Thermochemical alkane dehydrogenation catalyzed in solution without the use of a hydrogen acceptor

Article abstract of DOI:10.1039/a705105k

(PCP)IrH₂ [PCP = η³-C₆H₃(PBu^t₂) ₂-1,3] catalyzes the efficient (several hundred mol product/mol catalyst) dehydrogenation of alkanes under reflux to give the corresponding alkenes and dihydrogen.

Full text of DOI:10.1039/a705105k

View Article Online / Journal Homepage / Table of Contents for this issue
Thermochemical alkane dehydrogenation catalyzed in solution without the use
of a hydrogen acceptor
Wei-wei Xu,a Glen P. Rosini,a Mukta Gupta,b Craig M. Jensen,*b William C. Kaska,c Karsten
Krogh-Jespersena and Alan S. Goldman*a
a Department of Chemistry, Rutgers, The State University of New Jersey, New Brunswick, NJ 08903, USA
b Department of Chemistry, University of Hawaii, Honolulu, HI 96822, USA
c Department of Chemistry, University of California, Santa Barbara, CA 93106, USA
3
(PCP)IrH2 [PCP = h -C6H3(PBut2)2-1,3] catalyzes the
efficient (several hundred mol product/mol catalyst) dehy-
drogenation of alkanes under reflux to give the correspond-
ing alkenes and dihydrogen.
The high endothermicity of alkane dehydrogenation is
obviously a key factor involved in the challenge of developing
alkane dehydrogenation catalysts, particularly in developing an
acceptorless dehydrogenation system. In this respect cyclo-
octane is anomalous, since its dehydrogenation enthalpy of 23.3
kcal mol21 contrasts with ca. 28–30 kcal mol21 required for the
elimination of two secondary C–H bonds from more typical
alkanes.10 Extrapolation to typical alkanes, based on behavior
observed with cyclooctane, is therefore, a priori, quite tenuous.
We find, however, that the catalytic efficacy of (PCP)IrH2 with
other alkanes is at least comparable to that with cyclooctane and
the higher reflux temperatures of less volatile solvents permit
significantly greater rates. For example, in refluxing cyclo-
decane (201 °C), 170 and 360 turnovers cyclodecene are
obtained after 4 and 24 h, respectively (ca. 3:1 cis:trans;
1.0 mm catalyst; see Fig. 1). This rapid dehydrogenation occurs
in spite of the high dehydrogenation enthalpy of cyclodecane
(30 and 33 kcal mol21 to form cis- and trans- cyclodecene,
respectively).10
The decrease in the rate of alkene formation which invariably
occurs with time is probably not due to inevitable decomposi-
tion of the catalyst but may be caused by build-up of alkene
product. Accordingly, a fresh solution of catalyst in cyclo-
octane, to which had been added 10% COE, showed no increase
in COE concentration upon reflux. The mechanism by which
the alkene product inhibits the reaction remains to be deter-
mined.
