Kinetics of photochemical alkane dehydrogenation catalyzed by Rh(PMe3)2(CO)Cl: Implications concerning the C-H bond activation step

Article abstract of DOI:10.1016/S0022-328X(97)00260-X

The mechanism of photochemical alkane dehydrogenation catalyzed by Rh(PMe₃)₂(CO)Cl has been further probed with an emphasis on characterizing the initial C-H activation step and understanding the effect of added CO on selectivity. While pure cyclooctane and pure cyclohexane are dehydrogenated at the same rate (same quantum yields), cyclooctane shows much greater reactivity in mixtures of the two solvents. The product ratio (cyclooctene:cyclohexene) is highly dependent upon the partial pressure of CO, ranging from 12 in the absence of CO, to 75 in the limit of high CO pressure (>ca. 400 torr). The kinetic isotope effect for the dehydrogenation of c-C₆H₁₂/c-C₆D₁₂ is also found to be dependent upon CO pressure, ranging from 10 in the absence of CO to 4.2 under high CO pressure. The results support our earlier conclusion that the intermediate responsible for C-H activation is ground state [Rh(PMe₃)₂Cl]. It is also concluded that inhibition of the reaction by CO operates primarily via addition of CO to the intermediate alkyl hydrides, (R)(H)Rh(PMe₃)₂Cl. Addition of CO prior to C-H bond addition is apparently not a kinetically significant process, even under high CO pressure.

Full text of DOI:10.1016/S0022-328X(97)00260-X

Ž
.
Journal of Organometallic Chemistry 554 1998 41–47
Kinetics of photochemical alkane dehydrogenation catalyzed by
ž
/ ž /
Rh PMe3 2 CO Cl: implications concerning the C–H bond activation step
)
Glen P. Rosini, Sobhi Soubra, Maxime Vixamar, Sheyi Wang, Alan S. Goldman
Department of Chemistry, Rutgers – The State UniÕersity of New Jersey, Piscataway, NJ 08855-0939, USA
Received 9 January 1997
Abstract
Ž
. Ž
.
The mechanism of photochemical alkane dehydrogenation catalyzed by Rh PMe3 CO Cl has been further probed with an emphasis
2
on characterizing the initial C–H activation step and understanding the effect of added CO on selectivity. While pure cyclooctane and
Ž
.
pure cyclohexane are dehydrogenated at the same rate same quantum yields , cyclooctane shows much greater reactivity in mixtures of
Ž
.
the two solvents. The product ratio cyclooctene:cyclohexene is highly dependent upon the partial pressure of CO, ranging from 12 in the
Ž
.
absence of CO, to 75 in the limit of high CO pressure )ca. 400 torr . The kinetic isotope effect for the dehydrogenation of
c–C6H12rc–C6D12 is also found to be dependent upon CO pressure, ranging from 10 in the absence of CO to 4.2 under high CO
w
Ž
.
x
pressure. The results support our earlier conclusion that the intermediate responsible for C–H activation is ground state Rh PMe3 2Cl . It
is also concluded that inhibition of the reaction by CO operates primarily via addition of CO to the intermediate alkyl hydrides,
Ž .Ž .
Ž
.
R H Rh PMe3 2Cl. Addition of CO prior to C–H bond addition is apparently not a kinetically significant process, even under high CO
pressure. q 1998 Elsevier Science S.A.
Keywords: Dehydrogenation; C–H bond activation step; CO pressure
1. Introduction
Several years ago we proposed the catalytic cycle shown
w x
in Fig. 1 4 .
The photochemical dehydrogenation of alkanes cat-
Key features of the mechanism of Fig. 1 include the
Ž
. Ž
.
Ž
Ž ..
alyzed by Rh PMe3 CO Cl Eq. 1 could be consid-
Ž .
following points: i cleavage of the Rh–CO bond is the
2
Ž
ered the first reported example of efficient )ca. 100
catalytic turnovers alkane functionalization catalyzed
by a soluble low-valent transition metal complex for
the first example of alkane photodehydrogenation cat-
Ž .
only photochemical step in the cycle; ii ground state
.
Ž
.
RhL2Cl LsPMe3 undergoes addition to the alkane;
Ž
Ž
.
iii addition to the alkane is reversible.
