Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
2028588
'',,._
FIXED-BED FISCHER-TROPSCH CATALYST AND MANUFACTURE THEREOF
FIELD OF THE INVENl'ION
This invention relates to catalysts. More particularly, the invention
relates to a fixed-bed Fischer-Tropsch catalyst, to a method of manu-
facture thereof, and to a process for the production of aliphatic
hydrocarbons comprising wax.
. . .
BACXGROUND OF THE INVENTION
The Fischer-Tropsch process has been adapted for a number of different
applications. In the circulating fluidised bed (CFBI application,
finely divided iron particles are added to a stream of synthesis gas,
causing the particles of the catalyst to be entrained in the gas. The
synthesis gas is then allowed to react in a reactor before being fed
to a vessel in which the catalyst is allowed to separate from the
reaction mixture before it is returned to the reactor. In the fixed
fluidised bed (~) application, the catalyst is fluidised in the
reactor vessel by synthesis gas fed to the reactor, and is separated
from the reaction products before they leave the reactor. In the
fixed-bed (FB) application of the Fischer-'lrops~h process, the cata-
lyst is in the form of extruded pellets or granules which are packed
- in tubes located inside a tubular reactor. The reaction temperature
is generally lower than that in the CFB applicationr and for that
reason the FB application is also known as the low-temperature ~LT)
variant of the Fischer-Tropsch process.
- The F~ variant favours the production of longer-chain paraffins (Clo ~
Cloo), whilst the CFB variant yields mainly medium-range (Cs - C20)
paraffins and olefins. The present invention concerns the FB or LT
- Fischer-Tropsch process.
The preparation and composition of conventional fixed-bed iron cata-
~2 lysts for the Fischer-Tropsch process are described by Frohning et al
(C D Frohning, W Rottig and F Schnur, in J Falbe ~Ed), "Chemieroh-
stoffe aus ~ohle", George Thieme, Stuttgart, 1977, p 2341. In accord-
ance with the manufacturing process as described in this publication,
iron and copper-are dissolved separately in nitric acid at elevated
te~perature, whereafter the solutions are adjusted to lOO grams Fe per
.
2028~88
litre and 40 grams Cu per litre respectively. The solutions are then
stored separately with a small excess of nitric acid in order to
prevent precipitation of copper and iron by hydrolysis. Precipitation
- is achieved by first preparing, from the aforementioned separate solu-
tions, a boiling iron/copper nitrate solution containing 40 grams per
liter Fe and 2 grams per liter Cu and by then adding said solution to
a boiling solution of caustic soda or soda ash (the latter being pre-
ferredl. The resulting suspension is adjusted to a pH value of 7 - 8
;~ within a period of 2 to 4 minutes under intense agitation, in order to
expel carbon dioxide liberated from the solution. The suspension is
then filtered and the solids are washed with co~densate until they are
free of alkali. The solids are then slurried with steam condensate.
To this slurry an amount of potassium waterglass solution is added to
- produce, after this impregnation step, a solid which contains about 25
parts by mass of SiO2 per 100 parts of iron. Since technical potas-
sium waterglass solution usually contains SiO2 and potassium oxide in
a mass ratio of 2,5:1, the product obtained after the above impregna-
tion has too high a potassium oxide content, and the excess must be
- removed. For this purpose an accurately determined amount of nitric
acid is added to the slurry, whereafter the slurry is filtered. The
filter cake so obtained contains lby mass) 25 parts Si~2, S parts
potassium oxide and 5 parts copper per 100 parts Fe. The filter cake
is then pre-dried and extruded before it is finally dried to a resi-
~ dual water content of about 3 mass per cent. The extrudate is there-
- 25 after broken up and graded, the granules of particle size of 2 - 5 mm
being retained for use.
~' ~
The reduction of the catalyst is done batchwise at a temperature of
about 230~C over a period of one hour with hydrogen gas circulated at
~- normal pressure. Of the total iron present, about 25 - 3Q% is reduced
to the metallic form and about 45 - SQ% is reduced to Fe~II), whilst
the balance of the iron remains in the FelIII~ form. The reduced
catalyst is loaded in an atmosphere of inert gas, and for transporta-
' tion ~u~oses is covered with paraffin to protect the catalyst against
oxidation.