The development of effective systems for the functionalization
of alkanes is one of the major goals in the field of homogeneous
catalysis.1 Conversion of alkanes to the corresponding alkenes
is a particularly valuable functionalization reaction in view of
the great versatility of alkenes as organic feedstock. Crabtree
et al. first showed in 1979 that transition-metal phosphine
complexes could effect stoichiometric dehydrogenation of
alkanes;2 this was later extended to catalytic dehydro-
genation.3,4 The high endothermicity of dehydrogenation was
compensated for by either the use of a sacrificial hydrogen
acceptor or by UV irradiation.5
Such systems were all plagued by ligand decomposition,
however, and it was not until 1988 that the first ‘high-turnover’
( > ca. 100 turnovers) dehydrogenation system was discovered:
Rh(PMe3)2(CO)Cl under UV irradiation.6 This robust catalyst
was later found to yield efficient thermochemical dehydro-
genation through the use of sacrificial hydrogen acceptors
and, surprisingly, added dihydrogen; unfortunately, the pres-
ence of H2 atmosphere resulted in the hydrogenation of more
than one mole sacrificial acceptor per mole dehydrogenated
product.7
Most recently, the development of the first ‘high-turnover’
thermochemical catalysts not requiring H2 atmosphere was
Significant amounts of cyclooctane can also be dehydrogen-
ated to give COE and molecular hydrogen in reactions carried
out in a closed system under reduced pressure. In a typical
experiment, a solution of cyclooctane (4.0 ml) and (PCP)IrH2 (3
mg, 0.005 mmol) was sealed in a tube under 0.007 mbar of
argon and fully immersed in a 200°C oil bath. After 0.5 h, 10
turnovers of COE had been produced. The production of H2 was
confirmed by gas chromatographic analysis† of the vapor above
3
reported: complexes of the ‘pincer’ ligand h -C6H3(CH2P-
But2)2-1,3 (PCP), including (PCP)RhH2 or, more effectively,
(PCP)IrH2.8
The key to the effectiveness of the PCP-based catalysts
appears to be their long-term stability at very high temperatures
(e.g. 200°C).9 We considered that such temperatures are
sufficiently high, in principle, to overcome the large positive
enthalpy of dehydrogenation without the use of a sacrificial
acceptor.†10 This was indeed found to be the case and the
remarkable stability of the catalyst is maintained under such
conditions. Herein, we report the first example of efficient
alkane dehydrogenation catalyzed homogeneously without the
use of either light or a sacrificial hydrogen acceptor.
Refluxing a cyclooctane solution of (PCP)RhH2 (10 mm)
while passing a stream of argon above the condenser results in
negligible rates of cyclooctene formation, ( < 2 mm after 24 h).
As was found when sacrificial acceptors were used, the iridium
analog is much more effective for ‘acceptorless’ dehydrogena-
tion. Under similar conditions [2 mm (PCP)IrH2], the initial rate
of cyclooctene (COE) formation is 11 turnovers h21.‡
After 44 h, 104 turnovers are obtained; after 120 h, 190
turnovers are produced. These results may be compared with the
most effective soluble ‘acceptorless’ dehydrogenation catalyst
reported to date: IrH2(O2CR)(PCy3)2 (1.25 mm) was reported to
yield 1.41 turnovers h21 COE, with a maximum of 28.5
turnovers obtained after 48 h.11
400
300
200
PBut2
IrH2
PBut2
100
0
↓
C10H20
C10H18 + H2
reflux (201 °C)
5
15
t / h
25
0
10
20
Fig. 1 Cyclodecene formation (total, ca. 3:1 cis:trans) vs. time; (PCP)IrH2
in refluxing cyclodecane (1.0 mm, 201 °C)
Chem. Commun., 1997
2273

View Article Online
the reaction mixture. Increased levels of cyclooctane dehydro-
genation were not observed in solutions which were heated for
longer reaction times. However, additional dehydrogenation
activity was achieved in solutions which had been heated to
200 °C for 0.5 h by removing the resulting H2 by freeze–pump–
thaw degassing. A total of 36 turnovers of COE were obtained
in an experiment in which H2 was removed four times through
this procedure. These results demonstrate that under these
conditions the catalysis is limited by equilibrium constraints
rather than catalyst stability.
The PCP ligand clearly bears a close relationship (at least
formally) to the ligand set in the putative catalytically active
fragments M(PR3)2Cl (M = Rh,7 Ir12). In the case of the chloride
complexes, rhodium has proven to be the more effective metal
(for both transfer- and photo-dehydrogenation).13 Thus the
much greater efficacy of Ir vs. Rh in the case of the PCP
complexes presents an issue which merits attention.
