f the aby thry ,e rotatas hy
indicated that the true equilibrium structure was indeed planar,with C2v symmetry. But the very at nature of the bendingpotential function still implied the presence of large-amplitudeout-of-plane motions. As pointed out by Petersen and Klemperer,this system is of fundamental interest because we can think ofC2H2H2O as a rst step in the microsolvation of acetylene.
2 2 2
acetylene asymmetric CH stretch, the water symmetricOH stretch, and the water asymmetric OH stretch. The planarOH bonded C2v structure determined for acetylenewater meansthat its A rotational constant is expected to be relatively large,and roughly equal to the water B-value (i.e. 14.5, 9.1, or 7.3 cm1
for H2O, HDO, or D2O). But so far there has been no direct experi-mental determination of A. Levels with Ka = 0 and 1 represent dis-tinct nuclear spin modications, correlating with para- and ortho-H2O (or ortho- and para-D2O), and even in a cold molecular beamenvironment population is expected to remain in Ka = 1 since spin
Corresponding author. Fax: +1 403 289 3331.E-mail addresses: email@example.com (N. Moazzen-Ahmadi), robert.mck-
Journal of Molecular Spectroscopy 272 (2012) 1922
Contents lists available at
firstname.lastname@example.org (A.R.W. McKellar).water oxygen atom, with the heavy atoms in a linear congurationand an effective hydrogen bond length of 2.23 . The calculationsof Frisch et al.  suggested that the equilibrium structure mightbe bent, with the water H atoms out of the plane, but Petersenand Klemperer  concluded from their spectra that the effectivestructure was planar. The rst (and only previous) gas-phase infra-red spectra of acetylenewater were reported in 1992 by Blocket al. , together with further ab initio calculations. They arguedthat the complex was probably quasiplanar, with a relatively smallbarrier located at the planar geometry. Later, however, high levelcalculations  incorporating basis set superposition corrections
for C2H2H2O because it is free of complications due to Fermi res-onance. Although our results are generally consistent with earlierstudies, they raise a new mystery: the expected Ka = 1 transitions,which are present for C2D2HDO, appear to be completely missingfor C2D2H2O and D2O.
The rotational spectra studied by Petersen and Klemperer using the molecular beam electric resonance technique involvedfour isotopic forms: C2H2H2O, C2H2D2O, C2D2H2O, and C2D2D2O. The infrared spectra studied by Block et al.  using optother-mal detection with a molecular beam involved only the principalspecies, C H H O, but three different bands were observed: theHigh resolutionVan der Waals
In 198384, pioneering studies omolecular complex were publisheddifferent techniques: ab initio theorspectroscopy , and gas-phase purIt became evident that acetylene w0022-2852/$ - see front matter 2011 Elsevier Inc. Adoi:10.1016/j.jms.2011.12.003cetylenewater binaryee groups using ratherargon matrix isolationional spectroscopy .drogen-bonded to the
In the present paper, we report a study of the spectra of C2D2H2O, HDO, and D2O complexes in the 4.1 lm region of the m3fundamental band (asymmetric CD stretch) of C2D2. This repre-sents the rst gas-phase infrared observation of C2D2water com-plexes, and the rst observation of acetyleneHDO in any spectralregion. The results provide a measure of the vibrational red-shift ofthe OH vibration (OD in this case) which is less ambiguous thanInfraredDimer
C2D2HDO, but there is a puzzling absence of Ka = 1 for C2D2H2O and C2D2D2O. 2011 Elsevier Inc. All rights reserved.Infrared spectra of acetylenewater comand C2D2D2O
Mojtaba Rezaei a, N. Moazzen-Ahmadi a,, A.R.W. McaDepartment of Physics and Astronomy, University of Calgary, 2500 University Drive Nb Steacie Institute for Molecular Sciences, National Research Council of Canada, Ottawa,
a r t i c l e i n f o
Article history:Received 17 November 2011In revised form 12 December 2011Available online 29 December 2011
a b s t r a c t
Infrared spectra of C2D2wband using a tunable diodered shifts (27.7 to 28.0C2H2 vibration thanks to tbroadening leads to estimfor C2D2H2O, HDO, and
Journal of Molec
journal homepage: wwll rights reserved.xes: C2D2H2O, C2D2HDO,
West, Calgary, Alberta, Canada T2N 1N4rio, Canada K1A 0R6
r complexes are studied in the 4.1 lm region of the C2D2 m3 fundamentaler source to probe a pulsed supersonic slit jet. Relatively large vibrational1) are observed which are more easily interpretable than for the analogousbsence of Fermi resonance effects for C2D2. Noticeable homogeneous lineof upper state predissociation lifetimes of about 0.5, 0.9 and 1.1 nsO, respectively. Transitions involving Ka = 0 and 1 levels are observed for
lsevier .com/ locate / jms
achieve in terms of isotopic discrimination between H2O andD2O. In the upper trace, the strongest lines belong to C2D2D2O,
as marked by asterisks in Fig. 2. There is a corresponding Ka = 1Q-branch feature at 2418.353 cm1. However, no Ka = 1 transitionswere observed for C2D2H2O or C2D2D2O, a puzzling fact whichremains unexplained and is further discussed below.