. -- .
~ 35 The iron-based catalyst is used, in downflow conflguration, in a tubu-
.,
, 2
~.
lar reactor at 20 to 30 bar pressure at a temperature of
227 - 327~ C, to produce predominantly paraffinic waxes
(i.e. saturated hydrocarbons with a boiling point greater
than 370 ~C) from a synthesis gas with a molecular
hydrogen to carbon monoxide ratio of about 2.
One of the important quality objectives for the waxes
produced in the fixed-bed or low-temperature Fischer-
Tropsch process is that they should have a whiteness on
the Saybolt colour scale of at least 20. As will be
appreciated by those skilled in the art, the higher the
Saybolt number, the whiter the wax. To achieve this
minimum Saybolt number in the final product, it is
normally necessary subsequently to hydrogenate the wax in
the molten state under hydrogen pressure and with an
appropriate catalyst. It would therefore be advantageous
if a high Saybolt number, preferably at least 20, could
be achieved directly in the Fischer-Tropsch reaction,
thus to minimise, or if possible to eliminate, the need
for subsequent hydrogenation.
We have now found that the primary wax colour may be
improved considerably by the use of a catalyst in
accordance with the present invention.
OBJECTS AND SUMMARY OF THE lNV~NllON
It is an object of an aspect of the present invention to
provide a catalyst composition which is suitable for use
in a Fischer-Tropsch process, the catalyst being capable
of causing the production of waxes of improved colour
from synthesis gas.
~ 2~8~
._
3a
It is an object of an aspect of the invention to provide a
method of manufacturing an iron-based catalyst composition
suitable for use in a Fischer-Tropsch process in which
waxes of improved colour are produced from synthesis gas.
In accordance with one aspect of the present invention,
there is provided a catalyst composition used in a low
temperature variation of the fixed bed Fischer-Tropsch
process conducted at a temperature within the range of
227~C to 327~C, comprising iron and from 1 to 20 percent
by mass of activated carbon based on the mass of the iron
in the catalyst, wherein at least 50 percent of the
activated carbon particles are smaller than 100 microns.
.~
Preferably, the catalyst comprises from about 1 to about
80 per cent by mass of activated carbon based on the mass
of the iron in the catalyst. More preferably, the
catalyst comprises from about 1 to about 50 per cent by
mass, and still more preferably, between about 1 and about
20 per cent by mass, of activated carbon, based on the
mass of the iron in the catalyst.
While our results indicated that the improvement of
Saybolt number generally increases the higher the carbon
content of the catalyst, up to about lOOg of C per lOOg of
Fe, other considerations such as space-time yield
determine an optimum carbon content in the range 5 to lOg
of C per lOOg of Fe.
As will be apparent to those skilled in the art, for a
given installation, the incorporation of activated carbon
in an iron-based Fischer-Tropsch catalyst reduces the
available catalytically active iron per unit of reactor
volume. In the design of a new plant, however, allowance
could be made therefor. Selection of the activated carbon
content of the catalyst in accordance with this invention
for an existing reactor will therefore represent a
compromise between the requirement of high Saybolt number
of the wax product and a high throughput rate. We have
found that, for an existing installation, the optimum is
achieved with an activated carbon content in a preferred
range of 5 to lOg of activated carbon per lOOg of Fe in
the catalyst.
Conveniently, the activated carbon is intimately and/or
substantially uniformly mixed with or dispersed in the
iron and other ingredients of the catalyst composition.
We have found, surprisingly, that a bed of activated
carbon arranged downstream of the catalyst bed does not
have the same beneficial effect on the colour of the wax
as when the activated carbon has been mixed with the
catalyst, and particularly not as significantly beneficial
as when the activated carbon has been mixed intimately
with the catalyst.