Footnotes and References
* E-mail: goldman@rutchem.rutgers.edu; jensen@gold.chem.hawaii.edu
† For example, if DG≠ = 30.0 kcal mol (1 cal = 4.184 J), at 200 °C the
catalytic turnover rate would be 500 h21. Since the dehydrogenation
enthalpy of typical alkanes is on the order of 28–30 kcal mol21 10
reasonable rates could be obtained (in principle) even if the favorable reac-
tion entropy does not make any contribution in the transition state. Of
course, a significant concentration of alkene can only be obtained by
allowing hydrogen to escape, thereby permitting entropy to predominate
over enthalpy.
‡ Addition of liquid Hg to the solution has no measurable effect on the rate
of COE formation, indicating that the catalyst is not colloidal iridium: D. R.
Anton and R. H. Crabtree, Organometallics, 1983, 2, 855.
§ Density functional theory calculations including geometry optimization
employing the B3LYP hybrid functional; effective small-core potentials
and double-zeta basis sets on Rh and Ir (LANL2DZ model); all-electron
basis sets for main group atoms: D95(d) for C, P and Cl, 6-311G(p) for
dihydrogen and hydrides, and 3-21G for other H atoms. To model the PCP
structure, the phenyl groups in the M(PR3)2Ph complexes were held
coplanar with the Ir and P atoms.
¶ Similarly, methane C–H addition to Ir(PH3)2Cl is reported to be 29
kcal mol21 more exothermic than addition to Ir(PH3)2H (241.6 and 212.8
kcal mol21, respectively): T. R. Cundari, J. Am. Chem. Soc., 1994, 116,
340.
,
Although complexes Rh(PR3)2ClH2 are classical dihy-
3
drides,14 Milstein has reported the T1 of H2Rh[h -
HC(CH2CH2PBut2)2] to be 30–60 ms (240 °C, 400 MHz)15
strongly indicating the presence of a dihydrogen ligand. The
16
closely related (PCP)RhH2 is apparently also a dihydrogen
complex: we find a T1 value of 46 ms at 293 K (500 MHz)17 and
an H–D coupling constant of 33 Hz for (PCP)RhH(D).
1 For a recent review: B. A. Arndtsen, R. G. Bergman, T. A. Mobley and
T. H. Peterson, Acc. Chem. Res., 1995, 28, 154.
We have previously reported that addition of H2 to
Rh(PPri3)2Cl [to give H2Rh(PPri3)2Cl] is highly exothermic (!
33 kcal mol21).18 The fact that H2 addition to Rh(PCP) does not
even cleave the H–H bond suggests that addition is much less
favorable. Indeed, ab initio electronic structure calculations¶
yield remarkably different reaction energies for H2 addition to
the model fragments Rh(PH3)2Cl and Rh(PH3)2Ph: 228.5 and
24.2 kcal mol21, respectively (producing dihydride and
dihydrogen complexes, respectively, in accord with observa-
tion). In the case of Ir, the additions are much more exothermic
though the difference between chloride and phenyl complexes is
similar; respective values are Ir(PH3)2Cl: 254.1 kcal mol21 and
Ir(PH3)2Ph: 223.0 kcal mol21.¶ These computational results
strongly suggest a simple explanation for the very high catalytic
activities of the Rh(PR3)2Cl and Ir(PCP) fragments as compared
with the ‘converse’ pair [i.e. Rh(PCP) and Ir(PR3)2Cl]. For Rh,
the PCP complex adds H2 too weakly. This implies not only that
transfer of H2 from alkane to rhodium is unfavorable, but it also
suggests that C–H addition may be unfavorable, since H2
normally adds much more favorably than C–H bonds. For Ir, H2
addition to the chloride complex is highly exothermic; this may
inhibit thermal loss of H2 or even transfer of H2 to a sacrificial
acceptor. Thus, the similar affinities of the Rh(PR3)2Cl and
2 R. H. Crabtree, J. M. Mihelcic and J. M. Quirk, J. Am. Chem. Soc., 1979,
101, 7738.
3 D. Baudry, M. Ephritikine, H. Felkin and R. Holmes-Smith, J. Chem.
Soc., Chem. Commun., 1983, 788.