Reaction 1 is inhibited by added CO atmosphere 4 .
w x
w
x
alyzed by soluble metal complexes, 1–4 .
Clearly, the most straightforward explanation for this
observation, in terms of Fig. 1, involves addition of CO
to RhL2Cl, i.e. the reverse of the photochemical step.
The possibility that CO also inhibits dehydrogenation
via addition to the presumed rhodium alkyl hydride
intermediate was also considered. However, if CO adds
to two kinetically distinct intermediates in the cycle, the
inhibition would be expected to show an inverse sec-
h
n
alkane ™ alkeneqH2 ≠
1
Ž .
1
Since alkane dehydrogenation is a highly endother-
Ž
. w x
mic process ca. 23–32 kcalrmol 5 , the role of light
in this reaction is necessarily more than to merely
Ž
generate a reactive thermochemical catalyst via metal–
ligand bond cleavage, for example . Instead of simply
providing entry into a catalytic cycle, one or more
photochemical steps must be contained within the cycle.
.
w
x
ond-order term in CO . Since the inhibition very clearly
showed only an inverse-first order term Stern–Volmer
Ž
.
kinetics , it was assumed that CO addition to the alkyl
hydride is not a kinetically significant process 4 .
Herein we report on photokinetic studies designed to
provide a more in-depth understanding of the initial
w x
)
Corresponding author. Tel.: q1 908 4455232; Fax: q1 908
4455312; e-mail: goldman@rutchem.rutgers.edu
0022-328Xr98r$19.00 q 1998 Elsevier Science S.A. All rights reserved.

(
)
42
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
Fig. 1. Previously proposed mechanism for reaction 1.
Fig. 2. Dehydrogenation of 1:1 COArCHA under argon atmosphere.
steps in the catalytic cycle of reaction 1. The results
tion, 366 nm, the quantum yields were determined to be
Ž . Ž
.
provide further support for the points i – iii discussed
above. However, this work indicates that the assumed
mode of inhibition by CO, simple back reaction with
RhL2Cl, is not a kinetically significant process. In-
stead, addition to the rhodium alkyl hydride accounts
for the inhibition by CO. Assuming that the reaction of
free RhL2Cl with CO would be quite facile, this implies
that addition of C–H bonds to RhL2Cl is both kineti-
cally very facile and thermodynamically very favorable.
Ž . w x
0.097 4 4 . However, when a 1:1 COArCHA solvent
Ž
.
mixture was used, cyclooctene COE was formed
Ž
.
12.5-times as rapidly as cyclohexene CHE; see Fig. 2
4 . The mechanism of Fig. 1, as it would operate in a
two-solvent system, is shown in Fig. 3.
w x
In addition to such competition experiments, greater
reactivity of cyclooctane is also observed with pure
solvents under CO atmosphere. For cyclooctane, a plot
Ž
of 1rF vs PCO gives a straight line Stern–Volmer
.
dependence over the range 0–3400 torr with an inter-
cept of 9.4 and a slope of 0.080 torry1 4 . For cyclo-
w x
2. Results and discussion
hexane the rate of reaction under CO is too slow to
obtain good measurements using monochromatic light;
2.1. Photochemical dehydrogenation of cyclooctaner
Ž
.
using broad-band irradiation l)280 nm , however, a
cyclohexane
In the absence of CO atmosphere, irradiation of
Ž
.
solutions of 1 2.0 mM, l)280 nm, 508C in either
Ž
.
Ž
.
cyclooctane COA or cyclohexane CHA gives identi-
cal reaction rates for formation of the corresponding
1
Ž
.
alkene 0.27 mMrmin . With monochromatic irradia-
1 These rates were obtained with the use of a neutral density filter
Ž
.
