Journal of the American Chemical Society
Communication
Table 1. Tunneling Kinetics of s-cis,s-cis-Dihydroxycarbene
rotamerization, in agreement with the kinetic analysis. Neither
product can be monitored in our experiments because the CO2
concentration after pyrolysis exceeds our spectrometer limits
a
(
1cc) in a N Matrix
2
leak
fraction
initial 1cc:1ct
population ratio
b
and H is not IR-active. Continuously populating 1cc by laser
T/K
τct/h
τleak/h
τ /h
eff
2
irradiation of the other rotamers should lead to overall
3
7
2.5
5.0
0.0
.0
.5
0.331(1)
0.331(1)
0.312(1)
0.307(1)
0.278(1)
0.78(10)
0.73(9)
0.48(12)
0.82(22)
0.35(7)
0.363
0.359
0.325
0.331
0.291
0.13(2)
0.14(3)
0.10(7)
0.09(3)
0.19(18)
2.797(3)
2.106(3)
3.483(3)
4.255(5)
3.133(3)
depletion of 1 to form H and CO , but this was not observed
2
2
We attempted to probe the behavior of the dideuterated
1
1
2
2
2
5
species H -1 originating from the pyrolysis of H -oxalic acid
2
2
in order to further characterize 1cc and its tunneling decay
which should be quenched by dideuteration). However,
a
Standard errors of the least-squares fits are shown in parentheses in
(
multiples of the last significant digit. See the text and SI for a
2
b
bands of H -1cc were not observed after NIR excitation
because the energy of the O− H overtone of H -1tt is barely
too low (5258 cm = 15.0 kcal mol ) to overcome the
2
description of the fitting parameters. τeff = effective half-life for
2
2
2
overall 1cc decay.
−
1
−1
rotamerization barriers (1tt → 1ct = 15.9 and 1ct → 1cc = 16.1
−
1
kcal mol ).
best theoretical prediction for the 1cc → 1ct tunneling half-life
in the gas phase is 1.2 min, which was derived as follows: (a)
pCVTZ potential energy curve (Table S30) by 1.024 to
reproduce the converged FPA rotamerization barrier; (b)
WKB evaluation of the tunneling half-life on this scaled curve;
given by the B3LYP/cc-pVTZ level of theory (Table S29). The
resulting prediction is consistent with both the observed lack of
We tried to prepare monodeuterated 1 through ester
pyrolysis of several suitable precursors (Scheme S1), hoping
to use the OH group as an NIR antenna for remote
rotamerization. This would have allowed an indirect probe
of the leak process because the heavier H-1cc isotopomer
could not dissociate by QMT while rotamerization would still
occur at approximately 50% of the rate of the parent molecule.
However, in all of our attempts the carbene yield was too low
through this alternative generation method (see the SI for
details).
To conclude, we have isolated and characterized in solid N2
the higher-lying s-cis,s-cis rotamer of dihydroxycarbene, which
59
60
2
1cc bands in the Ar and Ne matrices and the measured 1cc half-
life of 17−22 min in solid N , whose appreciable polarizability
2
is known to stabilize reactive intermediates. Khriachtchev et al.
reported a similar stabilization of the high-lying HOCO·
19
may be the missing link in the reduction of CO with H to
2
2
4
form 1ct and subsequently formic acid. The 1cc species rapidly
rotamerizes even at 3 K and could be identified only in
20
to those reported for formic acid. The AE-CCSD(T)/aug-cc-
stabilizing N rather than noble gas cryogenic environments. A
secondary QMT pathway involving dissociation of 1cc to H +
CO is plausible according to theory and might be responsible
for the biexponential characteristics measured in our decay
pCVTZ-optimized complexes display O−H···N interactions
2
2
−
1
that yield binding energies of 1.2−1.8 kcal mol and
2
−
1
vibrational frequency shifts as large as 63 cm . Such
substantial interactions are also evidenced in the experimental
IR spectra of 1ct and 1tt, as both species display an O−H
2
profiles.
−
1
The QMT kinetics from 3 to 20 K was analyzed by means of
a simultaneous nonlinear least-squares fit of two 1cc and three
ASSOCIATED CONTENT
■
*
sı
Supporting Information
1ct high-quality IR bands (Table 1). The simplest integrated
rate equations capable of accurately fitting the entirety of the
kinetic data for 1cc decay and 1ct growth required four
Full-matrix IR spectra after pyrolysis of 6 in N
Ne; full matrix IR spectra after irradiation with NIR light
in N , Ar, and Ne; full assignments of 1 in Ne and N ;
2
, Ar, and
2
adjustable parameters: the rotamerization half-life (τ ), the
ct
1cc:1ct initial population ratio, a half-life for alternative 1cc
2
decay (τleak), and a branching fraction for such secondary
full PES around 1 at the FPA//AE-CCSD(T)/aug-cc-
pCVTZ level of theory; geometries with all important
bond lengths and angles; Cartesian coordinates of all
molecules on the PES; detailed results of tunneling
57
58
leaking. Unlike the cases of carbonic and oxalic acid, no
that if the rate constants for 1cc decay and 1ct growth are set
equal to one another, the spectral data cannot be fit precisely,
revealing the necessity of the “leak” component. As shown in
Table 1, leaking is responsible for roughly 10% of the decay,
which is too small to allow the parameters for this unknown
process to be pinned down to a high degree. Nonetheless, we
believe that the kinetic profiles provide compelling evidence
for a secondary, minor process of decay with a half-life less
than but on the same order of magnitude as τct.
2
of attempts to generate H
1
■
Corresponding Authors
Wesley D. Allen − Center for Computational Quantum
Chemistry and Department of Chemistry, University of
Georgia, Athens, Georgia 30602, United States; Allen
several levels of theory, including CCSD(T)/aug-cc-pVTZ//
B2-PLYP/aug-cc-pVTZ. To wit, the rate for the direct 1cc →
H + CO decomposition is in the same time regime as the
2
2
C
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX