High-Resolution Millimeter Wave Spectroscopy and Ab Initio

Jan 20, 2015 - Institute of Radio Astronomy of NASU, Chervonopraporna 4, 61002 Kharkiv, Ukraine. ¶. Ecole Nationale Supérieure de Chimie de Rennes, ...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

High-Resolution Millimeter Wave Spectroscopy and Ab Initio Calculations of Aminomalononitrile Roman A. Motiyenko,*,† Laurent Margulès,† Eugen A. Alekseev,‡ and Jean-Claude Guillemin¶ †

Laboratoire de Physique des Lasers, Atomes et Molécules, UMR CNRS 8523, Université de Lille 1, F-59655 Villeneuve d’Ascq Cedex, France ‡ Institute of Radio Astronomy of NASU, Chervonopraporna 4, 61002 Kharkiv, Ukraine ¶ Ecole Nationale Supérieure de Chimie de Rennes, CNRS, UMR 6226, 11 Allée de Beaulieu, CS 50837, F-35708 Rennes Cedex 7, France S Supporting Information *

ABSTRACT: The HCN trimer aminomalononitrile (H2NCH(CN)2, AMN) is considered as a key compound in prebiotic chemistry and a potential candidate for detection in the interstellar medium. In this view, we studied the rotational spectrum of AMN in the 120−245 GHz frequency range. The spectroscopic work was augmented by high-level ab initio calculations. The calculations showed that between two existing rotamers, symmetric and asymmetric, the most stable is the asymmetric conformation, and it is the only conformation observed in the recorded spectra. The symmetric conformation is 6.7 kJ/mol higher in energy and thus has a very low Boltzmann factor. The analysis of the rotational spectra of the A conformation has shown that the observed lines exhibit a doublet or quartet structure owing to two large-amplitude motions, C−N torsion and amino group inversion. To study the large-amplitude motions in detail, we calculated a two-dimensional potential energy surface and determined the barrier heights for the torsion and inversion, Vt = 12.5 kJ/mol and Vi = 19.1 kJ/mol. About 2500 assigned rotational transitions in the ground vibrational state were fitted within experimental accuracy using the reduced axes system Hamiltonian. The set of obtained spectroscopic parameters allows accurate calculation of transition frequencies and intensities for an astrophysical search of AMN.

■

INTRODUCTION Hydrogen cyanide (HCN) has been detected for a long time in the interstellar medium (ISM),1 in the atmosphere of Titan,2 and in comets.3 Adenine, an oligomeric compound of HCN with the (HCN)5 formula, is a constituent of DNA, RNA, and many coenzymes and can be easily formed in water starting from HCN.4 Adenine is thus often considered as a key compound of the prebiotic period of the Primitive Earth.4−6 Attempts to detect it in the ISM are unsuccessful up to date but the E-cyanomethanimine (HNCH−CN), a compound with (HCN)2 formula and an intermediate often postulated in the synthesis of adenine in prebiotic conditions,4,7−9 has been recently detected.10 In the hypothesis of the formation of adenine in one or both of these media by subsequent additions of HCN monomer on cyanomethanimine, the first step should give a compound with a (HCN)3 formula. Even if several isomers with a C3H3N3 formula can be drawn, the aminomalononitrile (aminopropanedinitrile, AMN) is often considered as the most likely derivative issued from the addition of HCN on cyanomethanimine.5,11,12 Furthermore, in all of the mechanistic studies on the synthesis of adenine from HCN, AMN has been postulated as an intermediate,5,6,9 and several reactions starting from the latter have given adenine.5,7−9,11 On the other hand, AMN is a useful intermediate in the synthesis of many heterocyclic systems.13−19 Such properties designate AMN as a key compound in prebiotic chemistry and an attractive candidate for the ISM. Its detection in the ISM, after © 2015 American Chemical Society

