2Bu*Nj*,< 1 | v* ,^V\
l?$ ■$.'■ .? «w $ 11
iV -a ■■<■.•<; • • -a .
m
For Reference
NOT TO BE TAKEN FROM THIS ROOM
THESIS
THE FISCHER'-TROPSCH SYNTHESIS
The Development of Methods and Apparatus,
and Preliminary Studies on Catalyst
by
J* Harry Donald, B.Sc,
an f
Donald Q,uon, B.Sc.
University of Alberta,
Edmonton, Alberta.
September,
■} h -
: "
mmMMWM ■ i ifte l ' ^
(ft */ Ht. . h V :■<?:.. /<• v
4.. N/. PilVU? V
.V! y ;y
S/Vfrt -jiV fyin
V*' ;
U \ 'X/
W> V.Mf
^ XYii 'JvV\
v V * iilr
1'l rV.ftj v 1*; {
^iavT v ^vV
}> C^ i;’ ^ .'I
v\ i* C ( V
■m. ft \
vv
1945
©X Mils
WOMSBiMtS
mtmamm
Digitized by the Internet Archive
in 2018 with funding from
University of Alberta Libraries
https://archive.org/details/fischertropschsyOOdona
.
*
.
'
.
UNIVERSITY OF ALBERTA
Faculty of Arts and Science
This is to certify that the undersigned have
read and recommended to the Committee of Graduate
Studies for acceptance, a thesis submitted jointly by
J. H. A. Donald, B.Sc. in Chemical Engineering, and
D. Quon, B.Sc. in Chemical Engineering, entitled:
THE FISCHER-TROPSCIi SYNTHESIS
The development of methods and apparatus, and pre¬
liminary studies on catalysts.
Professor
Professor
Professor
,r
THE FISCHER-TROPSCH SYNTHESIS
The development of methods and apparatus,
and preliminary studies on catalysts.
Submitt ed in Partial Fulfullment
of the
Requirements for the Degree of
MASTER of SCIENCE
by
J. Harry Donald, B.Sc.
and
Donald Quon, B.Sc.
Under the Direction of
Dr. E. H. Boomer.
University of Alberta,
Department of Chemistry,
Edmonton, Alberta. September, 1945.
.
'
■ V • .. ; i ■ :cf/j .
,
,
.
.
Acknowledgement
13 +*
4 16
The authors wish to take this opportunity
to extend sincere thanks to all those who have
assisted in this investigation:
To Dr. E. H. Boomer, director of the project,
whose assistance and advice were invaluable.
To J. G. Knudsen, for assistance in and
supervision of this research.
To the Research Council of Alberta, who
financed the project.
To the Board of Governors of the University
of Alberta, for the Research Scholarships which
enabled the writers to carry on this investigation.
To the Department of Chemistry for services
and supplies.
• -
, , J n r J n !
.
. ..
.
.
■
....
*
■
*
Table of Contents
Page
I Introduction ..* 1
II Literature Review ...« . 9
III Experimental Equipment ................ 34
IV Catalysts and their Preparation . 63
V Analytical Methods and Operational Tech¬
niques 67
VI Experimental Results .. 76
VII Summary and Conclusions .. 103
Appendix
A Preparation of Synthesis Gas . 105
B Design and Calibration of Flowmeters 111
C Short-Time Catalyst Testing in a
Static System .. 123
Bibliography . 132
List of Figures
Page
Figure I . . . . ^ 2cf
Figure II . 20
Figure III . 23
Figure IV .... . 31
Figure V . 37
Figure VI .. 38
Figure VII . 39
Figure VIII . 45
Figure IX . 50
Figure X .. 53
Figure XI . 54
Figure XII . 56
Figure XIII.!. 57
Figure XIV .. 91
Figure XV . 96
Figure XVI . 107
Figure XVII . 112
Figure XVIII . 118
Figure XIX . 119
Figure XX . 122
Figure XXI . 125
Figure XXII . 128
Figure XXIII
129
I INTRODUCTION
The discovery of petroleum and its applica¬
tions have been largely responsible for the great
industrial development of this century. So large has
been the consumption (and no decrease in the future
can be expected) that some oil authorities have pre¬
dicted a definite shortage within a period of forty
years. Extensive exploration is under way to discover
new oil fields but it is felt that the real solution
lies in the development of synthetic fuels and
supplementary reserves.
Even a preliminary examination of supplemen¬
tary reserves of liquid fuels in the form of oil shales
and sands, and coal, discloses quantities of products
that in principle remove for generations to come any
possibility of a shortage. The problem is not one of
scarce raw materials but one of economics, to provide
liquid fuels from coal, for example, at a cost not
substantially different from present costs based on
petroleum.
'
,
.
• .
. ' : •• '• v J'. :> , : : £;} OG -If
.
1 . M- . : ' ... . O'iv ...
.
• "'■■■■ . . c~
■ ■
.
.
• ' •... ; 0 :;
.
In the past few years, there has been a
growing interest in this problem. This is particularly
true of those countries with limited or no oil resources.
Spurred on by an endeavor to become economically self-
sufficient, Germany has been leading the field in
research, and commercial development. Japan, Britain,
and the United States, however, are now taking an
increasingly active interest in the possible production
of synthetic petroleum.
At the present time, two important processes
are receiving considerable attention.
1} Berguis Process - a catalytic hydrogena¬
tion of coal under high pressures and temperatures.
2) Fischer-Tropsch Process - an indirect
catalytic hydrogenation of coal and other carbonaceous
materials.
Both processes present excellent possibilities
and both have reached the commercial stage of develop¬
ment. However, only the latter process will be con¬
sidered in this report.
The Fischer-Tropsch synthesis consists of the
catalytic hydrogenation of carbon monoxide at medium
pressures (5-15 atm.) into a mixture of gaseous and
liquid hydrocarbons, suitable for the manufacture of
fuels. Comparatively low temperatures are employed
.
'
.
■■ - M ■ . . ■ .
.
, .. i
.e-,:;..,' i ; <>.u : ..
.
110 d . ■ ,
ioc ■
.» o
- Joaq ■ - , .
*
. ■ ' : : . ; ao-iq ,u . .. oo”- a
,
,
'
. 8
-3-
(190° to 250°C)- depending upon the catalyst. The
synthesis gas, a mixture of CO and H 2 often known as
water gas, is produced by the incomplete oxidation of
coal, coke, and other carbonaceous materials, such as
natural gas.
The production of synthetic hydrocarbons
from water gas was first carried out by P. Fischer
and H. Tropsch in 1925 - 26 (17). Working at atmos¬
pheric pressure and at temperatures of 2500 „ 300°G,
they obtained a product which was principally aliphatic
hydrocarbons, containing little or no oxygenated
compounds. Their research was so fruitful in the next
few years that in 1933 the Ruhrchemie A - G undertook
the construction of a plant at Holten, Germany, with
a capacity of 1000 tons per year. Much progress was
made in the years following. By the end of 1936, five
plants were completed or under construction, boosting
the production of primary oils to 145,000 metric tons
per annum (30). Germany continued to increase its
production of synthetic fuels until, in 1940, an
annual production of 1,000,000 tons of Fischer-Tropsch
liquids was reported (31). Undoubtedly this figure
was increased substantially during the war years.
In Japan, by 1938, two plants had been
reported giving an annual output of 150,000 tons.
.
dgW M . fl IjHM :
D;7.'',0 ■ . . , ■ v
.
. .
... ; ■. ; ■■. - . ■
. .
.
.
-
.i.■ ■ -.. ' ■ ■ - ; .. ■ . .
r ' . ■ ' ' : .. ■ ; J
*
o:
,
__
u
-
« .
- ...
*
• .
-4-
Later information revealed that the erection of three
new plants had increased the annual production by
260,000 tons. Small scale plants have been built in
both France and Britain.
The extensive development of the Fischer-
Tropsch synthesis has been promoted in Germany, not
particularly by economic, but by nationalistic motives.
An investigation by the U.S. Senate (27) into the
economics of this process indicates the relative costs
of gasoline produced by two methods.
1) Fischer-Tropsch synthesis
a. from coal - 19.2//gal. with the
present European design.
b. from natural gas - 8.8//gal. with
the present design. However a cost of only 4.8//gal.
is possible if large scale production were carried
out using American technology and natural gas at less
than 5/ per 1000 cu. ft.
2) Petroleum Refining
- 5.3 - 5.5//gal. with present day methods.
While the production of synthetic fuels from
natural gas could, under the most favorable circum¬
stances, compete economically with that from petroleum,
it cannot be considered as a long time solution to the
problem. Any real solution lies in the use of coal as
the primary raw material. The resources of natural gas
■ ; ■ . • ■ ■ ' • ■ ■
. ■ i .<
...
«
, .
.■■■: . 1
,
. .■ ■■ : j ■ , : . - ..:uo ,i' 1 ; I
,
,
.
*
f i£o 3 r iow olu o* q t,£ • . i ■ .
*
.
. .A
.
.
- 5 -
are of the same order of magnitude as petroleum and
further the economic and social value of natural gas
as an industrial and domestic fuel may rule out its
extensive use as a chemical raw material.
The Fischer-Tropsch process has attracted
the interest of several research organizations in the
United States. At present the Bureau of Mines, U.S.
Department of the Interior, is vitally interested in '
its development and is planning the construction of
several pilot plants. Research work is also being
carried out in a number of oil companies (the most
important of which are probably the Standard Oil of
Hew Jersey and the Gulf Oil Co.) but information
about their progress is unavailable.
Canada, up to the present time, has shown
little interest in the Fischer-Tropsch synthesis.
However, in 1944 the Research Council of Alberta
initiated a program of research in this connection.
This question is of particular interest to Alberta
which is enormously rich in reserves of coal and
natural gas, the basic raw materials for the manu¬
facture of synthetic fuels. Important industrial
development of this process would be of immense
economic value to this province.
■■■ • • : 1 ■ > '!■
'
; . 1
'i« : ' : r O'. ; o
? ' ... ,
•* * • •
,
■ - . : . v
■. i • v 5
. . • ' • . yi r, ■ Q
*
■ * i ‘ v
. .
_
' " • . • , ni
.
■
.
,
.
. .
'
*
- 6 -
In the research and development of the
Pischer-Tropsch process, the following general
problems must be solved:
1) Industrial development of laboratory
operations. This is a problem common to all new
processes. Economic design and operation depend
upon the availability of fundamental data on the
process. Much work remains to be done in this field.
2) Heat exchange. Since the reaction is
highly exothermic, and since the catalysts are very
sensitive to temperature, means must be devised to
dissipate the heat. At the present time, the allow¬
able space velocities (and hence gas velocities) are
low, resulting in poor heat transfer. Consequently
thin layers of catalyst must be used to properly dis¬
sipate the heat. All this means comparatively large
amounts of steel per unit of catalyst i.e. unit of
production.
3) Development of catalysts. This involves
the preparation, conditioning, maintenance, and
evaluation of new and more effective catalysts. Improve¬
ment of catalysts can lead to elimination of the heat
exchange problem. It is with the development of
catalysts that the present investigation is primarily
concerned
• . . ; P
• . , ! • •
j • 'p:' ' ■„> ’’ v' i 1 aj.C’iu
,
-
.. "'P 1 .Hr
.
.
... v ..’ . ; . ' . r , v.-\ -
. . 1 ;'» v! "■ : ;-.j ■ ; ;
.
.
: ' . t-BO
. . •' :r v. J.
.
V it ■ . ' . .... ...jo. Po i.p. i, . / .
, . . i
' ■ ; . . ,
• • ' . ■
. ' ■ ..' ..
-
- 7 -
The theory of catalysis is as yet poorly
developed and the problem can be attacked only upon
an empirical basis. Long term tests under conditions
approximating commercial production are necessary to
properly evaluate a catalyst. This testing is an
integral part of any long range program of fundamental
research on the whole question of catalyst development.
Research on Fischer-Tropsch catalysts at the
University of Alberta was begun in the summer and fall
of 1944 and only the preliminary work has been done on
f
the problem. This involved the construction and
assembly of the experimental equipment, the solution
of the operational difficulties, and the initial tests
upon a few catalysts.
In the course of this investigation, several
problems were encountered, which are related, but which
have no direct bearing upon the catalyst testing. Two
in particular, might be mentioned:
1) Preparation of the synthesis gas
2) Design and calibration of capillary flow-
meters. They are of such general interest that it was
felt that they should be included in an Appendix of
this report.
Preliminary experiments were also carried out
on the short-time testing of Fischer-Tropsch catalysts
. ■
• '■ L- :r-, ,'jr ' O'tqqu
,
; ... ao . .a . .. V .. i v.
%
: :
• l • ■ t ' J 18 ‘ I 1 . f 3 j 0 . .
o i; . .. ':! • ;;: ■
_
- .
■
' • ■ . - . ■ <} - , ■ .
■■ ■ . 1 ■ . . :
*
>
... .. : i -• J -:,o : no
- 8 -
in a static system at atmospheric pressure. A brief
discussion of this phase of the project is included
in the Appendix.
II LITERATURE REVIEW
Since its original discovery in 1926, the
Fischer-Tropsch synthesis has received wide-spread
attention throughout the world, with extensive investi¬
gations being carried out largely in Germany but with
important contributions from Japan, Russia, England
and the United States. As a result, the literature
on this process is extensive. Most of the original
papers have been published in foreign journals and,
except for their abstracts, have not been consulted.
In view of the importance of the process, however,
many of these papers have been translated and
reprinted in available journals. In addition, the
information has been brought up to date by several
comprehensive reviews (18, 23, 30). While the follow¬
ing literature review does not profess to be complete,
it is believed that the major contributions have been
included.
.
' i.i ' n
.
■
.
:j . ax •'x .. vv . I I-;. : ;-j : c mtt
.
■ ... ; . . . • . . .■■ Ia. u ; a ; • vn LL:;}
..... . ■ • • / j '-I 'rl'ia'. ■ L •
.
n ' *. I .■■■/ :,v . •
: :i?a
' "■ ... v, j 1 ''0 , . I 'lo v<sLv a I
. • ‘
- J’x
'a ;. .
"/
. . • >. .. ' .:
,
. ' J a.:. j ! I \
y/. . -I ■ J ■ fj ■ *i >[ 0; ' r r / I. ; I .
-
- 10 -
The production of hydrocarbons from water
gas was pioneered in 1913 by Sabatier (28). He found
that finely divided nickel catalyzed the reduction of
carbon monoxide to methane according to the following
reaction:
CO + 3H 2 CH 4 + H 2 0
However, no higher hydrocarbons were obtained.
In the same year the Badische Anilin and
Sodafrabrik organization patented a process (4) in
which saturated and unsaturated gaseous and liquid
hydrocarbons, alcohols, aldehydes, acids and ketones
were produced from water gas. The reaction took
place at elevated temperatures and pressures (300°
- 400° C, 100 - 200 atmospheres) over a methansl
catalyst impregnated with alkali and oxides of cobalt,
osmium or zinc..
In 1913 Franz Fischer and Hans Tropsch (16),
on passing water gas over an alkali iron catalyst at
400 - 450°C and 100 - 150 atmospheres, obtained a
mixture of aliphatic alcohols, aldehydes, ketones,
acids and esters. This product was called n Synthol !5 .
It was not until three years later, however, that
these two workers reported (17) the famous synthesis
which bears their name. By operating at lower
pressures and temperatures, and by using alkali-free
r
>v
.
■
' J {.' •'O ,, •..
O.v . ' - I
.
*
It .
.
■
■ : •
, •
'
■ -
don ai:OS..-.u d a..:: y.
• VI iJ
. . .'j . i ? .
. •. 1
>
.
■-
I ' ■ ’.o '.'i
iron and cobalt catalysts, they obtained a product
(given the name "Kogasin") which consisted almost
entirely of hydrocarbons ranging from ethane to solid
paraffin.
Since the original announcement, research
workers in many countries have studies this synthesis
and its implications. Attention has been devoted
chiefly to the following phases of the process - the
development of new and more active catalysts; the
determination of the optimum operating conditions for
existing catalysts; and the study of the mechanism of
the reactions involved. These will be dealt with in
turn.
Catalysts and their Properties .
Many catalysts have been suggested for the
Fischer-Tropsch synthesis. On the whole, however, the
main constituent of these catalysts is one or more
metals from the Eighth Group of the Periodic Table
(i.e. Hi, Co, and Fe) usually in the metallic form.
Before any fundamental research on Fischer-
Tropsch catalysts is possible, it is first necessary
to know the distinguishing characteristics of these
catalysts - their desired properties and their faults.
Eomarewsky and Riesz (23) state that Fischer-Tropsch
catalysts must comply with the following requirements:
..
: SI
,
.
*
■
. . . .
vcr:.. :
i d ; o / . t . 7
. ■
. •
■ -
. ■
*
.
:i. . .: d 20 . : r £ j
r , '
■
- ,
...... < 1
■ b 7 .: : . . ' O ' :
’
i ' • '7 7 7.7:: . ! ; ' 7 '
■
- 12 -
1) Ability to form a carbide from carbon-
monoxide
2) Hydrogenating ability
3) Polymerizing ability
All metals in the Eighth Group satisfy the
first two, and, to a certain extent, the third con¬
dition. These investigators further suggest that
multi-component or ”complex-action” catalysts are
the most suitable.
Although Fischer and Tropsch employed an
iron catalyst in the original synthesis, most of their
later work was carried out on cobalt and nickel
catalysts. The following table shows the yield and
life of some of the catalysts which they had studied
up to 1936 (9).
' ' . . /!• '' .*} ' !. " (
*>
'
i- 1 ' , 3. - 0\ <
• £ ..f; ' tf ,
,
■ 1 syio- :
■
• r.:• ,:i ^<5 i;
*
OO 'C
,
.
■
- 13 -
Table I
Yield and Life of Some Fischer-
Catalyst
Ip on:
Fe Cu Mn Silicagel 0.4 % KgCQ^
Fe Cu on Kieselguhr
Nickel:
Ni Th on Kieselguhr
Ni Mn A1 on Kieselguhr
Cobalt:
Co Th on Kieselguhr
Co Th Kieselguhr decomposed
Co Mn on Kieselguhr
Co Ni alloyed (Si)
Co Th Cu on Kieselguhr
Yield Life till de¬
crease to 80$
g/meter of of original
synthesis gas
activity
30-35
8 days
28
8 days
100
30 days
105
45 days
110
60 days
105
25 days
105
Min. 30 days
85
12 days
105
60 days
It is seen that the Co-Th-Kieselguhr catalyst
appeared to be the best at that time. However, in 1943,
Fischer and co-workers (13) report that further studies
have shown that a nickel-manganese-alumina catalyst was
clearly superior.
Other investigators have been actively engaged
in the development of new catalysts. Although attempts
have been made and are being made to develop short time,
positive test methods, there has been no success, and
' ■ ., ....: - I ’■
...... ■ .. •' *• fs.--.rn
v ... X.. '
. .
..
..
" .* "**-
: ;V:>. .L
c.
X •> \zi. XXX
. . ..
*
; ■ . : 0.. ..
X;.
-to f X T T
• . < g. X . . X . •
X .
or i
• ' X. ::’ X
;
i '. . .V;.
'.. .; ", .X oX
.10 t
1 '■ 1
£j'. .
*Ioa f . ■ - . ....X
.
’ . . ~ 0 l ......
■
_
,
. ■. r ....
.
■ . " • : : j
j V' .Xi . 'OX ct
■ .
■' : . ...... ■/ .
14 -
to date this problem can be attacked only empirically.
The value of a catalyst can be assessed only after
long continuous tests under conditions approximating
commercial operation. Cobalt and nickel catalysts have
received the most attention and are consequently more
developed than the iron catalysts. However, the
relatively high cost of cobalt and nickel has prompted
many research workers to study and develop the latter.
I. G. Farbenindustrie A.-G. has patented
the so-called '’Iso” process, using a robust iron
catalyst, operating at 200 psi. and 300°C. This is
probably a fused iron oxide catalyst which has been
reduced at 850°C. Little is known about this process
but development in Germany had apparently reached the
commercial stage. Recent information (private) suggests,
however, that this process is not yet successful.
The action of promoters on Fischer-Tropsch
catalysts is obscure and the only guide being followed
is that of trial and error. Foster (18) has reviev/ed
some of the catalysts which have been suggested. The
promoters used include alkali and alkaline earth metal
oxides, copper, thoria, and the oxides or sulphides of
uranium, chromium, tungsten, aluminium, and manganese.
Other proposed promoters are found in the patent litera¬
ture but the latter contains little information of value.
■■ . 0 i t. ,G C it
;■ ' ■ : :.v ^ Cf'X ?
% • •' i f 2 J •’;( GUOUISl £30 tfOZ
,
• ' . v..
•
■
•
0 t
}
*
o
-
.