2.2. Fitting of molecular parameters
The observed line positions and their assignments are listed in
Wavenumber / cm-12417.5 2418.0 2418.5 2419.0 2419.5
Fig. 1. Observed spectra showing the assignment of simple P- and R-branchtransitions to C2D2H2O (bottom), C2D2HDO (middle), and C2D2D2O (top). Theupper experimental trace contains C2D2D2O and HDO transitions, but not H2O,while the lower trace has C2D2H2O and HDO, but not D2O. Note the broader linewidths of the C2D2H2O transitions. The asterisk marks the C2D2HDO Ka = 1 Q-branch.
Wavenumber / cm-1
2418.4 2418.6 2418.8 2419.0 2419.2 2419.4
2417.2 2417.4 2417.6 2417.8 2418.0 2418.2
* * * * * *
Q* R(1) R(2) R(3)
Fig. 2. Expanded view showing Ka = 1 transitions of C2D2HDO (this is the samespectrum as the upper trace of Fig. 1). The gap at 2418.32 cm1 is obscured by C2D2monomer absorption (the P(12) line).
ularwhile C2D2HDO lines are also prominent and C2D2H2O linesare essentially absent. In the lower trace, the overall signal is lower(as indicated by the higher noise level), C2D2D2O is essentially ab-sent, and the strongest lines (in terms of peak height) are those ofC2D2HDO, followed closely by C2D2H2O. Since the C2D2H2Olines are almost twice as broad, its abundance here must actuallybe almost twice that of C2D2HDO. The distinct C2D2H2O and HDO lines are more obvious in the R-branch region (2418.62419.6 cm1) than in the P-branch where they partially overlap.
The prominent lines in both traces of Fig. 1 all have Ka = 0, asproven by the presence of R(0) and P(1) lines and the absence ofstrong Q-branches. However, there is a weak Ka = 1 subband whichrelaxation is relatively slow. (The ortho/para distinction is absentfor complexes containing HDO, but even here we might expectthe Ka = 1 population to be larger than expected on the basis ofthe rotational temperature, due to relatively slow rotational relax-ation among widely-spaced Ka levels.) Thus Petersen and Klemper-er detected both Ka = 0 and 1 levels, including R-branch microwave(e.g. JKaKc = 101000,211110) and Q-branch radio frequency (e.g.110111) ones. Similarly, Block et al.  detected Ka = 0 and 1 tran-sitions, though these were separately resolved only in the watersymmetric OH stretch a-type band. For the water asymmetricOH stretch b-type band, the Ka = 10 subband was observed,but, curiously, not the 01 and 21 subbands.
2. Results and analysis
2.1. Observed spectra
The spectra were recorded using a pulsed supersonic slit-jetapparatus at the University of Calgary as described previously[6,7]. The expansion gas was a dilute mixture of C2D2 (0.4%) inhelium with a backing pressure of about 7 atmospheres. To thiscould be added small additional amounts of H2O and or D2O. PGO-PHER was used for simulation and tting of the spectrum . Notethat we expect a-type parallel bands (DKa = 0) for C2D2H2O andC2D2D2O in the C2D2 in the m3 region (asymmetric CD stretch)since the a inertial axis of the complex coincides with the acetyleneaxis. The same is true for C2D2HDO except that there could also beweak b-type perpendicular transitions since the two axes are notquite coincident.