HCN and the cyanomethanimine, would be a determinant to improve our understanding of the chemistry of HCN in the Universe. How could AMN be synthesized in the ISM? The formation of cyanomethanimine in significant amounts by gas-phase synthetic routes is considered as not compatible with lowtemperature interstellar environments even for acid-catalyzed reaction pathways.20,21 However, even if the cyanomethanimine is formed in the solid phase, it has been detected in the gas phase, and it could react on the grain surfaces or in the gas phase after vaporization to form AMN. The reaction pathway could involve the radical addition of CN on cyanomethanimine, followed by abstraction of a radical hydrogen from another molecule or reaction with a hydrogen radical if the energy of the reaction can be released; see Scheme 1. Scheme 1. Proposed Reaction Pathway in the Synthesis of AMN from Cyanomethanimine

Received: December 18, 2014 Revised: January 15, 2015 Published: January 20, 2015 1048

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

Article

The Journal of Physical Chemistry A

Conventional Absorption Spectroscopy Experiment. Owing to the relatively high kinetic instability of AMN, the major part of experimental measurements of its rotational spectrum has been performed using the Lille BWO-based fast scan spectrometer26 and in the so-called “flow mode”. During the experiment, the absorption cell, which consisted of 1.2 m Pyrex glass tube with Teflon windows, was kept at room temperature, whereas the sample was cooled at a temperature of about −15 °C and evaporated outside of the cell. Then, the sample was continuously injected through a side opening at one end of the cell and pumped out through another side opening at the other end. The optimum gas pressure in the cell was kept close to 20 Pa. Using the fast-scan spectrometer, the spectral region of 120−180 GHz was covered by the measurements. In addition, we recorded the rotational spectrum of AMN in a relatively narrow frequency range of 225−242 GHz using the spectrometer based on solid-state sources. The major impurity observed in the recorded spectrum was the HCN monomer, but for the spectral analysis, this compound does not cause any problem because it has only one strong triplet line in the frequency range of the experiment at 177.261 GHz. We did not find any other significant impurity, at least not any having a permanent dipole moment, because all of the strongest lines observed were assigned to AMN. The recorded spectrum of AMN is very dense, which may be explained by relatively small values of rotational constants, as well as by a large partition function, owing, in particular, to five excited vibrational states lying below 300 cm−1.

The identification of a species in the ISM requires that its spectrum has first been assigned in the laboratory. We present here the results of the analysis of the AMN rotational spectrum as a support for future astrophysical observations. Our experimental work has been augmented by high-level quantum chemical calculations, which were undertaken to obtain information for use in assigning the rotational spectrum and investigating properties of the potential energy hypersurface. Two stable conformations, symmetric (S) and asymmetric (A), exist for AMN. Their structure and atom numbering are illustrated in Figure 1. The two conformations differ by

■

QUANTUM CHEMICAL METHODS The present quantum chemical calculations were performed using the Gaussian 09 suite of programs.27 The geometries were fully optimized using the “tight” convergence option, and the vibrational frequencies were calculated at the MP2 level. Peterson and Dunning’s correlation-consistent aug-cc-pVTZ basis set, which is of triple-ζ quality and includes both diffuse and polarized functions, was used. To analyze the LAMs in AMN, we computed a two-dimensional optimized potential energy surface (PES). Such calculation is rather costly; therefore, it was computed also at the MP2 level but with a smaller 6-311++G(2p, 2d) basis set. There are multiple ways to define LAM coordinates. For example, in the case of a similar problem for methylamine, the torsional angle may be defined as a simple dihedral angle28 involving a hydrogen atom of the methyl group or an average of six dihedral angles24 involving hydrogen atoms of both amino and methyl groups. In this case, the inversion angle for methylamine may be defined as the angle between the NH2 plane and the extension of the C−N bond.24 In the present study, we define LAM coordinates in a simpler way especially from the point of view of quantum chemical calculations. To describe two LAMs, we used two dihedral angles D8 (H2−C1−N7−H8) and D9 (H2−C1−N7− H9). For the computation, the angle D9 was varied from 0 to 355° with a step of 5°, and the angle D8 was varied from D9 − 100° to D9 + 100° also with a step of 5°. During the computations, both dihedral angles were kept fixed, whereas all others’ structural parameters (dihedrals, angles, and bond lengths) were optimized. Relative energies of the conformations and transition states (TSs) as a result of ab initio calculations are presented in Table 1. Calculated geometrical parameters and harmonic vibrational frequencies for both conformers are given, respectively, in Tables S1 and S2 in the Supporting Information. The resultant

Figure 1. Calculated structure and atom numbering of A (a) and S (b) conformations of AMN.

orientation of the NH2 group. In the S conformation, both hydrogens (H8 and H9) of the amino group are in syn configuration (±60°) relative to the H2 hydrogen. The S conformation has a plane of symmetry and thus belongs to the Cs symmetry point group. In the A conformation, one of the amino group hydrogens is in anti configuration (180°), and the other hydrogen is in a skew position relative to the H2 hydrogen. The rotation around the C−N bond connects the S and A conformations. Because of the large-amplitude character of the C−N torsion and NH2 wagging, the amino H atoms are symmetry-equivalent, and there are four equivalent configurations for the A conformation. A similar problem of two large-amplitude motions (LAMs) that complicate the rotational spectra exists for methylamine,22 ethylamine,23 and protonated methanol.24

■

EXPERIMENTAL SECTION Preparation of AMN. AMN was prepared as previously reported25 starting from AMN p-toluenesulfonate. The latter compound was purchased from Aldrich and used without further purification. The synthesis details are presented in Appendix A of the Supporting Information. 1 H NMR (400 MHz, CDCl3) δ 2.18 (d brd, 2H, 3JHH = 8.9 Hz, NH2); 4.78 (t, 1H, 3JHH = 8.9 Hz, CH). 13C NMR (100 MHz, CDCl3) δ 34.3 (d, 1JCH = 151.1 Hz, CH); 115.6 (s, CN). 1049

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

Article

The Journal of Physical Chemistry A

respect to the C1−N7 bond. On the calculated PES, there are two equivalent first-order saddle points that have the following values of dihedral angles: D8 = 0°, D9 = 180° or D8 = 180°, D9 = 0°. The harmonic force field calculation for the TS-inv state provides one imaginary frequency. Two equivalent secondorder inversion TSs are characterized by the dihedrals D8 = 90°, D9 = 270° or D8 = 270°, D9 = 90°. It is interesting to note that in the case of amino group inversion in methylamine, the values of the first- and second-order inversion barriers are very close29 and do not allow one to clearly determine the preferable tunneling pathway. The calculation performed for methylamine at the MP2 level and using the 6-311G++ basis set yielded the values of V(1) = 27.6 kJ/mol and V(2) = 27.9 kJ/mol, and the i i DFT/B3LYP calculations with the same basis set yielded the values of V(1) = 21.4 kJ/mol and V(2) = 21.7 kJ/mol. In the i i present case, the symmetry break for AMN leads to two distinctive barriers with rather different heights V(1) i = 19.1 kJ/ mol and V(2) = 21.5 kJ/mol. Thus, in the case of AMN, the i inversion through the first-order TS, which directly connects two minima of the A conformation, is preferred.

Table 1. Relative Energies with Respect to the A Conformer of AMNa conformation

ΔEa (kJ/mol)

A S TS-rot TS-inv (1) TS-inv (2)

0.0 6.72 (6.25) 12.5 (12.2) 19.1 (18.8) 21.5 (20.4)

a

The ZPE-corrected energies are reported in parentheses. For the definition of TSs, see the text and Figure 2.

