Note: Descriptions are shown in the official language in which they were submitted.
2088~02
1 FIELD OF THE INVENTION
2 This invention relates to an improved process for
3 reducing coke formation in hydrocracking of heavy oil, wherein
4~ a mixture of the oil, a solvent for asphaltenes, and an oil-
soluble metal compound, which inhibits coalescence of coke
6 precursors and forms catalytic particles in situ, is heated and
7 mixed at moderate temperature in a pre-treatment and is then
8 introduced into the reactor, wherein hydrocracking is conducted
9 with a prolific hydrogen flow to ensure mixing and efficient
light end stripping.
11 BACKGROUND OF THE INVENTION
12 The present invention was originally developed in
13 connection with hydrocracking of a heavy hydrocarbon feedstock
14 high in content of asphaltenes and sulfur moieties. More
particularly, the feedstock tested was vacuum tower bottoms
16 ("VTB") produced from distillation of bitumen. The invention-is
17 not limited in application to such a feedstock; however, it will
18 be described below with specific respect to it, to highlight the
19 problems that required solution.
Bitumen contains a relatively high proportion of
21 asphaltenes. When the bitumen or its vacuum tower bottoms are
22 hydrocracked, the asphaltenes produce coke precursors, from which
23 adherent solid coke evolves. The coke deposits on and adheres
24 to the surfaces of the reactor and downstream equipment. In
addition, since part of the feedstock is consumed in the
26 production of coke, the conversion of the feedstock to useful
27 products is reduced.
208~402
-
1 The present assignee is an Alberta government research
2 agency which was given a mandate to foster improvements in the
3 upgrading of bitumen and other heavy oils. Realizing the
4 conversion limitation and operating problems that coke deposition
inflicts, it initiated a research project to investigate the
6 mechanisms of coke formation and to look for improvements that
7 might be applied commercially.
8 The present processes were generated as a result of
9 this work. The research involved a progression of concepts and
experimental discoveries that came together to yield a process
11 characterized by a high order of conversion coupled with reduced
12 deposition of adhesive coke and reduced production of coke.
13 Searches and prosecution of the parents of this
14 application have identified the following relevant prior art:
U.S. 4,294,686 (Fisher et al) teaches that, when liquid
16 hydrogen donor oil is used along with hydrogenation in connection
17 with hydrocracking of bitumen vacuum tower residua, coke
18 deposition is allegedly eliminated.
19 However the present assignee and the assignee of the
above cited patent jointly conducted a large scale hydrocracking
21, test on bitumen residue using a liquid hydrogen donor process.
22 This test encountered serious coke production problems. It
23 appears that hydrocracking high asphaltene content feed such as
24 bitumen residue requires more than the presence of liquid
hydrogen donor oil alone.
26 U.S. 4,455,218 (Dymock et al) teaches use of Fe(CO)s as
27 a source of catalyst formed in situ for hydrocracking heavy oil
28 in the presence of H2. The reaction is allegedly characterized
29 by elimination of coking.
-
2088402
1 ~ U.S. 4,485,004 (Fisher et al) teaches hydrocracking heavy
2 oil in the presence of hydrogen, hydrogen donor material, and
3 catalyst comprising particulate Ni or Co on alumina.
4 U.S. 4,134,825 (Bearden et al) teaches forming solid,
non-colloidal catalyst in situ in heavy oil using trace amounts of
6 Fe added in the form of an oil-soluble compound such as iron
7 carbonyl. The metal compound is added to the oil and heated to 325-
8 415C in contact with hydrogen to convert it to a solid, non-
9 colloidal, catalytic form. This catalyst is then used in
hydrocracking the oil and it is stated that coke formation is
11 inhibited.
12 U.S. 4,592,827 (Galliasso et al) teaches injecting an
13 oil-soluble, catalyst precursor Mo compound and water into a heavy
14 oil stream moving to a heater, wherein the mixture is heated to a
temperature of 230C - 420C to effect decomposition of the Mo
16 compound. The heater product is then introduced into a
17 hydrocracking reactor.
18 SUMMARY OF THE INVENTION
19 In one aspect of the research work underlying the present
invention, coke was produced by hydrocracking a mixture of diluent
21 and bitumen vacuum tower bottoms ("VTB") and the coke composition
22 was studied microscopically. It was found that at progressive
23 stages of the evolution of the coke precursors into adherent solid
24 coke, there were present different species of isotropic and
anisotropic submicron and micron-sized spheroids. We have
26 identified these species with the following labels:
2088~02
1 ~ - isotropic sphere;
2 - basic isotropic particle;
3 - isotropic agglomerates;
4 - anisotropic spheres;
- basic anisotropic particles;
6 - anisotropic fine mosaic particles;
7 - anisotropic coarse mosaic particles; and
8 - anisotropic agglomerates.
9 It was further experimentally discovered:
- That the evolution of the coke precursors into coke
11 involved a coalescence process from the minute
12 isotropic species to the larger species; and
13 - That if the coalescence process was inhibited with
14 the major portion of the precursors remaining in
the isotropic and anisotropic agglomerate states,
16 then the deposition of adherent and solid coke was
17 significantly reduced and even virtually
18 eliminated.
19 These observations led to seeking out and identifying
compatible additives that would interfere with the coalescence
21 process and assist in reaching an end where, if any coke was
22 present, it would be present predominantly in the form of
23 agglomerate species, preferably in the isotropic state. It
24 was postulated that a well-dispersed, oil-soluble, metal compound
might be used to react in situ with sulfur moieties of the bitumen
26 VTB to produce colloidal, catalytic particles having wetting
27 characteristics that would enable the colloidal particles
- 2088402
1 to collect at the surfaces of the precursor spheroids and inhibit
2 the spheroids from coalescing. Furthermore, it was postulated
3 that an appropriate diluent might advantageously be used to
4 assist in dispersing this additive and in solubilizing the
5~ processor spheroids.
6 It was experimentally discovered that:
7 - if an oil-soluble Mo, Fe, Ni or Co compound
8 additive, for example iron pentacarbonyl or
9 molybdenum 2-ethyl hexanoate, which was
decomposable at hydrocracking temperature and
11 which was capable of forming particles in situ
12 that were catalytic with respect to hydrocracking,
13 was mixed with heavy oil (and preferably with a
14 diluent) at a moderate elevated temperature, that
was in the range 50 - 300C, preferably 80 - 190C
16 and which was less than the decomposition
17 temperature of the additive, for a period of time
18 sufficient to ensure substantially uniform
19 dispersion of the additive throughout the oil and
association of the additive with the asphaltenes;
21 and
22 - if the resultant mixture was heated to
23 hydrocracking temperature and reacted in a
24 reactor;
then the postulated mechanism appeared to take place.
26 Stated otherwise, inclusion of the additive in the
27 reaction mixture undergoing hydrocracking did have the desired
28 effect of reducing the deposition of adherent solid coke,
29 provided that the additive was well dispersed in the manner
2088402
1 described. Examination of cooled solid samples after
2 hydrocracking showed that the major portion of coke produced
3 under these conditions was in the form of isotropic agglomerates.
4 It is believed that at reactor temperature this coke would have
taken the form of minute spheroids of coke precursor. Chemical
6 analysis of the sample coke indicated that additive metal sulfide
7 was associated therewith in a significant amount and that most
8 of the metal sulphides were colloidal, typically being less than
9 0.1 nanometers in dimension.
In summary, in accordance with the invention an oil-
11 soluble, decomposable metal compound of the type described is
12 firstly well dispersed by mixing, preferably with the aid of a
13 diluent, at moderate elevated temperature (e.g. 100C) in the
14 heavy oil and becomes associated with the asphaltenes. When the
mixture is then subjected to hydrocracking temperature, colloidal
16 metal sulfide particles are produced which are thought to
17 accumulate at the surfaces of or inside spheroids rich in coke
18 precursors and interfere with their coalescence. Upon completion
19 of hydrocracking the coke precursors are found to be largely
transformed into isotropic agglomerates. It is further found
21 that the deposition of adhesive solid coke is significantly
22~ reduced.
23 Subsequent experimental work has shown:
24 - That if the additive is added to the oil at the
reactor inlet or at the pre-heater immediately
26 upstream of the reactor, so that prolonged mixing
27 at a proper moderate temperature is not carried
28 out, then hydrocracking is characterized by coke
29 fouling;
`_ 208840~
1 - That if prolonged mixing is done, but at a
2 temperature that is greater than the decomposition
3 temperature of the additive, then the catalytic
4 particles produced are relatively large (e.g. 5
microns to 4mm) and non-colloidal - in this case,
6 coke fouling occurs;
7 - That if bitumen is the oil used, it usually
8 contains sufficient solvent for asphaltenes, so
9 as not to require the addition of diluent or
solvent; and
11 - That a preferred procedure involves:
12 - mixing the oil, additive, and preferably an
13 asphaltene solvent, at a temperature in the
14 range 80 - 190C, which temperature is less
than the decomposition temperature of the
16 additive, for sufficient time to uniformly
17 disperse the additive,
18 - then digesting the product with mixing at an
19 increased temperature which is greater than
the additive decomposition temperature but
21 less than hydrocracking temperature, to
22 decompose the additive while maintaining it
23 in a well dispersed state; and
24 - then heating the mixture to hydrocracking
temperature and introducing it into the
26 reactor.
20s~so2
1 The test as to whether the dispersion and digestion
2 steps have been properly conducted for sufficient time, with
3 sufficient agitation and at an appropriate temperature is
4 affirmatively answered if the additive is converted into
catalytic metal sulphide particles of colloidal size.
6 When the phrase "decomposition temperature" is used in
7 this specification, it is intended to mean that temperature at
8 which less than about 10% by weight of the additive decomposes
9 during the course of the dispersion step.
Turning now to a second approach that was explored, it
11 was well known that asphaltenes precipitate when pentane is
12 added. Upon considering this known fact, applicants conceived
13 the notion of emphasizing the removal of light ends during
14 hydrocracking to determine the effect on coke formation.