The C2D2water spectra were rst detected during C2D2 dimerstudies  even though no water had been added to the expansiongas. Two overlapping linear-molecule-like patterns were observednear 2418 cm1, with effective values of B00 corresponding to about2420 and 2516 MHz. After some checking, we realized that the for-mer agreed perfectly with the known value of (B + C)/2 for C2D2D2O , and that the latter was entirely reasonable for C2D2HDO since the C2D2H2O value is 2624 MHz . But there wasno sign of C2D2H2O in this spectrum. Where did the HDO andD2O come from? Evidently, there were signicant D2O and HDOimpurities in the C2D2 sample gas, and/or there was isotope ex-change between C2D2 and trace amounts of H2O in our gas system.Assignment of the two observed bands was easily conrmed bydeliberately adding H2O and/or D2O to the expansion gas, but itproved surprisingly difcult to get good spectra of C2D2H2O. Partof the problem was that the C2D2H2O lines were signicantlybroader than those of the HDO and D2O species, with a correspond-ing decrease in peak height.
The two spectra in Fig. 1 illustrate the best we were able to
20 M. Rezaei et al. / Journal of Moleccan be assigned to C2D2HDO, as illustrated in Fig. 2. Each strongC2D2HDO line with Ka = 0, except for R(0) and P(1), has a weakshoulder feature located about 0.02 cm1 lower in wavenumber,*D2O
Spectroscopy 272 (2012) 1922Supplementary data section. Molecular parameters resulting fromthe analysis of these positions are given in Table 1. In the cases ofC2D2H2O and D2O, microwave ground state parameters were
ularavailable . We experimented with combined microwave andinfrared ts and concluded that it was perfectly satisfactory toleave the ground state parameters xed at the literature values.Since only Ka = 0 transitions were observed here, there were threedeterminable parameters: the band origin and the upper staterotational and centrifugal distortion constants, (B0 + C0)/2 and D0J .We used an estimate of 12 cm1  for the A-value of C2D2H2O,and appropriately scaled this for the other isotopic species. Allthe present t results were insensitive to A over a wide range of as-sumed values.
In the case of C2D2HDO, there were no previous data. Evidenceof asymmetry doubling was observed for P- and R-branch transi-tions with Ka = 1 and J greater than 4 or 5. In principle this is a mea-sure of the parameter (B C), but in practice the doubling was notsufciently well-resolved and so we simply xed (B C) at an esti-mated value. This left ve variable parameters primarily deter-mined by Ka = 0 transitions: the band origin and the upper andlower state rotational and centrifugal distortion constants,(B + C)/2 and DJ. In addition, there were three parameters deter-mined by the Ka = 1 transitions: the band origin difference, givenby (A0A00), and the Ka-dependent corrections to (B + C)/2, givenby D0JK and D
The ts were quite satisfactory, with rms deviations of less than0.0005 cm1 for C2D2D2O and for the Ka = 0 transitions of C2D2
Table 1Spectroscopic parameters for C2D2water complexes (values in cm1).a
C2D2H2O C2D2HDO C2D2D2O
m0 2418.5781(6) 2418.3705(4) 2418.2597(2)A0 12.0 7.51227(42) 6.01(B0 + C0)/2 0.0876373(68) 0.083907(38) 0.0806856(109)(B0 C0) 0.000707 0.0009 0.0010979D0JK 5.665 105 2.1(27) 105 5.8184 105D0J 16.(12) 107 4.4(36) 107 1.6(11) 107A00 12.0 7.53 6.01(B00 + C00)/2 0.087531188 0.083922(40) 0.080722194(B00 C00) 0.000707 0.0009 0.00109786D00JK 5.665 105 4.5(29) 105 5.8184 105D00J 2.669 107 5.3(37) 107 0.23 108
a Uncertainties (1r) are given in units of the last quoted digits. Parameterswithout uncertainties were xed at the indicated values, which are taken from Ref. except for the (B C) values for C2D2HDO, and all the A values, which areapproximate estimates. The t results were very insensitive to these estimatedvalues.
M. Rezaei et al. / Journal of MolecHDO. For C2D2H2O, the deviation was about 0.0009 cm1 andthere was some evidence for small random perturbations but itwas difcult to be sure because of the larger line widths in thiscase. In the case of the C2D2HDO transitions with Ka = 1, the devi-ation of about 0.0009 cm1 can be accounted for by the weaknessof the observed lines and the presence of incompletely resolvedasymmetry doubling.
