Hydroxymethylene (HCOH) and its d1-isotopologue (HCOD) are isolated in low temperature helium nanodroplets following pyrolysis of glyoxylic acid. Transitions identified in the infrared spectrum are assigned exclusively to the trans-conformation based on previously reported anharmonic frequency computations [P. R. Schreiner, H. P. Reisenauer, F. C. Pickard, A. C. Simmonett, W. D. Allen, E. Mátyus, and A. G. Császár, Nature453, 906 (2008); L. Koziol, Y. M. Wang, B. J. Braams, J. M. Bowman, and A. I. Krylov, J. Chem. Phys.128, 204310 (2008)]. For the OH(D) and CH stretches, a- and b-type transitions are observed, and when taken in conjunction with CCSD(T)/cc-pVTZ computations, lower limits to the vibrational band origins are determined. The relative intensities of the a- and b-type transitions provide the orientation of the transition dipole moment in the inertial frame. The He nanodroplet data are in excellent agreement with anharmonic frequency computations reported here and elsewhere, confirming an appreciable Ar-matrix shift of the OH and OD stretches and strong anharmonic resonance interactions in the high-frequency stretch regions of the mid-infrared.

Hydroxymethylene (HCOH), a singlet carbene involved in the photochemistry of its tautomer, formaldehyde,1–5 has been implicated in the formation of simple sugars (i.e., glycolaldehyde) in the cosmos,6,7 and has been the subject of numerous theoretical studies.6,8–13 Despite the large theoretical interest, experimental studies are lacking due to the highly reactive, transient nature of HCOH. The first indirect measurement of HCOH was reported by Sodeau and Lee.14 Low temperature matrix FT-IR and photolysis experiments were performed on isotopically labeled formaldehyde (FA), where it was postulated that formation of glycolaldehyde resulted from the reaction between HCOH and FA. The heat of formation of hydroxymethylene has been measured in photodissociation experiments of hydroxymethyl radicals,15,16 and furthermore, the energetics associated with cis- and trans-hydroxymethylene were reported using a sliced velocity map imaging technique.17 The first direct observation of HCOH (and HCOD) was reported recently by Schreiner et al.,18 where the pyrolysis products (vide infra) of glyoxylic acid (GA) were deposited in an 11 K Ar matrix. The resulting FT-IR spectrum was assigned to trans-HCOH(D) based on high-level anharmonic frequency computations. Interestingly, at 11 K, hydroxymethylene rearranged to formaldehyde via tunneling with a half-life of approximately 2 h, consistent with the computed 1.3 eV barrier to isomerization.18 Concurrently with the study by Schreiner and co-workers, the anharmonic frequencies and intensities were reported for both cis- and trans-HCOH and HCOD, using vibrational configuration interaction (VCI) theory.19 Both sets of anharmonic frequencies are in agreement with the matrix FT-IR spectra, with the exception of the OH and OD stretches, which deviate from computed values by ∼70 and ∼40 cm−1, respectively. We have isolated HCOH and d1-hydroxymethylene (HCOD) within helium nanodroplets and report here the rovibrational spectra in the CH and OH(D) stretching region.

