Low-temperature rhodium-catalyzed dehydration of primary alcohols promoted by tetralkylammonium and imidazolium halides

Article abstract of DOI:10.1021/om100716t

Rhodium complexes, promoted by imidazolium or tetraalkylammonium halide salts, catalyze the dehydration of primary alcohols with good conversion and selectivity.

Full text of DOI:10.1021/om100716t

Organometallics 2010, 29, 4001–4003 4001
DOI: 10.1021/om100716t
Low-Temperature Rhodium-Catalyzed Dehydration of Primary Alcohols
Promoted by Tetralkylammonium and Imidazolium Halides
George R. M. Dowson, Igor V. Shishkov, and Duncan F. Wass*
School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.
Received July 21, 2010
Scheme 1. Proposed Model Mechanism for Alcohol Dehydration
Summary: Rhodium complexes, promoted by imidazolium or
tetraalkylammonium halide salts, catalyze the dehydration of
primary alcohols with good conversion and selectivity.
With the growing scientific and social awareness of environ-
mental issues, chemical and energy companies are increasingly
looking to utilize alternative sustainable feedstocks. Ethylene
is the key current C2 synthetic intermediate in the petrochem-
ical industry, produced in large scale from nonrenewable
resources such as crude oil.1 Biomass-derived ethanol, or
bioethanol, is currently used as a fuel,2 but it also has the
potential to be used as a versatile C2 feedstock for commodity
chemicals3 or in the synthesis of more advanced fuels.4 Dehy-
dration of ethanol to ethylene is well-known and can be
achieved in the presence of a range of heterogeneous acidic
catalysts such as zeolites,5 polyacids,6 and acidic alumina.7
Such catalysts operate at temperatures of typically 300-
400 °C, with diethyl ether being a significant side product.
One of the most promising systems described to date is a
modified H-ZSM-5 zeolite catalyst which dehydrates ethanol
to ethylene with 42% yield and 99% selectivity at 170 °C.5d We
reasoned that a metal-based homogeneous catalyst system for
this seemingly simple transformation would be complemen-
tary to existing heterogeneous catalysts and potentially offer
advantages in certain applications. In this paper, we describe a
new transition-metal-catalyzed, homogeneous approach to the
dehydration of ethanol and its higher homologues and a
preliminary study into the mechanism by which the catalyst
operates.
of the well-known Monsanto process,8 save β-elimination in
our case leading to the desired product.
In a set of primary screening experiments, n-hexanol was
taken as a model for the dehydration of ethanol, since the
hexene or dihexyl ether dehydration products, being liquids,
could more easily and rapidly be analyzed by GC. Rhodium
complexes and hydriodic acid were utilized as components of
the catalytic system by analogy with Monsanto chemistry.
The results for the dehydration of hexanol are presented in
Table 1. We initially speculated that a wide range of simple
Rh complexes would show the desired reactivity. However,
although formation of hexenes was indeed observed with
only [Rh2(CO)4Cl2] and HI, the yields and selectivity were
extremely poor (run 1).
This disappointing result led us to investigate a range of
potential promoters. Halide salts are known to promote
carbonylation reactions with related species,9 and we found
that addition of 1-butyl-3-methylimidazolium chloride to the
reaction mixture gave significantly improved yields of hex-
enes (run 2). Addition of 2 equiv of PMes3 relative to
[Rh2(CO)4Cl2] led to further, modest improvements in cat-
alyst performance (run 3); we attribute this to improved
rhodium complex solubility rather than a ligand effect, since
in other related conditions no effect within error is observed
(for example, compare runs 13 and 18). Although some dihexyl
ether was also observed, good selectivity to the olefins is
achieved. The hexenes produced consisted of isomeric cis and
trans 2- and 3-hexenes with only minor amounts of 1-hexene.
This isomeric pattern is in line with what would be expected for
a standard metal hydride isomerization mechanism operating
Our approach is depicted in Scheme 1. Initially, the alcohol
should react with the acid HX to yield the corresponding
alkyl halide, which undergoes steps similar to the mechanism
*To whom correspondence should be addressed. E-mail: duncan.wass@
bristol.ac.uk.
(1) Arpe, H. J. In Industrielle Organische Chemie: Bedeutende Vor-
Und Zwischenprodukte, 6th ed.; Wiley-VCH: Weinheim, Germany, 2007.
(2) (a) Demirbas, M. F.; Balat, M. Energy Convers. Manag. 2006, 47,
2371. (b) Gnansounou, E.; Dauriat, A. J. Sci. Ind. Res. 2005, 64, 809. (c)
Hansen, A. C.; Zhang, Q.; Lyne, P. W. L. Bioresour. Technol. 2005, 96, 277.