Experimental work was therefore initiated to determine the effect
16 of stripping light ends (Boiling point ("B.P.") <220~C) from the
17 hydrocracking zone. Experimentation showed that coke formation
18 was reduced when light ends were consistently removed during
19 hydrocracking. To improve this, it appeared desirable to apply
mixing to the mixture during hydrocracking. Mixing would have
21 the further attribute of maintaining dispersion of the additive
22 metallic component.
23 To further elaborate on the foregoing, it had been
24 noted that coke formation is associated with phase separation.
It was postulated that, if the coke precursors became richly
26 concentrated in a distinct phase, then the coke formation process
27 would proceed rapidly and quantitatively. To impede this, it
28 appeared desirable to strip light ends and reduce phase
29 separation.
2088402
Therefore, as a second preferred aspect of the invention,
2 a tube reactor is used, preferably substantially free of internals,
3 and the hydrogen flow through the reactor is prolific and is
4 arranged to achieve mixing throughout the length and breadth of the
reaction zone. The prolific hydrogen flow functions to strip light
6 ends from the zone. Preferably, mixing and stripping is
7 accomplished by ensuring that the hydrogen flow is in the range of
8 8,000 - 20,000 SCF/BBL and is sufficient to provide the following
9 Peclet Number ("P.N.") regime in the reactor chamber:
Liquid:
11 axial P.N. = less than 2.0, preferably less than 1.0, most
12 preferably less than about 0.01
13 Gas:
14 axial P.N. = more than 3.0, preferably greater than 5Ø
In another thrust at reducing phase separation, a diluent
16 for solubilizing the asphaltenes was preferably added to the
17 reaction mixture. The diluent (or solvent) was a hydrocarbon
18 fraction having a B.P. of about 220-504C, preferably 220-360C. The
19 solvents used successfully had a high cot~value, as defined in the
paper "Oil Sands Composition and Behaviour" by Jean Bichard, (1987)
21 page 2 - 30 published by Alberta Oil Sands Technology and Research
22 Authority, Edmonton, Alberta, Canada.
23 The preferred diluent contained cyclic moieties that are
24 either aromatic or alicyclic but not aliphatic. For example, n-
hexane was not a good diluent but cyclohexane, decalin and benzene
26 were good diluents, the last being preferred. However, in the
27 hydrocracker, less expensive than these diluents are the 220C to
28 360C heavy aromatic fraction of the hydrocracker gas-
- 2088402
1 oil or the same fraction of coker gas-oil that has not been
2 stabilized.
3 It was hoped that the diluent would in addition
4 function usefully as a liquid hydrogen donor and, in combination
with the produced colloidal metal sulfide (which is catalytic in
6 nature) and the plentiful hydrogen, would create a regime that
7 would be favourable to high conversion of the high boiling (e.g.
8 greater than 504C+) fraction and low coke deposition.
9 Experimental runs indicated that when the combination of diluent
addition, well dispersed additive addition, and light ends
11 stripping with hydrogen was practised in the context of
12 hydrocracking of heavy oil containing asphaltenes and sulfur
13 moieties, exceptionally high conversion of the high boiling
14 hydrocarbons could be achieved, together with virtually no
adhesive coke deposition. When the diluent was omitted from the
16 combination, or the diluent was not a good solvent of asphaltenes
17 or when stripping of light ends was not sufficient, experimental
18 runs showed significant coke deposition. It is to be understood
19 however that diluent addition is only a preferred feature.
In summary then, dispersion is therefore preferably
21 achieved in a distinct step prior to heating to additive
22 decomposition or hydrocracking temperature, by mixing the heavy
23 oil plus additive plus diluent mixture in means such as a
24 continuous flow, stirred tank mixer, the mixture being maintained
at a temperature that is in the range 50 - 300C, preferably 80
26 - 190Cj but less than the temperature at which the additive
27 decomposes significantly, the residence time being sufficient to
28 ensure that the additive is substantially uniformly dispersed
29 throughout the mixture. It is preferable also that two or more
~ 2G88~Q2
continuous flow, stirred tank reactors in series be employed for
2 this mixing.
3 Broadly stated, in one aspect the invention comprises
4 a process for preparing a heavy hydrocarbon feedstock for
hydrocracking, said feedstock containing asphaltenes and sulfur
6 moieties, comprising: mixing the feedstock and an oil-soluble
7 metal compound additive at a temperature that is in the range
8 50C to 300C and less than the decomposition temperature of the
9 additive, to produce a product mixture; said additive being
selected from the group consisting of molybdenum, iron, nickel
11 and cobalt compound additives, said additives being adapted to
12 decompose and react, when heated to hydrocracking temperature,
13 with sulfur moieties in the feedstock to form metal sulfide
14 particles that are catalytic for hydrocracking; said mixing being
conducted for sufficient time to cause the additive to be
16 sufficiently dispersed so that the metal sulfide particles formed
17 upon hydrocracking are colloidal in size.
18 In another broad aspect, the invention comprises a
19 process for hydrocracking a heavy hydrocarbon feedstock
containing asphaltenes and sulfur moieties, comprising: mixing
21 the feedstock and an oil-soluble metal compound additive at a
22 temperature that is in the range 50C to 300C and less than the
23 decomposition temperature of the additive, to produce a product
24 mixture; said additive being selected from the group consisting
of molybdenum, iron, nickel and cobalt compound additives, said
26 additives being adapted to decompose and react, when heated to
27 hydrocracking temperature, with sulfur moieties in the feedstock
28 to form metal sulfide particles that are catalytic for
29 hydrocracking; said mixing being conducted for sufficient time
208~02
1 to cause the additive to be sufficiently dispersed so that the
2 metal sulfide particles formed upon hydrocracking are colloidal
3 in size; then further heating the product mixture to
4 hydrocracking temperature; introducing the heated product mixture
into the chamber of a hydrocracking reactor; temporarily
6 retaining the heated product mixture in the chamber, continuously
7 passing sufficient hydrogen through substantially the breadth and
8 length of the chamber contents to maintain mixing of the chamber
9 contents and stripping of light ends, and removing unreacted
hydrogen and entrained light ends from the chamber and producing
11 pitch containing colloidal metal sulfide.
12 In still another preferred aspect of the invention,
13 pitch is recycled from the downstream hot separator to the
14 reactor, to improve the conversion. In a more preferred aspect,
the separator product, containing heavy distillates and pitch,
16 is distilled to separately recover pitch; in conjunction with
17 this, fresh feed is added to the separator product stream
18 entering the distillation vessel, to reduce the separation of
19 asphaltenes from the pitch. The addition of fresh oil is
operative to reduce or prevent the production of adhesive
21 asphaltene lumps, which would otherwise appear in the
22 distillation vessel.
208~402
1 In a preferred embodiment, the invention involves the
2 following units and conditions in the hydrocracking operation,
3 having reference to Figure 44:
4 - Thermal hydrocracker:
operating temperature - 430 - 460C,
6 preferably 450
7 455C;
8 operating pressure - 1500 - 3000 psig,
9 preferably about
2000 psig;
11 - High pressure hot separator:
12 operating temperature - greater than about
13 350C;
14 operating pressure - reactor pressure;
_ Adding 5 - 15% fresh feed (heavy oil) to the
16 underflow from the hot separator;
17 - Low pressure hot separator:
18 operating temperature - l e s s t h a n
19 temperature of hot
separator;
21 operating pressure - 100 to 500 psig;
22 - Recycling 0 to 95% pitch from the low pressure
23 hot separator to the reactor.
14
~ 20~8~02
1~ DESCRIPTION OF THE DRAWINGS
2 Figure 1 is a photographic representation showing the
3 nature of isotropic sphere(s) and basic isotropic particles (b),4 magnified 1650X;
Figure 2 is a photographic representation showing the
6 nature of anisotropic spheres (s) and basic anisotropic particles
7 (b), magnified 1650X;
8 Figure 3 is a photographic representation showing the
9 nature of isotropic agglomerates (g) along with anisotropic
solids (a) and iron sulfide particles (S), magnified 1650X.