ODs1.0; 10% transmittance . Without such a filter the rates for
cyclooctane dehydrogenation are approximately 10-fold faster, as
would be expected; however, the rate of cyclohexane dehydrogena-
tion in the absence of CO atmosphere is only about a factor of 2
Ž
.
faster and reproducibility is poor . Under 50 torr CO both solvents
show the expected 10-fold difference. The unusual intensity-depen-
dence for cyclohexane may be attributed to residual CO being present
Ž
.
during high-intensity unfiltered irradiation: either due to a higher
steady-state concentration of CO or perhaps CO liberated by product
decomposition. Cyclohexane is expected to show a much greater
Ž
.
sensitivity to residual CO, considering that 50 torr CO ca. 0.5 mM
lowers the cyclohexane dehydrogenation rate by a factor of 60 while
the cyclooctane rate is lowered only be a factor of 2. Alternatively,
the intensity-dependence for cyclohexane may be attributed to reac-
tions other than CO addition competing with dehydrogenation. For
Ž .Ž . Ž
.
example, dimerization of RhL2Cl R H followed by loss of RH , if
it occurs, would be most significant when the alkyl hydride only
slowly undergoes b-hydrogen elimination and under conditions of
very low CO pressure and high irradiation intensity.
Fig. 3. The dehydrogenation mechanism of Fig. 1, adapted to a
X
X
Ž
.
two-solvent system R scycloctyl, XsH; or R sC6 D11, XsD .

(
)
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
43
assume that b-H elimination is more favorable for the
Ž
cyclooctyl hydride kinetically andror thermodynami-
.
cally .
The greater reactivity of cyclooctane cannot reason-
ably be attributed to a greater tendency to undergo C–H
bond addition; indeed Bergman has found that the inser-
Ž
.
tion of C5Me5 IrL into C–H bonds favors cyclohexane
w x
oÕer cyclooctane by a factor of 11 6 . Furthermore, a
greater tendency to undergo C–H bond addition could
not explain the selectivity enhancement effected by CO
Ž
.
vide infra .
The selectivity for cyclooctane observed in mixed-
Ž
.
Fig. 4. Hypothetical invalid excited state mechanism for alkane
photodehydrogenation.
solvent competition experiments requires that the photo-
generated species must be capable of an initial reaction
with cyclohexane, followed by cyclohexane loss and
reaction with cyclooctane to give cyclooctene. Signifi-
cantly, this rules out a mechanism in which the species
that activates the alkane must be a photoexcited state as
in Fig. 4.
The observations noted above can all be explained in
terms of the mechanism of Fig. 1 and Fig. 3. In
contrast, however, the results of competition experi-
ments under Õarying CO pressures are not easily recon-
ciled with this mechanism. Increasing CO pressure re-
sults in increasing the COErCHE ratio, from 12.5:1 in
the absence of CO, to ca. 75:1 in the limit of high CO
plot of 1rrate vs. PCO is also found to be linear. For
direct comparison, the rates of dehydrogenation of CHA
and COA were measured under 50 torr CO and under
identical conditions; they were found to be 0.13
Ž
mMrmin and 0.0042 mMrmin respectively cf. 0.27
mMrmin for both solvents in the absence of CO atmo-
.
sphere; see footnote .
The fact that pure cyclooctane and cyclohexane are
dehydrogenated with equal quantum yields in the ab-
sence of CO is consistent with a rate-determining step
under such conditions which does not involve the alkane,
i.e., photoextrusion of CO from 1. Thus both cyclooc-
tane and cyclohexane undergo dehydrogenation with
unit efficiency after photolysis of 1 in the absence of
Ž
.
pressure Fig. 5 . According to the mechanism of Fig. 3,
however, the COE:CHE product ratio is predicted to be
w
x
Ž
Ž ..
independent of CO Eq. 2 .
w x
CO 4 . In mixed-solvent systems a competition be-
tween alkanes take place. With pure solvents under CO
atmosphere, there is a competition between a reaction
with CO and the reaction with alkane. The greater
reactivity of COA vs. CHA is probably related to the
d COE rdt
d CHE rdt
COA k2 a k4 a k5a
k
k
y4 qk5 qk k
y2 Ž .
k
Ž
.
4
5
Ž
lower dehydrogenation enthalpy of COA 23.3 kcalrmol
s
w x
.
versus 28.2 kcalrmol for cyclohexane 5 andror,
perhaps, the more crowded geometry of the eight-mem-
bered ring which could facilitate access of the metal
center to a b-hydrogen atom. In either case, we may
CHA k2 k4 k5
k
y4 a qk5a qk4 a k5a
y2 aŽ .
Ž
.
2
Ž .
The observed dependence of the COE:CHE ratio on
PCO can, however, be reconciled with a mechanism in
which the respective alkane adducts react with CO in
competition with dehydrogenation. Such a mechanism
is shown in Fig. 6.
