Selective methylative homologation: An alternate route to alkane upgrading

Article abstract of DOI:10.1021/ja803029s

InI₃ catalyzes the reaction of branched alkanes with methanol to produce heavier and more highly branched alkanes, which are more valuable fuels. The reaction of 2,3-dimethylbutane with methanol in the presence of InI₃ at 180-200°C affords the maximally branched C₇ alkane, 2,2,3-trimethylbutane (triptane). With the addition of catalytic amounts of adamantane the selectivity of this transformation can be increased up to 60%. The lighter branched alkanes isobutane and isopentane also react with methanol to generate triptane, while 2-methylpentane is converted into 2,3-dimethylpentane and other more highly branched species. Observations implicate a chain mechanism in which InI₃ activates branched alkanes to produce tertiary carbocations which are in equilibrium with olefins. The latter react with a methylating species generated from methanol and InI ₃ to give the next-higher carbocation, which accepts a hydride from the starting alkane to form the homologated alkane and regenerate the original carbocation. Adamantane functions as a hydride transfer agent and thus helps to minimize competing side reactions, such as isomerization and cracking, that are detrimental to selectivity.

Full text of DOI:10.1021/ja803029s

Published on Web 08/13/2008
Selective Methylative Homologation: An Alternate Route to
Alkane Upgrading
John E. Bercaw,‡ Nilay Hazari,‡ Jay A. Labinger,*,‡ Valerie J. Scott,‡ and
Glenn J. Sunley*,#
Arnold and Mabel Beckman Laboratories of Chemical Synthesis, California Institute of
Technology, Pasadena, California 91125, and BP Chemicals Limited, Hull Research and
Technology Centre, Kingston Upon Hull, North Humberside HU12 8DS, England
Received April 24, 2008; E-mail: jal@its.caltech.edu; glenn.sunley@uk.bp.com
Abstract: InI3 catalyzes the reaction of branched alkanes with methanol to produce heavier and more
highly branched alkanes, which are more valuable fuels. The reaction of 2,3-dimethylbutane with methanol
in the presence of InI3 at 180-200 °C affords the maximally branched C7 alkane, 2,2,3-trimethylbutane
(triptane). With the addition of catalytic amounts of adamantane the selectivity of this transformation can
be increased up to 60%. The lighter branched alkanes isobutane and isopentane also react with methanol
to generate triptane, while 2-methylpentane is converted into 2,3-dimethylpentane and other more highly
branched species. Observations implicate a chain mechanism in which InI3 activates branched alkanes to
produce tertiary carbocations which are in equilibrium with olefins. The latter react with a methylating species
generated from methanol and InI3 to give the next-higher carbocation, which accepts a hydride from the
starting alkane to form the homologated alkane and regenerate the original carbocation. Adamantane
functions as a hydride transfer agent and thus helps to minimize competing side reactions, such as
isomerization and cracking, that are detrimental to selectivity.
Introduction
upgrading light hydrocarbons such as n-hexane to diesel-
range linear alkanes.5,6
The catalytic conversion of abundant but relatively inert
alkanes into higher value chemicals has been a longstanding
challenge for chemists and the petrochemical industry. While
there have been significant advances in our understanding
of C-H activation reactions over the last three decades, they
have not yet led to practical methods for alkane functional-
ization, because of limitations imposed by unfavorable
thermodynamics, low selectivity, and/or economic factors.1
In principle one could avoid these difficult functionalization
problems by simply converting a given alkane into another
alkane of higher value. Well-known examples, which are
extensively utilized by the petrochemical industry, include
isomerization of linear to branched alkanes2 and the alkylation
of isobutane with olefins.3 On a research scale, alkanes may
be coupled by mercury-photosensitized dehydrodimerization,
although selectivity is limited by the radical mechanism
involved.4 More recently the so-called alkane metathesis
reaction has been proposed as a potential method for
The dehydrative condensation of methanol to hydrocarbons
has attracted a good deal of interest over the years. Most often
this involves reaction over shape-selective solid acids, as in the
MTG (methanol-to-gasoline) and MTO (methanol-to-olefins)
processes.7 In contrast, Kim et al. reported that the reaction of
MeOH with zinc iodide at 200 °C led to formation of an alkane-
rich hydrocarbon mixture, with surprising selectivity to one
particular alkane, 2,2,3-trimethylbutane (triptane), in overall
yields of up to 20% (based on moles of carbon), corresponding
to as much as half of the gasoline-range fraction (eq 1).8,9 An
efficient synthetic route to triptane (research octane number )
(5) (a) Vidal, V.; The´olier, A.; Thivolle-Cazat, J.; Basset, J.-M. Science
1997, 276, 99–102. (b) Basset, J.-M.; Cope´ret, C.; Soulivong, D.;
Taoufik, M.; Thivolle-Cazat, J. Angew. Chem., Int. Ed. Engl. 2006,
45, 6082–6085.
(6) Goldman, A. S.; Roy, A. H.; Huang, Z.; Ahuja, R.; Schinski, W.;
Brookhart, M. Science 2006, 312, 257–261.
(7) (a) Chang, C. D. Catal. ReV.sSci. Eng. 1983, 25, 1–118. (b) Olah,
‡ Caltech.
G. A.; Molna´r, A. Hydrocarbon Chemistry, 2nd ed.; John Wiley &
´
# BP.
Sons: Hoboken, NJ, 2003; pp 117-122. (c) Haw, J. F.; Song, W.;
Marcus, D. M.; Nicholas, J. B. Acc. Chem. Res. 2003, 36, 317–326.
(d) Olsbye, U.; Bjørgen, M.; Svelle, S.; Lillerud, K.-P.; Kolboe, S.
Catal. Today 2005, 106, 108–111.
(1) See, for example,(a) Crabtree, R. H. J. Chem. Soc., Dalton Trans.
2001, 17, 2437–2450. (b) Labinger, J. A.; Bercaw, J. E. Nature 2002,
417, 507–514. (c) Goldman, A., Goldberg, K. I., Eds. ActiVation and
Functionalization of C-H Bonds, American Chemical Society: Wash-
ington, DC, 2004. (d) Fekl, U.; Goldberg, K. I. AdV. Inorg. Chem.
2003, 54, 259–320.
(8) Kim, L.; Wald, M. M.; Brandenburger, S. G. J. Org. Chem. 1978, 43,
3432–3433.
(9) The reaction can also be carried out by reacting methyl iodide with
Zn(OMe)2 (Diaconescu, P. L., unpublished results) or ZnO (Wal-
spurger, S.; Prakash, G. K. S.; Olah, G. A. Appl. Catal., A 2008, 336,
48–53) with similar outcomes, indicating that the various species
readily interconvert under reaction conditions.
(2) Ono, Y. Catal. Today 2003, 81, 3–16.
(3) Hommeltoft, S. I. Appl. Catal., A 2001, 221, 421–428.
(4) Brown, S. H.; Crabtree, R. H. J. Am. Chem. Soc. 1989, 111, 2935–
2946.
9
11988 J. AM. CHEM. SOC. 2008, 130, 11988–11995
10.1021/ja803029s CCC: $40.75
2008 American Chemical Society

Selective Methylative Homologation
A R T I C L E S
Scheme 1
112) would provide access to a valuable fuel component and
gasoline additive.
Both the isomerization of and initiation by alkanes implies
that they react with InI3 and thus enter the carbocation/olefin
pool, even if only in very low concentrations. This conclusion
was supported by an isotopic labeling study. Analysis of the
products from a reaction of 13C-MeOH, unlabeled 2,3-dimeth-
ylbutane (as an initiator) and InI3 by GC/MS showed that the
two major triptane isotopologues were singly labeled and fully
labeled (with much more of the latter). Fully labeled triptane
presumably arose via de noVo synthesis from MeOH, but a
singly labeled triptane molecule must have come from the
addition of a single methanol-derived CH2 group to the C6 olefin
generated by the activation of unlabeled 2,3-dimethylbutane.13
It should be noted that this methylatiVe homologation does
not suffer from the hydrogen deficiency of the original methanol-
to-triptane conversion; there is no requirement for the formation
of arenes or any unsaturated hydrocarbons, removing one
limitation on selectivity. If it were possible to convert large
quantities of light alkanes in this manner, using MeOH as a
methylating agent, it would constitute a route for the selective
conversion of relatively low value and abundant alkanes to more
valuable fuels by reactions such as eq 2. Isobutane and
2-methylbutane (isopentane) are produced on a large scale in
refinery operations;14 they are also significant byproducts from
methanol transformations such as MTG.7a While the latter
compound is used directly as a gasoline component, its
homologation to a higher octane, less volatile C6 or C7 branched
alkane would represent a significant upgrading of value. We
report here on work aimed at exploring this unprecedented and
potentially useful approach.
Our mechanistic studies on this complex reaction10 implicate
a carbocation-based route, wherein hydrocarbon growth proceeds
by successive olefin methylation and deprotonation, always
favoring the most highly substituted carbocations and olefins
respectively. Direct C-C bond formation from methanol (and/
or dimethyl ether; partial dehydration of MeOH to DME is rapid
under reaction conditions) apparently takes place only when
solids are present; for a completely homogeneous reaction
mixture an initiator is required, typically an olefin or a higher
alcohol. Alkanes are generated via hydride transfer from an
unsaturated hydrocarbon to a carbocation, with the resulting
multiply unsaturated intermediates eventually resulting in the
formation of arenes, as illustrated in Scheme 1.
According to this scheme, triptane yields are limited by two
constraints. First, the aliphatic pool contains both lighter and
heavier highly branched alkanes (and some olefins), along with
smaller amounts of less branched isomers; the selectivity for
triptane within that pool appears to be governed by the relative
rates of methylation and hydride transfer at the various stages
of product growth. Second, some fraction of the methanol feed
must be diverted to the aromatic pool to satisfy the stoichiometry
in hydrogen. The latter factor accounts for the observation that
the triptane yield is enhanced by additives containing P-H
bonds (phosphorus or hypophosphorous acid), which serve as
alternate hydride sources.11
More recently we have found that InI3 also functions as a
catalyst for this transformation.12,13 While many of the features,
including reaction conditions and typical triptane yields, are quite
similar for ZnI2 and InI3, suggesting that the mechanism for
the conversion of MeOH to triptane is basically the same for
the two systems, there are some significant differences, espe-
cially in the detailed product distributions. Much of the
difference in behavior can be accounted for by the fact that
InI3sunlike ZnI2sis able to activate (some) alkanes easily at
200 °C. In particular, alkanes that contain at least one tertiary
center are readily isomerized by the InI3 system, and are also
able to function as initiators for triptane synthesis.13 In contrast,
olefins (or olefin precursors such as higher alcohols) are required
to initiate ZnI2-catalyzed reactions at e200 °C.10
Results and Discussion
Homologation of 2,3-Dimethylbutane. The first series of
experiments were performed using reaction mixtures of InI3
(4.13 mmol), MeOH (12.4 mmol) and varying amounts of 2,3-
dimethylbutane (DMB), which were heated at 200 °C for 2 h.
