Rotational Spectrum, Conformational Composition, and Quantum

Article pubs.acs.org/JPCA

Rotational Spectrum, Conformational Composition, and Quantum Chemical Calculations of Cyanomethyl Formate (HC(O)OCH2CN), a Compound of Potential Astrochemical Interest Svein Samdal,† Harald Møllendal,*,† and Sophie Carles‡ †

Centre for Theoretical and Computational Chemistry (CTCC), Department of Chemistry, University of Oslo, P.O. Box 1033, Blindern, NO-0315 Oslo, Norway ‡ Institut de Physique de Rennes, Département de Physique Moléculaire, UMR 6251 UR1-CNRS, Université de Rennes 1, Bâtiment 11 C, Campus de Beaulieu, F-35042 Rennes Cedex, France S Supporting Information *

ABSTRACT: The rotational spectrum of cyanomethyl formate (HC(O)OCH2CN) has been recorded in the 12−123 GHz spectral range. The spectra of two conformers were assigned. The rotamer denoted I has a symmetry plane and two out-of plane hydrogen atoms belonging to the cyanomethyl (CH2CN) moiety. In the conformer called II, the cyanomethyl group is rotated 80.3° out of this plane. Conformer I has an energy that is 1.4(6) kJ/mol lower than the energy of II according to relative intensity measurements. A large number of rotational transitions have been assigned for the ground and vibrationally excited states of the two conformers and accurate spectroscopic constants have been obtained. These constants should predict frequencies of transitions outside the investigated spectral range with a very high degree of precision. It is suggested that cyanomethyl formate is a potential interstellar compound. This suggestion is based on the fact that its congener methyl formate (HC(O)OCH3) exists across a large variety of interstellar environments and the fact that cyanides are very prevalent in the Universe. The experimental work has been augmented by high-level quantum chemical calculations. The CCSD/cc-pVQZ calculations are found to predict structures of the two forms that are very close to the Born− Oppenheimer equilibrium structures. MP2/cc-pVTZ predictions of several vibration−rotation interaction constants were generally found to be rather inaccurate. A gas-phase reaction between methyl formate and the cyanomethyl radical CH2CN to produce a hydrogen atom and cyanomethyl formate was mimicked using MP2/cc-pVTZ calculations. It was found that this reaction is not favored thermodynamically. It is also conjectured that the possible formation of cyanomethyl formate might be catalyzed and take place on interstellar particles.

■

INTRODUCTION Methyl formate (HC(O)OCH3) is a typical interstellar hotcore molecule abundant in massive star-forming regions. It was found already in 1975 in Sagittarius B2 by means of its rotational spectrum.1 Methyl formate has later been detected toward low-mass star-forming regions2 and in cold clouds in the Galactic center.3 The spectra of vibrationally excited states,4,5 13 C-isotopologues,6 and 18O-isotopologues7 of this interstellar compound have been identified. A tentative identification of the deuterated species DC(O)OCH3 in Orion has also been reported.8 There are two rotameric forms of methyl formate. The O C−O−C chain of atoms is synperiplanar (sp) in one of these conformers9 and antiperiplanar (ap) in the second rotamer.10 The discussion above refers to the lower-energy OC−O−C sp form, which is about 25 kJ/mol lower in energy than the ap rotamer. Interestingly, a tentative detection in the interstellar medium (ISM) of the higher-energy ap conformer has been reported.10 The fact that methyl formate exists across a large variety of interstellar sources suggests that further esters of formic acid might also be formed in the ISM. Indeed, one additional such © XXXX American Chemical Society

ester, ethyl formate (HC(O)OCH2CH3), has recently been detected in two interstellar sources. 11,12 This finding encouraged us to undertake this first investigation of the rotational spectrum of yet another substituted methyl formate ester, namely, cyanomethyl formate HC(O)OCH2CN, hoping that this spectrum could form the basis for a potential future identification of this compound in the ISM. In cyanomethyl formate, one of the hydrogen atoms of the methyl group of HC(O)OCH3 is substituted by a cyano group. The cyano group is very prevalent in interstellar and circumstellar compounds and found in about 25 of the approximately 180 known compounds of this category.13 This functional group is found in neutral molecules, cations, anions, and radicals.13 The widespread occurrence of compounds containing this group in the ISM is one reason why cyanomethyl formate is suggested as a potential interstellar compound. Received: June 3, 2015 Revised: July 24, 2015

A

DOI: 10.1021/acs.jpca.5b05285 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

compounds, formic acid (HC(O)OH),14 and the cyanomethyl radical (CH2CN),15 which might lead to the formation of cyanomethyl formate (HC(O)OCH2CN).

Another reason for carrying out this research is that new and much more sensitive radio astronomy observatories such as, for example, the Atacama Large Millimeter/Submillimeter Array (ALMA) in Chile have created a demand for rotational spectra of candidate interstellar compounds such as the title compound. Cyanomethyl formate not only is of potential astrochemical interest but also has interesting conformational properties, which is associated with the rotation about its two C−O single bonds. The quantum chemical calculations discussed below indicate that there are three conformers of this compound. These three rotamers are shown in Figure 1 and given Roman

■

EXPERIMENTAL SECTION Compound and Spectroscopic Experiments. Cyanomethyl formate specified to be 99% pure was purchased from Aldrich and used as received. The rotational spectrum, which was studied using the Stark-modulated spectrometer of the University of Oslo, revealed no impurities. Cyanomethyl formate has a vapor pressure of roughly 60 Pa at room temperature. The spectrum was recorded in the 12−123 GHz spectral interval at a pressure of 5−10 Pa at room temperature, or with the MW cell cooled to about 0 °C. It was not possible to study the spectrum at lower temperatures, which would have enhanced intensities, due to the rather low vapor pressure of this compound. Salient features of the Oslo MW spectrometer, which has previously been described in detail,16 is its accuracy of 0.10 MHz for strong and isolated MW transitions and its resolution of about 0.5 MHz. Radio-frequency microwave double-resonance (RFMWDR) spectroscopy17 was also performed with this spectrometer to confirm certain assignments.