If one assumes steady state conditions for all of the
intermediates, the mechanism of Fig. 6 yields the fol-
lowing equation for the COE:CHE product ratio.
w
x
d COE rdt
w
x
d CHE rdt
w
x
w
x
COA k2a k4 a k5a
k
y2 Žky4 q k5 . q k4 k5 q k3Žky4 q k5 . CO .
Ž
s
w
x
w
x
CHA k2 k4 k5Žky2 aŽky4 a q k5a. q k4 a k5a q k3aŽky4 a q k5a. CO .
3
Ž .
Fig. 5. Ratio of COErCHE formation as a function of CO pressure.

(
)
44
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
Ž .
w x
In the limits of low and high CO pressure, Eq. 3
reduces to Eq. 4 and Eq. 5 , respectively,
genation is linear 4 . If we consider a scheme in which
CO reacts with both RhL2Cl ky1 and RhL2Cl R H
Ž .
Ž .
Ž
.
Ž .Ž .
Ž
w
x.
k3 CO to inhibit product formation, the steady state
kinetics yields the following expression f is the pri-
mary quantum yield for CO loss :
d COE rdt
d CHE rdt
Ž
.
COA k2 a k4 a k5a
k
k
y4 qk5 qk k
y2 Ž .
k
Ž
.
1rFobs s1rfqKX ky1 CO ky2 qk k r ky4 qk5
4
5
Ž
.
s
4
5
CHA k2 k4 k5
k
y4 a qk5a qk4 a k5a
y2 aŽ .
Ž
.
2
qKX k2 k3 CO qKX ky1 k3 CO
7
Ž .
= low CO pressure
4
Ž
.
Ž .
X
d COE rdt
d CHE rdt
w
x
.
Ž .
where K sfk2 k k5 COA . According to Eq. 7 , a
4w
x
plot of 1rFobs vs. CO will be non-linear unless one or
both of the first-order terms is much greater than the
COA k2 a k4 a k5a
k
y4 qk5
Ž
.
w
x
second-order CO term. In fact, the second-order term
is actually much greater than the first of the first-order
s
CHA k2 k4 k5 k3a
ky4 a qk5a
Ž
.
Ž .
Ž .
terms since Eq. 8 can be reduced to Eq. 6 .
= high CO pressure
5
Ž .
Ž
.
2
Thus, according to the mechanism of Fig. 6 the
COE:CHE ratio is predicted to vary with increasing CO
pressure, asymptotically approaching a value different
KX ky1 k3 CO 4KX ky1 CO
= ky2 qk k r ky4 qk5
8
Ž
.
Ž .
4
5
w
x
from that observed in the low- CO limit, in accord with
Ž .
Therefore, the dominant term in Eq.
7
must be
Ž
.
experimental results Fig. 5 .
X
w
x
Ž .
K k2 k3 CO ; in other words, those terms in Eq. 7
which represent the back reaction of RhL2Cl with CO
i.e. those terms containing ky1 are kinetically in-
significant. More specifically, Eq. 10 follows from Eq.
It should be noted that the observed dependence of
w
x
w
x
selectivity on CO , particularly the saturation in CO
Ž
.
Ž
Ž . Ž .
which can be attributed to Eq. 3 reducing to Eq. 5 ,
Ž .
.
requires that Eq. 6 is obeyed in the CO-saturation
Ž .
9 .
Ž
.
regime )ca. 400 torr CO . This implies that CO
addition is the major reaction undergone by the alkyl
hydride in this regime.
2
KX k2 k3 CO 4KX ky1 k3 CO
k2 4ky1 CO
9
Ž .
10
Ž
.
k3 CO 4ky2 qk k r ky4 qk5
6
Ž .
Ž
.
4
5
As noted in the introduction, it was previously re-
ported that a plot of 1rF vs. PCO for COA dehydro-
Ž
.
Eq. 10 implies that the major reaction of RhL2Cl is
formation of alkyl hydride rather than back reaction
Ž
with CO, even in the high pressure regime ca. 400–
3400 torr . In the same regime, the major reaction of the
alkyl hydride is addition to CO Eq. 6 . Thus, reaction
of alkyl hydride with CO is kinetically much more
important than the reaction of RhL2Cl with CO.
.