The control experiment (with no DMB present) contained
isopropanol as an initiator. The results of analysis by gas
chromatography (GC) are shown in Table 1. In all experiments
conversion of DMB was between 20 and 35%, while all the
MeOH/DME was consumed. Both the yield of triptane (in
milligrams) and the selectivity to triptane (stated as moles of
carbon in triptane per mole of total converted carbon) increase
with the amount of DMB added; the latter parameter is plotted
in Figure 1.15 It is noteworthy that the absolute triptane yield is
increased more than 3-fold by the addition of one molar
equivalent of DMB.
(10) Bercaw, J. E.; Diaconescu, P. L.; Grubbs, R. H.; Kay, R. D.; Kitching,
S.; Labinger, J. A.; Li, X.; Mehrkhodavandi, P.; Morris, G. E.; Sunley,
G. J.; Vagner, P. J. Org. Chem. 2006, 71, 8907–8917.
(11) Bercaw, J. E.; Grubbs, R. H.; Hazari, N.; Labinger, J. A.; Li, X. Chem.
Commun. 2007, 2974–2976.
(14) (a) Grayson, M., Eckroth, D., Eds. Kirk-Othmer Encyclopedia of
Chemical Technology, 3rd ed.; John Wiley & Sons: New York, 1980;
Vol. 12, pp 921-922. (b) See also: Alt, H. G.; Bo¨hmer, I. K. Angew.
Chem., Int. Ed. 2008, 47, 2619–2621.
(12) Kay, R. D.; Morris, G. E.; Sunley, G. J. PCT WO 2005023733, 2005.
(13) Bercaw, J. E.; Diaconescu, P. L.; Grubbs, R. H.; Hazari, N.; Kay,
R. D.; Labinger, J. A.; Mehrkhodavandi, P.; Morris, G. E.; Sunley,
G. J.; Vagner, P. Inorg. Chem. 2007, 26, 11371–11380.
(15) Detailed analytical results are shown in the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008 11989

A R T I C L E S
Bercaw et al.
Table 1. Yields of Triptane and Polymethylbenzenes
(Pentamethylbenzene (PMB) and Hexamethylbenzene (HMB)) As
a Function of 2,3-Dimethylbutane (DMB) Addeda
arise from triptane isotopologues with one and/or two unlabeled
carbons, further demonstrates that some scrambling has oc-
curred.)
DMB triptane
MeOH DMB recovered triptane selectivity PMB HMB
The quantity of aromatic products generated (Table 1)
provides further evidence for conversion of DMB into triptane.
Stoichiometrically, when DMB and MeOH react to form
triptane, the only byproduct is water. In contrast, direct
conversion of MeOH into triptane requires the equivalent of
one molecule of H2 (presumably delivered stepwise, as H+ and
H-), which is provided by formation of hydrogen-deficient
arenes.10 Thus, if some of the triptane being formed comes from
the homologation of MeOH with DMB, a reduction in the
amount of aromatics relative to triptane would be expected,
compared with a reaction in which there is only direct
conversion of MeOH into triptane. In nearly all reactions the
major aromatic compounds observed are pentamethylbenzene
(PMB) and hexamethylbenzene (HMB), so these two species
are used as the indicator for the total amount of aromatics
present.
InI3
(mmol) (mmol) (mmol)
triptane:
(%)
yield (mg)
(%)c
(mg) (mg) (PMB + HMB)
b
4.13 12.4
-
23
32
33
40
60
75
12
14
15
16
17
19
12
1
4
4
6
7
17
9
15
14
9
1.2
1.8
2.5
2.1
3.0
3.3
0
4.13 12.4 0.76
4.13 12.4 1.52
4.13 12.4 3.04
4.13 12.4 6.2
4.13 12.4 12.4
80
65
71
68
79
14
a All reactions were heated for 2 h at 200 °C. b Isopropanol (50 µL,
39 mg, 0.65 mmol) was added as an initiator. c Based on total converted
carbon.
The yields of triptane, PMB, and HMB and the ratio of
triptane to PMB + HMB for a series of experiments with
varying amounts of DMB are shown above in Table 1. Clearly
the ratio of triptane to aromatics increases as the amount of
DMB present increases, supporting the conclusion that triptane
arises from two competitive processes: direct conversion of
MeOH to triptane (which results in aromatic byproducts) and
the homologation of DMB (which does not). Some reduction
of arene yield is observed even at the lowest DMB levels,
indicating some homologation of the alkane to triptane.
Increasing the InI3 loading appears to reduce the contribution
of direct conversion of MeOH to triptane still further. A reaction
with equimolar MeOH and DMB but with twice as much In
(InI3:MeOH:DMB ) 1:1.5:1.5) gives the same overall triptane
yield, but the ratio of triptane to aromatic products (which in
this experiment include some tetramethylbenzenes) is 12.8,
compared to the value of 3.26 for the experiment shown as the
last entry in Table 1 (InI3:MeOH:DMB ) 1:3:3). The analogous
experiment using 13C-labeled MeOH in the same proportions
showed much more singly labeled than fully labeled product:15
the triptane (P-Me)+ signals at 91 and 86 m/z respectively
accounted for 3% and 50% of the total, indicating that under
these conditions homologation of DMB is the predominant route
to triptane. The cause of this effect is unclear; it may be
significant that at the higher indium loading some InI3 remains
undissolved at the reaction temperature, whereas solutions are
completely homogeneous at the lower indium level.
Figure 1. Triptane selectivity against number of mmol 2,3-dimethylbutane.