!~
2028538
'. _
With catalyst compositions containing high concentrations of carbon
and compositions in which the carbon had not been intimately mixed
with the other ingredients, especially where activated carbon
particles are still adhering to the outer surface of the catalyst
pellets, some carbon has been found to be washed through to the wax
knockout pots following the synthesis reactors, during the initial
stages of the runs in which such catalysts were tested. Because acti-
vated carbon particles absorb light, it was found that where they had
been washed through to the wax knoc~-out pots, fluctuating Saybolt
number readings were caused by activated carbon particles presenting
in the wax product. --
In order to achieve the desired physical strength of the catalystgranules, at least 50 per cent, preferably at least 75 per cent, and
more preferably at least 90 per cent, of the activated carbon parti-
cles should be smaller than about 100 micron.
Provided that the particles of the activated carbon are sufficientlyfine, for example at least 90 per cent are smaller than about 45
micron, the Instrom side crushing strength of the catalyst will be
better than the side crushing strength of a conventional catalyst
(i.e. one containing no activated carbon incorporated therein).
In this specification, unless ot~erwise stated, all particle sizes
refer to particle sizes as determined by ASTM sieves.
We have also found that catalyst compositions having side crushing
strengths equal to or better than conventional catalyst can be made
with activated carbon particles falling within the range of abou~ 850
microns to about 1200 microns, whilst activated carbon particles fall-
ing within the range of about 100 microns to about 850 microns and
those exceeding 1200 microns cause the catalyst pellets to have a side
crushing strength which is lower than that of a conventional FB cata-
lyst ~ position. Preferably, for ~ul~oses of catalyst compositionscomprising activated carbon particles falling with~n the range of 850
- to 1200 microns, at least 50 per cent, more preferably at least 75 per
-
s
2028588
_ cent and still more preferably at least 90 per cent of such activated
carbon particles fall within this range.
Surprisingly it has been found that, while some activated carbons give
better performance in terms of improving the Saybolt colour of the wax
product than others, all of the activated carbon products tested,
prepared- from a variety of starting materials, gave improved Saybolt
colours in comparison with conventional iron based Fischer-Tropsch
catalyst (i.e. catalyst into which no activated carbon had been incor-
- porated). Of the activated carbons tested by us, the product desig-
nated Ceca 2S, derived from pine wood, and supplied by the firm Ceca
in France, gave the most favourable performance.
For activated carbon produced from coconut shells, the CSC iodine
number of the activated carbon, at least over the range 600 to 1000
tested, had no significant influence on the Saybolt colour of the wax
product. For activated carbon derived from wattle, the Saybolt colour
of the wax product was found to be slightly lower for CSC iodine
numbers outside the apparent optimum range of 600 to 700. Our experi-
mental data on activated carbons obtained from Ceca and which had been
derived from pine wood showed a Ceca iodine number of about 130 to be
the minimum required for achieving a Saybolt colour of 20. The iodine
~ - numbers of the two former activated carbons were determined by the CSC
Method of Carbon Sales Company Inc, Tulsa, Oklahoma, U S A, whilst the
iodine number of the pine wood was determined according to the Ceca
method of the firm Ceca in France. It is to be expected that the
higher the iodine number, the longer it will take before the ability
of the activated carbon to Pnh~nce the colour of the wax is;depleted.
We have also found that steam activated carbon yields better results
- than acid activated carbon. m e carbon is preferably pre-activated
with steam at about 600~C prior to incorporation in the catalyst. In
situ activation of the car~u.. containing catalyst with steam is not
~ advisable, since hydrothermal sintering of the catalyst is expected to
s take place above about 300~C.
.
~ In accordance with another aspect of the invention, there is provided
.
a method of manufacturing an iron-based catalyst
composition suitable for use in a fixed-bed Fischer-
Tropsch process, comprising the step of adding, at any
stage prior to the extrusion stage, from 1 to 20 per cent
by mass of activated carbon based on the mass of the iron
in the catalyst.