2.3. Discussion of molecular parameters
The observed band origins correspond to vibrational red-shiftsof 20.666, 20.873, and 20.984 cm1 for C2D2H2O, HDO,and D2O, respectively, relative to the free C2D2 molecule m3 origin. There is a bit of ambiguity in the corresponding shift forC2H2H2O as determined by Block et al.  because of a large acci-dental Fermi resonance between m3 and m2 + m4 + m5 in C2H2[11,12]. As it turns out, their value of 33.98 cm1 is considerablylarger in magnitude than the shift of about 27.9 cm1 that wouldbe predicted by simply scaling our value using the ratio of C2H2 andC2D2 m3 frequencies. The C2H2H2O shift was based on adeperturbed C2H2 m3 origin (3288.66 cm1) as a reference point,which is entirely reasonable. Interestingly, shifts of 27.22 and40.16 cm1 are obtained using the actual (perturbed) C2H2 Fermidoublet band origins of 3281.90 and 3294.84 cm1 as referencepoints, the former in good agreement with our scaled predictionfrom C2D2H2O.
Signicantly larger red shifts were observed by Engdahl andNelander  for acetylenewater complexes isolated in an argonmatrix: 48.7 and 30.5 cm1 for C2H2H2O and C2D2H2O,respectively.
The observed changes in the rotational constant, (B + C)/2, dueto excitation of the C2D2 m3 vibration are small: +3.2, 0.5, and1.1 MHz for C2D2H2O, HDO, and D2O, respectively. For com-parison, the observed change in (B + C)/2 due to excitation of m3for C2H2H2O was +0.5 MHz . The change in sign as a functionof isotopic substitution is interesting and slightly unusual. As pos-sible explanation, consider that there may be two competing ef-fects here. The rst is normal anharmonicity associated with astretching vibration, which tends to give a negative change. Butwe know from the vibrational red shift that the upper state is sig-nicantly (21 cm1) more strongly bound than the ground state,and this second effect could tend to reduce the intermolecular sep-aration, giving a positive change in the rotational constant. The lat-ter effect could be more important for the lighter C2D2H2Ospecies, giving the observed trend.
The remaining signicant result from Table 1 is the observedchange in A due to excitation of m3 for C2D2HDO, which is0.018 cm1, or 532 MHz. The implication here is that m3 excita-tion may increase the amplitude of the low frequency bending mo-tions of water relative to acetylene in the complex, leading to asmall reduction in A. But we do not know much about the A rota-tional constant, so its difcult to be more illuminating. For com-parison, the analogous change in A observed for C2H2H2O was0.04 cm1 .
2.4. Line widths
All the present observed C2D2water transitions showed linewidths greater than the normal instrumental resolution(50 MHz), which is given by a combination of laser jitter andresidual Doppler broadening caused by the nonorthogonality ofthe laser beam and the expanding supersonic jet. We estimatethe extra broadening to be about 300, 170, and 150 MHz, forC2D2H2O, HDO, and D2O, respectively (full width at half maxi-mum, Lorentzian prole), corresponding to lifetimes of 0.53, 0.94,and 1.06 ns. In the case of C2D2HDO, there was some evidencefor J-dependence, with slightly smaller widths for higher-J transi-tions. The extra widths can be explained by predissociation dueto the nite lifetime of the excited upper state. Block et al.  re-ported an even larger width of 1200 MHz for C2H2H2O in theC2H2 m3 band region, whereas transitions were sharp and instru-ment-limited in the H2O m1 and m3 band regions. The reason ofcourse is that the C2H2 (or C2D2) m3 vibration couples directly tothe dissociation coordinate of the complex, whereas the H2O vibra-tions are only weakly linked to dissociation. The fact that thewidths decrease (lifetimes increase) with increasing D substitutionis expected, and arises because the CH vibrational amplitude andzero-point energy are lowered.
2.5. Absence of Ka = 1 for C2D2H2O and D2O
It is the lack of any Q-branch features which most convincinglymarks the complete absence of Ka = 1 in our C2D2H2O and D2Ospectra, since P- and R-branch lines with Ka = 1 might be obscuredby those with Ka = 0. Why is Ka = 1 present for C2D2HDO but not
Spectroscopy 272 (2012) 1922 21for C2D2H2O and D2O? The fact that Ka = 0 and 1 are distinct nu-clear spin species for the symmetric water molecules, but not forHDO, does not help here. In fact, it deepens the mystery because
spin symmetry should slow the thermal relaxation of the Ka-ladderin the supersonic expansion, leading to more Ka = 1 population.