. • t ;
»
*«
, .: :
*
•
.
' 0
■ 0
.... f-y
■
■
■
■ ■
■ 0 C ■- /' v
■ , ' , ' ■ .
.
- 15 -
Napthall (26) has offered explanations for
the action of some of these promoters. Copper facili¬
tates the action of the catalyst by lowering the
temperature necessary for the reduction of carbon
monoxide. It is effective with iron and cobalt but
not with nickel. Small quantities of alkali in iron
catalysts favor wax formation. Thorium, manganese,
and similar metals, when precipitated together with
the contact mass, appear to function by distorting
the metal surface enhancing the activity. Space
lattice distortion changes the inter-atomic distance
to permit cross linkages, thus affecting the chain
length and the iso-normal paraffin ratio.
In general, cobalt catalysts are promoted
by thorium or by rare earths while iron catalysts
are promoted by copper. The best carrier or support
for cobalt catalysts is kieselguhr. Iron catalysts
are not supported since all known carriers to date
exert only a detrimental effect.
In line with the general theory of solid
catalysts, there is evidence to indicate that the
synthesis takes place only at a comparatively few
active centres on the surface of the catalyst. Even
in the best Fischer-Tropsch catalysts, these few
active centres become deactivated rather rapidly.
The following possibilities have been suggested for
- 16 -
the loss of activity of cobalt catalysts:
1) Wax formation on the surface of the
catalyst - reversed by hydrogen treatment at 200°C.
2) Oxidation of the active centres by
water - reversed by hydrogen treatment at 400°C.
3) Carbiding - irreversible.
Underwood (31) has reported that, under
optimum operating conditions, the catalyst life is
from four to six months. This is only approximate
since life depends upon manner of operation, period
and manner of reactivation, minimum yield allowed and
economics. For instance, Fischer (15) states that a
nickel-manganese-alumina will remain active one to
two years provided that it is regenerated periodically.
In general, hov/ever, the life of Fischer-Tropsch
catalysts is relatively short and it is desirable to
lengthen it.
It might be mentioned that while cobalt
catalysts can be regenerated by a hydrogen treatment
(at either 200°C or 400°C depending upon the circum¬
stances) this is not possible for iron catalysts. If
for any reason, the latter loses its activity, it can¬
not be reactivated. Permanent loss of activity for
both cobalt and iron catalysts has been attributed to
the loss of active centres through volatilization as
carbonyls.
; . ' •„ ' ' ■ ‘.-ii, ■' ; ; • rf.T
ot. •
•c
. , i*. v" , » . m
• ■> " : ■ ' ....
, ' ' ‘ *. ■■ .
. . • . ■ ,
■ • t ■ . .
• ■ .'.:' -■ ' = : . l . , • ..
• •
•' “' , .
■
, .
;
:
:o ; • . ...
.
, co -r
- 17 -
Closely allied with the problem of deactiva¬
tion is the question of poisoning. Fischer-Tropsch
catalysts are characterized by their high susceptibility
to poisoning, particularly by minute quantities of
sulphur. Herington and Woodward (21) have investigated
the effect of small quantities of hydrogen sulphide on
a Co-Th-kieselguhr catalyst. The first addition of
hydrogen sulphide caused a marked increase in the yield
of liquid hydrocarbons. This continued until more
than 8 milligrams of sulphur had been added per gram of
catalyst. Further addition caused a drop in total
hydrocarbon yield but this could be offset by raising
the reaction temperature. The addition of hydrogen
sulphide in quantities over 34 milligrams of sulphur
per gram of catalyst resulted in a continuous decrease
in activity until complete poisoning was reached. How¬
ever, even as little as 10 milligrams of sulphur per
gram of catalyst is sufficient to prevent the periodic
regeneration of the catalyst. Sulphur poisoning is
irreversible. In commercial operation the maximum
tolerance for sulphur content is 0.1 grain per 100 cubic
feet. Myddleton (24) has reported on efforts to develop
sulphur resistant catalysts but the results have not
been conclusive.
Fischer-Tropsch catalysts are also poisoned
by minute quantities of oxygen gas. The latter appears
.
' . . : : : .. ;.0 ; :
■ ::
-
• ,.:;l lly.V.I OCtf.l
* , ■ •;
• •
.
»
‘
;
• -
r
’ ~ -
,
>
1 ' • ■
'
- 13 -
to attack the active catalyst centers preferentially.
However, cobalt catalysts which have been poisoned in
this way can be reactivated by a hydrogen treatment
at 400°C•
Operating Conditions for Fischer-Tropsch Catalysts
The various operating conditions-temperature,
pressure, space velocity, and composition of synthesis
gas - all have a marked effect on the behavior of
Fischer-Tropsch catalysts. Fischer and co-workers
have done considerable work on this phase of the
problem in an effort to determine the optimum operating
conditions for commercial plants.
Fischer-Tropsch catalysts are very sensitive
to temperature - the maximum range of variation being
15 centigrade degrees. Storch (30) reports that, even
for the most active cobalt catalysts, the reaction
rate is very slow below 175°C and above 225°C and the
production of liquid falls off sharply, with methane
formation predominating. The best cobalt catalysts
operate at a temperature of 195 - 200°C. Nickel-
manganese- alumina catalysts, as developed by Fischer
(14), operate between 190° - 210°C. As the catalyst
slowly loses its activity during operation, the temp¬
erature has to be raised slightly. Higher temperatures
favor methane formation with a corresponding loss in
■ ' rti
- . ■ : !■.. • roll
.
*
■- ; J j" : 0"s ro %
,
■ : • . T .
■
.r .
- •
■ ■ ■■•...
30 * ‘ v -!a s; ul , 0 e < :,'■■■ . - , 0 -
■ >r ‘ : ■ ro ■ .
- " ■ ■
i-' ' 'Lt 1 , \t
' • ■ • r eeeai
‘
*
*
.... .
- 19 -
yield of liquid hydrocarbons and furthermore causes
the catalyst to become irreversibly poisoned, possibly
due to carbon formation. In this connection, Hofer’s
observation (22) that the upper temperature limits of
the Fiseher-Tropsch synthesis and the lower temperature
limits of carbon formation (i.e. from the metallic
carbide) very nearly coincide, may be of significance.
For iron catalysts, the optimum temperature is around
250°C but with higher temperatures, carbiding and
carbonyl formation deactivate the catalyst irreversibly.
The synthesis pressure affects both the yield
and type of product, and the life of the catalyst.
Earlier work by Fischer and Tropsch (16) has shown that
high pressures (100 - 150 atmospheres) favored the forma¬
tion of ’’syntho!’ 1 or oxygenated compounds. On reducing
the pressure to 20 atmospheres or less, the reaction
produced hydrocarbons almost exclusively. Fischer and
Pichler ! s curves (13) showing the effect of pressure
on total and fractional yifcld of products using a
Co-Th-kieselguhr catalyst are reproduced in Figure I.
■■ ■ -Q ■. .v r ’ . '
i
t ' o-O.
-
■ ■ .
* .. ' -c . ' ■ . 0
' • ’■ ' . .. . 'j r :l v."\ lo >■; ■ .. ■ ij ■
' . ' - ■' . .
.-'.O.'J .v.y,
, • ■■ . .
i ' ' . . t ..
• ■
: ' ■ ■ ' : ' •
'
" '° r '- . - ' ; v. . ? *, j. . . u „
J .... .. v
Figure t * Effect of Pressure on Hydrocarbon Yields
Weeks of Operation
Figure II--* Effect of Pressure on Catalyst Life
- 21 -
A pressure of 5 - 15 atmospheres gives the greatest
degree of conversion and also the highest yield of
solid and liquid hydrocarbons.
The effect of pressure on the durability of
a cobalt catalyst was shown by Fischer and Pichler (12).
The results are shovm in graphical form in Figure II.
At both low and high pressures, the effectiveness of
the catalyst decreases markedly in the first few weeks.
A pressure of 5 - 15 atmospheres appears to be the
optimum for maintaining the life of the catalyst. These
results have been confirmed with iron. Komarewsky and
Riesz (23) report, however, that these pressures are
unsuitable for operating with nickel catalysts.
The space velocities employed are in compari¬
son with catalysts for other syntheses, extremely low,
being, in general practice, in the range of 100 - 150
per hour. This leaves much room for improvement.
Fischer and Pichler (12) have investigated the effect
of varying the throughput on the yield of liquid hydro¬
carbons obtained from a Co-Th-kieselguhr catalyst.
Their plotted results are shown in Figure III.
At the maximum point in the space-time-yield
curve, the yield of oil is only 60 grams per cubic
meter of gas indicating incomplete conversion. If
one-pass operation is practiced, such yields are
probably not economical. Aside from this, another
,
.
: I . :
. • r
. a ... :. ;.\/v
... • i
. , , A
■. : v:
. :> • ' '■ a £ -- a : .aa; , at./
,
■i ■ \:o
.
;
■
.
• i 1. •
•
.
A, ■ • .
> ; ■
•. ■:> i.;....
, > '!£•
. A.
7
, ■ ~■'
.
; i-.
• L
• •v 1
■ ■ ■ ' ■.
*
; V .
.
.
11 .<■
. •'
* •
r.;. / v ‘ -. '
~ . j,
. -A . • I ■
• o a! a
-
", . a a •
a "
- 22 -
factor limits the space velocities. Since the reaction
is strongly exothermic, a high space velocity presents
a serious heat exchange problem. If the heat of
reaction cannot be dissipated effectively, this causes
local heating of the catalyst, resulting in the sinter¬
ing and deactivation of the active centres. German
industry has attempted to solve the heat transfer
problem by using a relatively inert catalyst, high
space velocities, and recycling. However, the forma¬
tion of gaseous inert products (CH 4 , CO 2 ) leads to
serious recycling difficulties. Two stage operation
appears to be the compromise (30, 31).
. /
i .t
■
*
_
.. a; :• o a . t ' r> an
:± 1 . ■ "j -C: .. ,-i’d '■ . ..i ... ’ av J
jo : ' ...
. v ■
.
•'■q i
. . } c
Yield of Oil
./oxi'.. meter of gas gms./lOGO hrs./gm, of catalyst
- 23 -
- 24 -
Fischer and co-workers (15) have studied
the effect of varying the CO:Hg ratio in the synthesis
gas. They recommend that the stoichimetrical mixture
containing 33.3^ CO and 66.6 % Hg is the most desirable
one, giving the maximum yield of products per cubic
meter of raw gas per unit time. In addition this
ratio gives the highest degree of conversion to liquid
hydrocarbons.
The reaction rate decreases as the COiHg
ratio rises above this optimum. The yield of liquid
products increases as does the unsaturation. On the
other hand, a low CO:Hg ratio results in a greater
degree of conversion and more saturation. The effect
of the CO:Ho ratio can be somewhat minimized, however,
by the proper selection of a specific reaction temp¬
erature - this temperature being higher for CO rich
gases than for Hg rich gases.
Inert gases like nitrogen have only a diluent
effect and do not affect the reaction materially.
Type of Products
The type of products obtained in the Fischer-
Tropsch process depend upon both the operating con-
1 *
ditions and the catalyst employed. The effect of
varying the operating conditions has already been
qualitatively shown. The products obtained from the
. . ' ■ \ ' : :
/ .. .,v ;
0. L .... ■•....■•
.
.
. '
: ■ ' "■■■'■ ‘ ■ ... • i : , i -i .
. . x3- ■■
* • .
■ :«*0 ‘
• ■ ■ ■- o .• ■ o :1 .to
■ ■■ y ■ G'.;i q
■
* .t
' ...
j
.
. "Li.
- d .
-
•
,
•
•<;. /• . •... ^ Jj_o_
r : i :j -go* 'g; ... ■ r
.
.
' '■ d-;' ... : . r
'
«*
- 25 -
two general types of catalyst (iron; and cobalt,
nickel) show considerable differences. Most of the
literature deals with cobalt catalysts; data on iron
catalyst products are relatively scanty.
Napthali (26) has reported that cobalt
catalysts give a more saturated product than the iron.
Fischer (11) using a cobalt catalyst obtained the
following primary product or Kogasin.
Table II
Analysis of Kogasin
Fraction
B.Pt. °C
% by Tift •
% of Olefins
C 3 and C 4
Belov/ 30
8
50
gasoline
30 - 200
60
30
Diesel Oil
Over 200
22
10
Paraffin Wax
10
Kogasin is largely composed of straight chain
saturated and unsaturated hydrocarbons. The gasoline
fraction constitutes 60 % of the total product but has
a low octane number.
On the other hand, the U.S. Bureau of Mines
(32) find that at 7 atmospheres, and using a cobalt
catalyst, the product had the following approximate
composition by weight:
.
;: $
.
, - . .......
'
.
. t . • :q t i : .
. / ■ Jvo <: {[£) os.OO
»
•
• ... - * . .. ■
» V .
U b
co ; -
0C_: -
**’•*' * ■ «- >■• n . . •
. ■ ,
m
. :) :
... , •
- ■ i
: •’ ' • , .t-j ‘iOv . >
Percentage
10
Fraction
and
20
40 - 150°C B.Pt.
30
150 - 370°C B.Pt.(Diesel Gut)
40
above 370°C
These sets of data should be regarded as
only illustrative of the type of product obtained.
The character of the various fractions have been
revealed by several investigators (26, 30, 32). The
diesel oil cut may be removed by simple distillation.
It is of good quality, having a high cetane number due
to the straight chain character of the hydrocarbons in
this fraction. The paraffin wax differs in properties
from that obtained from petroleum but can be put to the
same uses. It is characterized by a high melting point
and contains relatively high proportions of long straight
chain hydrocarbons. The gasoline fraction may be in¬
creased by reforming and cracking the primary product in
conventional petroleum plant equipment. The octane
number can also be raised by these methods. Egloff,
Nelson and Morrell (7), by means of thermal treatment
and catalytic polymerizations, obtained 84,5% of 66
octane gasoline from a kogasin oil which originally had
an octane number of only 20.
• : .
. ; • .0
. .
' 0 , J':. ’:.0 B'los sr>,
’ 6xrf 0- p ; 16 '-‘'OfiX
' o ic ' . \o
. . - .
: : oo *:c- C? ,o
.
,
ft'iq t;M . : . d f ' •
.
*. 1
i r ■ x
-
* ■ ■ ’
’ *
■ • ' O ' I ■ ' J ...: 4:00
“ 27 -
4
Little is known about the product from iron
catalysts except that a higher percentage of olefins
(and hence a higher octane number) is found in the
primary product.
The development of new catalysts to direct
the reaction towards specific products has been
recently reported (6). Of particular interest is the
so-called "iso-synthesis M in which iso-paraffins are
produced directly. High octane gasoline fractions
are obtained which may be converted to aviation fuel.
The literature reveals the wide versatility
of the Fischer-Tropsch synthesis. A large variety of
products is possible through the proper choice of
catalysts and operating conditions.
Mechanism of the Reaction
The overall chemical reaction over Ni, Co
catalysts differs from that over Fe catalysts mainly
in the by-products produced. Along with hydrocarbons
the latter yield CC >2 while the former give principally
HgO. The reactions - assuming formation of paraffin
hydrocarbons only - may be represented by the follow¬
ing equations.
1) For the Co and Ni catalysts
nCO -»- ( 2n +1) H -—.> CnH 2n+2 +nH 2 0
2) For the Fe catalysts
2n CO + (n 4 1) H 2 -—» CnH 2n+2 4nC02
. • . • • ..
v . ' ■„> ' }
■
.
■ • o ' , ,j a ■ >
■O' •'.■■ 1 : 4 .. ■ . :,r
its t ii . ■•■ - . ■ ■
'
o • . i . c.■ ©d x ■ ■ ■
' C '•
■ d <1 -To ■ j ■': f. o
.
zd L- '1 .
. *** *' / • . • . . • i ... V— . ■*
•. ■ . a € f :
O . d: r.,.: • . ■,
.
... 1 ' . s 0,' blel , .. :
;
- ", ' j 'I ,":V - '■
■ -
. :: .j ,-m ’
■" •. .. 'j ■ ■,
- 28 -
This difference in behaviour has been one of the main
difficulties in the determination of the reaction
mechanism. Much research has been done on this question
and several theories have been advanced.
In 1926 F. Fischer and H. Tropsch (17) follow¬
ing their initial production of paraffin hydrocarbons
from water gas postulated that the catalytic process
involves the formation of a carbide, rich in carbon
(Fe^C^ or FegCg) which is later converted to methylene
groups. The overall effect may be represented by:
3 CO + 3H 2 —> C0 2 4 H 2 0 +- 2CH 2
The methylene groups polymerize in the presence of
hydrogen to form the paraffinic compounds. An Fe - Co
contact mass was used in their investigation and, as
a result, both CO 2 and H 2 0 were present. In a later
paper Fischer (10) reiterated his initial idea but
suggested an alternative. He stated that the methylene
groups first polymerize to extremely high molecular
weight hydrocarbons. These are subsequently cracked
into compounds with high vapour pressures and carried
off the catalytic surface. The actual mechanism is
suggested as follows:
CO + Me - > MeC + MeO
MeC +* MeO +■ 2H 2 ——2Me +• CHg 4 - HgO
However, it is probably more correct to assume that
the original carbide is merely reduced to a lower
' ■ - 1 \: i. ' 1'h
4
.
.
; ;: 1 .. . , fc
, 0': . d ; . ;.o- .v S ;
■ 1 ■ : ; ■- :■ '
,
. ■ ' .
< t '
■
v '
o'-.:;:.
.
'
. ■ : .
■ ■/ ■ . . : ,
• J ;?:■'/ . ) ■: aa ..C '•!
* ..
.
■ ■ ■ , • ■ a '
; : ■ : :Khl, ■!-;
' /. O'O
> . • < - 2 .... ,
- 29 -
carblde rather than to the metal. No attempt was made
to explain the COg formation.
In an endeavour to explain the formation of
oxygenated compounds, Elvin and Nash (8) advanced in
1926 an alternative theory. The hypothetical formation
of methanol is postuluted as the intermediate reaction
for obtaining both hydrocarbons and oxygenated compounds
(acids, alcohols, ketones and aldehydes). This, how¬
ever, is not supported by thermodynamic considerations.
This theory has found some support with Japanese investi¬
gators (3).
The presence of carbides in catalysts was
confirmed by experiments (22) involving the Garb idl¬
ing and reduction of the three metals, Ni, Go and Fe.
Up to certain critical temperatures, when treated with
CO, all three form metallic carbides which can be
reduced completely, giving CH^ as a product. Above
these temperatures (which are but slightly higher than
the optimum for the synthesis) elementary carbon is
formed which, by carrying away active centres and
producing irreducible residues, tend to poison the
catalyst. There is evidence to indicate that the
carbides are the real catalyst bodies.
Craxford (1, 2) in supporting and developing
Fischer’s carbide theory, has presented a detailed
mechanism of the reaction. He argued against the
' '
■*
■ ■
<
.
:; ■
,
■ a \ ■
} : ■ V
■ :■ ■;& . i v.; ./c
'
~
■
• . .... tv. • . .’(V .
. ;• . .. ,/ • i.
;
. : r ';
.
* ..
'
. • ■ '• ; ! • -V . . .
■
■
- 30 -
formation of hydrocarbons by simple carbiding and
reduction as expressed by the following reactions:
1) 2C o ■+• 2G0 -^ C o 2 C ~f~ C 0 g
2) OogC *f Hg > CHg-f 20o
First both reactions are too slow to account for the
rate of the overall reaction:
3) CQ + H 2 -^ CH 2 + H 2 0
Secondly over Co catalysts the production of HgO can¬
not be explained by equations 1) and 2).
As a result of experiments involving the
rate of ortho-para conversion of hydrogen on the surface
of an active Co-Th-kieselguhr catalyst, Craxford con¬
cluded that the reaction forming hydrocarbons proceeds
by way of molecular Hg. The following steps in the
reaction (shown diagrammatically in Figure IV) have
been suggested:
1) CO is chemisorbed to the surface and reduced
by molecular Hg to form the carbide.
2) The carbide is reduced by Hg to form methylene
groups.
3) In the presence of large amounts of chemisorbed
H 2 , the methylene groups are immediately reduced to
methane; with smaller amounts long chain hydrocarbons
are formed
. ' :• ;:o 1; ' :v .■
■ £flj -
V ' 1 \ >
.10
. • j i o ■■ nib Ido • d 1 *z £
■
; ■ , ;..'i;
,
...1 ad >i. ' o ' , 'y. :
■ .:: ■;r . a . : : — 1 d o> ..
. y' - a
. ■
. /X \ c • . ...
; ■ ..
■ '■ i ■■■’ l
?>• ■ :
w& . t 1(1
. .
■
... ..