Helium nanodroplet isolation spectroscopy was carried out using a custom apparatus that has been discussed in detail elsewhere.20,21 HCOH was formed via pyrolysis of glyoxylic acid, based on the method described by Schreiner et al.18 Experimental details, mass spectra of the HCOH doped droplet beam, and a survey vibrational spectrum through the CH and OH stretching regions are presented in the supplementary material.22 The ν1 band of formaldehyde is observed near 2785 cm−1,23 and two intense bands centered at 3494.10 and 3580.36 cm−1 (FWHM ∼0.6 cm−1) are assigned to the OH stretches of the two trans-conformations of GA based on previous infrared studies.24 The lower frequency band is assigned to the hydrogen bonded cis-configuration (tc-GA) of the OH group, while the higher frequency component is assigned to the “all-trans” structure (tt-GA). In addition, three sets of lines, split by ∼5.82 cm−1 (indicated by * in Figure S1 of the supplementary material),22 are also present in the survey spectrum. The two lower frequency sets of lines at ∼2700 and 2775 cm−1 are located near the previously measured vibrational band origins for trans-HCOH (tHCOH).18 Based on the variational computations carried out by Schreiner et al.,18 these two bands are assigned to a mixture of ν2 (CH stretch) and the ν3+ν4 combination band (in-phase HCO and COH bend + CO stretch). A pair of lines is observed in the OH stretching region blue shifted by ∼40 cm−1 compared to the Ar matrix FT-IR data.18 Anharmonic frequencies obtained from both variational18 and VCI19 computations predict a similar blue shift with respect to the matrix value, and it was inferred18 that the matrix imparted a non-negligible shift of the OH stretch. Based on these previous computations, as well as comparison with the experimental vibrational band origins,18,19 we assign these sets of lines to tHCOH.

High resolution scans of the transitions derived from tHCOH are presented in Figure 1. At the 0.4 K droplet temperature, the entire tHCOH population collapses to the 000 rovibrational ground state. Based on computed normal modes, the CH and OH stretches have transition dipole moment projections along the inertial a- and b-axes (Figure 1(a) inset), resulting in a/b hybrid bands split into a-type (101 ← 000) and b-type (111 ← 000) transitions separated by the difference in the A and B rotational constants. In general, the b-type component is broader than the a-type component, consistent with the larger droplet state density to which the rotor can couple upon excitation to the 111 level.25 The experimentally determined (A – B) values are collected in Table I, and compared to the computed (CCSD(T)/cc-pVTZ) ground state values; these are reduced by approximately 30%, which is typical for molecules isolated in superfluid He.26,27 From these computations, a lower limit to the vibrational band origin (000 ← 000) can be estimated by subtracting the computed (B + C) value (CCSD(T)/cc-pVTZ) from the centroid of the a-type transition, and these are indicated by the arrows in Figure 1 and collected in Table I. In addition to these three bands, we measured a single transition at 1057.48 cm−1, falling near the computed and experimental Ar-matrix ν6 band origin.18,19 A single c-type transition is expected for this a band, and we estimate a lower limit for the band origin to be 1046.69 cm−1 using the computed A + B rotational constant. A comparison of the experimentally determined vibrational band origins from the He droplet spectra to those computed and obtained from the matrix FT-IR spectra are collected in Table I. The band origins measured in this study are all within 20 cm−1 of the computed values. The satisfactory agreement between the computed and He droplet band origins is indicative of negligible solvent shifts, which is expected on the basis of a number of previously reported spectra of He-solvated hydrocarbon molecules.20,26 Because the low temperature of the droplets results in a sparse number of rovibrational lines, it is not possible to extract the true band origins.

FIG. 1.

Infrared spectra highlighting the (a) ν6, (b) ν2, (c) ν3 + ν4, and (d) ν1 vibrational bands of trans-HCOH. Assignment of these bands was aided by anharmonic frequency computations.18,19 Each band is split into a-type and b-type transitions, except ν6, which consists of a single c-type transition. Lower estimates of the vibrational band origins, indicated by the black arrows, are obtained using the computed (CCSD(T)/cc-pVTZ) rotational constants.

FIG. 1.

Infrared spectra highlighting the (a) ν6, (b) ν2, (c) ν3 + ν4, and (d) ν1 vibrational bands of trans-HCOH. Assignment of these bands was aided by anharmonic frequency computations.18,19 Each band is split into a-type and b-type transitions, except ν6, which consists of a single c-type transition. Lower estimates of the vibrational band origins, indicated by the black arrows, are obtained using the computed (CCSD(T)/cc-pVTZ) rotational constants.

Close modal
Table I.

Experimental and computed vibrational band origins and rotational constants (cm−1) for trans-HCOH and trans-HCOD. Relative intensities are given in parentheses.

trans-HCOH
AssignmentaMatrixbVariational computationbVCIcVPT2dHe droplete(A–B)f
ν6 1048.5(100) 1058.9 1060(100) 1051(100) 1046.69(41)g   
ν3 + ν5 h   2622(19)   h   
ν2 2703.3(42)(100i2706.5 2691(97)(100i2690(60)(100i2698.18(100) 5.80 
ν3 + ν4 2776.2(18)(43i2785.5 2776(20)(21i2774(26)(43i2771.72(42) 5.84 
ν1 3500.6(49) 3561.6 3553(43) 3553(56) 3541.01(29) 5.83 
trans-HCOD 
ν2 2588.1(100) 2626.8 2622(100) 2623(100) 2614.32(66) 4.12 
ν1 2675.9(34) 2682.8 2669(55) 2664(29) 2672.00(35) 4.15 
ν3 + ν4 2726.1(51) 2729.5 2713(33) 2722(73) 2720.55(100) 4.14 
3 2841.3(14) 2852.6 2851(17) 2853(34) 2838.75(35) 4.19 
trans-HCOH
AssignmentaMatrixbVariational computationbVCIcVPT2dHe droplete(A–B)f
ν6 1048.5(100) 1058.9 1060(100) 1051(100) 1046.69(41)g   
ν3 + ν5 h   2622(19)   h   
ν2 2703.3(42)(100i2706.5 2691(97)(100i2690(60)(100i2698.18(100) 5.80 
ν3 + ν4 2776.2(18)(43i2785.5 2776(20)(21i2774(26)(43i2771.72(42) 5.84 
ν1 3500.6(49) 3561.6 3553(43) 3553(56) 3541.01(29) 5.83 
trans-HCOD 
ν2 2588.1(100) 2626.8 2622(100) 2623(100) 2614.32(66) 4.12 
ν1 2675.9(34) 2682.8 2669(55) 2664(29) 2672.00(35) 4.15 
ν3 + ν4 2726.1(51) 2729.5 2713(33) 2722(73) 2720.55(100) 4.14 
3 2841.3(14) 2852.6 2851(17) 2853(34) 2838.75(35) 4.19 
a

Harmonic descriptions of the vibrational modes for trans-HCOH are ν1: OH stretch, ν2: CH stretch, ν3: in-phase HOC/COH bending, ν4: CO stretch, ν5: out-of-phase HOC/COH bending, and ν6: out-of-plane twist, and ν1: CH stretch, ν2: OD stretch, ν3: HOC bend, ν4: CO stretch, ν5: COD bend, and ν6: out-of-plane twist for trans-HCOD.

b

The variational computations were obtained at the AE-CCSD(T)/cc-pCVQZ level of theory.18 

c

Vibrational configuration interaction (VCI) computation.19 

d

Vibrational perturbation theory computations at the CCSD(T)/cc-pVTZ level of theory.

e

Reported values are lower limits for the vibrational band origins, which assume (B + C) = 2.284 cm−1 for HCOH or 2.107 cm−1 for HCOD obtained from CCSD(T)/cc-pVTZ computations. The He contribution to the effective moments of inertia about the inertial b- and c-axes is unknown for this system; therefore, we simply give lower limits but note that the band origins are likely to be centered ∼0.5 cm−1 to the blue of those reported here.

f

The computed (A – B) value is 8.365 cm−1 for HCOH and 6.156 cm−1 for HCOD at the CCSD(T)/cc-pVTZ level of theory. The experimental 2σ error is less than 0.003 cm−1.

g

Origin obtained using (A + B) = 10.785 cm−1 from CCSD(T)/cc-pVTZ computation.

h

Not observed.

i

Relative intensities of the two components of the ν2/(ν3 + ν4) Fermi dyad.