Because it is conceivable that addition of DMB might increase
the efficiency of the direct conversion of MeOH to triptane,
this trend does not by itself conclusively establish homologation
of DMB to triptane. Two additional lines of evidence do provide
support for homologation. First, a labeling experiment analogous
to that reported previously13 but with substantially more added
DMB was performed. A reaction mixture containing DMB (1.52
mmol), InI3 (4.13 mmol), and 13C-labeled MeOH (12.4 mmol)
was heated at 200 °C for 2 h, and products were analyzed by
GC/MS. Figures 2 and 3 show the MS patterns for the GC
fractions of DMB and triptane, respectively. For the former,
the major set of peaks from m/z ) 71-76 correspond to the
(P-Me)+ fragment ions. Of these, by far the largest is at 71
(12C5H11) and the next-largest at 72 (12C413CH11), with weaker
peaks resulting from mixed isotopologues and the fully labeled
isotopologue 76 (13C5H11). There is also a P+ peak at 86 m/z
for unlabeled DMB; the concentrations of other isotopologues
are too low for their parent ions to be observed. The observed
isotopologue distribution clearly demonstrates that activation
of DMB has taken place, leading to exchange of carbons with
the methanol-derived methylating species, as previously
documented.10,13
Variation of Reaction Conditions. The effects of changing
reaction time and temperature have been investigated using the
standard mixture of InI3 (4.13 mmol), DMB (6.2 mmol), and
MeOH (6.2 mmol). The triptane selectivities for reaction times
ranging from 15 min to 2 h at 200 °C are shown graphically in
Figure 4. For this (and all future) tables selectivity has been
redefined as moles of triptane formed per mole of DMB
convertedsi.e., the putative selectivity of the homologation of
DMB assuming it to be the only triptane-producing processs
which is a more useful figure of merit. The high triptane to
aromatic ratios observed in these reactions indicates that this
assumption is appropriate. Reaction times less than 15 min gave
yields and selectivities similar to those at 15 min, but the DMB
conversion was lower, as shown in Table 3. Even after only 5
min no unreacted MeOH/DME was detected (although it should
be noted that is difficult to detect very small quantities of the
highly volatile DME).
For the triptane fraction, the main signals again correspond
to (P-Me)+ ions; there is barely any detectable signal in the P+
region. The largest signal at 91 m/z is due to fully labeled
13C6H13, while the next largest, at 86 m/z, is due to singly labeled
12C513C1H13; weaker peaks are observed at intermediate values.
Table 2 shows the relative percentage of the different isotopo-
logues, as well as the values that would be expected for complete
statistical scrambling of carbon atoms in the reaction mixture
(in which 59.5% of the carbon atoms are unlabeled and 40.5%
are labeled). It is clear that the fraction of singly labeled triptane,
which would result from homologation of DMB, is far in excess
of the statistical value and conversion of DMB to triptane is
occurring. (The intensity of the 12C113C5 signal, which could
9
11990 J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008

Selective Methylative Homologation
A R T I C L E S
Figure 2. MS of DMB fraction from reaction between InI3, 13C-labeled MeOH and DMB.
Figure 4. Triptane selectivity against time.
Table 3. Effect of Time on Triptane Yield and Selectivitya
time
(min)
DMB
recovered (%)
triptane
yield (mmol)
triptane
selectivity (%)b
triptane:
(PMB + HMB)
5
10
15
30
60
73
70
67
66
58
56
0.51
0.56
0.62
0.64
0.62
0.59
30
31
31
30
24
23
>100
40
28
28
c
29
19
c
120
a Reactions were performed at 200 °C and used InI3 (4.13 mmol),
MeOH (6.2 mmol), and DMB (6.2 mmol). b Percentage of converted
DMB which becomes triptane. c Some tetramethylbenzene (TMB) was
observed in these reactions and has been included so the quoted figure is
the ratio of triptane to (TMB + PMB + HMB).
Figure 3. MS of triptane fraction from reaction between InI3, 13C-labeled
MeOH and DMB.
Table 2. Relative Statistical and Observed Percentages of
Isotopologues of (P-Me)+ for Triptane
of triptane produced. This is a consequence of side reactions of
DMB. We have previously demonstrated that both isomerization
into other C6 alkanes and cracking into lighter alkanes such as
isobutane and isopentane are catalyzed by InI3, and further that
the presence of MeOH slows down InI3-catalyzed alkane
isomerization and cracking.13 Because MeOH undergoes pro-
cesses other than providing methylating equivalents for DMB
conversion, such as direct conversion of MeOH to triptane and
the methylation of aromatic species, MeOH/DME is completely
consumed long before DMB (as noted above), so that no further
homologation can occur, but the isomerization and cracking
processes continue. (Triptane isomerization and cracking also
take place, but at a slower rate, as previously demonstrated.13)
molecular formula
statistical distribution (%)
observed distribution (%)
12C6H13
4.43
18.12
30.84
27.98
14.28
3.89
9.5
31.6
2.6
3.5
5.2
12C513C1H13
12C413C2H13
12C313C3H13
12C213C4H13
12C113C5H13
13C6H13
15.5
32.5
0.44
The decrease in triptane selectivity beyond 15 min corre-
sponds to a decrease in the recovered DMB at longer reaction
times, along with a much smaller decrease in the total quantity
9
J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008 11991

A R T I C L E S
Bercaw et al.
Table 5. Ratio of Triptane to (PMB + HMB) as Concentration of
DMB Increasesa
InI3
(mmol)
MeOH
(mmol)
DMB
(mmol)
DMB
recovered (%)
triptane
selectivity (%)b
triptane:
(PMB + HMB)
4.13
4.13
4.13
2.07
2.07
6.2
6.2
6.2
3.1
3.1
6.2
9.3
12.4
7.75
9.3
63
61
75
70
74
38
29
27
25
25
48
35
25
42
c
-
a Reactions performed for 30 min at 180 °C. b Percentage of
converted DMB which becomes triptane. c No PMB or HMB was
detected.