11 Here, the iron sulfide particles originated from the feedstock.
12 The coke sample studied by the microscope was generated from
13 thermal test without any iron additive (see Figure 17);
14 Figure 4 is a photographic representation showing the
nature of anisotropic agglomerates (a), anisotropic fine mosaic
16 (f), and anisotropic coarse mosaic (c), magnified 1650X;
17 Figure 5 is a photographic representation showing
18 anisotropic coke particles having grown via the coalescence of
19 smaller anisotropic spheres (c), magnified 1650X;
Figure 6 is a photographic representation showing
21 isotropic coke particles having grown via the coalescence of
22 smaller isotropic spheres (s), magnified 1650X;
23 Figure 7 is a photographic representation of the
24 reactor baffle after run CF-30 set forth in Example I (Table 2);
Figure 8 is a photographic representation of the
26 reactor baffle after run CF-9 set forth in Example I (Table 2);
27 Figure 9 is a photographic representation of the
28 reactor baffle after run CF-31 set forth in Example I (Table 2);
2~88402
1 Figure 10 is a bar chart setting forth coke composition
2 for runs CF-9, CF-31 and CF-30;
3 Figure 11 is a photographic representation of the
4 reactor baffle after run CF-A3 set forth in Example II (Table 3);
Figure 12 is a bar chart setting forth coke composition
6 for runs CF-A3 and FE-1 set forth in Example III (Table 4);
7 Figure 13 is a photographic representation of the
8 reactor baffle after run FE-1;
9 Figure 14 is a photographic representation of the coke
particles from run FE-1, which were mostly isotropic agglomerates
11 (A) associated with iron sulfides. Isotropic spheres (S) were
12 trapped among the agglomerates;
13 Figure 15 is a photographic representation of the coke
14 particles from run FE-1 showing isotropic spheres (S) which were
effectively prevented from growing into basic isotropic particles
16 by the iron derivative;
17 Figure 16 is a photographic representation of the
18 reactor baffle after run CF-38 set forth in Example IV;
19 Figure 17 is a plot showing nitrogen flowrate versus
coke production for Example V;
21 Figure 18 is a phase diagram for Example V;
22 Figure 19 is a plot showing pressure profiles for runs
23 involving different additives set forth in Example VIII;
24 Figure 20 is a bar plot showing hydrogen consumed for
various runs set forth in Example VIII;
26 Figure 21 is a bar chart setting forth coke composition
27 for a number of the runs set forth in Example VIII;
16
2088402
1 Figure 22 is a photographic representation of coke from
2 run CF-40, showing mostly a continuous sheet of basic isotropic
3 particles (B), magnified 1650X - see Example VIII;
4 Figure 23 is a photographic representation of the
reactor baffle after run CF-40;
6 Figure 23A is a photographic representation of the
7 reactor coil after run CF-41;
8 Figure 24 is a plot derived from Mossbauer spectroscopy
9 analysis of catalyst produced in accordance with the invention -
see Example III (Table 4);
11 Figure 25 is a simplified schematic of a pilot circuit
12 used to carry out the experimental runs reported on in Example
13 IX, with conditions shown thereon;
14 Figure 26 is a plot of pressures recorded between the
reactor and separator during run TRU 101 reported on in Example
16 IX;
17 Figure 27 is a plot of pressure differentials taken
18 across the reactor during run TRU 101 of Example IX;
19 Figure 28 is a plot of pressures recorded at the
entrance to the reactor during run TRU 101 of Example IX;
21 Figure 29 is a plot of pressure differentials taken
22 across the reactor during run B3-1 of Example IX;
23 Figure 30 is a plot of various pressures taken at
24 different points along the circuit during run B3-1 of Example IX;
Figures 31 and 32 are simplified schematics of the
26 segments of the pilot circuit used to carry out the experimental
27 runs reported on in Example X, with conditions shown thereon;
28 Figure 33 is a simplified schematic of the pilot
29 circuit used to carry out the experimental run carried out in
Example XI;
17
2088~02
1 Figure 34 is a plot of pressure logs for the run of
2 Example XI;
3 Figure 35 is a simplified schematic of the pilot
4 circuit used to carry out the experimental runs reported on in
Example XII, with conditions shown thereon;
6 Figure 36 is a plot of differential pressures across
7, the reactor, pressures at the heater, and digester temperature
8 of the circuit used for Example XII;
9 Figures 37(a) to 37(f) is a series of IR spectra
demonstrating the effect of change in temperature in the mixing
11 step for Example XIX;
12 Figure 38 shows asphaltene conversion versus pitch
13 conversion for experiments providing pitch conversions between
14 42 and 99% for Example XVII;
Figure 39 is a simplified schematic of the once-
16 through pilot circuit used in the first stage of run R 2-1,
17 described in Example XX, with conditions shown thereon;
18 Figure 40 is a simplified schematic of a modified form
19 of the pilot circuit of Figure 39, indicating the recycle of
pitch which was practised in the second stage of run R 2-1;
21 Figure 41 is a simplified schematic of a further
22 modified form of the pilot circuit of Figure 40, indicating the
23 recycle of pitch and addition of feed, which was practised in the
24 third stage of run R 2-1;
Figure 42 is a plot of differential pressure across the
26 reactor during run R 2-1;
27 Figure 43 is a plot of pressures taken at different
28 indicated points along the circuit during run R 2-1; and
18
2088~02
1 Figure 44 is a confocal micrograph depicting a particle
2 from run R 2-1 in a stage of fusion or coalescence of the outer
3 components trapping several particles in the central area. Sub-
4 micron size inorganic components of high reflectance are clearly
distinguished in several areas of the particle (Reflected mode,
6 647 nm, oil immersion, 2500X).
7 DESCRIPTION OF THE PREFERRED EMBODIMENT
8 The feedstock to the process is heavy oil. This term
9 is intended to include bitumen, crude oil residues and oils
derived from coal-oil co-processing that contain asphaltenes and
11 sulfur moieties. A typical feedstock could be vacuum tower
12 residues derived from Athabasca bitumen.
13 The feedstock is mixed with a catalyst precursor
14 additive and, preferably, a hydrocarbon solvent for asphaltenes.
The additive is an oil-soluble metal compound adapted
16 to decompose at hydrocracking temperature and to react with
17 sulphur moieties in the oil to form, in situ, metal sulphide
18 particles that are catalytic for hydrocracking and which function
19 to impede coalescence of coke precursors. The metal can be
selected from the group consisting of Fe, Ni, Co and Mo.
21 Preferred compounds are iron pentacarbonyl and molybdenum 2-
22 ethyl hexanoate.
23~ The hydrocarbon solvent for asphaltenes is preferably
24 a recycled stream having a boiling point in the range 220C -
504C, preferably 220 - 360C, and preferably having a high cot
26 e value, as defined in the paper previously mentioned "Oil Sands
27 Composition and Behaviour" by Jean Bichard.
19
2088402
1 ~ The preferred amount of additive added is in the range
2 0.0001 - 5 wt.%, based on the weight of the feedstock. Preferably,
3 we use about 0.002 - 0.5 wt. %. Typically, for the specific
4 preferred compounds we use:
molybdenum 2-ethyl hexanoate - 0.01 wt. %
6 molybdenum naphthanate - 0.007 wt. %
7 iron pentacarbonyl - 0.05 wt. %.
8 With respect to the solvent for asphaltenes, some
9 feedstock (e.g. crude Athabasca bitumen) may already contain
sufficient solvent so as to not require discrete solvent addition.
11 But in the cases where solvent addition is desirable, the preferred
12 weight ratio of solvent to feedstock is in the range 1:10 to 3:1,
13 preferably 1:4 to 1:1.
14 Mixing can be accomplished in a continuous flow, heated,
stirred tank mixer or by pumping the mixture from a tank, through a
16 preheater, and back to the tank. In any event, mixing is conducted
17 in accordance with the following conditions:
18 mixture temperature: within the range 50 - 300C,
19 preferably 80 - 190C, and less
than that temperature at which
21 more than about 10 wt. % of the
22 additive is decomposed during the
23 mixing step;
20~8~02
retention time: sufficient to ensure that the
2 additive is substantially
3 uniformly dispersed throughout
4 the oil and is associated
substantially at the molecular
6 level with asphaltene.
7 The process has become focused on use of iron
8 pentacarbonyl and molybdenum 2-ethyl hexanoate as the preferred
9 additives.
In the case of the iron pentacarbonyl, a relatively
11 large amount of it needs to be used to achieve satisfactory
12 conversions. However, if too much is used, it tends to form iron
13~ products that build up in the piping and result in blockages and
14 pressure surges. To properly use iron pentacarbonyl, we have
15 found it desirable to first well disperse it in the oil at
16 moderate temperature by mixing and then decompose the additive
17 in a higher temperature digestion step, again under mixing
18 conditions to keep the catalyst precursor dispersed.
19 By way of a typical example, for the case of using iron
20 pentacarbonyl as the additive and 504C+ bitumen vacuum tower
21 residuum as the oil, we use the following conditions:
22 additive amount: 250 ppm (based on oil)
23 solvent: 200 - 504C bitumen
24 fraction
solvent/oil ratio: 1:1.2
26 dispersion time: 20 minutes
27 dispersion temperature: 110C
2~88~02
1 dispersion vessel: 1 liter tank with
2 impellor operating at 800
3 rpm
4 digestion time: 60 minutes
digestion temperature: 250C
6 digestion vessel: 3.8 liter tank with
7 impellor operating at
8 1000 rpm.
9 By way of a typical example for the case of using
molybdenum 2-ethyl hexanoate as the additive with 430C+ bitumen
11 vacuum tower residuum as the oil, we use the following
12 conditions:
13 additive amount: lS0 ppm
14 solvent: 430 - 524C fraction
solvent/oil ratio: 1:2
16 dispersion time: 24 hours
17 dispersion temperature: 100C
18 dispersion vessel: 75 litres
19 digestion time: 60 seconds
no digestion
21 The mixture is then rapidly heated to about 450C -
22 455C and introduced into the hydrocracking reactor. In the
23 reactor, hydrogen is supplied at a rate sufficient to satisfy the
24 Peclet No. regime previously described, to ensure that mixing of
the reactor charge occurs and that light ends or volatiles are
26 stripped from the charge.
22
2088~02
1 By way of an example, we have typically used the
2, following conditions in hydrocracking the mixture produced by the
3 pilot plant when using the mixing treatments previously
4 described:
reactor size: 87" long x 1.77" diameter
6 reactor pressure: 1500 psig
7 reactor temperature: 455C
8 H2 rate: 68 l/min.
9 mixture rate: 2405 g/hr.
distillate flow rate: 1082 g/hr.
11 pitch flow rate: 1322 g/hr.
12["Conversion" is determined by calculating:
13100% (524C+ fraction in 524C+ fraction out)
14(524C+ fraction in)
where the 524C+ fraction includes coke but is mineral free.]
16In the case of the Fe(CO)s additive run, pilot plant
17 results based on the typical conditions described showed a
18typical conversion of 90% of the 524C+ fraction. The pitch was
19 analyzed and found to contain colloidal iron sulfide. Coke
production was about 1%.
21In the case of the molybdenum 2-ethyl hexanoate run,
22~ pilot plant results based on the typical conditions described
23showed a typical conversion of 90% of the 524C fraction. The
24 pitch contained colloidal molybdenum sulfide. Coke production
was about 0.3%.
26The invention as described will now be supported by
27 examples and data developed experimentally.
23
- 20~40~
EXAMPLES I - V
2 The following examples I - V are included to illustrate
3 some of the features investigated in the early work underlying
4 the present process.
All the tests in examples I - V were performed in a 1
6 litre, baffled, stirred autoclave. The charge, comprising
7 Athabasca vacuum tower bottoms (504C+) as feedstock, solvent
8 (otherwise referred to as "diluent") and additive (if used), was
9 introduced into the autoclave. The autoclave was sealed, purged
10 free of air, pressurized with nitrogen or hydrogen and heated to
11 430C. The reactor was stirred at 800 rpm, wlth a reaction
12 temperature of 430C and a reaction time of 105 minutes.
13 Properties of the Athabasca vacuum tower bottoms (VTB)
14 are given below.
wt. %
16 C 81.76
17 H 9.51
18 S 6.23
19~ N 0.78
API @ 16C: 2.43
21 IBP 504C.
22 Table 1 herebelow provides the composition (wt. %) of
23 diluents used during the experimental procedures.
