are zero because of assumed uniform and symmetric heating of the

reactor wall.

Normal tensile stresses in radial, angular and axial directions (0,, 0', and
0,, respectively) due to internal pressure p at a point (r,0,z) of the reactor wall

can be calculated from the following equations [66]

 

 

2 2
Ti 70
0’ = —p —— —— —1 (2-7)
7' 132—752 [ r2 ]
2 2
r- r
00 - P 2 l 2 J2— +1 (2'3)
7'0 —1°,' 1°
2
r.
a, = p 2' 2 (2-9)

24

It can be seen that these stresses are not functions of angular or axial positions.
That is what is expected because pressure is the same for all z and the normal
stress on the inner cylindrical wall of the reactor is independent of the angle 0.
Radial stress 0', is negative because it is compressive, angular and axial stresses

are positive because they are tensile.

Therefore, the maximum overall normal stress amax at a point, which is
calculated by Mohr’s circle from the shear and tensile stresses at that point, is a
function of the radial as well as the axial position. As discussed earlier, at the
hot end of the reactor vessel the shear stresses are lower in magnitude than those
at the ﬂanged-end, but the strength of the wall material is also lower. Thus, at
each point of the reactor wall amax should be calculated and compared with
rupture or yield strength of the wall material at that point. The results of
calculations for four diﬂerent sets of points of the reactor wall are presented in
Table 2.1. These four sets of points are two sets at the hot end, one on the
outside of the reactor wall (r0 ,L) and one on the inside (r,-,L ), and the other two
at the cooling coil, one on the outside of the reactor wall (r0 ,0) and one on the
inside (r,-,0). Again an operating internal pressure of 1100 psi was assumed for

calculating normal tensile stresses.

Table 2.1 Maximum overall normal stresses in the reactor wall

 

 

point a, o, a, r, r, 7,, d T / dz G a om”
(psi) (Kpsi) (Kiwi) (psi) (psi) (Kpsi) (‘ F/in) (Mpsi) (:4 in/in- ’ 1“) (Kpsi)

(r0 ,0) 250 0.52 0 0 0 11.8 395 13.1 5.83 11.9
(r,- ,0) 250 1.52 -1.1 o 0 5.15 395 13.1 5.83 4.79
(r0 ,L) 250 0.52 0 0 0 1.55 129 3.85 9.41 1.59
(r; ,L) 250 1.52 -1.1 0 0 0.58 129 3.85 9.41 1.52

 

 

 

 

 

 

 

 

 

 

 

 

 

25

From Table 2.1 it can be seen that even though the maximum shear
stresses and the maximum overall normal stresses are at the location of the
cooling coils (z=0), the weakest spots of the reactor wall are at the inlet of the
heated zone (z=L ), where the overall normal stresses are 15% lower than the
rupture stress amptm in 10000 h. The assumption that is made for calculating
G’ at 1000° C is that the Poisson’s ratio p. does not change with temperature.
This ratio is assumed to be equal to 0.3, as in ﬂange design, and G is calculated

from the following formula [72]

=__L_
2(1 +u)

(2-10)
where E is the Young modulus of elasticity. The value of u is 0.25 for a
perfectly isotropic elastic material, but for most metals it is closer to 0.33 [73].
Thus, even if Haynes-25 is assumed to be a perfect isotropic elastic material at
high temperature, the value of G will be only 4% higher than C for [1 equal to

0.3, and the overall normal stresses will still be less than ”rupture .

Close examination of values of 0', and 72,. in Table 2.1 reveals that at the
inner wall of the reactor, the major contribution to amax is from tensile stresses,
and thus the assumptions that were made for calculating 1‘2, do not play an
important role in evaluating the signiﬁcance of shear stresses. At the outer wall,
on the other hand, the major contribution comes from the shear stresses. The
assumption of no radial heat transfer is justiﬁed by the fact that 2" of the
reactor wall at the furnace inlet is insulated by ﬁre bricks from outside and thus
heat transfer to the surroundings by radiation or convection is minimal. Heat
transfer by convection to gas(es) inside the reactor is also small because of high

residence time of the gas(es) and the quartz lining inside the reactor. This has

26

been veriﬁed by recent experiments in which the temperature of a sample was
measured under vacuum and in the presence of 500 psi hydrogen gas, the
diﬂerence was found to be only about 10° C. The assumption of regarding the
reactor wall as a solid cylinder, explained in Appendix A, is also reasonable

because of the high thickness of the reactor wall.

Hence, the overdesign of the reactor wall, as calculated from Barlow’s
formula (Equation 2-1), is large enough to tolerate the additional normal stresses
induced by thermal shear stresses; and a length of 9" for cooling zone is

economical and safe (Figure 2.2).

2.2.1.5.2 Cooling Coils

The required total surface area A for heat transfer (and length of the
cooling coils LC) depends on the magnitude of the overall heat transfer coeﬂicient

ho , the temperature driving force AT and the heat load Q

A = —_Q__ (211)

The magnitudes of both heat transfer coefficient ha and temperature
driving force AT depend on the ﬂow regime inside the cooling coils [74]. Thus it
is necessary to determine whether the ﬂow regime is laminar or turbulent. This
is done by a simple energy balance which determines the mass ﬂux m of the

cooling water; the Reynolds number Re is then calculated from that:

 

12, = (2-12)

27

The temperature driving force AT is calculated from the temperature rise
of the cooling water and the temperature of the reactor wall. For this purpose
the inlet temperature T ,- of the cooling water was assumed to be at 70 ° F with a
temperature rise of 10°F. The reactor wall temperature T was assumed to be
at a constant temperature of 85°F. The details are presented in the Coolng

Coils section of Appendix A.

The important assumptions made are that i ) the overall heat transfer
coefficient ha is equal to the heat transfer coefficient across the liquid ﬁlm in the
cooling coils, because the thermal conductivities of the copper tubes and the
reactor wall are relatively high and thus impose negligible resistance to heat
transfer, and ii ) the temperature of cooling coils is uniform throughout, even
though heat is transferred from reactor wall to cooling coils only at the side of
cooling coils which is soldered to the reactor wall (which is less than half of the
cooling coil surface area A). The very high thermal conductivity of the copper
and the thick walls of copper cooling coils (3/64") justify these assumptions.

The results in Appendix A show that the ﬂow regime inside cooling coils
is turbulent and 2-3 laps of copper tubing around the reactor are required for

keeping the temperature of the ﬂanged end at room temperature. Six laps of

copper tubing were used.

2.2.1.6 Reactor Quartz Lining

 

The reactor wall goes through the thermal cycles of heating and cooling
from room temperature to reaction temperature or degassing temperature and
back. During these thermal cycles some metallurgical carbon from the reactor
wall alloy is deposited out on the reactor wall and reacts with hydrogen gas at
high temperature to yield methane. The rate of this reaction, although very

small, interferes with reaction rate measurements of graphite, as will be

28

discussed in Section 5.2. To reduce the reaction of hydrogen with this
metallurgical carbon, a quartz lining was inserted inside the reactor. This quartz
lining is a 22mm OD X 21" quartz tube closed at one end and 2" of one side at
the other end cut off to accommodate the hydrogen inlet stream. This quartz
tube ﬁts snugly into the reactor; for hydrogen to react with the metallurgical
carbon it has to diﬂuse through the very narrow space between the reactor wall
and the tube to the hot end, and then product methane has to diffuse in the
reverse direction to enter the hydrogen ﬂow stream. The tube reduces the

interference from the metallurgical carbon substantially.

2.3 q-Ray Detection System

In the experiments involving potassium carbonate as a catalyst,
determination of the amount of potassium in a sample at each point in time can
provide a great deal of information on the catalysis process, as will be discussed
in Chapter 5. One way of determining the amount (or relative amounts) of

potassium is by in situ neutron activation analysis (NAA).

NAA involves bombarding nuclei of potassium atoms in a sample with a
high ﬂux of neutrons (~ 1012-10l3 neutrons/cm2-s) inside a nuclear reactor, and
then monitoring the energy and intensity of radiation emitted by unstable
isotope(s) formed. The energy of emitted radiation serves as a ﬁngerprint of

unstable isotope(s) formed, and the intensity serves as a measure of the quantity.

The high ﬂux of neutrons interacts with the nuclei of potassium 41K
atoms, which constitute about 6.9% of the potassium found in nature, and yield
4("K nuclei. 42K is an unstable radioactive isotope which decays by a ,B-decay

mode with a half life t”, of 12.36 h [76]. The products of this decay are 3-

29

radiation (electrons) and 420a nuclei, of which 81% are in ground state, 18% are
in the lowest excited state (1.525 MeV, t.‘ = 0.82X10"l2 seconds) and 1% are in
excited states of higher energies with tg’s of less than 10'9 seconds. The excited
states of 420a all decay to ground state or lower excited states by releasing their

excitation energies as 'y-rays (photons).

The Mech-Tronies Nuclear ’y-ray detection system used consists of a
Harshaw 3%")(3W' NaI(Tl) scintillation crystal with a photomultiplier, a mode]
450 preampliﬁer probe assembly, a mode] 500 RC pulse ampliﬁer, a mode] 251
high-voltage power supply, a mode] 602 single channel analyzer, a model 1304

printer/interface, a model 706 sealer and a model 756 timer.

2.3.1 Theoretical Derivations

 

2.3.1.1 Mass-Radiation Relation

 

The radioactive decay of 42K is a statistical process and the intensity of
decay radiation (ﬂ-ray or 'y-ray) at a macroscopic level is thus directly
proportional to mass. Therefore, for a given sample a knowledge of weight
fraction of total potassium as 42K and the radiation rate or nuclear
disintegration rate at each point in time indicates how much potassium is
present in that sample at that particular time. The number of atoms [V42 of
radioactive potassium 4""K in the sample at any time is related to the nuclear

disintegration rate R:

E = .1?” (213)

where 6 is a proportionality constant equal to In 2/t%. The number of atoms

N042 of radioactive potassium 42K at the end of neutron bombarding of N0“1

30

atoms of 41K for a period of t,, is calculated from the following equation
42 41 774’ -ct.
N0 = N0 —6’(1—€ ) (2‘14)

where <I> is neutron ﬂux (1012-10l3 neutrons/ch-s) and r) is the cross section of
41K atoms for neutron absorption, which is equal to 1.5)(10-24 cm2. At time t
after the end of neutron activation some portion of N042 decays away and what
is left, N‘12 (not necessarily all in the sample), is calculated from the following

equation
N42 = N042 c-(t (2_15)

If no potassium is lost from the sample, N42 and N42 are numerically equal.
However, if some potassium is lost , say by vaporization from the sample during
this time, the radiation rate will be lower and the fraction f of potassium lost

from the sample is calculated from the following equation

N42_N42) 6 ct

f = ( N042

(2-15)

 

The total mass M of potassium in the sample at any time is then related to

initial total mass M0 of potassium according to the following formula

M = (1—1 )M. (2-17)

31

Substitution of appropriate values from Equations 2-13 through 2-16 into

the above equation yields

R ed
M = M 2-18
0 N04117CI> (l—e—‘t°) ( )

 

where N041 is calculated from M0 using the Avogadro’s number Nam, , potassium

molecular weight WK, and the fraction x of potassium as 41K;

N41 = XMO N600

2.3.1.2 Radiation Measurement

 

2.3.1.2.1 Problems

The only parameters that need to be measured during an experiment to
calculate the amount of total potassium M in the sample at any time from
Equation 2—18 are ta,t and R. The measurements of ta and t are
straightforward. The measurement of R, however, presents a great deal of
diﬂiculty because there is no way of counting number of nuclear disintegrations
directly. There are detection systems which detect a fraction of radiations by
counting the number of electrical pulses generated proportionally by ﬂ-rays or
'y-rays. Effective use of a detection system requires a knowledge of eﬂ'lciency of
its detector; that is, how many electrical pulses are generated and counted for
each B—ray or ’7-ray that enters the detector. This efficiency depends on the

nature and energy of the radiation as well as the type of the detector.

32

The problem of measuring the R is further complicated by the fact that
there are shielding materials between radioactive sample and the detector. The
shielding materials interact with radiation via several modes and result in
absorption of some or all of the radiation. If most of the radiation is absorbed
by the shielding materials, there may not be enough radiation entering the

detector to measure R reliably.

Thus the radiation rate as measured by the detection system B. is
substantially less than the actual nuclear disintegration rate R of the sample,
due to absorption of radiation by shielding materials and inefﬁciency of the
detector system to detect all the radiation reaching the detector. Therefore, the
relationship between R— and B should be established before Equation 2-18 can be
used to calculate M.

2.3.1.2.2 ﬁ-Rays

As mentioned earlier, 42K decays to 420a with emission of ﬁ-rays and ’7-
rays, both with different but discrete energies. The ,B-rays have much lower
penetrating ability than the '1- rays. The maximum energy that ,B-rays in 42K
can have is about 3.5 MeV, and the maximum range of penetration in Haynes-25
is about 0.075"[79]. This is much less than the Haynes-25 reactor wall thickness
of 9/ 16". Therefore, no ﬂ-rays reach the detector outside the reactor, and

detecting the ﬁ-rays as a means of measuring R is not possible.

2.3.1.2.3 ’y-Rays

Of the unstable 42K isotope formed by neutron bombardment, 18%
decays to an excited 42Ca nucleus, which subsequently decays to the ground state
and releases its excitation energy of 1.525 MeV as a 'y-ray. These ’y-rays have a

high penetrating ability even through metals, with a half value layer (the

33

thickness of shielding to reduce the intensity of the ’y-ray emission by half) of
1.174 cm for lead. As the calculations given in the Radiation System section of
Appendix A show, a 6 mg K2003 sample activated for 80 minutes emits enough
radiation to allow reliable counting rates (10 times the background). Therefore,

detection and counting the 1.525 MeV ’y—ray is a practical means for R

measurement.

2.3.1.2.4 Interaction and Absorption of '7-Rays

Any measurement of ’y-ray emission, along with a knowledge of detector
efﬁciency, can just determine the intensity of the radiation in the immediate
neighborhood of the detector. To relate this measurement to radioactivity of
42K in a sample and thus to total mass of potassium, it is necessary to estimate

the interaction and subsequent absorption of the ’y-ray by shielding material.

The extent of absorption generally depends on the energy of the ’y-ray,
the nature of the shielding material (atomic number Z and density p) and its
thickness. If a 'y-ray radiation of intensity 1,, passes through a shielding
material of thickness 1:, the radiation intensity is reduced to a value I which is

calculated from the following equation [76,77]

I = I, 5"“ (220)

where A is the attenuation coefficient of the material. Its magnitude depends
on the 'y-ray energy and nature of the shielding material only; that is, it is
independent of the thickness. Thus the extent of absorption of ’y-rays by
materials between a sample and the detector can be calculated from the

thickness of the shields and their values of A.

34

2.3.1.2.5 Attenuation Coefficients

Computation of values of A requires a knowledge of mechanism(s) of
interactions between ’y-rays and atoms or molecules of shielding materials.
These interactions can take place by eight diﬁerent mechanisms [77], but for the
'y-rays of interest practically only three mechanisms are responsible for the
absorption [78]. These three mechanisms are the photoelectric effect, Compton
eﬂect, and pair production. Hence, A can be written as a sum of three terms,

each of which is the contribution from one of these three mechanisms:

A=u+€+K (2'21)

The variables V, E and K are coeﬂ'lcients of photoelectric effect, Compton

scattering, and pair production, respectively.

2.3.1.251 Photoelectric Eﬂect

Photoelectric effect is the inelastic collision of a 'y-ray with an orbital
electron whereby all of the energy of the ’y-ray is transferred to the electron
(usually an inner K shell). The result is ejection of the electron from the atom
and production of an ion pair. The photoelectric coeﬂicient u is estimated from

the following relation [77,79]

z4.55

u = 5.05x10-33 me T
t

(222)

where A, is the atomic weight of the shielding material and p is the density of
the shielding material. The photoelectric effect is more important at lower

energies.

35

2.3.1.252 Compton Effect

Compton effect or Compton scattering involves a collision between a ’7-
ray and an electron in which a part of the energy of the 'y-ray (as opposed to all
of the energy in photoelectric effect) is imparted to the electron. The ’y-ray
energies from the collision in a new direction and with reduced energy. The

Compton scattering coefﬁcient E is estimated from Equation 2—23 [79]

z
E = men. 71— (2-23)
t

where ’7: is the Compton cross section per electron. The value of 1], depends on
the energy of the 'y-ray; for the 1.525 MeV '7-ray of interest it is roughly
1.6X10’25 cm2 [79]. Compton scattering is especially important for 'y-rays of
medium energy (0.5 to 1.0 MeV).

2. 3.1. 2.5. 3 Pair Production

In pair production, a 'y-ray passing close to a nucleus (i. e., in the vicinity
of a strong electromagnetic ﬁeld) suddenly disappears and an electron (negatron)
and a positron are created in its place. The rest-mass energy of these two
particles is 1.02 MeV and therefore only 'y-rays with energies in excess of 1.02
MeV can go through this process, with the difference in energy appearing as the
kinetic energies of the electron and the positron. The coefﬁcient K is computed

from the following relationship [77]

K = 1.19><10—29 Nm 22(E,—1.02) (2-24)

36

The above three equations are used to calculate the value of A for
Haynes-25 alloy. The contributions from the three mechanisms to attenuation
coefﬁcients of individual components (major components) in Haynes-25 alloy are
presented in Table 2.2. The attenuation coefficient A of Haynes-25 alloy is

simply the volume—weighted sum of the attenuation coefﬁcients of the individual

components:

11
A = 21);)».

i-l

(2—25)

where n is the number of components in the alloy, 12,- is the volume fraction of

component i, and k; is the attenuation coefﬁcient of component i.

Table 2.2 Attenuation coefﬁcients of lead and Haynes-25 component-

 

 

 

 

element p wt % 0, Z At 10003 f K X103 )x
g/cm3 cm‘1 cm“l cm”l cm"l
Co 8.9 53.55 0.555 27 59 1.79 0.392 2.70 0.397
Cr 7.18 20.11 0.257 24 52 0.960 0.319 2.13 0.322
W 19.3 15.54 0.0741 74 184 122.5 0.748 20.3 0.891
Ni 8.902 10.8 0.114 28 59 2.116 0.407 2.90 0.412
Pb 11.35 100 1.00 82 207 102.1 0.433 24.3 0.559

 

 

 

 

 

 

 

 

 

 

 

The attenuation coefﬁcients for other shielding materials used in this
system are readily available in the literature [79] and are tabulated in Table 2.3
along with the estimated attenuation coeﬂicient for Haynes-25 alloy from

Equation 2-25.

37

Table 2.3 Attenuation coeﬂicient A for various shielding media

 

 

Material Attenuation coefficient Ajcmﬁ
Air 0.0000513
Firebrick 0.025
Iron 0.380
Lead 0.578
Haynes-25 0.416

 

 

 

 

2.3.1.2.8 Absorption

For a sample of radioactivity lo, the fraction of radiation f I passing

through the shielding materials in Figure 2.3 can be calculated from

(-X:30A.z.)

I = 10 e (2—26)

where 9:,- is the thickness of the i th shielding layer and A,- is the attenuation

coefﬁcient of that layer.

Substitution of appropriate values for 2:,- from Figure 2.3 and A,- from

Table 2.3 yields
I
f, = — = 0.334 (2-27)
10

The details of the calculations are presented in Appendix A in the section on ’7—
ray calculations. Therefore, only 33.4% of ’7-rays emitted by 42K in the sample

pass through the shields and reach the detector.

rrrrrrrrrrrr

 

 

cccccc

 

 

 

 

 

 

 

3 3/4" Firebrick

 

 

..:1:6;_1////////////////////
.////////////////_//_&

 

 

 

 

39

2.3.1.2.7 Geometric Factor

The ’y-rays from a sample are given off in all directions and hence only a
fraction Id of the Trays are detected by the Na.I(Tl) detector. This fraction [,1
is equal to the fraction of the surface area of a sphere with radius Ld equal to
the distance of the detector from the sample, which is covered by the surface of

the detector. This is given by

 

 

/. = (2-28)

where Dd is the diameter of the detector crystal (3%"). This fraction turns out
to be 0.0051.

2.3.1.2.8 Sample Activation

The efﬁciency 11) of the NaI(Tl) detector for Trays is 4-5% [80] and the
measured background counting rate is 5-6 counts per second (c/s). Thus to
obtain a counting rate of 60 cps (i.e., ten times the background) the

radioactivity Io of the sample needed is

I _ 60 cps

. " 77.7; ‘2'”)

Note that Io and I? are related but are not equal; Io is the rate at which ’7—rays
are emitted from a sample, whereas P- is the rate of nuclear disintegrations (of
which only 18% emit 'y-rays). The required value of P- from the Sample

Activity section of Appendix A is 3.91X106 disintegrations per second (dps); or a

40

’y-ray intensity Io of 7 .04X105.

Now assuming that an experiment takes 2 hours and that during the
experiment 80% of the potassium is lost, say via vaporization, then the nuclear
disintegration rate needed at the beginning of the experiment I? lt-o is 590 140i

as calculated from the following equation

— 3.91x106 tips 2 ,,
R —_ .. 2-30
I"° (1— 0.8) c ( )

 

The samples containing potassium are allowed to be activated to only 100
pCi because of restrictions imposed by Ofﬁce of Radiation, Chemical and

Biological Safety at Michigan State University.

The time ta needed for activating a sample in the nuclear reactor to 100
uCi can now be calculated from Equations 2-13 and 2—14. For a typical 60 mg
carbon-catalyst sample of 10 wt%-K.2003 the amount of 41K is 0.234 mg, or
3.44X1018 atoms. Disintegration rate I? of 100 ”Ci corresponds to 2.38X10ll
atoms of 42K (Equation 2-13). Thus with a neutron ﬂux of 1012 neutron/cmz-s
an activation period of 25.3 hours is required. For samples of this size, therefore,
a neutron ﬂux of 1013 is used so that the activation time is reduced to 1.41 hours

or 84 minutes (see Appendix A, Sample Activation).

To prevent detecting ’y—rays from sources other than the sample itself (say
vaporized potassium which condenses on cooler parts of reactor vessel) at least 6"
of lead bricks are used to focus the detector. The ’y-ray absorption of these lead
shields is calculated from Equation 2-20 and Table 2.2, and is found to reduce
the intensity of the '7-rays from other parts of the reactor vessel by a factor of

7400 (see Appendix A, 'y-Ray Absorption).

41

2.3.2 Potassium Content Measurements

 

Therefore I? can be calculated from R;

1

 

if = 15’. 2-31
0-181/Jf d f I ( )
and can be substituted into Equation 2-18 to obtain
ct
M 8 (2-32)

 

 

= B.
Ma 0.18N041174’Wdf1 (l-e‘“‘)

Comparison of the terms in the denominator with Equations 2-13 and 2-

14 indicates the denominator is just B. lt-o: and thus

M R ct

a: 6 2'32
M. R I.-. ( )

 

if Id and it stay constant with time.

2.3.3 Biological Eﬁ'ects of '7-Rays

 

The biological effect of 100 ”Ci of '7-ray radiation is calculated using the

Gamma constant I‘ of 42K [78] which is

F = 0.14 Roentgens per hour per curie at 1 meter

= 0.13 rad/ h per curie at 1 meter.

This yields

42

I = 0.013 m rem/h at 1 meter
= 130 m rem/h at 1 cm

which is in the relatively safe range.

2.4 Gas Collection System

The devised gas collection system consists of three %"—OD stainless steel
Ushape cold traps (labeled A, B and C) each of which is connected on either
side to two miniature 2—way Brunswick Technetics solenoid valves (model A-

2013)

2.4.1 Solenoid Valves & Electronics

 

The response time to close or open these valves is 5-25 ms. Of the two
valves on one side of a cold trap, one is attached to the reactor effluent line and
is called the “reactor” valve, and the other valve is attached to the helium
carrier gas line of the gas chromatograph and is called the “GC” valve. The two
“reactor” valves or the two “GC” valves of a cold trap can only be activated
simultaneously, from a switch board of 12-pole 3-position rotary switches. The
switch board houses two of these switches, one for the “reactor” valves and one
for the “CC” valves. Each position of the three positions on the switches (called
positions A, B and C) activates the valves of the corresponding trap. Thus these
traps can be put in-line with the reactor effluent stream for gas collection, and
the GC helium stream for gas analysis by activating the corresponding valves.
A cold trap can be in-line with both the reactor efﬂuent stream and the GC

helium stream at the same time.

43

2.4.2 Cold Traps

 

The cold traps are inserted into liquid nitrogen to condense and collect
product gases such as CH4, CO and CO2 in the reactor effluent stream. The
cold-trap tubes should be long enough such that the tube ﬁttings connected to
the valves do not experience very cold temperatures, otherwise they might leak.
Stainless steel is used instead of copper to reduce the temperature drops at the
ﬁttings when the traps are inserted into liquid nitrogen. The cold-trap tubes are
29" long, and when inserted into liquid nitrogen there is at least 3" distance
between the liquid nitrogen level and the ﬁttings. Depending on the level of

liquid nitrogen 5.5"—11" of the cold trap is inserted in liquid nitrogen.

2.4.3 Silica Gels

 

The liquid nitrogen temperature alone is not enough for cooling and
condensing the product gases fast enough so that they can be collected
quantitatively, because of low surface area for heat transfer. To alleviate this
problem, 1.4-1.7 g of 6-16 mesh silica gel manufactured by J. T. Baker Chemical
company is added to each cold trap. This silica gel is highly porous and can
absorb more than 30% moisture. The problem with this amount of silica gel is
that the gas ﬂow rate through a trap cannot be raised very high; e.g., a ﬂow rate
of ~ 400 ml/min blows some of the silica gel out of the cold traps and to the

solenoid valves. Thus for experiments ﬂow rates of ~ 310 ml/ min are used.

44

2.5 Gas Analyzing & Calibration System

The gas analyzing apparatus includes a series 5750 F&M (Hewlett
Packard) research gas chromatograph (G0), a Sargent-Welch recorder model
XKR with an electronic integrator, 99.999%—pure helium gas, a six-port

calibration valve, and a calibration gas mixture.

2.5.1 Gas Chromatograph & Recorder

The GC is equipped with a thermal conductivity cell (TC cell) and two
60/80 Carbosieve S-II stainless steel columns, 5’ X1/8", installed in parallel. The
TC cell is operated at room temperature at a current of 150 mA. The
Carbosieve S-II column is suitable for analysis of permanent gases and light
hydrocarbons, and the recommended temperature programming is: “hold 2 min
at 35° C, then to 175 ° C at 20° C/min and hold,” with a helium ﬂow rate of 30
ml/min [81].

In proposed experiments, however, no hydrocarbon other than methane is
expected to be in the reactor efﬂuent gas and thus each gas analysis cycle is
expected to be completed in 8 minutes. The difﬁculty with this temperature
programming is that the CO peak and N2-02 peak overlap and cannot be
accurately quantiﬁed. Various trials shows that holding at 50-60° C instead of
35°C for 2 min andﬁow rates of 37-40 ml/min improves the resolution and
decreases the cycle time to 7.5 minutes. The pressure of helium carrier gas is 30
psi.