Table 6. Comparison of Triptane Yield and Selectivity for
Figure 5. Combined yield of 2-methyl and 3-methylpentane against time.
Table 4. Effect of Temperature on Triptane Yield and Selectivitya
Homologationa in the Presence and Absence of Adamantane
adamantaneb
(mg)
DMB recovered (%)
triptane yield (mmol)
triptane selectivity (%)c
temp
(°C)
DMB
recovered (%)
triptane
yield (mmol)
triptane
selectivity (%)b
triptane:
(PMB + HMB)
0
63
30
51
44
0.90
2.32
2.10
2.05
39
55
65
59
10 (1.2)
50 (5.9)
100 (12)
c
200
190
180
170
160
150
140
66
60
63
61
78
>90
>90
0.64
0.79
0.90
0.80
0.18
0.01
0
30
32
38
33
13
1
13
26
48
a Reactions were performed at 180 °C for 30 min and used 4.13
mmol of InI3, 6.2 mmol of DMB, and 6.2 mmol of MeOH. b Number in
parentheses is mol % of adamantane relative to DMB. c Percentage of
converted DMB which becomes triptane.
>100
d
nr
d
nr
d
0
nr
a Reactions were performed for 30 min and used InI3 (4.13 mmol),
MeOH (6.2 mmol), and DMB (6.2 mmol). b Percentage of converted
DMB which becomes triptane. c Some tetramethylbenzene (TMB) was
observed in this reaction and has been included so the quoted figure is
the ratio of triptane to (TMB + PMB + HMB). d not recorded.
Table 7. Comparison of Ratios of Side Products to Starting
Material and Triptane for Homologation in the Presence and
Absence of Adamantanea
ratio of hydrocarbons
no adamantane with adamantaneb
DMB: other C6 alkanes
Triptane: other C7 alkanes
(DMB + triptane): (isobutane + isopentane)
40:1
10:1
7:1
56:1
17:1
35:1
The combined yield of 2-methylpentane and 3-methylpentane,
two of the major isomerization byproducts, as a function of time
is shown graphically in Figure 5.
a Same reaction conditions as Table 5. b 50 mg of adamantane added.
The effect of temperature on the homologation of MeOH with
DMB is shown in Table 4. As the temperature is decreased from
200 °C, the extent of isomerization and cracking of DMB and
triptane decreases, which results in an increase in the triptane
yield and selectivity. At these temperatures the rate of activation
of the alkane and methylation of the olefin is sufficiently fast
to allow significant amounts of conversion (approximately 40%
conversion of DMB), but below 170 °C, conversion slows
dramatically and eventually stops altogether. Furthermore, at
and below 180 °C almost all of the triptane is formed by
homologation of DMB with little or no direct conversion of
MeOH to triptane, as shown by the high triptane:arene ratios;
a labeling experiment at 180 °C using 13C-MeOH and unlabeled
DMB confirms this, as virtually no signal arising from fully
labeled triptane is detectable.15 The maximum selectivity for
DMB homologation to triptane, 38%, was achieved at 180 °C.
In an effort to further suppress the direct conversion of MeOH
to triptane, several experiments were performed using higher
ratios of DMB to MeOH. The results (Table 5) show that this
does not significantly affect the triptane:aromatic ratio, but does
decrease the triptane selectivity, so that the optimum ratio of
MeOH:DMB appears to be 1:1.
The catalyst loading cannot be lowered substantially; for
starting ratios of MeOH:InI3 above 4:1 the yield of triptane drops
off sharply, regardless of the amount of DMB. This inhibition
is most probably a water effect (since, as noted above, partial
dehydration of MeOH to DME is rapid at any concentration),
as observed in the direct conversion of MeOH to triptane, where
high effective turnover numbers were achieved in cyclic mode
by removing water (use of DME instead of MeOH as feedstock
for methylative homologation gives essentially identical
results15).10,13 The same approach works for homologation: at
the end of a reaction all volatiles (including water) were removed
and the remaining InI3 dried under vacuum at 60 °C; a fresh
charge of DMB and MeOH was added to the reaction mixture,
and the reaction performed again, with no decrease in the
triptane yield or selectivity.15 The reaction can be thus continued
for a number of cycles if the volatiles are periodically removed
in this manner.
Effect of Added Adamantane on Homologation. Adamantane
has been shown to suppress cracking and thus enhance alkane
isomerization reactions that proceed through a mechanism
involving carbocations, presumably by acting as a hydride
transfer catalyst.16 Because cracking represents a major side
reaction in the alkane homologation reactions described above,
we examined the effect of adding a small amount of adamantane
to a reaction between DMB and MeOH catalyzed by InI3. Table
6 compares the triptane yield and selectivity in the presence
and absence of adamantane; there is a dramatic beneficial effect
on both, along with a significant increase in the amount of DMB
converted. The increase in selectivity occurs because adaman-
tane suppresses the isomerization of both DMB and triptane
and also greatly reduces cracking side reactions. Table 7 shows
that adamantane suppresses isomerization (of both C6 and C7
alkanes) as well as cracking (which leads to isobutane and
isopentane, among other products) competing with DMB
homologation. The effect involves a catalytic action of ada-
mantane, as 90-100% of added adamantane is recovered (by
(16) Iglesia, E.; Soled, S. L.; Kramer, G. M. J. Catal. 1993, 144, 238–
253.