In accordance with yet another aspect of the invention,
there is provided a process for the manufacture of
hydrocarbon waxes of improved whiteness, comprising the
step of contacting synthesis gas at a pressure of from 20
to 30 bar and at a temperature from 227~C to 327~C with
the catalyst composition of the invention.
DESCRIPTION OF THE DRAWING
The attached drawing is a graph, obtained from the results
of a non-limiting example, comparing the deterioration
over time of a conventional fixed-bed low temperature
Fischer-Tropsch catalyst, a catalyst composition
containing 5 grams of activated carbon per 100 grams of
iron and a catalyst composition containing 20 grams of
activated carbon per 100 grams of iron.
EXAMPLE 1
Two catalysts referred to respectively as catalyst A and
catalyst B, were prepared according to the description
given by Frohning. Before the final filtration of
catalyst A, powdered activated carbon, type 2S, derived
from pine wood, and supplied by the firm Ceca in France,
was added to the catalyst in a proportion of 5 grams of
activated carbon per 100 grams of iron, by stirring the
activated carbon into the catalyst slurry for a period of
5 to 6 minutes. A control catalyst B without any
activated carbon was prepared in accordance with the
conventional method.
Both catalyst slurries were filtered, extruded, dried and
broken into granules, and graded, retaining the granules
of particle size in the range 2 to 5 mm. After reduction
in flowing hydrogen, 20 litres of each catalyst were
charged to a low-temperature Fischer-Tropsch pilot plant
reactor and was synthesized from syntheses gas under down-
.,~,~
- 202858~
~_ flow conditions and at the temperatures given above. Table 1 reports
the Saybolt numbers of the product measured at regular intervals:
TABLE 1
Day CatalYst A Catalyst B
3 21 17
; 6 20 11
9 20 14
12 20 16
22 12
18 20 11
21 20 9
24 17 7
27 .16 9
; 15 33 18 4
36 18 3
39 15
42 12 -3
- 45 11 -4
48 15 -5
51 13 -2
g -11
The results in Table 1 show clearly that Catalyst A produced whiter
. wax, and the coloùr deteriorated at a.slower rate than with Catalyst
B, which contained no carbon.
EX~MP~E 2
The performance of Catalyst A packed into a reactdr tube was compared
with the performance of a reactor tube of the same dimensions partly
filled with the same mass of catalyst as in the case of Catalyst B,
except that the tube was provided with a layer of activated carbon
placed in the bottom of the reactor tube immediately below the cata-
.
:
.i
2a2~588
.
lyst bed, the mass of activated carbon being the same as the mass of
the activated carbon premixed with Catalyst A. The average diameter
of the catalyst granules was about 2-5 mm in both cases. The acti-
vated carbon particles packed in the bottom of the reactor tube in the
case of Catalyst B were also of about 2-5 mm average diameter.
Saybolt number determinations of the wax produced by Catalyst A and
Catalyst C are reported in Table 2. ~Catalyst C being Catalyst B
followed by the separate layer of activated carbon).
TABLE 2
lO DaY CatalYst A CatalYst C
'3 21 18
6 20 15
9 20 15
12 20 la
15 15 22 15
18 20 13
21 20 15
24 17 10
27 16 12
20 30 18 11
33 not analysed 5
36 18 6
39 15 5
42 12 ~ -1
25 45 11 -2
48 15 -8
51 13 0
54 9 not analysed
The results in Table 2 show that the separate layer of activated
carbon in the case of catalyst C did not yield the same improvement in
Saybolt colour as did catalyst A, containing uniformly premixed acti-
vated carbon.
.. .
- -~ 028~ 8 g
EXAMPLE 3
With the use of the same procedure as in the case of Catalyst A of
Example 1, a further catalyst sample (Catalyst Dl was prepared which
instead of 5 grams of activated carbon per 100 grams of iron, con-
tained 20 grams of activated carbon per 100 grams of iron. Afterfiltration, extrusion, drying, granulation, grading and hydrogen re-
duction as in the case of Catalyst A, 20 litres of catalyst D (of
granule size 2 - 5 mm) were placed in a low-temperature Fischer-
Tropsch pilot plant reactor, and wax was synthesised from syngas under
the same conditions as for Catalysts A, B and C. Saybolt colour de-
terminations were carried out on the wax produced with this catalyst,
as in the case of Catalysts A, B and C. The results obtained with
Catalysts A, B and D are given in Figure 1, which clearly shows that
Catalyst D yielded greatly improved results compared to Catalysts B
- 15 and A.