The recorder chart speed and voltage/full-range ratio can be varied from
0.05 to 20 cm/min and 10 to 100 mV, respectively. The voltage span of the
integrator is 1 to 1000 mV. Thus a wide range of peaks can be analyzed with

this recorder accurately, provided the GC column is not ﬂooded.

45

2.5.2 Calibration Setup

 

Two gas mixtures have been used for calibrating the GC signals: a Linde
helium gas mixture of 4.8% CH4, 4.9% CO and 4.9% CO2, and an Airco gas
mixture containing 5.20%, 5.33% and 5.26% by volume CH4, CO and CO2,

respectively.

A 6-port “43Y6” series Whitey ball valve (two-position) is used for
sampling and injecting the calibration gases into the GO. A 20"X1/8" and a
10.5"X1 / 8" 304-stainless steel tube are used for sampling 1.81 ml and 0.87 ml of
the calibration gases, respectively. The sampling tube is swaged to two opposite
ports on the hexagonal base of the calibration valve; two ports on one side of
these ports are connected to the calibration gas line, while the other two ports
are placed in-line with the helium carrier gas. This way, the sampling tube is
either in-line with the calibration gas stream or the carrier gas. By switching
between the two positions a predetermined volume of the calibration gas is

injected into the GC carrier gas stream and analyzed.

2.6 Gas Purifying System

The 99.999%—pure helium and hydrogen gases used in experiments still
contain some contaminating impurities such as oxygen and water vapor, and to
further purify these gases a purifying system has been devised. Linde molecular
sieve 3A (Alpha Products) with a nominal pore diameter of 3 A, which adsorbs
any molecules with an effective diameter less than 3 A, are used as purifying
agents. 500 ml Whitey sample cylinders are ﬁlled with 1/8" pellets of the
molecular sieves and are inserted in line with helium carrier gas and hydrogen

gas lines. Each cylinder is enough for purifying ﬁve 250 ft3-cylinders of

46

hydrogen or helium.

The molecular sieves are regenerated by heating them to 270° C in vacuo

with 1"X8' ﬂexible Samox—insulated electrical heating tapes for 45-60 minutes.

2 .7 Evacuation System

When samples are inserted into the reactor at the start of an experiment,
or hydrogen gas is replaced with oxygen gas for oxygen chemisorption during an
experiment, some air or hydrogen is trapped inside the reactor. It is essential to
remove the undesirable gases to avoid complications arising from the oxygen
present in the air or explosion of hydrogen/oxygen gas mixture at high
temperatures. A model MS-90-AB Veeco leak test station is attached to the
reactor vent line so that a vacuum of 1-10 qu can be obtained on the rough
pump of the station alone. Pressures as low as 10"5 Torr can be attained using
the diffusion pump of this station, but leaks at some joints makes it impossible
to obtain such low pressures in practice. Thus, only the rough pump is used in

experiments.

2.8 Flow & Pressure Control

2.8.1 Flowmeters

 

The only device to control the reactant gas ﬂow rate is a set of two
parallel Cole-Farmer variable area ﬂowmeters, one with glass ﬂoat ball and one
with 316 stainless steel ﬂoat ball. With this set up an air ﬂow rate range from

0.50 to 832 ml/ min can be obtained. The ﬂowmeters are calibrated for helium

47

and hydrogen gases by a soap-bubble ﬂowmeter.

2.8.2 Pressure Regulators

 

There are three different pressure levels in this complete reactor system;
reactor vessel pressure, pressure at the cold traps, and the pressure at the ﬂow
meters (which is virtually atmospheric). For each pressure level, pressure
regulators are used: Matheson two-stage high-pressure regulators at the
hydrogen and helium cylinders which feed the reactor, a two-stage Matheson
pressure regulator model 3104A which reduces the reactor pressure to the GC
carrier gas pressure of 30 psi, and a Fairchild-Kendall pressure regulator model

10 to reduce the pressure further to atmospheric pressure of the ﬂowmeters.

2.9 Rupture Disc

The high pressure hydrogen gas is fed from a high pressure cylinder (2500
psi) to the reactor at 1000 psi pressure through a two stage pressure regulator.
In case this pressure regulator fails or is misadjusted, or the reactant gases inside
the reactor explode, the reactor pressure will exceed the maximum allowable
working pressure (MAWP) and the reactor vessel will be damaged. Thus a
rupture disc is installed in the feed line such that it will rupture if the pressure is
raised above MAWP, thus releasing the pressure. For this purpose a Fike
stainless steel 316 Conventional Prebulged rupture disc (59") is used. The details
of the calculations for the speciﬁcations are presented in the Rupture Disc section

in Appendix A.

48

The stamped burst pressure of the purchased discs is 1448 psi and the
recommended ratio of operating pressure to burst pressure for this particular
rupture disc is 70% [82]. The ASME code concerning rupture discs requires
burst pressure not to exceed MAWP when the rupture disc is the primary or sole
relieving device [82]. This implies that the reactor wall can be raised only to a
temperature where the MAWP is 1448 psi, as opposed to 1100 psi MAWP at
1000° C. As calculations in the Rupture Disc section of Appendix A indicate,
this maximum operating temperature is 970°C, which is very close to the
original design temperature of 1000 ° C. The reason the temperatures are is that
in the latter case the reactor was overdesigned using Barlow’s formula, and in

the former case Lamé’s formula (Equation 2-28) was used with shear thermal

stresses being taken into account.

2.10 Sample Holders

For intrinsic reaction rate studies two different types of sample holders
were used as illustrated in Figure 2.4; a ceramic tube sample holder and a quartz
sample holder. Both of these sample holders are inert toward carbon, hydrogen
and methane, and with each one a 250 or 325 mesh 316 Stainless Steel 316
screen was used to avoid blowing the carbon out of the sample holder. The
ceramic sample holder consists of a ceramic tube which was made into a boat (1"
long and 1/8" deep) using Aremco high temperature 503 ceramic adhesive and
which holds about 50 mg of Raven 16. The quartz sample holder is a custom
designed boat (2’72" long and '72" deep) made from 5/8" OD quartz tube, and
holds 1.5 g of carbon.

 

Figure 2.4 Sample holders used for hydrogasiﬁcation studies.

50

2.11 Temperature Measurements

The temperature of carbon samples during experiments are measured
using two 30" subminiature (0.040" sheath diameter) Chromel Alumel
thermocouple probes with 304 Stainless Steel sheath, 11°C accuracy. These
thermocouples are inserted into the reactor through a two-hole Conax ﬁtting
(with Viton sealant), mounted on the reactor vessel cover disc, and the central
exit tube. One thermocouple is inserted inside carbon bed and the other
thermocouple is in the gas phase above sample holders. Chromel Alumel
thermocouples are suitable for high temperature operations in reducing as well as

oxidizing atmospheres.

Chapter 3

Characterization of Carbon

The reactivity of a carbon sample toward reactant gases depends on
various parameters such as the extent of crystallinity and graphitization of the
carbon, the amount and nature of impurities in the carbon, and the interfacial
surface area available for reaction. For instance, a carbon sample with a high
degree of order and crystallinity is less reactive than an amorphous carbon
sample (carbon with no order at an atomic level). The impurities such as iron
and sulfur in a carbon sample, on the other hand, may enhance or reduce the
gasiﬁcation rate, respectively, depending on the nature of reactant gases. A
carbon sample with a high surface area would react with gases faster than a
similar carbon with a lower surface area because more gaseous molecules would

attack the carbon surface when the surface area is higher and hence a higher

51

52

reaction rate is obtained. Furthermore, these parameters may also inﬂuence the
activity of a catalyst similarly. For example, the kinetic data obtained for
gasiﬁcation of a carbon might be very different from those for another type of
carbon loaded with the same catalyst due to different degrees of crystallinity or

impurity content.

Therefore, in kinetic studies of catalysis for carbon gasiﬁcation it is
essential to determine the extent of contributions from these sources to observed
reaction rates before the sole effect of a catalyst is evaluated. However, before
the extent of contributions from these sources to gasiﬁcation rates can be
measured, their relative magnitudes should be estimated; i. e., the magnitude of
surface area, the degree of crystallinity and the amount of impurities in a carbon
sample should be known or, in other words, a carbon sample should be

characterized ﬁrst.

For gasiﬁcation studies three kinds of carbons were used, as will be

discussed in Chapters 4 and 5; activated coconut charcoal, graphite and carbon

black.

3.1 Coconut Charcoal

The coconut charcoal used was a 50—200 mesh activated coconut charcoal

obtained from Fisher Scientiﬁc Company.

3.1.1 BET Surface Area

 

The surface area of this carbon was measured by volumetric adsorption of
nitrogen gas at liquid nitrogen temperature using BET analysis [83]. Nitrogen

gas adsorption partial pressures were 59 Torr, 147 Torr and 212 Torr and

53

adsorptions at each nitrogen gas partial pressure were carried out in triplicate
for the sake of accuracy. The calculated surface area was 577 m2/ g assuming a
coverage of 16.2 A2 by a nitrogen molecule. Large carbon particle sizes and high

surface area indicate a high degree of porosity of this particular type of carbon.

3.1 .2 Ash Content

 

The ash content (the residue left behind after burning the carbon in air)
was determined by burning three samples of this carbon (1.7363 g, 3.3485 g and
5.4090 g) at 850°C in air for 11 h in 60 mm ID evaporating dishes. The
samples were then cooled down in a desiccator and subsequent weight

measurements showed an average ash content of 3.70%.

3.1.3 Impregnation with K2C03

The carbon samples were loaded with K2003 by wet-impregnation with a
mixture of K2003 solution/acetone and subsequent drying. First, 200 ml of
1.799 M solution of potassium carbonate was prepared from analytic reagent
Mallinckrodt granular anhydrous potassium carbonate. Then 8.0 ml, 16.0 ml
and 24.0 ml of this solution were diluted to 60 ml with doubly distilled water in
200 ml beakers and were each mixed with 20.0 g of carbon to yield 9.04 wt%,
16.6 wt% and 23.0 wt%-K2003 samples, respectively. After 5 minutes of
mixing, 2 ml of acetone was added to each beaker and the beakers were covered

and were left overnight.

The samples were dried at 90 ° C overnight without mixing and the result
was a large degree of recrystallization of potassium carbonate. Thus 45-50 ml
distilled water was added to each beaker with 5 minutes of mixing to redissolve
the potassium carbonate crystals. The samples were then slowly dried by

heating them at 90°C for 8 h with stirring every 30-45 minutes. Following

54

that, the samples were kept at 60° C for one week and stirred once a day, and
then kept at 120° C for two days. The samples were ﬁnally cooled down and

were stored in 20 oz french square glass bottles with screw caps.

3.1 .4 Reactivity

 

These KQCO3-impregnated samples were highly reactive toward 500 psi
hydrogen gas at 865 ° C (see Section 5.3). The 23.0 wt%-K2003 sample was so
reactive that more than half of the sample gasiﬁed before the steady state
temperature of 865 ° C was reached. Thus, no useful kinetic data at low
conversions could be obtained from this carbon and, therefore, it was decided to
abandon this type of carbon for further studies, and to use other carbons such as

graphite for the experiments.

3.2 Graphite

Since coconut charcoal was too reactive, it was decided to use a less
reactive carbon sample; a carbon sample with a higher degree of crystallinity and
lower surface area. For this purpose a non-porous -325 mesh Alpha Products

graphite powder with a purity of 99.5% was tried.

3.2.1 BET Surface Area

 

The BET surface area of this carbon was determined using the same
technique that was used for coconut charcoal with one minor modiﬁcation; the
graphite sample was degassed at 125° C for 30 min before nitrogen adsorption
experiments. The nitrogen adsorption partial pressures were 49 Torr, 144 Torr

and 215 Torr, and the surface area was calculated to be 6.58 m2/ g. However, it

55

has been reported that this calculated surface area is too low and the effective
coverage area by nitrogen molecules should be taken as 20 A2 per molecule
instead of 16.2 A2 that has traditionally been used [84]. A proposed explanation
is that nitrogen molecules are localized at graphite lattice sites so that each
nitrogen molecule ﬁlls four of the unit hexagonals, an area of 21 A2. Thus the

actual surface area was 8.12 m2/g.

3.2.2 Impregnation with K2C03

 

This graphite powder was loaded with K9003 to 1.00 wt%, 4.73 wt% and
9.04 wt%-K2003 samples by wet-impregnation with a mixture of K2003
solution/acetone and subsequent slow drying. First, 0.8 ml, 4.0 ml and 8.0 ml
samples of a 1.799 M potassium carbonate solutions were added to three 100 ml
beakers using a 1 ml in 1 / 100 pipette and were diluted to 40 ml with doubly
distilled water. Then 40 ml acetone was added to each beaker, stirred and 20.0
g graphite was added. After stirring the solid-liquid mixture well, the beakers
were covered and were saved overnight. Samples were dried slowly at 65°C
with stirring them every 15 to 30 minutes until the powders dried enough to
agglomerate. After drying the samples completely at 120°C overnight, they
were ground in a pestle and mortar and stored in 20 oz french square glass

bottles with screw caps.

A blank graphite sample was also prepared along with other samples in
an identical manner except for the fact that no potassium carbonate solution

was used.

3.2.3 Reactivity

 

The measurement of gasiﬁcation rates of this carbon at 865°C was,

however, a problem. As will be discussed in Chapter 4, the gasiﬁcation rate of

56

fresh carbon was so low that there were signiﬁcant interferences from the
reaction of the reactant hydrogen gas with metallurgical carbon that deposited
out on the reactor wall during thermal cycles from room temperature to reaction
temperatures and vice versa. These reaction rates in general were higher than

the gasiﬁcation rates of graphite samples.

Thus it was decided to insert a quartz tube lining into the reactor vessel
to reduce the reaction of hydrogen gas with the metallurgical carbon. It was
also decided to ﬁnd an alternative carbon sample that has higher reactivity than

this rather inactive graphite sample, but still less reactive than the coconut

charcoal.

3.3 Carbon Black

The carbon black used for further studies was Fisher Scientiﬁc lampblack
(Raven 16 carbon black) which is a graphitic furnace black. This carbon black
is non-porous and consists of spherical particles with average diameters of 60-70
nm as determined by electron microscopy. The geometric surface area

corresponding to 65 nm particles is 48.5 m2/ g.

3.3.1 Ultimate Analysis & Ash Content

The ash content of this carbon black was reported to be 0.3 wt%; this
was veriﬁed by an independent ultimate analysis test carried out by Commercial
Testing & Engineering Company. The elemental analysis of the ash is

presented in Table 3.1.

An ultimate analysis revealed that this particular carbon black is 97%

carbon; the remainder is impurities such as moisture, sulfur, etc., as can be seen

57

in Table 3.2.

Table 3.1 Elemental analysis of ash in Raven 16

 

 

Element Concentration (ppm)
Fe 450

Ca 250

Na 200

Mg 30

Al 30

Ti 10

K, Si remainder

 

 

 

 

Table 3.2 Ultimate analysis of Raven 16

 

 

Impurity wt%
Moisture 0.46
Carbon 96.92
Hydrogen 0.27
Nitrogen 0.29
Sulfur 1.54
Ash 0.31
Oxygen (difference) 0.21

 

 

 

 

Sulfur constitutes almost half of the impurities in the carbon and is
present as S bonded to carbon black (free S<0.03 wt%). Volatile loss upon

heating between 120 ° C and 950 ° C in inert atmosphere is 0.9 wt%.

3.3.2 Impregnation with K2003

 

Carbon black samples were impregnated with KQCOa solutions to
potassium carbonate loadings of 1, 5 and 10 wt%-K2003. These samples were
prepared by ﬁrst diluting 0.80 ml, 4.20 ml and 9.00 ml of a 1.799 M potassium

carbonate solution to 90 ml with equal volumes of doubly distilled water and

58

acetone and then adding 20.0 g of carbon black to each solution. The solid-
liquid mixtures were mixed thoroughly, covered and left overnight (14 h), then
were heated in an oven at 80 ° C with occasional stirring until dry. The samples
were then heated at 120° C for 24 h, ground in a mortar and pestle and stored
in 20 oz french square glass bottles with screw caps. The actual potassium
loadings of these samples were 1.02 wt%, 4.96 wt% and 9.99 wt%-K2003 which
compares rather well with the NAA measurements of 1.1 wt%, 5.1 wt% and 10.0
“#:7049003.

A blank sample was also prepared. This sample was treated similar to
the impregnated samples except for the fact that no potassium carbonate

solution was added.

For some later experiments some carbon black samples underwent
treatments such as hydrogasiﬁcation, degassing and combustion before
impregnation. For impregnating these samples smaller proportions of acetone
were tried and it was found that volumes of acetone as little as 1/3 of the
volume of water used is enough for wetting hard-to-wet carbon surfaces like
degassed carbon. For samples which were oxidized in oxygen the carbon surface
could be readily wetted and no acetone was needed. This reduced the drying
time period considerably. Furthermore, it was found that depending on the
treatment prior to impregnation less water could be used to make a moderately

thick slurry which could be dried much faster.

3.3.3 BET Surface Area

 

The surface area of this carbon black, fresh or gasiﬁed up to 70%
conversion, was measured using the same technique that was used for coconut
charcoal and graphite. The surface areas of the 10 wt%-K2003 samples at

various levels of gasiﬁcation at 865°C in 500 psi hydrogen gas were also

59

calculated after washing the samples with doubly distilled water thoroughly to
remove residual catalyst, which normally blocks some of the pores making them
inaccessible to N2 gas. The adsorption partial pressure of nitrogen gas in all

cases was less than 200 Torr. The results are presented in Figure 3.1.

It can clearly be seen from Figure 3.1 that the surface area increases
linearly from 20 m2/g to 400 m2/ g at 60% conversion. This increase is much
higher than what is expected from a simple reduction in the diameters of the
carbon particles (geometric surface area of 65 nm particles after 50% conversion
is only 61.1 m2/g), and development of a high degree of porosity is the only
plausible explanation. This phenomenon has already been observed for

combustion of carbon blacks [85,86].

BET analysis of N2 adsorption isotherms has been used as a fairly
accurate means of measuring the total surface area of both porous and non-
porous carbon blacks [87-91]. Application of t-method to porous carbon blacks
with substantial volumes of pores with diameters from 4 A to 15 A has shown
that the classical BET analysis of N2 adsorption data is within 15% of the
results from the t-method [92]. Further studies of porous furnace blacks by t-

method have showed much better agreement [93,94].

However, it has been suggested that the classical BET surface area may
be incorrect due to capillary condensation of nitrogen gas in the micropores [95].
But, the application of a modiﬁed BET method to low temperature adsorption
of nitrogen gas on coconut charcoal has shown no sign of capillary condensation
for adsorption pressures of less than 150 Torr [96]. Furthermore, the application
of Polanyi-Dubinin Potential Energy equation to C02 adsorption isotherms at
-78°C on coconut charcoal and some carbon blacks also has shown no sign of
capillary condensation of C02 in the micropores, and the surface area thus

calculated agreed very well with the BET area of C02 adsorption [97]. Hence,

60

Dow

 

3:228:50 got?» as 2 53mm “0 dose. 835m Hmmhz H.” oazwmh

E zo_mmm>zoo
on on as om

noowxlxt. o...
concoo 580T

     

O
I

O

 

lllLilLlLLILLILLLJILLILL

 

00.

com

con

0
0
d’

O
O
I!)

coo

(IS/aw) vasv aovaans

61

the BET surface area from C02 adsorption isotherms seems to be a more reliable

means of measuring the surface area.

The BET surface area of a 72% hydrogasiﬁed sample was measured, after
washing off all the residual catalyst, from adsorption isotherms of both low
temperature N2 adsorption and 002 adsorption at -78°C. The surface areas
were essentially the same, suggesting that there was no capillary condensation of
nitrogen in the pores that were developed during gasiﬁcation. Therefore, the
measured N2 BET surface areas presented in Figure 3.1 are fairly accurate at all
conversions, and, in addition, indicate that loading carbon with potassium
carbonate does not change the course of development of pores in a carbon

sample during hydrogasiﬁcation.

3.3.4 Reactivity

 

As expected, this carbon lampblack was more reactive than the graphite
powder because of lower crystallinity and higher surface area, and still less
reactive than the coconut charcoal due to its much lower surface area. The
hydrogasiﬁcation rates of this carbon black, catalyzed and uncatalyzed, were
“ideal” (higher than 60% conversion in 8 h) and could be measured with
reasonable accuracy (conversions obtained from graphical analysis were within
10% of the weight measurements). Furthermore, steady state conditions could
be reached at less than 10% conversion, as opposed to more than 50%
conversion for the charcoal gasiﬁcation. Therefore, it was decided to use this
carbon black for further studies of KQCO3—catalysis of hydrogasiﬁcation, and

thus this carbon was further characterized.

62

3.3.5 )CRD Analysis

 

The X-ray diﬂraction analysis of this carbon black was conducted by Mr.
Xiao—Wen Qian at the Department of Physics at Michigan State University,
using the analysis discussed in detail by Heckman [98]. The average interlayer
spacing of the layers was found to be 3.54 A which is larger than 3.354 A
interlayer spacing of ideal graphite layers. The average diameter La of the
carbon layers and the average distance L, between the top layer and the bottom

layer of a crystallite were calculated to be 37 A and 30 A, respectively.

The diffraction pattern and crystallite parameters indicated that there is a
deﬁnite structure and order in this carbon black, and thus this carbon is not
amorphous. However, the size of the crystallites are much smaller than the
crystal sizes of graphite (on the order of um) and from the interlayer spacings it
can be inferred that the stacking of the layers and their relative orientations in

this carbon black are completely random [99].

3.3.6 Microstructure of Carbon Black

The term microstructure refers to the internal (paracrystalline)
organization of the carbon within the particles of the carbon black. By the late
1940’s two different models had been proposed for microstructure of carbon
blacks: i ) ‘random crystallite orientation model’ based on X-ray results which
assume that crystallite (parallel layer planes of carbon) orientation is random
within particles and that dimensions of all crystallites are similar to the average
values found by X-ray technique [100,101]; and ii ) ‘tangential layer plane
orientation model’ based primarily on dark-ﬁeld electron microscope studies
[102,103]. The major shortcoming of the latter model is that it does not
distinguish clearly between crystallites and single layer planes and makes no

allowance for distribution in crystallite dimensions.

63

In 1960’s, with the advancement of high resolution electron microscopy,
some new models were suggested. Donnet & Bouland [104] suggested a model
consisting of uniform concentric crystallites throughout a carbon black particle.
However, this model could not explain the mode of oxidation of carbon blacks;
namely, combustion of thermal and furnace blacks were from inside out [105].
Based on this observation and further X—ray and electron microscope studies of
oxidized carbon blacks, Heckman & Harling [105] proposed ‘concentric crystallite
model’ in which outer crystallites of a carbon black particle are larger, more
perfect and their layer planes are generally oriented parallel to the particle
surface. In contrast, the central portion is made up primarily of small, imperfect
crystallites, single layer planes and possibly disorganized carbon (carbon not
incorporated into layer planes). Hess et al. [106] supported this model with their
conventional and diffracted beam electron microscopic results. Donnet et al.
[107] improved this model to incorporate their speculated closed porosity within
the particles. Heidenreich et al. [108] proposed a similar model on the basis of
results obtained using the lattice fringe imaging techniques which give a more

detailed picture of the organized parts of the carbon.

However, independent X—ray studies of carbon blacks by Ergun [109] led
to the conclusion that the fundamental microstructural unit of carbon black is
the graphitic layer, not the crystallite. Ergun gave convincing evidence (from
atomic radial distribution functions) that the crystallites of carbon blacks as
derived from X—ray analysis are not true microcrystals, but are in reality a
chance parallel alignment of successive carbon layers. He also concluded that the
crystallite layer size La , derived from the width of the X—ray (hk) reﬂection, is
less than half that of the distorted graphitic layer, and that layer stackings may
be more extensive than the proﬁles of (001) reﬂection indicate. Ergun’s

paracrystalline model was composed of large anisotropically distorted layers

64

oriented in parallel stacks having a greater interlayer spacing than graphite, due
to interstitials and gross distortions. From this model it is expected to observe
crystallites with a variation or distribution in the interlayer spacing under
electron microscope, rather than a single interplanar distance. This has been

shown to be indeed the case [110].

Ban & Hess [111] obtained the ﬁrst direct images (high-resolution phase-
contrast transmission electron microscopy) of the defective and distorted, but
generally concentric, graphitic layers in normal carbon blacks. Based on the
above observations, Harling & Heckman [112] corrected their earlier model [105]
to ‘concentric layer orientation model’ which incorporated many of the major
aspects of the earlier paracrystalline and concentric crystallite models. According
to this model, each carbon layer has one or more planar portions which when
stacked roughly parallel and equidistant to adjacent planar domains will
function as a crystallite plane in its ability to diffract X-rays. The central
portion is composed of the smallest layers with the greatest curvature and,
therefore, with a more open structure. This model can best explain the up-to-
date experimental observations such as the combustion of furnace carbon blacks

and is still current.

Hence, it seems that the calculated crystallite dimensions presented in the
XRD Analysis section cannot provide any deﬁnitive information about the
degree of microcrystallinity or order of the carbon black particles. However,
they do indicate that the high temperature treatment (1000 ° C) does not induce
any detectable changes in the microstructure of the carbon particles, otherwise
the interlayer spacing would decrease and would have a narrower distribution
[110], or would be reﬂected in the crystallite dimensions [109]. These effects were
not observed; the diffraction pattern of the fresh carbon and the heat treated

carbon were identical.

65

3.3.7 XPS Analjsis

 

The surface of the carbon black was examined by XPS, which chemically
characterizes the outermost 10 A of the surface [113]. The surfaces of carbon
black samples that were degassed at 1000° C under vacuum for 15 h or
hydrogasiﬁed to 47% conversion at 865 ° C were also analyzed by XPS. It can be
seen from the XPS spectra of these three samples (Figure 3.2) that there is no
trace of nitrogen on the surfaces of these carbons (N18 binding energy of nitrogen
is 402 eV [114]) in spite of the fact that 0.29 wt% of the carbon is nitrogen.
Sulfur, which constitutes half of the impurities, could be identiﬁed only on the
surface of the fresh carbon, implying that bonds between sulfur and the carbon

are weak and are readily broken upon heating to high temperatures.

In contrast, oxygen is present on the surface of all three carbons despite
the fact that it is present in the bulk of carbon in amounts much less than
sulfur. However, carbon is known to adsorb oxygen even at room temperature,
and since the samples were exposed to air for a few days before they were
analyzed by XPS, the measured concentrations of oxygen on the surfaces of these
carbon are probably higher than those under experimental conditions. The
XPS-measured surface compositions of sulfur, oxygen and carbon are presented

in Table 3.3.

3.3.8 Oxygen Surface Complexes

 

As discussed in Chapter 1, the oxygen present on the surface of carbon
may be in the form of acidic, basic or neutral complexes depending upon the
conditions under which they are formed. The inﬂuence of these acidic and basic
groups on hydrogasiﬁcation rates as well as the catalysis by K2003, discussed in
Chapters 4 and 5, is the cornerstone of the studies presented here. Thus, the

formation and the nature of oxygen surface groups need to be characterized ﬁrst.