-31-
Legend: oooooo Yan der Waal Adsorption
- Chemisorption
(1) Chemisorption of CO
Reduction of Chemisorbed CO
by H 2 to give Carbide
H 2 h 2 o
: c :
O o
(2) Reduction of the Carbide to Chemisorbed Methylene
I-I 2
o
o __
( 3 )
a
Presence of a Large Amount of Chemisorbed H 2
Hj
H
CH 2
CH,
b. Presence of a Small Amount- of Chemisorbed iL
CHo CHo
-OHo — CE ft --CE*~
h-2
CE a -GE a -CE a -S CHa-CHa-GHj
CH a -OE 2 -CH 3 H CE a -CE a -CE 2
o
o
o
o
CEa-CHo-CHs E
CE a -CE a -CE a
CE a -CE=CE a .
>
Figure IY - Mechanism of the Synthesis of Hydrocarbons
(Craxford)
-32-
This mechanism supports Fischer's idea of
the formation of high molecular weight hydrocarbons
subsequently followed by hydrogenation or 'cracking'.
Further experiments involving the breaking down of
ethane, propane, and butane (mixed with the requisite
amount of H 2 ) into CH^ over a freshly reduced Fischer
catalyst substantiated this view.
While Co catalysts are usually considered
very active for the water gas shift reaction:
C0 + H 2 0 —¥ co 2 +h 2
the primary oxygenated compound is water rather than
C0 2 . No other explanation for this phenomenon is
given other than to assume the reaction is inhibited
by the operating conditions of the Fischer - Tropsch
synthesis. Vi/hen the temperature is raised to 250 -
300°C, C0 2 is again produced exclusively, with no
hydrocarbons save methane.
To conclude then, while the Craxford
mechanism is valuable, it is nevertheless complicated
and is so far limited to Co catalysts.
Herington and Woodward (21) conclude that
the Fischer - Tropsch catalysts possess two types of
active centres:
(a) Type A (probably cobalt carbide) which is
responsible for the formation of methylene groups.
, ■ ; : ■ ' O' , ill -
• . ; - . . * •• n.’i
.
■
... a * ■■ . . J-'i.u
.
, .
'
: ... 0, ; ...
.
■ - • :.G £'.:c \ ;v<-> .• no uilJ ;c
\
... :
........... , o
. .. .; :
. . ' N or.
■ '' V . . .. :. i>:.
. r -
f
33-
(b) Type B (probably cobalt metal) which is
responsible for the adsorption of Hg to liberate the
products from the catalytic surface.
Myddleton (25) has offered an explanation for
the difference in the by-products with Ni, Go, and
Pe catalysts on the basis of the chemical bonding of
the CO to the surface. With dual bonding to adjacent
atoms (as in the case of cobalt and nickel) HgO is
produced whereas with single attachment (as with Fe)
CO 2 is produced. Further experimental evidence is
needed to support this claim.
In general, the carbide theory, advanced by
Fischer and developed by Craxford, appears to be the
most acceptable at the present time. This mechanism
has been developed primarily for Go catalysts and
little work in this connection has been done with
Fe catalysts. It is almost certain, however, that
the formation of iron carbides plays an important role
in the mechanism of the latter.
■'ov. 1 . \ ... v?:'! ■: r ( ■ i
. . ■ , . . to I.b .
,_o0 ,1 ■ : . 16 ( r - , | :
.
.
£, • ’
■ ■ I .'. :
' , ' . . . '
■ ; . • oV>,\c i\, /;; ... an out .■
■ ' : ■ v. •.fi
." ■ ■ i . . f .
;
,
, ' ' : f .
• '.J - - ■ , ... . ;• . , *.V iC . :/ ',■><.{i
•i.
■c
*
-34-
IIX EXPERIMENTAL EQUIPMENT
The Fischer-Tropsch process is composed of*
essentially two distinct stages - the preparation of
the synthesis gas and its conversion into suitable
hydrocarbons. The description of the equipment may
also be divided in this manner. The preparation of
the synthesis gas will be described in detail in the
Appendix. The emphasis in this section will be
placed on the equipment used in the actual synthesis.
Catalyst testing is carried out on a pilot
plant basis and the equipment is operated continuously
for long periods - extending into several months. The
following requirements must be satisfied - the passage
of water gas at extremely low space velocity (100-150
per hour) over a catalyst column at medium pressures
(100 psi.) and temperatures (200° - 250°C) with good
control in general but particularly with respect to
temperature.
Much of the present design, the reaction
converter, and the recovery system in particular, has
. ' O tt
; {.d • ' ' -
„ "
~ .
..
t- • • •'> : ..-d sr^.L' cJ./ C
. :ii o • x ■ if. i if' # .< i ... . j: ...'
■
;
*
- ' ' . < a 7
: • ■ ' .. ' ' - lo ..J uo o
. •
• .' ■■ * • ■ . . i... 7 : O vj
-35-
been based on the experience of the U.S. Bureau of
Mines (Department of the Interior) Pittsburg, Pa.
However, at the University of Alberta, much more
automatic control (especially in the control of gas
flow) had to be incorporated into the design. This
was necessary because only a comparatively small
staff was available for work on this project. There
are six individual catalyst testing units, each of
which may be operated nearly automatically^and
independentlyof each other.
(a) General Flow S heet
Before giving a detailed description of
the individual pieces of equipment, a short disciission
of the general flow sheet (Figure V) will be first
presented. This will afford an overall picture of the
whole process and will bring out the relationships
between the various pieces of equipment. Only a brief
descriptive paragraph on the synthesis gas production
will be given here.
Methane and oxygen, in the volumetric ratio
2 : 1, are passed simultaneously into gasholder A,
the flow being measured by two capillary type flowmeters.
The mixed gases are passed at a controlled rate over a
nickel catalyst heated to 1000°G in the electric furnace
B. Here the conversion into synthesis gas occurs, the
„ .
: ■■ . v v <:
• • S l ■ \ri iO 0 i !.orf .
, -'1 •
. .; .■ i . s :-' ■ r■■■■.' . ■:
... : ' .
. ' - $$% . M, . J ' ■. |
\:l a: ' ■ ■ >qo ■' srjv ' :
111 "r
’
- ■ : ■ * ■ „• . m ^
■ • . . . . • .
. *1
. . . ,
; ' - . i
-
.
. . . ;
^ 1
J®
c ; .
1
-36-
main reaction being
2 CH 4 + 0 2 -* 2 CO +4 H 2
The exit gas is purified (of H 2 S and C0 2 ) by a caustic
scrubber C and is then stored in gasholder D. This
low pressure gas is compressed to 100 psi. guage and
held in the high pressure storage tank E. The pressure
here is maintained constant by automatic control, the
movement of a Bourdon tube actuating a relay and magnetic
switch in the power supply to the compressor. High
pressure synthesis gas, after being dried in the alumina
dryer F, is passed into a header common to all six units.
The gas flow to each unit is indicated by the
pressure drop across a capillary-tube flowmeter, the
pressure drop being indicated on a mercury manometer,
equipped with high, low and intermediate contacts.
The manometer contacts are part of an electrical
circuit which operates a motor driven control needle
valve G s designed to keep the flow constant within
narrow limits. The gas passes into the reaction
chamber of the converter H containing a column of
©
catalyst. The catalyst bed (approximately 16 inches
in length) is supported inside a inch iron pipe.
1
The catalyst temperature is held constant by a liquid
boiling at constant pressure in a jacket surrounding
the reactor. This outer jacket forms part of a closed
system which also includes the water-cooled reflux
--V
.
;:C : v ’ ■ --o-:
, . d o :: ... - - . 'OX
.
. . : ; . - '".l
;
.
■ : .u , : Li . ::.o *;q
,
•, ' ■ 'i ■ ' .,c : e; •/ ; :
■ o s Si , ■ nc . • i ■■■:. . ..... . • . . . •
,
■ o n vr
.
: '.ot 'jt
.
. - ' : • ■ . t ■ . '
■ • r <. . J : o
-37-
Fi qxto T - General flow sheet for testing of Fis olier
Tropscli c-t^lysts.
38
Figure VI - Front View of Catalyst Testing Units
- 39
Figure VII - Main Panel Board and.
Equipment for Synthesis Gas Preparation
-40-
condenser K and the ballast chamber L. The tempera¬
ture of the boiling liquid is changed by varying the
pressure of an inert gas in this system.
The liquid and solid products of the synthesis
are partially separated and collected in the two pots
M and N. The hydrocarbons with boiling points higher
than 100°C (at 100 psi.) are collected in a steam-
coil- jacketed pot M; the hydrocarbons with boiling
points between 0°C and 100°C (at 100 psi.) in the ice-
jacketed pot N. Pressure is maintained constant in
the unit by a back pressure regulator P. The gaseous
products, in passing through this valve, are reduced
from 100 psi. to about atmospheric pressure. Water
and carbon dioxide are removed by a calcium chloride
dryer R and a caustic scrubber S respectively. The
volume of the residue gases is measured by a Precision
Wet Test Meter T.
For purposes of reduction and reactivation
of the catalyst, and for the maintainence of a reducing
atmosphere over the catalyst during 11 shut-downs 1 *, a
hydrogen supply is included in the flow sheet. Hydrogen
from a high pressure storage cylinder is discharged at
a controlled rate and at pressures slightly above
atmospheric through a two stage pressure-reduction
valve. Passage of the hydrogen over metallic copper
at 250° - 3Q0°C in the furnace J serves to remove any
* ' ‘ . : ■. ; '. ■:
' ■ '_3 ,■
* • ..
'7 ' ■; ...
■ • ...
-
‘'.O'
..
,
:X: '-ti
.
—
*
;: o .. ;.x: o : ■ n :Q
-
■ - ■ ■ . . 1 ■/
■
*
' .
■lu . . . .&/ :c, . O
• y - > ; ... \ ■;> •, .. 1-; ■ 7 ...
*
.
■ ■ v . r ‘. ... ,. c '. V. . ■.
-41-
oxygen, while activated charcoal in the scrubber V
preferentially adsorbs any high molecular weight
gases^leaving relatively pure hydrogen gas. A common
header services all units.
(b) Description of Synthesis Equipment
This section is not intended to give a
detailed description of all the major pieces of syn¬
thesis equipment used in these investigations.
General descriptions will be given of common equip¬
ment and details will be given only for equipment
that presents new design or new unique features. The
equipment required in the preparation of the synthesis
gas will be discussed in the Appendix. Special
features incorporated into the design to control the
experimental conditions - temperature, pressure and
rate of flow - will be dealt with in the next sub¬
section on w Instrumentation and Controls. 9 *.
Since the Fischer-Tropsch process involves
the synthesis of gas under pressure, two types of
storage are required - a low and high pressure storage.
A conventional water-seal type gasholder D serves as
the low pressure reservoir (Figure V) and a pressure
tank E as the high pressure reservoir.
Gasholder D has a capacity of approximately
15 cu. ft. and the floating drum is counter-balanced
. • ■ , ,
■'0 .1
. .
, : ■ ■ : ... ly*sq
*
' ■
: - ■
1-
♦
; ' . .Vo 7 i.:
J.
). •; :
• • u
.• ::j." 'V:.,
.
0 0
i.O . ,7...
■•j'..
7 7.; 7 I. '
!
' r -j
’
..
r.
.
yf'yi:
. : - . ■ \7,
J . ■' -j'
■;;.o
: ' •
■
Hi vo
•
i
'
'0 : ■ •X.i *.(
— y. .••if..
-
. .
W7 17 V j
* .
o . j, $
' ■;
• 0 . ;,
i.l
'• C 0
-
' 3
- -•
"MO .
■ :
*
'
o
if
, ■ v.
-42-
to maintain a gas pressure of about one inch of water*
Two safety devices, in the form of upper and lower
limit switches, have been incorporated into the design*
The lower limit switch consists of a "microswitch’*
and is attached to the outer fixed drum of the gas¬
holder. It is actuated by the floating drum when it
reaches its lower position and breaks the electrical
circuit of the compressor motor, thus eliminating the
possibility of water being drawn accidentally into
the gas compressor. The upper limit switch consists
of a microswitch attached to the top of one of the
gasholder guides and is actuated by the floating drum
when near the full position. A warning bell is sounded
and steps are taken to reduce or discontinue synthesis
gas production.
The high pressure reservoir E consists of a
standard compressed air receiver, approximately 15
inches in diameter and 60 inches long, of about 6 cu.
ft. capacity. The pressure is indicated by a Bourdon
tube pressure guage and is maintained constant to
within 0.5 to 1 psi. by means of an adapted Mercoid
Pressure Control operating on the compressor motor
circuit.
The gas is compressed by a small single-
acting compressor with a 2 inch stroke and a lj inch
piston. It is fitted with a solid plunger-type
'■) f.:X
... - ... ■
.... .
. . J 1 .
,
U.
.... 0
■ c ..
\J: . ,;.7o
3,' 7 ' 77.
: . : ■;
"7 oo o
1: '
■t
■ / -•
' 0 . 1
'. ■ 7
■ . : ■■•... . 1 ,.
■ ■
. i ■ i
3 .1 J . V.'
.
07; ....
.
o
7 7 . ■ : ; , ■
ft
o. r. .j.t
.' . :7 , : ",
. o'. ;
r j
. '
: 0
0 . ,
; ■ . J .jo ■
( . • .
- I f. ■
, U .
.;. ..
"
; ..
. .
ov:'o
■■ 'o. . ~
»
2
■■ ' ■ .
■I , 1 .,.'.;
■ • * - ■ * X
' 7 'G
.
7 : : •„
;... 0oi '■
7.7
I o' od cw
,
•
-O : 7 ■ ' '
< 1 :
1 .. : a
o ■J.:cvr.
■ >.,■ ...■ ’ .
% .. *
\ .... L
07 J
1
•: . 1 Vl
'
j . r "\
... ; . • . 7 .
' .
; . 00 ; >J.
i.6'uOO
r / .
■
•
,
'
,
y
: '
.. .• .■ V
*
' : • ' ;..
■
.
o • . o
..
,
- 43 -
piston, an external packed gland, and poppet type
valves. There are no piston rings. Lubrication is
supplied by gravity through the intake valve. The
compressor, driven by a J h.p. 1700 r.p.m. induction
motor through a belt drive, operates intermittently,
always under pressure, transferring the gas from the
gasholder to the pressure storage in response to the
mercoid control. A check valve in the discharge
line allows the compressor to be unloaded and prevents
leakage from the high pressure storage.
The AlgOg drier F is a two foot length of
2 inch standard iron pipe, filled with dehydrated
alumina and is wound with a heating coil, the whole
being encased in Q5% magnesia pipe insulation. Water
is removed from the synthesis gas at room temperature.
The heating coil permits the reactivation of the alumina
at a temperature of 300°C.
The six converter units, identical in both
design and construction, are connected to common
headers but isolated one from the other so that the
operating conditions may be varied in each. They are
constructed of standard-weight, butt-weld iron pipe
and standard malleable iron pipe fittings. Each
unit consists of principally three sections:
(1) the catalyst chamber (2) temperature control
system (3) liquid product recovery system. A
: ■
j • 1 . • 7
- ,'.'1^ ■' i : .'.. . ■ ■ ; -j
. . . L . , . . •
• ■ - •' , ■ • ' J ..... . . ' O': : . NO 'on;
' N ... ......
• • O' so' t ■ ■. , n ■ ■ • | .. p ; 'd £.;
. .
- -' ■ Cft s ■ J )i i ■ . . . ..
* £
iOl QQl 3 . >. ■ ■ ■ ) ■
il > : <
■ ■
«■
.
' ■ . .
n ■ . p i
) 'v
I
...
8
. .
;
■ ■ . ■ '. ’•
.4 . . ...
'. ■, i ., 0
ua ii ■ lo
.
k.« 1
1 i’i o
■ ■ ,, .
.. .. i.
>0
’’ _NN ...
. 0
0 ■ . : I )
• ■ N."„ : . •. .
- 44 -
drawing of the converter and its accessories is shown
in Figure VIII.
The catalyst tube A is a | inch standard
pipe 27 inches long. Screwed on the upper end is the
catalyst chamber head B, machined from a 2 inch
hexagonal steel bar. A thermocouple well C of j inch
steel tubing, screwed into the head, extends the
length of the chamber. Provision is also made in the
head for the introduction of the catalyst into the
converter. A piece of 3/8 inch pipe D with a fine
wire screen welded to its upper end, is screwed into
the lower end of the column and serves as a support
for the catalyst. The gas enters the unit through a
i inch pipe E and passes down through the catalyst
bed. Connection is made to a Bourdon-type pressure
guage by means of a J inch tee which also serves as
a connection for the gas feed line.
A heating jacket F, 20 inches long of 22-
inch standard pipe with ^ inch steel plate ends^is
welded to the catalyst tube. A liquid, heated
electrically by a resistance coil encircling the jacket,
is kept boiling within the jacket maintaining the
catalyst tube at a constant and uniform temperature.
The boiling temperatures are controlled by the pressure
in the heating jacket. The temperature of the liquid
i 3 observed by means of a thermocouple inserted in the
. ■. ' J ' .o voA ^
.1 . . '
' ' . )[ f • 0 d t tf . ! .
,
,
' . :
\ . :
. '.. . : :i> / i ■ ■ o . ./
... ■ m
-■ , ■ .
*
f • ' o ■
- -
' v< . r or:.';
J/i. •
. .
x.X k
,
£ 7 ■ '■ o
... ...
•
;: 1t
■ is
..... 'j .
. .
j'J..
' •'/ a ; £
.
.
0
; ' 1 ■ ■ . C‘ : VI ...
-45-
-46-
thermocouple well G. The liquid level is indicated
by an open end sight gauge H. When necessary the
valve J shuts off connection between the chamber and
the gauge, allowing pressure or vacuum to be built
up in the former and allowing the latter to be drained.
To prevent excessive heat loss, the jacket and
catalyst column are covered with 85^ magnesia pipe
insulation K with asbestos cement at each end. A J-
inch pipe L with its upper end enclosed by a water
condenser M is connected to the jacket and serves as
a reflux column for the boiling liquid. At the upper
end of this column is a Fenwal thermoswitch U which
breaks the heating circuit if the hot vapours rise
too high^in case of failure of the cooling water
supply. Here also are connections to: (1) a compound
»
Bourdon tube pressure gauge (range of 30 in. Hg vacuum
to 60 psi.) (2) a ballast tank made from an 18 inch
length of 3 inch pipe and (3) a natural gas header
common to all six units and connected to a high
pressure gas storage. The liquids used in the con¬
verters depended upon the temperatures required, and
were tetralin, dowtherm and a paraffin oil of initial
boiling point of 360°C (700 mm Hg).
The recovery of liquid products is accom¬
plished in two stages - in steam-heated and ice-
jacketed condensers R and S resulting in a rough
*
■
■. . ’ Q'.-rcJ s>,' . .
'
' . ' . v' : ■ ,-v , . / .
^ OlU'jlO :> ,!■ x .-dy O
1 . :
‘I ' f-0 : .. • 71 , Jt .t ■ -
' b. ’.o . i cm . ••.. ; |nS:;ncu)
c !
ol . ■; ■ ....
i
■ '' V.. '-0 ;-j ' ' : . • :
'
■
,
■ ■ ' , ,• ; ■ - J • j.
0 '~ i . 1 ' . - • . : . i ‘j
\ ■
" 1 ■' ' -■■ ~ :' ‘i' V
.
« ' * ; ; V
■ • ■ I.: Z O v ;: r i
: r i - : ov.’d r:' r.
: ' •• ' . . u • .. r, L,-i ' . ‘
-47-
separation of heavy waxes and low boiling hydrocarbons.
Both pots made from 2j inch pipe, 7 and 8 inches long,
are only slightly different in design. Condenser R
has several turns of ^ inch copper tubing encircling
it and is insulated with 85$ magnesia pipe insulation.
Steam passing through the copper tubing maintains the
temperature at approximately 100°C. Both pots are
equipped with J- inch drain valves to permit the
removal of the liquid products. Pipe unions, connect¬
ing the pots to. one another and to the adjoining
equipment, enable the removal of either one of the
condensers.
The residue gases are bled off continuously
by the Cash back-pressure control valve T. These are
standard Cash model 1935, size in which the
makers replaced the usual bronze seat by a metal-com¬
position seat. The control desired, from 100 - 0
psi. gauge, and the small quantities of gas flowing
necessitated experimentation with the seat. The
material known as ”vulcanized fibre’* was found most
suitable, giving perfect control without drifting
and complete shut-off. It is probable that the valve
floats continuously when passing the small amount of
gas involved.
The expanded gases are treated for removal
of water and carbon dioxide, the residue gases after
; ' • or-
' - •’ ■: : •:- ,,oo :.i rioa
. ■
■
'.'s. 1 .!
'/ 4* —
■ o v. . ruo.i
u;:i
0. f -r or oG 8 „::>SA!
.
0 0 o’;: i‘
'-JJC 'l .iJ
■
> i ■ o _ ; .h? c .
- ■’ ■ v :j
hio l i .
A " \
' : .... , -
..oo oo
'&&&
•
4.
'O ; ..
•
0 . '
•
.
o r
v ■ ■■ o<:o
1 *" ...
, ■ i
/ ; '
■OsO..
! ‘ :■
. r.
■.
. ■
■*
- • ■
-
■ '
' -i
„ ‘
•
■3
’ . . . • ■ •.
0 .
-.a :0 0 ;
0 7- •
t.
■ i :
f V
•J'
■ -*
• - .
v •) or..
oc!r
•
• Oi.r.. ;.;j .to
ojoui'::
*
i v •: o :■
'-o’
.'.'0o
■’■'o ; ofMi
-48-
measurement by a standard Precision Wet Test Meter
being discharged to the atmosphere. Water is adsorbed
in calcium chloride in Schwartz ground glass stoppered
U-tubes; carbon dioxide is removed in a standard
potassium hydroxide solution contained in a fritted
glass type gas washing bottle (500 cc). Two driers
and two scrubbers in parallel, suitably connected by
stop-cocks so that either drier or scrubber can be
used, make up this phase of the system. A continuous
gravity-type gas sampler (not shown on the general
flow sheet), placed immediately before the gas meter,
gives a representative sample of the residue gases.
The rate of sampling is regulated by a capillary placed
in the liquid discharge. A 10$ NagSO^ and 10$ HgSO^.
solution is used as the confining liquid.
Saran tubing (a vinylidene chloride plastic)
is used for connections betn/een the various pieces of
the equipment. Its strength (at low temperatures only)
and its flexibility make it an ideal conduit particu¬
larly where many short circuitous leads are required.
Saran tubing and fittings are used to connect together
the header, the flowmeter, the mercury manometer, the
control valve, the catalyst column and the- pressure
gauge in the high pressure synthesis gas stream.
They also form an integral part of the temperature
control system, connecting the methane leader and the
ballast tank to the main assembly.
■ 52 ' : , ,
.
. 0 .
' »
"
I .0 i !J3’|
'
1 1 ■;
■ ! :;
.
’ ' ■' ■ ©‘iqe* . ' vl
' . ..
• • • ■
* ■ ' ' J ■■ ■' - 0 ::
■
,
.
' 1
. ' '.t' ' i\J: c;
:
;; ■ ■ ■ ' ... :■ ;
’
* : ■' ' ' ' List ;■
-49-
(c) Instrumentation and Control
The electric power required for operation
is supplied to a main panel board and from here to
the individual units. The instrumentation of the
main panel board is shown in Figure IX (a). Power
from a 110 volt line passes through a two pole line
switch A to the various units. A fuse B inserted in
the line protects the power supply. Across this
supply and in parallel are a light C and a relay D
which together serve as an alarm system. When the
power is on, the light glows and the relay is
actuated, opening the battery circuit containing the
bell E. A power failure causes the light to go out
and the bell to ring. The other bell F is common to
all six convertors and rings when the automatic
temperature and pressure controls on any of the units
fail to function properly. The latter alarm is also
connected to a microswitch G, which is actuated when¬
ever the gasholder for the synthesis gas rises too
high.
The thermocouple system is shown in Figure
IX (b). The cold junction H, the constantan header
K, and the potentiometer L are common to all units.
Individual copper leads are used for each thermo¬
couple. These are connected to the main circuit
through a pair of single gang eleven position rotary
switches M and H.
- - ■ -. -.. :
*
' ' ■ ' , . . . , ^ ^ • ,
" ' ■■ ■' " • • • . • t ' .
' ’ • l; '• .
.
' - ■ ■ v. : ■ : ;■ s .,.- rr ..
■ ' '* ; ' ' ' ' ■ , ..... . ... ;
-
.C '
■ cv :..e ....
■
'
: f ' 0 • • ,± ■
• •> •••;!. f-r , , ,
* ~ ° ■■■ ;;i '.) f,
' ' ; - ■ ■ .. i ?• • ...
' v :; ■ : ; •
. j. ,
■ , 00' • f ■
■ o ■
.
: ■ ■ ■.
x70'/r; Xi? -ja© :
*
> ;
< • o ■ i, . ■
*
U)
Figure H - Main panel board (a) Electrical supply
and alarm system. (b) Thermocouple system.
-51-
One rotary switch connects the thermocouples giving
the catalyst bed temperatures; the other, the jacket
temperatures. Each set of thermocouples (R and S)
is connected as shown.
Each individual unit is designed to maintain
certain predetermined operating conditions continuously
for long periods of time. There are three main opera¬
tional controls:
1) Automatic control of synthesis gas flow. This
flow must be steady and easily measurable.
2) Temperature control - the maximum range of
deviation in catalyst temperature is 15 centigrade
degrees.
i
3) Pressure control - a constant pressure of
100 psi. must be maintained in the reaction system
and at the same time the residue gases must be bled
off continuously.
These controls will be discussed in turn.
1. Flow Control
The chief difficulties of automatic flow
control arose as a result of the extremely small
flows encountered. A gas space velocity of 150/hour
(normally used in the Fischer-Tropsch synthesis)
corresponds to an absolute flov/ of only about 0.4
cc./sec. for the volume (70 cc.*s) of catalyst space
i).;
' ■ ' : • ' ' .■) •. ■ ' '
■
* ; t
■
■ ' a ■ ino;t$£ i qq . is*j . ■ . :
.
. ' ■ ' . ;
* .■■■•■ . .. ; : : .
•• • oq: ■
• ■
;\C: r ;, ; ' t ; j o • •
. ■ .
■ ■ ■ CfG.iJj ■ ■ . ■ t ■ U . : rid }
0
■ od...ia• o - ■ LX .: , a• ; a o , ■
' " ■ ol '... jo : r .... .. : ; t . \ •- P -
n. : . LOv.df! . 0
. ,
• * . ■ . 1 • . . v) ’4 f 0 *. ■ -
- ;;if . : a
,
1 < ., ....
ad:..
•-*
'i.o
' 0 ;at.:.
00 X
• . l-
■
.m ...
... 0 El
•0
• o
».;• >
. ' .
/
\ - -
! . 0 0
OV
-52-
used, when measured at the operating pressure of
100 psi. gauge. The literature reveals no previous
attempts to use automatic flow control on such low
rates.
For automatic control, there must be a
primary measuring instrument which will respond
quickly to changes in the rate of flow. This
response must then operate a control instrument to
correct the flow. In this case, the measuring
instrument is a capillary type pressure flowmeter.
A diagrammatic drawing of the latter, together with
a description of its design and calibration will be
shown in the Appendix. It is sufficient at this
point to state that the rate of gas flow is indicated
by a pressure drop across the capillary, this being
measured by a mercury manometer.
The instrumentation and wiring is shown in
Figure X (a). The two mercury arms of the Merriam
3-contact manometer A are connected to an electrical
circuit operating a small electric motor B. The
latter is a 1/50 H.P. reduction gear motor, induction
type, delivering 29 r.p.m. on the output shaft. The
driving mechanism of the valve stem is linked to the
motor shaft by a 24 to 1 worm and worm wheel, the
valve drive rotating at 1.2 r.p.m.
£ ,
■ '• ■ " ): .od '.v; ,:dr : : jj ?;,<
*
f-'uo • ,. -
i-.
• . :■
.1 -y/.';x>y, a od\^-rof:o :.k;r
)
■’ / '' • •••' : :t ' ! x S.. u'i. • } . L' *J/
J)
; - \, '. .V
; ' o J; t ae£>- : ■
,
'
• ^ ■ ■ ■ "■ .. ■ . . ■ . r#
- p :
: : > • • . i'll • (t$ - ' . '
• •' ' '
■ " 1 •’ 0 '■ ■ ' : ■ y. ' J; *.
. .
* * * . ‘ ■ v .
• • ' • ' •••■ ' r i :')0' ; .'."/I:*/:".'
O ' ' v> . . ...■ . L ■
o . ! ;.■! ;■! ,t‘; ' •. • j '
54
- 55 -
The motor, the control valve, and the flow¬
meter are all mounted on an angle iron, which is in
the form of an inverted L. Their spatial relation¬
ships are indicated in a photograph - Figure XII.
The valve and mechanism are illustrated in
Figure XIII. The brass valve drive shaft B is mounted
in a steel bearing block C. Female threads (24 to the
inch) at its lower end connects the drive shaft to a
brass thrust shaft D. The latter, with one of its
hexagonal sides bearing against the brass bar S, is
prevented from rotating. For every revolution of the
drive shaft, there is imparted to D a vertical move¬
ment of approximately 1/24 of an inch. This is in
turn transmitted to the steel valve stem F. The
needle of the latter has a very small taper of only
0.034 (inches /inch), thus permitting fine control.
There is, however, a definite lag (fixed by means of
the threaded nut on top of the valve stem) between
the action of the motor and the response of the valve
stem. This lag improves the control characteristics
of the valve by matching the lag in the flowmeter-
manometer system. A gland G and the packing H keep
the valve pressure-tight. The valve body K is
machined from a lj inch hexagonal brass bar. The
synthesis gas enters at L and leaves at M, making the
valve self-cleaning.
• * ' ' ■ ' " J n r, 0::.i c r.ii 6‘X8’ *'U; J’Ofi'i
' ■ . .
*
: ‘J :■
’ -
• j. .
0 .. .
V
• . f
' • ■' :
*
■
-
O <: r'J
•
• : ,oo "'O.,
.
: > \.o
f.
- »
•3 ' > v o.o . ■
6 til
.
J. a;'
<
n
■
; o •' •:
' 1 o .■
■
’ • .
.
. .,o
J,' V . 1 •
►
. .
. . -.1 o
.
•
.' / ‘ o.
J ' Oo 0.0
■ 0 0
.i
;ol: O
’
\p,
•
, < .
’
- ■
,
•
' i-r.i :o.
56
-57-
- 58 -
There is provision made to prevent the
valve from binding in either the extreme open or
closed positions. The drive shaft B operates a brass
cam N. This cam is machined from a !-§- inch hexagonal
brass bar and has internal threads, four to the inch,
fitting similar threads on the drive shaft B. Rotary
motion'is prevented by the brass bar E. As the drive
shaft turns, the cam moves up or down. At either
extreme of its travel, the cam actuates the microswitch
P, breaking the motor circuit and ringing the warning
bell.
The motor operating circuits are fed through
(F vc^.Xol')
an operating switch C -"a 3 gang, 6 pole, rotary type.
Only three poles are used, each alternate pole being
open. In two of these three switch positions, the
motor is operated manually to either open or close the
valve. In the third position, the relays D and E and
the manometer A are brought into the circuit and the
motor is operated by automatic control. These revers¬
ing relays are connected such that the circuit through
one actuates the field of the other. Thus when one is
open, the other is necessarily closed and short circuits
are prevented. A fuse further protects the motor and
circuit.
Considerable difficulties have arisen with
the control value, due largely to the smallness of the
'
.
.... \. . 1 . ‘
• ‘.. i i.Vi
* : ; . ;/ •
arts'
■j ... Y ... i . 1 at. •' r...V .
• • ' ■ * ' V' \.J .
r r
■ . . ■ » „ *
. ■ Y ’ , ..,
. ‘ . ....
• ‘ • ■ y ; ; !... r
- ■ ’ . ir - 0 < ..
a
r. ,
.
O' . : . . . o !
,
1 0
.
’ ■■■ Y . . .
■ =. • V ■ .
■ iv Y.
.a
L . .
...' •' .- :.
- 59 -
flows involved. The slight coasting of the motor, the
lag in response of the mercury manometer, the inertia
of the mercury are all important factors when dealing
with such low rates of flow and were overcome in
part by test and experience.
Upon actual trial runs, it was found that
the maximum allowable pressure drop across the control
valve were only about one pound per square inch. If
this value we^fc exceeded, the mercury in the manometer
started to hunt i.e. it continued to oscillate over a
range of several inches. With a high pressure drop
across the valve, the valve stem operated in a nearly
closed position, resulting in a proportionately large
percentage of throttling per increment of vertical
motion. If the pressure drop was less than one pound
per square inch, however, the control was excellent.
The valve stem quickly found, and remained at, the
equilibrium position corresponding to the existing
flow and pressure.
It became clear that satisfactory automatic
flow control required that the pressure variation in
the supply reservoir be not more than a pound per
square inch. This was accomplished by means of pressure
actuated, position relays located in the relay circuit
of a magnetically operated compressor starting switch.
The movement of the Bourdon tube (in an adapted mercoid
if
'V.
: , ..C
: - ” . .. . - 0'-. i,..; aO
. ... 7 .... J. 0 . . /
r: r O\ • •'/
0 -- ■■>Qy
. . . . \ ' ■ ; X • Oj - .
. ■ rp 3 '. - ■' rocL ■
• : ' . ! ■ ' '
. .
V. ■
.... ... \ - J . V . .. ' . .'-j- c>
.... • . ■. ■
.
.
. ’ • bo fill 0‘ . , axro , . ia avl "
.
.
'
,
; 1 ■' ■' -iJ r j . .' ' ...»
. • '
- 60 -
pressure control gauge attached to the reservoir) is
magnified and translated into a lateral motion of a
metal contact strip. The latter moves between two
f
electrical contacts which, in turn, are connected
through two relays to the magnetic switch relay. As
the pressure in the reservoir drops, one contact is
made starting the compressor. As the pressure is
raised, the other contact breaks the circuit, stopping
the compressor. The supply pressure was maintained
constant within half a pound per square inch. The
service required of the compressor motor is a difficult
one. A future design would involve improvements in the
flow control valve and continuous compressor operation
with control by an unloading device.
2. Pressure Control
In normal operation, the pressure in the
system must be kept at 100 psi. gauge. This is done
using a Cash type 1935 back pressure regulator. This
can be adjusted to maintain the desired pressure dif¬
ferential between the system and the discharge - in
this case 100 psi. Other pressures can be used but
only if the same for all converters. Improvement in
the flow control value to the extent necessary to
permit wide variations in pressure between converters
does not seem possible. Back pressure regulators
■ '■ ; o
•
M ::
• • '• . < ■'
A' . ..
-3;, '. -r
• - •
V . , .
•
.
s'©* ad • i.* aae
' . . ■ •: . '
' O
* ' " : ' i
, .
. .
,
' ■
3 . ’ • " - a
:
^ (
• ■ l . t . o
3 c.r f ■ • • ; ] • ,
-
• ‘ ■ oil oiii
. . ’i c' /.. .
♦
- 61 -
between the synthesis gas header and the individual
units would be the solution.
3. Temperature Control
Temperature control is obtained by the
method devised by Downs. The temperature of the
catalyst tube is kept constant by an outer jacket of
boiling liquid. Three liquids have been used - dow-
therm when operating at 250°C, tetralin at 200°C and
for activations, a paraffin fraction having an initial
boiling point of 360°C. The electrical heating
circuit is shown in Figure X (b). The outer jacket
is heated by a resistance coil K, (total resistance
of 20 ohms) which is connected in series with a
rheostat* L. The heat input to the jacket is manually
controlled to be slightly above that required to keep
the fluid boiling, the vapors being condensed in a
water-jacketed condenser. Connected to the latter is
a ballast chamber whose purpose is to eliminate surges
due to changes in the vapor position. The volume of
the ballast is about ten times that of the condenser.
The temperature of the boiling liquid is
controlled by varying the pressure in the condenser
system. An inert atmosphere of methane is maintained
above the liquid. Pressures greater than atmospheric
are obtained from a high pressure methane storage;
,
,
.
, ’ .
... . : Jiii.
.
'
.
O ■ . j . 0
o .■ ■■ 3 . i'- 4 % :-r
•I
.
I 1 i I -(id- f • ... ‘v ;j ■
*
- 62 -
vacua from a water pump. It is possible to vary the
pressure in the system from 26 inches of mercury
vacuum to 40 psi. gauge. For dov/therm this corres¬
ponds to a temperature range of 180°C to 330°C, and
for tetralin 130° - 280°C.
There are two thermocouples in each con¬
verter - one in the boiling liquid, the other in the
catalyst bed. The latter may be adjusted to give
temperature readings throughout the whole length of
the catalyst bed.
An alarm mechanism is incorporated into the
electrical heating circuit to prevent the liquid from
boiling out of the jacket. This may arise from one of
two causes: the stoppage of the water flow or the
supply of too much heat to the boiling liquid. A
reversible thermal switch M in the top of the con¬
denser is set to break contact in the heating circuit
at 70°C. If any hot vapors reach the top, the
thermoswitch opens breaking the circuit. The light
N goes off and the bell rings. Thus in the event of
water failure, the action of the thermal switch would
prevent any undue temperature rise and, by intermittent
action, maintain approximate temperature control.
There is no warning bell if the heating coil
burns out but there is an increased brightness in the
signal light.
• • c ; , n ■-... .
.. ;m .ji: ,0'i L :. .• . ■ . •. .. ... ; 2
,
. . •:
■ . i •: • ■. • o. . ■: •. ■
. : : '; ■ .. ■■ ’ . ■ ! ■. v
'
, , . .
. " ■
, . . o i ;•
.
,
... .3
■ lo oixwe 1 • i . •.
' l > i .... ... ' f .! ", .... ..
. • .
. .
.
■ ■ ' J , ••
V :
' •' ■ . . • ' ' •
■
.
IV CATALYSTS AND THEIR PREPARATION
In these investigations five catalysts in
all were tested. Except for two of these, very little
is known of their preparation. Two catalysts, a
Co-Th and a Ni-Mn, were prepared by the Harshaw Chemical
Co. of Cleveland, Ohio, and a third was an ammonia iron
base catalyst supplied by the Alberta Nitrogen Products
Ltd. of Calgary, Alberta. The other two catalysts,
one of pure iron and the other iron-copper, were pre¬
pared by an under-graduate student (29) as part of
another project, following the procedures laid down
by the U.S. Bureau of Mines.
The Bureau of Mines No. 10 Iron catalyst was
prepared according to methods developed by the U.S.
Bureau of Mines. Two solutions were made up separately
at 70°C. Solution I consisted of 1012 grams of
C.P. FeCNO^Jg* SHgO dissolved in 3 litres of distilled
water; solution II was 571 grams of C.P.K 2 CO 3 dissolved
in a litre of distilled water. With both solutions at
70°C, solution II was added slowly to I over a period
.. . v:
■, ;• ■: t r .
' ' * ** < . 4- -
J u 0
.1
-•
■ o
.
. .• .
A • •-
0 • > .i
.
■ ir u;i
■ • : ;V
,
. -1
.
"f
n-;
v ... r. i
-
..
U
.
' r. .■
J-'T
J
...
0 O'.
■
•J !.'
-u
.
i
...
>V. ' l
r
0 .'(0
/
:)
.r’.y
-
, . n;
jx
; : "
.J .. ■ •
c”.
'. ’■ '• •
■
:'■£ 1
. ■
. i.
.
.
,
Oft
.. o/ryv
1
■ ■■
i; . (fif
.
\
i.
. ..
1 /
•• ' lO .
■ '• .. ■
' 0 I
«
■ .
9 -
*
*♦
1
■
'
) $
I :
.
< r
yo oi ;Xv."j
C -; a
•
- 64 -
of 20 minutes. The temperature of the mixture was
to be maintained at 70°C throughout the preparation.
The slurry was made up to 6 litres, the supernatant
liquid decanted. The slurry was again made up to 6
litres, the supernatant liquid again decanted. This
was to be repeated until the nitrate concentration was
less than 1 part in 16,000 as indicated by a
diphenylamine test. Nitrate in larger concentrations
turns a diphenylamine solution a distinct blue. Accord¬
ing to instructions sixteen decantations should have
been sufficient but, in practice, even after thirty
decantations, the wash water gave a weak nitrate test.
At this point washing was considered complete.
The precipitation and washing was carried
out over a period of a week, during which time the
temperature of the slurry may have varied over a range
of 10 - 15 degrees. It is thought that this long
period of digestion may have had a detrimental effect
upon the catalyst.
The slurry was filtered in a large Buckner
filter and washed several times on the filter paper.
The cake was placed upon a porous plate and allowed to
dry slov/ly at room temperature for five days, open to
the atmosphere. It was then placed in an electric
oven. The temperature was raised slowly to 150OC over
a period of five hours and held for an additional 24
.
' • -i. ' j°OV d:<i tiznl&uctlu$i od c;i
. '•id'll 3 o .
-o.
.. ’• rv: .. a .
■ ••• '
, • 0 cyJ ■■‘As. ..
- - • •. ;
• r ' a 1
. ...C
. .. , ■ 'i iZ . ..... .'.
fci . , I
- • a:
• - '1
.