Figure 2 presents the infrared spectrum of the isotopically labeled, d1-trans-hydroxymethylene (HCOD). Vibrational spectra were collected on mass channel 30 u to discriminate against the undeuterated species. Four pairs of transitions (Figure 2) are identified in the spectrum near the computed frequencies of tHCOD. The increased reduced mass upon deuteration decreases the rotational constants, with the computed (CCSD(T)/cc-pVTZ) value of (A – B) reduced by ∼26%. Assuming a similar droplet effect for both HCOH and HCOD, we predict a splitting of 4.28 cm−1 between the a- and b-type lines in the spectrum of tHCOD, based on the computed reduction in the rotational constants. The observed splitting is 4.15 cm−1 on average, and the band origins are in good agreement with anharmonic frequency computations. Therefore we assign these four bands to the ν2, ν1, ν3 + ν4, and 2ν3 transitions of tHCOD, with the lower estimate of the band origins and a zero-order description of the vibrational modes presented in Table I. Overall, the computed frequencies obtained using both the variational and VCI methods are within 1% of the experimentally determined band origins for He-solvated tHCOH and tHCOD.

FIG. 2.

Infrared spectra highlighting the (a) ν2, (b) ν1, (c) ν3 + ν4, and (d) 2ν3 vibrational bands of trans-HCOD. Assignment of these bands was aided by anharmonic frequency computations.18,19 Each band is split into a-type and b-type transitions. Lower estimates of the vibrational band origins, indicated by the black arrows, are obtained using the computed (CCSD(T)/cc-pVTZ) rotational constants.

FIG. 2.

Infrared spectra highlighting the (a) ν2, (b) ν1, (c) ν3 + ν4, and (d) 2ν3 vibrational bands of trans-HCOD. Assignment of these bands was aided by anharmonic frequency computations.18,19 Each band is split into a-type and b-type transitions. Lower estimates of the vibrational band origins, indicated by the black arrows, are obtained using the computed (CCSD(T)/cc-pVTZ) rotational constants.

Close modal

From Lorentzian fits of the a- and b-type lines in the He droplet spectrum, we obtained the relative projection of the vibrational transition dipole moment for each transition. Using the CFOUR computational chemistry and spectroscopy package,28 we computed anharmonic frequencies at the CCSD(T)/cc-pVTZ level with the ANO2 basis set,29,30 making use of second-order vibrational perturbation theory (VPT2).31 The results from these computations are summarized in Table I, along with the previously reported computations. The predicted frequencies for ν1 and ν6 are in good agreement (within 12 cm−1) with the experimentally determined band origins; additionally, these computations are also consistent with the other two sets of calculated anharmonic frequencies.18,19 Previous variational computations indicated a large degree of mixing between ν2 and ν3+ν4,18 and the presence of an appreciable Fermi resonance was confirmed by the harmonic derivative analysis.32 Standard deperturbation and diagonalization of an effective Hamiltonian yields frequencies within 8 cm−1 of the experimental band origins. Furthermore, the effective Hamiltonian matrix element that mixes the dressed ν2 and ν3 + ν4 states is 38 cm−1, in pleasing agreement with the experimentally inferred value of 34 cm−1.33 Similarly good agreement with the observed level positions is achieved for HCOD, and it is clear that the VPT2 computations do a serviceable job of predicting the relative intensities of the Fermi tetrad (again obtained by a deperturbation + diagonalization treatment) comprising ν1, ν2, 2ν3, and ν3 + ν4. The relative intensities obtained with this approach are in vastly better agreement with experiment in comparison to previous computations. For the normal isotopic species, the projections of the transition dipole moment onto the a- and b-axes computed by the GUINEA module of CFOUR are [81:100], [52:100], and [48:100] for ν1, ν3 + ν4 and ν2, respectively, compared to the experimental values of [67:100], [45:100], and [49:100]. For tHCOD the transition dipole moment projections onto the a- and b-axes are [77:100], [36:100], [55:100], and [56:100] for ν2, ν1, ν3 + ν4, and 2ν3, respectively, which compare favorably to the computed values of [84:100], [49:100], [64:100], and [56:100].

