Efficient thermochemical alkane dehydrogenation and isomerization catalyzed by an iridium pincer complex

Article abstract of DOI:10.1039/a900631a

((i-Pr)PCP)IrH₂ is found to be a remarkably effective solution-phase catalyst for the 'acceptorless' thermochemical dehydrogenation of cycloalkanes (and isomerization in the case of cyclodecane), and the first such catalyst reported to effect the dehydrogenation of n-alkanes.

Full text of DOI:10.1039/a900631a

Efficient thermochemical alkane dehydrogenation and isomerization catalyzed
by an iridium pincer complex
Fuchen Liu and Alan S. Goldman*
Department of Chemistry, Rutgers University, Piscataway, NJ 08854-8087, USA.
E-mail: goldman@rutchem.rutgers.edu
Received (in Bloomington, IN, USA) 22nd January 1999, Accepted 9th February 1999
(
i2PrPCP)IrH2 is found to be a remarkably effective solution-
dehydrogenation cycle may be even more sterically disfavored
by the tert-butyl groups. Accordingly we synthesized† and
investigated the i–PrPCP analog of 1, i.e. (i-PrPCP)IrH2 (2).
Vigorously refluxing (151 °C) a cyclooctane solution of 2
(1.0 mM) leads to the formation of 47 mM cyclooctene (47
turnovers) after 30 min. The reaction rate subsequently levelled
off; 105 mM cyclooctene was present after 15 h. This initial
dehydrogenation rate ( > 94 turnover h21) may be compared
with 11 turnover h21 resulting from cyclooctane dehydrogena-
tion catalyzed by 1 under similar conditions,5 and 1.41 turnover
h21 reported for IrH2(O2CCF3)(PCy3)2.7
Cyclodecane, with a reflux temperature of 201 °C, undergoes
dehydrogenation at a much greater rate than cyclooctane: for
example 460 turnovers cyclodecene are obtained after 1 h of
reflux, (cis:trans ratio of 4.6:1; see Table 1). This compares,
for example, with 170 turnover after 4 h obtained from the use
of catalyst 1.5 The isopropyl derivative 2 is thus an extremely
efficient catalyst for acceptorless dehydrogenation of cycloalk-
anes, yielding turnover rates a full order of magnitude greater
than the tert-butyl derivative.
The dehydrogenation enthalpy of cyclooctane, and probably
that of cyclodecane as well, is anomalously low.9–12‡ Accord-
ingly, with n-alkanes, the t-BuPCP catalyst 1 gave no observable
yields of acceptorless dehydrogenation product.§ By contrast, 2
is able to catalyze the acceptorless dehydrogenation of n-
alkanes, though the efficiency is less than that with cyclode-
cane. Refluxing an n-undecane solution of 2 (1.0 mM; 196 °C)
yields 35.4 mM undecenes, as a mixture of internal isomers (not
fully characterized), after 0.5 h. Unfortunately, the yield quickly
levelled off and after 1 and 45 h the respective concentrations of
undecenes were 42.3 and 44.1 mM. This represents the first
report of a homogenous catalytic system for the dehydrogena-
tion of n-alkanes.
To determine whether product concentrations level off due to
catalyst decomposition or product inhibition, we allowed an n-
decane solution of 2 to reflux (174 °C) until the decene
concentration reached a constant value.§ Volatiles (decenes and
decane) were then removed and fresh n-decane was added; upon
refluxing, the kinetics of decene production were identical to
those of the initial run indicating that the catalyst did not
decompose. This result is similar to that obtained previously for
1-catalyzed dehydrogenation of cyclooctane.5 In the case of n-
alkanes, however, the removal of product, if sufficiently rapid,
could have an additional and very significant benefit. When n-
alkane dehydrogenation is conducted (at 150 °C) using a
phase catalyst for the ‘acceptorless’ thermochemical dehy-
drogenation of cycloalkanes (and isomerization in the case of
cyclodecane), and the first such catalyst reported to effect
the dehydrogenation of n-alkanes.
Alkanes are the world’s most abundant organic resource, yet
methods for their conversion to useful chemicals remain
extremely limited. Alkenes, by contrast, are probably the single
most useful and versatile class of feedstock for commodity-
scale organic chemical syntheses. Methods for the dehydroge-
nation of alkanes to give alkenes are therefore of great
interest.1
In general, the yields and selectivity afforded by heteroge-
neous alkane dehydrogenation catalysts are very low.1 Soluble
transition metal complexes, however, have held promise in this
respect since the early 1980s when Crabtree and Felkin reported
the development of catalysts for transfer-dehydrogenation.2
R5
R7
R1
R3
H
H
H
H
catalyst
heat
R1
R3
R5
R7
+
+
R6
R8
R2
R4
R2 R4
R6 R8
The initially reported catalysts suffered from ligand degrada-
tion which severely limited turnover numbers. Systems were
later developed that afforded much higher turnover numbers;3
of these, the most important in the context of the present work
was the ‘pincer’ complex, (t-BuPCP)IrH2 [1; t–BuPCP
=
2,5-C6H3(CH2PBut2].4
PR2
(RPCP)IrH2
H
Ir
1: R = t-Bu
2: R = i-Pr
H
PR2
Recently [eqn. (1)] it was reported that refluxing alkane
solutions of 1 resulted in dehydrogenation with evolution of
dihydrogen; no sacrificial hydrogen acceptor was required.5
Although this was by far the most efficient catalyst to date for
such ‘acceptorless’ dehydrogenation,6,7 the observed rates were
still fairly low by the standards of practical catalytic systems.
(RPCP)IrH2
alkane
alkene + H2
↑
(1)
reflux
To a significant extent, the pincer geometry ‘holds back’ the
phosphinoalkyl groups such that they are further removed from
the metal, and the molecule is sterically less congested, relative
to a hypothetical analogous complex of monodentate ligands
[e.g. Ir(PR2Me)2(Ph)H2]. Nevertheless, Milstein and coworkers
recently reported that H2 adds to the carbonyl complex
Table 1. Dehydrogenation and isomerization of cyclodecane catalyzed by
2a
t/h
cis-Cyclodecene trans-Cyclodecene cis-DECb trans-DEC
1
4
8
378
619
692
700
82
145
163
163
16
38
42
43
10
59
79
81
(
i-PrPCP)Ir(CO)8 whereas, in contrast, we find that the t-BuPCP
analog undergoes no observable addition of H2 (800 Torr). Thus
in spite of the ‘pincer effect’ the tert-butyl groups are apparently
still bulky enough to inhibit H2 addition; this led us to suspect
that alkyl hydrides and other presumed intermediates of a
20
a Refluxing cyclodecane; 1.0 mM 2; values are turnover numbers (equal to
concentration/mM). b DEC = diethylcyclohexane.
Chem. Commun., 1999, 655–656
655

sacrificial acceptor, the major kinetic product is the terminal (a)
olefin which then undergoes isomerization to the more stable
internal olefins.13 Presumably the same kinetic product is
formed in the absence of an acceptor but isomerization is much
more rapid relative to dehydrogenation; accordingly, added dec-
1-ene is rapidly isomerized ( > 99% in 2.5 min) in refluxing n–
undecane. Thus, efficient dynamic removal of product from the
mixture might lead not only to high turnover numbers, but also
to high regioselectivity for a-olefins in the case of n-alkanes.
Returning to the 2-catalyzed dehydrogenation of cyclode-
cane, two intriguing products other than cis- and trans-
cyclodecene are observed to accumulate in significant concen-
tration (by GC and GC–MS); GC–MS and 1H NMR data
indicate them to be diethylcyclohexanes (DEC). When cis- and
trans-1,2-DEC were independently generated by the hydro-
genation of 1,2-diethylbenzene; 1H NMR, GC and GC–MS data
were in complete agreement with those of the cyclodecane
reaction products.