9
11992 J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008

Selective Methylative Homologation
A R T I C L E S
a
Table 8. Summary of Homologation Reactions between MeOH and Isopentane Catalyzed by InI3
InI3 (mmol) MeOH (mmol) isopentane (mmol) time (min) temperature (°C) isopentane recovered (%) triptane yield (mmol) triptane selectivity (%)b triptane + DMB selectivity (%)c
4.13
4.13
4.13
6.2
12.4d
18.6
6.2
6.2
6.2
60
300
750
180
200
200
52
56
76
0.60
0.70
0.36
23
26
25
41
45
40
a All reactions contained 0.367 mmol of adamantane. b Percentage of converted isopentane which becomes triptane. c Percentage of converted
isopentane which becomes either DMB or triptane. d A control reaction using InI3 (4.13 mmol), MeOH (12.4 mmol), adamantane (0.367 mmol), and
isopropanol as an initiator, but no isopentane, produced only around 5 mg of triptane.
a
Table 9. Summary of Homologation Reactions between MeOH and Isobutane Catalyzed by InI3
isobutane
(mmol)
isobutane
recovered (%)
triptane yield
(mmol)
triptane
selectivity (%)b
triptane + DMB
triptane + DMB +
InI3 (mmol)
MeOH (mmol)
time (min)
selectivity (%)c
isopentane selectivity (%)d
3.49
3.49
3.49
5.2
10.5
15.8
5.2
5.2
5.2
60
300
720
48
53
56
0.09
0.15
0.17
4
6
7
5
8
10
22
36
42
e
a All reactions were performed at 200 °C and contained 0.367 mmol of adamantane. b Percentage of converted isobutane which becomes triptane.
c Percentage of converted isobutane which becomes either DMB or triptane. d Percentage of converted isobutane which becomes either isopentane,
DMB, or triptane. e This reaction also produced 10 mg of PMB and 7 mg of HMB, indicating that some direct conversion of MeOH to triptane was
occurring.
GC) in all cases. Varying the loading of adamantane between
1-12 mol % (relative to DMB) had little effect on triptane yield
or selectivity.
previous results which indicate that as the amount of MeOH
present is increased, the activation of alkanes becomes slower,
as well as the observation that isopentane is activated more
slowly than DMB.13 No aromatic species are detected by GC
in these experiments, which is indicative of almost complete
suppression of the direct MeOH to triptane pathway; but MeOH
is completely converted in all experiments, which is why the
triptane yield decreases while the triptane selectivity remains
almost constant (using the definitions given earlier) as the
amount of MeOH is increased. Clearly, as in DMB homologa-
tion, there are side reactions that use up MeOH without
involving isopentane.
The increased DMB conversion does not appear to be a
consequence of adamantane increasing the rate of homologation;
rather, it suppresses side reactions that use up the methylating
agent. As noted earlier, although the synthesis of triptane
requires only one equivalent of MeOH per DMB, in nearly all
experiments (with or without adamantane) using either a 1:1 or
1:2 ratio of DMB to MeOH, the MeOH is completely consumed,
while only a fraction of the DMB has been converted.15 At very
short times, reactions can be stopped and analyzed before all
the MeOH was consumed; the results15 demonstrate that MeOH
consumption is slowed by the addition of adamantane. This
suggests that the increased conversion in the presence of
adamantane is achieved by reducing the quantity of MeOH
“wasted” by side reactions, such as the homologation of alkanes
which are not on the pathway to triptane, and thus allowing the
MeOH to be utilized more efficiently for homologating DMB.
Furthermore, it has been demonstrated that addition of ada-
mantane greatly suppresses the direct conversion of MeOH to
triptane,17 so under these conditions methylative homologation
becomes even more favored.
Homologation of Lighter and Less Branched Alkanes with
MeOH. The upgrading of isobutane and isopentane, which are
available in much larger quantities, would be significantly more
practical than the upgrading of DMB. The results of a series of
reactions of isopentane with MeOH in the presence of InI3 and
adamantane are summarized in Table 8. Homologation of
isopentane to DMB and triptane is observed, along with products
resulting from isomerization and cracking. The ratios of DMB
to other C6 alkanes and triptane to other C7 alkanes are around
2.5:1 and 5:1 respectively, while the ratio of DMB + triptane
to isobutane is approximately 5.5:1, depending on the exact
reaction conditions. If adamantane is not present, more isomer-
ization and cracking occurs.15
In contrast to DMB homologation, there is some consumption
of adamantane in these reactions: only about 50% of the starting
adamantane is recovered (by GC) at the end, probably because
of the somewhat more stringent conditions (higher temperature
and/or long reaction time) required. Also, the mass balance
achieved in these reactions is not as good as that for DMB
homologation: only around 60% of the total carbon present at
the start of the reaction is accounted for at the end. This is due
in part to the volatility of isopentane.
Similarly, isobutane is homologated to isopentane, DMB, and
triptane, with the first being the major product, as shown in
Table 9. Here the selectivity for the conversion of isobutane
into higher branched alkanes increases with higher MeOH to
isobutane ratios; as above, that also requires longer reaction
times. The 2:1 ratio of MeOH to isobutane is probably optimal;
although the selectivity appears to be higher at 3:1, significant
quantities of aromatics are observed for that reaction, indicating
some direct conversion of MeOH to triptane. Since the
calculated selectivity (as defined previously) assumes no such
conversion of MeOH, it somewhat overstates the actual ho-
mologation selectivity.
As with isopentane, there were substantial problems with mass
balance in all isobutane reactions because of its high volatility.
It should be noted that such losses will result in underestimating
the selectivity for homologation, as the apparent conversion of
starting branched alkane will be artificially high. Hence, the
selectivities reported in Tables 8 and 9 are lower limits.
The selectivity (either combined, or to triptane alone) is
essentially independent of the starting MeOH:isopentane ratio,
although as the amount of MeOH present increases, the rate of
reaction decreases significantly, so that higher temperatures and
longer reaction times are required. This is consistent with
The above examples all involve homologation of alkanes that
can form olefins that are on the pathway to triptane. Our earlier
observation that InI3 can catalyze isomerization of other alkanes,
so long as they contain at least one tertiary center,13 suggests
(17) See footnote d to Table 8. Additional experiments demonstrating this
conclusion will be published in a later paper.
