COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Enhanced selectivity in the conversion of methanol to
,2,3-trimethylbutane (triptane) over zinc iodide by added phosphorous
2
or hypophosphorous acid{
John E. Bercaw, Robert H. Grubbs, Nilay Hazari, Jay A. Labinger* and Xingwei Li{
Received (in Cambridge, UK) 11th April 2007, Accepted 6th June 2007
First published as an Advance Article on the web 20th June 2007
DOI: 10.1039/b705470j
The yield of triptane from the reaction of methanol with zinc
iodide is dramatically increased by addition of phosphorous or
hypophosphorous acid, via transfer of hydride from a P–H
bond to carbocationic intermediates.
which trimethyl phosphite could modify the chemistry: 1) by
2+
increasing the acidity of the medium; 2) as a ligand for Zn ,
changing the nature of the catalyst; 3) as a more efficient
methylating agent, in an Arbuzov-like reaction; and 4) as a
reducing agent. The absence of a similar effect for trimethyl
3
1
Methanol is expected to play an increasingly important role as an
energy source and chemical intermediate. One approach that has
attracted recent attention is the dehydrative conversion of
phosphate appears to rule out the first.§ P NMR spectroscopy
demonstrates that the water liberated by dimethyl ether formation
rapidly hydrolyzes trimethyl phosphite to a mixture of phosphor-
ous acid and its monomethyl ester. Both of these exist almost
1
methanol to hydrocarbons; in particular, conversion of methanol
5
to light olefins (MTO) over zeolitic materials has been the object of
2
intense mechanistic study. In contrast, the reaction of methanol
entirely as the phosphoryl tautomer and hence would not be
available for either the second or third function, suggesting the
reducing agent explanation is the most likely. Indeed, at the end of
7
over zinc iodide gives a highly-branched C alkane, 2,2,3-
3
1
trimethylbutane (triptane, Eq. 1), in surprisingly high selectivity
3
up to 20% yield on a moles carbon basis). We recently reported
the reaction the only significant P NMR signal is that of H PO .
3
4
(
According to this interpretation, phosphorous acid should work
just as well as trimethyl phosphite, while hypophosphorous acid,
which has two P–H bonds, should be equally good or even better.
Indeed, as shown in Table 1, the results using H PO and P(OMe)
extensive studies implicating a carbocation-based mechanism for
this transformation, involving successive methylation of lighter
olefinic intermediates and hydride transfer to the resulting
carbocations to generate alkanes, along with multiply unsaturated
3
3
3
are identical (after correcting for the additional methyl groups
provided by the latter), while H PO gives additional enhance-
ment, up to 32% yield. The highest yield (36%, nearly double the
baseline case) was obtained with H PO by reducing the reaction
temperature (at the cost of a much longer reaction time).
NMR spectra (Fig. 1), followed over the course of the latter
reaction, shows that H PO is indeed oxidized, first to H PO and
ultimately to H PO
4
species that end up mainly as methylated benzenes. We report
3
2
here that the yield is significantly enhanced by the addition of
certain phosphorus reagents, whose unusual mode of operation is
consistent with the previously proposed mechanistic explanation of
selectivity.
3
2
31
P
3
2
3
3
ð1Þ
3
4
.
As part of our ongoing research program on this system, we
examined the effect of water-sequestering agents, since water
a
Table 1 Triptyl yields from the reaction of methanol and zinc iodide
in the absence and presence of P–H bonded additives
b
(
produced at the earliest stages of reaction by dehydration of
Triptyl
yield,
mg
Triptyl yield,
% based on
MeOH
Triptyl yield,
% based on
total C
methanol to dimethyl ether) was observed to inhibit conversion.
Several such additives, including trimethyl orthoformate, dimethyl
Additive, mol%
rel. to MeOH)
(
2 5
carbonate, P O , and trimethyl phosphate had little or no effect. In
—
PO(OMe) , 6.8%
66
69
108
65
100
102
89
122
129
19
20
31
18
28
29
25
35
36
18
16
24
17
23
21
23
32
36
contrast, addition of trimethyl phosphite (7 mol% relative to
methanol) resulted in a marked increase in yield, from 18% to 24%,
even after accounting for the additional carbon provided in the
additive (Table 1).
3
P(OMe)
P(OMe)
P(OMe)
P(OMe)
3
3
3
3
, 6.8%
, 1.7%
, 3.4%
c
c
c
, 10.2%
We can conceive of at least four mechanisms (besides water
removal, which does not appear to have a beneficial effect) by
H
3
H
3
H
3
a
PO
PO
PO
3
2
2
, 6.8%
, 7.4%
, 7.4%
d
Yields are given as total ‘‘triptyls’’, triptane plus triptene, as these
are not cleanly separated by our routine GC analytical procedure.
The relative amounts can be readily distinguished by C NMR
Arnold and Mabel Beckman Laboratories of Chemical Synthesis,
California Institute of Technology, Pasadena, California, USA 91125.
E-mail: jal@its.caltech.edu; Fax: 1-626-449-4159; Tel: 1-626-395-6520
13
b
spectroscopy. Except as noted, reactions were carried out as
{
Electronic supplementary information (ESI) available: Experimental
4
previously described for 3 h at 200 uC, using 790 mg MeOH with
31
details, additional tabular results, P NMR spectra, and discussion of
mechanistic alternatives. See DOI: 10.1039/b705470j
3
indicated mol% of additive. Reaction time 2 h. Reaction carried
2 mol% ZnI as catalyst, 2.6 mol% i-PrOH as promoter, and the
2
c
d
{ Present address: Division of Chemistry and Biological Chemistry,
Nanyang Technological University, Singapore, 639623.
out at 175 uC for 24 h; no i-PrOH added.
2
974 | Chem. Commun., 2007, 2974–2976
This journal is ß The Royal Society of Chemistry 2007
Bercaw, John E.
