Published on Web 03/14/2003

Conformational Isomerization of Formic Acid by Vibrational Excitation at

Energies below the Torsional Barrier
Mika Pettersson,*,† Ermelinda M. S. Maçôas,†,‡ Leonid Khriachtchev,† Rui Fausto,‡ and
Markku Räsänen†
Laboratory of Physical Chemistry, P.O. Box 55, FIN-00014, UniVersity of Helsinki, Finland,
and Department of Chemistry-CQC, UniVersity of Coimbra, P-3004-535, Coimbra, Portugal
Received November 27, 2002; E-mail: petters@csc.fi

Understanding the structural and dynamical conformational

properties of molecules is of fundamental and practical value in
chemistry. One of the most fascinating issues in this respect is the
folding of proteins, which represents a very important but simul-
taneously extremely complicated and multidimensional problem.
The most detailed information can be obtained by studying small
molecules. In this respect, formic acid (FA), being the simplest
organic acid, represents a good model system for studying the
Published on March 14, 2003 on http://pubs.acs.org | doi: 10.1021/ja0295016
Downloaded by PORTUGAL CONSORTIA MASTER on July 10, 2009

conformational properties of carboxylic acids and the carboxylic

group in general. It is known that interconversion of different
conformers can be induced in the solid state by vibrational
excitation.1 Coupling of the initially excited state with other modes
of the isolated molecule and with lattice phonons leads to the flow
of energy into the reaction coordinate, finally promoting the change
of the geometry of the molecule from one conformation to Figure 1. The quantum yield for the trans f cis isomerization of formic
another.1,2 Usually, it is natural to assume that to induce the acid via pumping various vibrational transitions. The line connecting the
conformational conversion, the initially excited state should lie well points is for guiding the eye. The position of the barrier for the isomerization
from the ab initio calculations is indicated in the figure with a vertical line,
above the reaction barrier.2,3 In this communication, we show that and the estimated uncertainty is indicated by a shaded region. The points
in FA, due to tunneling, the conversion from the lower energy trans correspond to the following vibrational modes: 4651 cm-1, νOH + COH-
to the higher energy cis form can be induced with high efficiency CO def.; 4184 cm-1, νOH + τOH; 4174 cm-1, νCH + CO-COH def.; 3552
at energies well below the reaction barrier. cm-1, νOH; 3516 cm-1, 2νCO; 2955 cm-1, νCH.
In the experiments, FA (KEBO lab, >99%) was mixed with forms is 1396 cm-1, the experimental value being 1365 ( 30 cm-1.7
argon (AGA 99.9999%) in a glass bulb with a ratio FA/Ar ) An additional factor to consider is the effect of the environment
1/2000. The mixture was deposited on a CsI window held at 15 K on the energetics. We have estimated the effect of the medium on
in a cryostat (APD DE 202 A). After deposition, the sample was the barrier height within the PCM (polarized continuum model)
cooled to 8 K. The IR pumping was carried out with pulsed (∼5 solvation model in GAUSSIAN 98 using a dielectric constant of
ns) narrowband IR radiation of an optical parametric oscillator 1.63 for argon.8 According to these results, the barrier height is
(Sunlite with an infrared extension, Continuum, fwhm ≈ 0.1 cm-1). reduced by 111 cm-1 from the gas phase to solid argon. Combining
The absolute wavelength was determined with a wavemeter these computational results, we reliably estimated the barrier height
(Burleigh WA-4500), and the pulse energy was measured with a for the solid-state trans f cis conversion in an Ar matrix to be
pulsed energy meter (Molectron). The pulse energies were 0.5-1 3810 ( 100 cm-1.
mJ, and the repetition rate was 10 Hz. The IR spectra were measured Cis FA was prepared by pumping various vibrational transitions
with a FTIR spectrometer (Nicolet 60 SX) with a resolution of 1 of trans FA.9 Cis FA decays back to trans FA in a time scale of
or 0.25 cm-1. The experimental setup was arranged in such a way minutes even at 8 K due to tunneling.9,10 In effect, under continuous
that IR spectra could be recorded during irradiation. Suitable IR pumping, a photoequilibrium is established, and the equilibrium
interference filters were used to suppress the effect of the glowbar concentrations of the two forms [cis]eq and [trans]eq are given by
radiation on the IR induced isomerization. It was also verified that the pumping efficiency and the tunneling rate for the backreaction.
the room-temperature radiation from the surroundings had a The pumping efficiency is determined by the radiation intensity,
negligible effect on the results. absorption cross section, and the quantum yield for the isomeriza-
An important value for the present work is the barrier height for tion, the first two values being experimentally measurable. The
the trans f cis conversion. There are various estimates for the backreaction rate can also be measured.10 Thus, the quantum yield
barrier in the literature, and the high level ab initio calculations for the isomerization at different excitation wavelengths can be
give ∼50 kJ/mol (4500 cm-1).4,5 We performed additional calcu- extracted.
lations within the GAUSSIAN 98 program package.6 Our value The results of these experiments are shown in Figure 1,
for the barrier height at the CCSD(T)/aug-cc-pVTZ//MP2/aug-cc- presenting the quantum yield for pumping several different
pVTZ level is 3921 cm-1 (corrected for the zero-point-energy (ZPE) fundamental and combination vibrations in the energy interval
using the calculated frequencies at the MP2 level). At the same 2900-5000 cm-1. The position of the barrier for isomerization
level of theory, the energy difference between the cis and trans determined from the ab initio calculations is given in the figure as
† University of Helsinki.
well. The isomerization efficiency remains surprisingly high, being
‡ University of Coimbra. about 20% down to ∼4100 cm-1. There is very little difference in
4058 9 J. AM. CHEM. SOC. 2003, 125, 4058-4059 10.1021/ja0295016 CCC: $25.00 © 2003 American Chemical Society
 COMMUNICATIONS