■

RESULTS AND DISCUSSION Quantum Chemical Methods. The Gaussian 0918 suite of programs was used for the MP219 calculations, whereas the Molpro20 set of programs was employed for the CCSD computations. The frozen-core approximation was assumed in all calculations and the default convergence criteria of the two programs were always used. Dunning’s21 correlation-consistent triple-ζ (cc-pVTZ) and quadruple-ζ (cc-pVQZ) basis sets were used in the MP2 calculations, whereas the cc-pVQZ basis set was chosen for the CCSD computations. The calculations were performed employing the Abel cluster of the University of Oslo. Computational Conformational Analysis. The conformational characteristics of cyanomethyl formate can perhaps best be described by the values of the C2−O4−C5−O6 and O1−C2−O4−C5 dihedral angles. Two MP2/cc-pVTZ scans were performed. In the first of these, the rotation about the O4−C5 bond was studied by stepping the C2−O4−C5−O6 dihedral angle in 10° intervals, while keeping the O1−C2− O4−C5 chain of atoms in the sp conformation (about 0°). All structural parameters but the C2−O4−C5−O6 dihedral angle were optimized. The resulting potential function, which is shown in Figure 2, has two minima corresponding to conformers II and I, respectively. The geometries of I and II were optimized and their dipole moments, 14N nuclear quadrupole coupling constants, harmonic and anharmonic vibrational frequencies, Watson’s quartic and sextic centrifugal distortion constants,22 rotation−vibration constants (α’s),23 and differences between effective (r0) and equilibrium (re) rotational constants were computed; the results are listed in Tables S1 and S2 of the Supporting Information. It is seen from these two tables that conformer II has a MP2/ cc-pVTZ electronic energy that is 0.77 kJ/mol lower than that of I. This difference becomes 0.18 kJ/mol (II lower-energy form) after correction for the harmonic zero-point vibrational energies given in the same two tables. The C2−O4−C5−O6 and O1−C2−O4−C5 dihedral angles were 180.0 and 0.0° for I and 77.7 and 0.2° for II. The transition state for conversion

Figure 1. Models of the three conformers of cyanomethyl formate. Atom numbering is indicated on conformer I. Microwave spectra of I and II have been assigned, and the energy difference is 1.4(6) kJ/mol with I as the lower-energy conformer. III is 23.6 kJ/mol higher in energy than I according to CCSD/cc-pVQZ calculations. Its rotational spectrum must consequently be extremely weak and it was not assigned.

numerals for reference. Atom numbering is given on conformer I. The O1−C2−O4−C5 and the C2−O4−C5−C6 dihedral angles can conveniently be used to characterize the conformational properties of this compound. The O1−C2−O4−C5 dihedral angle is exactly 0° in I, 0.5° in II, and −179.6° in III, whereas the C2−O4−C5−C6 dihedral angle is exactly 180° in I, 80.3° in II, and 78.1° in III, according to the CCSD/ccpVQZ calculations discussed below. The vast majority of the interstellar and circumstellar compounds have been identified by their rotational spectra and microwave (MW) spectroscopy was therefore chosen to investigate the rotational spectrum of this compound in the 12−123 GHz spectral range. Moreover, rotational spectroscopy is a superior conformational-analysis method due to its unparalleled accuracy and resolution. High-level quantum chemical calculations can now be used to predict rather accurate rotational and centrifugal distortion constants, which are useful in the assignment procedures of complex rotational spectra. Reliable parameters not readily available from experiment can sometimes be obtained from these calculations as well. One example is reaction mechanisms and we have also used quantum chemistry to investigate a possible gas-phase reaction between two known interstellar B


Article


Figure 3. Relative electronic energy as a function of the C2−O4−C5− C6 dihedral angle. The O1−C2−O4−C5 chain of atom is held in the ap conformation. The minimum corresponds to III; see text.

Figure 2. Relative electronic energy as a function of the C2−O4−C5− C6 dihedral angle. The O1−C2−O4−C5 chain of atom is held in the sp conformation. The minimum at 77.7° corresponds to conformer II, and the other minimum at 180° corresponds to I. The maximum at 129.7° has an energy that is only 3.55 kJ/mol higher than the energy of II. The relatively flat portion of the potential curve between about 65° and 180° implies that both I and II will undergo large amplitude vibrations about the O4−C5 bond. Indeed, the MP2 anharmonic torsional vibrations are 47.7 cm−1 in I and 62.2 cm−1 in II; see text.