Note that Block et al.  easily observed Ka = 1 for C2H2H2Owith a population approximately equal to that of Ka = 0. Theireffective rotational temperature (the J-temperature) was about5 K, as compared to our values of around 34 K, depending onthe amount of water in the gas mix. This temperature differencecannot explain the absence of Ka = 1 for C2D2H2O in our spectra,particularly since the effective Ka- temperature should be consid-erably higher. In the case of C2D2D2O, the Ka = 1 spin weight is lessfavorable (weights are 2:1 in D2O, and 1:3 in H2O, for Ka = 0:1). Butthis would be more than compensated for by the smaller A-value ofC2D2D2O if spin relaxation were rapid. To summarize, if spinrelaxation is normal (i.e. slow) in our supersonic expansion, thenwe should easily observe Ka = 1 for C2D2H2O. But if spin relaxationwas somehow super fast in our expansion, then we should at leaststill see Ka = 1 for C2D2D2O.
The conclusion appears to be that Ka = 1 levels in C2D2H2O andD2O, but not C2D2HDO, are massively perturbed and/or predisso-ciated in the m3 excited state (predissociation being a kind of per-turbation, or course). It seems amazing that there could be aviolent perturbation of Ka = 1 when few or no effects are presentfor Ka = 0. Of course Ka = 0 and 1 correspond to different nuclearspin states for the symmetric isotopologues, so the available back-
a cleaner determination of the vibrational red shift induced in acet-ylene by complexation with a water molecule, together withparameters which could aid in a search for microwave spectra ofC2D2HDO. No Ka = 1 transitions were detected for C2D2H2O andD2O. This is a signicant anomaly for which the only explanationseems to be the presence of large perturbations which, however, donot affect C2D2HDO.
We thank L. Murdock for technical assistance. The nancial sup-port of the Natural Sciences and Engineering Research Council ofCanada is gratefully acknowledged.
Appendix A. Supplementary data
Supplementary data for this article (Tables A1A3) are availableon ScienceDirect (www.sciencedirect.com) and as part of the OhioState University Molecular Spectroscopy Archives (http://library.o-su.edu/sites/msa/jmsa_hp.htm). Supplementary data associatedwith this article can be found, in the online version, atdoi:10.1016/j.jms.2011.12.003.
22 M. Rezaei et al. / Journal of Molecular Spectroscopy 272 (2012) 1922ground perturbing states are completely different. But then why isnot C2D2HDO also affected? Worth mentioning in this context isanother K-related acetylenewater mystery, the apparent absenceof Ka = 01 and 21 subbands in the C2H2H2O spectrum accompa-nying the H2O m3 band .
In conclusion, we report here observations of the weakly-boundcomplexes C2D2H2O, HDO, and D2O in the region of the m3vibration of C2D2 (2400 cm1), obtained using a tunable diode la-ser to probe a pulsed supersonic jet expansion. These are the rstreported infrared spectra of these species, and the rst observationof an acetyleneHDO complex in any region. The results are ingood general agreement with previous investigations, and provideReferences
 M.J. Frisch, J.A. Pople, J.E. Del Bene, J. Chem. Phys. 78 (1983) 4063. A. Engdahl, B. Nelander, Chem. Phys. Lett. 100 (1983) 129. K.I. Petersen, W. Klemperer, J. Chem. Phys. 81 (1984) 3842. P.A. Block, M.D. Marshall, L.G. Pedersen, R.E. Miller, J. Chem. Phys. 96 (1992)
7321. D. Tzeli, A. Mavridis, S.S. Xantheas, J. Chem. Phys. 112 (2000). M. Dehghany, M. Afshari, Z. Abusara, C. Van Eck, N. Moazzen-Ahmadi, J. Mol.
Spectrosc. 247 (2008) 123. M. Dehghany, M. Afshari, R.I. Thompson, N. Moazzen-Ahmadi, A.R.W. McKellar,
J. Mol. Spectrosc. 252 (2008) 1. PGOPHER, a Program for Simulating Rotational Structure, C.M. Western,
University of Bristol, UK, . C. Lauzin, N. Moazzen-Ahmadi, A.R.W. McKellar, J. Mol. Spectrosc. 269 (2011)
124. S. Ghersetti, K. Narahari Rao, J. Mol. Spectrosc. 28 (1968) 27. W.J. Lafferty, A.S. Pine, J. Mol. Spectrosc. 141 (1990) 223. J. Vander Auwera, D. Hurtmans, M. Carleer, M. Herman, J. Mol. Spectrosc. 157
Infrared spectra of acetylenewater complexes: C2D2H2O, C2D2HDO, and C2D2D2O1 Introduction2 Results and analysis2.1 Observed spectra2.2 Fitting of molecular parameters2.3 Discussion of molecular parameters2.4 Line widths2.5 Absence of Ka=1 for C2D2H2O and D2O
3 ConclusionsAcknowledgmentsAppendix A Supplementary dataReferences