levels below them are also strongly delocalized. For instance, the
square of the wave function for the third highest state gives 84 and
16% for the probabilities for the system to be in trans or cis wells,
respectively. This state lies ∼500 cm-1 below the top of the barrier,
that is, energetically slightly below the trans OH-stretching
fundamental and the carbonyl stretching overtone. Dynamically,
this means that at these energies torsional randomization is fast,
comparable with the free rotor states, in accordance with the
observed high quantum yield for the isomerization. The next two
states are already strongly localized in either well with a probability
of ∼99%. These states lie at ∼900 cm-1 below the top of the barrier,
that is, near the CH stretching. The strong localization means
dynamically reduced tunneling probability, and it is in accordance
with the reduced quantum yield for the isomerization.
Figure 2. The calculated torsional potential of formic acid. The vibrational
energy levels and wave functions are from the variational calculations. The The obtained results together with the results of other studies
blue and red curves refer to the trans and cis wave functions, respectively. point out the importance of tunneling in the conformational
dynamics of the OH group.9,10,13,14 Moreover, it is remarkable that
the quantum yield between different pumped modes in the same the excitation energy is so efficiently transferred to the isomerization
energy region. Remarkably, the efficiency decreases only slightly coordinate. The initially excited state can relax via cascading
to about 5% at ∼3500 cm-1 (corresponding to the excitation of vibrational relaxation, and this process is possibly accompanied by
the OH stretching or carbonyl stretching overtone) despite these phonon emission.15 At some stage of the relaxation process, the
Published on March 14, 2003 on http://pubs.acs.org | doi: 10.1021/ja0295016

energies being reliably below the barrier. The quantum yield for molecule is transferred to a high overtone state of the torsion, and
Downloaded by PORTUGAL CONSORTIA MASTER on July 10, 2009