24 It is noteworthy that according to the relative content
25 of condensed dicycloparaffins and benzocycloparaffins, diluent
26 B has the most hydrogen donor capability and diluent C has the
27 least.
24
2~8~4~2
1 TABLE 1
2 Diluents
3 Hydrocarbon Type A B C
4 Paraffins 13.02 16.38 13.10
Uncondensed
6 Cycloparaffins 7.32 6.29 5.51
8 Condensed
9 Dicycloparaffins 5.20 13.03 3.80
Condensed
11 Polycycloparaffins 0.49 1.27 0.15
12 Alkylbenzenes 18.07 15.25 11.25
13 Benzocycloparaffins32.29 37.54 20.36
14 Benzodicycloparaffins4.77 3.80 5.53
Naphthalenes 15.86 6.11 19.49
16~ Naphthacycloparaffins1.61 0.26 7.73
17 Fluorenes 0.82 0.00 6.21
18 Phenathrenes/Anthracene 0.61 0.00 6.18
19 EXAMPLE I
20This example illustrates the effect of different
21 diluents. The autoclave was charged with 109 grams of bitumen
22 and 220 grams of diluent A, B or C. A nitrogen overpressure of
23 0.55 MPa was applied and the contents were thermally cracked at
24430C for 105 minutes.
25The results of the tests are shown in Table 2. The
26 reactor was opened and Figures 7, 8 and 9 show the coke deposited
27on the baffles for experiments CF-30, CF-9 and CF-31,
28 respectively.
29It is noteworthy that experiment CF-31 produced as much
coke as experiment CF-9 but that the coke was most easily
31 dislodged from the baffles and reactor surfaces. Moreover,
`_ 208~402
1 although experiment CF-31 produced nearly twice as much coke as
2 experiment CF-30, the coke was most easily dislodged. The
3 surfaces of the reactor and baffles of experiment CF-31 were
4 least fouled.
The coke from the three experiments was examined
6 microscopically and the results are shown in Figure 10. It was
7 noted that when the agglomerate content (which was anisotropic)
8~ was relatively high (Experiment CF-31), the coke deposition and
9 adhesion was least intense in spite of the fact that diluent C
had the least hydrogen donor capability.
11 TABLE 2
12 Test conditions: 430C, 105 min., 800 rpm,0.55 MPa
13 initial N2 pressure
14 Diluent to Vacuum Tower Bottoms ratio is 2:1
Experiment No. CF-30 CF-9 CF-31
16 Diluent type B A C
17 Yield (wt. % vacuum tower bottom corrected for diluent)
18 H2 0.21 0.07 0.08
19 C~ - C4 10.3 10.8 14.8
C5 - 200C 42.3 57.1 55.2
21 200 - 360C -8.6 -37.2 -43.5
22 360 - 504C 22.8 26.5 30.6
23 504C+ (coke free) 26.3 33.8 34.6
24 Coke 4.3 7.7 7.7
Conversion to 504C & coke 73.7 66.2 65.4
26 Mass Balance 98.2 97.3 96.9
` 26
2~8~02
1 EXAMPLE II
2 This example illustrates the effect of hydrogen
3 overpressure.
4 The experimental conditions and results are shown in
Table 3. Experiment CF-A3 is compared with experiment CF-9.
6 Figure 11 shows coke deposition on the baffles for
7 experiment CF-A3. Compared to Experiment CF-9 (Figure 8), the
8 coke yield and deposition of experiment CF-A3 was least.
9 Figure 12 shows results from a microscopic examination
of the coke obtained from experiment CF-A3. It is to be compared
11 with those results shown in Figure 10 for experiment CF-9. The
12 results are similar.
13 In both experiments, over 80% of the coke components
14 were of the anisotropic type. The agglomerate concentration for
experiment CF-A3 was not significantly more than that of
16 experiment CF-9.
17 This example teaches that abundance of hydrogen alone
18 does not neutralize the adhesiveness of the coke precursors nor
19 does it selectively modify the coke composition.
- 2Q88~2
1 TABLE 3
2 CF-9 CF-A3
3 Diluent A A
4 Yield (wt. % VTB), corrected for diluent
H2S 0.07 2.2
6 Cl - C4 10.8 11.6
7 Cs - 200C 57.1 44.0
8 200 - 360C -37.2 -18.0
9 360 - 504C 26.5 30.1
504C+ (coke free) 33.8 27.7
11 Coke 7.7 3.1
12 Conversion to 504C & coke66.2 72.3
13 Mass Balance 97.3 98.5
14 Selectivity to Cs - 504C 77.6
Conditions: CF-9: 430C, 105 min., 800 rpm, 0.55 MPa N2 initial
16 pressure,
17 diluent: VTB ratio 2:1
18 Conditions: CF-A3 430C, 105 min., 800 rpm, 6.8 MPa H2 initial
19 pressure
diluent: VTB ratio 2:1
21 EXAMPLE III
22 This example illustrates that coke containing much
23 agglomerate is not adhesive.
24 The results and conditions of experiments CF-A3 and FE-
1 are shown in Table 4.
26 Figure 13 shows no coke deposited on the baffles for
27 experiment FE-1. Compared to Experiment CF-A3 (Figure 11), the
28 coke yield and deposition of Experiment FE-1 was least.
28
20~8402
_
1 The coke from Experiment FE-1 was observed to be minute
2 particles loosely settled in the bottom of the reactor.
3 TABLE 4
4 Test conditions: 430C, 105 min., diluent/vtb - 2:1,
800 rpm
6 Experiment No. CF-A3 FE-1
7 Diluent A A
8 Gas/Pressure (MPa) H2/6.8 H2/6.8
9 Additive (metal, wt. % VTB) - Fe/0.5
10 Yields on VTB, wt. %, corrected for diluent
11 H2S 1.2
12 Cl - C4 11.6 6.6
13 C5 - 200C 44.0 37.7
14~ 200 - 360C -18.0 -9.1
360 - 504C 30.1 31.2
16 504C+ (coke free) 27.7 31.4
17 Coke 3.1 1.6
18 Figure 12 shows results from a microscopic examination
19 of coke obtained from experiments CF-A3 and FE-1. The results
are very different. The coke from experiment FE-1 is over 80%
21 isotropic agglomerate.
22 Figures 14 and 15 for Experiment FE-1 showed that solid
23 particles were all loosely associated with one another. Coke
24 composition showed that over 97% of the components were of the
isotropic type - see Figure 12. Isotropic agglomerates accounted
26 for 80% of the coke composition.
29
2a8~4~2
This data for experiment FE-1 indicated that the
2 adhesiveness of the coke precursors was effectively neutralized
3 by the highly dispersed iron compound. Where isotropic spheres
4 were concentrated (see Figure 15), the isotropic agglomerates
5 effectively prevented the spheres from coalescing into basic
6 isotropic particles.
7 It is noteworthy also, that additive present as iron
8 sulphide amounts to approximately 1/3 the weight of the coke but
9 is not so evident.
EXAMPLE IV
11 This example further illustrates that the choice of
12 diluent is desirable.
13 The experiment CF-38 was done according to the teaching
14 of U.S. 4,455,218 (Dymock et al). The experimental conditions
were identical to those shown in Table 4 for experiment FE-1.
16 Whole Athabasca bitumen was used instead of Athabasca VTB and no
17 diluent was added. The whole bitumen contained about 60 wt. %
18 hydrocarbon boiling at temperatures greater than 504C. 0.5%
19 (metal) of iron pentacarbonyl was added on the basis of
equivalent 504C+ content in the bitumen.
21 The coke yield was 7.9% (504C+ basis) and this coke
22 adhered very strongly to surfaces of the reactor and baffles.
23 Figure 16 shows the coke deposited on the baffles.
24 EXAMPLE V
This example illustrates the effect of the rates of
26 continuously removing highly volatile components from the
27 reacting fluids.
_ 2Q~40~
1 The one liter autoclave was fitted with a dip tube for
2 sparging N2 or H2 into the reacting liquid, an outlet permitting
3, continuous flow of product gas, and cold trap condensers to
4 remove volatile products from the gas stream before collecting
the latter in a sample bag for analysis. Experiments were
6conducted in the above described reactor at 430C for 105 minutes
7 under 550 kPa pressure both without gas flow and with gas flowing
8 continuously into and out of the reactor. In each experiment,
9110 g of Athabasca 504C+ vacuum tower bottoms (VTB) and 220 g of
a diluent were used. Table 5 presented herebelow gives the
11 reaction conditions and experimental results.
12TABLE 5
13Autoclave Test Results
14 Experiment No. 1 2 3 4 5 6
Diluent Type A A B B B B
16 Nitrogen Flow
17 Rate (1/min) 0.01.89 0.00.21.05 2.16
18 Yield (wt.% VTB):
19 Cl - C4 15.516.5 10.310.08.2 13.8
C5 - 504C 42.338.1 56.556.556.6 51.4
21 504C+ pitch
22 (coke free) 34.638.8 26.327.030.0 29.7
23 Coke 7.93.8 4.33.8 3.4 3.1
24 Condensate
recovered from
26 purge gas (wt. ~
27 VTB) -- 47.5 -- 3.1 23.9 39.7
~ 2088902
1 TABLE 6
2 Simulated Distillation Results of Condensate from Experiment No. 5
3 % Off Temp. C % Off Temp. ~C
4 IBP 34 55 146
57 60 156
6 10 70 65 164
7 15 84 70 176
8 20 94 75 190
9 25 98 80 201
111 85 210
11 35 116 90 219
12 40 123 95 229
13 45 131 FBP 262
14 50 139 -- --
Figure 17 shows the amount of coke produced as a
16 function of the rate of flow of nitrogen. As shown for diluents
17 A and B, the amount of coke produced decreased as the rate of
18 flow of nitrogen was increased. At high rates of flow of
19~ nitrogen, the amount of coke produced for the experiment using
diluent B (the best hydrogen donor solvent) was not very
21 different from that for the experiment using diluent A (the worst
22 hydrogen donor solvent).