66

 

aooo .
‘n

Fresh carbon

v

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O
E
3
o
U
'5:
Wm Vi‘t‘vv A A V r 1 P: .v
- - - - ‘7» ’2.
’°°°-° ”new. (Mfg, levl
0.0
. c. - . V - - .7 -
. . c
or i
4 m gas1f1ed carbon ‘
" <
.
U
C
3
O
U
* v4
°|s
1
w;v -;t "A r 4":7 -‘wA-v ' " ‘1
me e Imam. My (an 0.0
10003 I ' 1 ' . -
, 4
‘1:
Deqassed carbon
b ‘ 1
.3
C
3
O
U
h 4
AM
1000.0 landing chore! 10V) 0.0

Figure 3.2 ESCA scan of fresh and heat treated Raven 16.

67

Table 3.3 Surface concentration (atomic %) of non-metallic elements in Raven 16

 

 

Element Fresh carbon 47%-gasiﬂed carbon Degassed carbon
C 94.52 94.11 95.75
0 4.63 5.89 4.26
S 0.85 0.00 0.00
Others 0.00 0.00 0.00

 

 

 

 

 

 

Acidic or basic surface groups were introduced onto the surface of the
carbon via combustion in oxygen at 400° C and 800° C, respectively. However,
no attempt was made to chemically analyze and identify individual oxygen-
containing surface groups (initially present in the fresh carbon or introduced via
combustion) at different stages of the experiments. Instead, pH measurements of
suspensions of carbon samples were used as a gross and semi-quantitative means
of identifying acidic or basic character. Furthermore, within the pH range of
solutions of these experiments, the pH is expected to be related to the oxygen
group concentrations, and thus ‘relative’ changes of the concentrations can be
estimated from changes in pH. It has been shown that the pH values of these
suspensions closely follow the acidic or basic nature of the samples in a
qualitative way [115]; i.e., basic surface groups give rise to a basic pH value in

the suspension and acidic groups give rise to an acidic pH value.

3. 3. 8.1 Procedure

 

3.3.8.1.1 Combustion

Combustion of carbon samples (oxygen chemisorption) was carried out in
the reactor used for hydrogasiﬁcation studies (described in detail in Chapter 2).
Oxygen chemisorption, either at 400°C or at 800°C, was carried out by ﬁrst

raising the temperature to the desired chemisorption temperature under vacuum

68

(~ 10 qu) followed by addition of about 0.33 cc (STP) 02 gas per milligram of
carbon sample. The reactor was isolated (two hours at 800 ° C or 12—15 hours at

400 ° C), then evacuated, and the next step was carried out.

3.3.8.1.2 pH Measurement

pH measurements were made in parallel 25 ml Erlenmeyer ﬂasks to ensure
that changes in pH were solely due to the presence of carbon. Carbon samples of
18—24 mg were added to the same ﬂask every time and 10 cc of a 0.1 M KCl
solution was added to each ﬂask with a pipette. Potassium chloride solution
was used instead of pure distilled water because oxygen groups on carbon
surfaces produce a bigger change in the pH of KCl solution than of distilled
water [115]. The ﬂasks were then attached to water-cooled condensers and the
solutions were boiled for 2-3 hours on a hot plate. After boiling the ﬂasks were
sealed with glass stoppers and cooled to room temperature. The solutions were
then transferred to glass bottles and the pH of each was measured by a Fisher
Accumet Model 825 MP pH-meter equipped with a n-Metrohm glass pH

electrode or a Cole Parmer epoxy-body pH electrode.

3. 3. 8.2 Results

 

The results of pH measurements of various carbon samples in 0.1 M KCl
solution are tabulated in Tables 3.4-3.8. The change in pH of the KC] solution,
ApH, is a measure of the acidity/basicity of the oxygen complexes on the carbon
surface; a positive ApH, for instance, demonstrates the presence of basic groups
because the pH of the K01 solution is increased in the presence of carbon. For a
few carbon samples, however, the oxygen surface groups could not be
characterized from the pH measurements because of either the hydrophobic

nature of the carbon surface or the presence of potassium on the surface. The

69

Table 3.4 ApH of 0.1 M KCl solution for fresh, degassed and gasiﬂed carbons

 

 

 

Treated Sample ApH
Fresh carbon did not wet
Degassed carbon did not wet
Gasiﬁed 24% in H2 +2.9
Gasiﬁed 24% in H2 +2.6
Degassed and gasiﬁed 32% in H2 +2.6
Heated at 400 ° C in vacuo for 2 h did not wet
Heated at 800° C in vacuo for 2 h +0.5
Heated at 800 ° C in helium for 2 h +0.6

 

 

 

 

Table 3.5 ApH of 0.1 M KCl solution for carbons oxidised at 400 ° C

 

 

Treated Sample ApH
Fresh carbon oxidized for 2 h -0.75
Fresh carbon oxidized for 15 h -2.1
Gasiﬁed 42% in H2, then oxidized for 2 h -0.4

 

 

 

 

Table 3.6 ApH of 0.1 M KCl solution for carbons oxidised at 800 ° C

 

 

Treated Sample ApH
Fresh carbon oxidized for 2 h +1.6
Fresh carbon oxidized for 2 h +1.44
Degassed carbon oxidized for 2 h +0.3
Gasiﬁed 22% in H2 +2.7
Gasiﬁed 22% in H2 +2.5
Gasiﬁed 43% in H2, then oxidized for 15 min +2.4
Gasified 43% in H2, then oxidized for 2.5 h +2.7

 

 

 

 

surface of a fresh untreated carbon or a degassed carbon sample did not wet for
reasons that will be discussed later. Addition of a small amount of acetone to
the solution resolved this problem, but it was observed that acetone inﬂuenced
the pH value of any carbon suspension, although these changes were small. The

pH values of the suspensions of these hard-to-wet samples were virtually the

70

same as that of the KCl solution. Potassium, whether in carbonate or reduced
metallic form, yields a basic solution and if present on the surface of a carbon
increases the pH value of its suspension drastically, thus making it impossible to
characterize surface groups. Therefore, the only meaningful pH measurements
were those of samples that contained no catalyst and were wet without acetone;

i. e., oxidized and/or partially gasiﬁed pure carbon samples.

Examination of the ApH values in Tables 3.4 and 3.6 also shows that the
pH measurements are fairly reproducible. The ApH of duplicate suspensions of a
24% gasiﬁed sample in Table 3.4 and those in Table 3.6 indicate that the ApH

measurements agree within 0.3 pH units.

Tables 3.4, 3.7 and 3.8 show that a partially gasiﬁed carbon black,
irrespective of the treatment preceding the gasiﬁcation, contained basic surface
groups, provided that gasiﬁcation proceeded to the extent that carbon surface
inﬂuenced by treatments had been gasiﬁed away (> 10—20 wt%). This
observation is further illustrated in Figure 3.3; different points represent carbon
samples that have been treated diﬂerently (degassed, oxidized in oxygen or
hydrogasiﬁed individually or in combination) and then all have been
hydrogasiﬁed for at least an additional 10% conversion. These carbon samples

all exhibit a basic surface character.

The surface of a carbon sample oxidized at 400° C had an acidic
character, as can be clearly seen from Table 3.5. Table 3.6 shows that when
oxidation took place at 800° C, basic groups were formed on the surface of the
carbon. It can also be seen that time periods as low as 15 min were enough to
saturate the surface with oxygen groups at 800° C, whereas 15 h was necessary
at 400° C. Finally, an interesting result is the observation that a fresh carbon
wetted easily once it was heated to 800°C in vacuo or in helium. Similar

treatment at 400 or 1000 ° C did not give the same result (Table 3.4).

 

 

(1L

71

.9 52mm 5:53 E 888388 commxo ho 25.82 n6 ouswmrm

o5 zo_mmm>zoo
8 on om on 9. on cm 2 o

.dqdddqdld-dddd1dddddud‘ddddde-Edd‘d‘d.dddqdddlﬂd-dd‘dwdddddqlldd‘qu-ddddddldﬂ

 

lmlllllmllllllj

hill

an.

 

IIJIJAIII

QIALAJI

 

 

 

0.0

o...

o.N

05

0+

Noumos ISM w £0 10 Hdv

72

Table 3.7 ApH of KCl solution after combustion at 400 ° C and hydrogasifying

 

 

Treated Sample ApH
Oxidized then gasiﬁed 8% in H2 +2.7
Oxidized then gasiﬁed 11% in H2 +3.2
Oxidized then gasiﬁed 31% in H2 +2.9
Oxidized then gasiﬁed 38% in H2 +2.5
First gasiﬁed 43% in H2, then oxidized and then gasiﬁed additional 17% in H2 +2.4

 

 

 

 

Table 3.8 ApH of KCl solution after combustion at 800 ° C and hydrogasifying
Treated Sample ApH

 

 

First oxidized then gasiﬁed 30% in H2 +3.3
First gasiﬁed 22% in H2, then oxidized and then gasiﬁed additional 22% in H2 +2.9
First gasiﬁed 43% in H2, then oxidized and then gasiﬁed additional 16% in H2 +2.7

 

 

 

 

i 3. 8. 3 Discussion

Examination of Tables 3.5 and 3.6 reveals that after oxidation of a
carbon sample at 400 or 800 ° C, acidic or basic groups are formed on the surface
of carbon consistent with literature results. The important result is the
observation that carbon surfaces show basic character after being gasiﬁed in pure
hydrogen for more than 10-20% conversion, irrespective of their treatment prior
to gasiﬁcation ( Tables 3.4, 3.7 and 3.8). The presence of nitrogen, sulfur,
hydrogen, or ash in the carbon cannot explain this basic character, because the
XPS spectrum of a degassed or a 47% gasiﬁed sample (Figure 3.2) did not show
any trace of the elements on the surface, and the ash content of 0.3 wt% is too

10‘” t:0 account for such a high pH. Thus, it can be concluded that the oxygen
surface groups are solely responsible for basic and hydrophilic character of the

surface.

73

The results presented in Table 3.4 show hydrophobicity of the surface of
fresh carbon sample, in spite of a surface coverage by oxygen groups of nearly
13% as measured by XPS. This hydrophobicity was also observed for a carbon
sample degassed at 1000°C as well as for a fresh carbon heated in vacuo at
400 ° C, but not for a sample heated in vacuo at 800 ° C. This apparent anomaly
is explained by two different eﬂects. It may be that a small amount of residual
hydrocarbon material (from which this particular carbon was produced) was
adsorbed on the surface, making the XPS-detected oxygen surface groups present
on the fresh carbon inaccessible and thus the surface hydrophobic even at
400 ° C. These organic deposits were desorbed or decomposed at some
temperature between 400 and 800° C, thus exposing oxygen groups and making
the surface hydrophilic. Further evidence for this argument was a yellow ﬁlm
deposit on the inner quartz lining of the gasiﬁcation reactor which was observed
upon heating carbon samples to high temperatures. This deposit was soluble in
acetone but not in water. Once fresh carbon was wash with acetone and dried,
the surface could be wetted in water easily. The prolong heating at 1000° C for
degassing was apparently severe enough to remove oxygen groups from the
surface, hence the lower surface oxygen coverage and the hydrophobic surface of

the degassed carbon.

The GC analysis of the gas collected continuously during degassing at
1000° C of a fresh carbon sample indicated the release of large amounts of CO
from thermal decomposition of oxygen groups of the carbon sample, 4.94 cc CO
and 0.207 cc 002 per gram of carbon. This ratio of 00/002 is virtually the
equilibrium value at 1000°C, and consequently it is not possible to deduce
whether the oxygen groups decomposed to CO or C02. The amount of oxygen
in the evolved gas could cover about 78% of the total surface area of the initial

fresh carbon if each atom of oxygen was assumed to cover 8.3 A2. But, the

74

XPS-measured oxygen surface coverage of fresh carbon was about 13%, implying
that oxygen was released from groups both in the bulk carbon near the surface
and from the surface. Apparently, the time it took for the oxygen-containing
gases to diffuse out of the carbon particles and carbon bed into the gas stream

was long enough to establish equilibrium.

Heating a fresh carbon sample at 865° C in helium for 3 hours produced
1.21 cc CO and 0.046 cc C02. This amount of oxygen desorbed corresponds to a
hypothetical coverage of 19% of the surface area as opposed to the XPS-
measured oxygen coverage of 13% of the surface of fresh carbon, implying that
most of the oxygen was from the surface. In this case, the 00/002 ratio is
higher than the equilibrium value and so it follows that the oxygen groups near

the surface of the fresh carbon were mainly non-acidic.

However, the pH measurements of a fresh carbon sample heated at 800 ° C
(Table 3.4) indicates slight basicity of the surface (ApH = +0.5 - +0.6), and
thus it can be concluded that the oxygen groups near the surface of the original
fresh carbon are mostly C=O carbonyl groups. Furthermore, the pH
measurements of the degassed samples after about 32% gasiﬁcation (Table 3.4)
indicate the presence of much higher concentration of basic surface oxides (ApH
= +2.6), possibly chromene structures, which are believed to be responsible for
the low reactivity of this sample after degassing. The pH measurements of other
carbon samples that were not degassed yielded similar pH values after
gasiﬁcation (Table 3.4), suggesting that neither degassing nor gasiﬁcation
decomposed the stable basic groups within the bulk of carbon black. This is
consistent with the results obtained by Blackwood [40], who attributed the low
reactivity of high-temperature treated char samples to the stable chromene
structures. The fact that ApH of the suspension of a carbon sample after 32%

gasiﬁcation is 2 units higher than that of a fresh sample implies that the

 

HE”

bu

75

concentration of basic oxygen surface groups is on the order of 100 times higher
in the former case. However, the BET surface area of the 32% gasiﬁed carbon is
8—9 times that of the fresh carbon (Figure 3.1) and the XPS—coverage of surface
by oxygen is 1.1-1.3 times that of the fresh carbon (Table 3.3), so the net overall
coverage of the surface of 32% gasiﬁed carbon by oxygen is only 10 times that of
the fresh carbon. It thus seems that 10 times more basic oxygen groups are

inside the fresh carbon particles than are on their surface.

In summary, it thus can be concluded that oxygen groups initially present
near the surface of the fresh carbon are mainly carbonyl groups and those in the
bulk carbon are dominantly basic in character. Furthermore, combustion of the
carbon in oxygen at 400 ° C and 800 ° C establishes acidic and basic complexes on
the carbon surface, respectively, as expected. The signiﬁcance of these results
will be discussed in Chapters 4 and 5 in relation to the activity of potassium

carbonate catalyst and the reactivity of carbon toward hydrogen.

Chapter 4

Uncatalyzed Hydrogasiﬁcation

The ultimate objective is to study the inﬂuence of addition of potassium
carbonate and oxygen surface complexes on kinetics of hydrogasiﬁcation of
carbons. Thus, the hydrogasiﬁcation kinetics of carbons in the absence and in
the presence of potassium carbonate and oxygen surface complexes must be
Compared and any observed differences be carefully examined. For instance, if
the addition of some oxygen complex on surface of a carbon sample enhances the
hYdrogasiﬁcation rate of that carbon, that oxygen surface complex would be
deﬁnitely involved directly or indirectly in the sequence of events taking place on
the surface of carbon during hydrogasiﬁcation. On the other hand, if addition of
Potassium carbonate does not result in an enhancement of the hydrogasiﬁcation

rate, the carbonate most likely will not be involved in a catalysis process, at

76

77

least in the mode or the way it is added to the carbon.

However, this task necessitates that the measured hydrogasiﬁcation rates
be the true intrinsic reaction rates; i. e., the reaction rates should not be masked
by mass transfer limitations within the carbon ﬁxed bed as well as the pores of
carbon particles that develop during hydrogasiﬁcation (see Section 3.3). If there
is a substantial mass transfer resistance within the carbon ﬁxed bed, the rate of
mass transfer of product gas(es) from within the ﬁxed bed into the bulk gas
stream would be much less than the hydrogasiﬁcation rate deep inside the bed
under the bulk gas conditions and, therefore, the partial pressures of gases at
carbon-gas interface would be different from those in the bulk gas. Thus, the
reaction environment at carbon-gas interface would be different from the bulk
gas monitored conditions and, consequently, any correlation of the bulk gas
reaction conditions with the observed reaction rates would be meaningless. In
addition, this would also mask the true enhancement caused by the catalyst and
make the apparent activation energy Ea’ much lower than the true activation

energy E, depending on the extent of mass transfer resistance.

There are some other factors such as temperature variation inside the
carbon ﬁxed bed that need to be considered before comparing the measured
reaction rates. The temperature of the carbon ﬁxed bed could vary within the
bed or in different experiments because the carbon sample size varies or the
reactor vessel temperature may not be uniform throughout the furnace, and the
location of the sample holder may not be exactly the same location for each
experiment. The hydrogasiﬁcation rate is strongly dependent on reaction
temperature as is evident in a reported value of 63 kcal/g-mol [116] for the
activation energy Ea of hydrogasiﬁcation of Raven 16. Thus, if the temperature
of the carbon ﬁxed bed varies signiﬁcantly within the bed, the reaction rate

would also vary and might result in different overall reaction rates.

78

Hence, the mass transfer limitations and temperature variation within
carbon ﬁxed bed are discussed ﬁrst and then the inﬂuence of oxygen surface
groups are presented in this chapter. The inﬂuence of addition of potassium
carbonate with or without the addition of oxygen surface complexes are
presented in Chapter 5. However, a brief review of the literature on
hydrogasiﬁcation of carbons, in the absence of any external catalysts, seems

appropriate at this point.

4.1 Background

4.1.1 Hydrogasiﬁcation Products

 

Reaction of hydrogen gas with carbon produces compounds that contain
hydrogen and carbon; i. e., hydrocarbons. At temperatures above 800°C the
heavy hydrocarbons (C4 hydrocarbons or higher) are thermally unstable and
their formation is thus thermodynamically unfavorable. The 02 hydrocarbons
such as ethylene and acetylene are thermodynamically much more favorable than
higher hydrocarbons. Nevertheless, methane formation is by far the most

thermodynamically favorable reaction at temperatures below 1000 ° C [117].

But, the question of how much of these hydrocarbons are actually formed
under a given set of conditions is also a kinetic question; e.g., if the formation of
methane is kinetically very slow and that of ethylene is very fast, the only
reaction product would be small amounts of ethylene. However, it has already
been shown that at temperatures below 1200°C methane formation is also
kinetically favorable; i.e., methane is by far the main product detected [118-122]
with very small amounts of ethane [119]. In contrast, formation of small

amounts of ethane and ethylene have also been reported in the temperature

 

6
CV

79

range of 600 ° -880 ° C over thin ﬁlms of pyrolytic carbon [123].

The hydrogen pressures in these studies have been sub-atmospheric and
since the kinetics depends on the the partial pressures of reactant gases, a
substantial change in the partial pressures might alter the kinetics and thus
change the composition of the product gas mixture. However, for hydrogen
pressures of up to 40 atm and over a temperature range of 650 ° -870 ° C methane
has been found to be by far the main product of hydrogasiﬁcation of chars
[124]. Therefore, the relative kinetics of formation of various hydrocarbons are

not inﬂuenced signiﬁcantly by changes in the hydrogen partial pressure.

Hence, it can be concluded that for carbon blacks, which have a graphitic
structure similar to chars, methane would be by far the main product of
hydrogasiﬁcation below 900 ° C and hydrogen partial pressures of up to 40 atm.
Therefore, the hydrogasiﬁcation rate of carbon and the rate of formation of
methane can be taken to be equivalent and, thus, the problem of studying the
kinetics of hydrogasiﬁcation is reduced to just studying the kinetics of methane

formation.

4.1.2 Thermodynamic Eﬂuilibrium

The thermodynamic equilibrium of a chemical reaction determines the
maximum amount of the products obtainable under the given set of conditions.
That is, no matter how long a reaction is allowed to proceed, the amount of
product(s) cannot exceed a certain limit which is prescribed by the
thermodynamic equilibrium of that reaction. Starting a chemical reaction with
no product present, if at any time there appears to be more product than what
is predicted from the thermodynamic equilibrium, either the equilibrium
calculations are erroneous or some other reactions are taking place as well under

the given set of condition. In contrast, if for a given chemical reaction the

80

product concentrations (partial pressures) are less than the equilibrium values,
that reaction cannot be ruled out as a formulation of chemical changes because
very slow kinetics could be responsible for low concentrations. Therefore,
thermodynamic equilibrium calculations cannot explain or account for all the
events taking place during a chemical reaction; instead, they provide some

criteria on some physical limitations of what could take place.

The thermodynamic equilibrium of carbon-hydrogen—methane has been
investigated experimentally [115—117] and theoretically [127,128] over the wide
temperature range of 300°-2100°C. The results have shown that the

hydrogasiﬁcation of carbon

C + 2H2 —'> CH4 AHQS '0 = “‘18 kCal/ngZC (4‘1)

is an exothermic reaction and, thus, the equilibrium concentration of methane
decreases as the reaction temperature is raised (at 870 ° C heat of reaction AH is

-22 kcal / g mol). For this reaction the equilibrium constant KP is

 

PCH
KP = 2 ‘ (4-2)
ﬂy,

in accordance with the law of mass action, if it is assumed that methane and
hydrogen are ideal gases and thus their fugacities are equal to their partial
pressures. This assumption is valid because of the nonpolar nature of both
hydrogen and methane and the high reaction temperature, and, furthermore, the
high pressure of hydrogen (up to 200 atm) has been shown to have no effect on

the value of K1, at a given temperature [125]. In the above equation, pay] and

if:

81

pH2 are the partial pressures of methane and hydrogen, respectively. The

equilibrium constant KP is normally evaluated from the heat of reaction AH by

the so called van ’t Hoff equation [129];

d anp = AH
d Tm, RT,§,,

 

(4-3)

where R is the universal gas constant and Tm, is the reaction temperature.

The temperature dependence of AH is calculated from the speciﬁc heat

capacity Cp data of the reaction components at various temperature [129]:

Tn:
AHTrn = A1125 '0' + f ACp dT (4'4)
298 'K
with
A0,, = 214-010,, (4-5)
a

where V,- is the stoichiometric coefficient of the i th component of the reaction
and Cpn’ is its speciﬁc heat capacity. The heat capacity data for methane,
hydrogen and graphite are readily available in the literature [130] and

substitution of these data into the above equations yields

82

7365

 

an, IT," = 28.46 + — 5.42 1nT,,,,+ 0.00181Tm, + 2.65x10-7T,§,, (4-6)

T171!

29440

2
ram

— 1.1x10—10T,2,,+

where Tm, is in Kelvin and KP is in psi.

The fraction xeq of hydrogen converted to methane at equilibrium is
calculated from Equation 4—2 by ﬁrst computing the mole fraction of methane

yo”4 in terms of xeq by a complete mass balance

 

 

YCH, = (4'7)
1+2[1 —1]

Substituting mole fraction times total pressure for the partial pressure of each
gas into Equation 4—2 and solving for xeq (assuming that only hydrogen and
methane exist in the gas phase; i.e. , no other hydrocarbon or inert gases exist in

the reactor) yields;

= 1—’\ / ————I 4-8
Xeq 4pr +1 ( )

This equation clearly shows that the equilibrium concentration of methane

depends on the value of K1) which in turn depends on the kind of carbon used.

At 870° C and 500 psi inlet hydrogen pressure (total reaction pressure p)

xeq and You, are equal to 0.53 and 0.36, respectively. That is, if the

hydrogasiﬁcation reaction is allowed to achieve equilibrium under above

83

conditions, 36% of the reactor effluent gas stream would be methane.

However, these values are based on the calculations for graphite (data for
Cp and AHQS .0 are for graphite), and it has been shown that the value of the
equilibrium constant KP depends on the type of the carbon used [125,126,131].
These studies, nevertheless, have indicated that the equilibrium concentration of
methane is the lowest in the presence of graphite. Thus, as long as the methane
concentration in the reactor effluent gas stream is much below the above
concentration of methane, the hydrogasiﬁcation reaction is far away from
equilibrium and the reaction kinetics can be studied (provided, of course, that

there is not signiﬁcant mass transfer resistance within the carbon ﬁxed bed).

4.1 .3 Kinetics

 

While the thermodynamic equilibrium sets the maximum possible
conversion of the reactants in a chemical reaction under a given set of
conditions, the kinetics of that reaction demonstrate how fast that maximum
possible conversion is reached under those conditions. Kinetic expressions can be
derived either empirically or from a proposed plausible mechanism for that

reaction.

For hydrogasiﬁcation of chars, Blackwood [124] proposed an empirical
ﬁrst order kinetic expression for pressures up to 40 atm, temperatures of 650°-
870 ° C and under the conditions far from the equilibrium (i.e., in the presence of

very small amounts of methane);

R CH, = ’9 CHJ’ H2 (4‘9)

where ICC”4 is the reaction constant and RCH‘ is the rate of formation of

84

methane per unit mass carbon. This expression does not give any indication of
the mechanism, nor does it readily suggest how the equilibrium relationship
would appear in a general rate expression. For instance, at equilibrium where
the reaction rate is zero this rate expression predicts a deﬁnite reaction rate. A
true rate expression should predict the equilibrium condition and illustrate how

the reaction rate varies as the equilibrium is approached.

Several mechanisms [132-135] have been proposed (discussed in Section
4.1.4) and some kinetic expressions have been derived from these mechanisms.

Zielke & Gorin [132] suggested the following expression

2
apH2

-———— 4-10
1 +pr2 ( )

RCH. =

which reduces to Equation 4-9 at high hydrogen pressures (a and b are
constants). This expression, however, does not accommodate the inﬂuence of
methane concentration on the reaction rate. This effect is signiﬁcant for near the
equilibrium conditions, and thus this expression fails to illustrate how the

equilibrium is approached.

Blackwood [133] obtained a kinetic expression which incorporates the
inﬂuence of methane concentration as well and, thus, satisﬁes the equilibrium
requirements and predicts a zero order reaction in hydrogen at high hydrogen
pressures. Further studies by Blackwood [134] and the reinterpretation of
Blackwood’s experimental results by Shaw [135] lead to a slightly different

expression;

_ I“ 119113-19 2PCH,
1 +k3pn,+k4pcn,

 

RCH, (4-11)

85

where kl, k2, k3 and 16,, are constants related to the speciﬁc rate constants of the
elementary reaction steps in the proposed mechanism. The assumption made in
deriving the above expression is that in the proposed mechanism (see reactions
4-15 and 4-16), the concentration of intermediate species reaches a steady state
value. At low methane concentrations and high hydrogen pressures this

expression again reduces to Equation 4-9.

4.1.4 Mechanism

 

Zielke & Gorin [132] have proposed a mechanism based on the
hydrogenation of the exposed edges of the graphitic carbon layers as an example

of a “reasonable” mechanism to explain their experimental observations;

H. ,H Heir/H
C=C + H2 —. 0-0 (4-12)
\ \
C c o C
Her I?)
C-C + H9 —+ CH3 CH3 (4-13)
/ \ / \
o 0
CH3 CH3 ll
C C + 2112 —. C C + 2011,, (4-14)

In this mechanism, on the average, an active site is generated for each one
that is consumed. The kinetic expression 4.10 is derived from this mechanism
by assuming that the intermediate product of the ﬁrst step (Equation 4-12) is at

a steady state and that the last reaction is much faster than the other steps.

86

In contrast, Blackwood [134] and Shaw [135] have suggested the following

two step mechanism:

0 + H, —» (0H,) (4-15)

((3H,) + H, -> CH4 (4-16)

where (CH2) is a surface complex which is formed by the chemisorption of
hydrogen. If the surface concentration of this (CH2) intermediate species is

assumed to establish a steady state value, the kinetic expression 4.11 would be

obtained.