the isomerization then drops strongly about 3 orders of magnitude the isomerization can take place.
at ∼3000 cm-1 (corresponding to the C-H stretching fundamental) In summary, it has been shown that formic acid undergoes
but is still reliably measurable. Therefore, the principal result of photoinduced conformational isomerization when the photon energy
these studies is that IR induced rotamerization takes place at is significantly lower than the energy barrier for this process. The
significantly lower energies than the torsional barrier height. quantum yield for the isomerization slightly below the top of the
The natural explanation for these observations is that the barrier is comparable with the quantum yield above the barrier and
isomerization process occurs by tunneling. Indeed, we have decreases strongly at lower energies. The conformational isomer-
previously observed that cis FA tunnels quite efficiently back to ization occurs most probably via the tunneling mechanism assisted
trans FA even from the vibrational ground state.9,10 To gain more by solid-state energy relaxation.
insight into the mechanism, we have calculated the torsional Acknowledgment. This work was supported by the Academy
potential for the FA at the MP2/aug-cc-pVTZ level. At a given of Finland. E.M.S.M. acknowledges a Ph.D. student grant of the
H-O-C-H dihedral angle, the other coordinates were fully Portuguese foundation for science and technology.
optimized. The obtained potential was fitted with a cosine series.
References
n
1
∑1 Vn(1 - cos nφ)
(1) Räsänen, M.; Kunttu, H.; Murto, J. Laser Chem. 1988, 19, 123-146.
V) (2) Roubin, P.; Varin, S.; Verlaque, P.; Coussan, S.; Berset, J.-M.; Ortéga,
2 J.-M.; Peremans, A.; Zheng, W.-Q. J. Chem. Phys. 1997, 107, 7800-
7808.
For this potential, the vibrational Schrödinger equation was (3) Räsänen, M.; Murto, J.; Bondybey, V. E. J. Phys. Chem. 1985, 89, 3967-
variationally solved in the sine and cosine basis sets using the 3970.
(4) Goddard, J. D.; Yamaguchi, Y.; Schaefer, H. F., III. J. Chem. Phys. 1992,
Hamiltonian of Lewis et al.11 The inertial constant as a function of 96, 1158-1166.
the torsional angle needed for the calculation was obtained using (5) Su, H.; He, Y.; Kong, F.; Fang, W.; Liu, R. J. Chem. Phys. 2000, 113,
1891-1897.
the method of Pitzer,12 and the resulting data set was also fitted (6) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
with a cosine series to be used in the vibrational analysis. The A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
geometries needed for the calculation of the inertial constants were Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
taken from our ab initio calculations. M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
The fitted potential and the obtained energy levels and wave Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.;
Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.;
functions are shown in Figure 2. The quality of the potential can, Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
to some extent, be judged by comparing the calculated and experi- Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith,
T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.;
mental torsional frequencies. The calculated torsional frequency Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M.
for the trans FA is 614 cm-1 and for the cis FA is 513 cm-1, while W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98, revision A.9; Gaussian, Inc.: Pittsburgh, PA, 1998.
the experimental values are 635 and 503 cm-1, respectively;9 that (7) Hocking, W. H. Z. Naturforsch. 1976, 31a, 1113-1121.
is, the agreement is reasonably good. The most relevant for the (8) Hallam, H. E.; Scrimshaw, G. F. In Vibrational Spectroscopy of Trapped
Species; Hallam, H. E., Ed.; John Wiley & Sons: London, 1973.
present study are the states near the top of the barrier and the extent (9) Pettersson, M.; Lundell, J.; Khriachtchev, L.; Räsänen, M. J. Am. Chem.
of delocalization of these states. Because the calculated potential Soc. 1997, 119, 11715-11716.
(10) Pettersson, M.; Macoas, E. M. S.; Khriachtchev, L.; Lundell, J.; Fausto,
is determined at a lower level than our best estimates for the barrier R.; Räsänen, M. J. Chem. Phys. 2002, 117, 9095-9098.
height and ignoring solvation, it yields somewhat different energet- (11) Lewis, J. D.; Malloy, T. B., Jr.; Chao, T. H.; Laane, J. J. Mol. Struct.
1972, 427-449.
ics as well. To make meaningful comparisons with the higher level (12) Pitzer, K. S. J. Chem. Phys. 1946, 14, 239-243.
calculations and with experiments, we discuss the energetics with (13) Nobuyuki, A.; Kudoh, S.; Takayanagi, M.; Nakata, M. Chem. Phys. Lett.
respect to the barrier top. This is a reasonable approach because 2002, 356, 133-139.
(14) Nobuyuki, A.; Kudoh, S.; Takayanagi, M.; Nakata, M. J. Phys. Chem. A
we are mainly interested in the region near the top of the barrier. 2002, 106, 11029-11033.
It is seen in Figure 2 that the two torsional states just below the (15) Bondybey, V. E. AdV. Chem. Phys. 1981, 47, 521-533.
barrier are strongly delocalized over the two wells. The next two JA0295016

J. AM. CHEM. SOC. 9 VOL. 125, NO. 14, 2003 4059