I EXAMoeLE 4
Catalyst samples cont~ining 5 grams of Ceca 2S activated carbon per
100 grams of iron were prepared by the same method as for Catalyst A
of Example 1, using activated carbon of different iodine numbers
(determined by a method provided by Carbon Sales Company, Inc., of
~ Tulsa, O~lahoma, U.S.A.). Further sets of catalyst samples were pre-
- pared, using activated carbon samples of varying iodine number but
derived from wattle wood and coconut shells.
TABLE 3
Influence of iodine number of various activated carbons on the Saybolt
- number of the wax produced
Wa = activated carbon produced from wattle wood, (Carbon
Developments IPty) Ltd, Randburg, South Africa).
~-. Co = activated carbon produced from coconut shells (Carbon
Developments (Pty) Ltdl.
~, Pi = activated carbon produced from pine wood (Ceca).
St = standard Fe-based F-T catalyst.
: .
~'
-
.
2 ~
ActivatedIodine Saybolt Activation
carbonNumber Number Method
Wa400 (CSC) 14 Steam
600 ICSC) 20 "
700 ICSC) 21 n
800 (CSC~ 15 ll
- 900 ICsc) 19
1000 (CSC~ 18 ~ " .
,
Co600 (CSC) 21
700 ICSC) 20 n
800 (CSC) 18
900 (CSC) ~ 21
1000 (CSC) 22 ll
Pi100-110 (Ceca) 17 Acid
l 5 110-120 (Ceca) 19 Acid
130 (Ceca) 20 Steam
St - 16
Table 3 reports Saybolt numbers measured after about 500 hours of the
start of a run. The results show that Saybolt numbers are substan-
tially independent of iodine number for the activated carbon derivedfrom coconut shells over the range of iodine numbers used, whereas for
the wattle-wood derived carbon the desired Saybolt numbers were
achieved only for runs in which the iodine number of the activated
carbon was 600 or 700. The results also appear to indicate that, with
. 25 activated carbon derived from pine wood, a minimum iodine number of
130 is required. Furthermore, carbon activated by steam appears to be
marginally better than carbon activated by acid..
~ .
EXAMPLE S
~A number of F-T catalysts incorporating lOg of activated carbon
-30 Iderived from coconut shells, Carbon DeYelopments (Ptyl Ltd~ were
11
202~588
prepared with activated carbon of different particle size ranges using
the method described for Catalyst A in Example 1. These were tested
for Instrom side crushing strength by the ASTM D4179-82 method as well
as for a drop test in which 500 cc of each catalyst was dropped ten
times through a two meter long pipe of 50 m~ diameter, and the percen-
tage of fines passing through a 2 mm ASTM sieve determined thereafter.
The results reported in Table 4 show that up to a particle size of 45-
75 micron, the side crushing strength was unaffected by the incorpora-
tion into the catalyst of activated carbon.
- TABLE 4
Influence of particle size of the activated carbon on the Instrom side
crushing strength of the catalyst.
Particle size Side Crushinq DroP Test
Ranqe Strenqth (kql(% Fines obtainedl
1 - 38 0,60 + 7,7
- 75 0,60 + 6,9
- 106 0,40 + 14,0
300 - 400 0,22 + 65,0
500 - 550 0,34 + 51,0
600 - 850 0,36 + 47,0
850 - 950 0,63 + 30,0
21500 0,20 + 28,0
Conventional 0,60 + 23,0
The comparative side crushing strength for a standard catalyst (i.e.
into which no carbon was incorporatedl was 0,60 kg, whilst the percen-
~ tage fines obtained for the drop test was 23%.
The claims which follow are to be considered an integral part of thedisclosure.
..
12