None of these mechanisms, however, consider the possibility of the
formation of products intermediate (for instance, C2 hydrocarbons) to the
formation of methane. Cao & Back [123] have modiﬁed the mechanism
proposed by Zielke & Gorin to allow the intermediate formation of ethylene and

ethane:

(CH) + H, —» (0H,) + H (417)
(CH2) + H, -» (CH3) + H (4-18)
(CH2) + (CH2) _* (C2H4) "" C2H4 (4'19)

(CH3) + H, —> CH4 + H (420)

87

(CH3) + (CH3) _’ (C2H6) —’ C2Hs (4'21)

According to this scheme ethylene, ethane and methane are all primary products
distribution of which depends on the competition between abstraction from
hydrogen or dimerization of the intermediate radicals which have a degree of
mobility on the carbon surface. Their experimental results led to the conclusion
that at high temperatures ethane and ethylene undergo secondary reactions in

the gas phase to form free radicals which eventually convert to methane.

4.2 Intrinsic Nature of Reaction Rates

The samples holders (Section 2.10) are mounted at the end of the exit
tube positioned at the center of the reactor vessel and, thus, the carbon samples
in these boats could experience different temperatures because they are positioned
at different locations inside the reactor. The effect of this temperature variation
and the depth of the carbon bed were examined to ascertain that the rate
measurements are free from any masking effects due to temperature variations

and mass transfer resistances.

4.2.1 Mass Transfer Resistances

The magnitude of the mass transfer resistances during hydrogasiﬁcation
of carbon black at 865°C and 500 psi hydrogen gas was studies by measuring
the reaction rates of a 75 mg and a 260 mg carbon sample, distributed evenly in
the quartz sample holder. It is evident from Figure 4.1 that the depth of the
carbon bed had no effect on the measured reaction rat%; the rates are essentially

the same. Therefore, for small sample sizes of carbon, 30—80 mg, (which had a

88

.wmma cow was 0 . new as m: Seam Ao somusocvswoassm: mew—30 moocwumwmoa commas: 3.32 sz van—mfg

g zo_mmw>zoo

on on on 0* on ON 0 _. o

 

 

 

 

..:....:1:1....11....:11.21.111.31:51:211....0.0
mod
w
on D m
D n
0 min. 0 1.0.4
CD D
ODD—U DOD
e u 3.0
0 8 m
D .m .
O mom
...v\ Fl Soon non c0900! 0
:0 F\—.' SwaOU U08 COEDO' D u
0.0—

(0 w: ﬁ-ulw/VHo 1141)

3.018 NO LLOVBEI

89

typical bed thickness of ~1/8") mass transfer resistances within the carbon ﬁxed

bed are absent.

Another piece of evidence for the absence of mass transfer limitations is
the magnitude of the activation energy E, (Figure 4.2). E, was evaluated by
assuming a simple rate expression of some order in hydrogen concentration; 6. g.
Equation 4-9. This assumption is justiﬁed by the fact that methane
concentration is very small. The value of the activation energy calculated over a
temperature range of 800 ° -900 ° C was reported to be 63 kcal / g mole which is in
direct contrast to a value of about 30 kcal/g-mole reported by Gardner et al.
[38]. If the reaction rates were limited by external mass transfer (i.e., by
diffusion of reactant gases to the outside surface of the carbon bed), the observed
activation energy would have had a typical value of 1-2 kcal/g-mole [136].
Obviously, this is not the case, because the measured activation energy is one

order of magnitude higher.

In contrast, if diﬂusion inside the carbon bed were signiﬁcant, the
measured activation energy could have a value as low as one-half of the intrinsic
activation energy depending on the severity of the diffusion limitations [137]. In
this case, as the reaction temperature is raised and the reaction rates become
faster the diffusion limitations become more important and thus the magnitude
of apparent activation energy decreases. That is, the Arrhenius plot would not
yield a straight line. Again, this does not appear to be the case, as evident from
Figure’4.2. However, this point is studied further from a theoretical point of

view.

The effectiveness factor for the carbon bed was calculated by ﬁrst

calculating the Thiele modulus d) [138] which is deﬁned as

 

= _ (4-22)

90

To

.mm wmma com as 2 55m .«0 2558:332va ho uoE msmcofaax «4‘ 0.3—mum

A 76340 a
as

ed

Tm

 

 

.l

mi
1

‘

 

co_m._o>coo Rmnl
co_Eo>coo RNNI

.. 06
i 06
.H 0d
1 odw

 

.. odN

 

(0 am t‘>-u!w/"I-Ia It“)

31372:! NOLLOVEU

91

for a heterogeneous reaction with ﬁrst order kinetics such as the

hydrogasiﬁcation reaction (see Section 4.1.3)

ern = krin CA (4'23)

where CA is the molar concentration of reactant A, krin is the rate constant,

Rm, is the reaction rate, and Rh is the hydraulic radius deﬁned as three times
the ratio of volume to external surface area. The measured reaction rates
presented in Figures 4.1 and 4.3 are less than 10 ml CH4/min-g initial carbon, or
8.18X10-4 mole Hz/min—g carbon. The density of the carbon black particles
used is 1.9 g/ ml [139] and the bed porosity of fresh carbon was measured to be
80% and assuming a bed porosity of 90% at higher conversions (due to the
development of porosity, see Section 3.3.3) a value of 0.19 g carbon/ml for the
bulk density of the carbon bed is obtained. Thus the reaction rate Rm, at most
reaches a value of 2.59X10—6 mole Hg/sec-ml. At 865 ° C and 500 psi hydrogen
pressure the concentration of hydrogen gas is 3.64)<10‘4 mole HQ/ml.
Substitution of these values into Equation 4-23 yields a value of 7.12X10’3 sec”1
for krin.

The experimental diffusion coeﬂicient DAB of the binary system of H2-
CH4 at 1 atm has been reported to be 0.625 ch/sec at 00° C [140] and 0.726 A
cm2/sec at 25°C [141]. At 865°C and 500 psi hydrogen pressure DAB is
calculated from the lowest of the above two values to be 0.156 ch/sec using the

following conservative relation [142] for temperature and pressure dependence of

DAB

 

(4-24)

92

.NE E3 000 was 0 . mam as 3a.. comaaommmmwoémn :0 :oSwEwccg can 2 Exam he 8ch 0i oaawmh

95 20.2928

an on ma 0N up 0 F n

0

 

 

r-‘d‘d‘ddlIJI-GGJICJJJIIJ.‘ddl‘l‘dd‘d.dlllldd‘dlld-Iddddldddld.d‘GG‘l‘dj‘Jd‘dJ‘l“

£80.30...

A... .55 35238 0.65..
€235.35 2.5823 "tonal

<0.

 

-0.0

0.N

0.?

0.0

0.0

0.0 —

ON.

0;;

(0 w: Swim/"Ho M)

31378 NOLLOVBH

93

The effective diffusivity D, is related to DAB according to the following

relation [143]

6
D, = in“, (425)

Where 6,, is the porosity or void fraction and T, is the tortuosity factor of the
carbon bed. Since the unit of D, is based on unit cross sectional area of carbon
bed, and not unit cross sectional area of the pores in carbon bed, the porosity E?
which is a measure of how much of a unit cross sectional area of carbon bed is
available for gas diffusion is introduced, and is taken to be 0.9. In addition, the
rate of diffusion into the carbon bed is measured perpendicular to the surface of
carbon bed, whereas the molecules diffuse along the direction of the carbon bed
pores which, in general, are at an angle with the perpendicular direction. As a
consequence, the diffusion rate in the perpendicular direction is lower than that
in the direction of the pores. The tortuosity factor 1', is introduced to account
for the effect of difference in the directions of the pores and the perpendicular
direction. There is no experimental value available for 7,; however, for loose
powders and beds of glass spheres having bed porosities of 0.35 to 0.43
tortuosity values of 1.4 to 1.6 are reported [143]. For a bed of fresh carbon the
value of T, is expected to be even less because the value of 1', is known to
decrease with an increase in porosity. But, the development of porosity during
hydrogasiﬁcation tends to complicate evaluation of a value for 7', since the
carbon particles cannot be treated as spheres any more. The reported values for
r, of various porous catalyst supports with various degrees of porosity are less
than 10 and typical values of 2-6 are recommended [143]. Hence, a value of 10

seems to be a rather conservative estimate for the tortuosity factor 7, of the

94

carbon bed during hydrogasiﬁcation. Substitution of these values into Equation

4-25 yields a value of 0.0140 ch/sec for De.

The hydraulic radius Rh of the carbon bed is calculated by treating the
carbon bed as a half cylinder of 1 / 8" radius. For the carbon bed in ceramic tube

the value of R}, is then 0.748 cm.

Substitution of appropriate values into Equation 4-22 yields a value of
0.178 for the Thiele modulus 45. The corresponding effectiveness factor for a ﬁrst
order reaction, for the case of low concentration of products and a stoichiometric
ratio of products to reactants of 1 to 2, is virtually 1.0 [138], indicating that
diffusion limitations are not dominant. It should be noted that the assumptions
made in calculating d) are rather conservative and the actual value of ¢ is most
likely lower for the value 7', is lower. Therefore, the reaction rates can still be
faster and not be inﬂuenced appreciably by diffusion resistances, implying that

the measured reaction rates are indeed the intrinsic reaction rates.

When large carbon samples (1.5 g) are used in the quartz tube, the
hydraulic radius R, of the carbon bed is 5.40 cm which makes the (1) rise to 1.3.
The corresponding effectiveness factor is about 0.75, indicating that internal
diffusion resistances are not negligible. This result was veriﬁed experimentally;
the conversion versus time behavior of these large samples were identical to those
of small samples in the ceramic sample holders. The temperatures of the large
samples are higher at some parts closer to the reactor wall and thus the reaction
rates are higher (see Section 4.2.2 and Figure 4.1). But the observed overall
reaction rates showed no enhancement, pointing to the fact that there were some

mass transfer resistances present.

The above analysis focused on the mass transfer resistances in the carbon
bed, i.e., the voids among the carbon black particles and not inside the

particles. As discussed earlier (Section 3.3.3), as hydrogasiﬁcation reaction

95

proceeds, some degree of porosity is developed inside the carbon particles and,
thus, the diffusion resistances inside these pores need to be considered. For a
complete and accurate evaluation of these resistances, however, a knowledge of
the pore size distribution is essential. The pore sizes would determine whether

the diffusion is in the domain of Knudsen diffusion or it is a bulk diffusion.

No attempt was made to determine the pore size distribution; instead, the
diffusion resistances are analyzed for Knudsen domain since Knudsen diﬂusion
coefficients are typically one order of magnitude lower than the bulk diffusion
coefficients. This would reveal the highest possible resistances to methane

molecules diffusing out of the individual carbon black particles.

The Knudsen diffusion coefficient is calculated from the following

equations

 

(4-26)

 

(427)

where r, is mean pore radius, S is the total surface area of carbon particles,
Tm, is the temperature of carbon particle, W is the molecular weight of the
diffusing gas, 6, is the porosity of a carbon particle, and p is the bulk density of
a carbon particle. The mean pore radius r, deﬁned in Equation 4-27 is in fact
an empirical factor which is introduced because the internal geometries of the
pores are not known. The carbon black particles have densities of 1.9 g/cm3

[139], and at 50% conversion the measured BET surface area is 350 m2/ g (Figure

3.1) and 6, has a value of 0.5. Substitution of these values into Equation 4-27

96

yields a mean pore radius of 15 A. For a reaction temperature of 865° C, the
Knudsen diffusion coeﬂicient for hydrogen and methane calculated from

Equation 4-26 is 0.0347 ch/sec and 0.0123 ch/sec, respectively.

Carbon black particles have a mean diameter of 700 A (Section 3.3) and
the corresponding hydraulic radius is 350 A. Substitution of the Knudsen
diffusion coefficient for methane, a porosity of 0.5 (at 50% conversion) and a
tortuosity factor of 10 for the pore systems inside the carbon particles into
Equation 4-25 yields an effective diffusivity of 6.15X10"4 cmg/sec. The
corresponding Thiele modulus calculated from Equation 4-22 is 4.0X10"‘3 which
implies that mass transfer resistances during hydrogasiﬁcation inside the carbon

particles are virtually non-existent.

4.2.2 Temperature Variation of Carbon Bed

As discussed earlier, a given carbon sample might have different reactivity
in the ceramic than in the quartz sample holder because of the temperature
variation in the carbon bed. When the furnace temperature is set at 900 ° C,
the thermocouples inserted at the center of the reactor at the inlet of the central
exit tube register temperatures of 865:};2 ° C, indicating a temperature difference
between the reactor wall and the sample holder in the reactor. The heating of
the carbon bed is achieved by radiation from the reactor wall and, thus, the
closer a sample to the wall the higher its temperature. The quartz sample holder
is much longer than the ceramic sample holder, and once inserted inside the
reactor, its end would be much closer to the wall. This implies that the carbon
particles near that end would have much higher reaction rates; for instance, if
the carbon particles at this end were at 880 ° C instead of 865 ° C, their reaction
rates would be 1.44 times higher, if Ea: 63 kcal/g-mole (Figure 4.2). The

average reaction rates, however, would be less than 1.44 higher because not all of

97

the carbon particles are located at that end.

The results presented in Figure 4.3 demonstrate this point clearly; when
the carbon sample in the quartz tube is conﬁned to only 1" of the boat to
resemble the sample conﬁguration in the ceramic boat, the reaction rates are
virtually the same; but, when the sample is distributed evenly in the quartz
boat, the reaction rates are much higher than the samples in the ceramic boat.
The increase in reaction rate is more than 1.44 times, suggesting that the reactor
wall temperature is higher than 880°C. If the reactor wall temperature is
900° C and the temperature of the carbon bed is assumed to vary linearly from
865°C at the end mounted on the exit tube to 900°C at the end which is
almost touching the ﬂat end of the reactor, the average reaction rate would be
expected to be 1.57 times higher. Therefore, the sole temperature diﬂerence in

the carbon bed can account for the observed difference in the reaction rates.

It, thus, can be concluded that the temperature variations inside the
carbon ﬁxed bed could be signiﬁcant and they should be avoided in the

experiments data which are to be compared.

4.3 Hydrogasiﬁcation of Carbon Black

4.3.1 Procedure

 

To study the effect of presence of oxygen complexes on the carbon surface,
samples were treated by degassing and oxygen chemisorption prior to
hydrogasiﬁcation. Degassing and gasiﬁcation as well as oxygen chemisorption
were carried out in sequential combinations in the reactor to avoid the exposure
of the sample to air between steps. Degassing of the carbon samples was carried

out by raising the temperature to 1000 ° C, holding for about 15 hours, and then

98

cooling to the desired temperature, all under vacuum (1-10 11 Hg).

Oxygen chemisorption was carried out at either 400° C or 800° C. The
temperature was ﬁrst raised to the desired chemisorption temperature under
vacuum. About 0.33 cc (STP) 02 gas per milligram of carbon sample was
introduced into the reactor, which was then isolated. After oxidation (two hours
at 800 ° C or 12—15 hours at 400 ° C) the reactor was evacuated and the next step

was carried out.

Gasiﬁcation procedure depended on whether or not it was preceded by
degassing and/or oxygen chemisorption. For a fresh sample of carbon, the
temperature was ﬁrst raised from room temperature to 450 ° C under vacuum in
about 45 minutes. The reactor was then pressurized with hydrogen gas to 500
psi, and the temperature was increased to the steady state value of 865 ° C in an
additional 50 minutes. At the end of gasiﬁcation, the reactor was evacuated and

was either cooled to room temperature or to the temperature of the next step.

When oxygen chemisorption at 400 ° C preceded the gasiﬁcation, the same
gasiﬁcation procedure was adapted. In the case of oxygen chemisorption at
800°C, the reactor was evacuated after chemisorption, hydrogen gas was
introduced, and the temperature was raised to 865°C in 20 min. When
gasiﬁcation was preceded by degassing, the temperature was lowered to 800° C

after degassing and the above procedure was used.

For all hydrogasiﬁcation and oxygen chemisorption experiments Linde
UHP or Airco grade 5 (99.999% pure) hydrogen gas and Airco grade 2.6 oxygen
gas were used. The hydrogen gas (which was further puriﬁed by passage
through a 3A molecular sieve trap, see Section 2.6) ﬂow rate used was 310:1:10

ml/min.

99

4.3 .2 Results

 

A set of raw reaction rate data (reaction rate versus time) for
hydrogasiﬁcation of fresh carbon (as-received carbon black) is presented in
Figure 4.4. The start of experiment (time = 0) was taken as the time hydrogen
ﬂow was begun at 450° C. The reaction rate increased as reaction temperature
was increased, reaching steady steady after 50 minutes. In this particular
experiment and all other experiments, carbon conversion was less than 10%
when the steady state temperature was reached. Furthermore, in all cases the
total conversion of carbon as calculated from integration of the raw reaction rate
data was close to the conversion determined by weighing the residual carbon
samples after experiments, showing that all methane was accounted for. The
experimental results were also fairly reproducible, as some duplicate experiments

in Figures 4.6 and 4.7 indicate.

The manipulated kinetic results (reaction rate versus conversion) of
carbon black, fresh or after degassing, with or without oxygen chemisorption, at
400 ° C and 800 ° C, are presented in Figures 4.5 through 4.8. In all cases, except
for the degassed samples, methane evolution rate decreased at steady state
temperature as carbon was consumed (Figures 4.5-4.7). Degassed carbon black,
on the other hand, showed a relatively constant reaction rate, but exhibited
much less activity than the fresh carbon irrespective of the combustion
treatment (Figures 4.5 and 4.8). In contrast, heating a carbon sample for 3
hours at 865 ° C in helium before gasiﬁcation had no signiﬁcant effect on reaction

rate (Figure 4.5).

The reactivity of carbon black, fresh or degassed, toward hydrogen was
consistently enhanced by oxygen chemisorption prior to gasiﬂcation. Oxidation
at 400°C enhanced the rate of methane evolution from fresh carbon

considerably (by a factor of 2 at low conversions) and from degassed carbon

100

am wmma 00m. was 0 . new as 3 55$ .6 330 8a.. :oSsoEmmwocPE spam «as ouzmmh

EEC m2:

 

 

000 00¢ 000 com 00p 0
qjdﬁlldlIdIl-l‘d‘ld-‘qu‘dldJ-l#ddddd‘ddqqd4lddlddql‘l‘JJl“ L
n

A

l

l

l

I

l

J

O .u
s as L

L

A

O .U

n

U

l

A

U

A

1

J

L

l

J

l

 

 

0.0

0."

0.N

0.0

0:...

0.0

0.0

0.5

(0 W! f5—U!uJ/"’Ho M)

31378 NOLLOVEIH

101

.mm wmma 000 «Ed 0 o 000 as 2 aged 00 Ba.— aoﬁeocmmdwoaczc :0 20852798; 00 Scam 04, 250mb

00 20.2928

 

 

05 00 00 0+ 00 0N 0. 0 .
4.1.1.1..I..<...<..<1+lli.141111411..........ii:_i.<.1.. 00
0.N

DD
DD

D D .

n. n. . o.‘

D _U U n1 I .

D n‘ 0.0

I

D 0.0

I 0.0..

n 0 as 0.000 00 OI E 0300...! I GNP
coaaoo nasal a

g 9 .6. 0.82 so 3888.. a
of

 

‘(0 W! ﬁ-U!w/*’H0 M)

31%! NO LLOVBH

102

.mm wwma 000 was 0.000 em 2 5330 A0 3a.— ccscéwmmwoavzc :0 0.00v as. 0059558 .00 800m 04. 250mm

E 206658

0m. 00 00 0* 00 0N 0 —

 

 

O
1 dildddddd [IqudldII‘I ddﬂdd‘GId qu‘dddd‘d ‘qidqd‘Jl JId‘dJI‘ lddddddq
. . a _ q 1 a 00

C.
an

: 0.N .6“— 00528 coeoo snot...

: 3 .8 0302.8 :oEoo causal

; 2 .2 3586 55 .3293 «of
c0960 58......

c N no». 03038 coﬂoo scout—l

COOPI

 

0.N
0..v
0.0
0.0
00?
0:0.—
0::
0.0—

0.0—

(0 2111! ﬁ-in/VHO M)

31378 NOUOVBB

33 Ema 000 was 0.000 as 3 0.8/am .00 30.— ccsaocwmmwogsk; :o O .000 90 00595500 .00 835 bzv 0.59%

g zo_mmm>zoo

 

103

 

on 8 on 3 on ca 2 o.
14:11:31....11......1.....:...::.......1:.:..:::: oo
o o.u
ee )
trauma.“ 0.
. t... m. D m a
D .7
0.x!- 0 0 0 W m
I I 0 u as m. m
D _
I 5 W
0 ed M a
.1
3325 55 .323; new... 0
(

03028 c0000 £000...
352.6 5.: .383: Roi
contoo casual

.533me 55 .3892 soul

00?

{DOII

0&—

 

104

a: was 8... sea 0 . new

as A0 .0005 3 00230 0003000 .00 35 0053060020.»; :0 0059558 .00 800m 0.... 950$

03 20.9928

 

 

on on on o4 on on o. o.
-.-.<<.<.q.-.i..< oo

II I II I
”g od

II
I
I I I 3.
I I
I I I
I I 90
I

I 0.0

0.80 so 352.8 55 3333.:

0.03 «0 0302.8 :05 00000000...
00000000!

c0900 soot!

0.0F

ICOI

0N—

 

(0 mu: ﬁ-urw/tHo 1w)

31W NOLLOVEH

105

slightly (Figure 4.6). Oxidation at 800 ° C was not as effective (Figure 4.7). The
oxidations were carried out at different levels of conversion by ﬁrst gasifying a
sample to a desired conversion, followed by chemisorbing oxygen, and then
gasifying it further. It can clearly be seen that the carbon black is activated to a
greater extent by oxygen at lower conversions. At higher conversions, the

enhancement, although small, was consistently observed.

4.3.3 Discussion

 

The addition of oxygen surface groups by partial combustion enhances
the reactivity of the carbon samples and is more effective at lower conversions
(Figures 4.6 and 4.7). This enhancement, however, is not as high as that
observed by Cao & Back, who attributed the enhancement to the oxide surface
complexes formed from the oxygen present in the gas stream [39]. The
formation of surface complexes requires chemisorption of the oxygen on the
carbon surface and, therefore, the number of surface complexes is directly
proportional to the active surface area (ASA) and not to the total surface area
(TSA). The ASA of the pyrolytic carbon used by Cao & Back (which had no
structure) was essentially equal to the TSA [144]; the carbon black used in this
study, on the other hand, is highly graphitic and only a small fraction of the
carbon atoms are located at the edge of the basal planes, where the active sites
are believed to be located. Hence, relatively much less oxygen chemisorbs on this
carbon, and the reactivity enhancement is expected to be much less compared to

the pyrolytic carbon.

As discussed in Section 3.8.3, from pH measurement experiments it was
inferred that the oxygen complexes present on and near the surface of fresh
carbon are mainly C=O carbonyl groups, and the oxygen groups in the bulk of

carbon are dominantly basic in character (and are not destroyed by prolonged

106

degassing at 1000°C). The oxygen groups ﬁxed on the carbon surface via

combustion at 400 ° C are acidic and at 800 ° C are non-acidic.

The rate data in Figure 4.8 shows much lower reaction rates for degassed
samples, even after ﬁxing oxygen groups on the surface by partial combustion.
The possibility that the degassing temperature of 1000°C might have caused
graphitization of the carbon, and hence lower reactivity, was ruled out because
the diffraction pattern of this carbon was identical to that of the fresh carbon

(see Sections 3.3.5 and 3.3.6).

The thermal desorption of oxygen surface groups into CO and 002 is
known to produce highly reactive ‘nascent’ sites on the carbon surface [145,146].
The ASA generated by the decomposition of the initial oxygen surface groups
has obviously been destroyed by the high temperature degassing of the carbon
sample. The high temperature degassing of carbons and graphite has been
reported to cause surface annealing and to change the ASA [147], resulting in
much less chemisorption of oxygen and fewer active sites. The small reaction
rate, following the thermal destruction of active sites, is attributed to existence

of the basic oxygen groups as was suggested by Blackwood [40].

However, if the basic oxygen groups are responsible for a very small
portion of the activity of the carbon, then the large activity of a fresh carbon
sample must be associated with either some other oxygen complexes or some
other active sites which are not related to oxygen in the carbon, or both. In the
former case, these active oxygen complexes are decomposed, followed by thermal
annealing of the generated active sites, or transformed into basic complexes by
degassing. The basicity of the surface of a carbon sample gasiﬁed after
degassing does not differ from that of a gasiﬁed fresh carbon (Table 3.4),
Suggesting that the transformation of other oxygen groups into basic groups is

unlikely. Thus, thermal annealing of active sites seems to be responsible for the

107

observed large loss of the activity of a degassed carbon sample. Once these
active sites are thermally destroyed, less oxygen is chemisorbed and less active
sites are generated. This explains the much lower reaction rates of the degassed
carbon samples even after oxidizing. At this point, however, it is not clear
whether these active sites have their origin exclusively in thermal decomposition

of oxygen complexes or some other sources.

The higher enhancement of reactivity observed for oxygen groups formed
at 400° C can be attributed to their lower stability at the reaction temperature
of 865 ° C. These oxygen groups decompose at temperatures higher than 750 ° C;
the desorption of CO and C02 is known to produce highly reactive ‘nascent’
sites on the carbon surface [146] and these sites are postulated to act as active
sites for gasiﬁcation. The oxygen complexes formed at 800°C, on the other
hand, are much more stable at this high temperature and would remain on the
surface occupying the active sites. However, the reactivity is still higher than
that of the fresh carbon because oxidation of the surface is known to increase the
ASA [145,146] and the added ASA compensates for the active sites occupied by

the oxygen complexes.

The reaction rate of the fresh carbon has been observed to be higher at
the early stages of the gasiﬁcation than at higher conversions (Figures 4.1, 4.5-
4.8). This is attributed to the presence of C=O carbonyl groups on the surface
of the fresh carbon samples. At early stages of gasiﬁcation, decomposition of
these carbonyl groups creates the reactive ‘nascent’ sites and thus enhances the
reactivity of the sample. At higher conversions, these active sites are consumed
or thermally annealed and less carbonyl groups are present to generate active

Sites. The reaction rate therefore slowly decreases.

108

It may be argued that high initial activity is due to the electronic effect of
carbonyl groups on the carbon lattice, so that upon their slow desorption the
activity also decreased slowly. If this were the case, the reaction rate would
decrease after 3 hours of heating a sample at 865°C in helium, due to both
desorption of carbonyl groups and the thermal annealing of the active sites. The
reaction rate of such a sample, however, showed no noticeable decrease (Figure
4.5). The desorption of oxygen surface complexes creates more active sites and
the thermal annealing tends to reduce the number of the active sites. These two

competing processes apparently compensate and no net effect is observed.

From the above results a mechanism for the observed enhancement of the
hydrogasiﬁcation rates by the surface oxygen complexes is postulated: oxygen
surface groups enhance the reactivity of the carbon in hydrogen by generating
active ‘nascent’ sites on the surface upon decomposition, not by the so called
“electron-transfer” mechanism in which the chemisorbed oxygen alters the
electronic state of the crystalline lattice. If the “electron-transfer” mechanism
was responsible for the activation of the carbon, the enhancement would be
observed only as long as oxygen was bound to the surface. However, the result
presented in Figure 4.5 suggests the opposite; i.e., the enhancement was observed
following desorption of the surface complexes. It should be noted that this is
not to say that all the active sites are exclusively associated with the oxygen
groups in the carbon and that the degassing deactivation is purely an annealing
process of these sites. There may very well be other active sites associated with

impurities in the carbon.

Chapter 5
Catalyzed Hydrogasiﬁcation

In Chapter 4, the reactivity of carbon black toward hydrogen gas and the
inﬂuence of oxygen surface complexes in the absence of potassium carbonate or
other catalysts were discussed. In the present chapter the effect of the presence
of different amounts of K2CO3 and the role of oxygen surface groups in the
catalysis process are discussed. This is done by ﬁrst presenting the results of

kinetic rate measurements and then comparing them with the results presented

in Chapter 4.

109

110

5.1 Background

5.1.1 Thermodynamic Equilibrium

 

By deﬁnition, a catalyst is a substance that increases the rate of a
reaction and can be recovered chemically unchanged at the end of the reaction
[148]. The catalyst provides an alternate mechanism which is faster than the
mechanism in the absence of the catalyst. The equilibrium constant for the
overall reaction is, however, determined by the Gibbs free energy of the reaction,
which is calculated from free energies of formation of the reactants and products,
and therefore is independent of the reaction mechanism. That is, the presence of
a catalyst does not change the thermodynamic equilibrium constraints of the

reaction; instead it merely enhances the reaction rate.

KQCO3 cannot be recovered chemically unchanged, because some
interactions with carbon form intermediate species (for instance, free potassium
metal) which vaporize at the high reaction temperatures (see Section 5.1.3). In
this respect, K2003 is not a catalyst in the traditional sense of the word.
Nevertheless, the presence of KQCO3 on carbon enhances the methane evolution
rate signiﬁcantly [38], and, furthermore, methane is by far the main product,
similar to the case of the uncatalyzed hydrogasiﬁcation (see Section 4.1.1).
Hence, it seems that the interaction of K2003 with carbon and the products thus
formed (CO, 002, etc.) are not substantial enough to alter the course and the
overall reaction of the hydrogasiﬁcation of carbon. Instead, the interaction
forms some active intermediate species which catalyze hydrogasiﬁcation.
Therefore, the equilibrium constant K, assumes the same value for the catalyzed

and the uncatalyzed hydrogasiﬁcation of carbon.

111

5.1.2 Kinetics & Mechanism

 

The catalyzed gasiﬁcation of carbon is an enormously complicated
reaction, and its kinetics are too complex to allow any theoretical derivation of a
kinetic expression. The major difﬁculty is a lack of knowledge of overall
mechanisms. Even though some intermediate species (for instance, >C-O-K
[12,21,22] or some non-stoichiometric potassium oxides [24,25]) have been
identiﬁed, the exact sequence of event leading to formation of these species and
their exact role in the overall mechanism is not fully understood. Thus, the
various proposed mechanisms [12,14-30] are only in the realm of speculation.

Since the reaction rate depends on the nature of the carbon used as well
as the nature and loading of the added catalyst, any empirical kinetic
expressions have very limited applicability. Very few attempts have been made
to study the kinetics of catalyzed hydrogasiﬁcation of carbons, and probably the
most detailed study of the kinetics are those by Gardner et al. [38]. They found
a kinetic expression of 3/2 order in hydrogen for coal chars at pressures up to
1000 psi and temperatures as high as 1000° C in the presence of KHCO3 (which
had about the same activity as K2003). They also found the following relation

for the activation energy E,

E, = E, | H + 2.43x (5-1)

In the absence of the catalysts they found

E, = E, I... — 2.43x (52)

Thus the activation energy for hydrogasiﬁcation of their coal char was not

112

affected by the presence of the catalyst at low conversions, but was reduced at

higher conversions.

5.1.3 Carbon-K2C03 Interaction

 

Potassium is not active as a catalyst in its carbonate form; active
intermediates are produced via interactions of the carbonate with carbon. The
extent of the interaction depends on the degree of dispersion of the carbonate on
carbon surface. For instance, if carbon particles and KQCO3 grains are mixed
mechanically so that there is a very small contact area between carbon and the
carbonate, carbon-KQCO3 interaction and thus the activity would be minimal.
The carbonate is activated substantially as long as there is a good degree of
dispersion of the carbonate molecules on the carbon surface, and the
impregnation method to achieve this high level of dispersion is immaterial [12].
When highly dispersed, the carbonate evenly spreads into a thin ﬁlm on graphite
[149] and other carbon surfaces [24,33] at temperatures as low as 550°C
(compared to its melting point of 891°C) followed by reaction with carbon

surface and formation of active species.

The exact nature of the intermediate species formed during interaction of
the carbonate and carbon surface has been studied extensively for over a
century; but still remains a point of great controversy. Roscoe & Schorlemmer
[150] and Fox & White [151] have proposed the carbothermic reduction of
potassium and sodium carbonates to free alkali metals and carbon monoxide at

temperatures above 800 ° C

M2003 + 20 —> 2M + 300 (53)

where M stands for potassium or sodium. Verra & Bell [19] have presented some

113

results in support of this reaction. Furthermore, McKee et al. [15,18] have
suggested this reaction as the ﬁrst and the rate controlling step of the catalysis
during steam and C02 gasiﬁcation of graphite at temperature above 700 ° C and
coal and chars at above 600°C. Sancier [35] has also used this reaction as the

ﬁrst step in explaining his in situ e.s.r. results.

However, in some other studies in addition to the CO (Reaction 5-3) some
C02 has been detected, the origin of which is a point of dispute. For instance,
Yuh & Wolf [23] observed Reaction 5-3 for graphite at temperatures above
700°C, but for Illinois No. 6 coal they also detected CO2 in the temperature

range 500 ° -700 ° C and attributed this observation to the following reaction

Freriks at al. [31] have shown that the decomposition temperature depends on
the dispersion level of KQCO3 and above 700° C the products were C02 and a
potassium containing surface complex the nature of which could only be
speculated. In addition, Moulijn et al. [30,33,152] have observed, by using in
situ FTIR technique, that at 500°C KQCO3 disappears on carbon surface and

CO2 and CO evolve. They have explained their ﬁndings by

K2003 —> K20 + (:02 Tm, <900 °K (5-5)

with further reaction of the oxide with carbon

K20 + c-» 2K + 00 T,,,, >900 ”K (5-6)

114

They showed, in addition, that if the contact between KQCO3 and carbon is
poor, say because of dry mixing, no interaction is observed at temperatures
below the melting point of K2C03, and that part of the free alkali metal thus
formed further reacts with carbon to form surface complexes which thermally
decompose at above 800°C (no elaboration was made on the nature of these
complexes). The excess free metal (above the saturation ratio K/ C of 0.18 atoms
of K per atom of C, or KC5'5) evaporated readily at above 500 ° C. Huhn et al.
[153] and Yokoyama et al. [154] have proven the existence of the sequence of

Reactions 5-5 and 5—6 by experiments with 13C-labeled KQCO3.

The ﬁndings of Saber et al. [27] are in direct contrast to Reaction 5-5.
From studies with K213CO3 they concluded that below 730° C the evolved C02
does not originate from the carbonate, because the carbonate did not decompose
to a large extent. Instead, any 13C02 evolved arises from isotope-exchange
reactions. Above 730 ° C, where the carbonate did decompose substantially, they
suggested that oxygen surface complexes stabilize the free metal formed via
formation of surface compounds in which K/O ratio is 2, irrespective of the

initial oxygen surface concentration.

Huttinger & Minges [20] support the formation of K20 on coal (Reaction
5.5) at above 500° C, but ruled out the feasibility of the formation of K20 and
the free potassium metal on graphite based on thermodynamical considerations.
Instead, they proposed a scheme which involves formation of KOH followed by
formation of a potassium/oxygen oxide (K, O, with y <:c) which they believed
to be the active intermediate species. In agreement with this scheme, Wood at
al. [24,25] inferred that the carbonate in contact with char undergoes a chemical
and physical transformation to form a molten potassium oxide ﬁlm that covers
the char surface. Their measured equilibrium vapor pressure of free potassium

metal (by Knudsen cell diffusion technique) over the temperature range of 450°-

115

700° C above this ﬁlm was orders of magnitude below that appropriate to the

carbothermic reduction of the carbonate to free metal.

Interpreted in a somewhat different direction are the studies by Mims &
Pabst [155], who report C02 evolution following interaction of K2C03 with
carbon or coal at ~450 ° C. They believe that [12] the interaction of carbon with
the carbonate involves bonding of potassium to surface anionic groups on carbon
(phenolates being the most likely candidates) to form alkali phenolates (C-O—K

complexes) and thus stablizing oxygen bonded to carbon:

K2C03 + 2C—OH -+ 2K+ o- C + H20 + co2 (5-7)

Hoshimoto et al. [156] have veriﬁed the existence of these C-O-K complexes on
carbon surface by FTIR technique and by showing that the amount of surface

oxygen and potassium were in a 1:1 proportion.

The above studies have mostly concentrated on evaluating the individual
reactions and determining the primary and secondary reactions (CO and C02
evolution). Very few researchers, most notably Sams ct al. [157], have paid
much attention to the kinetics of the loss of the free potassium formed via
interaction, with carbon surface. These researchers have observed rapid
Potassium loss at above 700° C following the reduction of the carbonate with
Carbon, with the vaporization rate being faster under a reducing atmosphere
tllan that under oxidizing conditions. They have concluded that the interaction
of the carbonate and carbon is a prerequisite for the potassium vaporization,
and, furthermore, the gasiﬁcation rate depends only on the K/ C ratio (as long as
it is below the saturation ratio of 0.025) and is independent of the heat

treatment history of the carbon.

116

In diversion from the above popular schools of thought regarding the
nature of active species and their formations, some researchers have speculated
the involvement of potassium-carbon intercalation compounds in the catalysis
process of steam and C02 gasiﬁcations [28,30]. Kapteijn et al. [158] have
illustrated (using X-ray analysis) the formation of potassium-graphite
intercalation compounds from mixtures of KQCO3 with activated carbon or coal
(but not with graphite) in nitrogen at 830°C. However, no experimental
evidence has yet been provided to actually support either the formation of the
intercalation compounds under gasiﬁcation conditions or, more importantly,

their involvement in the catalysis process.

In fact, Tromp & Cordfunke [159], by using a high temperature XRD
technique, have shown that no formation of potassium-carbon intercalation
compounds can be expected during KQCO3-catalyzed carbon gasiﬁcation because
they are not thermodynamically stable at gasiﬁcation temperatures (680°-
880° C). Ferguson et al. [160] have provided further evidence for the thermal
instability of potassium-graphite intercalation compounds under gasiﬁcation

conditions using SEM and in situ XRD techniques.

From the above studies, in summary, it appears that when KQCO3 is
highly dispersed on carbon, it spreads into a thin ﬁlm at temperatures as low as
500 ° C, followed by reduction to metallic potassium via interaction with carbon
Surface groups. The free potassium metal thus formed reacts with oxygen
Surface groups to form surface complexes such as C-O—K which prevent the
Vaporization of potassium and, hence, retain the free potassium metal on carbon
Surface. In an oxidizing atmosphere such as C02 or H20, where a constant
S1lpply of oxygen is available, an alkali oxide liquid layer with excess alkali metal
is formed. This oxide layer reduces the vapor pressure of the alkali metal by

Orders of magnitude and, therefore, reduces potassium losses from the surface at

117

the high gasiﬁcation temperatures. In reducing environments, no such oxide ﬁlm
can be formed and thus more potassium is lost via vaporization. Also,
formation of potassium-graphite intercalation compounds seems to be highly

improbable at the high gasiﬁcation temperatures.

However, the mechanisms or even the nature of active intermediate
species and the exact sequence of events taking place on carbon surface during
gasiﬁcation are yet far from being understood. In addition, although the
importance of the presence of oxygen on carbon surface has been demonstrated,
no attempt has yet been made to study the nature and the functionality of
surface oxygen groups as related to the catalysis process. That is, it is not
known whether different oxygen surface complexes (acidic, basic, etc.) behave

similarly, or whether their inﬂuence on the catalysis depends on their nature.

5.1.4 Effect of Carbon Impurities on Reactivity

The Ravon 16 used for hydrogasiﬁcation studies has various impurities
such as sulfur, nitrogen and ash (see Section 3.3.1). The ash content is too small
to cause any deactivation of the catalyst via formation of aluminosilicates
(kaliophilite) [161-163], and other elemental impurities such as iron are too small
to increase the reactivity noticeably. The nitrogen impurities also do not create
any complications because they readily leave the carbon surface at high
temperatures, as evident from XPS results of Section 3.3.7. The hydrogen
impurities are not signiﬁcant because, ﬁrstly, they are very small and, secondly,
they are in the same form as the hydrogen complexes formed during
hydrogasiﬁcation, namely, C-H bonds. The effect of oxygen impurities will be

discussed in the following Sections.

118

Sulfur, however, is present in substantial amounts, 1.54 wt%, and could
form potassium sulﬁde which is not active. Sulfur exists in carbon either as
highly stable C-S-C structures in which sulfur may be present in the aromatic
rings of the carbon layers, or as sulﬁdes and hydrosulﬁdes [164]. The C-S—C
structure is formed by addition to unsaturated sites and does not decompose
completely even at 1200° C. Upon thermal decomposition this structure yields
mostly carbon disulﬁde C82. The sulﬁde and hydrosulﬁde groups, on the other
hand, are formed by substitution through interaction with certain oxygen groups
which come off as carbon monoxide on high temperature decomposition, and are

relatively unstable (start to decompose at a temperature as low as 400 ° C).

The XPS results of Section 3.3.7 indicate that samples of Raven 16 degassed
at 1000° C or hydrogasiﬁed at 865 ° C show no trace of sulfur on their surface.
It thus follows that the sulfur present in Raven 16 is not in the form of stable
C-S-C structures. The sulfur impurities thermally decompose to hydrogen
sulﬁde, H28, or are substituted by hydrogen to form H28. Any potassium
sulﬁde, K28, formed during hydrogasiﬁcation via reaction with the carbonate, is

reduced to free potassium metal by hydrogen under reaction conditions [20]

K28 + H2 -+ 2K + H28 (5—8)

Therefore, the presence of sulfur does not interfere with the activity of KQCO3.

119

5.2 Intrinsic Nature of Reaction Rates

For kinetic studies of catalyzed hydrogasiﬁcation, a ceramic boat (described
fully in Section 2.10) was used. This sample holder was not deep enough for the
carbon samples to exhibited any mass transfer limitations within the carbon bed
during the uncatalyzed hydrogasiﬁcation reactions (Section 4.2.1). However, in
the case of catalyzed hydrogasiﬁcation, where the reaction rates are much faster,
the mass transfer resistances might become signiﬁcant, and thus diffusion
limitations need to be examined. The temperature variation of carbon bed
associated with using sample holders of different geometries and conﬁgurations
(see Section 4.2.2) is absent because only one type of sample holder is used for

these studies.

5.2.1 Mass Transfer Resistances

Typical sets of raw data of methane evolution rate versus time for fresh
carbon as well as carbon samples impregnated with KQCO3 to 5 wt% and 10
wt% are presented in Figure 5.1. The maximum measured rate of the 10 wt%-
sample is about 30 ml CH4/min-g initial carbon which is three times faster than
the methane evolution rate used for the analysis of diffusion limitations in
Section 4.2.1. Similar analysis would predict the magnitude of mass transfer
resistances during hydrogasiﬁcation of the impregnated samples. However, a few
assumptions must be made because of the presence of the catalyst in the carbon
bed. The ﬁrst assumption is that the porosity development during the
hydrogasiﬁcation (see Section 3.3.3) is not altered by the presence of the
catalyst. This assumption is justiﬁed by the observation that the surface area

development is not altered by the presence of the catalyst.

120

a: wmma 0cm was 0 . mow as S seam 80.: 3a.. 005296 28508 33034 H6 ouswmh

2.5 us:
can con omu 00m 0m... P 00 F on - 0

Il‘ l qdlﬂlﬁd - I I d d q .4 d d dld .d 1 d d -ld ‘\ o

  

    

O
'-

O
N
(a mu: 6-U1w/1’Ho M)

0
yr)

.3960 58...! o
nooux 5; ml +
nooux 5; o... e o...

 

 

3.1318 NO LLOVEH

121

The second assumption is that the presence of potassium carbonate on the
surface of the carbon does not inﬂuence mass transfer processes within the voids
and pores of the carbon bed. This assumption is plausible because the vapor
pressure of free potassium metal (see Section 5.1.3) and partial pressures of other
gases such as CO and C02 are very low compared to those of hydrogen and
methane. Furthermore, the catalyst has a high degree of dispersion on the
carbon surface and readily spreads into a very thin ﬁlm on the carbon surface
during reaction, thus not blocking the pores and voids appreciably. These two
assumptions make the values of 6p, T, and thus D, independent of the catalyst

loadings of interest in the experiments.

The value of hydraulic radius R, of the ceramic boat sample holder used for
the catalyzed experiments (see Section 4.2.1. and Equation 4-22) is not changed
by the presence of the catalyst. Hence, the presence of the catalyst merely
increases the magnitude of the rate constant km, by a factor of 3, which in turn
results in a Thiele modulus of 0.308 and an effectiveness factor of 0.96 if the
catalyzed hydrogasiﬁcation reaction exhibits ﬁrst order kinetics. However, as
discussed in Section 5.1.2, the kinetics of K2CO3-catalyzed hydrogasiﬁcation is

3/ 2 order in hydrogen concentration. For a second order reaction

ern = k2 CA2 (5.9)

T2771

where k3,, is the rate constant of the second order reaction [138] and CA is the

concentration in the bulk gas, the Thiele modulus ¢ is deﬁned as

R k2 C
45 = _"_ .LC‘. (5-10)
3 1),

 

122

These two equations indicate that the value of the Thiele modulus does not
depend on the order of the reaction, at least for reactions with simple kinetics.
Thus, for a given hydrogen partial pressure an N-fold increase in the reaction
rate simply increases the value of (15 by a factor of W. Hence, the value of ()5

equals 0.308 even for the 3 / 2 order hydrogasiﬁcation.

For a given value of 45 less than 0.5 the effectiveness factors for a second
order reaction is at most 0.05 units lower than those for ﬁrst order reactions
[138]. Therefore, the effectiveness factor for the catalyzed hydrogasiﬁcation of
3/2 order is expected to lie in between those of ﬁrst order and second order

reactions; e. g., at least about 0.91. This implies that diffusion resistances again

are negligible.

However, the assumptions made regarding the estimation of D, may not be
completely valid and thus the diffusion resistances were examined
experimentally. Figure 5.2 represents the results of such a study in which the
carbon bed depth was varied by a factor of three and the reaction rates were
compared. The measured reaction rates of the two samples are essentially the
same within the experimental error (see Figure 5.3). It clearly shows that
reduction in the carbon bed thickness does not result in the enhancement of the

measured reaction rates, implying that diffusion resistances are insigniﬁcant.

The absence of substantial mass transfer resistances is further supported by
the results of activation energy measurements (Figure 5.4) [116]. As mass
diffusion resistances become more signiﬁcant, the magnitude of the apparent
activation energy E,’ gradually decreases [137,138,165]. For (I) greater than 3
the reaction kinetics are controlled by diﬂusion resistances and the value of E,’
reduces to half the value of the intrinsic activation energy E, . In this range the

Arrhenius plot would yield a straight line.

123

am 33 8m Ba 0 . 8m

as son 2 522m 033.35 00 comamocmmawoésn mats—o mmoasaﬂmma 3.023: mde «.0 ouswmh

g zo_mmu>zoo

 

8. cm on E. 8 on 9. on cm 2 o.
1.1(41ll.21...1.1.1-4111...2.11111L00
an H
e no.0
C”
m 4 ..o.2
D l
U
DI mod—
‘ L
8.8
n. U
mesa
U L
4 ..o.on
9: can... a”
as 2.6.. 4“

 

 

 

0.nn

(o tun! 5—U!w/"Ho Iw)

31W NOLLOVEEI

124

.45 E3 OS was 0 . new as 838% mOOmvToxSB 0H .00 95.8 00585335»: .00 ﬁzwﬁoscoaaom 0.0 ouswmh

00 zo_mmu>zoo

00 F 00 00 0s. 00 on 0* 0n ON 0 F 0 .

   

 

 

1-24.11.21.21.1.4.2.4411.11.11... o
4
4 %1.0.0
4 4 3.2 w.”
a... m. o
4 4 god" .Wn
g. I w
0 40.0N m..
... .D
o 4 3.3 m.
... mu
4 4 mod” (0
U
0 moan
.93

 

 

BIVH NOLLOVEB

.«E 33 00m. .800: comm—8,08 9500 as 2 :96.”— 00 :omamommmmmoém; .50 .020 2.32:4 v.0 0.53%

fix???

 

125

 

 

 

0.0 4.0 «.0 0.0 0.0 0.0 4.0

1 a q [a d 4 All 4 A 4 J a a J J a J a J 1 4 J J
4 no;
.. 4 0.N
1 n
r .. 0.4
. .
.T .U 0.0
n. 1. o.m
ﬂ .. 0.0..
1 L 0.0N
a .
0. nooux no... .. 1 0.9.
1 £09.00 LOP—Ll O .

b b n F f F [If p L .l P P p b p - lp . f r n p . Doom

 

(0 4w! f5-U!uJ/"'H:> M)
3m; Nouavaa

126

The Arrhenius plot of Figure 5.4 for 10 wt%-K2C03 samples at 22%
conversions is a straight line, and the calculated value for (15 at 865 ° C, based on
conservative assumptions, is however only 0.308. This value is not high enough
to obtain a straight Arrhenius plot with substantial diffusion limitations
present. Hence, it follows that strong diffusion limitations within the carbon bed
of KQCO3-impregnated samples are absent, at least at lower conversions. It
would be worthwhile to measure E, at higher conversions where higher degrees
of porosity are present. Furthermore, a value of 51.3 kcal/gmole-°K for E,

indicates that external mass transfer limitations are absent (see Section 4.2.1).

Therefore, the measured reaction rates for the catalyzed hydrogasiﬁcation
reactions for small sample sizes (~80 mg) are free from masking by signiﬁcant

diffusion limitations.

5.3 Hydrogasiﬁcation of Coconut Charcoal

The hydrogasiﬁcation procedure and conditions for KQCOg-catalyzed
hydrogasiﬁcation of coconut char were the same as the non-catalytic
hydrogasiﬁcation of carbon black discussed in Section 4.3.1 (865 ° C and 500 psi
H2).

Hydrogasiﬁcation of samples of 150-200 mg (22.9 wt%-K2C03) in ceramic
sample holders yielded very high methane evolution rates, as can be seen in
Figure 5.5. On one hand, the bulk density and the BET surface area of coconut
char is higher than those of carbon black (see Sections 3.1.1 and 3.3.3) indicating
larger particle sizes and a much higher degree of micro-porosity in coconut
charcoal. On the other hand, the measured reaction rates in Figure 5.5 are

about four times faster than those of the KQCO3-impregnated carbon black

127

.30 9mg 00m. 05w 0 . mew as E09223 2588 80.0 3a.. 02.8850 2552: 328an 0.0 onawmh

3.5 m2:
8. 8 8 3 om

 

_ ddd Jqu ‘jded ddqj; ‘14 1‘Jld *dqddd
. e . 1 . o .

 

llLlllJllJllllJJJijlJJJ'IlJJL'lIAlIIILJLL o

 

 

O

O
F)

O
(D

O
O)

0NF

0...:

00..

0pm

(0 1!”! Swim/VH0 um)

31%! NOLLOVBH

128

samples and, thus, it follows from the discussion in Section 5.2 that diffusion

limitations are deﬁnitely signiﬁcant.

Furthermore, most of the carbon is hydrogasiﬁed before the steady state
temperature of 865°C is established, making the determination of the actual
solid carbon temperature at low conversion complicated. The high ash content
of the coconut charcoal of more than 3.7% also introduces an element of

uncertainity into the interpretation of the role of the catalyst.

Therefore, it was decided that the coconut charcoal was not a suitable

choice for studying the catalysis of hydrogasiﬁcation by K2003, and the

experimentation with this char was not pursued.

5.4 Hydrogasiﬁcation of Graphite

After using coconut charcoal, it was decided to use a carbon sample which
had lower surface area and lower activity than the coconut charcoal so that mass
diffusion resistances could be eliminated and the hydrogasiﬁcation rates at low
conversions under steady state temperatures could be measured. For this
purpose, graphite powder samples impregnated with up to 9 wt% K2003 were
tried (for characterization of this graphite see Section 3.2). The gasiﬁcation

procedure and conditions were the same as those for coconut charcoal.

The measured reaction rates for graphite samples were very low (even for 9
wt%-K2003 samples), but much too high to be accounted for by sample weight
loss during hydrogasiﬁcation. With a 9 wt%-K2003 sample in the reactor,
maximum methane evolution rates (irreproducible) of only 2 ml/ min were
measured, and in experiments with no carbon sample inconsistent rates up to 0.9

ml/min were obtained. This indicated that in the former case, only a fraction

129

(which could not be determined with acceptable accuracy) of methane was from
the graphite sample. The origin of the extra methane detected was determined
to be the metallurgical carbon from the reactor vessel wall and the stainless steel
screen used to keep the particulate carbon bed in place, which deposited out
during thermal cycles from room temperature to the reaction temperature. The
irreproducibility of the hydrogenation rate of these carbons made the accuracy of

the measurements of the graphite hydrogasiﬁcation rates highly uncertain.

Hence, it was decided to insert a quartz tube lining inside the reactor to
eliminate contact between the high temperature bulk hydrogen gas and the wall
of the reactor vessel in the hot zone. In addition, it was decided to use a more
reactive sample so that the rate measurements of hydrogasiﬁcation of carbon
samples were not complicated by the methanation of the metallurgical carbon.
Thus, further study of graphite hydrogasiﬁcation was abandoned and carbon

black Raven 16 was used for further studies.

5.5 Hydrogasiﬁcation of Carbon Black

5.5.1 Procedure

 

Hydrogasiﬁcation, oxygen chemisorption, and heat treatment procedures for
KQCO3-impregnated carbon black samples were basically the same as those for
the uncatalyzed samples (Section 4.3.1). However, minor modiﬁcations had to be
made in a few cases to accommodate the presence of K2003 on the carbon
surface; in experiments with degassed carbon black, degassing was carried out
before impregnation with K2C03, but in the case of heat treatments at 865° C
for three hours in He, the K2CO3 impregnation was done ﬁrst. Furthermore, in

experiments involving oxygen chemisorption on the carbon surface, the

130

chemisorption was done before the impregnation with K2003 (in one experiment
oxygen was chemisorbed on carbon surface at 400° C after K2003

impregnation). The impregnation procedure and details are discussed in Section

3.3.2.