9
J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008 11993

A R T I C L E S
Bercaw et al.
Table 10. Products from the Homologation of 2-Methylpentane
with MeOH
Scheme 2
product
yield (mg)
isobutane
isopentane
DMB
23.5
10.7
1.76
267
73.0
12.1
5.03
5.25
27.4
8.48
17.3
2-methylpentane
3-methylpentane
2,4-dimethypentane
triptane
2-methylhexane
2,3-dimethylpentane
3-methylhexane
C8 alkanes
process shown in Scheme 2. The carbocation intermediates will
undergo competitive skeletal rearrangements (which lead to
isomerization) and scissions (which lead to cracking); the
resulting side products and the carbocations and olefins derived
therefrom will participate in similar chemistry, using up the
MeOH that does not participate in the direct homologation route.
that alkanes such as 2-methylpentane could be homologated as
well. A reaction with MeOH (6.2 mmol), 2-methylpentane (6.2
mmol), InI3 (4.13 mmol), and adamantane (0.367 mmol), heated
at 180 °C for 30 min, exhibited the GC analysis shown in Table
10. A quantity of 2,3-dimethylpentane (27.4 mg), the expected
primary homologation product, was formed, and 267 mg (3.1
mmol, 50%) of 2-methylpentane was recovered, corresponding
to a selectivity for homologation of about 9%. The low
selectivity corresponds to a considerably higher level of
isomerization and cracking than that found in reactions of DMB;
we have observed that both 2-methylpentane and 2,3-dimeth-
ylpentane (which can undergo isomerization without a change
in the length of the main carbon backbone) are more reactive
for these transformations than DMB or triptane.13
Despite these side reactions, 2,3-dimethylpentane comprises
approximately 50% of the C7 fraction and highly branched C7
alkanes (2,3-dimethylpentane, 2,4-dimethylpentane and triptane)
comprise approximately 75% of the C7 fraction. Furthermore,
we observe C8 products in the homologation of 2-methylpentane
(in contrast to the homologation of DMB), indicating that further
methylation of C7 species has occurred. Our original explanation
for high triptane selectivity was based in part on the expectation
that methylation of a more substituted (and hence more electron-
rich) olefin will be faster.18 Hence, since 2,3-dimethyl-2-pentene
(the olefin in equilibrium with 2,3-dimethylpentane) is trisub-
stituted, whereas triptene (the olefin in equilibrium with triptane)
is only disubstituted, further methylation of 2,3-dimethylpentane
(the major C7 product from 2-methylpentane) should be more
extensive than methylation of triptane (the major C7 product
from DMB), as observed.
Attempts to homologate hexane resulted in only small
quantities of C7 products; even that required extremely long
reaction times.15 We believe this is because there are only
primary and secondary C-H bonds in hexane, which are more
difficult to activate than tertiary C-H bonds; also even if hexane
is activated, it will only form a disubstituted olefin, so the rate
of methylation is expected to be slow.
Mechanism. The mechanism proposed for the homologation
of DMB with MeOH begins with the activation of DMB by
InI3 to give the 2,3-dimethylbutyl carbocation, the same as the
first step in alkane isomerization catalyzed by InI3.13 The 2,3-
dimethylbutyl carbocation will be in equilibrium with 2,3-
dimethyl-2-butene, which is methylated to give the triptyl
carbocation. The latter presumably obtains a hydride from
another alkane to give triptane; most often that will come from
DMB, which is in greatest abundance, resulting in the chain
The facile transfer of H- between alkanes and carbocations
is a key feature of this chain mechanism, which must be
operating (or something very similar) to permit homologation
to proceed efficiently even though the steady-state concentration
of olefins, produced via alkane activation by InI3 followed by
deprotonation, must be very low indeed. It is notable that ZnI2
does not catalyze alkane homologation at these temperatures;
carbocations are formed as intermediates in the ZnI2-catalyzed
conversion of MeOH to triptane,10 so a chain process would
seem possible in principle, once reaction has been initiated.
Nonetheless, reactions in which a small amount of an olefin
(2,3-dimethyl-2-butene) as initiator was added to a solution
containing DMB, MeOH, and ZnI2 showed no conversion of
DMB to triptane at all. Our explanation for this observation is
related to the presence or absence of olefins. Transfer of H-
from an olefin to a carbocation is particularly favorable as it
generates an allylic carbocation (ultimately leading to arenes).10
In ZnI2-catalyzed reactions, a macroscopic amount of olefin is
required to generate the initial carbocation at 200 °C, and olefin
concentrations remain significant throughout, so that transfer
of H- from alkane never competes effectively, and no chain
process can take place. In contrast, with InI3-catalyzed homolo-
gation olefin concentrations are always extremely low, allowing
transfer from the large excess of DMB to dominate and the chain
process to proceed efficiently.
We attribute the beneficial effect of adamantane to its ability
to act as an efficient hydride transfer agent, analogous to the
interpretation of previous observations.16 Reactions shown in
Scheme 3 effectively catalyze the transfer of hydride between
the triptyl carbocation and DMB, thus improving the efficiency
of the chain process that is central to homologation, while
suppressing side reactions by partially quenching the carboca-
tions that lead to them. The primary carbocation derived from
DMB, which leads to isomerization via methyl shifts, is rapidly
trapped by adamantane (AdH) to reform DMB and generate
the adamantyl cation (Ad
+). Of course, AdH can also transfer
hydride to the tertiary DMB-derived cation, which at first would
seem to be chain-inhibitory, but that is reversible: the predomi-
nant fate of the resulting Ad+ (which does not appear to undergo
any side reactions of its own, under these conditions) will be
to take H- back from another molecule of DMB and start a
new chain. Also, of course, the driving force for transfer of H
from AdH to a (less stable) primary carbocation is greater than
that for transfer to the tertiary carbocation; the latter process is
(18) Mayr, H.; Schneider, R.; Irrgang, B.; Schade, C. J. Am. Chem. Soc.
