Oligomerisation of alkenes by radical initiation

Article abstract of DOI:10.1021/op060253y

The use of di-tert-butyl peroxide (DTBP) as initiator for the radical oligomerisation of 1-octene and pentene, typical Fischer-Tropsch-derived products, was studied in the temperature range 100-200°C. Using this approach, the favourable product distributi

Full text of DOI:10.1021/op060253y

Organic Process Research & Development 2007, 11, 286−288
Oligomerisation of Alkenes by Radical Initiation
Michele Cowley*
Fischer-Tropsch Refinery Catalysis, Sasol Technology Research and DeVelopment, P.O. Box 1,
Sasolburg 1947, South Africa
Abstract:
We consider that the addition of a radical initiator could
The use of di-tert-butyl peroxide (DTBP) as initiator for the
radical oligomerisation of 1-octene and pentene, typical Fis-
cher-Tropsch-derived products, was studied in the temperature
range 100-200 °C. Using this approach, the favourable product
distribution of radical oligomerisation (viz. less branched
products in contrast with the catalytic oligomerisation of
alkenes) can be obtained, whilst operating at less severe
conditions than normally required for the radical oligomeri-
sation reaction in thermal oligomerisation. The less branched
products find application as plasticizer and detergent alcohols,
poly(r-olefin) (PAO) lubricants, and high-cetane distillate.
Dimerisation and trimerisation of the alkenes were observed
in the temperature range 100-200 °C at 10-20 bar pressure,
which coincides with the activation temperature range of the
di-tert-butyl peroxide. Although technically feasible, the use of
an initiator to lower the operating temperature of thermal
oligomerisation could not be justified economically.
decrease the severity of the operating conditions required
for thermal oligomerisation. In this study we present our
initial results using pentene and octene as feedstocks with
di-tert-butyl peroxide (DTBP) as radical initiator.
Experimental Section
The oligomerisation reaction was carried out in an
autoclave (1-L 316 SS Parr pressure reactor). The temper-
ature and pressure were adjusted as required (see Supporting
Information, Table S3, for full details), and a stirring speed
of 720 rpm was used to ensure effective mass transfer. Where
appropriate, oxygen was replaced by removing the air under
vacuum followed by the addition of argon gas to the reactor.
This procedure was repeated three times to ensure maximal
removal of oxygen. In a typical experiment, alkene (300 mL;
2.8 mol pentene or 1.9 mol octene) was stirred at 720 rpm
and heated. At the desired temperature, DTBP (10 mL; 0.05
mol) was added, and a sample was taken at the time
indicated.
Analyses. Product quantification was carried out using
an HP 5980 gas chromatograph equipped with an FID, on a
PONA column (dimensions 50 m × 0.2 mm × 0.5 µm) at
Introduction
In previous work, thermal oligomerisation of Fischer-
Tropsch (FT)-derived alkenes was shown to be effective in
providing a diesel product with acceptable cetane number
and an oil product with acceptable viscosity index. The
relative simplicity of the thermal process enabled it to be
1
1
80 kPa pressure, with 0.2 µL sample and a split ratio of
:100. After an initial hold time of 10 min the temperature
was increased from 35 to 290 °C at a heating rate of 4 °C/
min with a final hold time of 30 min.
1
preferred over the catalytic process using BF
Unfortunately, severe reaction conditions had to be employed
i.e., temperatures of 335-385 °C, pressures of 50-180 bar,
3
as catalyst.
Results and Discussion
(
Baseline Experiment. A baseline experiment with octene
as feed was performed without adding initiator (DTBP). The
results are given in Table 1 (experiments 1, 2). The results
in Table 1 indicate that no oligomerisation took place at short
contact times (1 h) in the absence of an initiator. At longer
contact times (16 h), some oligomers were observed but were
negligible in comparison with the results observed in the
presence of initiator, being over 2-3 orders of magnitude
lower. The contribution of nonradical initiator-induced
reactions could therefore be neglected under these reaction
conditions and in the time scale of the experiments.
Evaluation of the Inertness of the Reactor. It is known
that metal surfaces have varying influences on hydrocarbon
reactions. Some metals may promote radical reactions (e.g.,
and residence times of 0.8-3.0 h).
The use of radical initiators for polymerisation of alkenes
has been well established for many years.2 Thermal
treatment does not cause chain branching of alkenes,
resulting in a product that is less branched than that of the
-4
5
3
BF -catalysed process. Products with a low degree of
branching could provide desirable properties for applications
6
7
such as plasticiser and detergent alcohols, synthetic poly-
8
9
(
R-olefin) (PAO) oils, and high-cetane diesel fuel.