(3) (a) Danner, H.; Braun, R. Chem. Soc. Rev. 1999, 28, 395. (b)
Kvisle, S.; Aguero, A.; Sneeden, R. P. A. Appl. Catal. 1988, 43, 117. (c)
Ndou, A. S.; Plint, N.; Coville, N. J. Appl. Catal. A: Gen. 2003, 251, 337. (c)
Gucbilmez, Y.; Dogu, T. Ind. Eng. Chem. Res. 2006, 45, 3496. (d) Inaba, M;
Murata, M.; Saito, M.; Takahara, I. React. Kinet. Catal. Lett. 2006, 88, 135.
(4) Tsuchida, T.; Yoshioka, T.; Sakuma, S.; Takeguchi, T.; Ueda, W.
Ind. Eng. Chem. Res. 2008, 47, 1443.
(5) (a) Talukdar, A. K.; Bhattacharyya, K. G.; Sivasanker, S. Appl.
Catal. A: Gen. 1997, 148, 357. (b) Takahara, I.; Saito, M.; Inaba, M.; Murata,
K. Catal. Lett. 2005, 105, 249. (c) Phillips, C. B.; Datta, R. Ind. Eng. Chem.
Res. 1997, 36, 4466. (d) van Mao, R. L.; Nguyen, T. M.; Maclaughlin, G. P.
Appl. Catal. 1989, 48, 265.
(8) (a) Yoneda, N.; Kusano, S.; Yasui, M.; Pujado, P.; Wilcher, S.
Appl. Catal., A 2001, 221, 253. (b) Ellis, P. R.; Pearson, J. M.; Haynes, A.;
Adams, H.; Bailey, N. A.; Maitlis, P. M. Organometallics 1994, 13, 3215 and
references cited therein.
(9) (a) Maitlis, P. M.; Haynes, A.; James, B. R.; Catellani, M.;
Chiusoli, G. P. Dalton Trans. 2004, 3409. (b) Smith, B. L.; Torrence,
ꢀ
(6) Lee, K. Y.; Arai, T.; Nakata, S.; Asaoka, S.; Okuhara, T.;
Misono, M. J. Am. Chem. Soc. 1992, 114, 2836.
G. P.; Murphy, M. A.; Aguilo, A. J. Mol. Catal. 1987, 39, 115. (c) Murphy,
ꢀ
M. A.; Smith, B. L.; Torrence, G. P.; Aguilo, A. J. Mol. Catal. 1986, 303,
(7) Knoezinger, H. Angew. Chem., Int. Ed. Engl. 1968, 7, 791.
257.
r
2010 American Chemical Society
Published on Web 08/25/2010
pubs.acs.org/Organometallics

4002 Organometallics, Vol. 29, No. 18, 2010
Dowson et al.
Table 2. Dehydration of Primary Alcoholsa
Table 1. Dehydration of 1-Hexanola
salt (amt
(mol %))
conversn of
alcohol (%)
selectivity to
alkenes (%)b
run
alcohol
22
23
24
ethanol
ethanolc
ethanol
Bu4NBr (21.6)
Bu4NBr (21.6)
Bu4NBr (21.6)
Bu4NCl (2.7)
Bu4NBr (21.6)
Bu4NCl (2.7)
Bu4NBr (21.6)
Bu4NCl (2.7)
Bu4NBr (21.6)
Bu4NBr (21.6)
83
80
51
53
56
67
conversn of
hexanol (%)
selectivity to
hexenes (%)c
run
salt (amt (mol %))b
25
26
ethanold
ethanole
68
67
88
86
1d
2d
3
4
5
6
7
8
9
10
11
12
13
14
15
16
none
4
13
22
24
19
5
48
49
80
11
18
65
78
18
34
72
50
92
92
>99
93
>99
63
62
60
28
94
71
79
96
29
84
ImCl (14.4)
ImCl (14.4)
ImCl (21.6)
ImCl (28.8)
ImCl (43.1)
ImBr (14.4)
ImBr (21.6)
ImBr (43.1)
Im(PF6) (14.4)
ImMeCl (14.4)
Bu4NBr (14.4)
Bu4NBr (21.6)
Bu4NCl (14.4)
Bu4NI (14.4)
Bu4NBr(21.6)
Bu4NCl (5)
Bu4NBr (21.6)
Bu4NBr (21.6)
ImCl (14.4)
LiBr (21.6)
27
28
n-propanol
n-butanol
77
95
96
>99
a Reaction conditions: 3 mL of alcohol, 0.065 mol % of [Rh2(CO)4Cl2],
0.13 mol % of PMes3, 6 mol% of 57% aqueous HI, 48 h, 110 °C. Conver-
sion and selectivity were determined by GC analysis of products or bromi-
nated products. b The other product is di-n-alkyl ether; 1-2% butenes was
also observed in the case of ethanol (see main text). c 6 mol % ethyl iodide
was employed instead of HI. d 0.13% of [Rh2(CO)4Cl2] and 0.26% PMes3.
e No PMes3 added.
bromide are more active for dehydration than are those with
chloride, although the latter give better selectivity. Intrigu-
ingly, a mixture of bromide and chloride salts can be used to
improve selectivity while maintaining conversion (run 16).