Ž
Ž ..
2.2. Photochemical dehydrogenation of cyclohexane-d12
rcyclohexane
In an attempt to gain more information about the
alkane activation step, reactions were conducted using a
mixture of perprotio and perdeutero cyclohexane
Ž
.
50:50 . In the absence of CO, a very high isotope effect
is observed, kHrkD s10. This provides further evi-
dence against selectivity being determined by the rela-
tive rates of oxidative addition, and further argues
Ž .
against the mechanism of Fig. 4, since C–H D addition
Ž
isotope effects are generally small certainly much less
.
than 10 in all known cases .
The effect of added carbon monoxide on the isotope
effect was examined. As in the analogous COArCHA
Ž
competition experiments, the product ratio in this case
Ž
Fig. 6. Proposed photodehydrogenation mechanism two-solvent sys-
X
X
.
.
tem; R scycloctyl, XsH; or R sC6 D11, XsD .
CHE-h10 :CHE-d10 was found to be dependent upon

(
)
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
45
Thus a rather strongly inverse KIE is derived for
alkane loss from the species which undergoes the kineti-
cally significant reaction with CO. Inverse KIEs are not
common, but certainly not unprecedented. In general
they are observed when an HrD atom is transferred to a
site in which the vibrational frequencies are signifi-
Ž .
cantly increased, and in particular when H D is trans-
w x
ferred from a metal to a carbon atom 7 . In the case of
C–HrD elimination, inverse KIEs seem to imply en-
dothermic eliminations, presumably since significant
C–HrD bond formation is found in a later transition
Ž
w x
state 7 ; for examples of normal KIEs for reductive
Ž .
w x.
elimination of C–H D bonds see 8 . Jones has mea-
sured a KIE of 0.52 for C–H elimination of benzene for
Cp)RhL to give the h –benzene complex 7 . In the
particular case of C6 H12rC6 D12 elimination from a d8
metal center, Bergman has observed an inverse KIE of
Fig. 7. Ratio of the rates of dehydrogenation of C6 D12 rC6 H12 vs.
CO pressure.
2
Ž
. w x
Ž
.Ž
. w x
0.7 for elimination from Cp)IrL HrD cyclohexyl 9 .
Thus, taken together, these results strongly suggest that:
CO pressure, decreasing from 10 to 4.2, with saturation
Ž .
Ž .
Ž
.
i CO inhibition occurs via attack on an alkane adduct;
ii the adduct is specifically an oxidative addition
observed at ca. 400 torr CO Fig. 7 . These results can
also be interpreted in terms of Fig. 6. In this case since
the alkane adducts should react with CO at essentially
equal rates, we can assume that k3 sk3a which simpli-
fies the kinetic analysis.
product, rather than a loosely bound solvated or s-bound
species — indeed, Moore and Bergman have shown
Ž
.
that for reasons as yet unclear C6 D12 coordinates
Ž .
considerably more strongly to the Rh I center of
To account for the COA:CHA competition results, as
well as the high isotope effect observed in competition
experiments in the absence of CO, we must assume that
reductive elimination from the alkyl hydrides is much
Ž
. Ž
. w x
Cp)Rh CO ca. 1.0 kcalrmol 10 and thus the ki-
netic isotope effect for dissociation of the s-bound
alkane would presumably be normal.
Ž
Independent of the isotope effect, it may be noted
that loss of alkane is much slower than CO addition
faster than the subsequent rate-limiting step i.e. there is
a pre-equilibrium between the respective alkyl hydride
Ž
Ž .. Ž
in the saturation regime )ca. 400 torr,
.
Eq.
6
isotopomers :
2
w
x.
corresponding to ca. 4 mM CO . Based on diffusion
limitations for CO addition, an upper limit for the rate
of alkane exchange can be estimated as ca. 106 sy1; this
implies an activation barrier DG/ of ca. 10 kcalrmol
for alkane exchange which seems more consistent with
an alkyl hydride than with a s-bound or solvated
species.
ky2 4k k r ky4 qk5 or ky2
k
y4 qk5 4k k
Ž
.
Ž .
4
5
4
5
11
Ž
.
Ž
.
Ž .
In the limits of low and high CO pressures, Eq. 3
Ž
.
Ž
.
substituting CHE-d10 for COE reduces to Eq. 12 and
Ž
.