*
Corresponding author. Telephone: +27 16 960-3133. Fax: +27 11 522-
4
507. E-mail: michele.cowley@sasol.com.
(
(
1) De Klerk, A. Energy Fuels 2005, 19, 1462.
2) Rust, F. F.; Seubold, F. H.; Vaughan, W. E. J. Am. Chem. Soc. 1948, 70,
5.
9
(3) Whitmore, W. F.; Gerecht, J. F. J. Am. Chem. Soc. 1950, 72, 790.
(4) Garwood, W. E.; Haddonfield, N. J. U.S. Patent 2,937,129, 1960.
(5) Hurd, C. D.; Goodyear, G. H.; Goldsby, A. R. J. Am. Chem. Soc. 1936,
10
autooxidation), whereas others decrease the rate of the
1
1
reaction. The possible effect of the metal on carbon
5
8, 235.
(6) Wilson, A. S. Plasticisers Principles and Practice; The University Press:
Cambridge, 1995; p 129.
(8) Chester, A. W.; Garwood, W. E.; Tabak, S. A. U.S. Patent 4,618,737, 1986.
(9) Claude, M. C.; Martens, J. A. J. Catal. 2000, 190, 39.
(7) Jakobi, G.; L o¨ hr, A. Ullmann’s Encyclopedia of Industrial Chemistry, 5th,
completely rev. ed.; VCH: Weinheim, New York, 1987; Vol. A8, p 339.
(10) Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic
Compounds; Academic Press: New York, 1981.
2
86
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development
10.1021/op060253y CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/21/2007

Table 1. Initial experiments with 1-octene and pentene, at a
temperature of 150 °C and a pressure of 10 bar
Table 2. GC analysis of pentene feed and product in mass %
pentene
(feed)
pentene
temp reaction initiator feed benzeneg productsd
a
component
compounds
3-methyl-1-butene
acetone
(product)
expt # (°C) time (h) (mol) (mol)
(mol)
(%)
C
4
1.6
0.5
0.6
0.0
0.0
0.1
1
2
3
4
5
6
7
8
9
0
1
2
3
150
150
150
150
150
150
150
150
150
150
150
150
150
1
16
1
1
1
3
16
1
16
1
1
0.00
0.00
0.05
0.05
0.05
0.05b
1.0
1.0
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.9
1.0
1.0
1.68
1.68
0.0
0.1
7.1
8.7
7.5
0.0
4.8
10.4
12.1
2.5
2.9
2.0
4.0
c
c
c
isopentane
1-pentene
3.2
5.3
0.7
1.5
e
a
a
2-methyl-1-butene
pentane
trans-2-pentene
2-methyl-2-propanol + cis-2-pentene
2-methyl-2-butene
2-methyl-propanal
1-propanol
2.6
0.8
6.1
23.4
18.2
6.3
0.1
0.2
12.7
45.6
16.4
10.2
0.2
b
0.05
0.05b
0.05b
f
b
1
1
1
1
0.08
0.08b
0.05
0.05
0.1
1
64
1.68
1.68
C
C
5
naphtenes
compounds
0.8
0.3
0.1
0.6
1.2
0.1
6
methylethyl ketone (MEK)
a
2-pentanone
C compounds
7
0.0
0.0
0.1
0.1
In the presence of air. All other experiments were conducted in an argon
b
(
99.999%, Fedgas) atmosphere. DTBP (98%, Sigma-Aldrich). All other experi-
ments were conducted with DTBP (91%, Polifin, now Sasol Polymers). More
detailed analyses of the DTBP are given in the Supporting Information (Table
S1). c Experiments 3-5 were conducted with pentene (Sasol Synfuels, Secunda)
as feed. All other experiments were conducted with 1-octene (98%, Sigma-
Aldrich, by analysis 99.3%) as feed. The feed compositions, as analysed by
GC, are listed in Table 1 (octene) and Table 2 (pentene). A more detailed
composition of the pentene is given in the Supporting Information (Table S2).
C compounds
xylenes
0.0
0.0
1.3
0.6
8
C
C
C
10-dimers
0.0
0.0
0.0
0.0
34.2
0.6
0.0
3.8
100.0
10
H
H
22
11
24
d
undefined
total
All material heavier than the feed was considered product. Only the liquid
e
100.0
products were analysed by GC. Experiment was conducted with a glass liner
in the autoclave. f Ethanol added. Benzene (99%, NT Laboratory Supplies)
g
a
Experiment 19. Reaction conditions were 150 °C, reaction time 16 h, 10
mL of initiator, 10 bar pressure, and 300 mL (2.8 mol) of pentenes as feed, and
air was removed from the reactor.
formation previously noted11-13 was not considered a threat
because the operating temperature was much lower and such
effects were not seen with stainless steel during thermal
cooling the samples following reaction was evaluated, and
the results are reported in Table 1 (experiments 10, 11). The
oligomers obtained were similar when an excess ethanol was
added to terminate the reaction, indicating that no further
reaction took place after the sample had been drawn from
the reactor. Hence, for subsequent experiments, no ethanol
or other reaction-terminating agents were added to samples
taken.