The unprecedented isomerization of cyclodecane to cis- and
trans-1,2-DEC might seem to suggest the involvement of a
cyclodecane C–C bond activation step. It is difficult, however,
to envisage such a mechanism that would account for the
selectivity of this ring-contraction. Furthermore, the formation
of DEC’s appeared to be a secondary reaction since the ratio of
the DEC’s to cyclodecenes significantly increased with time
(Table 1). This leads us to propose, for the overall isomerization
reaction, the mechanism of Scheme 1 which receives strong
additional support from the following observations:
ization proceeding via dehydrogenation and a secondary olefin
reaction.
Support by the Division of Chemical Sciences, BES, OER,
US Department of Energy is gratefully acknowledged. We
thank Johnson-Matthey for a generous loan of iridium.
Notes and references
† The protonated i-PrPCP ligand was synthesized according to Milstein8 and
reacted with [Ir(cyclooctene)2Cl] in refluxing toluene for 3 days. The
resulting (i-PrPCP)IrHCl was isolated and converted to (i–PrPCP)IrH2 using
the procedure previously reported for (t-BuPCP)IrH2.4
‡ The dehydrogenation enthalpy of cyclooctane is ca. 23.8(5) kcal mol21 as
determined by either direct measurement of hydrogenation or on the basis
of enthalpies of formation.9,10 The value for cyclodecane is less certain.
Based on available data for enthalpies of formation it would appear to be
rather high, at least 30 and 33 kcal mol21 for formation of cis- and trans-
cyclodecene respectively.9–11 However, direct hydrogenation measure-
ments yield values of 220.7(1) and 224.0(9) kcal mol21, respectively, in
acetic acid solvent.12 Enthalpies of other cycloalkene hydrogenations in
acetic acid, determined by the same workers, are ca. 1.3 kcal mol21 less
negative than reliable values in hydrocarbon solvent.9,12 Thus, 22 and 25
kcal mol21 are probably the best estimates for the dehydrogenation of
cyclodecane to give cis- and trans-cyclodecene respectively.
§ An attempt to dehydrogenate n-decane using catalyst 1 gave no detectable
decene. It is difficult to set an upper limit on the amount of decene produced
since as many as eight major isomers could be formed. Nevertheless, the
reaction of n-decane catalyzed by 2, unlike that of 1, gave very easily
detectable decene GC peaks indicating total concentrations of 14.4 and 16.1
mM after 0.5 and 5 h respectively.
(i) A mixture of 2 (1.0 mM) and cyclodecene (200 mM) in n-
undecane solvent was refluxed (196 °C). After 30 min, 4.2 mM
diethylcyclohexane had appeared (ca. 2:1 trans:cis);¶ in
addition, ca. 50% of the cyclodecene had been transfer-
hydrogenated to cyclodecane (with commensurate formation of
undecenes), while 98 mM cyclodecene remained. This rate of
DEC formation (8.4 turnover h21; 149 mM time-averaged
concentration of cyclodecene) is proportional to the time-
averaged concentration of cyclodecene present in the reactions
that began with cyclodecane only; i.e. this result is entirely
consistent with the DEC being derived from cyclodecene (not
directly from cyclodecane).
(ii) trans,trans-1,5-cyclodecadiene has been reported to
undergo a thermal Cope rearrangement to give trans-1,2-di-
vinylcyclohexane.14∑ Extrapolating the rate data to 201 °C gives
a rate of 1.2 s21,14** which is clearly consistent with Scheme 1
and the lack of build-up of observable quantities of the
cyclodiene. Presumably, cyclodecadienes other than the
1,5-isomers are initially formed and these undergo isomeriza-
tion to the 1,5-diene which then undergoes rearrangement.
(iii) Since divinylcyclohexane is not formed in an observable
quantity, the mechanism of Scheme 1 requires that this
intermediate is rapidly hydrogenated under the reaction condi-
tions. Accordingly, when a cyclodecane solution of 2 (1.0 mM)
and added vinylcyclohexane (100 mM) was refluxed, after only
10 min 100% conversion to ethylcyclohexane was observed.