23 It is noteworthy in Table 5, that for those conditions
24 providing the least amount of coke, the amount of condensate
recovered from the purge gas was highest. This was true for both
26 diluents A and B.
_ 20~8~02
Table 6 shows results of simulated distillation of the
2 condensate from experiment 5. About 90% of this condensate boils
3 at temperatures less than 220C.
4 This example teaches that coke production is reduced
if the low boiling products are removed continuously (stripped)
6 from the reacting fluids. Moreover, it teaches that coke
7 production is reduced if the low boiling products are removed
8 from the diluent.
9 These observations are consistent with the model that
has asphaltenes separate as another liquid phase from the
11 reacting fluids. In analogy with the common experiment that has
12 pentane added to bitumen to yield solid asphaltene as a
13 precipitate at room temperature, such an experiment done at high
14 temperature is expected to yield asphaltene as a separate liquid
phase. Moreover, it is expected that this separate liquid phase
16 will be rich in the asphaltenes that thermally crack to form
17 coke.
18 This phase separation is shown schematically in Figure
19 18. The three components of this Figure are respectively
labelled asphaltenic, aromatic and paraffinic and alicyclic to
21 represent those fractions having boiling points 504C+, 220
22 504C and 220C, respectively. The arrow indicates the evolution
23 of the composition of whole bitumen as might occur for example
24 IV.
EXAMPLE VI
26 This example illustrates the effect of using hydrogen
27 for continuously removing highly volatile components from the
28 reacting fluids.
2088402
1 A continuous flow system consisting of a preheater, a
2 2-litre stirred reactor and a product collection system was used.
3 The baffles and stirrer were similar to those of the previous
4 examples. A mixture of Athabasca VTB, diluent A and preheated
hydrogen were pumped through the preheater into the bottom of the
6 stirred reactor. Products were removed though a dip tube with
7 its entrance set at 60% of the reactor's height.
8 The experimental conditions and results are shown in
9 Table 7 for experiments 7, 8 and 9. In each experiment the
hydrogen flow rate was 12 slpm. In experiment 7, the liquid
11 hourly space velocity is twice that of experiment 8 and of
12 experiment 9. The temperature of the reacting fluids is 20C
13 higher than that of experiment 8.
14 Noteworthy is that the amount of coke produced in
experiment 8 was less than that of experiment 7 and that almost
16 no coke at all was produced in experiment 9, in spite of the
17 increased severity of hydrocracking from experiment 7 to 8 to 9.
18 Such a result is expected if one considers that the highly
19 volatile fractions of the reacting fluids are removed with
increasing efficiency as conditions are changed from experiment
21 7 to 8 to 9. In experiment 9 the pipe connecting the reactor to
22 the product collection vessel became plugged at the completion
23 of the experiment.
24 In experiment 10, two one-litre reactors were placed
in series with the entrances to the dip tubes adjusted at 50% and
26 70% of reactor height. The conditions and results are shown in
27 Table 7.
34
2088402
1 TABLE 7
2 Continuous Bench Unit Test Results
3' Diluent Type : A
4 Pressure : 10 MPa
Experiment No. 7 8 9 10
6 Reactor 1 (1) 1.2 1.2 1.2 0.5
7 Reactor 2 (1) -- -- -- 0.7
8 Reaction Temperature
9 (C) 440 440 460 440
Liquid Hourly
11 Space Velocity (hr~l) 1 0.5 0.5
12 Hydrogen FLow rate
13 (slpm) 12 12 12 16
14 VTB Concentration
in feed (wt.%) 63.10 65.02 45.53 47.3
16 Conversion (Wt. % VTB
17 to coke and 504C ) 69.0 78.3 98.1 79.8
18 Yield, Wt. % VTB
19 C~ - C4 5.2 7.9 17.3 7.2
C5 - 200C 16.3 19.1 31.4 15.1
21 200C - 360C 23.4 31.1 43.1 24.2
22 360C - 504~C 20.3 18.4 9.3 24.8
23 Coke 4.6 3.1 0.1 4.6
24 504C+ Pitch
(coke free) 31.0 21.7 1.9 24.8
26 Total distillate
27 C5 - 504C 60.0 68.6 83.8 64.7
20~8~02
1 It is noteworthy that the conversion was similar to
2 that of experiment 8. This was expected given the different
3 liquid hourly space velocities and different number of reactors.
4 However, the amount of coke produced in experiment 10 was higher
than that produced in experiment 8 in spite of the higher rate
6 of flow of hydrogen of experiment 10.
7 This example teaches that hydrogen flow and reactor
8 temperature may be used skilfully to remove (strip) low boiling
9 products from the reacting fluids to reduce the amount of coke
that is produced. Moreover it teaches that for one or more
11 hydrocracking reactors in series, a configuration having one
12 reactor only produces the least amount of coke. Moreover, it
13 teaches that if several hydrocracking reactors are placed in
14 series, then least coke is produced if volatile hydrocarbons are
removed from the fluids as they pass from one reactor to the
16 next.
17 EXAMPLE VII
18 This example illustrates that by skilful use of reactor
19 configuration, severity of reaction, stripping of volatile
components and additive, high conversions of VTB to distillate
21 products can be obtained with acceptable production of coke and
22 minimal fouling of the reactor.
23 The continuous flow system of experiments 7, 8 and 9
24 of Example VI was used. The additive was iron pentacarbonyl.
The conditions and results of experiments 11 and 12 are shown in
26 Table 8.
36
20884û2
1 The conditions for experiment 12 were much more severe
2 than those of experiment 11. Nevertheless, all surfaces in the
3 reactor, pipes and collection vessel remained free of fouling by
4 coke and the coke that was produced was a fine friable matter
S that settled in the product collection vessel.
6 The results of a microscopic examination of the coke
7 produced in experiment 12 are shown in Table 9. 74% of the coke
8 was in the form of agglomerates. 23% of the coke was in the form
9 of isotropic spheres but these spheres were isolated and trapped
in a matrix of agglomerates.
11 This example teaches that high conversions with minimal
12 fouling of the reactor may be obtained when the coke that is
13 produced is mostly agglomerates.
37
- 2088402
TABLE 8
2 Experiment No. 11 12
3 Reactor 1 (1) 1.2 1.2
4 Reactor 2 (1) --- ---
Reaction
6 Temperature (C) 450 450
7 Liquid Hourly Space
8 Velocity (hr l) 1.05 0.73
9 Hydrogen Flow rate
(slpm) 8 12
11 VTB Concentration in Feed
12` (wt. %) 47.5 47.8
13 Catalyst (wt. % of Fe
14 based on VTB) 0.5 0.5
15 Conversion (wt. % VTB
16 to coke and 504C ) 67.8 82.1
17 Yield (wt. % VTB)
18 Cl - C4 19. 1 22.9
19 200 - 350C 19.9 27.6
350 - 504C 21.9 21.1
21 Coke 2.4 2.8
22 504C+ Pitch
23 Coke Free 32.2 17.9
24 Total distillate
Cs ~ 504C 60.9 71.6
26 Note that if H2 flow was not increased in experiment 12, one
27 would expect that a 14% increase in conversion should be
28 accompanied by much higher coke yield than the amount recorded.
2088402
1 TABLE 9
2 Coke Composition of Run No. 12
3 Vol %
4 Basic Isotropy 3
Isotropic Spheres 23
6 Isotropic Agglomerates 42
7 Basic Anisotropy 0
8 Anisotropic Fine-Mosaic 0
9 Anisotropic Coarse-Mosaic 0
Anisotropic Spheres 0
11 Anisotropic Agglomerates 32
12 EXAMPLE VIII
13 This example also compares various additives and
14 various metal compounds.
A series of tests using the following additives:
16 - fine, Alberta coal char,
17 - oil soluble nickel naphthanate,
18 - oil soluble cobalt naphthanate, and
19 - oil soluble molybdenum naphthanate
were carried out to compare their relative effectiveness in
21 preventing coke formation and deposition. Iron pentacarbonyl was
22 used as the bench mark for comparison.
23 All tests were performed under common reaction
24 conditions:
0.5 wt. % (metal on vacuum tower bottoms) additive,
26` Athabasca vacuum tower bottoms (33.3%), diluent
27 (66.7%), 6.8 MPa initial hydrogen pressure, 800 rpm
28 stirrer speed, 430C, and 105 minutes reaction time.
39
_ 20884~2
1 In the case of Alberta coal char, the amount added was
2 equivalent to 4% of the vacuum tower bottoms.
3 Pressure profiles presented in Figure 19 and the
4 hydrogen consumption results presented in Figure 20, showed the
following observed order for hydrogen consumption:
6 molybdenum additive (CF-40) 68%
7 nickel additive (CF-41) 39%
8 cobalt additive (CF-41) 27%
9 iron additive (FE-1) 26%
and coal char (CF-43) 21%.
11 Product distributions presented in Table 10 showed the
12 following order of additive for:
13 - vacuum tower bottoms conversion
14 molybdenum > iron > coal > nickel > cobalt
- selectivity to Cs - 504CC
16 nickel > iron > cobalt > molybdenum > coal char
17 - coke formation
18 nickel < cobalt < coal char < iron < molybdenum.
19 Figure 21 shows the effectiveness of the various
additives in converting the coke precursors to form the non-
21 depositing isotropic agglomerate coke particles. Although
22 experiments using additives containing molybdenum consumed the
23 highest amount of hydrogen, over 90% of the coke was basic
24 isotropic particles. In Figure 22, coke from CF-40 appeared as
a continuous sheet of basic isotropic particles. The coke from
26 CF-40 was evidently more densely packed than the coke from FE-
27 1 using the iron additive (Figures 14 - 15).
`- 2088402
1 As pointed out earlier, it was discovered that, to
2 prevent the coke from depositing on the reactor walls, the
3 additive must selectively transform the coke precursor spheres
4 into isotropic agglomerates. The lack of isotropic agglomerates
in coke from experiment CF-40 suggested an explanation for the
6 deposition of adherent coke on the reactor baffles (Figure 23).
7 In contrast, the reactor baffles in experiments with iron
8 pentacarbonyl (Figure 13) did not have any adherent coke.