characteristic C2−O4−C5−O6 and O1−C2−O4−C5 dihedral angles are +78.3 and −179.1°, respectively, for III. Computations of optimized CCSD/cc-pVQZ structures, principal inertial-axes dipole moment components, total dipole moment, and energy differences of the three conformers I−III were carried out. Calculations of vibrational frequencies were not done because they would be too costly at this high level of theory. The CCSD structures, which are assumed to be close to the Born−Oppenheimer equilibrium structures, and dipole moments of the three forms are listed in Table 1, and further results of these calculations are listed in Tables S5 − S7 of the Supporting Information. The rotational constants calculated from the CCSD structures in Table 1 are given in Tables 2 and 3 below together with their experimental equivalents. These tables also contain the MP2 quartic centrifugal distortion constants taken from Tables S1 and S2, as well as their experimental counterparts. Some of the results of these quantum chemical calculations warrant comments. The CCSD computations predict conformer I to be the rotamer with the lowest electronic energy, being 0.55 kJ/mol lower in energy than II, and 23.60 kJ/mol lower in energy than III. The C2−O4−C5−O6 and O1−C2− O4−C5 dihedral angles are 180 and 0°, 80.3 and 0.5°, and +78.1 and −179.6° in the three cases. These results are close to what was found in the MP2/cc-pVTZ computations above, but the energy order of I and II is opposite from what was found in the MP2/cc-pVTZ calculations. It would be of interest to see what role the basis set plays for the energies of I and II. MP2 calculations of optimized structures of I and II were repeated using the cc-pVQZ basis set. Calculations of vibrational frequencies at this level of theory were too costly to be performed. The results are given in Tables S1 and S2, respectively. It was found II is lower in electronic energy than I, just as in the case of the of the MP2/cc-pVTZ calculations (see above). The MP2/cc-pVQZ electronic energy difference is 0.56 kJ/mol compared to 0.77 kJ/mol found the MP2/cc-pVTZ calculations. The CCSD prediction that I and II are much lower (∼23 kJ/ mol) in energy than III is in accord with the propensity of formic acid esters to prefer sp conformations for the C2−O4− C5−O6 link of atom, as exemplified by HC(O)OCH3,10 HC(O)OCH2CH3,24,25 and HC(O)OCH2F.26 Conformer III

between conformers I and II was calculated (Table S3) to have an electronic energy that is only 3.55 kJ/mol higher than the energy of II. This maximum occurs at 129.7° for the C2−O4− C5−O6 dihedral angle. The rather flat part of the potential function between about 65° and 180° indicates that the torsion about the O4−C5 bond will be a large amplitude vibration in both I and II. The other maximum of Figure 2 occurs when the C2−O4− C5−O6 and the O1−C2−O4−C5 dihedral angles both are 0.0°. The energy of this maximum is 28.55 kJ/mol higher than the electronic energy of II. This high energy is presumably due to the fact that the carbonyl oxygen atom and the cyano group are brought into close contact in this conformation resulting in a comparatively large repulsive interaction. In the second scan, the C2−O4−C5−O6 dihedral angle was again changed in 10° portions, whereas the O1−C2−O4−C5 link of atoms was ap (approximately 180°). The ensuing potential function is drawn in Figure 3. This function is very different from the first function (Figure 2) and has only one minimum corresponding to conformer III instead of the two minima (conformers I and II) found in the first scan, were the O1−C2−O4−C5 link of atoms was sp. The energy of the two barrier maxima are 8.57 kJ/mol at 0.0° and 6.25 kJ/mol at 180°. The much lower barrier at 0° (8.57 kJ/mol) than in the previous case (28.55 kJ/mol) may reflect the absence of repulsion between the carbonyl oxygen atom (O1) and the cyano group in the O1−C2−O4−C5 ap conformation. MP2/cc-pVTZ calculations of the same series of molecular parameters as those performed above for I and II were also undertaken for III with the results shown in Table S4 of the Supporting Information. Interestingly, III is as much as 23.45 kJ/mol higher in electronic energy than II (from Tables S2 and S4). This difference becomes 21.46 kJ/mol after correction for the harmonic zero-point vibrational energies. The corresponding energy difference in methyl formate is 25 kJ/mol.10 The C


Article


Table 3. Spectroscopic Constantsa of Conformer II of Cyanomethyl Formate at Different Vibrational States

Table 1. CCSD/cc-pVQZ Parameters of the Conformers I, II, and III of Cyanomethyl Formate I

II

Bond Distance (pm) 119.3 119.1 109.1 109.1 134.0 134.5 142.4 142.1 146.5 147.1 108.8 108.7 108.8 108.5 115.0 115.0 Angle (deg) O1−C2−H3 125.7 125.7 O1−C2−O4 124.9 125.1 H3−C2−O4 109.4 109.2 C2−O4−C5 113.6 114.9 O4−C5−C6 107.6 111.5 O4−C5−H8 110.5 110.8 O4−C5−H9 110.5 106.1 C6−C5−H8 109.9 109.1 C6−C5−H9 109.9 109.2 H8−C5−H9 108.3 110.2 C5−C6−N7 178.1 178.6 Dihedral Angle (deg) O1−C2−H3−O4 180.0 179.7 O1−C2−O4−C5 0.0 0.5 H3−C2−O4−C5 180.0 −179.8 C2−O4−C5−C6 180.0 80.3 C2−O4−C5−H8 60.0 −41.5 C2−O4−C5−H9 −60.0 −161.0 Dipole Moment (debyea) μa 2.26 2.60 μb 0.09 3.49 μc 0.00b 0.66 μtot 2.26 4.40 Electronic Energy Diffrencesc (kJ/mol) ΔE 0.0 0.55 O1−C2 C2−H3 C2−O4 O4−C5 C5−C6 C5−H8 C5−H9 C6−N7

III Av (MHz) Bv (MHz) Cv (MHz) ΔJ (kHz) ΔJK (kHz) ΔK (kHz) δJ (kHz) δK (kHz) ΦJ (Hz) ΦJK (Hz) ΦKJ (Hz) ΦK (Hz) ϕJ (Hz) ϕJKc (Hz) rmsa std (MHz)a Na

118.4 109.9 135.4 141.3 147.3 109.0 108.5 115.1 124.6 122.0 113.4 115.9 112.0 111.6 106.5 108.8 109.3 108.5 179.3

ground

first excited torsion (v21)

equilibriumb

7608.5723(26) 2608.94813(71) 2198.40201(72) 6.44242(73) −36.4432(42) 99.941(11) 2.13646(33) 17.343(13) −0.09001(29) 1.0602(25) −5.416(20) 13.169(49) −0.03629(16) −1.0158(98) 1.323 0.12 641

7852.3358(51) 2549.4614(15) 2170.8155(15) 7.4003(39) −51.5050(99) 163.91(15) 2.44371(50) 22.079(20) −0.1504(33) 2.1423(60) −12.290(31) 24.2(12) −0.06238(27) −1.946(17) 1.523 0.14 655

7704.9 2608.1 2206.0 5.30 −24.6 60.8 1.74 13.0 −0.0461 0.371 −17.3 43.8 −0.0191 −0.452

a Defined in the footnote of Table 2. The spectra are listed in Tables S17 and S18 of the Supporting Information. bRotational constants from CCSD/cc-pVQZ structure in Table 1. MP2/cc-pVTZ centrifugal distortion constants from Table S2. cϕK preset at zero in the leastsquares fit; see text.