Some experiments were carried out on partially hydrogasiﬁed carbon black
which was subsequently impregnated with K2003. For these experiments, fresh
carbon black samples were hydrogasiﬁed to a desired conversion at 865 ° C in 500
psi H2 and then were impregnated to the same level of potassium content that a
fresh 10 wt% sample would have after gasifying to that conversion. These latter
samples are referred to as “10 wt%” samples, although, the actual potassium
carbonate loadings after 47% and 65% conversion were 11.8 wt% and 14.3 wt%

of the impregnated samples, respectively.

Reaction rates were measured by collecting product gases in cold traps and
subsequent analyzing by gas chromatography. The product gases were collected
at steady state temperature of 865 ° C for a period of 10 sec to 1 min depending

on the activity of the samples.

5.5.2 Results

 

5.5.2.1 Catalyzed Hldrogasiﬁcation

 

5.5.2.1.1 Fresh Carbon Black

Raw hydrogasiﬁcation rate data for the four carbon black samples studied
are reported in Figures 5.1 and 5.6 as methane evolution rate versus reaction
time. These results are typical of repeated experiments which exhibited good
reproducibility (Figure 5.3). The start of experiment in the Figures 5.1, 5.5 and

5.6 (time = O) is taken as the time hydrogen ﬂow is begun (at 500°C) and

131

.um Ema cow was 0 . mow as 2 5.5m 80.0 3a.. comes—96 2:232: 8303< 0.0 ouswmh

3.5 m2:
com ooe com com oo_

 

 

Jlldd-dqldJJdlddtdl-lqddiddl‘lq-JJGJ(*Jldld.1‘I‘dddlql‘d‘lql‘dlqjl‘n

b0.0.0 e

 

5
~----

 

c0900 52...... o
nooux 53 T p

o .

ILIJIIJJJIIJII

LJLJJJIIIJJILILJI

IJJLJIJJI

 

 

.00

0...

0.N

0.0

0.¢

0.0

0.0

0.5

0.0

(o mu: 5-Ulw/J'HO It“)

31%! NOLLOVEH

132

heating the carbon sample is started. The methane evolution rate increases as
reaction temperature is increased, approaching steady state of 865°C after 50
min reaction time. In all experiments carbon conversion is less than 10% when

the steady state temperature is reached.

The conversion of carbon in each experiment is calculated by integrating the
curves of methane evolution rate versus time. Complete conversion of 1 g of
carbon to methane should yield 2036 ml of methane at 1 atm pressure and 25 ° C
temperature, and the ratio of the curve integration to 2036 ml indicates the
extent of the reaction. Conversion of carbon to other compunds is negligible
because methane is by far the main product. For hydrogasiﬁcation experiments
with fresh and 1 wt% samples, the calculated conversions compared well at all
burn off levels with conversions determined by weighing residual samples after
experiment, showing that all methane is accounted for. For 10 wt% samples
gasiﬁed to more than 85%, however, weight conversions were consistently 9-14%
lower than the rate conversions. This is due to very short gas collection
periods. Cold trap pressures are much lower than the pressure of the reactor
efﬂuent stream, and thus Within a given period of time more gas is passed
through the traps than expected from efﬂuent ﬂow rate. For gas collection
periods of l min, used for fresh and 1 wt % samples, this effect does not appear
to be signiﬁcant; but, for a gas collection period of 10 seconds used for 10 wt%
samples, this effect is severe. In these cases the gasiﬁcation rates are normalized
to weight conversions. The methane evolution rates reported in the Figures are

all normalized.

Surprisingly enough, for 10 wt% samples gasiﬁed to 20-70% conversions,
the rate conversions are lower than weight conversions 3-7%. For instance, in
one experiment carried out at 776° C and 500 psi the rate conversion was 29%

whereas the weight conversion was 66%. In this particular case, the sample

133

spontaneously burned (the sample became red hot) at room temperature as soon
as it was taken out of the reactor and was exposed to the atmosphere. The
thermocouple above the sample momentarily registered temperatures higher than
100°C. It thus seems that at low conversions some highly active complex is
formed on carbon surface which readily reacts with oxygen even at room
temperature. Therefore, for these samples the measured rates were normalized
not to measured weights, but, instead, by reducing them by 10%, similar to the
case of samples gasiﬁed to high burn off levels. The normalized rates are

reported in Figure 5.3.

The results show that methane evolution rate decreases at steady state as
carbon is consumed for fresh and 1 wt% KQCO3 loaded carbon samples (Figure
5.6). The addition of 1 wt% catalyst is seen to have little catalytic effect. In
contrast, the methane evolution rate for the 5 wt% and 10 wt% loaded samples
increases as carbon is consumed, reaching a peak at 60% conversion for both
loadings (Figures 5.1 and 5.3). It is interesting to note that the initial steady
state methane evolution rate for the 5 wt% loaded sample is nearly the same as
that for the fresh carbon sample without catalyst. This observation cannot be
explained satisfactorily at this point because the 10 wt% sample behaves

differently.

5.5.2.1.2 Treated Carbon Black

Reaction rate data of differently-treated “10 wt%”-carbon samples are
presented in Figures 5.7-5.10. Treatments such as oxidizing, degassing and
gasifying, individually or in combination, were used to obtain more insight into

the nature of the carbon surface during gasiﬁcation.

134

 

_ 8.

.aomEQDcOo mamas, «I wmmm 00m was 0 . new as 3a.. 2558063 swam—3A5 Nam ouzwmh

g zo_mmm>zoo

om om oe om _
A DDDDD
D...
+ R DDDD
+ o D D ++++¢.
D + 4
+ D D D +
+ 4
4
+ + 4
+ +
4
4 4

g 2 .2 0.89 so 8858..
g n .8 0.8m an o... 5 36oz:
nooux 5; o...

D
+
‘

 

mp

0N

mm

on

mm

0.?

(o ”1!“! LUNA/"Ho It“).

31%! NOllOVEH

135

.aofao svmmmmos no «5 End 00m. was

0 .www as men.— =owﬁ85mmw :o ”8.23% emu—38 was 0 .00v as :OSsExO .00 $8.0m 06 van—mum

00 2055528

 

 

25.638050 N0 55 @030. 63390....
53 0— 3 uovoo. uco uommumool
vouuomoon.

on 0% 0n 0N 0..
...+1+.+-..4-++¢«+4...¢4++...........-4.-4..44¢4..+¢4+.44. .
0 0 O O O U
.m
I .._
I4- II I u
D l4__.4D u 4 u
D D I I A
D C D‘ M
_I I— D,.D m
D D m
D I U
‘ .m
M
D I wanna. :05 voﬁoﬂEoso N0 6030000! U
u

OI‘D

 

 

0.0

0.N

0..v

0.0

0.0

0.0P

0&—

(o mu: 6-U!w/"Ho M)

31318 N0 LLOVBEI

136

.cofmu soammdw was? .00 «I E

O .mcw as 88 cowaaofmdw :0 comic?“ $338 was 0 00¢ an nosaExo .«o .6me 06 van—ma

E zo_mmu>zoo

 

 

mm mm mm. mm mm mzw
11 .11. 211 .1 . 11 1.11111 11211.1 . 21 0
m
0 GOOD ﬂ H
J
a a . . a.
O O O D M
4
O 1 or
P .
D. 4
§ 1
5 mm?
I
Now
No 5 Rm 852on a .
nouocmoag 55 N0 5 No 252.6! D U
vouocmoLaET e x mm

 

 

(a mu: B—U!w/"'Ho Iw)

31W NOLLOVBH

137

.mm 5 O .. wow .6 :of8 0059mm

0&3 00 3d.— coSaoEmmw co cowﬁcsm amen—sumo was 0.00v as :oSaExo .00 Scum 0H6 ouswmh

 

 

 

$206658
00. mm 00 mm 00 mm. 05 m&.
ﬂa.¢+..a.q-.a.1.¢...q¢.qJ-Jq.a-.qu..O
4 m
4 has
‘ X H
0 m3.
4 x o o m
0 ..
o o .
x o 0 mod
x 4 x m
4 x 4 m
.u 0.0
:5: or. 2 nosocooaé 5.: .5 232.6... x mod.
.8 882.6 :2: ...53 or. 2 8.2935... 4 ”
noo~x1=g Oval. 0 ._
«3.8

 

31W NOLLOVBEI

(o 'uu: ﬁ—usw/VHO Iw)

138

Generally, such treatments did not change the trend of reactivity of samples
versus time or conversion, though the absolute rates were inﬂuenced. Heating in
helium (at 865 ° C for 3 h) after impregnation or degassing prior to impregnation
reduced the reaction rates of the samples signiﬁcantly, with degassing effect
being more severe (Figure 5.7). Oxygen chemisorption at 400 ° C either before or
after impregnation did not enhance the reactivity of a degassed fresh carbon
sample (Figure 5.8). However, the reactivity of a pure carbon sample gasiﬁed to
47% or 65% conversion and then oxidized at 400°C before impregnation did
increase to some extent (Figures 5.9 and 5.10). Again, the experimental results

were fairly reproducible.

5.5.2.2 Catalyst Loss

5.5.2.2.1 Fresh Carbon

The surface of carbon black carbothermically reduces K2C03 to free
potassium metal which is readily lost from the carbon bed via vaporization (see
Section 5.1.3). The total amount of potassium in the reaction zone of the
reactor (sample, sample holder and the quartz lining of the reactor around them)
during hydrogasiﬁcation was calculated from the ’y-ray emission rate versus time
(Equation 2-15) recorded by the detector. The curves of total amount of
potassium in this zone versus time are given in Figure 5.11 for the 5 wt% and 10

wt% samples.

The actual amount of potassium on the carbon black surface, as opposed to
the reaction zone, is of greater interest. Thus, following gasiﬁcation, carbon was
removed from the sample holder and the potassium content of the sample itself
was measured by NAA. This sample catalyst loading is plotted in Figure 5.12 as

a function of conversion for the 10% sample along with the amount of

139

dosages hair Do 3:585 a: wwma

000 was Chow as coswocmgwemomn mat—6 macs cogs?— Eoc mmo_ Esmmwﬁom .36 25mg

 

3.5 m2:
8» 8» Ba com on. 8. on o
1.41111144.1111.4411.41.1«1.111.111-O

 

 

I
l
...
.

.l
I
j
A

J
I
n
J

     

 

0D

noomxlﬁs o... p on

noomxuﬁz m: o

 

 

-00

 

(o '11“! 6/>1 6“1) ONIOVO'I wmssvmd

140

 

«E 33 000 saw 0 . new as comamommmawoém: mains 3 53am UosmOTmOOuvH 80.0 $2 Esmmmgom «H6 0.25er
E zo_mmm>zoo
00p 00 0.0 04 cm 0
I - I I I I I I I 11I - I I I I I I I I I - I I I I I I H I I - I I I I I I I IlI - I I I I I I I I o
4 op
O

x x x x mm

Anoouxlku; 03 ocon comoooml
; n o: c... 38.. noomxlﬁz o...
noouxlﬁz nu

noouxlx; o...

IOIﬂ

 

   

j
n
l
j

I
n
I
U

..l
1

 

 

jjlllllnlllllll

 

0N

on

0*

on

 

05

-8

(o w: 5/>i 6w) smovm wmssvroa

141

potassium in the reaction zone. It is seen that potassium loss from the reaction
zone via evaporation at 865 ° C in hydrogen is slow, but the potassium loss from
the sample is signiﬁcant. The catalyst loading is seen to decrease linearly as
carbon is consumed. The residual loading for residual 10% (pretreated in

helium) and 5% loaded samples are also given.

5.5.2.2.2 Treated Carbon Black

The catalyst loss from treated (degassed, oxidized or heated in He) carbon
samples at gasiﬁcation temperatures are shown in Figures 5.13 and 5.14. It is
seen that a degassed carbon sample impregnated to 10 wt%-K2003 lost catalyst
during gasiﬁcation much faster than a sample that was not degassed, to the
extent that the amount of the potassium on the carbon reduced to a level less
than that of the 5 wt%-K2C03 sample (Figure 5.13). Carbon samples gasiﬁed to
47% or 65% conversion and then impregnated also lost catalyst during further
gasiﬁcation much faster than 10 wt% sample (open symbols near 80%
conversion). When the 47%-hydrogasiﬁed sample was heated in helium at
865 ° C for 3 h, the potassium content reduced to that of the degassed sample at
that conversion (open circle at 47% conversion on the middle line in Figure
5.13). In all cases, samples degassed or hydrogasiﬁed to moderate or high
conversions prior to impregnation lost catalyst much faster than the untreated

samples.

Oxygen chemisorption at 400°C seemed to partially compensate for the
effect of degassing on catalyst loss, but had little effect on the partially gasiﬁed
samples (Figure 5.14). Comparison of Figures 5.13 and 5.14 reveals that, after
10—20% conversion, the rate of catalyst loss (slope of potassium loading vs.
conversion) was not inﬂuenced by oxygen chemisorption; the oxidation of

degassed samples merely reduced catalyst loss at the early stages of the

142

.A0 . 33 “56.8250 .00 m_m>m_ 2.2.5.»

as 835% coneso .00 23:8 Eswmmdaoa :o $08.33: 235383 :3: .«o .08me «4.0 vuzwmh

as zo_mmu>zoo
8. 8 8 9. on

0

 

-IJIIIIIIIII-IIIIIIIIII-IIIIIHIII-IIIIIIIwWJIquIIIIIIJII

< %//ﬂl

I
I

/

m
/

:5: 0r. 3 33:093.»: :05 uommomool

x; mu
:5; or. 3 unaccooaé 5.: .xno 3:39.31:
:53 or. 2 escapaé 5.: .x? 3589.31:

53 0p...

 

     

 

 

0
0w
0N

on

(o W! 5/>l 5w)
SNICIOV'I WHISSVlOd

143

.A0 . mowv commagcg .00 £9.6—

msota> as 838% c0353 00 .2828 Esmmmﬁoa :o O .00v um :oSaExo 00 Scam «In mV.—numb

00F
X

 

00

E zo_mmm>zoo
om as ou

 

:5; or. 2 sang 26 Bones .xnm 6322::
:5; or. 2 snag n5 832.6 a“? @3231...
:53 0... 3 £09.95 use No? “.3038 .uowmomool

.53 0....

.3144 4‘41-1111 «dd44-14J «1:114 «lad1u1d1dd #1-11 ud4144lqd

%

DDKX

0

 

0

0?

ON

0
I")

O 0
IO <-
(o MU! 5/>1 5w)

0 0
[N (D
ONIOVO'I WHISSVlOd

O
a)

144

gasiﬁcation.

5.5.3 Discussion

 

The main objective of this study was to elucidate some general features such
as the extent of the inﬂuence of oxygen surface complexes on KQCO3-catalysis of
hydrogasiﬁcation. To this end, the KQCO3-catalysis of gasiﬁcation of fresh
carbon black was studied ﬁrst and then the effect of some treatments such as
degassing were studied. Here, the general features of the catalysis of gasiﬁcation
of fresh carbon are discussed ﬁrst, and then the inﬂuence of various treatments

are discussed with the reinterpretation of the general features in the light of

these results.

5.5.3.1 Hydrogasiﬁcation of Fresh Carbon Black

Because gasiﬁcation is a heterogeneous reaction, the representation of
gasiﬁcation rate should be based on either a total surface area or active site
density. The gasiﬁcation rates measured for all samples are represented as
speciﬁc rate (methane evolution rate per unit surface area) based on the total
surface area measured as a function of conversion (Figure 3.1). Total surface
area is used instead of active surface area because measurement of active surface
area is extremely complicated; the magnitude of the area depends on the
conditions and gases used and they do not necessarily represent the catalyzed
gasiﬁcation conditions. The speciﬁc gasiﬁcation rate is plotted in Figure 5.15 for
each loading as a function of carbon conversion. The speciﬁc rates for the
uncatalyzed and 1% loaded samples decrease strongly during reaction (by a
factor of ﬁve), while those for the 5% and 10% loaded samples remain

essentially constant or increase slightly in the conversion range of 15-65%.

145

.comﬂoiau :ofao mama? 33 000 was 0.80 as 2 52mm 00 3a.. comes—96 2:233: 206on mad 95me

g zo_mmu>zoo
8 8 2 on

0

 

 

noouxlﬁz 2..
noouxsxi ......
0.. :33. 52......
noouxnﬁz T

 

 

-1IIIIIIIlII-I%IIIIIII-IIIIIIIII-IIIIIII1#

I+DO

IJJJJAJ

I A

ljlljjlll

LI_ILJL14JJIA

 

 

00.0

00.0

0F.0

09.0

0N.0

0N0

0n.0

(Zw—ugw/VHQ lw)
31W NOLLOVEH OlleBdS

146

A convenient and meaningful way of comparing catalyst activities under
various conditions is to adapt the relative reaction rate parameter versus
conversion (Figure 5.16). The relative reaction rate parameter is deﬁned as the
ratio of the gasiﬁcation rate of a sample at a given conversion to the reaction
rate of an uncatalyzed, untreated, fresh carbon sample at the same conversion.
When this parameter is equal to one (horizontal dashed line at the bottom of the
graph), it implies that a sample has the same activity as a fresh carbon; or in
other words, that the catalyst is inactive at that conversion. In Figure 5.16 the
relative reaction rates of 1 wt%, 5 wt% and 10 wt% samples are plotted. It is
clearly seen that the effect of adding catalyst is not noticeable for the 1 wt% and
is initially quite small and increases dramatically during reaction for 5 wt% and
10 wt% samples. This latter observation is extremely important and was the
starting point and the cornerstone of further study. The totality of the research
presented here was carried out as an attempt to explain this simple and basic

observation.

These results provide insight into the role of potassium carbonate in
enhancing hydrogen gasiﬁcation rate. First, some factor other than the mere
presence of the carbonate catalyst is also involved; otherwise, the reactivity
should decrease as potassium content of the carbon sample is decreased by
potassium loss from the sample (Figure 5.12). This is not observed; the
reactivity increases as more potassium is lost from the carbon sample. Second, it
is clear that the catalyst does more than simply increase the reactivity of
existing reactive sites on the carbon surface, for if this were the case, the ratio of
catalyzed to uncatalyzed reaction would be constant during gasiﬁcation; the
enhancement factor. Instead, the catalyst must become more active as new
surface area is developed and interacts with the catalyst. The interaction

produces new reactive sites on the carbon surface, resulting in an overall increase

a: was cow as o . saw

as 286.8250 c918 mama»; moi: cosh—38:: 8 «>532 woes.— comeaowmmmwevmm 0H6 ouswmh

g zo_mmm>zoo
os om on as on cm 0.

 

-IlI11IlI-IIIIdl1qIIJI-llquII Iﬁl‘wd-IIII-Idlqu

|||||| I \' ‘l

I
I
I
1

147

or

m—

0.0.

mm

31w NOIlOVEEl BALLV'EIH

 

noouxaﬁs 2.. -
noouxnﬁa m... p
noouxnﬁs T. a

JILLLIJLJLLLJJL'IIjlllllllljl

 

 

148

in absolute reaction rate as gasiﬁcation proceeds. In the absence of catalyst, no
new active sites are produced, and the absolute gasiﬁcation rate decreases slowly,

possibly by thermal annealing of active sites.

The constant or slowly increasing value of speciﬁc rate for the 5% and 10%
loaded samples indicates that in both cases the catalyst present is able to
activate, maintain the activity of, and even increase the activity of new surface
area formed in gasiﬁcation. This is surprising, as it appears not enough catalyst
is present to cover the extended area formed. A simple calculation shows that, if
one assigns a molecular area of 10 A2 to a potassium atom on the carbon
surface, the initial catalyst loadings of 5% (K/C=0.009) and 10% (K/C=0.019)
correspond to a monolayer coverage of potassium on areas of 43.6 m2/ g and 87.2
mQ/g, respectively. These areas covered by catalyst are even smaller at high
conversion, as approximately 40% of the catalyst is lost at 60% carbon
conversion (Figure 5.12). It thus appears that the reactive area which the
catalyst activates is a small fraction of the BET area, and that the reactive area
is formed at least in proportion to the BET area during gasiﬁcation. This is in
agreement with the general consensus [166] that gasiﬁcation occurs at edge sites

on the carbon surface.

The increase in speciﬁc rate for the 10% over the 5% catalyst loading is not
easily explained by a simple surface coverage of reactive sites. The ratio of
observed speciﬁc rates of 10% and 5% samples is approximately three over the
entire range of carbon consumption. It is not clear at present as to how this
additional catalytic effect, above that predicted by a linear increase with loading,
arises. One explanation is that some fraction of the catalyst on the carbon is
deactivated at the beginning of gasiﬁcation or cannot be activated during
gasiﬁcation [167]. For instance, if two weight percent of catalyst is not active,

the ratio of active catalyst for the 10% and 5% catalyst becomes 8:3, a ratio

149

close to the value of three for the observed reaction rates. This concept of
inactive catalyst also explains the lack of observed catalytic activity for the 1%
loaded carbon sample. The deactivation of catalyst has been suggested to result
from reaction with the impurities in the starting carbon or possibly from
intercalation of potassium into carbon. The possibility of intercalation at
gasiﬁcation conditions, however, has been discounted at high temperatures for
gasiﬁcation in other environments [159,160,168]. It may well be that a minimum
amount of catalyst must be present on the carbon surface in order to provide an
environment in which active catalyst species can form; this explanation is the

most plausible for the observed gasiﬁcation behavior.

The rate per catalyst atom present on carbon increases several-fold during
gasiﬁcation, while the catalyst loading per unit area decreases. This seems to
indicate that either just a fraction of the catalyst is necessary to activate the
carbon surface or the catalyst is activated gradually as higher conversions are
approached. In the latter case, most of the catalyst is probably in an inactive
form and as new surface area is developed it is activated via interaction with the
new surface. Wigmans et al. [169], have noted a similar phenomenon, stating
that the quality of an active site increases during gasiﬁcation as catalyst is
activated from an inactive state. One explanation is thus that the observed
increase in gasiﬁcation rate is dictated solely by activation of the catalyst. In
this case, the carbon surface area would be inconsequential as long as sufﬁcient

area is available for the activated catalyst to spread onto.

If catalyst activation is a slow process, one would expect that helium
pretreatment at high temperature would facilitate catalyst activation, and
therefore result in higher gasiﬁcation rate. The results of such an experiment are
shown in Figure 5.17, in which the sample was held at 865°C in helium for

three hours and then gasiﬁed. It is seen that gasiﬁcation rate was lower after

150

.3» H8 000 was O.m©w as m: :oDam soaacweaﬁw no .335an :o acmanoS .23; ..o Quota 5H6 ouzwmh

 

 

3.5 m2:
omN DON 0m: 00— on
.1 u1+q.q4111_.+11_1.
..T/I \\,,
# 9++1-
I \\
a \+ I
1“ ......
x \
I «T
D \
.1 .+
a \
(TM?

5 n o: s “.38.. .25 noouxlﬁs 2 3 .883:

noouxaﬁa o...

0.v

 

 

0.

0N

0n

(0 '11“! 5—U1w/7Ho Iw)

HIVH NOIIOVEll-l

151

pretreatment, indicating that catalyst is either not activated in helium or only

activated as carbon is consumed.

5.5.3.2 Potassium Catalyst Loss

 

The results of potassium catalyst loss measurements (Figures 5.11 and 5.12)
indicate that catalyst loss by evaporation out of the reaction zone is relatively
slow at 865 ° C in hydrogen. However, catalyst loss during gasiﬁcation by
migration onto the sample holder and by evaporation and condensation onto the
quartz insert is signiﬁcant, as indicated by the results presented in Figure 5.12

for catalyst loading measurements following gasiﬁcation.

The loss of catalyst is apparently associated with the extent of gasiﬁcation,
and not merely with time of exposure to high temperature. This was illustrated
by the experiment in which a 10% loaded sample was held in ﬂowing helium at
865° C for three hours. Catalyst loss in this experiment was only 4% (2.4 mg
K/g initial carbon) after the extended exposure. The catalyst loss during the
gasiﬁcation which followed, however, was quite similar to that for samples not
heated in helium. The ﬁnal catalyst loading is given by the open circle in Figure
5.12. Sams et al. [157] have concluded that KQCO3 is not lost from carbon
sample appreciably below its melting point of 891 ° C as long as it is in the form
of carbonate, and that the formation of free potassium metal is a prerequisite for
rapid loss of the catalyst. That is, the interaction of K2CO3 with carbon surface
is necessary for the reduction and subsequent loss of the catalyst. Thus, it seems
that on surface of fresh carbon insufficient number of active sites are present for
substantial interaction with K2CO3 and the catalyst remains mostly in form of
carbonate, or the interaction with carbon surface forms potassium containing
complexes that stablize the potassium on the surface. At higher conversions,

however, more surface area is developed and more active sites are available, so

152

that more catalyst is activated and is lost from the sample. This subject will be

discussed in more detail in the following Section.

5.5.3.3 Eject of Treatments on Catalysis

 

The interpretation of the results relevant to the role of oxygen groups in the
alkali-catalyzed gasiﬁcation are more conclusive if the catalyst losses and the
catalyst activities are studied together. Some of earlier results are also

reinterpreted within the framework of the ﬁndings of this Section.

The comparison of potassium loadings in Figures 5.13 and 5.14 shows that
a degassed carbon loses potassium much faster than a fresh carbon sample, and
that oxidation of the surface at 400° C improves the potassium retention ability
of the carbon to some extent. This improvement is observed throughout
gasiﬁcation although the rate of the catalyst loss is essentially the same after
20% conversion. The corresponding reaction rates in Figure 5.8, however, show
a slight decrease following oxidation. Both the added catalyst retention and
lower rate suggest the deactivation of a portion of the catalyst via formation of
stable surface compound(s) of potassium upon oxidation of the surface. Harker
[170] has attributed the catalytic activity of alkali-metal salt in oxidizing
atmospheres to its ability to destroy the acidic surface oxides by
neutralization/decarboxylation or attack by free metal. So at this point it is not
possible to speculate on the nature of these stable and inactive compound(s).
The catalyst that did not interact with these particular oxygen groups remained

active.

The results look somewhat different at 47% and 65% conversion. The
oxidation did not seem to improve the catalyst retention ability of the carbon,
but it did enhance the reaction rates as can be seen in Figures 5.9 and 5.10. In

these cases potassium was lost rapidly from the carbon samples and its level was

153

reduced to that of a 10 wt%-degassed sample when gasiﬁed to the same
conversion (open triangles and circles around 80% conversion in Figure 5.13;
crosses and asterisks above 80% conversions in Figure 5.14). Apparently, at
high conversions the stable inactive surface compounds of potassium were not
formed, at least to the same extent as at low conversion. Instead, some other
intermediate species were formed that were active. The higher activity at the
same potassium content in these cases as compared to the degassed samples
suggests that either the potassium complexes formed on the surface were
different or the portion of the potassium lost from the surface was involved in a
mechanism that created more active sites. In the former case, the observation

that potassium contents were the same then is just a coincidence.