We are confident in the comparison of the ν2 and ν3 + ν4 band intensities in the tHCOH spectrum, given the similar experimental conditions with which these bands were recorded; however, the ν1 and ν6 intensities cannot be compared to the others directly. The ν1 band falls near a problematic region of the laser, due to MgOH impurities in the PPLN crystal that result in a dramatic reduction in the output power; ν6 was collected with a quantum cascade laser (QCL), where the output power is ∼20 times less than the idler output of the OPO. Additionally, vibrational relaxation of the ν6 mode results in fewer evaporated He atoms, leading to a smaller laser induced depletion signal when compared to the higher frequency bands. The experimental relative intensities for tHCOD are far more reliable, given that all of the bands were recorded with similar laser conditions. Regardless of these caveats, the relative band intensities for He-solvated tHCOH and tHCOD are compared in Table I to the previously reported Ar-Matrix18 and computed intensities from the VCI19 and VPT2 computations. The relative intensities of ν2 and ν3 + ν4 (100:42) are approximately identical in both the He droplet and Ar-matrix studies when these two modes are considered independent of all other transitions. The ν3 + ν4 combination band intensity relative to ν2 is underestimated by a factor of two in the VCI computation, while a ratio of (100:43) is obtained from the VPT2 computation for these two modes (Table I), which is nearly identical to the experimental value. One drastic difference between the VCI computation and both the matrix FT-IR and He spectra is the absence of the ν3 + ν5 combination band predicted for tHCOH. This mode was found to be very sensitive to the computational method employed, with VMP2 predicting this mode 75 cm−1 higher in energy compared to the VCI value.19 

From the matrix isolation spectra reported by Schreiner et al.,18 it was determined that the ratio of HCOH to FA was 1:5.5, using the relative intensities of bands assigned to FA and HCOH scaled to the corresponding computed harmonic intensities. In a similar fashion, we estimate a ratio of 1:2.4. A combination of the rapid cooling inherent to He droplets,34–36 and the millisecond timescale of the experiment could indeed yield the larger ratio of HCOH to FA observed in the He droplet experiment. The survey scan (Figure S1 of the supplementary material22) did not reveal any lines that could be assigned to the cis-conformation of hydroxymethylene (cHCOH), although the VCI computations19 predict the ν3 + ν5 combination band to be the only transition with significant intensity (>10 km/mol) in the surveyed region. The higher energy cHCOH was also not observed in the matrix spectrum, and this was rationalized as being due to the CO2 extrusion process preferentially forming tHCOH.18 

Lastly, we comment on the variable width associated with each vibrational band. For example, the a-type line widths for the ν1 and ν3 + ν4 bands of tHCOH are 0.065 and 0.37 cm−1, respectively. This dramatic difference clearly cannot be attributed solely to the different state density of droplet phonon modes at the energy of the oscillator, as an opposite trend would be expected. Several previous reports of He-solvated hydrocarbon molecules, such as C2H2,37 C2H4,38 and C2H5,39 have discussed vibrational state-dependent broadening in the high-frequency CH stretch region. Although a clear physical picture has yet to emerge, it is becoming increasingly evident that transitions to vibrational levels strongly mixed by mechanical anharmonicity are homogeneously broadened to an extent that is far greater than observed for more weakly coupled modes. Indeed, for tHCOH, the observed transitions within the ν2/(ν3 + ν4) Fermi dyad are broader than the analogous transitions within the ν1 OH stretch, despite the latter mode lying ∼800 cm−1 higher in energy. Invoking a simple model where a single bright state (i.e., CH stretch) is coupled to the bath of droplet states, through some series of doorway states (i.e., lower frequency vibrations and rotations), stronger coupling of the bright state to the doorway states could give rise to a more rapid decay of the vibrational resonance into the helium bath. Time resolved experiments are clearly needed in combination with theory to elucidate the rich vibrational dynamics operative in this dissipative superfluid environment.