−179.6 −179.6 0.9 78.1 −44.2 −162.5

has a Boltzmann factor relative to I as low as about 8 × 10−5 at room temperature, which means that III will have an extremely weak MW spectrum that would be impossible to detect with our spectrometer. This rotamer is therefore not considered further. The existence of two conformers such as I and II with similar energies is typical for substituted formate esters and has already been found for HC(O)OCH2CH324,25 and HC(O)OCH2F.26 The C2−O4−C5−C6 chain of atoms takes the expected 180° angle in I, and the unusual value of 80.3° in II (Table 1). The corresponding C−O−C−C dihedral angle is 180° and 81.7° (rα0-structure) in the corresponding two conformers of HC(O)OCH2CH3,25 whereas the C−O−C−F dihedral angle is 180° and 84(1)° in the two conformers of HC(O)OCH2F.26

0.78 3.12 2.30 3.95 23.60

1 debye =3.33564 × 10−30 C m. bFor symmetry reasons. cRelative to Conformer I. a

Table 2. Spectroscopic Constantsa of Conformer I of Cyanomethyl Formate at Different Vibrational States Av (MHz) Bv (MHz) Cv (MHz) 2Pc (10−20 u m2)c ΔJ (kHz) ΔJK (kHz) δJ (kHz)d ΦKJ (Hz)f rmsg std (MHz)h Ni

ground

first excited tors (v21)

second excited tors

third excited tors

lowest bend (v20)

equilibriumb

19304.2(23) 1775.2182(43) 1644.0460(44) 3.466(4) 0.20014(32) −2.010(16) 0.0254(16) −0.58(12) 1.191 0.13 330

18710.7(47) 1777.9569(74) 1650.9341(77) 5.140(9) 0.21687(54) −3.3927(52) 0.0241(29) −e 1.271 0.14 238

18114.4(78) 1781.622(13) 1659.412(13) 7.01(2) 0.25357(56) −6.1409(42) −e −e 1.388 0.15 214

17354.5(79) 1786.461(13) 1670.405(13) 9.47(2) 0.3166(15) −11.074(83) −e −e 1.762 0.21 70

19752.9(76) 1777.827(14) 1644.608(14) 2.56(1) 0.20690(74) −2.915(23) −e −e 1.511 0.17 158

19731.7 1784.1 1653.0 3.14 0.193 −1.34 0.0218 −0.280

a A-reduction, Ir-representation.22 Spectra listed in Tables S12−S16. bApproximate equilibrium rotational constants from CCSD/cc-pVQZ structure (Table 1) and approximate equilibrium MP2/cc-pVTZ centrifugal distortion constants (Table S1). cPlanar moment defined by Pc = (Ia + Ib − Ic)/2, where Ia, Ib, and Ic, are the principal moments of inertia. Conversion factor: 505379.05 × 10−20 MHz u m2. dFurther quartic constants preset at zero; see text. eFixed at zero; see text. fFurther sextic centrifugal distortion constants preset at zero; see text gRoot-mean-square deviation defined as rms2 = Σ[(νobs − νcalc)/u]2/(N − P), where νobs and νcalc are the observed and calculated frequencies, u is the uncertainty of the observed frequency, N is the number of transitions used in the least-squares fit, and P is the number of spectroscopic constants used in the fit. hStandard deviation of the fit. i Number of transitions used in the fit.

D


Article


almost exclusively oriented parallel with its a-inertial axis (Table 1). Only pile-ups of a-type R-branch transitions separated by almost exactly the sum of the B and C rotational constants were expected for this rotamer. The weakness of the spectrum is presumably largely caused by the existence of numerous vibrationally excited states. The MP2 calculations predict five normal vibrations for each conformer (Tables S1 and S2) with frequencies below 500 cm−1. The torsions about the O4−C5 bond, which are as low as 47.7 cm−1 in I (Table S1, anharmonic frequency) and 62.2 cm−1 in II (Table S2), will contribute most to the reduction of intensity, because the Boltzmann factors are much larger for these two normal modes than for the other fundamentals. This simple nature of the spectrum of I greatly facilitated its assignment and the spectrum of this conformer was first assigned. The CCSD rotational and the MP2 quartic centrifugal distortion constants shown in Table 2, last column, were used to predict the approximate frequencies of the spectrum of I, which was assigned in a straightforward manner. The assignments of several of these transitions were confirmed using RFMWDR spectroscopy. The higher-K−1 members of these series of aR-transitions have rapid Stark effects, which were also very useful for their assignments. No b-type lines were observed, presumably because μb is as small as 0.09 D resulting in very weak b-type transitions. c-type lines were excluded because μc is zero for symmetry reasons. A total of 330 aR-transitions with J-values between 4 and 35 with a maximum value of K−1 = 12 were assigned. These lines are listed in Table S12 of the Supporting Information. No splittings due to quadrupole coupling of the 14N nucleus with rotation were observed. The transitions were least-squares fit to Watson’s A-reduction Ir-representation Hamiltonian22 using Sørensen’s program Rotfit.32 The spectroscopic constants obtained in this manner are shown in Table 2, second column. Significant values different from zero could only be obtained from our collection of aR-branch lines for three of the quartic centrifugal distortion constants, namely, ΔJ, ΔJK, and δJ, and one sextic constant, ΦKJ. The remaining centrifugal distortion constants were preset at zero in the least-squares fit. It was not possible to determine the dipole moment of this conformer as well as that of conformer II because no low-J lines with a clearly resolved Stark pattern that could be used for quantitative measurements were observed. Comparison of experimental and theoretical parameters is in order. The value of twice the planar moment, 2Pc, where Pc = (Ia + Ib − Ic)/2 and Ia, Ib, and Ic are the principal moments of inertia, is characteristic for compounds containing a symmetry plane and two out-of-plane hydrogen atoms attached to an sp3hybridized carbon atom, which is the case for I. The CCSD value of 2Pc (Table 2, last column) is 3.14 × 10−20 u m2, whereas the corresponding value is 3.466 (same order of magnitude and units) for the ground vibrational state (second column, same table). The experimental value of 2Pc (3.466(4) × 10−20 u m2; Table 2) is larger than the theoretical counterpart. This increase is presumably caused largely by the torsional vibration about the O4−C5 bond.33 The CCSD rotational constants (Table 2, last column), which are close to the equilibrium values, are larger than the ground-state rotational constants (same table, second column) by 427.5, 8.9, and 9.0 MHz in the cases of A0, B0, and C0, respectively. This is what one would expect because the equilibrium bond lengths (re) are generally shorter than the ground-state bond lengths (r0), resulting in smaller principal