Eq. 13 , respectively
d COE-d10 rdt
A ‘qualitative’ understanding of the effect that vary-
w
x
ing CO has on product selectivity is most easily
achieved if we posit that product formation occurs via a
rate-determining irreversible reaction of the alkyl hy-
d CHE-h10 rdt
C6 D12 k2 a k4 a ky2 k5a
k
y4 qk5
Ž
k
.
.
s
Ž
dride which occurs with rate constant k4. For purposes
C6 H12 k2 k4 ky2 a k5
y4 a qk5a
Ž
of this discussion we will assume that this step is
b-hydrogen elimination; however, it can also be a com-
posite reaction such as reversible b-hydrogen elimina-
tion followed by irreversible loss of olefin, with an
oÕerall rate constant of k4. Eq. 12 and Eq. 13 then
reduce to Eq. 14 and Eq. 15 , respectively by assum-
s0.10 no CO
12
13
Ž
.
Ž
Ž
.
.
d COE-d10 rdt
d CHE-h10 rdt
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
C6 D12 k2 a k4 a k5a
k
y4 qk5
Ž .
s
.
ing ky4 <k5 in the case of 1:1 solvent mixtures.
C6 H12 k2 k4 k5
k
y4 a qk5a
Ž .
s0.24 high CO pressure
Ž
.
Ž
.
Ž
.
Dividing Eq. 12 by Eq. 13 gives the kinetic
Ž
.
isotope effect KIE for reductive elimination.
2 The solubility of CO in alkanes is estimated to be approximately
ky2rky2 a s0.42
w
x
w x
equal to that in benzene; under 800 torr CO s0.0072 M: see 11 .

(
)
46
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
d CHE-d10 rdt
K2 a k4 a
K2 k4
ments were performed with a temperature-programmed
s
s
s0.10 no CO
14
Ž
.
Ž
.
Ž
Varian 3400 using a 50-m HP-1 cross-linked methylsil-
d CHE-h10 rdt
d CHE-d10 rdt
d CHE-h10 rdt
.
icone gum phase capillary column and a flame ioniza-
k2 a k4 a
k2 k4
tion detector. Authentic cyclooctene and cyclohexene
were used to generate calibration curves which encom-
passed the experimental concentration range. Baseline
separation of the cyclohexane isotopomers was achieved;
the identity of C6 D10 was confirmed by comparison
with authentic C6 D10 generated from dedeuterogenation
of C6 D12. Irradiations were conducted using a 500 W
Hg-arc Oriel lamp, and the desired wavelengths were
s0.24 high CO pressure
Ž
.
.
15
Ž
w
x
In the low- CO limit, the alkyl hydride isotopomers
exist in a pre-equilibrium and the product ratio reflects
Ž
Ž
.
.
the equilibrium constant K2a rK2 and the ratio of
w
x
rates of b-H elimination k4a rk4 . In the high- CO
Ž
achieved using filters obtained from FJ Gray Corning
7-83 for 366 nm and 0-56 for l)280 nm .
limit, the alkyl hydride most frequently reacts with CO
.
Ž
.
unproductively ; elimination of alkane is kinetically
insignificant. In this case, the product ratio reflects the
3.2. Typical photolysis conditions
Ž
.
relative rates of C–H addition k2a rk2 and the relative
Ž
.
rates of b-H elimination k4a rk4 . Thus both limits are
dependent upon the relative rates of b-H elimination;
however, in the high-pressure case b-H elimination is
predominantly in competition with CO addition, while
in the low-pressure case it occurs predominantly in
competition with elimination or exchange of alkane.
Stock solutions of 2.0 mM 1 were prepared in cy-
clooctane, cyclohexane, and d12-cyclohexane. In gen-
eral, 1.5-ml samples were prepared using the appropri-
ate mixture of stock solutions. The samples were placed
in a quartz cuvette sealed to a ballast used to maintain
constant partial gas pressures and equipped with ports
for attachment to a vacuum line and for removal of
microliter samples for GC analysis. The samples were
2.3. Conclusions
Ž
placed under a total pressure of 800 torr mixtures of Ar
Varying CO pressure significantly affects the inter-
.
and CO , and then irradiated at 508C. The samples were
Ž
molecular selectivity cyclooctanercyclohexane or
then analyzed at various times by GC for alkene forma-
tion. Product ratios COE:CHE and CHE-d10 :CHE-h10
were found to remain constant with time.