In summary, rotationally resolved spectra are obtained for tHCOH and tHCOD solvated in He nanodroplets, and the experimentally determined band origins agree with anharmonic frequency computations, and these should be only slightly perturbed from the true gas-phase values.26,27 The isolation of this transient species in He droplets should lead to future studies probing the reactivity of HCOH, for example, with formaldehyde, providing valuable insight into the role of HCOH in the formation of simple sugars and other prebiotic chemical species.7,11,13

This work was supported by the Office of Energy Research, Office of Basic Energy Sciences, Chemical Sciences, Geosciences and Biosciences Division of the U.S. Department of Energy (DOE) under Contract No. DE-FG02-12ER16298. J.F.S. would like to thank the Robert A. Welch Foundation of Houston, TX (Grant No. F-1283) for financial support.

1.
R. R.
Lucchese
and
H. F.
Schaefer
,
J. Am. Chem. Soc.
100
,
298
(
1978
).
2.
D.
Townsend
,
S. A.
Lahankar
,
S. K.
Lee
,
S. D.
Chambreau
,
A. G.
Suits
,
X.
Zhang
,
J.
Rheinecker
,
L. B.
Harding
, and
J. M.
Bowman
,
Science
306
,
1158
(
2004
).
3.
B. C.
Shepler
,
E.
Epifanovsky
,
P.
Zhang
,
J. M.
Bowman
,
A. I.
Krylov
, and
K.
Morokuma
,
J. Phys. Chem. A
112
,
13267
(
2008
).
4.
B. N.
Fu
,
B. C.
Shepler
, and
J. M.
Bowman
,
J. Am. Chem. Soc.
133
,
7957
(
2011
).
5.
P.
Zhang
,
S.
Maeda
,
K.
Morokuma
, and
B. J.
Braams
,
J. Chem. Phys.
130
,
114304
(
2009
).
6.
M. R.
Hoffmann
and
H. F.
Schaefer
,
Astrophys. J.
249
,
563
(
1981
).
7.
D.
Ley
,
D.
Gerbig
, and
P. R.
Schreiner
,
Org. Biomol. Chem.
10
,
3781
(
2012
).
8.
Y.
Osamura
,
J. D.
Goddard
,
H. F.
Schaefer
, and
K. S.
Kim
,
J. Chem. Phys.
74
,
617
(
1981
).
9.
10.
P. S.
Peters
,
D.
Duflot
,
A.
Faure
,
C.
Kahane
,
C.
Ceccarelli
,
L.
Wiesenfeld
, and
C.
Toubin
,
J. Phys. Chem. A
115
,
8983
(
2011
).
11.
V. G.
Kiselev
,
S.
Swinnen
,
V. S.
Nguyen
,
N. P.
Gritsan
, and
M. T.
Nguyen
,
J. Phys. Chem. A
114
,
5573
(
2010
).
12.
L.
Koziol
,
V. A.
Mozhayskiy
,
B. J.
Braams
,
J. M.
Bowman
, and
A. I.
Krylov
,
J. Phys. Chem. A
113
,
7802
(
2009
).
13.
H. M.
Buck
,
Int. J. Quantum Chem.
112
,
3711
(
2012
).
14.
J. R.
Sodeau
and
E. K. C.
Lee
,
Chem. Phys. Lett.
57
,
71
(
1978
).
15.
L.
Feng
,
A. V.
Demyanenko
, and
H.
Reisler
,
J. Chem. Phys.
120
,
6524
(
2004
).
16.
L.
Feng
and
H.
Reisler
,
J. Phys. Chem. A
108
,
9847
(
2004
).
17.
C. P.
Rodrigo
,