A C−O−C−X, dihedral angle of about 80°, where X is a substituent, thus seems to be normal for one of the conformers of substituted formates. Formation of Cyanomethyl Formate in the ISM. The chemistry in the ISM is very complex involving both the gas phase and compounds adsorbed on “dust” particles.27−30 One way to generate cyanomethyl formate from compounds already known to exist in the ISM could be a reaction between interstellar formic acid (HC(O)OH) 14 and interstellar cyanomethyl radical (H2CCN)15 in the gas phase. HC(O)OH + H 2CCN → HC(O)OCH 2CN + H

(1)

Unrestricted MP2/cc-pVTZ calculations for this reaction listed in Tables S7−S10 yielded an electronic reaction energy corrected for harmonic zero-point vibrational energy (ΔEreact) of +73.5 kJ/mol. Reaction 1 is therefore likely to be nonspontaneous. The transition-state structure, which was also computed, is shown in Figure 4. This state is +84.0 kJ/mol higher in energy

Figure 4. Model of the transition state of the reaction between formic acid (HC(O)OH) and the cyanomethyl radical (CH2CN).

than the combined energies of formic acid and the cyanomethyl radical. This is a significant activation energy, which, together with the value of ΔEreact (73.5 kJ/mol) implies that this gas phase reaction is rather unlikely. However, a surface reaction could be catalyzed, leading to the formation of cyanomethyl formate from HC(O)OH and H2CCN. Other surface reactions might also lead to cyanomethyl formate. There exists a large reservoir of a variety of cyanides in interstellar space,13 which may provide the necessary reagents leading to cyanomethyl formate. MW Spectrum and Assignment of the Ground-State Spectrum of Conformer I. Survey spectra revealed a very dense and relatively weak rotational spectrum with absorption lines occurring every few megahertz throughout the investigated spectral range, 12−123 GHz. This is not surprising because the rotational partition function of I is 1.17 × 105 (from the ground-state rotational constants) and the vibrational partition function is 2.06 (from the MP2/cc-pVTZ harmonic vibrational frequencies) at a temperature of 300 K. The corresponding values are 1.33 × 105 and 1.92 for II. Most of the observed transitions are likely to be due to II, which has sizable components of the dipole moment along both the a- and the b-axis (Table 1). b-type transitions would be especially numerous for a compound with the rotational constants similar to those predicted for II (Table 3, last column). A minor part of the spectral lines should belong to I, because it is calculated to be a near prolate rotor (Ray’s asymmetry parameter31 κ ≈ − 0.94) with a dipole moment E


Article


relative intensity measurements yielded 173(30) cm−1. The experimental α-constants are −454.7(79), − 2.609(15), and −0.562(15) MHz compared to their MP2 equivalents −380.43, −2.11, and −0.316 MHz, respectively (Table 1S). Assignment of the Spectrum of Conformer II. This conformer was predicted to have its major dipole moment component (3.49 D; Table 1) along the b-axis. Searches were first performed for the b-type Q-branch transitions, which are among the strongest transitions of the spectrum, using the CCSD and MP2 spectroscopic constants shown in the last column of Table 3 as starting points. These transitions were soon found. Successful searches for R-branch transitions of the a- and b-varieties were then undertaken. The assignment of several aR-branch lines were confirmed by RFMWDR experiments. The Stark effects of high-K−1 lines is very rapid, and this facilitated the assignments of these transitions. No quadrupol splittings due to 14N were seen. The assignments of the aR-, bQ- and bR-transitions were gradually extended to include additional transitions involving higher and higher values of the J-quantum number. All Rbranch transitions up to J = 56 could be least-squares fitted within their experimental uncertainties. Attempts to include transitions with even higher values of J failed presumably because they are too weak due to very small Boltzmann factors. However, bQ-branch transitions could be fitted within their experimental uncertainties only up to J ∼ 40. For higher values of J, deviations of up to a few megahertz, more than10 times larger than the experimental uncertainties, were encountered in most cases. Inclusion of octic centrifugal distortion constants did not improve the fit. These transitions were consequently omitted from the fit. The same problem with these transitions was encountered both for the A- and the S-reduction of Watson’s Hamiltonian,22 but the fit using the A-reduction seemed to be slightly better than the fit with the S-reduction Hamiltonian. The reason for the unusual behavior of the high-J b Q-branch lines is assumed to be comparatively small higherorder vibration−rotation interactions. The c-component of the dipole moment is 0.66 D according to the CCSD calculations (Table 1). Searches for c-type transitions were undertaken, but none of them could be unambiguously assigned due mainly to the high spectral density with frequent overlapping transitions and the interference of Stark lobes. A total of 641 transitions shown in Table S17 of the Supporting Information were used to determine the Areduction spectroscopic constants displayed in Table 3, second column. All quartic and all sextic constants but ϕK were determined. A value of ϕK significantly different from zero could not be obtained. The value of this constant was hence preset at zero in the least-squares fit. Comparison of the experimental spectroscopic constants (Table 2, second column) with the CCSD rotational constants and MP2 centrifugal distortion constants, show that the CCSD A and C rotational constants are larger than the experimental equivalents by 96.3 and 7.6 MHz, respectively, whereas B is slightly smaller by 0.8 MHz. However, MP2 calculations (Table 2S) predicts that the equilibrium value of A is smaller than the ground-state value by 8.9 MHz, whereas the equilibrium values of B and C are 32.92 and 23.77 MHz, respectively, larger than the r0 values. Table 3 reveals that the experimental quartic centrifugal distortion constants are all much larger than the MP2 counterparts. The MP2 sextic constants deviate very much form experiment. It is concluded that the MP2 predictions are