¶ Both our dehydrogenations and our independent synthesis gave mixtures
of the two DEC’s; we are unable to assign their respective stereochemistry
with high confidence. However, the ratio changes with time during the
cyclodecane dehydrogenation and we tentatively assign the product that
relatively increases in concentration as the more stable trans-DEC.
∑ The product of the uncatalyzed reaction of 1,5-trans,trans-cyclodecadiene
is exclusively the trans-DEC isomer in contrast with our observation of a
mixture of trans and cis. However, MOPAC calculations indicate that the
1,5-cis,trans-cyclodecadiene is the more stable diene; molecular modeling
suggests that this isomer would yield cis-1,2-diethylcyclohexane. Fur-
thermore, Cope rearrangements can be metal-catalyzed which might
represent another pathway leading to the cis isomer.
** The measured rates (40–90 °C) give the following activation parameters:
DH‡ = 24.32 kcal mol21; DS‡ = 27.81 cal mol21 K21 14
.
1 P. N. Rylander, in Ullmann’s Encyclopedia of Industrial Chemistry, ed.
B. Elvers, J. F. Rounsaville and G. Schulz VCH Verlagsgesellschaft,
Weinheim, 1989, p. 494.
2 D. Baudry, M. Ephritikine, H. Felkin and R. Holmes-Smith, J. Chem.
Soc., Chem. Commun., 1983, 788; M. J. Burk, R. H. Crabtree, C. P.
Parnell and R. J. Uriarte, Organometallics, 1984, 3, 816.
3 J. A. Maguire and A. S. Goldman, J. Am. Chem. Soc., 1991, 113,
6706.
4 M. Gupta, C. Hagen, R. J. Flesher, W. C. Kaska and C. M. Jensen,
Chem. Commun., 1996, 2083; M. Gupta, C. Hagen, W. C. Kaska, R. E.
Cramer and C. M. Jensen, J. Am. Chem. Soc., 1997, 119, 840.
5 W. Xu, G. P. Rosini, M. Gupta, C. M. Jensen, W. C. Kaska, K. Krogh-
Jespersen and A. S. Goldman, Chem. Commun., 1997, 2273.
6 T. Fujii, Y. Higashino and Y. Saito, J. Chem. Soc., Dalton. Trans., 1993,
517.
7 T. Aoki and R. H. Crabtree, Organometallics, 1993, 12, 294.
8 B. Rybtchinski, A. Vigalok, Y. Bendavid and D. Milstein, Organome-
tallics, 1997, 16, 3786.
9 NIST Standard Reference Database Number 69, 1996, http://webboo-
k.nist.gov/chemistry/
10 D. R. Stull, E. F. Westrum and G. C. Sinke, The Chemical
Thermodynamics of Organic Compounds, Robert E. Kreiger Publishing,
Malabar, FL, 1987.
11 J. B. Pedley and J. Rylance, Computer Analysed Thermochemical Data:
Organic and Organometallic Compounds, University of Sussex,
Brighton, UK, 1977.
12 R. B. Turner and W. R. Meador, J. Am. Chem. Soc., 1957, 79, 4133.
13 F. Liu, E. B. Pak, B. Singh, C. M. Jensen and A. S. Goldman, J. Am.
Chem. Soc., in press.
In summary, we describe the first homogeneous catalyst
system to dehydrogenate alkanes at rates that might be
considered suitable for a practical and useful system. It would
appear that efficient removal of the olefin product, by either
physical or chemical means, might lead to dramatically
increased yields and even regioselectivity. The formation of
DEC represents a novel example of catalytic alkane functional-
(
i-PrPCP)IrH2
201 °C
(
i-PrPCP)IrH2 –2 H2
(
i-PrPCP)IrH2
+ 2 H2
Cope Rearrangement
14 P. S. Wharton and D. W. Johnson, J. Org. Chem., 1973, 38, 4117.
Scheme 1
Communication 9/00631A
656
Chem. Commun., 1999, 655–656