9 This example teaches that appropriate selection of
additive may inhibit coke production and may inhibit deposition
11 of adherent coke when an appropriate diluent is used. Such an
12 additive will maximize the fraction of coke that is in the form,
13 isotropic agglomerate. Oil soluble additives containing iron or
14 cobalt or nickel or combinations of these are preferred.
41
- 2d8~402
1 TABLE 10
2 Test Conditions: 430C, 6.8 MPa H2 initial pressure, 105
3 min, 800 rpm
4 additive added = metal concentration of
0.5 wt. % VTB
6 Diluent/VTB = 2:1
7 Experiment No. CF-A3 CF-43 FE-1
8 Additive -- Coal char t4%) Fe
9 Diluent type A C A
H2 consumed (wt. %
11' initial H2) 21 21 26
12 Yield, wt. % vacuum
13 tower bottom
14 H2S 2.2 2.3 1.2
C1 ~ C4 11.6 9.8 6.6
16 C5 - 200C 44.0 36.6 37.7
17 200 - 360C -18.0 -7.7 -9.1
18 360 - 504C 30.1 24.4 31.2
19 504C+ (coke free) 27.7 33.0 31.4
Coke 3.1 1.1 1.6
21 Conversion to 504C
22 & coke 72.3 66.2 68.6
23 Selectivity to
24 Cs - 504 C 77.6 80.5 87.0
Mass Balance 98.5 99.0 98.1
42
208B~02
1 TABLE 10 (Continued)
2 Experiment No. CF-41 CF-42 CF-40
3 Additive Ni Co Mo
4 Diluent type C C C
5' H2 consumed (wt. %
6 initial H2) 39 27 68
7 Yield, wt. % vacuum
8 tower bottom
9 H2S 2.0 1.9 3.3
0 Cl - C4 6.1 8.0 8.5
11 Cs - 200C 34.6 36.4 42.4
12 200 - 360C -2.3 -8.9 -11.8
13 360 - 504C 24.9 27.3 29.6
14 504C+ (coke free) 34.8 35.3 28.2
Coke 0.4 0.9 1.8
16 Conversion to 504C
17 & coke 65.6 64.7 71.8
18 Selectivity to
19 Cs - 504C 87.2 84.7 83.8
Mass Balance 99.6 99.5 98.9
21 EXAMPLES IX - XII
22 These examples as a group support the assertions that:
23 1. Additive dispersion needs to be accomplished at
24 less than decomposition temperature and requires
prolonged mixing at moderate elevated temperature
26 to achieve uniform dispersion of the additive
27 through the asphaltenes;
28~ 2. Digestion leading to additive decomposition needs
29 to be accomplished under mixing conditions; and
43
- 20~84~2
1 3. The combination of the described additive
2 selection, preferred use of solvent, dispersion
3 and digestion steps, and stripping and mixing
4 during hydrocracking, come together to create
colloidal catalyst particles which enable high
6 525C+ conversion associated with little adhesive
7 coke formation.
8 Stated otherwise, if the additive is not well
9 distributed at the molecular scale before significant
decomposition occurs, then there is a likelihood that relatively
11 large, non-colloidal, micron or larger sized catalyst particles
12 will be produced, accompanied by adherent coke formation and low
13 conversion. In the same vein, if decomposition of the additive
14 takes place without mixing to maintain dispersion, again non-
colloidal catalyst can be produced and coke formation and low
16 conversion follow.
17 EXAMPLE IX
18 This example (relating to runs TRU 101 and B 3-1) shows
19 the desirability of properly dispersing the additive by mixing
it for a prolonged period at an elevated temperature that is well
21` below the decomposition temperature of the additive; otherwise,
22 when the mixture is subsequently rapidly heated to hydrocracking
23 temperature, severe fouling will occur in the heater or at the
24 reactor inlet and cause plugging, which is characterized by
pressure surges in the circuit.
26 Figure 25 shows the circuit used for these tests.
27 Figures 26 - 28 show the pressure logs taken during run TRU 101
28 at points indicated on Figure 25.
2 0 ~ 0 2
1 A mixing and dispersion vessel ("mixer"~ was provided
2 with a pump and return line, so that the feed could be circulated
3 and mixed. Hydrogen from a source was added to the line taking
4 the product from the mixer. The mixture passed through a heater
to raise its temperature to hydrocracking temperature. The
6 heater product was then introduced into a hydrocracking reactor.
7 The reactor product was passed through a hot separator to produce
8 pitch.
9 Following were the conditions relating to the first run
(TRU 101):
11 (a) Feedstock: Cold Lake crude vacuum bottoms (430C)
12 containing 70% by wt. 525C+ residuum and 300 ppm
13 wt. molybdenum as molybdenum ethyl hexanoate;
14 (b) Dispersion: 24 hours at 135C, later raised to
150C, with mixing and circulation;
16 (c) Hydrogen flow: 14,000 SCF/barrel;
17 (d) Reactor conditions:
18 pressure - 13.6 MPa
19 temperature - 455C.
The mixer was initially operated at 135C for 17.1
21 hours from start. The mixer temperature was then raised to 150C
22 (which was less than the decomposition temperature of the
23 additive). After 25.9 hours from start, a first pressure pulse
24 was observed at PT455, suggesting that minor plugging occurred
downstream at the entrance to the hot separator. After 85.3
26 hours from start, a pressure pulse to 16.3 MPa was observed at
27 PT 320, suggesting that minor plugging occurred between PT 320
28 and PT 340. As the plug freed itself, pressure pulses were
29 observed at DP 450, suggesting that the plug was being pushed
20~02
1 through the reactor and downstream to the hot separator. After
2 99 hours from start, the pressure at PT 400 pulsed to 15.2 MPa,
3 suggesting that a plug had formed at the inlet to the reactor.
4 After 108.9 hours from start, the pressure at PT 400 and upstream
jumped to 21.6 MPa because a strong plug had formed at the
6 reactor inlet.
7 These results bring up the following observations:
8 - That plugging was not a problem when dispersion
9 was conducted at 135C - but it did become a
problem at 150C; and
11~ - That decomposition of the additive was taking
12 place in and adjacent to the heater under non-
13 mixing conditions. This led to the formation of
14 large iron particles that plugged the piping.
The same circuit was later used for run B 3-1.
16 Following were the conditions relating to this run:
17 (a) Feedstock: Cold Lake crude vacuum bottoms
18 containing 100 ppm wt. molybdenum as molybdenum
19 ethyl hexanoate dispersed in 200 - 360C gas-oil;
(b) Dispersion: 24 hours at about 105C with mixing
21 and circulation;
22 (c) Hydrogen flow: 14,000 scf/barrel;
23 (d) Reactor conditions:
24 pressure - 13.6 MPa
temperature - 450C.
46
2~88~02
1 The B 3-1 run was continued for 225 hours. It involved
2 the following changes relative to run TRU 101:
3 - the dispersion temperature was lower;
4 - the concentration of additive was considerably
reduced; and
6 - the reactor temperature was slightly lower.
7 The pressure logs from run B 3-1 are shown in Figures
8 29 - 30.
9 Smooth, plug-free operation was observed, substantially
10 throughout the test. After about 160 hours the pressure upstream
11 of the separator pulsed briefly to about 19.0 MPa as a plug
12 formed and then broke down. Plugging and fouling of unit
13 surfaces were significantly less severe in run B 3-1 than in run
14 TRU 101.
The runs indicate the desirability of dispersing at a
16 temperature that is significantly less than the decomposition
17 temperature and then heating rapidly to hydrocracking
18 temperature.
19 EXAMPLE X
This example shows that if the additive is provided in
21 high concentration in oil and if dispersion is practised at a
22 high temperature that exceeds decomposition temperature, then
23 poor results follow.
24 In this test, dispersion and decomposition were carried
25 out in one step at a first site and the mixture product moved to
26 another site for hydrocracking. A concentrate (4% Fe by wt.) was
27' formed at the first site, to facilitate transportation. The two
28 circuits used are shown in Figures 31 and 32.
47
2088402
1 The conditions of the runs are shown in conjunction
2 with the Figures.
3 The results of making several runs with this system
4 were as follows:
TABLE 11
6 MB 1-4 MB 5-8 10 MB 11-14 MB 15-18
7 CONDITIONS
8 - feed rate, kg/hr3.420 3.272 2.966 3.121
9 - LHSV 0.97 0.93 0.85 0.89
- reactor temperature,
11 C 439 439 451 450
12 - H2 treat gas rate,
13 1/min 40 40 40 60
14 - additive concen-
tration, wt%Fe on
16 524C+ residuum0.49 0.16 0.095 0.12
17 RESU~TS
18 - 525C+ pitch conversion,
19 volume % 57.2 58.1 68.4 67.7
- CCR removal, wt% 24.9 27.4 31.7 33.7
21 - desulfurization, wt% 20.928.7 37.3 35.1
22 Electron microscope analysis of solids from the
23 produced pitch indicated FeSx particles typically having a
24~ diameter of 5~m.
The pitch conversion (57 to 68%) was relatively poor
26 and coke was produced in the reactor circuit.
48
- 2~884~
1 EXAMPLE XI
2 This example is additive to Example X and shows that
3 if a bitumen/additive is only digested at decomposition
4 temperature, without preliminary low temperature mixing, then
poor results follow even if digestion involves mixing.
6 Figure 33 shows the pilot circuit and some conditions
7 used in this experiment. Figure 34 shows the pressure logs from
8 the run.
9 In this test, the following pertained:
- feed: Athabasca bitumen, composition: 45% 220
11 - 524C, 55% 525C+;
12 - feed rate: 2.815 kg/hr.;
13 - additive: iron pentacarbonyl - 33%wt. in
14 llght gas oil;
- additive rate: 32.9 ml/hr.;
16~ - additive concentration: 5000 ppm with respect to
17 525C+ fraction;
18 - hydrogen: 34 standard liters/minute (4000
19 SCF/BBL);
- digester temperature: 250C;
21 - reactor conditions: 450C 10.2 MPa.