axis moments of inertia and larger rotational constants for the equilibrium rotational constants compared to the ground-state rotational constants. The re- and r0-rotational constants have been calculated at the MP2 level (Table S1). The differences between them are 183.2, 10.0, and 8.3 MHz, in the cases of A, B, and C, respectively. The MP2 differences compared to the differences between the CCSD and r0 rotational constants just given, are satisfactory in the cases of the B and C rotational constants, whereas a much larger discrepancy is seen for A. Comparison of the experimental and MP2 centrifugal distortion constants (Table 2) reveals good agreement in the case of ΔJ, whereas the absolute value of the experimental ΔJK is 33% larger than the MP2 equiv. There are also significant differences in the cases of δJ and ΦKJ. The MP2/cc-pVTZ procedure is obviously not sufficiently refined to predict accurate values for the centrifugal distortion constants and the differences between the equilibrium and ground-state rotational constants. However, more elaborate methods could not be employed due to lack of computational resources. Vibrationally Excited States of I. The ground-state spectrum was accompanied by several weaker satellite spectra with Stark and RFMWDR behavior very similar to that of the ground-state spectrum. Detailed assignments were obtained for four such spectra, which are assumed to belong to vibrationally excited states of I. These spectra are listed in the Supporting Information, Tables S13−S16 and their spectroscopic constants are displayed in Table 2. The maximum value of the principal quantum number J was 31 for all these excited states. The maximum value of K−1 was 17 found for the first excited state of the torsion (v21). The number of transitions of each excited state is given in Table 2. Three of the excited states, whose spectroscopic constants are listed in columns 3−5 of Table 2, are assumed to belong to successively excited states of the torsion about the O4−C5 bond of I (v21) because 2Pc increases upon successive excitation of this mode.33 The increase in 2Pc and the variation in the rotational constants from one excited state to the next are not constant, which would have been the case for a completely harmonic vibration. It is therefore concluded that the torsion has a significant contribution from anharmonicity. The frequency of the first excited state of the torsion about the O4−C5 bond (v21) was determined by relative intensity measurement to be 43(15) cm−1, whereas its anharmonic MP2 value is 49.8 cm−1 (Table S1). The increase in 2Pc upon excitation can be used to get a rough value of the torsional vibration. The torsional frequency, ω, is given roughly by ω = (67.5 × 10−20 u m2)/Δ cm−1,33 where Δ is the increase in 2Pc upon excitation. Using 1.67 × 10−20 u m2 for Δ obtained from the ground and first excited state (Table 2), one gets 40 cm−1 for the torsional vibration. The vibration−rotation constants (the α’s)23 are obtained as αA = 593.5(52), αB = −2.7387(86), and αC = −6.8881(89) MHz by subtraction of the first-excited-torsional-state rotational constants in Table 2 from the ground-state rotational constants in the same table. The corresponding MP2 values are (Table 1S) +572.56, −2.45, and −6.15 MHz. The agreement between the experimental and theoretical vibration−rotation constants is good given the approximate nature of the MP2 procedure. The value of 2Pc decreases for the last excited state (Table 2, column 6). A decrease is expected for a bending vibration that maintains the Cs symmetry.33 The MP2 anharmonic frequency of this lowest bending fundamental (v20) is 159.6 cm−1, whereas F


Article


A possible route to interstellar cyanomethyl formate could be a reaction between interstellar formic acid (HC(O)OH) and the interstellar cyanomethyl radical (CH2CN). A gas-phase reaction was modeled with the MP2 method, but it turned out that this reaction is not thermodynamically favorable. A catalyzed reaction on interstellar dust might be feasible. Other reaction paths other than the reaction between formic acid and the cyanomethyl radical are also conceivable given the large inventory of cyanides in interstellar space.