Addition of a phosphine ligand has less effect for these salts;
indeed, triphenylphosphine actually acts as a catalyst poison
under the same conditions (run 19). Neither dihexyl ether nor
olefinic products were observed in control runs when no HI
was employed, whereas the absence of the rhodium complex
in the reaction (run 17) led to formation of dihexyl ether
(20% yield) as the only product, presumably by simple acid
catalysis. The use of LiBr in place of imidazolium or tetra-
lkylammonium salts (run 20) gave very low yields of the
hexenes. Dehydration of hexanol can be also carried out
using 6 mol % of n-hexyl iodide in place of aqueous HI (run
21), supporting the mechanism depicted in Scheme 1. Mon-
itoring an ongoing run based on conditions in run 14 shows
smooth conversion of hexanol consistent with first-order
kinetics for an 18 h period, the final conversion figure
actually being achieved within this time (see the Supporting
Information for a typical graph). The selectivity vs time to
hexenes compared to dihexyl ether is invariant within error.
The optimized catalytic system from these primary screen-
ing experiments was extended to other alcohols (Table 2).
Dehydration of n-propyl alcohol and n-butyl alcohol pro-
ceeds smoothly, affording good yields of olefins with excel-
lent selectivity. Results for ethanol were initially more
disappointing, with poor selectivity being observed for tetra-
butylammonium bromide (runs 22 and 23). However, using a
mixture of bromide and chloride salts, as before, gave
improved selectivity (run 24) and yields were improved by
using higher catalyst loadings (run 25).
17e
18d
19f
20
20
83
4
4
70
0
74
78
60
56
21g
Bu4NBr (21.6)
a Reaction conditions: 3 mL of hexanol, 0.065 mol % of [Rh2(CO)4Cl2],
0.13 mol % of PMes3, 6 mol % of 57% aqueous HI, 110 °C, 48 h. The molar
percentage is given relative to hexanol. Conversion and selectivity were
determined by GC. b Im = 1-methyl-3-butylimidazolium. c The other product
is di-n-hexyl ether. d No ligand used. e No Rh catalyst used. f 0.13 mol %
PPh3 added. g 6 mol % hexyl iodide in place of HI.
under thermodynamic control. Consistent with this hypoth-
esis, the same isomer distribution within error was observed
when pure 1-hexene was refluxed with [Rh2(CO)4Cl2] and
HI. Further screening revealed that the optimum amount of
the imidazolium salt is 20-30 mol % relative to the alcohol
(runs 3-6); the yield of hexenes was decreased if higher salt
amounts were used. The use of 1-butyl-3-methylimidazolium
bromide gave improved conversion (runs 7-9), although
selectivity to olefins is reduced by around 30%. The role of
halide seems crucial so that only very poor results are
obtained with Im(PF6), despite this being a more hygro-
scopic additive (run 10); this suggests a coordinative role for
the halide, in line with the proposed role of such species in
carbonylation catalysis.9 We believe N-heterocyclic carbene
ligated species formed via the imidazolium salts can also be
ruled out: 1-butyl-2,3-dimethylimidazolium chloride (run
11) cannot form such species but gives very similar results
to the monomethylated derivative.10
Good results are also obtained if simple tetraalkylammo-
nium halides are utilized. For example, tetrabutylammo-
nium bromide gives an increase in yield and selectivity
compared to imidazolium bromides (compare runs 8 and
13). Again, a halide anion effect is observed so that tetra-
butylammonium chloride gives a lower yield but higher
selectivity to olefin; in contrast, for tetrabutylammonium
iodide, a lower yield is accompanied by a preference in
selectivity toward dihexyl ether. In general, salts containing
It is not clear why the dehydration of ethanol gave lower
yields of olefin as compared to n-propanol or n-butanol. A
possible explanation could be product inhibition via stronger
binding of the π-acidic ethylene at the rhodium center; rhodium
ethylene complexes are observed by NMR spectroscopy in a
typical run (vide infra). Added trimesitylphosphine has no
effect within error for ethanol dehydration (run 26).
Studies are underway to define the mechanism of this reac-
tion, and our preliminary investigations point to a number of
(10) So-called “abnormal’” carbenes cannot be ruled out by this
observation but, given the very similar performance of tetraalkylammo-
nium halides, are unlikely to be important for these catalysts. Abnormal
carbenes: Schuster, O.; Yang, L.; Raubenheimer, H. G.; Albrecht, M.
Chem. Rev. 2009, 109, 3445.
(11) (a) Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds.;
Wiley-VCH: Weinheim, Germany, 2002. (b) Weldon, T. Chem. Rev. 1999,
99, 2071. (c) Dupont, J.; de Souza, R. F.; Suarez, P. A. Z. Chem. Rev. 2002,
102, 3667. (d) Parvulescu, V. I.; Hardacre, C. Chem. Rev. 2007, 107, 2615.

Communication
Organometallics, Vol. 29, No. 18, 2010 4003
roles for the crucial tetraalkylammonium salts. First, they
solubilize the various reaction components; the use of similar
salts as ionic liquids11,12 or phase-transfer agents13 is well-known
to lead to unique reactivity. No phase separation is observed at
the reaction temperature. Second, they generate the active
catalytic species. Treatment of [Rh2(CO)4Cl2] with 1 equiv of
Bu4NBr in hexanol leads to virtually instantaneous conversion
to [Bu4N][Rh(CO)2Br2] (νCO 2064 and 1990 cm-1). Addition of
HI leads to equally rapid formation of [Bu4N][Rh(CO)2I2] (νCO
2059 and 1987 cm-1), which persists as the only Rh-CO species
observable by IR spectroscopy throughout a 48 h run. Forma-
tion of such types of complexes by this reaction has already been
described in the literature.14 Addition of 1 equiv of EtI to
[Bu4N][Rh(CO)2Br2] in the absence of additional salt leads
directly to the production of ethylene; intermediate Rh-Et
species are not observed by NMR spectroscopy at room
temperature, suggesting that β-elimination of these species is
rapid. Both free ethylene at 5.26 ppm and signals consistent with
η2-coordinated ethylene at 5.23 ppm (ratio of free to coordi-
nated species 6:4) are observed. Surprisingly, in the absence of
additional salt, only one turnover to produce the olefin is
observed, even with a large excess of EtI or EtOH. Only upon
addition of a minimum of a 4-fold excess of Bu4NBr to rhodium
is catalytic turnover achieved. This suggests a third role for the
added salt: to regenerate [Bu4N][Rh(CO)2X2] (X = Br, I) by
facilitating the reductive elimination of HX from [Bu4N]-
[RhH(CO)2X3] in the final step of our proposed catalytic cycle.
This hypothesis is consistent with the report of Bunel and co-
workers, which demonstrated the facile disproportionation of
[RhH(CO)2X3]- species to [Rh(CO)2X2]-, [Rh(CO)2X4]-, and
H2 in the absence of added salt. It also has parallels in halide-
promoted rhodium-catalyzed methanol carbonylation at low
water concentrations, where halide (usually iodide) is believed
to stabilize catalytically active rhodium species from decom-
position and precipitation.9b,c It is noteworthy that such
stabilization effects are particularly important at low CO
pressure, the CO concentration in our case, of course, being
effectively zero. Work is ongoing to define the specific
mechanism by which salts facilitate catalyst regeneration
and the origin of the subtle selectivity changes observed
depending on halide.
(12) Promoting effect of ionic liquids: (a) Sliger, M. D.; P’Pool, S. J.;
Traylor, R. K.; McNeil, J., III; Young, S. H.; Hoffmann, N. W.;
Klingshirn, M. A.; Rogers, R. D.; Shaughnessy, K. H. J. Organomet.
Chem. 2005, 690, 3540. (b) P'Pool, S. J.; Klingshirn, M. A.; Rogers, R. D.;
Shaughnessy, K. H. J. Organomet. Chem. 2005, 690, 3522.
(13) (a) Yadav, G. D. Top. Catal. 2004, 29, 141. (b) Albanese, D.;
Landini, D.; Maia, A.; Penso, M. J. Mol. Catal. A: Chem. 1999, 150, 113. (c)
Jin, G.; Morgner, H.; Ido, T.; Goto, S. Catal. Lett. 2003, 86, 207. (d) Jeffery,
T. Tetrahedron 1996, 52, 10113.
(14) (a) Vallarino, L. M. Inorg. Chem. 1965, 4, 161. (b) Fulford, A.;
Bailey, N. A.; Adams, H.; Maitlis, P. M. J. Organomet. Chem. 1991, 417,
139.
(15) Roe, D. C.; Sheridan, R. E.; Bunel, E. E. J. Am. Chem. Soc.,
1994, 116, 1163.
In summary, this new homogeneous catalyst for ethanol or
other primary alcohol dehydration gives olefins in good
selectivity and opens possibilities for the conversion of
bioderived alcohols to higher value products.
Acknowledgment. We thank the EPSRC (to I.V.S.)
and BP (to G.D.) for funding.
Supporting Information Available: Text and figures giving
experimental details. This material is available free of charge via
the Internet at http://pubs.acs.org.