4 M. J. Burk, R. H. Crabtree, C. P. Parnell and R. J. Uriarte,
Organometallics, 1984, 3, 816.
5 M. J. Burk, R. H. Crabtree and D. V. McGrath, J. Chem. Soc., Chem.
Commun., 1985, 1829.
6 (a) K. Nomura and Y. Saito, J. Chem. Soc., Chem. Commun., 1988, 161;
(b) T. Sakakura, T. Sodeyama, M. Tokunaga and M. Tanaka, Chem.
Lett., 1988, 263; (c) J. A. Maguire, W. T. Boese and A. S. Goldman,
J. Am. Chem. Soc., 1989, 111, 7088.
7 (a) J. A. Maguire, A. Petrillo and A. S. Goldman, J. Am. Chem. Soc.,
1992, 114, 9492; (b) K. Wang, M. E. Goldman, T. J. Emge and
A. S. Goldman, J. Organomet. Chem., 1996, 518, 55.
8 (a) M. Gupta, C. Hagen, R. J. Flesher, W. C. Kaska and C. M. Jensen,
Chem. Commun., 1996, 2083; (b) M. Gupta, C. Hagen, W. C. Kaska,
R. E. Cramer and C. M. Jensen, J. Am. Chem. Soc., 1997, 119, 840; (c)
M. Gupta, W. C. Kaska and C. M. Jensen, Chem. Commun., 1997,
461.
9 M. A. McLoughlin, R. J. Flesher, W. C. Kaska and H. A. Mayer,
Organometallics, 1994, 13, 3816.
10 NIST Standard Reference Database Number 69, 1996,
http:// webbook.nist.gov/chemistry/
11 (a) T. Aoki and R. H. Crabtree, Organometallics, 1993, 12, 294; (b) T.
Fujii, Y. Higashino and Y. Saito, J. Chem. Soc., Dalton. Trans., 1993,
517.
12 J. Belli and C. M. Jensen, Organometallics, 1993, 12,294; T. Fujii, Y.
Higashino and Y. Saito, J. Chem. Soc., Dalton Trans., 1993, 517.
13 T. Sakakura, T. Sodeyama and M. Tanaka, New J. Chem., 1989, 13,
737.
14 R. L. Harlow, D. L. Thorn, R. T. Baker and N. L. Jones, Inorg. Chem.,
1992, 31, 993.
15 A. Vigalok, Y. Ben-David and D. Milstein, Organometallics, 1996, 15,
1839.
16 S. Nemeh, C. Jensen, E. Binamira-Soriage and W. C. Kaska,
Organometallics, 1983, 2, 1442.
Ir(PCP) fragments for H2 (viz., 228.5 and 223.0 kcal mol21
)
may help explain their similarly high catalytic activities. In the
case of the Rh complex, however, H2 is needed to cleave the m-
Cl bridge and catalyst decomposition is observed at very high
temperatures ( >> 100°C).7 In the case of the Ir(PCP) unit,
formation of analogous anion-bridged dimers is not possible
and presumably the rigidity of the PCP ligand endows the
complex with resistance to decomposition that is highly unusual
for a transition-metal phosphine complex. Apparently, ligand
stability and the appropriate (but unremarkable) energetics of
oxidative addition to the Ir(PCP) fragment are the key factors
resulting in the unprecedented ability to effect efficient alkane
dehydrogenation without the use of light or a sacrificial
acceptor.
17 This value (which is not minimized) is sufficiently low to permit
characterization as a dihydrogen complex: D. G. Hamilton and
R. H. Crabtree, J. Am. Chem. Soc., 1988, 110, 4126.
18 K. Wang, G. P. Rosini, S. P. Nolan and A. S. Goldman, J. Am. Chem.
Soc., 1995, 117, 5082.
This research was supported by the Division of Chemical
Sciences, BES, OER, US Department of Energy and the US
Department of Energy Hydrogen Program.
Received in Bloomington, IN, USA; 17th July 1997; 7/05105K
2274
Chem. Commun., 1997