rather unreliable in the present case. Accurate predictions of quartic and especially sextic centrifugal distortion constants definitely require amuch more elaborate procedure than the MP2/cc-pVTZ method. Vibrationally Excited State. The spectrum of one vibrationally excited state was assigned in the same manner as described above for the ground-vibrational-state spectrum. A total of 655 transitions were assigned for this excited state (Table S18). The result of the least-squares analysis is given in Table 3. Relative intensity measurements yielded 60(15) cm−1 for this vibration, which is assumed to be the first excited state of the torsion about the O4−C5 bond (v21). The MP2 anharmonic value is 62.2 cm−1 (Table 2S). The rotational−vibrational constants obtained from columns 2 and 3 of Table 3 are αA = −243.7635(57), αB = 59.4867(17), and αC = 27.5865(17) MHz, whereas the MP2 method predicts −158.65, +46.44, and +21.65 MHz, respectively. Interestingly, the centrifugal distortion constants are generally much larger for the excited state than for the ground vibrational state. Energy Difference. The intensities of several transitions of the ground vibrational state of conformer I were compared with the intensities of transitions of the ground state of II observing the precaution of Esbitt and Wilson.34 The energy difference was calculated as described by these workers.34 The statistical weight of II was assumed to be 2, whereas I was assumed to have a statistical weight of 1. It was found that I is lower in energy than II by 1.4 kJ/mol. The uncertainty corresponding to three standard deviations was estimated to be ±0.6 kJ/mol. The MP2 method above corrected for zero-point vibrational energies predicts that II is lower in energy than I by 0.18 kJ/ mol, whereas the CCSD method predicts I to be lower in energy than II by 0.55 kJ/mol (Table 1). The CCSD electronic energy difference (0.55 kJ/mol) and the experimental energy difference of the ground vibrational states (1.4(6) kJ/mol) are in satisfactory agreement.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b05285. Microwave spectra of the ground and several vibrationally excited states of conformers I and II. Results of the theoretical calculations, including electronic energies, molecular structures, dipole moments, harmonic and anharmonic vibrational frequencies, rotational and centrifugal distortion constants, rotation−vibration constants, and 14N nuclear quadrupole coupling constants (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*H. Møllendal. Tel: +47 2285 5674. Fax: +47 2285 5441. Email: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Anne Horn for her skillful assistance. This work has been supported by the Research Council of Norway through a Centre of Excellence Grant (Grant No. 179568/V30). It has also received support from the Norwegian Supercomputing Program (NOTUR) through a grant of computer time (Grant No. NN4654K).

■

CONCLUSIONS The rotational spectra of two rotamers, denoted I and II, of cyanomethyl formate (HC(O)OCH2CN) have been assigned. Conformer I has a symmetry plane consisting of all atoms but the two hydrogen atom attached to cyanomethyl group. The cyanomethyl group is rotated 80.3° out of this plane in rotamer II. Conformer I is lower in energy by 1.4(6) kJ/mol relative to II. The ground-vibrational-state spectrum as well as spectra of four vibrationally excited states of I have been assigned, whereas the spectra of the ground vibrationally state and one excited state were assigned for II. A large number of transitions have been used to determine accurate spectroscopic constants for both rotamers. These constants should be able to predict very accurate frequencies for transitions occurring outside the investigated spectral interval (12−123 GHz) and thus facilitate a potential identification of interstellar cyanomethyl formate. The spectroscopic work has been augmented by advanced quantum chemical calculations at the MP2/cc-pVTZ and CCSD/cc-pVQZ levels of theory. The CCSD structures of the two rotamers, which are close to the Born−Oppenheimer equilibrium structures, were derived. The MP2 calculations predict centrifugal distortion constants and vibration−rotation constants that are only in fair agreement with their experimental counterparts. More refined and costly methods must be used to obtain reliable predictions of these parameters.

■

REFERENCES

(1) Brown, R. D.; Crofts, J. G.; Gardner, F. F.; Godfrey, P. D.; Robinson, B. J.; Whiteoak, J. B. Discovery of Interstellar Methyl Formate. Astrophys. J. 1975, 197, L29−L31. (2) Cazaux, S.; Tielens, A. G. G. M.; Ceccarelli, C.; Castets, A.; Wakelam, V.; Caux, E.; Parise, B.; Teyssier, D. The Hot Core Around the Low-Mass Protostar IRAS 16293−2422. Scoundrels Rule! Astrophys. J. 2003, 593, L51−L55. (3) Requena-Torres, M. A.; Martin-Pintado, J.; Rodriguez-Franco, A.; Martin, S.; Rodriguez-Fernandez, N. J.; de Vicente, P. Organic Molecules in the Galactic Center. Hot Core Chemistry Without Hot Cores. Astron. Astrophys. 2006, 455, 971−985. (4) Kobayashi, K.; Ogata, K.; Tsunekawa, S.; Takano, S. Torsionally Excited Methyl Formate in Orion KL. Astrophys. J. 2007, 657, L17− L19. (5) Takano, S.; Sakai, Y.; Kakimoto, S.; Sasaki, M.; Kobayashi, K. Detection of Methyl Formate in the Second Torsionally Excited State (vt = 2) in Orion KL. Publ. Astron. Soc. Jpn. 2012, 64, 89/81−89/89. (6) Carvajal, M.; Margulès, L.; Tercero, B.; Demyk, K.; Kleiner, I.; Guillemin, J. C.; Lattanzi, V.; Walters, A.; Demaison, J.; Wlodarczak, G.; Huet, T. R.; Møllendal, H.; Ilyushin, V. V.; Cernicharo, J. Rotational Spectrum of 13C2-Methyl Formate (HCOO13CH3) and Detection of the Two 13C-Methyl Formate in Orion. Astron. Astrophys. 2009, 500, 1109−1118. (7) Tercero, B.; Margulès, L.; Carvajal, M.; Motiyenko, R. A.; Huet, T. R.; Alekseev, E. A.; Kleiner, I.; Guillemin, J. C.; Møllendal, H.; G