' . .. t sd-f oe„b
V v • .J .i .. . . * . ' L * 1 ... • ‘ U -
■ d ■■■ r‘:'
. , .*■
, . j '■ . ;.; : ./ " . if {8)
so . ■ ■ . •
.
. ' ■
■ ' ; ■ ' ... . !> : J " ' ;
, •
- . ‘ ;... . 8 .1
• . \. . ..■•./ .. ■. i '
. " r • ... ■■ .0 ,
■ . • 's '■ \ \ . .o. •. ;i'_ r.-Q_r .
.
' - - ‘ . ,
- 65 -
hours, The dried catalyst was ground to pass through
a 50 mesh sieve. A yield of 193 grams was obtained.
Nine grams of small flake graphite was added as a
lubricant and the catalyst was pelletted into small
tablets 7/32 inches in diameter and 3/32 inches thick.
The graphite has no catalytic action; it is inert.
The second catalyst. Bureau of Mines No. 24
Fe-Cu, was prepared by the co-precipitation of Fe(0H)g
and CuCO^ from a sulphate solution with KgCO^. It
was prepared in a manner almost Identical with that of
the iron catalyst above. Similar solutions were used,
the only difference being that solution I with the
same iron concentration had CuSO^ added to give an
Fe-Gu ratio of 4:1. Precipitation and washing was
the same except that the slurry was blown 10 - 15
minutes with air before each decantation. Some of the
iron is oxidized to the ferric state. Sixteen decan¬
tations were performed. The catalyst was mixed with
graphite and pelletted.
Nothing is known of the two Harshaw catalysts,
A5 and A124, except that cobalt and nickel are the
bases with thoria and manganese respectively as pro¬
moters. The pellets were In the form of cylinders,
3/32” diameter and 1/8 1 ’ long.
The ammonia catalyst Is prepared by fusing
iron oxide, prepared from "drum scrap" steel sheet.
, •
... . 2
. : ' ■ ' ' . :1 . ".,0 i ; ;e.I
il ' r . • '4-. • ... . . .
X : . X ■ ' .u. ..... X ; ...
. ..: o r, ' .: :> . .
: . •. ,,./ X • ■
■ . t
*
.
.
.
. . . / c- X
,
»'.ev >. • 0 'hi soft oj
: . . ~ ... .
.
- OX nwol
>
u:
. :
■ •
*
. • '. ■ ; ■. . ...■ -
.
■ . • . X :: ; : .
, ■
’ ' . . .
- 66 -
with suitable promoters, crushing and screening the
resultant mass to J” - 3/8” mesh. The exact procedure
is not known. The following analysis of this
ammonia catalyst was supplied by the maker:
Si0 2 - 1.4/
MgO - trace to
0 .2/
K 2 0 - 1.7/
CuO - trace to
0 .2/
A1 2 0 3 -1.7/
Fe - 67.7/
For present purposes, this catalyst was crushed again
and that portion retained between 6 and 10 mesh sieves
tested as a Fischer-Tropsch catalyst.
-
,
.
■ . - 'j V ■
. .
-
' " - ; • y '
. * .
-
- -
V
-*
, •• . . ■
V ANALYTICAL METHODS AND OPERATIONAL TECHNIQUES
The procedure followed in testing the
catalysts was based primarily on the U.S. Bureau of
Mines technique. These methods were developed
empirically and slight deviations are permissible
provided that a careful record is kept at all times
of the experimental conditions. In general* the
methods of handling the iron type and the cobalt type
catalysts differed considerably.
After the units had been tested and the
leaks reduced to a negligible quantity* the catalyst
chambers were filled with catalyst pellets, forming
a column approximately 16 inches long and 0.6 inches
in diameter. In some cases porcelain chips were used
to support from below and cover the top of the
catalyst bed* primarily to insure uniformity of
temperature over the length of the catalyst column.
: . . . ..
V
• <•/- . • « •!- --
't,
- •' ... : ... 11 c": . , l '.vTi;
<
,
■■ =. .... - : ;. .... ..vo\:
' ' . ' . ...
.
-■ “0 0 .J . ... . :: ,. . . :
. o
.. ci hi '■ :: 3 J ao - ■c. i / ..... ....
* • .
■ ' ■ i - ) . ■ -v ■ , .
.. ■:
, (>
•- VO
-68-
Reduction and Conditioning of the Catalyst
The system was first flushed with hydrogen
and the rate of flow adjusted to give a space velocity
of 20 - 25/hour. (Space velocities are based on the
volume of the empty catalyst chamber, not on the
actual volume of the catalyst). In the first experi¬
ments, a slo?/ stream of methane passing through the
outer jacket constituted the heating medium. The
heat supply was controlled manually with a rheostat.
The temperature of the catalyst bed was raised quickly
to 100°G and then slowly over a period of 20 hours
to 360° - 380°C. During this time, the jacket temp¬
erature was kept 20 - 30 centigrade degrees above the
catalyst temperature, insuring the proper rate of
heating. Reduction was completed by holding the
catalyst temperature at 360° - 380°C for 4 hours.
There were two difficulties encountered in
the use of a methane heating medium. First, it was
difficult to raise the catalyst temperature at a
uniform rate and secondly, there was a considerable
temperature differential over the length of the
catalyst column, the top and bottom portions being
30 - 40 centigrade degrees cooler than the middle.
For these reasons, in the later experiments, a high
boiling point liquid was used in the outer jacket as
the heating medium. This liquid was the residue
. o': ' •
■
. ; r-
I: i ■: »
■ D
-
.
*
) . v
,
' ■ =' '
'
.
.
t
.. j .
.. .;.v..o ■; o\";,
’ ’ i . \ -
■ ' ■: : ' • -
■
' . . i :r o t
* ......
\ ..., '.v
V*
Oi..- -
% dt •
... ■ • •.
: ■■ '
.. _ .
■ . •• ■ o
. , . 0
0
.
. :• r O ' ' O'
)c'.: -
... J 0
0
..ooi: or
o- .
,. r , ■.
•
‘it
■ o •.. ■ : ■, . : ■:.
* . . ,* ' ; ; ;i
f,
;
o ■ ■
*
' *
. ;
.0 '
... :
' 0 - .
*
• o :
- 69 -
obtained after distilling off all the lighter fractions
(up to 375°C) from a gas oil or petrolatum.
The conditioning procedures for iron and for
cobalt catalysts were quite different. In the former
case, after reduction, the temperature was reduced to
220°C and synthesis gas at 100 psi. and space velocity
of 150/hr. was admitted, replacing the hydrogen. The
temperature was adjusted to give a contraction of 13$
(20$ in the case of a water gas having a C 0 :H 2 ratio
of 1:1)* Yftien it became steady at this value, the
temperature was raised slowly over a period of 48
hours to 250°C or to a temperature such that the con¬
traction was approximately 40$. The catalyst was now
completely conditioned and ready for operation. After
a cobalt catalyst had been reduced, it was allowed to
cool to 160°C and synthesis gas admitted at one
atmosphere and space velocity of 100/hour. The temp¬
erature was adjusted to give 20$ contraction and held
for 4 hours. The temperature was raised until 30$
contraction and again maintained for 4 hours. This
was repeated until a 60$ contraction was reached.
The most active cobalt catalysts operate at
-x $ contraction =
Vol.of Feed Qas - Vol.of COg - free Residue Gas ^ .
Vol. of Feed Gas
: •. ' • ■ IS J.‘ . I . , :S • , , ' " 0
- '■ :
-
. . . ■:.» , ' c
■
... €
J :> .
.';;
3 V...
«
: ' ,
•
• .....
.
0....'
o;
’
*
>. ; if ' ' ....
J
. j. ... ...
,
; .. o.i
,
■
0 . : ' ,
'.... • c‘ir
'
: '
*ieqo 'io to . d & .
• ' -
I : ■■' . . . . • ■ '
.
.
■ ■ ' ’ ... ... ; ■ .. . .
’ ' .... ■
lb
3 J a.C.rOu !• i
.. .0 .
.Or
.
,
0 ■ too
,
- 70 -
approximately 190°C. The synthesis gas pressure was
raised to 100 psi. and production started with an
initial contraction of about 70^.
Method of Operation
When the automatic controls were function¬
ing properly, the operation of the process (as
distinguished from the analytical work involved)
required comparatively little attention. All the
units were started on production Sunday midnight and
operated continuously until Saturday morning. During
week-ends, the temperature of the catalyst was lowered
15 centigrade degrees and a slight but positive flow
of hydrogen maintained over the catalyst. A normal
weekly run was approximately 125 hours. At the end
of the week, the liquid and solid hydrocarbon products
were removed from the pots, weighed, and stored for
analysis.
Iron catalysts were operated with a gas
space velocity of 150/hour for the first four weeks
and 100/hour thereafter. Cobalt catalysts utilized
a space velocity of 100/hour throughout their whole
life.
Good iron catalysts maintain their activity
for periods of months, while cobalt catalysts become
slowly deactivated (noted by a decrease in the per-
,
, ; ■ -
■ o ■ y • ' ■ i
■ I' ■ ■' ' ■ ■ . ■ . ' . V . /
,
'
*
I* .. i , • . i « ■;
*
.
O'. ........ ’ .
> - ■ ■' . , : ■ i ■ v W, ... .
. i I. t . ■
.
: ■ f J • ' . . .•
♦
■ 1 ■ J 0 , :0'lJ ,
, ' . - . . : OC
' .. ; .... .. ,
- 71 -
centage contraction). Low temperature reactivations
of the latter were carried out periodically, usually
once a week. The pressure was allowed to fall to
atmospheric and the synthesis gas replaced by hydrogen.
The temperature was raised 15 degrees above operating
temperature and hydrogen passed over the catalyst at
a space velocity of 200 - 300/hour for two hours.
The unit was put on production again by cooling to
operating temperature and replacing the hydrogen by
synthesis gas at 100 psi.
Iron catalysts are never reactivated.
Aside from the analyses, the work of the
operator normally consisted of the following:
(a) the preparation of the synthesis gas, (b) a
periodic checkup of the recovery system, especially
with regards to the operation of the scrubber and the
residue gas proportional sampler, (c) check on the
proper functioning of the automatic controls, and
(d) periodic readings of the experimental conditions.
In the project's initial stages, however, considerable
difficulties were encountered in setting up the
various operational controls (particularly in the
regulation of the gas flow) and this, together with
the maintenance of the equipment, required a large
part of the operator's time.
;cf. x4' 5®t>jBXq©‘ ■ ■ , |
■
. o::.v: :
. . ' .' '
‘\ ‘S rjc ; .. '■ v 00g -
. i ' cic o .
■ 1 .
m
• ■ ’
.
W? .. . • jjj,, » V,
OOqCS • L. ■
• ... ; . ■
■•'JO • . •}, . . . .r - o/r.5j:ae$l
-l ' ■ ' ' . O';... )i -o ' I (l
R - o :.. U. 1 :• •.;• , ;
• ■ - • -i v . , ! ' t o■ .>iiy ...
oo x ’-i i(-y o
■ ■ >© o ■ ro
I jppj/ 1 ' * K.
' ' ' '. ... .,
*
1
.
- 72 -
Flowmeter readings were recorded every hour
(and the contraction calculated) while temperature
and pressure readings were taken every three or four
hours. By means of these and other operating data,
it was possible to follow the progress of the catalyst
testing and to note any changes in the experimental
conditions. A sample operating sheet is shown on the
next page.
Analyses and Material Balances
Examination of the flowsheet indicates one
feed and three product streams. The synthesis and
residue gases flowed continuously while the light and
heavy liquid products were removed from the ice-
jacketed and steam-jacketed pots once a week. Overall
material balances, therefore, had to be based upon at
least a week’s operation.
The synthesis gas was analyzed once every
two or three days in a standard Burell Gas Analysis
Apparatus. The percentages of COg (by KOH), Hg and
CO (by oxidation over CuO) were determined directly,
Ng by difference. A careful check was kept on the
H 2 :C0 ratio which ideally is 2;1 but in actual practice
ran slightly higher, around 2.03.
The residue gas contained water vapor, COg,
hydrocarbons (CH 4 , CgHg, & 3 Hq), Ng, and unconverted
: • ' '■ •. ■: ' ; .Ci . ' ■\’C.
' , ■■■•■,
- ■
• : O : 0.. w j;
■
.
-
• o 1 :■ ' i:; o ■ " ; : . o ' / .
, . • •
■■11 ■'■'■■> J j' ‘
: ■ ■ ■ ' ' -• -■ -O'-
' ' . ' . . :jr|
&ml * ■ i . .
,
'
. .
I . ■ ' V. ) L)
Oi-- .
•>A
* ' : • ; • , . - Va , . ■ ‘
Exp*t No.
Catalyst
— 2.
tore Fe (_%■«* M ' *
9 !• fc ^ •
OPERATING SHEET
-FRZ-T--
Date:
1 /h'
i
Time
B AM
9
10
11
12
1_PM
2
3
4
5
6
7
8
9
10
11
12
1 M
2
3
4
5
i 6
7
tressures
Reactor D s ^-*
gauge
1004
|
lot
lot
IM
let A
|ro-t
Jacket ^ -
*3-3
3Vi'
33N
R'C
3 So
3M o
u-%
>
Barometer
Tot
I t 2.
lox
lox
16 x
a o-2_
TEMPERATURES
- . , n
1
_
Catalyst 1
141
att
141
14*1
I4t
r
>
: 041
Catalyst 2
t.4£
Uf
144
x(A
U4
14^
at*'
Catalyst 3
xsl
XiN
XsN
2 TI
lSA
-
Jacket °C
24£
l
245
^44
144
x44
—
N4jT
Gasmet er °C
0 %
49
lx
ir
13'
IT-
TO
FLOP
Gasmet er-cu. ft
-
/
T-
/
o' 7
'b 0
V
'v' 7
r
S 7
r
V
v
A
„o
V
vT
V
pV 2
<4 J
3' -
A-
x^ V
Or
t\-
A
*
A’
4’
,A
/
r>
/
*3
X <1 0
4°
✓c
piormeter L
Ml
S.’M
2 ..C0
Ml
i u
1 -6|
I-1C
” R
t-^sT
Mt
<
INI
l?4
If 3'
” Total
3 Si
3 c n
344
3.^
<1 01
3.7/
| % Decrease
'ii-i
51-0
34 4
34'0
2S%
34 b
34-4
341
iPl
34 o
| Samples
.
5cM3
!
’5«3|
:ceH
f
1
-
Reactor No. - S
NoMinal Pressure - t«=c pA.
Nominal Temperature - 2 . 4 , 5 'cl
.Nominal Plow Rate - s. Ht
Notes:
Notesi
Notes:
!
- 74 -
Hg and CO. The water was removed by a CaCl^ U-tube
drier, and determined by weighing. It was necessary
to weigh these driers only once a week since the
water content of the residue gases was comparatively
low. The C0 2 was removed by a caustic scrubber, and
determined by titration. The scrubber liquid con¬
sisted of 250 cc’s of a 3N solution of KOH. Only
half of the total strength of the KOH could be
effectively utilized since, beyond that point, some
COg passed unabsorbed through the scrubber. An
aliquot part (10 cc’s) of spent solution was titrated
against IK HC1 to both the phenolphthalein and methyl
orange end points. The difference represented the
molar equivalent of COg.
A proportional sampler collected a daily
2-litre sample of carbon dioxide-free residue gas.
This sample was analyzed in a Burell Apparatus for
co 2 , Hg and CO (over CuO), hydrocarbons (burning with
oxygen over a hot platinum wire) and Kg (by difference).
The hydrocarbon content was expressed in terms of
methane and ethane although it was realized that some
higher hydrocarbons may have been present.
The liquid products collected each week
were allowed to accumulate until in sufficient
quantities to be analyzed. In both of these fractions,
considerable water was present which was removed in a
■ ' ... ... ■ ■ ..v:
/ . • j O • : . ;
*
.
j
.
T .
■
iii GO -B
■
.. * ' .id:
.
.
•
• 1,0
r * •
OQ ';...
■
•
•
.. "
.
O'! ;
t,
... ...
*•: :
. : 1 ...
■
)
.
C' 1 . •' '■ '
•
..
• :
...
rtr i
. .r \'
. ' .
t , . .
... • >■..
J:. . .
*
'•••••' • ' • r O ' . 1
: ■ • C .. ■ 1 : .
■ ■ 1 ’ . . v
- 75 -
separatory funnel. The liquids were filtered to
remove solid impurities.
The water in each case was tested for the
following:
a) acidity - hy titration with a weak solution
of NaOH.
b) a distillation to detect the presence of
alcohols and other low boiling fractions,
c) density - accurately by density bottles.
The oil fractions collected were tested
for:
a) density
b) aniline point - to determine the degree of
unsaturation.
c) refractive index.
d) an A.S.T.M. distillation, modified because
of the small quantities involved.
The wax fraction was tested only for density.
•' . : ■' ; «. i . ....
: j .. .... . ■ .i. >. '. ■
.
,
. U- • ;i.:' v/ i
: 0:: J Ic , /-j:
. : • : ■. • . . 2 :
: ». ;■ . i: • ... • V , LiJ,
. •
7.0 ■ fri ■ :
. /
.
■ ‘ , ■ • . ‘ ) ■ . •. • • ■ < • ‘ ••
. .) T. • . . : J
V. - U .
.
, . , .
-
: ' ..0 ... . ./ i >1 '. ‘ •. . ) ...
- 76 -
VI EXPERIMENTAL RESULTS
Since the present investigation involves
the testing of specific catalysts, each must be
reported separately with respect to treatment and
results. Five catalysts were tested:
1) Pure iron
2) Iron-copper
3) Promoted iron (ammonia)
4) Cobalt-thoria
5) Nickel-manganese
The reduction, conditioning, behavior, and yields
will be discussed for each catalyst.
The feed or synthesis gas, common to all
units, was found to have the average composition:
CO 2 0,3% by volume
Eg 64.4^
CO 31.5 fo
Ng 3,8%
Hg/CO 2.04:1
.... . . J ' . ■ -iq '
v ■•.: . . "’0 i M
■. ■■. ■' ; : , ; .'J .0 Li;i; ■
‘
; H
• :
■
. ,
. . i. ■ • ,
* • -■ '
' o r ■ . . 3 • ■
-• '
,
• .
- 77 -
This gas was used in the testing of all catalysts and
the daily variation was on the average not more than
about 2/6.
During the coiirse of the experimental runs,
it was suspected that the feed gas may have contained
some catalyst poisons, either oxygen or sulphur.
Sensitive tests failed to detect the presence of
either. In testing for oxygen, the feed gas was pas¬
sed over white phosphorus freshly distilled and
deposited from the red allotrope. Ho luminescence *
was visible indicating the absence of oxygen in
quantities greater than one part in a million (20).
In testing for sulphur, it was assumed that, if present,
sulphur would occur as hydrogen sulphide. A measured
quantity of gas (several cubic feet) was passed through
a caustic scrubber, the solution acidified and titrated
with an 0.005 N-Ig solution using starch as an indica¬
tor. Close agreement of the titration of the sample
and a blank indicated a negligible amount, at the most,
of sulphur in the gas.
1. Iron Catalyst (Bureau of Min e s #10)
Source of catalyst - prepared at the University of
Alberta.
Weight of catalyst - 71,5 gms.
Volume of chamber occupied by catalyst - 62.2 cc.
■ -V: uOi;^ Vi-V/ tfCil
.
, <r
;■ : .. ■ ...
,
a oat. - . ' ■ Qi ■ ' k oXi ■ i
A ..
.
.
.
2>dr: p. &£ . . ' v : ' .
. ■ .
,
'
.
• .y. $
«
, • • •
, ■
,
. ' ■ ....
.
, . . -
: •• , : ai jyooo 'ioan&'.o lo ")rr,/IoV
-78-
Reduction ; The catalyst temperature was brought up
slowly over a period of 20 hours to 360 - 380°C in
an atmosphere of hydrogen. This temperature range
was maintained for 4 hours. Heating was by transfer
across a slow stream of methane passing through the
outer jacket. Although the middle three quarters of
the catalyst column was at 380°C the top and bottom
portions were as much as 50 degrees lower.
Conditioning : On reducing the catalyst temperature
to 225°C and admitting synthesis gas- at 100 psi. gauge,
no apparent contraction occurred. The temperature
was raised in 5 degree increments over a period of
48 hours to 250°C. The conditioning then proceeded
as follows:
Temp.
Percentage
Total hrs.
oc
Contraction
operation
250
13 $
78
254
18 $
86
260
20%
92
268
Initially 30$ increasing
to 38$
Operation:
In order to maintain the necessary con-
traction.
operation was carried on at
268°C, even
though thi
s temperature is higher than
that recommended
by the Bureau of Mines. Table III shows the operating
.