22 The estimated pitch conversion was 75%.
23 The pressure drop across the reactor increased slowly
24 during the run and then precipitously after 33 hours. The
circuit became inoperable after 36 hours as the pressure recorded
26 at PT23A increased.
27 Examination of material filtered from the product pitch
28 contained iron sulfide particles sized 1 - 2~ .
49
2088402
EXAMPLE XI I
2 This example shows that if appropriate dispersion is
3 conducted at a mild or moderate temperature that is well below
4 additive decomposition temperature and decomposition is conducted
with mixing, then good conversion and coke reduction results
6 follow.
7 Figure 35 shows the circuit and conditions used for
8 this run. Figure 36 shows various logs from the run.
9 In this test, bitumen and iron pentacarbonyl were mixed
at a temperature of about 100C for about 30 minutes in an
11 impellor-equipped first vessel, to disperse the additive, and
12 then mixed at a temperature of about 250C for about one hour in
13 an impellor-equipped second vessel to decompose the additive
14 while keeping it dispersed.
The following Table 12 sets forth other conditions and
16 the results of the run:
17 TABLE 12
18 Dispersion Vessel Temperature - 100C
19 Digester Temperature - 252C
Fe ppm of 525C+ - 2500
21 Hydrogen Athabasca Reactor Fe(CO)s 525C Days
22 1/min. bitumen TempC LGO/hr conversion
23 kg/hr
24 68.0 2.960 450 11 84 5
68.0 2.960 455 11 87 3
26 55.3 2.405 455 8.9 90 4
27 68 2.960 455 8.9 92 3
2088~02
1 ~ Smooth, plug-free operation was observed for the first
2 100 hours of operation. At that point the pump failed.
3 Following repair, operation of the circuit was fairly smooth,
4 although small fluctuations in pressure drop across the reactor
were recorded. Pitch conversion increased slowly from 84% to 92%
6 over a run duration of about 300 hours.
7 Examination showed the iron of the additive to be
8 present in the pitch in the form of colloidal iron sulphide
9 particles.
EXAMPLE XIII
11 This example provides data showing the extent of
12 decomposition of molybdenum naphthanate ("Mo-naph") and
13 molybdenum ethyl hexanoate ("Mo-HEX") at different temperatures.
14 More particularly, infrared spectra of samples of
bitumen containing either Mo-naph or Mo-HEX were measured over
16 time at temperatures of 130C, 200C and 300C. In the following
17 table, the % decomposition of each of these catalyst precursors
18 is expressed as a fraction (%) of the respective spectral
19 components that had disappeared by a given time.
51
- 20~8~Q2
1~ TABLE 13
2Relative Disappearance of Mo-Carboxylates
3in Feed During Heating
4Precursor
Sampling Mo-Fraction, wt%a Disappearance, %b
6 Temp., C Time, Min. NAPHC HExd NAPHC HExd
7 130 30f 32.1 32.3 0.0 0.0
8 150 35.4 36.7 0.0 9.8
9 1110 29.9 35.9 10.4 32.2
3030 31.3 37.5 13.9 42.4
11 3990 33.2 37.5 15.1 42.6
12 5820 33.2 38.0 20.9 43.8
13 200 5f 30.6 33.1 0.0 0.0
14 35 30.1 36.4 20.4 33.9
32.4 37.0 22.3 44.8
16 150 36.1 - 31.1
17 240 36.6 - 34.5
18 330 35.5 - 35.9
19 960 - 15.7e _ 65.3
300 Of~g 32.3 33.8 0.0 0.0
21 5 30.8 36.5 50.9 94.6
22 10 33.8 33.1 77.2 97.8
23 20 35.0 - 82.5
24 30 - 34.8 - 96.9
35.5 36.7 84.2 96.2
26 70 36.1 38.4 83.5 96.4
27 a GPC, SX-4/CHCl~; void vol. 100 ml; fr. vol. 25 ml
28 b DRIFTS - carboxylate region
29 c _ 43000 ppm Mo/CLVB
d _ 35000 ppm Mo/CLVB
31 e Product: 16.3% insolubles
32 f Reference sample
33 g Sampled at 100C
34 h Sampled at 260C
52
2088402
1 EXAMPLE XIV
2 This example supports the assertion that the catalytic
3 particles produced by the process of the invention are colloidal
4 in size.
A sample of pitch produced in run CFE-1 was examined
6 by X-ray diffraction and Mossbauer spectroscopy.
7 The X-ray diffraction analysis revealed the presence
8 of FeS2.
9 The spectrum from the Mossbauer analysis is shown in
Figure 24. The supporting data are set forth in Table 14 below.
11 Notable in the spectrum is the breadth of each of the peaks.
12 Such breadth is indicative of very fine, colloidal particles,
13 typically less than 10 nanometers in dimension.
14 Prior to this test, microscopic examination of samples
of pitch obtained from experiments done in accordance with the
16 invention showed no evidence of iron sulphide particles, even
1; though chemical analyses typically showed more than 20% by weight
18 iron sulphide in the pitch. This evidence indicated that the
19 catalyst particles were submicroscopic.
2088~02
1 TABLE 14
2 Channel number: 512
3 Folding point: 257.5
4 Geom. Effect: O
5Results of Fit July 8, 1988 17:16:00
6gO46 CFE-1 deposited 5.0 x 0.1 24/6/88
7 Theory: 4
8Number of Parameters: 26 Number of Iterations: 1
9 Chi : 1.9297
~ Initial Final Error Check
11 BASELINE4609769 4609769 25.2337 1.000 1.000
12 Total Area0.0266 0.0266 0.0028 1.733 1.729
13 Mag Field 124.956024.9560 0.0250 0.985 0.998
14 Quad Mag 10.1166 0.1166 0.0061 0.986 0.996
Shift Mag 10.58790.5879 0.0031 0.984 1.024
16 Width Out 10.60000.6000 F I X E D
17 3:2:1 corr1.0000 1.0000 F I X E D
18 WNat Mag 10.6000 0.6000 F I X E D
19 Mag Field 228.660328.6603 0.0050 2.034 2.045
Quad Mag 20.0882 0.0882 0.0072 3.743 3.687
21 Shift Mag 20.59850.5985 0.0220 0.695 0.888
22 Width Out 20.60000.6000 F I X E D
23 3:2:1 corr1.000 1.0000 F I X E D
24 Area Mag 20.3342 0.3342 0.0204 1.353 1.385
WNat Mag 20.60000.60000 F I X E D
26 Mag Field 332.000032.0000 F I X E D
27 Quad Mag 3-0.2200 -0.2200 F I X E D
28 Shift Mag 30.26000.26000 F I X E D
54
2088402
1 TABLE 14 (Continued)
2 Width Out 3 0.7500 0.7500 F I X E D
3 3:2:1 corr 1.0000 1.0000 F I X E D
4 Area Mag 3 0.0000 0.0000 F I X E D
WNat Mag 3 0.2600 0.2600 F I X E D
6. Quad Split 1 0.7555 0.7555 0.0218 0.999 0.991
7 Iso Shift 1 0.2700 0.2700 0.0126 0.748 0.755
8 Width 1 0.6000 0.6000 F I X E D
9 Area 1 0.3624 0.3624 0.0121 1.946 1.964
EXAMPLES XV - XIX
11 These examples are based on experimentation using
12 molybdenum naphthanate as the additive or catalyst precursor.
13 In the experiments, vacuum tower bottoms derived from
14 bitumen were used as the feed. ~he characteristics and
composition of the feed were as follows:
- 2D8`8~02
TABLE 15
2 Distillation Wt. % IBP- 430C
3 IBP - 525C 24.0
4 +525C 76.0
5 Elemental Composition Wt. %
6 Carbon 83.6
7 Hydrogen 9.7
8 Nitrogen 0.8
9 Sulfur 5.9
Oxygen --
11 H/C 1.4
12 TLC/FID Class Composition,
13 Hydrocarbons 75.0
14 Asphaltene (includes Preasphaltene) 25.0
The circuit used for the runs reported was that of
16 Figure 25. 300 ppm of molybdenum, as molybdenum naphthanate, was
17 added to the feed tank. The feed was stirred and pumped around
18 the loop at 200C for 3 hours before the experiment. The tests
19 were 12 to 15 hours duration.
EXAMPLE XV
21 This example shows that high conversion of asphaltenes
22. with minimal production of solid coke was achieved when the
23 invention was practised with molybdenum naphthanate as the
24 additive.
An asphaltene-rich feedstock of Cold Lake vacuum
26 residuum, IBP greater than 430C, was charged to a 0.01m3 surge
27 tank. 300 ppm of molybdenum, as molybdenum naphthanate, was
56
2088~02
added to the tank which was equipped with a stirrer and recycle
2 pump, and mixed therewith under a nitrogen blanket at 200C to
3 form a homogeneous mixture. The mixture was then pumped through
4 the process heater into the reactor. Its temperature was
increased to 455C in the process heater. Hydrogen was admixed
6 with the mixture at the entrance to the process heater. The
7 hydrogen was supplied at a rate of 10,000 - 12,000 SCF/BBL and
8 at a pressure of about 2,000 psig. The process heater consisted
9 of a 2.9 mm I.D. 6100 mm long coil immersed in tin at about the
hydrocracking temperature.
11 The volume of the hydrocracking reactor was 669 cc.
12 It was a stainless steel cylinder 25 mm I.D. and 1370 mm high.
13 The following conditions applied to the reactor
14 operation:
Volumetric flow of H2/liquid = 10,000 SCF/BBL
16 Liquid Peclet No. = about 0.25
17 Gas Peclet No. = about 6
18 (The Peclet Nos. were determined from tracer studies
19 using Xe133 and Il31 )
The LHSV was 0.4 to 1.0 h 1. It usually required 10
21 12 hours for the reactor to reach steady state operating
22 conditions. The hydrocracking took place at a temperature of
23 455C and pressure of 2000 psig. The reactor effluent comprising
24 a mixture of gases and liquids was fed to a hot separator where
gases and liquid were separated.