1
oligomerisation. The potential effect of carrying out radical
oligomerisation in a metal autoclave was investigated by
performing some reactions with a glass lining using pentene
as feed. The results are reported in Table 1 (experiments
3-5). It is clear that the results obtained with and without a
glass liner were similar. All subsequent reactions were,
therefore, carried out without a glass liner present in the
autoclave.
Effect of Addition of a Nonreactive Diluent (Benzene).
The effect of the addition of a diluent was investigated with
octene as feed. It is known that the use of a solvent has an
effect on the half-life of peroxides. In nonpolar solvents, such
as benzene, the DTBP half-life is longer than in polar
solvents such as chlorobenzene. From the results in Table 1
it is clear that the addition of benzene (experiments 12, 13
compared with 8) decreases the amount of oligomers formed.
Oligomerisation Reactions of Pentene and Octene.
Additional experiments using DTBP as initiator were then
conducted at other temperatures (see Supporting Information,
Table S3). Typical products resulting from the reaction of
pentene are shown in Table 2 (experiment 19) and for octene
in Table 3 (experiment 43) and discussed below.
Effect of Oxygen. The effect of oxygen was studied using
octene as feed, and the results are presented in Table 1. More
oligomerisation products were obtained in the absence of
air (Table 1, experiments 8 and 9) than in the presence of
air (experiments 6 and 7). Increasing the reaction time from
3
h (experiment 6) to 16 h (experiment 7) resulted, as
expected, in an increase in oligomerisation products but still
less than that obtained in the absence of air at a reaction
time of 1 h. Subsequent experiments were, therefore, carried
out in the absence of air, since the presence of air can lead
14,15
to the formation of partial oxidation products,
in unwanted side products.
resulting
Effect of Ethanol Addition (Determination of the
Extent of Termination of Reaction). The addition of
ethanol to determine the extent of the termination of the
reaction was investigated using octene as feed. The pos-
sibility that the radical reactions were not terminated by
Pentene. The pentene feed and product data in Table 2
show that C10 dimers made up 34.2 mass % of the product
and compounds >C10 made up 3.8 mass % of the product.
Although the feed contained 1.6 mass % C
4
compounds
(
alkanes and alkenes) no C compounds were reported in
4
(11) Dunstan, A. E.; Hague, E. N.; Wheeler, R. V. Ind. Eng. Chem. 1934, 26,
the product. The acetone content decreased from 0.6 to 0.1
mass %. The acetone in the product is much lower than in
the product obtained from octene feed. This is probably
because the pentene run was done at milder conditions (lower
3
07.
(
(
(
(
12) Hurd, C. D.; Eilers, L. K. Ind. Eng. Chem. 1934, 26, 776.
13) Frolich, P. K.; Wiezevich, P. J. Ind. Eng. Chem. 1935, 27, 1055.
14) De Klerk, A. Ind. Eng. Chem. 2003, 42, 6545.
15) Iring, M.; T u¨ d o¨ s, F. Prog. Polym. Sci. 1990, 15, 217.
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
287

Table 3. GC analysis of octene feed and product in mass %
and C16 (1-decene < x <C16), with a carbon chain length
longer than the feed material, also indicates oligomerisation.
octene
(feed)
octene
component
(product)a
8
The formation of C cycloalkanes (1-ethyl-2-methyl-
cyclopentane and 1,2-dimethyl-cyclohexane) can be ex-
plained as follows: Cyclisation only occurred when octene
acetone
0.0
0.0
4.9
11.3
0.1
2
1
1
-methyl-2-propanol
-hexene
0.0
was used as reagent, the linear C
8
olefin being converted to
,3-dimethylcyclopentane
0.0
0.0
0.0
C
5
and C rings with substituents, and was not observed when
6
di-tert-butyl peroxide
0.0
pentene was used as feed. This phenomenon could be due
to the radical attacking itself, which would also explain the
lack thereof in the pentene.