.
... . ... v :
' • ' J •' - ' '
. ' . v
, : ■>. . ;i - ,1 •. ■' r . : ,-IC
-
, r ... . .i
: ■ :. .
.
■, .. !Li :
Uq pOX
a . . . . ■ i b .. °< ;
' 0 ;• H a ■
r • • • ■’ t ■. •.»
' ' . '
*
'
; i :; '
.
. ; - . . ... ..
..'. '■!. , .0 ■
.
. • • . : • - .. • a! . :
■ n y :o
: • ■;
- 79 -
conditions and contractions for each successive
week of testing.
Table III
Weekly Operation with
the pure Iron
Catalyst
Week
No.
Temp.
°C
Space
Velocity
hrs.“1
Contraction
. .
Flow
Control
1
Activation
2
269
150
39
Good
3
269
150
38
Good
4
265
150
35
Good
5
265
150
35-40
Unsteady
6
253
150
25
Good
7
265
150
35
Good
8
266
100
55
Good
9
266
100
52 .
Fair
Reactivation and Operation : Although the catalyst
was giving the proper contractions, it was evident
that this was due mainly to the methane formation and
the water gas shift reaction. At the end of nine
weeks, the catalyst was given a high temperature
reactivation. Hydrogen at one atmosphere and space
velocity of 20 - 25/hour was passed through the
: ' •. • ' :j ' . ,o.i . ■ ; ){•
•
‘ ^ v. . .
,
.
r . ■/"' 0
.- ■
... ..... >
.
f
. 1 • i
■
....
0 ••);
■
:x •.)
V J 1
-
.“.co
' S .
03 X
■ .
.
-
•■.000 ;J
. ■ •
| .
u€ "
■ .
■ .
,
r n t . ■:
•' j y ;. . j
■■ • i'' r : , : •< ...
. ,
' ■; ■;t ; : . •- to . . - ' ", Oi... . '/
- 80 -
catalyst bed. The temperature of the latter was
raised quickly to 360°C and held for 4 hours.
After reactivation, synthesis gas at 100
psi. was admitted and the unit operated at 250°C for
two more weeks. Although contractions were high,
there was no increase in liquid hydrocarbon yields
and it was decided that no useful purpose would be
served in further continuing the test.
Yields The yields of the various products are shown
in Table IV.
Table IV
Weekly Yield of Products from the pure Iron Catalyst
gms./cubic metre of gas at N.T.F.
Week Liquid Fraction Off - gas
No. Heavy Light CH^ GgH^ GOg **2®
1 Activation
2
03
•
O
7.7
41
62
w
■w
3.9
3
0.8
9.8
40
43
28.6
4
1.4
9.4
29
48
6.8
5
1.7
9.1
26
48
18.8
6
5.1
5.4
21
11
108
14.8
7
1.4
8.0
31
44
180
22.2
8
4.9
38
32
153
26.9
9
1.2
35
34
145
14.7
Reactivation
10
0.1
6.5
55
22
222
38.0
11
3.6
5.9
46
24
154
25.4
•* Figures are unavailable for the first few weeks since
much CO 2 was passing through the scrubber unabsorbed.
•+
- 81 -
Analyses : Pour distinct products from the synthesis
were analyzed namely, the residue gas, the light
liquid fraction, the heavy liquid fraction, and the
water.
a) The residue gas had the following average
composition (COg unadsorbed):
Component
co 2
%
CO
ch 4
°sP«
Volume Per cent
1S.1
55.3
5.4
8.8
5.6
4.8
There were no significant trends in the gas composi¬
tion during the course of the catalyst testing except
that, after the high temperature reactivation, the
gaseous hydrocarbons and carbon dioxide content
increased slightly.
b) The light liquid fraction analyzed as follows
1) Density at 25°C - 0.7164 gms/cc.
2) Aniline point - 66.9°C.
3) Refractive index - 1.4112
4) Distillation - see Figure XIV
5) Odour indicates possible oxygenated com¬
pounds .
6) Average molecular weight - 130.
Si.j
. . .
,
‘0- M ;; . •. if') .
, .
. dX : .'o a- .f ‘ a
. ; ■ a ' .X;'
Ci.
-. ad "ir-i *....a.
V ' ■ • ■ ■ • " ; -a;
a a X . '... . 1 M ;
■
.
. , . . • ' ' .■
. ..
.. ..
, ••
» - ■
' - l :.I:
ad t> l t *
■
.
. • . 4 J.\r. ( •;
r i....
a : • • o
- 82 -
c) The heavy liquid fraction showed the follow¬
ing properties:
1) Density at 25°C - 0.7734.
2) Wo solid products precipitated in this
fraction at room temperature.
d) The water had the following characteristics:
1) Density at 25°C - 0.9968 indicating
practically pure water.
2) Traces of acids: Acid normality - 0.004.
3) Traces of a low boiling point liquid on
distillation.
Discussion of Results
1) Even at the beginning of the testing, there
were positive indications of the poor behaviour of
this iron catalyst. These may be listed as follows:
a) Formation of appreciable quantities of
water - not usually expected from active iron catalysts.
b) Low yields of liquid hydrocarbons - about
a sixth of the yields obtained by the U.S. Bureau of
Mines.
c) Comparatively high yields of gaseous
hydrocarbons indicating the methane reaction.
d) High COg production, indicating the water
gas shift reaction.
e) In order to obtain the contraction
normally expected from this catalyst, operating
- 83 -
temperatures 15° - 20° higher than those reported as
optimum in the literature were required*
There is a possibility that the inactivity
of the catalyst may have been due to poisoning either
by sulphur or oxygen. However, delicate tests have
failed to reveal the presence of either in the
synthesis gas. In the case of sulphur poisoning, un¬
less there is a comparatively high concentration of
hydrogen sulphide in the feed gas (which is unlikely),
the poisoning should take some length of time. In
the initial stages, the literature reports that the
liquid hydrocarbon yields actually increase. However,
this catalyst was inactive from the very beginning.
Initial oxygen poisoning is unlikely but a possibility.
Some parts of the catalyst may have been
improperly reduced. However, thi3 is insufficient
to account wholly for its poor behavior.
The third possibility is that the catalyst
was originally inactive due to contamination in its
preparation. Further, the long time of digestion
(several days) may have affected the physical condi¬
tion of the catalyst, perhaps in the state of sub¬
division .
2) Until the eighth week the total liquid yields
remain substantially constant, at which time they
dropped sharply with a corresponding increase in
-
• v . : - vv'i'.iiJ.'cL'iiia.Oq .e rl c
' •• / : O 0 < ' O
.
- 1 .. :'sV‘ £dc7 s*. o '
.
■
■ d-' ) l : ,1m. ' ■ r.J: >■;.[ . : m-y ,
■ ' ■' 1 I ■ ■' . i
'
- .j ed . ' .. g '
■■ ' •- - : - . ■;j . b ....
■ #:
Vttd . : :
,,
. * \ 'O: :• ;i o
■. o n.
.
.
. „ b, ■, .m
' ■ ■ ■ ' X : .
-84-
methane production. The few original active centres
were apparently deactivated either by poisoning,
carbiding or carbonyl formation.
3) High temperature reactivation failed to
restore the activity of the iron catalyst. This is
in agreement with the literature. The subsequent
high contractions are due almost entirely to methane
formation and the water gas shift reaction.
4) The liquid hydrocarbon products were entirely
oil. Ho wax was precipitated.
5) Correlation of the aniline point and the
average molecular weight indicates that the light
liquid product contained some unsaturated and pos¬
sibly some cyclic compounds. The refractive index
tends to substantiate this.
2• Iron-Copper Catalyst (Bureau of Mines #2 4)
Source of catalyst - prepared at the
University of Alberta.
Weight of catalyst - 76.3 gms.
Volume of chamber occupied by catalyst - 58 cc.
Reduction : This catalyst was reduced in the same way
as the iron catalyst described previously, i.e. at
360 - 380°C for 4 hours in an atmosphere of hydrogen.
The same difficulties were encountered.
.
■
l
.
. ■ \..v :■ • • ■ ■ ■ ,0 ; ! .V " , .
*
-
:ig ■ ■/:. ■ >' J;
* ? ■ ■' ;• f'
' . -q ■ ■: U-'
■
*
• ; ' • :. ■ ;u . ,oi\t * : • • ,
• ;
■ P :
. "■
. ; .
' •' Jv, r •.... ' . •/
■ ~ ■■ -o =■ rru
■ - ' . . . ■ ■ ; .
.
*1
0 ", , !:■ :: . ;
* *
.: ■' ,: • '
>.
. ...
i. *; :': r -anf
■ ' o ■ ..' ’..o'.
-86-
Conditioning : Synthesis gas at 100 psi. was admitted
at (catalyst temperature of 223°C. The progress of
the conditioning is recorded below:
Temp.
°C
Percentage
Contraction
Total hrs.
operation
223
0
230
11
16
237
16
34
245
20
58
250
25
62
Despite the low contraction, it was decided
to operate in the region of 250°C, since the litera¬
ture reveals that at slightly higher temperatures
carbiding and carbonyl formation quickly deactivate
the catalyst.
Operation : Operation proceeded at a pressure of 100
psi. gauge and a gas space velocity of 150/hr. The
results are shown in tabular form below:
,
v. ; o-.
. . . u. ... ... o; ..
' %Oi
. 0
-
.
•V2
■■ r ; ■ :) . ■ :. C ■. l,... , r :
-
? . . ■; .. -it qo o '
■ ■ ■ ■ as ' ■ 1 | ■ ■ jo*
o ■ ■ ■ i .. id* o
»
■ ‘ -0 L : , 50* ' B noj , , • 0
. .
. .
J. J .; .
■
-86-
Table V
Weekly Operation with the Fe-Cu Catalyst
Week
No.
Temp.
°C
Contraction
<sf
/0
Plow
Control
1
Activation
2
250
23-25
Fair
3
250
22-25
Fair
4
255
29-31
Fair
5
255
25-35
Unsteady
6
255
30-40
Unsteady
The contraction is apparently temperature
sensitive. The unsteadiness of flow during the last
two weeks of operation obscured the contraction but
indications point to increasing contraction.
Reactivation and Operation : At the end of the sixth
week, in an endeavour to improve its activity, the
catalyst was given a high temperature reactivation at
360°C for 4 hours. Although the resulting contraction
was higher (58 - 60^), there was no improvement in the
liquid hydrocarbon yield. It was concluded that the
catalyst had been permanently deactivated. Testing
was discontinued.
Yields: The yields per cubic meter of synthesis gas
of the variour fractions for each successive week are
shown in Table VI
. >/v,.
■
*
r/j
.
t:< t«.i vS: - .
i
0 ;Si
•
'll '
: . . .
.j'
. . • ' i
.... '• .; .; .
' . ' . 1
> *
^3*
G
0 : ■ -
*
' ~ . ; • ■ • ' ’’.I..’ , . ,\7 ;{■
* . . - *
\ • ‘ £ tli . ^ . ;...
, jc
• - .' ;
. C-°0Q5
' ■ , - ' ■■
. ■ • ■ . . ■
■ ; . - • .j:, 0
' .. , .. -
' ■ 7 - ; iU o \ ; \o
: . /< ,
-87-
Table VI
Weekly Yields of Products from Pe-Cu Catalyst
Sms./cubic metre of gas at N.T.P .
Week
No.
Liquid
Product
Off-gas
H 2 0
Heavy
Light
ch 4
°£= 6
co 2
1
Activation
2
6.6
5.8
8.0
11.0
53.7
25.3
3
5.6
5.3
9.3
9.7
62.6
24.7
4
6.2
7.7
26.2
20.1
86.2
23.1
5
5.1
2.9
22.4
32.8
89.5
6
3.7
32.5
23.1
154.5
3.7
High
Temperature Reactivation
7
5.5
19.6
40.8
197
28.6
8
8.9
17.4
37.6
149
19.0
Analysis of Products : (a) Table VII shows the change
in composition of the residue gas during the catalyst
testing.
-88-
Table VII
Composition of Off-Gas from Fe-Cu Catalyst
- Volume % -
Week
No.
co 2
CH 4
C 2 H 6
CO
H 2
H 2
1
Activation
2
4.4
1.8
1.3
24.3
63.0
5.2
3
5.4
2.2
1.2
24.8
63.1
3.3
4
8.7
6.3
2.6
18.4
62.1
1.9
5
11.9
7.6
5.8
14.0
57.9
2.8
6
20.3
9.4
3.5
13.3
51.3
2.2
High Temperature Reactivation
7
29.7
5.8
6.6
4.8
50.0
3.1
8
22.8
6.1
6.9
4.4
57.6
3.2
(b) The light liquid
fraction analysed
as follows:
1) Density at 25
|°C - 0.7537 gms./cc.
2) Aniline point - 63.2°C.
3) Refractive index - 1.4189.
4) Distillation - see Figure XIV.
5) Odour - suggestive of oxygenated organic
compounds•
6) Average molecular weight - 130.
•; -•' o.
Vi:: O’
.
-
.
0 ...
-89-
(c) The heavy liquid fraction exhibited the following
properties:
1) Density at 25°C - 0.7856 gms./cc.
2) White solid at room temperatures.
(d) The water showed the following characteristics:
1) Density at 25°C - 0.9840 gms./cc.
(Pure water - 0.9940 gms./cc.)
2) Traces of acid: Acid normality - 0.004.
3) Traces of a low boiling point liquid on
distillation.
Discussion of Results :
1) The overall behavior of the Pe-Gu catalyst was
very similar to that of the pure iron with respect
to water formation, low yields of liquid hydrocarbons,
high yield of gaseous hydrocarbons, high carbon
dioxide production, and lower contraction than that
normally expected at the operating temperatures. The
low activity of this catalyst may be explained in the
same manner as above.
2) As testing proceeded, the activity of the catalyst
decreased continuously. Table VI shows the progressive
decline in liquid and solid product yields and the
increase in production of carbon dioxide and gaseous
hydrocarbons. It is clear that the reactions favoring
the water gas shift and methane formation became pre¬
dominant after a few weeks 1 operation.
o
-90-
3) As before high temperature reactivation failed to
improve the activity of the catalyst.
4) Approximately half of the condensible hydrocarbon,
product was solid wax at room temperature.
5) The aniline point and average molecular weight
indicated some unsaturation in the light product.
3. Iron base Ammonia Catalyst .
Source of catalyst - Alberta Nitrogen Products Ltd.
Weight of catalyst - 187.6 grams.
Volume of chamber occupied by catalyst - 75 cc.
Reduction : Reduction with hydrogen was the same as
that for the other four catalysts. The temperature
was raised over a period of twenty hours to 362°C and
maintained for four hours. In the final stages, a
high boiling point gas oil was used as the heating
medium and no difficulties were encountered in main¬
taining a fairly constant temperature throughout the
whole catalyst bed.
Conditioning : Temperatures were lowered to 220°C and
conditioning was commenced at 100 psi. gauge pressure.
At 249°C a contraction of 3% was obtained which remained
unchanged even with temperatures as high as 292°C. The
test was then discontinued.
Yields : No liquid products, either water or oil, were
obtained. Only a small amount of methane in the residue
' ■
' ■ i. ; .. 1 ';ZO‘£q Jx* ’>"
'
*
■.. . o . ,o
. '
'
..
• -i . .1 ■ .. .. r ..
V'l'Gd' G;i,
•
• ::: f r
. i Cj
;:>r.
■
.. / if-.. ' .0
■
o > . .
...... ... ,
-
.
1
; x
' 1 ‘ , ! . 1 1... ' . :
■
■■ : \
' ■ ' ... 1 '; 0 r.
■ S' ' ■ ' ra 0 • : 0 ( ii
*
■ ■ • 1 • . i . ;• v:
90
80
70
60
50
40
30
20
10
0
- 92 -
gas indicated any sign of hydrocarbon synthesis.
Discussion : This catalyst, reduced and conditioned
as described above, has proven to be an inert Fischer-
Tropsch catalyst. Possibly the temperatures were not
high enough to reduce sufficiently the fused iron
oxide mass. When this catalyst is used for the
synthesis of ammonia, reduction is carried out at
temperatures approximating 500°C. With the present
equipment, this temperature would be very difficult
to obtain.
This catalyst is of the type used in the
^iso^-synthesis which involves a relatively inert
catalyst, low yields per pass, high space velocities,
and recycling. Such a catalyst would probably be
difficult to test in a one-pass system.
4. Cobalt-Thoria Catalyst
Source of catalyst - Harshaw Chemical Co., Ltd.
Weight of catalyst - 51.4 grams.
Volume of chamber occupied by catalyst - 58 cc.
Reduction : The catalyst was reduced with hydrogen
as described above by raising the temperature slowly
over a period of 20 hours to 360 - 380°C and held for
4 hours. A stream of methane through the jacket
acted as the heat transfer medium.
• ■ •'I
uoj t£ hob • . ..... ; oev
. ■ .. I..; '■e. - a , ■,
ad-: Cun t ... £> . ■.
,
» . ,J • . . ■ .; i, ■. -jqmoJ
. ■' ' ■. ■ . 6 o\. . .i -i - • u
■
. . ' : . ' \ J . . .
.
• ■ < .... . , J . !.,■ .. : ■- ' ' .
. .
•6,1. (■ .!
.
-• . ' i O OQ ■ . S t ' a . .
■ ' '. , rc<:L- .. . ■ ,o .
' .. .; .... • ;u
,
* ■ '
- 93 -
Conditioning : Synthesis gas was admitted at atmospheric
pressure and space velocity of 100/hr. to the catalyst
maintained at 161°C. No perceptible contraction was
obtained. The catalyst temperature was raised slowly
over a period of 175 hours to 190°C and the first
contraction of 5$ was obtained. Over a period of
six days, the temperature was raised to 19S, 206,
210 and 215°C with corresponding increases in contrac¬
tion of 8 - 9, 10, 16 and 17$ respectively. No
further increases in temperature were made in view of
reports that temperatures above 215°C favor carbon
formation and catalyst deactivation. The desired
60$ contraction was not obtained at atmospheric
pressure.
Operation : Operation was begun at 216°C, a space
velocity of 100/hr. and a pressure of 100 psi. gauge.
A high initial contraction of over 60$ was attained
which decreased quite sharply over a period of 40
hours to approximately 32$ (Figure XVa). Subsequent
operation at temperatures of 209°C and 200°C show
the same effect of deactivation, the contraction be¬
ing lower throughout the operation for lower tempera¬
tures .
Low Temperature Reactivation : At the completion of
each run, usually at the end of each week, the
catalyst was reactivated by passing hydrogen through
■
« '
,
.
.
i •• ' > • I ’ . ■ , • -
*
■
- ’ t
-
. ■ ; ■ - .
a t: i;;
. ; •: ' • o 'J ■ '■
♦ .. ■ ' ■ \;
1 ’ !
■ ■■' . . ■ ' ' 1. V C . i, :j .. ,0 J
. r .
■X :
;cidJUfVi
• ■
1 1 ■• • 1 . ) ■ : ; ,
; ■ - : . '.cl:
. •, c* r ' L'jj i , H j.'
• • . C- . o
*'■■■ f ■ . , \ . ;o
! Ov
. j ' - ' . ’ ' \J <?.;,■ ?(} [ ' : .
.
: / ' •: •
■> ■ ; . . •
' • ■
. .) , .. . t r. •
- 94 -
the bed at temperatures fifteen degrees above opera¬
ting temperatures. The catalyst activity was revived.
High Temperature Reactivati on: At the end of the
eighth week, the catalyst was given a high temperature
treatment i.e. hydrogen was passed through it at 350 -
370°C for four hours. This treatment increased the
contraction even above that obtained initially. This
effect may be observed in Figure XV b.
Yields:
Table VIII
Wk
Ho
Yields
of Products
from Co
-Th Catalyst
- .
h 2 o
Oper.
. Temp.
. °C
Vol. of
Feed Gas
Cu.M #
Liquid
Product
Light
Off-Gas
Heavy
ch 4
c 2 E 6
co 2
1
Activation
2.3
11.3
2
Activation
5.2
19.0
3
215
0^89
20.2
7.2
19
13
4.9
78.8
4
210
0.85
21.5
7.2
14
5
4.7
25
5
• *200
0.85
6.6
2.3
13
3
0.5
30
6
213
0.67
11.0
3.3
21
14
27
7
213
0.47
13.2
2.8
25
5
8.9
46
8
213
0.92
14.5
4.0
41
23
14.6
42
High Temperature Reactivation
•
9
200
0.62
17
3.6
15
11
9.1
71
10
200
0.75
17
2.0
23
6
12.1
63
Measured at U.T.P
- 95 -
Yields shown above are those actually obtained during
each week's operation. Yields computed upon a unit
volume of gas serve no useful purpose because of rapidly
changing yields other than to show that greater yields
per unit volume are obtained during the initial stages
of each run when conversion is high.