26 Table 16 provides typical results for the process.
57
2088402
1 TABLE 16
2 Reaction Temperature, C 455 455
3 LHSV, h~1 0.41 1.03
4 Pressure, psig 2000 2000
5 Product Yields wt. % on feed
6 H2S 4.41 3.88
7 C1 ~ C3 8.00 9.01
8 C4 - 195C 20.30 6.88
9 195 - 350C 46.00 39.73
350 - 525C 21.42 35.21
11 +525C 0.11 5.76
12 Coke 0.00 0.86
13 C4 - 525C, vol. % 108.42 96.44
14 Pitch Conversion, wt. % 99.2 91.2
Asphaltene Conversion, wt. % 100.0 84.4
16 HDS, % 82.8 72.7
17 H2 Cons., wt. % of feed 2.5 1.9
18 The above hydrocracking tests were conducted on Cold
19 Lake vacuum bottoms described in Table 15 and the precursor
concentration was 300 ppm Mo on feed. After each test, all units
21 of the experimental circuit were opened, examined and found to
22 be free of coke or other fouling.
23 It will be noted that the Mo run was conducted
24 successfully without solvent, even though VTB's were used as the
feed.
58
- 2088~
1 EXAMPLE XVI
2 This example supports the assertion that the catalyst
3 from Example XV was colloidal.
4 Hydrocracking residuum was dispersed in methylene
chloride and the mixture was injected into a gel permeation
6 column. The molybdenum containing component was found to have
7 an apparent molecular weight range 400 to 3000 with respect to
8 this particular gel permeation column calibrated with respect to
9 polystyrene. This range corresponds to colloidal particles of
diameter greater than 0.002 micron but less than 0.01 microns.
11 EXAMPLE XVII
12 This example shows the effect of preferential
13 association of catalyst precursor with the asphaltenic fraction
14 of bitumen residue feedstock.
Table 17 shows data from two tests, one with catalyst
16 and one without catalyst. These tests demonstrated the
17 differences on asphaltene conversion and coke yield, in
18 particular. Although the pitch conversions for the two
19 experiments were similar, the asphaltene conversions differed by
a factor of 2; the catalyst selectively converted the asphaltene.
59
_ 2Q88qO~
1 TABLE 17
2 No Catalyst 300 ppm Mo
3 Reactor Temperature; C 455 455
4 LHSV; h h 3.63 3.65
Pressure; psig 2500 2500
6 H2 flow rate; scf/bbl 7900 7800
7 Product Yields wt. % on feed
8 H2S 1.94 2.40
9 Cl - C3 2.59 2.22
0 C4 - 195C 5.16 3.55
11 195C - 350C 22.40 20.20
12 350 - 525C 31.78 35.82
13 +525C 36.25 36.09
14 Coke 6.5 0.79
Pitch Conversion, % 52.9 52.6
16 Asphaltene Conversion, % 23.1 58.5
17 HDS, % 31.8 39.3
18 H2 cons., wt. % of feed 0.42 0.91
19Additional evidence of the effect of catalyst precursor
on selective asphaltene conversion and coke suppression is shown
21in Table 18 where the composition of two +525C hydrocracking
22 residua (pitch) are compared.
20~84~2
1 TABLE 18
2 Fraction Pitch I Pitch II
3 Yield % Sulfur % Yield % Sulfur %
4 Maltenes 63.2 3.9 41.5 4.7
Asphaltenes 36.6 5.8 33.4 6.3
6 Preasphaltenes 16.3 6.2
7 Coke 0.2 -- 8.3 6.7
8 Pitch I was derived from a test containing molybdenum
9 naphthanate catalyst precursor. Pitch II was derived from a test
not containing molybdenum naphthanate catalyst precursor.
11 Figure 38 shows that asphaltene conversion was favoured
12 by the presence of the catalyst for a broad range of pitch
13 conversion, 42 to 99%. In the presence of catalyst the process
14 units remained clean and free of coke. In the absence of
catalyst, the process units became fouled by coke.
16 EXAMPLE XVIII
17 This example shows that the process operates
18 successfully over a broad range of concentration of precursor in
19 the bitumen residuum.
61
20~402
1 TABLE 19
2 30 ppm Mo 300 ppm Mo
3 Reaction Temperature 455 455
4 LHSV/ hl 1.03 1.03
Pressure; psig 2000 2000
6 H2 flow rate; scf/bbl 16,400 13,400
7 Product Yields wt. % on feed
8 H2S 2.86 3.88
9 Cl - C3 8.43 9.01
0 C4 - 195C 12.13 6.88
11 195~ - 350C 36.92 39.73
12 350C - 525C 34.37 35.21
13 +525C 5.88 5.76
14~ Coke 0.43 0.86
C4 - 525C 83.42 81.82
16 C4 - 525C; vol. % 100.52 96.44
17 Pitch Conversion, % 91.6 91.2
18 Asphaltene Conversion, % 87.4 84.4
19 HDS, % 53.6 72.7
H2 Cons., wt. % of feed 1.66 1.90
21 EXAMPLE XIX
22 This example shows that the catalyst precursor,
23 molybdenum naphthanate, decomposes at temperatures greater than
24 about 300C in the absence or presence of bitumen residuum.
62
2088~2
1 Figures 37a and 37b show that the catalyst precursor
2 is stable at temperatures less than 250C. Figure 37c shows
3 that the catalyst precursor begins to decompose and polymerize
4 slowly at a temperature of 300C. At higher temperatures the
decomposition was more rapid and coke was produced.
6 Figures 37d and 37e show that the catalyst precursor
7 dissolved in bitumen residuum was stable at temperatures less
8 than 250C. Figure 37f shows that the catalyst precursor
9 dissolved in bitumen began to decompose slowly at temperature of
300C
11 Injection of the catalyst precursor into bitumen
12 residuum at 350C produced coke containing molybdenum.
13 EXAMP~E XX
14 In accordance with a preferred embodiment of the
invention, the heavy distillate and pitch mixture leaving the
16 hot separator (which treats the reactor product) is subjected to
17 distillation, to produce pitch. Part of this pitch is recycled
18 to the reactor. In so doing the following things are
19 accomplished:
(1) a greater rate of stripping of light ends is
21 obtained without increase of hydrogen flux, the
22 light ends having been removed from the recycle
23 stream. This reduces coke formation and
24 consumption of catalyst;
(2) the active catalyst being in its colloidal form
26 in the recycle stream accumulates in the reactor,
27 to provide a higher steady-state concentration
28 therein than would be obtained without recycle.
63
2088~02
1 This reduces catalyst consumption by typically 50%
2 from that obtained without recycle; and
3 (3) residence time of pitch is selectively increased
4 thereby increasing overall liquid yield and
improved stability of operation.
6 In addition, a small amount of fresh feed is added to
7 this recycle stream, thereby accelerating the mixing of the
8 stream and cooling it before it is mixed with the additive-
9 containing feedstock.
This example demonstrates the advantages of practising
11 these preferred features.
12 More particularly, Figures 39 - 41 show the circuits
13 and conditions used in a 3-stage test run (R 2-1) which is now
14 described. The run lasted a total length of 490 hours.
Common conditions of run R 2-1 were as follows:
16 Feed: Cold Lake vacuum bottoms containing 150 ppm
17 molybdenum ethyl hexanoate dispersed in 200 -
18 360C gas-oil;
19 Mixing: circulation and mixing for at least 24 hours
at 105C under an atmosphere nitrogen blanket
21 was practised, before the feed was processed;
22 Hydrocracking conditions:
23 pressure - 13.6 MPa
24 temperature - about 450C
H2 flow - about 15,000 scf/barrels
26 Distillation:
27 conducted in accordance with ASTM D-1160
28 distillation.
64
~0884 0~
1 During the first stage, consisting of 96 hours of
2 operation, the test was conducted on a "once through" basis, i.e.
3 without pitch recycle, as shown in Figure 39. In the second
4 stage, unconverted pitch was recycled back to the reactor to
contribute 15% by weight of the feed. This second stage process
6 is shown in Figure 40 and lasted for 390 hours. Recycling of
7 unconverted pitch improved fresh feed pitch conversion from 90%
8 (in the first stage) to 98% (in the second stage).
9 Compared to run B 3-1 (Example IX), run R 2-1 never
experienced any significant plugging or pressure pulses. This
11 is indicated by the pressure logs set forth in Figures 42 and 43.
12 However, during the distillation of the hot separator
13 product to recover unconverted pitch for recycling, it was noted
14 that significant lumping of pitch (similar to agglomeration)
occurred in the distillation pot. These lumps were hard to break
16 up and they adhered strongly to the distillation vessel.
17 ~he lumps were determined to comprise unconverted
18 asphaltene and molybdenum sulfide formed by the additive.
19 In the third stage of the test, involving the last 150
20. hours of the run, a portion of fresh feed was added to the hot
21 separator product, prior to introducing it to the distillation
22 vessel. This arrangement is shown in Figure 41. Also, in this
23 third stage the molybdenum hexanoate concentration was 150 ppm
24 (metal).
25 ` It was determined that, in the second stage, 3164 grams
26 of hot separator product produced 89.4 grams of lumpy solids and
27 28.7 grams of residue adhered strongly to the distillation pot.
28 This was equivalent to 4.1% of the charge.
2û88~02
1 In the third stage, 3200.1 grams of hot separator
2 product plus 601.5 grams of fresh Cold Lake vacuum bottoms
3 produced no lumps and only 6.3 grams of residue adhered to the
4 distillation pot. This was equivalent to 0.2% of the hot
separator product.
6 In conclusion then, the test showed:
7 - That the conditions of the process yielded 490
8 hours of operation free of plugging and fouling;
9 - That pitch conversion increased significantly with
recycling of unconverted pitch; and
11 - That adding a portion of fresh feed into the
12 distillation unit for pitch separation resulted
13 in reduction of asphaltene separation in the
14 distillation step. In other words, the addition
of some fresh feed to the hot recycle pitch
16 accelerated its dispersion in the feed stream to
17 the reactor.
18 At the completion of the test, the reactor and hot separator were
19 opened and all unit surfaces were observed to be clean and free
of fouling. Liquid collected from the reactor was filtered. The
21 solid material so obtained was a fine dust consisting of
22 microscopic agglomerates. The solid material so obtained is
23 shown in Figure 44.
66