Article


(28) Garrod, R. T.; Widicus Weaver, S. L.; Herbst, E. Complex Chemistry in Star-Forming Regions: An Expanded Gas-Grain WarmUp Chemical Model. Astrophys. J. 2008, 682, 283−302. (29) Klemperer, W. Astronomical Chemistry. Annu. Rev. Phys. Chem. 2011, 62, 173−184. (30) Schmitt-Kopplin, P.; Gabelica, Z.; Gougeon, R. D.; Fekete, A.; Kanawati, B.; Harir, M.; Gebefuegi, I.; Eckel, G.; Hertkorn, N. High Molecular Diversity of Extraterrestrial Organic Matter in Murchison Meteorite Revealed 40 Years After its Fall. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 2763−2768. (31) Ray, B. S. The Characteristic Values of an Asymmetric Top. Eur. Phys. J. A 1932, 78, 74−91. (32) Sørensen, G. O. Centrifugal Distortion Analysis of Microwave Spectra of Asymmetric Top Molecules. The Microwave Spectrum of Pyridine. J. Mol. Spectrosc. 1967, 22, 325−346. (33) Herschbach, D. R.; Laurie, V. W. Influence of Vibrations on Molecular Structure Determinations. III. Inertial Defects. J. Chem. Phys. 1964, 40, 3142−3153. (34) Esbitt, A. S.; Wilson, E. B. Relative Intensity. Rev. Sci. Instrum. 1963, 34, 901−907.

Cernicharo, J. Microwave and Submillimeter Spectroscopy and First ISM Detection of 18O-Methyl Formate. Astron. Astrophys. 2012, 538, A119/111−A119/113. (8) Margulès, L.; Huet, T. R.; Demaison, J.; Carvajal, M.; Kleiner, I.; Møllendal, H.; Tercero, B.; Marcelino, N.; Cernicharo, J. Rotational Spectrum and Tentative Detection of DCOOCH3-Methyl Formate in Orion. Astrophys. J. 2010, 714, 1120−1132. (9) Curl, R. F., Jr. Microwave Spectrum, Barrier to Internal Rotation, and Structure of Methyl Formate. J. Chem. Phys. 1959, 30, 1529−1536. (10) Neill, J. L.; Muckle, M. T.; Zaleski, D. P.; Steber, A. L.; Pate, B. H.; Lattanzi, V.; Spezzano, S.; McCarthy, M. C.; Remijan, A. J. Laboratory and Tentative Interstellar Detection of Trans-Methyl Formate Using the Publicly Available Green Bank Telescope PRIMOS Survey. Astrophys. J. 2012, 755, 153/151−153/111. (11) Belloche, A.; Garrod, R. T.; Müller, H. S. P.; Menten, K. M.; Comito, C.; Schilke, P. Increased Complexity in Interstellar Chemistry: Detection and Chemical Modeling of Ethyl Formate and n-Propyl Cyanide in Sagittarius B2(N). Astron. Astrophys. 2009, 499, 215−232. (12) Tercero, B.; Kleiner, I.; Cernicharo, J.; Nguyen, H. V. L; Lopez, A.; Munoz Caro, G. M. Discovery of Methyl Acetate and Gauche Ethyl Formate in Orion. Astrophys. J., Lett. 2013, 770 (L13), 16. (13) The Cologne Database for Molecular Spectroscopy, http:// www.astro.uni-koeln.de/cdms/molecules (Accessed April 2015). (14) Zuckerman, B.; Ball, J. A.; Gottlieb, C. A. Microwave Detection of Interstellar Formic Acid. Astrophys. J. 1971, 163, L41−L45. (15) Agúndez, M.; Fonfria, J. P.; Cernicharo, J.; Pardo, J. R.; Guélin, M. Detection of Circumstellar CH2CHCN, CH2CN, CH3CCH, H2CS. Astron. Astrophys. 2008, 479, 493−501. (16) Samdal, S.; Grønås, T.; Møllendal, H.; Guillemin, J.-C. Microwave Spectrum and Conformational Properties of 4-Isocyano1-butene (H2CCHCH2CH2NC). J. Phys. Chem. A 2014, 118, 1413−14099. (17) Wodarczyk, F. J.; Wilson, E. B., Jr. Radio Frequency-Microwave Double Resonance as a Tool in the Analysis of Microwave Spectra. J. Mol. Spectrosc. 1971, 37, 445−463. (18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Revision B.01; Gaussian, Inc.: Wallingford, CT, 2010. (19) Møller, C.; Plesset, M. S. Note on the Approximation Treatment for Many-Electron Systems. Phys. Rev. 1934, 46, 618−622. (20) Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M.; et al. MOLPRO, a Package of Ab Initio Programs, version 2010.1; University College Cardiff Consultants Limited: Cardiff, Wales, U.K., 2010. (21) Dunning, T. H., Jr. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron Through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (22) Watson, J. K. G. Vibrational Spectra and Structure; Elsevier: Amsterdam, 1977; Vol. 6. (23) Gordy, W.; Cook, R. L. Microwave Molecular Spectra; Techniques of Chemistry; John Wiley & Sons: New York, 1984; Vol. XVIII. (24) Riveros, J. M.; Wilson, E. B., Jr. Microwave Spectrum and Rotational Isomerism of Ethyl Formate. J. Chem. Phys. 1967, 46, 4605−4612. (25) Peng, Z.; Shlykov, S.; Van Alsenoy, C.; Geise, H. J.; Van der Veken, B. Joint Analysis of Ethyl Formate in the Gas Phase by Electron Diffraction and Microwave and Vibrational Spectroscopy Supplemented by Ab Initio Calculations of Force Fields. J. Phys. Chem. 1995, 99, 10201−10212. (26) Lopata, A. D.; Kuczkowski, R. L. Fluoromethyl Formate. Synthesis, Microwave Spectrum, Structure, Dipole Moment, and Anomeric Effect. J. Am. Chem. Soc. 1981, 103, 3304−3309. (27) Garrod, R. T.; Herbst, E. Formation of Methyl Formate and Other Organic Species in the Warm-up Phase of Hot Molecular Cores. Astron. Astrophys. 2006, 457, 927−936. H


Rotational Spectrum, Conformational Composition, and Quantum

Recommend Documents