1
4
3
-octene
-octene
-octene
99.3
0.0
53.6
0.1
0.1
0.6
octane
0.3
0.5
trans-2-octene
cis-2-octene
0.1
2.7
radical rearrangement
•
R -CH(OR)-C H
2
8
0.0
1.0
6
1
-ethyl-2-methylcyclopentane
0.0
0.3
•
self-attack, minus OR
•
nonane
0.1
0.0
R-C OR-CH3
8
C
16
C
H
16
32
H
0.0
0.5
1
,2-dimethylcyclohexane or 1,2-methylethylcyclopentane
>
32
0.0
22.4
1.8
1
-decene < x < C16
0.0
This study has demonstrated that oligomerisation of
pentene and octene for applications such as plasticiser and
detergent alcohol precursors and poly(R-olefin) (PAO)
lubricants can be technically feasible. Due to the fact that
free radicals have a low tendency to isomerise, the oligomers
formed via this process can be less branched than in the
catalytic process. Experiments with pentene resulted in higher
alkene conversion (and thus more oligomer product pro-
duced) per mole of decomposed DTBP than with octene.
Without further optimisation studies, 38 mass % of product
could be obtained from pentene at 150 °C and 10 bar, 16 h
reaction time and 10 mL of initiator with a productivity of
total
100.0
100.0
a
Experiment 43. Reaction conditions were 200 °C, reaction time 16 h, 45
mL of initiator, 20 bar pressure, and 300 mL (1.9 mol) octene as feed, and air
was removed from the reactor.
temperature (150 vs 200 °C) and pressure (10 vs 20 bar)
and with a smaller amount of initiator (10 mL for the pentene
experiment vs 45 mL for the octene experiment) while having
the same volume of feed, resulting in less decomposition
product formed. The fraction reported as 2-methyl-2-propanol
+
cis-2-pentene is only slightly higher in the product,
possibly due to the decrease in cis-2-pentene and less
formation of 2-methyl-2-propanol due to the addition of a
smaller volume of initiator compared to the case of the octene
feed. The only aldehyde detected in the feed (2-methyl
propanal), decreased from 0.17 to 0.05 mass %, while the
only alcohol detected in the feed (1-propanol), increased from
20 mol C feed consumed per mole DTBP decomposed. For
5
both pentene and octene feedstocks it was found that there
is a linear dependence between the amount of feed consumed
(and thus oligomer products formed) and the initiator
decomposition.
0
(
0
.07 to 0.21 mass %. The two ketones detected in the feed
MEK and 2-pentanone) increased from 0.06 to 0.13 and
.00 to 0.08 mass %, respectively. MEK, being a minor
Conclusions
It was shown that the severe temperatures of thermal
oligomerisation could be lowered by the addition of a radical
initiator. Pentene and octene can thus be upgraded by radical
oligomerisation to precursors for specific applications. The
oligomerisation rates obtained in these unoptimised condi-
tions were, however, not sufficient to make the process
economically viable. A rough order of magnitude economic
calculation, based on the best results obtained, indicated that
the productivity of the initiator was an order of magnitude
too low to generate a positive cash flow.
product of decomposition of DTBP, indicates some decom-
position of DTBP despite the relatively low acetone content
in the product.
The o-xylene (0.17 mass %) and m,- p-xylenes (0.47 mass
%) detected in the product by GC analysis probably
originated from the wash solvent.
Octene. Limited octene double-bond isomerisation was
observed. This could be due to premature termination of the
octene radicals. Theoretically, at a temperature of 200 °C
(
as calculated by Gibbs-energy minimization with Aspen
Acknowledgment
using the Redlich-Kwong-Soave equation of state), the
ratio of 1-octene/3-octene/4-octene/cis-2-octene/trans-2-
octene should be 1:19.2:9.3:3.2:14.7 at equilibrium.
Although no DTBP was detected in the GC analysis of
feed or product (the injector temperature was higher than
the decomposition temperature of DTBP), one of the main
decomposition products of DTBP, acetone (4.9 mass %) was
detected in the product. An important by-product of initiation
and minor product of DTBP composition, 2-methyl-2-
propanol (11.3 mass %) was also detected in the product.
I thank Sasol Technology Research & Development for
permission to publish this work. The contribution of R.
Ferreira (338/99) is gratefully acknowledged.
Supporting Information Available
GC-FID analysis of the DTBP used (Table S1); more
detailed GC-FID of the pentene feed and product (Table S2);
a table of experimental conditions for the oligomerisation
experiments (Table S3). This material is available free of
charge via the Internet at http://pubs.acs.org.
C
16 alkenes (0.5 mass %) were formed, indicating
dimerisation of the C feed. In the product 22.4 mass %
compounds >C16 32 were found, indicating oligomerisation.
The presence of 1.78 mass % compounds between 1-decene
8
Received for review December 5, 2006.
OP060253Y
H
288
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development