.
C6 D12rC6 H12 of 1-catalyzed alkane dehydrogenation,
with saturation being observed at ca. 400 torr CO. In
conjunction with previous mechanistic studies, this ef-
fect can only be explained by a mechanism in which
CO attacks the respective alkane adducts — even if CO
addition to the different adducts is assumed to occur at
Acknowledgements
Ž
equal rates as is the case in the C6 D12rC6 H12 compe-
We thank the Division of Chemical Sciences, Office
of Basic Energy Sciences, Office of Energy Research,
US Department of Energy for support of this work.
A.S.G. thanks the Camille and Henry Dreyfus Founda-
tion for a Teacher Scholar Award and the Alfred P.
Sloan Foundation for a Research Fellowship.
.
tition . An inverse isotope effect for loss of alkane from
the adduct is calculated, kh-12rkd-12 s0.42; this strongly
Ž
.
suggests that the adducts are alkyl hydrides deuterides
rather than solvated or s-bound species. Since the reac-
tion of CO with other species such as solvated or
s-bound RhL2Cl does not occur to a kinetically signifi-
cant extent, it may be inferred that such species are not
present in significant concentration relative to the alkyl
hydrides. This implies that formation of the alkyl hy-
drides, i.e. oxidative addition to the RhL2Cl photo-
product, is both kinetically very facile and thermody-
namically very favorable.
References
w x Ž .
1
a
M.J. Burk, R.H. Crabtree, D.V. McGrath, J. Chem. Soc.,
Ž .
Ž
.
Chem. Commun. 1985 1829. b M.J. Burk, R.H. Crabtree, J.
Ž
.
Am. Chem. Soc. 109 1987 8025.
w x
Ž
.
2
K. Nomura, Y. Saito, J. Chem. Soc. Chem. Commun. 1988
161.
w x
3
T. Sakakura, T. Sodeyama, M. Tokunaga, M. Tanaka, Chem.
3. Experimental
Ž
.
Lett. 1988 263.
w x Ž .
4
a
J.A. Maguire, W.T. Boese, A.S. Goldman, J. Am. Chem.
Ž
.
Ž .
Soc. 111 1989 7088.
b J.A. Maguire, W.T. Boese, M.E.
3.1. General procedures
Ž
.
Goldman, A.S. Goldman, Coord. Chem. Rev. 97 1990 179.
D.R. Stull, E.F. Westrum, G.C. Sinke, The Chemical Thermo-
dynamics of Organic Compounds, Robert E. Kreiger Publishing,
Malabar, FL, 1987.
w x
5
All samples were prepared under an argon atmo-
sphere in a Vacuum Atmospheres Dry-Lab glovebox.
Solvents were distilled under vacuum from NaK alloy.
w x
6
Ž
.
A.H. Janowicz, R.G. Bergman, J. Am. Chem. Soc. 105 1983
3929.
Ž
. Ž
.
Ž .
Rh PMe3 CO Cl 1 was prepared as described previ-
2
w x
7
Ž .
W.D. Jones, F.J. Feher, J. Am. Chem. Soc. 108 1986 4814.
w x
ously 4 . Gas chromatographic analyses for all experi-

(
)
G.P. Rosini et al.rJournal of Organometallic Chemistry 554 1998 41–47
47
w x Ž .
Ž .
w
w
x
8
a L. Abis, A. Sen, J. Halpern, J. Am. Chem. Soc. 100 1978
10 A.A. Bengali, R.H. Schultz, C.B. Moore, R.G. Bergman, J. Am.
Ž .
2915. b R.A. Michelin, S. Faglia, P. Uguagliati, Inorg. Chem.
Ž
.
Chem. Soc. 116 1994 9585.
Ž
.
x
22 1983 1831.
11 Linke, W.F., in Solubilities of Inorganic and Metal Organic
Compounds, 4th Edition, 1958, Vol. 1, D. Van Nostrand,
Princeton, NJ, p. 453.
w x
9
J.M. Buchanan, J.M. Stryker, R.G. Bergman, J. Am. Chem.
Ž
.
Soc. 108 1986 1537.