In either case, the implication is that the carbon surface structure is
different at these higher conversions. The surface structure is different due to
the fact that signiﬁcant porosity develops as carbon is gasiﬁed. The drastic
increase in surface area during gasiﬁcation, from 15—20 m2/ g in fresh carbon to
about 400 mQ/g at 60% conversion (Figure 3.1), is much too high to be
accounted for by a simple reduction in the carbon particle diameter, and
porosity development is the most plausible explanation. These different
structures very likely were also responsible for the lower enhancement of the
reaction rates of the uncatalyzed carbon samples when oxidized at higher
conversions (Figures 4.6 and 4.7). As discussed earlier (Section 3.3.8), oxygen
groups already present in this particular carbon were different on the surface
than in the interior; this could be due to the difference of the carbon structure
when they Were formed. The oxygen surface groups formed upon oxidation are
thus most likely different at the higher conversions. This is supported by the pH
measurements. The pH value of a suspension of a carbon sample that was

oxidized at 400° C for two hours after 42% gasiﬁcation was slightly less acidic

154

than that of a fresh oxidized carbon (Table 3.5). However, the surface area of
this sample is about 14 times larger than the fresh sample, and it is expected
that if the same acidic groups were formed to the same extents on both surfaces
then the pH of the suspensions would be at least one unit lower in the former
case. This is not observed. The reason is probably the existence of still large
amounts of the stable basic groups that did not oxidize during the oxidation
period of two hours. The acidic groups on the surface formed during oxidation
at 400 ° C coexist with basic groups that were already on the surface, thus giving
the slightly acidic pH.

Some of the catalyst added at high conversions interacted with the basic
groups and formed stable species which retained potassium to the end, similar to
a 10 wt%-fresh degassed sample (Figures 5.13 and 5.14). The observation that
the potassium content of a “10 wt%”-47% gasiﬁed—carbon sample after 3 hours
of heating in helium at 865° C (lower open circle at 47% conversion in Figure
5.13), was the same as a 10 wt%-degassed sample after 47% gasiﬁcation
supports this idea that stable potassium surface complexes are formed with the
basic groups, not with the groups introduced via oxidation at 400° C at these
high conversions. The non-basic groups, on the other band, formed catalyst
species that were not stable but were active and were, in addition, different from
those formed at the lower conversions. These species were responsible for the
higher reaction rates of these samples. It should be noted that this enhancement
is too large to be accounted for by just the enhancement of the uncatalyzed
gasiﬁcation and it indicates that the catalyst is deﬁnitely involved in this

process.

The basic oxygen groups did not enhance the reactivity of the uncatalyzed
carbon signiﬁcantly, but it seems that in the presence of potassium carbonate

they play an important role in the catalysis process. Perhaps, potassium

155

carbonate interaction with these groups destabilizes and desorbs the oxygen
groups, generating active sites by a process similar to that of the non-basic

oxygen groups in the uncatalyzed case.

It has already been shown that the reaction rate per unit surface area
(speciﬁc reaction rate) of this particular carbon black maintains a fairly constant
value after 20% conversion for 5 and 10 wt%-KQCO3 fresh carbon samples
(Section 5.5.3.1). This can be explained in terms of interaction of the catalyst
and the basic oxygen surface groups. As carbon is gasiﬁed and higher surface
area is developed, more basic oxygen groups within the carbon are exposed. The
newly exposed oxygen groups interact with the catalyst present on the surface
and form the active intermediate species. Thus as more surface area is
developed, active sites are generated and the speciﬁc reaction rates remain

constant.

Fresh carbon samples that were not degassed or gasiﬁed prior to
impregnation showed higher activity and better retention of potassium. As
discussed earlier, the oxygen groups near the surface of this particular carbon
were mainly C=O carbonyl groups that were lost upon degassing or gasiﬁcation
to high conversions. So it seems reasonable that the presence of these carbonyl
groups and the higher reactivities are somewhat related. The interaction of the
alkali carbonate and these groups to form a potassium surface oxide (C-O-K)
deﬁnitely is a possibility. This surface oxide has already been shown to be stable
at temperatures as high as 700° C and is directly proportional the reactivity of
char samples in steam gasiﬁcation [12]. Further evidence for the importance of
the initially present non-basic oxygen groups is provided by careful comparison
of Figures 5.13 and 5.18. The reaction rates of a 10 wt%-degassed carbon
sample are very close to those of the 5 wt%-sample (Figure 5.18). However, the

potassium content of the degassed sample is much higher up to 50% conversions,

156

.2228 Ego—-moog “0333.93; .00 O . mom .3 mega.— aowemommmaw 35330 wﬁm charm

E zo_mmu>zoo

 

 

on on on 9. on om o. o
HHVHVV.I1I.I.I.I....I.I.I_I.I..HH....I.I .. ......thwwﬂ .oldl.“
m

U

L

.A

l

I

I

4

L

x

L

._ n o: s 3.8; noouxukt. o... a .U

809:5; o... .. .

noouxuﬁs ml . U

0009753 or 3 nouocoeaEm :05 commence! a 1

 

 

 

0F

09

0N

nu

BlVH NOLLOVBH BALLV'BH

157

as can be seen in Figure 5.13. The immediate conclusion is that a large fraction
of the catalyst is inactive in the degassed sample, because of the absence of the

carbonyl groups.

When the severity of the heat treatment is reduced by heating an
impregnated sample at 865° C for 3 h in helium, the reduction in reactivity is
much less than the degassed sample (Figure 5.18). In this case the potassium
loss is only 4% at the end of the heat treatment, which is not enough to account
for the relatively large drop in reaction rate from that of the fresh impregnated
carbon. In this less severe treatment fewer oxygen groups are lost and thus the
reduction in reaction rate is less. It is worth mentioning that the potassium
content in this case was between that of the fresh and the degassed samples

throughout the reaction period.

The amount of oxygen lost as CO and 002 from the carbon samples during
degassing at 1000° C was about four times higher than that during heating for
three hours in helium at 865 ° C. But the loss in reactivity was not proportional
to the oxygen losses (Figure 5.18). At present, this cannot be satisfactorily be
explained, but a few speculations can be made. One speculation is that when the
carbon sample is cleaned of most of its oxygen by degassing, the oxygen in the
carbonate plays a larger role in the catalysis process. Another speculation is
that a signiﬁcant portion of the oxygen groups that were removed from the
sample by degassing were inactive groups, probably ether groups. In either case,
it can be said with conﬁdence that not all the oxygen present in the carbon is
involved in the catalysis process of the alkali carbonate, somewhat contrary to
the results of some studies that claim the reactivity of a carbon sample is
directly related to its oxygen content [27] or the oxygen trapped on its surface
[156]. The involvement of other active sites and the oxygen in the carbonate in

the catalysis process have already been proposed in the literature [159]. As

158

stated earlier, the results of the present study do not rule out the possibility of
involvement of other active site that are not necessarily associated with oxygen.
These results nevertheless suggest that most of the activities are associated with

the oxygen groups.

The fact that surface oxygen groups play a signiﬁcant role in K2003-
catalysis of both steam gasiﬁcation (oxidizing environment) and
hydrogasiﬁcation (reducing environment) strongly suggests that the dominating
steps in the catalysis process are the interactions between the catalyst and
carbon. The enhancement of reaction rates by oxygen surface groups are greater
in the catalyzed case, implying that oxygen inﬂuences the interaction of KQCO3
with either carbon or reactant gases. If oxygen inﬂuences the interaction of the
catalyst with carbon, then enhancement of the reactivity should be independent

of the gaseous environment.

In contrast, if oxygen were to inﬂuence the interaction of the catalyst with
reactant gas, then oxygen would have to interact with the catalyst or the
reactant gas ﬁrst. The interaction of reactant gas with intermediate species
formed from interaction of the catalyst and oxygen surface groups would then be
strongly dependent on the nature of the reactant gas; i.e., if this intermediate
were more reactive toward oxidizing gases, it would be less reactive toward
reducing gases and vice versa. This is not observed; both steam and H2 gas
gasiﬁcation rates are enhanced by presence of oxygen. On the other hand, if
oxygen surface groups interact with reactant gas ﬁrst to form surface
complex(es) which further react with the catalyst, this complex would be
expected to be similar to oxygen-hydrogen complexes formed during steam
gasiﬁcation. In this case this complex would readily decompose to CO and H2
and thus methane formation rate would not be enhanced. Again, this is not

observed.

159

Hence, it seems that oxygen plays a signiﬁcant role in the interaction of the
catalyst with carbon in KQCO3-catalyzed gasiﬁcation of carbon. This suggests
that the study of oxygen inﬂuence on KQCO3-catalysis of hydrogasiﬁcation is a
better choice to probe the intermediate steps because the sources of oxygen are
limited to carbonate itself and oxygen impurities in the carbon. In steam or
CO2 gasiﬁcation, there is a constant supply of oxygen to the carbon surface from
the reactant gases and, therefore, it is more difficult to control the formation and
the nature of oxygen groups thus formed and to follow oxygen through reaction

sequences.

Chapter 6‘

Conclusions 8' Recommendations

The results of this study lead to some important conclusions on the role
of oxygen surface complexes on alkali-catalyzed hydrogasiﬁcation of carbon.
However, there are still many unanswered questions, and some recommendations

are thus made within the framework of the present study to resolve some of the

relevant questions.

160

161

6.1 Conclusions

6.1.1 Uncatalyzed Hydrogasiﬁcation

Based on the results of the study of inﬂuence of various oxygen complexes
on hydrogasiﬁcation rate of carbon black (in the absence of potassium
carbonate) a mechanism is postulated. According to this mechanism, oxygen
surface groups enhance the reactivity of the carbon toward hydrogen by
generating ‘nascent’ active sites upon thermal decomposition into CO and CO2,
and not by the “electron transfer” mechanism. These ‘nascent’ active sites act

as the sites for hydrogenation of the carbon surface.

It was found that oxygen complexes near surface of fresh and untreated
carbon black Raven 16 are mainly carbonyl groups, and that oxygen complexes
in the bulk of the carbon are dominantly basic. Carbonyl groups and acidic
groups readily desorb from the surface and create the ‘nascent’ active sites and
therefore are very active. In contrast, the basic groups, possibly chromene

groups, are fairly stable and thus inactive.

As higher conversions are obtained, the carbonyl groups are lost from the
carbon surface and thus many fewer active sites are generated, resulting in a
sharp drop in the activity of the carbon surface. On the other hand, the total
surface area of carbon particles increases drastically due to the development of a
high degree of porosity, thus exposing the basic oxygen complexes in the carbon
bulk. The low activity of the basic oxygen complexes and the high surface area
together compensate for the effect of loss of carbonyl groups, hence the slow
decrease of overall hydrogasiﬁcation rate. In addition, prolong degassing desorbs
the carbonyl groups and thermally anneals the ‘nascent’ active sites, resulting in

signiﬁcant loss in reactivity of carbon.

162

6.1.2 Alkali-Catalyzed Hydrogasiﬁcation

Potassium carbonate is an effective catalyst (enhanced reaction rate 25
times for the 10 wt%-K2003 sample). However, there seems to be a minimum
amount of potassium carbonate, between 1 wt% and 5 wt%, needed before any
catalytic activity can be observed. The interaction of potassium carbonate, in
excess of this minimum amount, with carbon involves a small fraction of the
carbon surface area which reduces the carbonate to free alkali metal. Subsequent
formation of active species generates active sites on the carbon surface in such a
way that keeps the activity per unit surface area fairly constant. In the absence
of K3003, the carbon surface is deactivated rapidly due to the thermal annealing
and consumption of active sites. A signiﬁcant portion of the potassium is lost
during hydrogasiﬁcation from carbon bed via vaporization (free potassium
metal) or migration onto sample holder surface, but is retained in the reaction
zone (on the surface of sample holder, quartz reactor lining or the tip of the exit

tube)

The catalyst interacts with different oxygen surface groups to form
intermediate species which vary in terms of activity and stability. The results
lead to the conclusion that not all oxygen groups are involved in the catalysis
process. The oxygen surface complexes that do participate in the catalysis have
different activities depending on their natures. The interaction with 0:0
carbonyl groups forms stable and highly active species which are possible
candidates for the formation of C-O-K surface groups. The basic oxygen surface
complexes, on the other hand, interact with the catalyst and produce relatively
unstable species that enhance hydrogen gasiﬁcation. It is this formation of
active species from basic oxygen groups, otherwise inactive, that is responsible

for signiﬁcant enhancement of hydrogenation rate at higher conversions.

163

In contrast, the interaction of acidic oxygen groups with the catalyst
could not be well characterized and seems to depend on the conversion level or
the structure of carbon surface. Probably, at different conversion levels,
combustion ﬁxes different proportions of carbonyl groups and various acidic
groups which interact distinctively with the catalyst, resulting in a variety of

observed hydrogenation rates.

Furthermore, potassium retention on the carbon surface and catalyst
activation seem to be two distinctive phenomena. It is possible to retain

potassium on the carbon surface without activating it.

Presence of K2003 does not seem to alter the course of the porosity
development within carbon particles during hydrogasiﬁcation, and the impurities

such as sulfur and nitrogen do not seem to play any role in the catalysis.

6.2 Recommendations

There is some degree of speculation involved in the interpretation of the
results reported in the present work, and further veriﬁcation using other
techniques is advisable. Here, it has been shown that, ﬁrst, oxygen plays an
important role in alkali-catalysis and, secondly, various oxygen groups on carbon
surface have distinctive interactions with the carbon surface. These results
provide a rather unique means for further studies of the inﬂuence of oxygen on

alkali-catalysis.

Various oxygen surface groups should be introduced onto carbon surface
by various techniques such as combustion or oxidation by nitric acid, followed
by impregnation with potassium carbonate. These samples should then be

heated in helium to reaction temperature. During the heat up, the carbonate

164

and oxygen surface groups will interact and form some intermediate species.
The nature of these intermediate species can be probed by adapting a variety of
techniques such as XPS and FTIR. Since intermediate species formed from
different oxygen groups have different activities, these techniques should identify
the active species and eventually reveal more information about the catalysis
mechanism. These experiments should also be carried out in hydrogen at
various levels of conversion to examine the effect of hydrogen on the interaction
and also the eﬁect of change in carbon surface structure. However, extreme care
should be taken to avoid exposing samples to air before analysis, for after
reaction the samples may readily interact with air even at room temperature
(Section 5.5.2.1.1) to form oxygen surface groups which might interfere with the

analysis of surface groups.

A knowledge of the nature of the most active intermediate species and the
chemistry involved could perhaps lead to the development of a new processing
method of coal or new additives so that the oxygen present in coal can be put
into a more active form. Economic feasibility of such a processing cannot be
evaluated at this point simply because it is not yet known what is involved and

exactly what needs to be done.

Another recommendation is to investigate the phenomenon discussed in
section 5.5.2.1.1 where some rather active intermediate species, which
spontaneously react with air at room temperature, seem to be formed at
moderate conversions. If this phenomenon is understood, perhaps after the
initial high temperature gasiﬁcation, the rest of the carbon can be gasiﬁed in
steam (oxidizing environment) at much lower temperatures so that methane

formation is more favored.

Appendix A

Appendix A

Reactor Design Calculations

Flange Design

The thickness of the ﬂange ring can be calculated from Equation A-l [69]

 

g2 = — 3P 2(s+1)(A’ —an)+(s —l)A—”— —(s+3) for U = 1 UV”
I 4W80'9 U2
where
12 l+3s
’ = 1K
K12- " 1+ 20+.)
21+ K2

 

= (s-lxxf -1)

L = Poisson ’3 Ratio [1
s

and

P = effective load, lb ,

165

166 Appendix A

09 = tangential stress in plain ring, psi

r1 = radius of bolt-circle, in

r0 = inside radius of ﬂange ring, in

g, = thickness of ﬂange ring, in
weakest spot of the ﬂange ring; i.e. , the spot at which tangential stress becomes equal
to the yield strength of the material. Inside radius r0 was taken to be equal to inside
radius of the cylindrical reactor vessel which is 9/16". The bolt-circle radius was taken
to be the distance from the center of the ﬂange ring to the outer edge of the bolt holes.
The 0'0 was set equal to yield stress of the alloy at room temperature which is 63000
psi [70]. The value of Poisson’s ratio [1 could not be found for this particular alloy and
a value of 0.3 was used; the range of Poisson’s ratio is limited between -1 and 95 [66]
and varies from 0.25 to 0.42 for most metals and alloys [71]. The effective load P is
the sum of gasket sealing stress, the weight of the reactor vessel and the stress due to

internal hydrostatic pressure. The sealing stress can easily be calculated

Sealing stress = O-ring perimeter X sealing stress per inch of O—ring
= 13/8" X 71' X 1001b,
= 511 lb f

The stress due to internal hydrostatic pressure is calculated from
Internal pressure stress = O-ring inside area X inside pressure
= (1%”/2)2 x 1r x 1100 psi
= 1350 lb 1
where a safety factor of 10% is included for the operating pressure inside the reactor.
The weight of the reactor can be estimated by assuming a solid rod of 2" OD X 40"

long as the reactor

weight stress = solid rod volume X alloy density

167 Appendix A

= (2'72)2 x 1r x 40" x 0.3301bm/in3
= 41.5 lb ,

where the density is 0.330 lbm/in3 [70]. The sum P of all of the above stresses is equal
to 1900 lb, and is a fairly conservative approximation. Substitution of the appropriate

values into Equation A-l yield the required thickness of ﬂange ring:

 

 

 

 

= 7 4" _—
1 7/16"
1
1+3(—
2
A' = 4 ln4+ 0'3
42—1 2(1+-—1—
0.3
=2.75
1 2
2(1+ —)4
A" = l 0'3 ln4
—-1 42—1
(0.3 )( )
=5.49
3(19001b
2: f) 2—1— 1275-! 1 —1——1 5'49 —1— 3
c; ( + )(. n )+(0.3 )(12H0.3 +)

41r(6j—3-)(63000 psi)

= 0.0655 in2

a 3‘, =0.256"

168 Appendix A

The thickness of the ﬂange ring is not very sensitive to the magnitude of
Poisson’s ratio 11; the thickness varies from 0.244" to 0.267" when Poisson’s ratio is
varied from 0.1 to 0.5, respectively. Thus it is decided to use a plain ring ﬂange of 95"
thickness, which is operated quite safely after introducing several conservative

assumptions discussed above.

The deﬂection of the ﬂange ring under the applied stress is calculated from

Equation A—2 [69]

_23(8 —1)Prg

21r32E

 

[A' (K,2 —1)—(K,2 A" yum] (A-2)

where E is Young’s modulus of elasticity and y is deﬂection of ﬂange ring.

The Young’s modulus of elasticity E for Haynes-25 alloy at room temperature is

34.2X106 psi [70]. Substitution of appropriate values into Equation A-2 yields

_3((—)—1)woow, ( " 77,)”

 

2. 75(42 —1) — (42 —5. 49)ln4]
=2111-632(34.2><106 p8: )(W‘ )3

= 0.00098"

The thickness and the deﬂection under stress of the cover disk is calculated
similarly using Equations A-1 and A—2; but, in this case the diameter of the hole at the
center of the disk, for mounting a Conax ﬁtting needed for insertion of thermocouples
into sample bed, is only 56". The yield strength and Young’s modulus of elasticity for
stainless steel 316 are 30000 psi and 28>(106 psi, respectively [70]. The calculated disk
thickness is 0.611". A l"-thick cover disk is used for which the deﬂection under stress is

found to be 0.00014".

169 Appendix A

Thermal Stresses

Temperature Proﬁle

 

The temperature proﬁle can be derived by solving Equation 2-5 subject to two

boundary conditions

d dT
_ = o _
7;» z) (A 3)
T 2 room temperature Tm, 77 °F at z = 0
T = furnace temperature T, at z = L

The value of L is set to 9" and Equation A-3 is solved by integrating both sides of the

equation w.r.t. z to yield

K— = q (A-4)

where q is the constant of integration and by deﬁnition is the negative of axial heat
ﬂux in the reactor wall. The thermal conductivity n as a function of T [64] is

substituted in the above equation and both sides are integrated again

170 Appendix A

r, L

rch =f qdz (A-5)

0

a“!

where

r: = 59.06 + 0.0761 T, Btu-in/(ftz-h- ° F)

and solving for q yields

59.06( T, )+
q = L

(If—7'73)

 

0.0761
2 03-6)

 

For a furnace temperature T! of 1832 °F (1000° C) the ﬁnal temperature proﬁle,

the temperature gradient and the heat ﬂux q are

T = —776+ 852.9 ‘\ /l+8.35% (A—7)

dT = 3561

 

 

dz (A-8)
‘\ / .353—
1+8 L
2.307 105
q = —3<— (A-9)

L

respectively.
For a cooling length of 9" the temperature gradient varies from a maximum of
395° F/in at the cooling coils site (2:0) to 129 °F/in at the furnace inlet (z=L) and

the heat ﬂux q is 178 Btu/in2-h.

171 Appendix A

Shear Strain

 

The thermal expansion coeﬂicient a by deﬁnition is

a = —— (A-lO)

where V is the volume of the control volume under study. Consider a cylindrical shell
of a very small thickness Ar and length A2 with radius r and located at point 2 along
the reactor z-axis as the control volume (cylindrical coordinate). The volume of this

control element is then 27rr ArAz. Substitution into Equation A-lO then gives

 

_ 1 dr 1 dAr 1 dAz
a—rdT+ArdT+Asz (All)
In cartesian coordinates with a control volume of A13, Equation A-lO reduces to
3 dAx
= __ A-12
0' A: dT ( )

Thus if it is assumed that thermal expansion in the z direction is not inﬂuenced by
thermal expansion in other directions, then the last term on the right hand side of
Equation A-ll can be replaced by above equation. Since the reactor wall is thick it can
also be assumed that the ﬁrst two terms on the right hand side are equal in magnitude

(assume a solid cylinder). Then Equation A- 11 yields

ﬂ = If. (A-13)
dT 3

 

172 Appendix A

Therefore, the shear strain can be estimated by substituting Equation A-8 and

the above equation into Equation 2-4

 

— = 3 (A-l4)

From this equation it can be seen that the points of maximum strain are the points on
the outside of the reactor wall at the site of cooling coils (ro,0). At these points the
modulus of elasticity G is also a maximum and because E is a maximum over the

temperature range of interest and u is fairly constant with temperature, [73]

E

= 2(1 + p) (A45)

Cooling Coils

 

The heat ﬂux inside the reactor wall calculated from Equation A-9 is 178 Btu/in2-h. The heat
duty Q of the cooling coils is calculated by multiplying the heat ﬂux q and the cross sectional

area Am, of the reactor wall:

 

 

Q = qXAm (A-16)
Btu 2" 2 7Z8" 2
= 178 ——
= 452.919.

h

173 Appendix A

The cooling water ﬂow rate is then calculated from an energy balance on cooling
water with heat capacity C", and mass ﬂow rate m and a temperature rise AT, of

10°F:

 

Q
m = —— (A—17
0,,AT, )
_ 452 Btu/h
(1 Btu/lbm— 'F) x 10°F
lb
= 49.7 J—
h

where heat capacity of cooling water is taken as 1 Btu/lbm- ° F.

Mass ﬂux ﬁt inside the cooling coils is then calculated by dividing the mass ﬂow

rate m by the internal cross sectional area Ac of 54" OD copper cooling coils

 

° m
m = A’18
4c ( )

__ 49.7 1b,, /h
" 5732"
1r( 2 )2
= 2592 lb,,, /in2—h

where inside diameter d of copper cooling coils is 5/32". The Reynolds number R, is

then calculated from the following equation

R, = d’" (A-19)
uu

 

_ 5/32 in x 259210m/z'n2—h
— 0.1916 15,, /in —h

 

174 Appendix A

= 2114

where the viscosity of the cooling water 11,, at 75° F is 0.95 centipoise or 0.1916 lbm/in-
h.

This value of Re is in the region between laminar and turbulent regimes [74].
However, this required cooling water velocity inside the cooling coils is only about 1.65
ft/s which is much lower than what is obtained in practice and thus the actual value of
Re during operation is somewhat higher than the above value. Hence, the ﬂow regime
can be assumed to be turbulent so that a more realistic estimation of the heat transfer
coefficient he can be obtained. The heat transfer coefficient for turbulent ﬂows inside a

circular tube is calculated from the following equation [74]

0.7 /3
d 0.3 Cr ”v _
”[1433] IR, [ n )1 (A 20)

C

 

RC
ho = 0.0237—

 

where L“ is the length of cooling coils and Kc is the thermal conductivity of cooling
water which is 0.349 Btu/ft-h- ° F at 75 ° F. Generally, a velocity of 3 ft/s for the ﬂuid
velocities in heat exchangers is assumed and thus this value is used for cooling water
velocity inside cooling coils. This yields values of 4710 lbm/inz-h and 3840 for ti: and
Re, respectively; the temperature rise of cooling water is reduced from 10° F to 55°F.
As a ﬁrst approximation, assume a ratio Lcc / d of greater than 50, so that the ﬁrst

bracket in the above equation can be ignored. Substituting appropriate values into the

equation to ﬁnd ho;

 

 

 

 

 

' 3 lb “/3
Btu 1—"‘—— 2.30 '"
h, = 0.023 ,, (3840) -
(5_)(.__1 f‘) 0.349 ————B"‘
32 12" , ft—h—°F ,

175 Appendix A

Btu

= 852
ftz—h— “F

 

The temperature difference driving force AT in turbulent regime is taken to be

equal to the log-mean temperature differences AT 1,, [74]

(TI-Ti) _(T-To)
T— i
T—T,

AT”:

 

(A-21)

 

= (85 'F—75.5 'F)—(85 'F—70 'F)
1n(85 'F—75.5 °F)
(85 'F—70 'F)

 

 

=12?

where it is assumed that the temperature of cooling coil wall is constant and uniform
even though heat is transferred only from the side that is soldered to the reactor wall
and not from the sides that are exposed to air; i.e., all of the tube surface area of
cooling coils is available for heat transfer. This assumption is valid since copper has
very high thermal conductivity and, in addition, the copper tubes used were thick-

walled.

The total heat transfer surface area A is then calculated from Equation A-22

A = Q A-22
hATln ( )

452 Btu/h
(852 Btu /ft2—h— ‘F) x (12 °F)

 

= 0.0442 1:2

176 Appendix A

The length of cooling coils L,, is then calculated from Equation A-23

-4 _
L... - 1,, (A23)

00M2ft

1r meg—rig—

 

=L%ft

This length of cooling coil corresponds to about 2 laps of copper tubing around the

outside of the reactor. It is decided to use 6 laps of cooling coils.

After the construction of the reactor the actual cooling water velocity was
measured and was found to be about 2 ft/s as opposed to the value of 3 ft/s that is
used in Equation A—20. Thus the actual value of R, is 2560 and the corresponding
temperature rise of cooling water is about 8.3 ° F. This yields a value of 616 Btu/ftz-h-
°F for h, from Equation A-20 and 10.3°F for AT 1,, from Equation A-2l. Values of
0.071 ft2 and 1.74 ft are obtained for A and L,,, respectively. Thus only 3.5 laps of

copper cooling coils are needed for the actual operation.

_—__J

177 Appendix A

’y-Ray Detection System

q-Ray Absorption

 

There are 6" of ﬁre-bricks, two iron sheets of 1/16" thickness, 9/16" of Haynes-25
alloy, 0.5" hydrogen gas and 6" of air between the detector crystal and a sample at the
reaction site inside the reactor. The detector is “focused” on the reaction site by
inserting at least 6" of lead bricks between the detector crystal and the areas around
the sample (Figure 2.3). The extent of absorption of the ’1—rays by these shielding
materials is calculated from the attenuation coeﬂicients A in Table 2.3 and Equation 2-

20 which can be rearranged to

L = e'A’ (A-24)
Io

Substitution of A’s from Table 2.3 and above mentioned thicknesses into the

above equation yields

-(0.380 cm’1)(2%€-"i)(2X1/16")

‘1— liron = e
10

= 0.886

2.54 cm ..