Analysis of Products
(a) There were no significant trends in the composition
of the residue gas during the course of the test. A
typical analysis is shown below (Sample #6R23):
iponent
Volume %
co 2
0.5
H 2
57.9
CO
29.4
CH 4
6.5
C 2 H 6
1.9
N 2
3.8
The results of the analysis of the light
and heavy products and the water revealed:
(b) Light Fraction:
1) Density @ 25°C - 0.7207 gms./cc.
2) Refractive index - 1.4093
3) Aniline point - 74.0°C
4) Distillation - see Figure XIV
5) Odour - suggestive of oxygenated organic
compounds
6) Average molecular weight - 85.
' .. ' VG« ibX .
, ! :
'■■■ - - : o . V.-;. . ‘ . ■. i.. : ":.r.ri jy
- : ■ *1 ■ ie o . . to
.' - • . I’ 7 ;: • ■ r • ......
.
.i '
& ■- . ■ o ~ , ;:;
-
' 'i
,
,
• ’ ) l ■ l ' , .
*
Percentage Contraction Percentage Contraction
- 97 -
(c) Heavy Fraction:
1) Density @ 25°C - 0.7668 gms./cc.
2) This fraction was a thick suspension of wax
in oil.
(d) Water:
1) Density - same as pure water.
2) Acidity - 0.004 N.
3) Distillation - very similar to that of pure
water.
Discussion : The behavior of this Co-Th catalyst dif¬
ferent in several respects from that of proven cobalt
catalysts.
1) In the conditioning at atmospheric pressure,
the desired contraction could not be obtained even
with a temperature as high as 214°C.
2) Operation indicated a very short life catalyst.
The initially high contraction dropped very rapidly
- 20$ in 20 hours. A drop of 10$ in four days is
normally expected from a good catalyst.
3) The yields of liquid hydrocarbons are about a
quarter to a fifth of that reported in the literature.
These Ioy/ values may be explained partly by operation
for considerable lengths of time at low contractions,
normally the catalyst is reactivated when the contrac¬
tion drops from 70 to 60 per cent. During these tests
contractions as low as 20 - 30 per cent were allowed.
; '■ o ■■ ’ o ;
.■. . ■ ,■
.
-
-
.
-
,
■ 1 £ ' . .. ■ ,; j; '
• O- '•• •'<! 1( .»v .
. ' • , ■ ,
,
• ' s n . . •
■ .
-
- 0 O'... •>! .? :it... . . •' o'
t i ■ ;
,
0 ... ■ • 1 J. ( l".
w
• ’ '
O 0 i. u j J
oJcf:
: ‘ >
. . VO -'J o ■„ o t
j
1 *• ' ' **
■ .0.;
, .
0
: . •
r
oo
. :<) I
o.o.' ■■ :
L ; i 0 0
- 98 -
Low liquid yields account for the high gaseous - liquid
hydrocarbon ratio in the products.
4) A high gaseous - liquid hydrocarbon ratio was
obtained. This may be due to the method of condition¬
ing. The desired 60$ contraction would not be obtained
at atmospheric pressure even with a temperature as high
as 214°C. Pressure operation at 216°C gave an
immediate 60$ contraction. This stidden increase in the
rate of reaction, as exhibited by the increase from
17 - 60$ contraction, may have favored, to some extent,
the production of gaseous in preference to liquid
hydrocarbons. This is the "methane reaction" reported
by Herington and Woodward (21).
Low temperature reactivation restored the
catalyst to approximately its original activity. High
temperature treatment did not increase the activity
much above that obtained by low temperature reactiva¬
tion. The relative contractions were much higher but
the liquid yield increase only slightly. The yield
of liquid hydrocarbons dropped sharply during the fourth
and fifth weeks suggesting that the catalyst had been
poisoned possibly by oxidation. Reactivations restored
the catalyst’s activity as the yields in the succeeding
weeks showed a gradual increase in both liquid and
gaseous hydrocarbons.
- \ ;• . ' . ,.,r 11 o ,I
.
,
.
’
. .v.v- / o.t*x© icrcomd-a ifa
V
,
. ■ • : ' .. .
■ ,. . •
r •
• i j , -po. , o i-:o : I :ji:J:: 0 'zxi .’©riJ
i r <j d
a ■'
’
.. ..
:c: • •
.
3 , ■ . r ■ 1 ■ . •. r;;; •>. ‘ ; ■<:
11
' ; ■ BO
l
.
3 £ 0
. i
r toe i . O'. ■- or. > or' , a:. -
,
- 99 -
A definite increase in carbon dioxide produc¬
tion indicated a change in the catalyst favoring the
water gas shift reaction. The high temperature
reactivation did not improve the catalyst in this
respect.
This catalyst produced an oil with a much
lov/er average molecular weight than that from the two
iron catalysts but at the same time produced a consid¬
erable quantity of solid waxes. The aniline point and
refractive index indicate that the oil is highly
paraffinic in character. The distillation curve, the
average molecular weight and the denisty show that
the oil from the cobalt catalyst was more highly satura-
ted than that from the two iron catalysts. This
agrees with reports in the literature.
The short life of the catalyst may be due to
the presence of only a few highly active centres upon
the catalytic surface. The condition may be due to
improper reduction or improper activation. During
operation a few centres may become deactivated by
waxing, this effect being accentuated because of the
small number. On the other hand, the short life may
be an intrinsic property of this particular catalyst.
I
■ - \ 7 • ..
'
» ' : - •
: . . ■■ ....
. '0 7 : ; .J. . J ; 0.
/ . '•• ’ • . 0 ■ ■ ) ■ 7;: KO 0 r
. ' . .. . : 7 orl.:/ 7-7 0 i:.C :v.J:
.
■
• . HQ i , ' " ■ • . .
• ; j '
. -
, . ' • . ' . ■
.
• ' V, : 7 >■": . 7.; O . .. :• . . : \i
. .
J*"!.
.
' - • ’ 7. 7 : v 'iJ,, „
-100
5. The Hickel-Manganese Catalyst
Source of catalyst - Harshaw Chemical Co.
Weight of catalyst - 67.6 grams
Volume of chamber occupied by catalyst - 70cc.
Reduction : The catalyst was reduced with hydrogen at
a temperature of 358°C. A gas oil heating medium was
used.
Conditioning : The catalyst temperature was lowered
to 160°C and synthesis gas was admitted at a space
velocity of 100/hr. and at atmospheric pressure. A
very small contraction was detected. Temperatures
were raised slov;ly and corresponding increases in the
contraction were obtained.
Temperature
°C
%
Contraction
Total hrs
Operation
172
5
5
183
8
24
198
15
35
205
23
38
211
31
42
218
41
48
230
57
54
Operation : Pressure was raised to 100 psi. and opera¬
tion commenced at 230°C with a 73$ contraction. This
dropped to 59$ over a period of 10 hours.
A low
'V.. .i
. ' ■- . J .
I. . •#
■ , kli '. ! ! ©. 04
.
■
• vd'2-
.
■
■i ■ ■ . ■ ‘.i
■
0 : •! jj
: 0 :
- •
.
• i *.i, ' O I;.',' ' ti>* ■: , .
■ :
y
• ■ ■ • ■ ■ ■ . :«©£ ;
* '• ■•’ . * * - *’ ■> 0
■
- 101 -
temperature reactivation was then carried out at 240°C.
Hydrogen was passed through the catalyst during the
shutdown over the week end. Operation was recommenced
under the same pressure conditions but only 47$ con¬
traction was obtained. This decreased to zero over a
period of IS hours.
The catalyst viras given another low tempera¬
ture reactivation and operated at atmospheric pressure
at 231°C. A contraction of 11$ was obtained and it
decreased in time. Increasing temperatures to 239,
254 and 262°G increased the contraction to 13, 20
and 22$ respectively. These contractions however,
decreased with continued operation. This concluded
tests upon the Hi-Mn catalyst.
Yields : Wo yield of liquid hydrocarbons, except one
and a half cubic centimetres of light oil, produced
during conditioning, was obtained. The residue gas
analysis revealed high ethane and methane formation.
Water and some carbon dioxide was produced.
D iscussion : This catalyst was activated according to
the method prescribed by the U.S. Bureau of Mines for
cobalt catalysts. Although the desired initial con¬
tractions were obtained, the decrease in contraction
with time and the extremely low yield of liquid
hydrocarbons indicated that, operating under these
conditions, this Wi-Mn catalyst was ineffective in
promoting the Fischer-Tropsch synthesis.
0..
i: j .v
■
• • .■ :
0 . . '
■ .
. \ ,/ i:
■
,
'
■. -
",
vX.iC:
■ ■ -
■ 1
. ,
.
fiO-j:
uO
,
t
„ I
0 . ' • ’ ,;C
. o L' \;.cr’
. i.
' ,o . o
J
.
■ . J».
' ;
i .'"■rf ;.' 1 .j
r,.:. ■■ . ■ ■ i ■ , i . ■ ■ o.,'.
. •
,
■ ■ . •
.
*
,
: ; \ •.v .. , ..:o . I,’.. : ;.u
■t *
.
, :j . ■: i
‘
. I:: 0
’
"
- 102 -
Low temperature reactivations failed to
restore the activity of the catalyst for either
pressure or atmospheric operation. The loss of
activity may be due to one or more reasons:
i
1) Poisoning by oxygen or sulphur. Delicate
tests fail to indicate either in quantities large
enough to be detrimental.
2) Too high a temperature employed. To obtain
the desired contraction, the temperature had to be
raised to 230°C. This is considerably higher than
the optimum reported by Fischer (15) (190° - 210°C).
The high yield of gaseous products, methane and
ethane, with little or no liquid products also
supports this vieif.
3) Loss of active centres by carbonyl formation.
Komarewsky and Riesz (23) report that nickel catalysts
soon lose their activity at medium pressures due to
the loss of active centres in the formation of nickel
carbonyl. Nickel carbonyl is very volatile and thus
removes the active centres irreversibly.
• r ■ ,• ‘ CCCCC - ■; ,'i C0,.C
■ •
; :■ i- c c c x x':. *,■: cc 3
■
.
,
.
■
f - rs - 1 ■ ■ (<.. • i ■ '
" ■ ..... ■ : V .:. X ; V . ■ ! .m
*
■ _ 9 j [ ottbQ' ■ 3 . . ,
.
• n - ;■ :
. ■■ ... 0
".!• -v .r, :: i: . rv '.. .v ;;aei:x. xi
I a S f D } O f ■ . •
'
• X • : ■.
■
.... . ■ VC" . 1 .
- 103 -
VII SUMMARY AND CONCLUSIONS
Five catalysts have been tested for their
ability to promote the Fischer-Tropsch synthesis.
Their performance and characteristics have been studied
and discussed. Under the conditions of the tests,
none of them has proven to be an effective Fischer-
Tropsch catalyst. Poisoning is a possible explanation
of their poor behavior. However, sensitive tests have
failed to detect the presence of either oxygen or
sulphur in the synthesis gas. The catalysts exhibit
some signs of poisoning but the evidence is confusing
and inconclusive.
In the preparation of the Bureau of Mines’
catalysts, the digestion and washing stages extended
over several days with possible temperature fluctuations.
This deviation in procedure may have affected the
activity of the catalysts.
The percentage contraction of the synthesis
gas in passing over a catalyst is not a positive
criterion of catalyst preformance. In practically all
.
» - •: . : ; , : [ , . . .
*
. 1 -.i ... \V: . ''
• .
■’..
. 1. i'
’
' '
'
].
J
■ i . . . .
!.
v iooo’oj
: < \J '.j .
V ■;o ,
„
•:0.L'Va
« •
■ ; l: Vo
•
: ' ' o
. ; ■ . ■ .
•
"0.0'
'
o', V
1
Q$
V oo' O
: ;o : . ■ :
.. J. ci mo
. . ,i‘, t.-i oa
*
‘.O'
■
■J : on.i. o:
• ■ . • ■' 0
■ ■ - ■ ' 0
. - o
0 ■' '' ' HO M O •„ J ' t ,;VVc 0'
c ' 1/0 ' -■ , •/ ; C 0 ' i; £
- 104 -
cases, although the normal contraction was reached,
the liquid hydrocarbon yields were much lower than
expected. With comparatively inactive catalysts, the
methane reaction and the water gas shift reaction
appeared to predominate, giving high yields of gaseous
products.
Some difference has been found in the products
from the two types of catalysts. The cobalt catalyst
yielded a more paraffinic product than the iron. The
former yielded a high 'percentage of wax as might have
been expected. The pure iron catalyst yielded only
oil, while the iron-copper gave some waxes.
The testing of Fischer-Tropsch catalysts in a ^
dynamic system requires special operational techniques.
These catalysts are characterized by temperature
s
sensitivity, low space velocities, pressure operation,
and high sensitivity to poisoning. No difficulty has
been found in maintaining and controlling the operating
temperature. Initial difficulties were experienced in
maintaining the pressure in the units constant. How¬
ever, this was remedied by modifying the Cash-type
back pressure regulator, replacing the composition
seats with a hard vulcanized fibre. The special feature
of the experimental equipment was the use of automatic
control on these extremely low rates of gas flow. This
has on the whole performed quite satisfactorily.
: ... r : ■ ■■ ■ •. , ■ ' 1 . \ >1 ; .....' ;
■
, .
■ ■ • .1 ' . •. .' ..o;;.. „ • i, a.£fj Jom
. j ' , ; • ■
. ,
s ri a «*/< 1 net 3 ■ ■ ■ t
>
■< ■
- -«
o
9
,
0 .
• •
T ■ ■ • ■
e " ,
• J
■
-
•
‘.-1 J V V
•
Li
. ■;: -
:; -i ■; v. .'
.:.
0 .'o 5
a }■
cmi e .'
.. ibid
;
.
V> ! : : :::
■ .... :J
. ' o
Q
.
::oo u-’7
. ,0,,
1 10ci ".Mi
j ] L
j;.-: colmL:
... .' . • ; ' : .
. ;• '. ,.. . :'c .. " .,, . . V ,t:>G ■ '
.
‘
■ ■ " ■ . ;• . .
.
a . oo wnJ . t or; Iovm
. ■
.
•
v. >
O'l e
•
■ . ..
• v
' :5£ i ■
>ii.C V .
■
J ... i
5*
i
o.r 'is-
APPENDIX
A Preparation of Synthesis Gas
B Design and Calibration of Flowmeters
C Short-Time Catalyst Testing in a Static
System
..
i.
*
- 105 -
A - PREPARATION OF SYNTHESIS GAS
The Fischer-Tropsch process, unlike the Ber-
gius process, is an indirect conversion of carbonaceous
material to hydrocarbons. This gives the former pro¬
cess a distinct advantage in enabling the use of a
wide variety of raw material. All types of coal, or
coke, can be used in preparing the synthesis gas, a
mixture of CO and H2. The waste gases of the large
steel industries have been suggested as a possible
source and natural gases have been used on a small
scale. Large scale industrial development in both
Germany and Britain has been based on the incomplete
combustion of coal and coke. However, production from
natural gas, the apparent cheapest raw material on this
continent, is still in the experimental stage.
The simplest and most convenient method of
preparing synthesis gas for laboratory purposes is by
the partial oxidation of natural gas with oxygen,
_
.
. ■ - ' ■
::
.
.
.
'
\
i
c
■»
.. 'J . " .
-
.
,
.*
■<- - .■
. •' : r
-
: .
3
«
' ■( j '
*
.
< i '--J "0
’
- 106 -
carbon dioxide or steam:
CH4 + 1/2 02 v
CO
-*■ 2H2
(1)
CH4 ■+• CO2 ^—
- v 2C0
+ 2H2
(2)
CH4 -t H20 —
A CO
^ 3H 2
(3)
These reactions give
H2:C0
ratios
of 2, 1, and 3 respective-
ly. Since a synthesis gas with a 2:1 ratio was chosen
for the present program of catalyst testing, reaction (1)
was employed.
The preparation of water gas is a batch opera¬
tion. Methane (high methane natural gas) and oxygen are
fed simultaneously into a water-seal type gasholder, A,
of about 10 cu. ft. capacity (Figure XVI). The rates of
flow of each gas, measured by glass capillary flowmeters
B and C, are carefully controlled to give a 2:1 methane-
oxygen ratio. A high pressure storage, maintained by
pumping natural gas from a commercial gas line, serves
as the methane supply; the usual commercial cylinders
were purchased to provide an oxygen supply.
When gasholder A is full, the methane and
oxygen supplies are disconnected. The gas mixture is
then passed back through flowmeter C to the reaction
chamber D. The latter is essentially a 3/4-inch quartz
tube, 18 inches long, and heated by a Multiple Unit
Electric Furnace, operating on 220-volt supply. A
6-Inch catalyst bed is located in the middle of the
tube, with 5 inches of porcelain chips at both ends.
,
. ' • .... •
.
■ A .. ■ . A" . .. d
5 •— . '
. .
I
. ' '
\
.
■ .1 , ,
: 5 ■ v. 11:1 ;■ i
■
....
«
■ A ..
,
■
-
1 f
107
- 108 -
Thus the catalyst, well inside the furnace, is held, in
its entirety, at a reasonably uniform temperature. (This
precaution is necessary since the water gas reaction re¬
verses over the catalyst at lower temperatures.) The
thermocouple well, a l/4-inch quartz tube, containing a
Pt - Pt 10$ Rh thermocouple extends to the catalyst
bed. Passing over the nickel catalyst at 900° - 1000OC,
the CHq and 02 are converted to CO and Hg, with small
amounts of C02, H20, and H 2 S. Water is removed in the
water trap S; CO 2 and H 2 S in the caustic scrubber F.
The caustic scrubber consists of three sec¬
tions: the absorption column F, and upper and lower
storage vessels 0 and K. The first is a 12-inch length
of 1-1/2 inch glass tubing, packed with 6-mm. diam.
glass Raschig rings. Glass wool at the top serves to
give an even distribution of the scrubbing liquid, pre¬
venting channelling. Compressed air is used to elevate
the solution (25$ NaOH) from K to J, This liquid flows
by gravity at a controlled rate through the scrubber.
The synthesis gas is purified on passing counter-
current ly through the scrubber.
The synthesis gas, carbon monoxlde, hydrogen,
and a little nitrogen, is stored in a water-seal type
gasholder, G, of approximately 15 cu. ft. capacity.
Flow is maintained through this system by a difference
; ' ,
. , i
• ' : ' 7
;i i Jiiti .. I \ro ' '..-.arv'
' -- , ■ . •
- ■
.
-
: ■ •
? .
.
...
\
• -
c
■
'
- ..... .... ■■ .. ...
- : • ;. ' ‘ ‘ . ' ’ ... . L : .;) 'i ,' yjti'I?
... :
. ■ ' ' ,
( t ...
- 109 -
of pressure of 5 - 6 inches of water between the two
gasholders.
The catalyst used is pure nickel on porcelain,
as developed at the University of Alberta in 1926 (la).
It is prepared by soaking porcelain chips of 4 - 8 mesh
for a period of 24 hours in a saturated solution of
nickel nitrate. The chips are removed and heated
strongly in a nickel crucible until brown fumes are no
longer evolved. The catalyst, now NiO, is placed in the
reaction chamber and reduced by heating slowly to 1000OC
in an atmosphere of hydrogen.
Oxygen and sulfur, even in very small quan¬
tities, are serious poisons to Fischer-Tropsch catalysts.
Both attack the active centers, permanently deactivating
i*eiwova\
the catalyst. Oxygen^\presents no serious problems. The
reducing atmosphere in the synthesis gas furnace re¬
moves any excess oxygen that may be present. The stor¬
age and connecting lines are always under a slight
positive pressure and any leaks would be outward. Sul-
re moved
fur^ however, presents greater difficulties. Fortunately,
in the reducing atmosphere of the furnace, all the sulfur
compounds are converted to H 2 S, which is later absorbed
in the caustic scrubber. Complete removal of the sul¬
fur is essential, as even minute quantities (0.1 grains
per 100 cu. ft. of gas is the upper allowable limit)
can quickly poison a catalyst.
' ; , ‘ \Ci
.
:
■
-
'
V . ‘
'
.J' ■ . . V
. .
*
-
.
. .
-
:o i.--:
’ .
1
.. ■
.
.
•
■*-
■
)
-
•• n
-
■
.• . ■ i , : •
: ■ ■
; ' ,: '■ : . v ..
.
'
-
-
- 110 -
Little difficulty was experienced in the
manufacture of this gas. Though methane and oxygen were
mixed in an exact 2:1 ratio, an analysis of the syn¬
thesis gas revealed a product slightly high in hydrogen.