1990, 112, 4454–4459.
9
11994 J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008

Selective Methylative Homologation
A R T I C L E S
Scheme 3
difficult case, the combined selectivity to higher branched
alkanes from homologation of isobutane exceeds 40%; a process
involving separation and recycle of the intermediate products
might be effectively applied. Methylative homologation also
affords product mixtures with much lower aromatic content than
most processes for direct conversion of methanol to hydrocar-
bons over zeolites or other catalysts, another highly desirable
feature. This route thus offers considerable promise for upgrad-
ing relatively low-value, lighter branched alkanes to species that
are both less volatile and higher octane, and hence significantly
more valuable fuels.
Experimental Section
General. InI3 (purchased from Alfa Aesar), ZnI2 (purchased from
Sigma-Aldrich), MeOH and other organic compounds were reagent-
grade commercial samples used without further purification. GC
analyses were performed on an HP model 6890N chromatograph
equipped with a 10 m × 0.10 mm × 0.40 µm DB-1 column. GC/
MS analyses were performed on an HP model 6890N chromato-
graph equipped with a 30 m × 25 mm × 0.40 µm HP5-1 column
and equipped with an HP 5973 mass selective EI detector.
approximately thermoneutral (in the gas phase).19 One function
of adamantane is thus basically a ”repair” mechanism, capturing
the isomeric carbocation before it can undergo skeletal rear-
rangement and returning it to the parent DMB pool. Further-
more, by decreasing the lifetime of the triptyl carbocation (and
others) adamantane will lower the rate of cracking. On both
counts, then, the addition of adamantane inhibits the rates of
side reactions relative to methylative homologation.
The mechanism for the homologations of isopentane and
isobutane would look much the same, starting from an earlier
point in the growth sequence. In each case, methylation of the
olefin derived from the starting alkane takes place so as to
generate the most substituted carbocation; thus, DMB is a
significant product, along with triptane, from the homologation
of isopentane, and likewise isopentane is one of the products
from isobutane.
Standard Reaction Protocols. All reactions were performed in
thick-walled pressure tubes equipped with Teflon stopcocks (Kontes
valves), rated up to 10 bar. The procedure for alkane homologation
reactions is based on the procedure reported earlier for MeOH to
hydrocarbon conversions using ZnI2 and InI3.10,13 In a typical
experiment, the tube was equipped with a stir bar and charged with
InI3 (2.05 g, 4.1 mmol), MeOH (0.25 mL, 6.2 mmol), and DMB
(0.807 mL, 6.2 mmol). (The InI3 was weighed out in a glovebox
due to its hygroscopic nature; however the reactions were carried
out under an atmosphere of air.) If adamantane was utilized, it was
added at this stage. The pressure tube was then placed in a preheated
oil bath behind a blast shield and stirred at the appropriate
temperature for the desired period of time. After heating, the tube
was removed from the bath and allowed to cool to room temperature
and then placed in an ice bath. The stopcock was removed, and
chloroform (1.0 mL), containing a known amount of cyclohexane
as an internal standard, was pipetted into the reaction mixture
followed by water (0.5 mL). The stopcock was replaced, the mixture
was shaken vigorously and the organic layer separated. A small
aliquot was diluted with acetone or tetradecane for GC analysis.
Conclusions
The results presented above demonstrate that lighter branched
alkanes can be converted to more valuable products through a
novel process involving methylative homologation. Product
distributions, as well as the observation that homologation yields
are substantially improved by the addition of adamantane as
hydride transfer agent, are consistent with a mechanism in which
InI3 activates branched alkanes to produce tertiary carbocations.
The latter are in equilibrium with olefins which undergo chain
growth by reacting with methylating species generated from
methanol and InI3, as previously shown for the direct conversion
of methanol to triptane.13 Although carbocationic mechanisms
are not usually associated with highly selective transformations,
selectivities as high as 65% based on converted alkane were
achieved for the homologation of DMB to triptane. Selectivities
in homologations of the abundant branched alkanes isopentane
and isobutane are not quite so high, but even for the most
In reactions involving isobutane, all reagents except isobutane
were loaded into the tube. The tube was then degassed using three
consecutive freeze-pump-thaw cycles and frozen in liquid
nitrogen. The desired amount of isobutane was condensed into the
tube using a calibrated gas bulb, and the tube was allowed to warm
to room temperature and then heated as described above. For
reactions involving isopentane all components were cooled at 0
°C to minimize evaporation, and the isopentane was then added.
Acknowledgment. We thank Professor Enrique Iglesia for
useful discussions. This work was supported by BP through the
MC2 program.
Supporting Information Available: A complete listing of
experimental results (in Excel format), along with an explanatory
guide to the Excel tables and details of a labeling experiment.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(19) (a) Solomon, J. J.; Field, F. H. J. Am. Chem. Soc. 1975, 97, 2625–
2628. (b) Meot-Ner, M; Solomon, J. J.; Field, F. H. J. Am. Chem.
Soc. 1976, 98, 1025–1026. (c) Kruppa, J. H.; Beauchamp, J. E. J. Am.
Chem. Soc. 1986, 108, 2162–2169.
JA803029S
9
J. AM. CHEM. SOC. VOL. 130, NO. 36, 2008 11995