_I_ —(0.025 cm'1)( 1,, )(6)

lo lfirc—bricl: = C

178 Appendix A

 

 

= 0.683
-(0.416 cm )( 2 “1H°"°)(9/16 )
‘1'- lHayncs—25 = e
0
= 0.552
1 -(o.578 ...—1x2 5:. )<°——'1'—6 )
f llead = e
= 0.000135
1 —(0. 0000513 cm“)( 2' 5°, °"')(0" )
_ lair = e
[a
= 1.001

Above calculations show that ’7-ray absorption by air or hydrogen is negligible
and 6" of lead shielding is eﬂective for “focusing” the detector. The overall absorption
by shieldings between the detector and the sample can now be calculated by

rearranging Equation 2-26 and using Equation 2-27:

I I
fl = I— lovcrall = T liron>< %lfirc— brickx— {blames-25 (A‘25)
a o
= 0.886 X 0.683 X 0.552

= 0.334

179 Appendix A

Thus only 33.4% of the ’y-rays emitted from a sample at the reaction site inside
the reactor will pass through the shielding materials between the detector crystal and

the sample.

Sample Activit}!

 

The radiation rate I, from a sample or nuclear disintegration rate R is then
calculated by taking into account the geometric factor [4 (Eq. 2-28), the detector
efficiency 1]) for '7-rays and the fraction [1 of the radiation passing through the shielding
materials. The desired counting rate is ten times the background counting rate of 5-6

cps and thus the disintegration rate R is calculated by modifying Equation 2-29:

— _ 60 cps

’ 0.1891711 (A26)

The factor 0.18 in the denominator is to account for the fact that only 18% of the
nuclear disintegrations yield 1.525 MeV '7-rays. Substitution of the appropriate values
of f d from Equation 2-28, I, from Equation A-25 and a value of 5% for the detector

eﬂiciency in the above equation yields

 

R: 60 cps
0.18X0.05X0.0051X0.334

= 3.91x106 cps

= 106 110i

180 Appendix A

Sample Activation

 

Once the desired level of radiation and the amount of potassium in a sample are
known, the activation time t, is calculated from Equations 2-13 and 2-14. The
maximum allowable level of activity for a sample however is limited to only 100 uCi by
Oﬂice of Radiation, Chemical and Biological Safety at Michigan State University. Thus

for 60 mg of a 10 wt%-K2CO3 sample the number of atoms of 41K is

 

41
0.060 gXlO%x 0-565 9 K X0069 9 K
N41 = 1 g carbonate 1 9 K N
o 600

41 g 41X
1 mol 41K

 

= 3.44X1018 atoms

Combining Equations 2-13 and 2-14 and solving for t, yields

t, = — —ln [1— ————] (A-27)

For a Neutron ﬂux <I> of 1012 and a cross section 77 of 1.5)(10—24 cm2, above

equation yields

1 1n 1_ 3.91x10° cps
0.056 11'1 (3.44X1018 atoms )(1.5X10"24 cm2/ atom )(1012neutrons/ cm2—s)

ta:

= 25.3 h

181 Appendix A

For such small samples a neutron ﬂux of 1013 is used instead which reduces the

activation time to 1.41 h or 84 min.

182 Appendix A

Rupture Disc Sizing

To size the rupture disc it is necessary to determine if the ﬂow through the
ruptured disc is sonic or subsonic and then the gas ﬂow rate Q, (SCFM) and area A“,

is calculated [82]. The ﬂow through ruptured disc is sonic if

Bid—l k+1 (A28)

where k is the ratio of speciﬁc heats and p, is the exhaust pressure [82]. For hydrogen
gas I: is equal to 1.41 [82] and if the exhaust pressure p, is taken to be atmospheric, for

a stamped burst pressure of 1450 psi the ﬂow will be sonic ﬂow:

1.41

2 1.41 — 1
—— = 0.527
1.41 + l

The gas ﬂow constant C for sonic ﬂow is then calculated from the following

M— _.00101 <
1450 psi

 

equation [82]

k+1

-1
___ _l" A-29
0 3—2—5600 k 1 ( )

1.41+1 _
2 1.41- 1
l.41+ l

 

183 Appendix A

= 0.09915

For sonic ﬂow the area A", and volumetric ﬂow rate Q, are related to each other

[82]

 

A _ 9.
rd ‘—
22772Kp0

\fWH( Tm + 460) (A-30)
where K is a constant set to 0.62 by ASME Code regarding rupture discs and W” is

molecular weight of hydrogen gas. Substitution of appropriate values into the above

equation yields

 

 

Q?) g
A = 2.016 — 77 “F 460
" 22772X0.62X1450 ¢x0.09915 V mole ( + l

A“, = 1.02x10-5 Q,

The smallest size Fike Prebulged rupture disc is '12" in diameter and the
maximum gas ﬂow rate through this rupture disc is then calculated from the above

equation

Qu—

_ ___ (4-31)
1.62X10’5

= 12120 SCFM

184 Appendix A

Thus for a full 250 ft3 and 2500 psi hydrogen gas cylinder it takes 1.5 minutes to
reduce the gas cylinder pressure to 1450 psi by venting through the rupture disc to
atmosphere, assuming that the 1/8" tubing upstream from the rupture disc does not
limit the ﬂow from the cylinder to the rupture disc. In reality, in case of pressure
regulator failure or misadjustment, the ﬂow in the 1/8" tubing will be a sonic ﬂow and
will be less than the sonic ﬂow through the rupture disc. This implies that the reactor

will be safe from any over pressurizing by gas cylinders.

For Fike Conventional Prebulged rupture discs the recommended ratio of
operating pressure to burst pressure is 70% [82]. This implies that for an operating
pressure of 1000 psi the rupture disc burst pressure should be at least 1430 psi. On the
other hand, the ASME Code for the case where a rupture disc device is used as the
primary or sole relieving (as is the present case) requires that rupture disc burst
pressure not to exceed the maximum allowable working pressure (MAWP) [82]. Thus
the reactor should be operated at a temperature such that MAWP is higher than the
burst pressure; i.e., MAWP should be at least 1448 psi which is the stamped burst

pressure of the purchased 95" rupture disc.

The reactor is originally designed for a MAWP of 1100 psi at 1000°C using
Barlow’s formula (Equation 2-1). When MAWP is set to 1448 psi, the operating
temperature of the reactor vessel should be reduced to a limit where the rupture disc
fractures before the reactor is damaged. To this end, the normal tensile and
compressive stresses in the reactor wall have to be calculated for an internal presure p
of 1450 psi and then be coupled with thermal shear stresses. The resulting overall
maximum normal stresses am”, calculated from Mohr’s circles will determine the

maximum allowable operating temperature.

The normal tensile and compressive stresses are calculated from Equations 2-7
through 2-9. The thermal shear stresses are assumed to have the same magnitude as
those at 1000 ° C operating temperature (Table 2.1). The Mohr’s circle results indicate
that on,“ is equal to 2136 psi and 1737 psi at (r,-,L) and (r,,L ), respectively. The
rupture stress ”rupture in 10000 h is about 2140 psi at 970°C and 2650 psi at 950° C

185 Appendix A

[64]. Therefore, an operating temperature of 950 ° C or lower is quite safe.

This temperature is very close to the original designed temperature of 1000 ° C for
a MAWP of 1100 psi. The reason is that the original design was based on Barlow’s
design formula (Equation 2-1) which is more on the safe side than the Lamé’s formula

which is used for the above analysis.

LIST OF REFERENCES

ll]

[2]
[3]
[4]
[5]
[6]
l7]
l8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]

[16]

[17]

LIST OF REFERENCES

Seglin, L., R. Geosits, B. R. Franko, and G. Gruber, ACS Adv. Chem. Ser. 146,
1 (1975).

Gruber, G., Ibid, 31.

Tessi du Montay, o. M., and o. R. Marechal, British Patent No. 2548 (1867).
Tomita, Akira and Yasuhatsu Tamai, J. Cute. 27, 293 (1972).

Tomita, A. and Y. Tamai, J. Phys. Chem. 78, 2254 (1974).

Tomita, A., N. Sato and Y. Tamai, Carbon 12, 143 (1974).

McKee, D. W., Ibid, 453.

Nahas, N. 0., Fuel 02, 239 (1983).

Kubiak, H., H. J. Schroter, A. Sulimma and K. H. van Heek, Ibid, 242.
Huttinger, Klaus J. Ibid, 166.

Mims, C. A. and J. K. Pabst, ACS Div. Fuel Chem. Prepr. 25(3), 263 (1980).
Mims, C. A. and J. K. Pabst,Fuel 62, 176 (1983).

Juntgen, Harald, Fuel 62, 234 (1983).

Huttinger, Klaus J. and R. Minges, Fuel 65, 1122 (1986).

McKee, D. W. and D. Chatterji, Carbon 13, 381 (1975).

McKee, D. W. and D. Chatterji, 12th Biennial Conference on Carbon pp. 189-
191, Pittsburgh, Pennsylvania (1975).

McKee, D. W. and D. Chatterji, Carbon 16, 53 (1978).

186

[18]
[19]
[20]

[21]

[22]

[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]

[31]

[32]
[33]
[34]
[35]

[36]

[37]
[38]

[39]

187

McKee, D. W., C. L. Spiro, P. G. Kosky and E. J. Lamby, Fuel 62, 218 (1983).
Veraa, M. J. and A. T. Bell, Ibid 57, 194 (1978).
Huttinger, K. J. and R. Minges, Ibid 65, 1112 (1986).

Nlims, C. A. and J. K. Pabst, Proc. Int. Conf. on Coal Science, pp. 730-736, Ver-
lag Gluckauf GmbH, Essen (1981).

Mims, C. A., K. D. Rose, M. T. Melchior and J. K. Pabst, J. Am. Chem. Soc.
104, 6886 (1982).

Yuh, S. J., and E. E. Wolf, Ibid 63, 1604 (1984).

Wood, B. J., R. H. Fleming and H. Wise, Ibid 63, 1600 (1984).
Wood, B. J., R. D. Brittain and K. H. Lau, Carbon 23, 73 (1985).
Long, F. J., and K. W. Sykes, J. Chim. Phys. 47, 361 (1950).
Saber, J. M., J. L. Falconer and L. F. Brown, Ibid 65, 1356 (1986).
Wen, W. Y., Cata. Rev. Sci. Eng. 22, 1 (1980)

Wigmans, T., H. Haringa and J. A. Moulijn, Fuel 62, 186 (1983).
Wigmans, T., J. vanDoorn and J. A. Moulijn, Ibid 62, 190 (1983).

Freriks, I. L. C., H. M. H. vanWechem, J. C. M. Stuiver and R. Bouwman, Ibid
60,465 (1981).

Yuh, S. J., and E. E. Wolf, Ibid 62, 252 (1983).

Cerfontain, M. B. and J. A. Moulijn, Ibid 62, 256 (1983).
Sancier, K. M., Ibid 62, 331 (1983).

Sancier, K. M., ACS Div. Fuel Chem. Prepr. 28(1), 62 (1983)

Huffman, G. P., F. E. Huggins, R. G. Jenkins, A. Piotruswski, F. W. Lytle and
R. B. Greegor, Fuel 65, 1339 (1986).

Wood, R. E. and G. R. Hill, ACS Div. Fuel Chem. Prepr. 17(1), (1972).
Gardner, N., E. Samuels and K. Wilks, Adv. Chem. Ser. 131, 217 (1974).

Cao, J. R. and M. H. Back, Carbon 23, 141 (1985).

[40]
[41]
[42]
[43]
[44]
[45]

[46]

[47]

[43]
[49]
[50]

[51]

188

Blackwood, J. D., Ibid 12, 14 (1959).

Hashimoto, K., K. Miura, J. J. Xu and A. Watanabe, Chem. Lett. 5, 555 (1985).
Sams, D. A., Tevan Talverdian and F. Shadman, Fuel 64, 1208 (1985).
Kolthoﬁ', I. M., J. Am. Chem. Soc. 54, 69 (1932).

Villars, D. S., Ibid 69, 214 (1947).

Villars, D. 8., Ibid 70, 3655 (1948).

Studebaker, M. L., E. W. D. Huffman, A. C. Wolfe and L. G. Nabors, Ind. Eng.
Chem. 48, 162 (1956).

Boehm, H. P., E. Diehl, W. Heck and R. Sappok, Angew. Chem. Inter. Edit.
3(10),669(1964)

Voll, M. and H. P. Boehm, Carbon 9, 481 (1971).
Studebaker, M. L. and R. W. Rinehart Sr., Rubb. Chem. Tech. 45, 106 (1972).
Barton, S. S., G. L. Boulton and H. Harrison, Carbon 10, 395 (1972).

Puri, B. R., D. D. Singh, J. Nath and L. R. Sharma, Ind. Eng. Chem. 50, 1071
(1958)

Puri, B. R. and R. C. Bansal, Carbon 1, 457 (1964).

Barton, S. 8., D. Gillespie and B. H. Harrison, Ibid 11, 649 (1973).

Garten, V. A., D. E. Weiss and J. B. Willis, Aust. J. Chem. 10, 309 (1957).
Barton, S. S. and B. H. Harrison, Carbon 13, 283 (1975).

King, A., J. Chem. Soc. 1937, p. 1489.

Garten, V. A. and D. E. Weiss, Aust. J. Chem. 8, 68 (1955).

Garten, V. A. and D. E. Weiss, Ibid 10, 309 (1957).

Walker Jr., P. L. and J. Janov, J. Coll. Inter. Sci. 28, 449 (1968).
Dubinin, M. M., Carbon 18, 355 (1980).

Gray, H. R.,“Teating for Hydrogen Environment Embrittlement: Experimental
Variables,” ASTM Special Technical Publication 543, pp. 133-151, Los Angeles,
Calif, 25-30 June 1972.

[32]

[63]

[64]

[65]

[66]

[37]

[68]
[39]

[70]

[71]

[72]

[73]

[74]

[75]

[75]

[77]

[78]

189

Chandler, W. T. and R. J. Walter, Ibid, pp. 170-197.

Bhattacharyya, S., E. J. Vesely, Jr. and V. L. Hill, J. Materials for Energy Sys-
tems 3(4), 12 (1982).

Simmons, Ward F. and Howard C. Cross,“The Elevated- Temperature Properties
of Selected Super-Strength Alloys,” ASTM Special Technical Report No. 160, pp.
20 and 133-142 (1954).

Seely, Fred B. and James O. Smith,“Advanced Mechanics of Materials,” 2nd ed.,
Chapter 10, John Wiley & Sons, New York, NY. (1952).

Ibid, pp. 69 and 220-228.

Waters, E. O., D. B. Wesstrom, D. B. Rossheim and F. S. G. Williams, A.S.M.E.
Transactions, FSP-59-4 59, 161 (1937).

Private communication with the manufacturer.
Waters, Everett 0. and J. Hall Taylor, Mech. Eng. 49, 531 (1927)

Perry, Robert H. and Cecil H. Chilton,“Chemical Engineers’ Handbook,” 5th ed.,
pp. 23-38-23-53, McGraw-Hill Book Co., NY. (1973).

Baumeister, T., E. A. Avallone and T. Baumeister III,“Marlcs’ Standard Handbook
for Mechanical Engineers,” 8th ed., pp. 5-5 and 6-11, McGraw- Hill Book Co.,
NY. (1978).

Seely, Fred B. and James O. Smith,“Advanced Mechanics of Materials,” 2nd ed.,
pp. 49-67, John Wiley & Sons, New York, NY. (1952).

Dieter Jr., George E.,“Mechanical Metallurgy,” p. 36, McGraw Hill Book Com-
pany Inc., New York, 1961.

McCabe, W. L. and J. C. Smith,“ Unit Operations of Chemical Engineering,” 3rd
ed., pp. 312-320, McGraw Hill Book Company, New York, NY. (1976).

Bird, R. 3., Warren E. Stewart and Edwin N. Lightfoot,“Transport Phenomena,”
pp. 317-319, John Wiley & Sons, Inc., New York, NY. (1960).

Lederer, C. M. and V. S. Shirely,“Table of Isotopes,” 7th ed., pp. 101-102, John
Wiley & Sons Inc., New York, N.Y., 1978.

Chase, G. D. and J. L. Rabinowitz,“Priciples of Radioisotope Methodology,” 2nd
ed., pp. 136-152, Burgess Publishing Co., Minneapolis, MN 1965.

Oﬂ'ice of Radiation, Chemical and Biological Safety,“Radiation Safety Study
Guide,” pp. 56-58, Michigan State University, East Lansing, N11, 1982.

[79]

[Sol
[81]

[82]

[83]

[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]

[92]

[93]
[94]
[95]
[96]
[97]
[98]
[99]

[100]

190

Lamarsh, John R.,“Introduction to Nuclear Engineering,” pp. 75-87, Addison-
Wesley Publishing Co., 1975.

Private communication with the manufacturere.
Supelco Inc., 00 Bulletin 7600, pp. 2—4, 1983.

Fike Metal Products Corporation, Catalog No. 7388,“Rupture Discs, Explosion
Vents 8 Other Pressure Relief Devices,” 1983.

Brunauer, Stephen, P. H. Emmett and Edward Teller, J. Am. Chem. Soc. 60,
309 (1938).

Pierce, Conway and Bland Ewing, J. Phys. Chem. 68, 2562 (1964).
Ladd, W. A. and M. W. Ladd, Rubb Chem. Tech. 34, 697 (1961).
Donnet, J. B. and J. C. Bouland, Rev. Gen. Caoutchouc 41, 407 (1964).
Emmett, P. H. and Martin Cines, J. Phys. Coll. Chem. 51, 1329 (1947)
Anderson, R. B. and P. H. Emmett, J. Appl. Phys. 19, 367 (1948).
Janzen, Jay and Gerard Kraus, Rubb. Chem. Tech. 44, 1287 (1971).
Klyne, R. A., B. D. Simpson and M. L. Studebaker, Ibid 46, 192 (1973).
Pausch, J. B. and C. A. McKalen, Ibid 56, 440 (1983).

de Boer, J. H., B. G. Linsen, Th. van der Plas and G. J. Zondervan, J. Catal. 4,
649 (1965).

Voet, A., T. G. Lamond and D. Sweigart, Carbon 6, 707 (1968).

Smith, W. R. and G. A. Kasten, Rubb. Chem. Tech. 43, 960 (1970).
Lamond, T. G. and C. R. Price, J. Coll. Inter]. Sci. 31, 104 (1969).
Turkdogan, E. T., R. G. Olsson and J. V. Vinters, Carbon 8, 545 (1970).
Lamond, T. G. and H. Marsh, Carbon 1, 281 (1964).

Heckman, F. A., Rubb. Chem. Tech. 37, 1245 (1964).

Walker Jr., P. L., American Scientist 50, 259 (1962).

Riley, H. L., Chem. 5 Ind. 58, 391 (1939).

[101)

[102]

[103)
[104]
[105)

[106]

(107)
[108]
[109]

[110]
[111)

[112]

[113]

[114]
[115]

[116]
[117]

[118]
(119)
(120)

[121)

191

Bisco, J. and B. E. Warren, J. Appl. Phys. 13, 364 (1942).

Hofmann, A. R., G. Rudorff, R. Holst, W. Ruston, A. Russ, and G. L. Ruess, Z.
Anorg. Chem. 255, 195 (1947).

Hall, C. E., J. Appl. Phys. 19, 271 (1948).
Donnet, J. B. and J. C. Bouland, Rev. G’én. du Caoutchouc 41, 407 (1964).
Heckman, F. A. and D. F. Harling, Rubb. Chem. Tech. 39, 1 (1966).

Hess, W. M., L. L. Ban, F. J. Echert, and V. Chirico, Rubb. Chem. Tech. 41, 356
(1968).

Donnet, J. B., J. Schultz and A. Eckhardt, Carbon 6, 781 (1968).
Heidenreich, R. D., W. M. Hess and L. L. Ban, J. Appl. Cryst. 1, 1 (1968).
Ergun, Sabri, Carbon 6, 141 (1968).

Marsh, P. A., A. Voet, T. J. Mullens, and L. D. Price, Rubb. Chem. Tech. 43,
470 (1970).

Ban, L. L. and W. M. Hess, Proc. Annu. Meet. Elect. Microsc. Soc. Am. 26th,
256 (1968).

Harling, D. F. and F. A. Heckman, Mater. Plast. Elastomeri 35, 80 (1969).

Briggs, D. and M. P. Seah,“Practical Surface Analysis,” pp. 5-8, John-Wiley &
Sons, Chichester, England 1983.

Ibid, pp. 519-520.
Garten, V. A. and D. E. Weiss, Aust. J. Chem. 8, 68 (1955).

Treptau, M. H., H. Zoheidi and D. J. Miller, ACS Div. Fuel Chem. Prepr. 31(3),
150 (1986).

Smith, J. M. and H. C. Van Ness,“Introduction to Chemical Engineering Thermo-
dynamics,” 3rd ed., p. 390, McGraw Hill Book Co., New York, 1975.

Bone, W. A. and D. S. Jerdan, J. Chem. Soc. 71, 41 (1897).
[bid 79, 1042 (1901).
Pring, J. N. and R. S. Hutton, Ibid 89, 1591 (1906).

Bone, W. A. and H. F. Coward, Ibid 93, 1975 (1908).

[122)
(123)
U24]
U25]
0261
[1271
[1281

(129]

[130)

[1311
[132)
[1331
[1341
[1351

U36]

[1371
[138]

(139]

(140)

[141]

u42]

192

Pring, J. N., Ibid 97, 498 (1910).

Cao, J. R. and M. H. Back, Carbon 20, 505 (1982).

Blackwood, J. D., Aust. J. Chem. 12, 14 (1959).

Pring, J. N. and D. M. Fairlie, J. Ind. Eng. Chem. 4, 812 (1912).

Ibid, J. Chem. Soc. 101, 91 (1912).

Browning, L. C. and P. H. Emmett, J. Am. Chem. Soc. 73, 581 (1951).
Saunders, S. W., J. Phys. Chem. 28, 1151 (1924).

Levine, I. N.,“Physical Chemistry,” p. 157, McGraw Hill Book Co., New York,
1978.

Himmelblau, David M.,“Basic Principles and Calculations in Chemical Engineer-
ing,” 4th ed., pp. 583-586, Prentice-Hall Inc., Englewood Cliffs, New Jersey, 1982.

Squires, A. M., Trans. Inst. Chem. Engrs. 39, 3 (1961).

Zielke, C. W. and E. Gorin, Ind. Eng. Chem. 47, 820 (1955).
Blackwood, J. D., Aust. J. Chem. 15, 397 (1962).

Blackwood, J. D. and D. J. McCarthy, Ibid 19, 797 (1966).

Shaw, J. T., National Coal Board Gasiﬁcation Note No. 29 (1977).

Satterﬁeld, C. N.,“Mass Transfer in Heterogeneous Catalysis,” p. 231, Krieger
Publishing Company, Huntington, N.Y., 1981.

Ibid, pp. 133-139.
Thiele, E. W., Ind. Eng. Chem. 31, 916 (1939)

Mantell, C. L.,“Carbon and Graphite Handbook,” p. 21, Interscience Publishers,
John Wiley & Sons Inc., N.Y., 1968.

Wilke, C. R. and C. Y. Lee, Ind. Eng. Chem. 47, 1253 (1955).

Boyd, C. A., N. Stein, V. Steingrimson, and W. F. Rumpel, J. Chem. Phys. 19,
548(1951)

Marrero, T. R. and E. A. Mason, J. Phys. Chem. Ref. Data 1, 3 (1972).

[143)

[1441
[14s]
[1461
[1471

[148]

[149)

[150]
[151]
[152]
[153]

[154)

(155]

[156]

[157]
[158]
[159]
[160]
[151]

[162]

[163]

193

Satterﬁeld, C. N.,“Mass Transfer in Heterogeneous Catalysis,” pp. 33-42 and 157,
Krieger Publishing Company, Huntington, N.Y., 1981.

Graham, D., J. Phys. Chem. 61, 1310 (1957).

Hart, P. J., F. J. Vastola and P. L. Walker Jr. Carbon 5, 363 (1967).
Phillips, R., F. J. Vastola and P. L. Walker Jr. Ibid 8, 197 (1970).
Dollimore, J ., C. M. Freedman, B. H. Harrison and D. F. Quinn, Ibid, 587.

Levine, Ira N.,“Physical Chemistry,” p. 516, McGraw Hill Book Co., New York,
1978.

Mims, C. A., R. T. K. Baker and J. J. Chludzinski, ACS Div. Fuel Chem. Prepr.
28(1),71(1983)

Roscoe and Schorlemmer,“Treatise on Chemistry,” (1883).
Fox, D. A. and A. H. White, Ind. Eng. Chem. 23, 259 (1931).
Cerfontain, M. B. and J. A. Moulijn, Fuel 65, 1349 (1986).
Huhn, F., J. Klein and H. Juntgen, Fuel 62, 196 (1983).

Yokoyama, S., K. Miyahara, K. Tanaka, J. Tashiro, and T. Takukuwa, Nippon
Kagalcu Kaishi 6, 974 (1980).

Mims, C. A. and J. K. Pabst, ACS Div. Fuel Chem. Prepr. 25, 258 (1980).

Hashimoto, K., K. Miura, J. J. Xu, and A. Watanabe, Chem. Lett, 555-556
(1984)

Sams, D. A., T. Talverdian and F. Shadman, Fuel 64, 1208 (1985).
Kapteijn, F., J. Jurriaans and J. A. Moulijn, Fuel 62, 249 (1983).

Tromp, P. J. J. and E. H. P. Cordfunke, Thermochimica Acta 77, 49 (1984).
Ferguson, E., R. Schlogl and W. Jones, Fuel 63, 1048 (1984). 2

Kuhn, L. and H. Plogmann, Fuel 62, 205 (1983).

Formella, K., P. Leonhardt, A. Sulimma, K.-H. van Heek, and H. Juntgen, Fuel
65, 1470 (1986).

Bruno, G., L. Carvani and G. Passoni, Ibid, 1473.

[164]

[165]

[1661
[167]
[1681
[1691

[170]

194

Puri, B. R. and R. S. Hazra, Carbon 9, 123 (1971).

Froment, G. F. and K. B. Bischoﬂ',“Chemical Reactor Analysis 8 Design,” pp.
178-184, John Wiley & Sons, New York, N.Y., 1979.

Wood, B. J. and K. M. Sancier, Cata. Rev. Sci. Eng. 26, 233 (1984).
Wigmans, T., R. Elfring and J. A. Moulijn, Carbon 21, 1 (1983).
Rao, Y. K., A. Adjorlolo and J. H. Haberman, Carbon 20, 207 (1982).
Wigmans, T., J. C. Goebel and J. A. Moulijn, Carbon 21, 295 (1983).

Harker, H, Proc. 42" Biennial Conf. Carbon, pp. 125-139 (1959).

    

 
 

RRRRR

"llilllgﬂ)][][Iliuqui‘yllllulljxu]'“