C. C.
Zhou
, and
H.
Reisler
,
J. Phys. Chem. A
117
,
12049
(
2013
).
18.
P. R.
Schreiner
,
H. P.
Reisenauer
,
F. C.
Pickard
,
A. C.
Simmonett
,
W. D.
Allen
,
E.
Mátyus
, and
A. G.
Császár
,
Nature
453
,
906
(
2008
).
19.
L.
Koziol
,
Y. M.
Wang
,
B. J.
Braams
,
J. M.
Bowman
, and
A. I.
Krylov
,
J. Chem. Phys.
128
,
204310
(
2008
).
20.
M. Y.
Choi
,
G. E.
Douberly
,
T. M.
Falconer
,
W. K.
Lewis
,
C. M.
Lindsay
,
J. M.
Merritt
,
P. L.
Stiles
, and
R. E.
Miller
,
Int. Rev. Phys. Chem.
25
,
15
(
2006
).
21.
A. M.
Morrison
,
T.
Liang
, and
G. E.
Douberly
,
Rev. Sci. Instrum.
84
,
013102
(
2013
).
22.
See supplementary material at http://dx.doi.org/10.1063/1.4874850 for experimental methods, mass spectra, and survey infrared spectrum.
23.
P. L.
Raston
and
G. E.
Douberly
,
J. Mol. Spectrosc.
292
,
15
(
2013
).
24.
K. L.
Plath
,
J. L.
Axson
,
G. C.
Nelson
,
K.
Takahashi
,
R. T.
Skodje
, and
V.
Vaida
,
React. Kinet. Catal. Lett.
96
,
209
(
2009
).
25.
A. M.
Morrison
,
P. L.
Raston
, and
G. E.
Douberly
,
J. Phys. Chem. A
117
,
11640
(
2013
).
26.
J. P.
Toennies
and
A. F.
Vilesov
,
Angew. Chem. Int. Ed.
43
,
2622
(
2004
).
27.
C.
Callegari
,
K. K.
Lehmann
,
R.
Schmied
, and
G.
Scoles
,
J. Chem. Phys.
115
,
10090
(
2001
).
28.
CFOUR, a quantum-chemical program package by
J. F.
Stanton
,
J.
Gauss
,
M. E.
Harding
,
P. G.
Szalay
 et al For the complete citation and current version, see http://www.cfour.de.
29.
J.
Almlöf
and
P. R.
Taylor
,
J. Chem. Phys.
86
,
4070
(
1987
).
30.
L.
McCaslin
and
J.
Stanton
,
Mol. Phys.
111
,
1492
(
2013
).
31.
I. M.
Mills
, in
Molecular Spectroscopy: Modern Research
, edited by
K. N.
Rao
and
C. W.
Matthews
(
Academic
,
New York
,
1972
), pp.
115
.
32.
D. A.
Matthews
and
J. F.
Stanton
,
Mol. Phys.
107
,
213
(
2009
).
33.
G.
Herzberg
,
Molecular Spectra and Molecular Structure
(
Van Nostrand Reinhold Company, Inc.
,
New York
,
1945
).
34.
D. S.
Skvortsov
and
A. F.
Vilesov
,
J. Chem. Phys.
130
,
151101
(
2009
).
35.
A.
Lindinger
,
J. P.
Toennies
, and
A. F.
Vilesov
,
J. Chem. Phys.
110
,
1429
(
1999
).
36.
M. Y.
Choi
and
R. E.
Miller
,
J. Am. Chem. Soc.
128
,
7320
(
2006
).
37.
K.
Nauta
and
R. E.
Miller
,
J. Chem. Phys.
115
,
8384
(
2001
).
38.
C. M.
Lindsay
and
R. E.
Miller
,
J. Chem. Phys.
122
,
104306
(
2005
).
39.
P. L.
Raston
,
J.
Agarwal
,
J. M.
Turney
,
H. F.
Schaefer
, and
G. E.
Douberly
,
J. Chem. Phys.
138
,
194303
(
2013
).

Supplementary Material