However, by increasing the proportion of oxygen in the
mixture, the proper H2:C0 ratio was obtained. This
ratio could be maintained with a deviation of not more
than four per cent.
,
.
$
' :
-
.
e ■ r ;
.
- 111 -
B - DESIGN AND CALIBRATION OF FLOIMET3RS
One of the commonest methods of determining
the rate of flow of fluids is to measure the pressure
drop caused by the insertion of a restricted opening
into the line. Thus orifice and venturi meters are popu¬
lar industrial flow measuring devices. However, for
laboratory work where the rate of gas flow is small,
capillary flowmeters have proven the most satisfactory.
Their behavior is dependable and predictable; the pres¬
sure drop easily measurable.
Since the gas flow in the Fischer-Tropsch
synthesis is measured under 100 psi. pressure, a flow¬
meter of sturdy construction must be employed. Two
other features are incorporated into the design—a
separate capillary holder which may be conveniently re¬
moved from the main flowmeter block; and provision for
bypassing the gas while this is being done.
A schematic drawing of the whole flowmeter
is shown in Figure XVII. The main body, A, is machined
• L
.
t . !
-
'
.
0 ' J, ' ' '
■
l
.. o EMS
. . 1 ' o \ '*j.. ;;j-oI
.. 1 ’ . : ‘ .< o
I *
' : . 1 , 'u.
.
-
- 112 -
Figure XVII - ^a'pillury Flowmeter led
- 113 -
from a 2-inch hexagonal steel bar and has an overall
length of 6 inches. In the interior of the block a
brass capillary holder B is threaded and sealed by a
garlock gasket. The gas passages and pressure taps are
located as shown on the diagram.
Three valve stem assemblies, C, D, and E
(from l/4-inch needle valves) direct the gas flow. The
brass needle seats directly onjfco the steel edged open¬
ings in the flowmeter block. The capillary holder
chember B is closed by a threaded steel plug G, which is
tightened with a dolly bar. Pressure taps at H and K
are connected to a Merriam mercury manometer, which
indicates the pressure drop across the capillary. Lead
gaskets seal the valve stems and the steel plug to the
flowmeter block.
The brass capillary holder B is threaded
snugly into chamber E. One end is notched to permit the
use of a screwdriver for tightening. The capillary L
(a short length of thermometer stem) is held tightly in
place by the brass gland and packing consisting of a
short piece of rubber tubing.
In normal operation the bypass valve D is
closed and the other two (C and E) open. The gas enters
at M, passes into the hollowed section E, through the
capillary, and discharges through outlet N. On closing
valves G and E and opening valve D, the cep illary is
. .... :
-
*
c *X> •.
»
C v •
■■ : V- ''t;
•
'
- '
*
«
•
. J O'. ■
.
■ ih n
.
♦
C '
>
. '' . V ’
' v ' . • , , ' ,
-
. / i • '-X '- ;i
,
.
.
,
'
f * v • . • 'v _ ‘ i' „ a '• .a<) V JL /
r
- 114 -
taken out of the gas stream and may be removed and
cleaned without interruption of gas flow to the reactor,
manual flow control being necessary, of course, during
this operation.
Capillary tubing of the proper length and
diameter had to be chosen to give a suitable pressure
drop for normal operational flows. Govier (.19) has
combined in one equation the various factors which
determine the pressure drop across a capillary tube with
square ends, through which a fluid is flowing .
H 3 ..JA- (81pQ, + #& )
ZTgI>4 ' 7T
H s pressure drop across capillary - cms. HgO
D = diameter of capillary - cms.
1 st length of capillary - cms.
p - viscosity of flowing gas - poises
= density of flowing gas - gm./c.c.
Q, s rate of flow (under flowing conditions)
c.c.* s/sec.
This equation reveals several interesting
features about capillary flowmeters:
1) If the end effects are small, then the rela¬
tionship between Q, and H is linear. With the low space
velocities used in the synthesis (of the order 0.4 c.c.*s
per sec.), this condition is certainly satisfied. There¬
fore the second term in the above equation may be neglected.
■
« t
; ■
.
. , '
*
r. o
tv
fJ
< -
■ ..
' . ' ■
,
1 ■
-
1 t ,
1 .. :.
r
• ‘ C . .• ,
- 115 -
2) The quantity Q, in the equation refers to the
volume of gas as measured under flowing conditions,
i.e., at the average of the upstream and downstream
pressures. Since the viscosity of a gas is almost
independent of pressure (up to about 10 atmospheres),
then the capacity of the flowmeter is roughly proportional
to the pressure of the flowing gas.
The design equation may he expressed in the
following form:
1 = TT-gP* (4H)
128 ^uQ,
In order to determine the proper length of
capillary to he used, it was decided that synthesis
gas flowing at 100 psi. gauge, with a space velocity of
150/hour (or 0.4 c.c.*s/sec. at 8 atmospheres) should
give a pressure drop of approximately 4 inches of mer¬
cury. All the variables except 1 and D (and u in the
case of the gas mixture) are now known or readily cal¬
culable. Data on the viscosity of CO - H 2 mixtures
are very scanty and the relationship between viscosity
and composition is obscure. It was finally necessary
to estimate the value of u, knowing the behavior of
CO - H 2 and N2 - H 2 mixtures over limited ranges.
Convenience with regard to the size of the
flowmeter block limited the length of the capillary to
a maximum of 12 cms. Preliminary calculations
t ' • ' ■ *£ 3 yi ..■■,/ ., : •; nKtrrrm*
-
.
) ®
.
' ■■ ji [n ;t
i v ■
'
.
.
M r")
* <“
• 3;
„
OJ.CO
.v; .... ■..
v.
*
t ■' : ■ ; ot
:
*
j - ■' ‘ ■ 0 ' ’ ' 1 •, - .C ■' •/
■ . ' ■ .
- 116 -
indicated that only thermometer capillaries had suffi¬
ciently small diameters to permit this requirement to
be satisfied. By sucking up and weighing a measured
length of mercury in the capillaries, it was possible
to determine the average diameters. Knowing these
data, the proper length of capillary in each case was
determined and the cut capillaries inserted into flow¬
meters.
These capillaries were designed approximately
and calibrated absolutely under normal operating con¬
ditions. The calibrations were carried out for hydrogen
and synthesis gas at atmospheric pressure, and for the
latter at 100 psi. gauge.
It is obvious from the design equation that
a capillary designed for a space velocity of 150/hour
for synthesis gas at 8 atmospheres would only permit a
space velocity of 20 - 30/hour for synthesis gas or
hydrogen at atmospheric pressure. The latter rate is
satisfactory for hydrogen, since it is used only in
reducing^conditioning, and reactivating the catalysts,
where low space velocities are permissible. These
capillaries are, however, not suitable for operating
with water gas at 1 atmosphere—the allowable space
velocity being too low for effective catalyst testing.
Nevertheless, a calibration for this last case was
carried out in order to determine whether or not the
'• ■ ■
;
•
.1 if •
«
■
:/
-
' '
.
' ..f ... <
, ' ■ , ' . u' . ■ . i.:. i j , ;y ■ ■ ■, f )r :r
... ' ■
.
’
•
a V i
* . . / /'
" ■ v ' • ' ■ .
■j - ' ’ - !
■ . : - ' . •
J L .
*
. , ' o '.
• ..
C ' ;
.
<
.
J0 - j • - • '• . . .. .
1 i
■ ' 1' - ‘ ■ •'
a:.;.: j . ; .. •. ....
•• t ' ' .......
- 117 -
effect of pressure on the behavior of the flowmeter is
predictable from theoretical considerations.
The calibrations proved to be a rather
formidable task because of the small flows involved.
Any small leaks in the system were sufficient to change
the flow by several hundred per cent. It was found
necessary to check for tightness before every set of
experimental runs. After passing through the flowmeter,
the rate of gas flow was measured accurately in a
Precision Wet Test Meter. A stopwatch was used to note
the time required for the passage of a 1/10 of a cubic
foot of gas (corresponding to one complete revolution
of the meter)* This was done in duplicate for three or
four different rates of flow for each capillary and for
each set of conditions. Since some single runs took as
long as two hours, the calibrations were tedious and
time consuming.
The rates of flow (measured at atmospheric
pressure) were then plotted against the pressure dif¬
ferentials. A typical set of curves for one capillary
is shown in Figure XVIII.
Since the rate of flow Q,' is in this case
measured under discharge conditions, the graphs show a
slight curvature. In order to test the validity of
the design equation, it is necessary to calculate the
flow rates as measured under the actual flowing conditions.
-
'
'
, :
•
'
.
o
,
r
♦
'
*
■, :
.
■
•
■*
. .
118
1
7
8
Discharge - litres/hour
Flow - litres/hour
Manometer Differential - inohes of Hg#
- 120 -
These results were then plotted—rate versus
pressure differential. A typical set of graphs is
shown in Figure XIX. Six capillaries were calibrated
and 18 curves drawn. An examination of these curves
led to the following conclusions:
1. In all cases the linear relationship between
the rate of flow Q, (under flowing conditions) and the
pressure drop^H was confirmed.
2. The ratios of the slopes of the curves for
hydrogen (1 atmosphere):synthesis gas (1 atmosphere):
synthesis gas (8 atmospheres) for all of the six capil¬
laries were consistent. This is to be expected since
these ratios depend only upon the flow characteristics
(viscosity and density) of the gas being measured.
The following table shows this clearly—the ratio of
the slopes being based on a value of unity for the
hydrogen curve.
TABLE IX
Correlation of Slopes of Calibration Curves
llary No.
Ratio
He
1 atm.
of Slopes
CO: 2Hg
1 atm.
of Curves
C0:2B 2
100 psi
5
1.0
0.59
0.57
6A
1.0
0.59
0.57
6B
1.0
0.61
0.55
10
1.0
0.60
0.57
11A
1.0
0.60
0.56
11B
1.0
0.60
0.56
w ' ' 1$' . ,)
:
i '
. l< ' \
*
* , . ■■ : ,
,
• ■ - -J . >
' ' !- '
1 ,
- - '■ • ' . ■ .
- 121 -
The average deviation is ]e ss than 1$. This
indicates the correctness of the calibrations.
3. The slope of the curves Q, « f 04 H) for syn¬
thesis gas changed only slightly with an increase in
pressure from 1 atmosphere to 8 atmospheres—there being
a decrease in slope of only about 7$. It is therefore
possible to interpolate and obtain curves for any
intermediate pressures without incurring any appreciable
error.
4. The effect of gas leaks on the calibration
curves was clearly demonstrated. A leak in the dis¬
charge system causes the graph to shift to the right.
A bypass leak through the capillary holder or through
the bypass valve changes the slope of the graph. These
two effects are illustrated in Figure XX.
5. The results did not agree with the calculated
values of the basic equation. Deviations were found
as high as 25$. Therefore, while the equation is of
much value in the design of capillary flowmeters, with
flows of this order^the latter should always be
calibrated absolutely.
■
,
•" ■>, V * .
,
■
■
.
: . -
■ /. .
-
>2.
. '
■
; • • ■&
■
’ .. a .
. . ;
r
’
’ : ' ' v, ' .
- ' ' .
■
Flow (measured under flowing conditions) - litres/hour
122 -
- 123 -
C - SHORT-TIME CATALYST TESTING IN A STATIC SYSTEM
Introduction
The proper evaluation of a Fischer-Tropsch
catalyst requires weeks of testing under conditions
approximating commercial operation. This is time con¬
suming and it is desirable to devise a short-time test
which can he used to predict the catalyst behavior.
Specifically, what is sought is a correlation between
the results of these short-time tests and those
obtained from long term dynamic tests.
An investigation of this type was started
but is still only in the preliminary stages and no
conclusive results are as.yet available. The method
used involves the reaction of synthesis gas in a static
system at 1 atmosphere over a reduced but not conditioned
catalyst. As the adsorption and reaction proceeds,
giving hydrocarbon products (mainly methane but pos¬
sibly some liquids), the pressure in the system drops.
Pressure-time readings are recorded and plotted.
- 124 -
Assuming the initial reaction to be only
CO + 3H 2 -
—> CH 4 -+ H 2 0
or 2C0 + 2H 2 -
—> CH 4 -h C0 2
the slope of the curve gives the initial rate of me¬
thane synthesis. Ultimately, a correlation between
these rates and catalyst behavior in liquid production
will be sought, the absence of any such correlation
vitiating this method of evaluation.
Apparatus
The individual pieces of equipment listed
below are represented in their proper places in
Figure XSI.
A - Storage chamber for the synthesis gas.
B - P 2 O 5 drier for drying the synthesis gas.
C - Mercury manometer (vacuum on its right arm)
to measure the absolute pressure in A.
D - Activated charcoal to purify the hydrogen.
E - Reaction chamber.
F - Copper wire basket, containing the catalyst
pellets.
G- - Mercury manometer.
H - Paraffin wax constant temperature bath, heated
by a hot plate and two knife heaters (con¬
trolled through a rheostat).
K, L, M - Three-i/ay stopcocks
. J.
' j • 5
' . I i
r
■2 ■' •' . "
/■: ',V '
i. : ■
..... .:
■ ! ’ '
> ■■■■. .
' •
.
■■■ ' ■ > . ' ;■ . , laoD-i^o ot
.
■
-■
-
.
*
-
-125-
- 126 -
Procedure
The copper wire basket F was filled with
catalyst pellets, weighed, and placed in the reaction
chamber E. The latter was surrounded by an air bath.
The whole system, including the storage chamber A, was
evacuated. Purified hydrogen was admitted into E and a
slight but positive flow of hydrogen was maintained,
the latter leaving at N. The catalyst was heated to
360° - 370° C and held at this temperature foe 4 hours.
The storage chamber A was filled with synthesis gas and
the pressure measured on manometer C. After the
catalyst had been reduced, the air bath was replaced
by a paraffin wax bath H and the temperaturereduced to
185°C. The reaction chamber E was evacuated free of
hydrogen, and synthesis gas from A admitted. The
final pressure in A was recorded. On admission of the
synthesis gas into E, the pressure was read on the
manometer G and the stopwatch started. Pressure-time
readings were taken, at 1-minute intervals at first,
and at longer time intervals when the pressure
readings became steadier. The temperature was main¬
tained constant by adjusting the rheostat which was
connected in series with the knife heaters. Each run
took from 30 to 60 minutes, depending upon the catalyst.
Duplicate runs were made at 185°C, 190°C, and 195°C.
.
- ' I
■ K':
.
.
-
-
■
,
, ;
■
■ ■
'
? ' ' , .r ;j
. - . ' . ,
■ -
" J ,
i ” .J'X
‘
y i ■ <} ... I
*
-
• ' ' ■ • '
• ■
' ■' ' ' ■ 3
■ ■ . ' ,
- 127 -
Results
Three catalysts were given preliminary test,
namely;
(1) Fe-Cu - Bureau of Mines #24.
(2) Co-Th - Harshaw Chemical Co. Ltd.
(3) Ni-Mn - Harshaw Chemical Co. Ltd.
A comparison of the activity of each may he seen in
Figure XXII. The effect of temperature upon the re¬
activity of the cohalt catalyst is shown in Figure XXIII.
Discussion
With respect to the Co-Th catalyst in parti¬
cular, successive runs at the same temperature gave
duplicate results. Similar performance was obtained
after reactivation for four hours with hydrogen.
Whether this catalyst could he reactivated after cooling
to room temperature was not investigated. Tests upon
the Fe-Cu catalyst were inconclusive.
At the temperatures tested the Fe-Cu was
much less active than the other catalysts. This is
probably not a fair comparison since iron catalysts
operate normally at a temperature of 250°C. The ni¬
ckel catalyst revealed a higher activity than the
cobalt but the per cent contraction was less. Since
the contraction in the case of the nickel catalyst was
approximately one half, the methane reaction is
indicated. The contraction in the case of the cobalt
- 128 -
Time— minutes
Time - minutes
- 130 -
catalyst is 65 per cent, and therefore cannot be ex¬
plained wholly by this reaction. Adsorption or the
formation of higher hydrocarbons are the only two alter¬
natives. Figure XXIII indicates that the temperature
coefficient of reaction is comparatively low. This
would be the case if, as has been suggested by some
authors, diffusion were the rate controlling factor.
Insufficient data have been obtained to
determine the order of reaction. The same is true at
this stage of the work with respect to the value of
these rate measurements as indicative of catalyst per¬
formance in the Fischer-Tropsch synthesis.
'
" • ' •
. ^ MMH r^K . j ra j M EB. i "H. 1 Sjfl'
,
. ' ■ . .. •. ‘i : . ,
,
•jp
'
,
* ' • ' ' a
'
— ' ‘ . ■ ... ■:
* ■ • • t o ok ■: •. c
- 131 -
BIB LI OGRAPHY
(la) Boomer, E.H,, Can. Pat. 289,830, (1926).
(1) Craxford, S.R., Trans. Far. Soc., 35, 946-958
(1939). “
(2) Craxford, S.R. and Rideal, E.K., Jour. Chem.
Soc., Part II, 1604-1614 (1939).
(3) Domai. S., Bull. Chem. Soc. Japan, 16., 213
(1941), C.A. 36, 739 (1942).
(4) D.R.P. (German patent) 293, 787 (1913).
(5) D.R.P. , 293, 797 (1913).
(6) D.R.P., 705, 528 (1938).
(7) Egloff, G., Nelson, E.F., and Morrell, J.C.,
Ind. Eng. Chem., 29, 555-559 (1937).
(8) Elwin, O.E. and Nash, A.W., Nature, 118 .
154 (1926).
(9) Fischer, F., Brennstoff-Chem., 16, 2 (1936).
(10) Fischer, F., J. Inst, of Fuel, 10, 10 (1936).
(11) Fischer, F., Ber., 71, 56-67 (1938).
(12) Fischer, F. and Pichler, H., Brennstoff-Chem.,
20, 41-61 (1939).
(13) Fischer, F. and Pichler, H., Brenn. Chemie,
21_, 285-8 (1940).
(14) Fischer, F., Roelen, 0., and Feist, ¥., Brenn.
Chemie, 13, 461-8 (1932).
(15) Fischer, F., Roelen, 0., and Feist, W., Re¬
finer, 22, 429-35 (1943).
(16) Fischer, F. and Tropsch, H., Brenn. Chemie.,
4, 276-85 (1923).
(17) Fischer, F. and Tropsch, H., Brenn. Chemie, 2>
97-104 (1926).
* •:
<; 4 c
t
c. , .
- 132 -
(18) Foster, ti.L,, Oil and Gas Journal, 43, No* 15,
99-101; No* 17, 46-49; No. 18, 66-69; (1944).
(19) Govier, G.W., "The production of formaldehyde
by low pressure oxidation of high-methane
natural gas", Thesis, Part I, Appendix A,
University of Alberta (1945).
(20) Hamon, F., Mayer, A., and Plantefol, L., Ann.
physio, physicochim. biol., j6, 452-63 (1930).
C.A. 25, 4292 (1931).
(21) Herington, E.F.G. and Woodward, L.A., Trans.
Far. Soc., 35, 958-67 (1939).
(22) Hofer, L.J.E., "The preparation a.nd properties
of metal carbides", Bureau of Mines (U.S.
Department of the Interior) Booklet, July
(1944), p. 1, 39.
(23) Komarewsky, V.I. and Riesz, C.H., Refiner, 25 .
415-22 (1944).
(24) Myddleton, ¥.1., J. Inst, of Fuel, 11, 477-92
(1937).
(25) Myddleton, W.W., J. Inst. Petroleum, 30, 211
(1944).
(26) Napthali, M., Refiner, 17, 47-51 (1938).
(27) Russell, R.P., Hearings before a Subcommittee
of the Committee on Public Lands and Sur¬
veys, United States Senate on Synthetic
Liquid Fuels, p. 39 (1944).
(28) Sabatier and Sendereus, Compt. rend., 154 . 514
(1902).
(29) Shaw, A.G., "Preparation and Testing of Fischer-
Tropsch Catalysts", Unpublished report,
Dept, of Chemistry, University of Alberta,
(1945).
(30) Storch, H.H., "The Chemistry of Coal Utilization",
edited by H.H. Lowry, John Wiley & Sons,
New York, N.Y., 1945, Vol. II, pp. 1797-
1845.
(31) Underwood, A.J.V., Ind. Eng. Chem., 32, 449-53
(1940).
5
1
S * ?
: ' 1 ^
c *
*
.
*s
- 133 -
(32) U
(33) U
.S. Bureau of Mines (Department of the Interior),
Pittsburgh, Penn., Private Communication
(1944).
.S. Steel Corporation, "Methods of the Chemists
of the U.S. Steel Corporation for the
Sampling and Analysis of Gases”, Booklet,
Third Edition, p. 106 (1927).
r .
. ■ ' ;
. '
. ' :• ;
„■ .
«
.
B29752