PES is shown in Figure 2. For a convenient graphical representation, the PES is plotted as a function of two

■

GROUP THEORY Four equivalent configurations of the A conformation may lead to complications in the description of its rotational spectrum. The A conformation of AMN does not possess any symmetry; however, for the description of tunneling motions, a molecular symmetry group may be used. By labeling the atoms from 1 to 9, as shown in Figure 1, one can see that the following permutation-inversion operations, E, (89), (35)(46)*, (35) (46)(89)*, are feasible owing to the tunneling trough the barriers of finite height. The (35)(46)(89)* operation is feasible owing to the torsional tunneling, the (35)(46)* operation corresponds to the tunneling through inversion barrier, and finally, the (89) is the product of two others operations mentioned above, that is, it corresponds to the combination of torsional and inversion motions. These four permutation-inversion operations represent G4 molecular symmetry group having four irreducible representations, A1, A2, B1, and B2. Thus, from a molecular symmetry group point of view, owing to four structurally equivalent configurations, each rotational level of the A conformation of AMN has four-fold degeneracy, which may be removed by tunneling through torsion and inversion barriers. Therefore, in addition to usual quantum numbers, the rotational levels of AMN may be now labeled according to the irreducible representations of the G4 group. For the parent isotopic species of AMN, the elements of the G4 group for AMN exchange are two nitrogen nuclei having spin IN = 1, two carbon nuclei having zero spin, and two hydrogen nuclei having spin IH = 1/2. Our calculations show that in this case, the statistical weights for the irreducible representations A1, A2, B1, and B2 are correspondingly 1, 1, 3, and 3. According to the ab initio calculations, the A conformation of AMN has three nonzero dipole moment components, μa = 1.1 D, μb = 3.3 D, and μc = 1.0 D. The operations of the G4 group that contain the inversion of coordinates change the sign of only the μa projection, keeping the signs of μb, and μc. Therefore, in the ground vibrational state, the electric dipole transitions owing to μb can occur only between two sublevels having the same symmetry, that is, A1 ↔ A1, A2 ↔ A2, B1 ↔ B1, and B2 ↔ B2. Consequently, the a- and c-type transitions have the following selection rules, A1 ↔ A2 and B1 ↔ B2, and the transitions between A and B levels are forbidden.

Figure 2. Calculated PES for AMN as a function of two independent coordinates D9 + D8 and D9 − D8. See the text for the definition of the coordinates.

independent coordinates D9 + D8 and D9 − D8, and it is repeated over 720° for the D9 + D8 coordinate. As follows from the analysis of Table 1 and Figure 2, the A conformer is the most stable one. For both A and S conformations, only positive vibrational frequencies were obtained, indicating that they correspond to a minimum (stable conformation) on the potential energy hypersurface. The S conformation is less stable than the A by about 6.7 kJ/mol (zero-point energy (ZPE)corrected value: 6.2 kJ/mol). The calculated PES contains six distinct minima; four of them correspond (A1, A2, A3, and A4) to the A conformation and two others (S1 and S2) to the S conformation. In the adopted coordinate system, the internal rotation around the C1−N7 bond keeps fixed the D9 − D8 coordinate either at 120 or at 240° and varies the D9 + D8 coordinate. Thus, it connects two A configurations and one S configuration. The TSs related to the C1−N7 internal rotation (TS-rot) were characterized using the “QST3” procedure implemented in the Gaussian 09 software package at the MP2 level of theory and using the augcc-pVTZ basis set. Two saddle points, exhibiting only one imaginary frequency, have been found. The calculations showed that both TSs have the same energy of 12.5 kJ/mol (ZPEcorrected value: 12.2 kJ/mol) with respect to the most stable A conformation. The TSs that correspond to the inversion of amino group hydrogens (TS-inv) were searched for using the “TS” procedure of the Gaussian 09. As a result, two different saddle point geometries were obtained. The first-order TS is characterized by planar orientation of all hydrogen atoms with 1050

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

Article

The Journal of Physical Chemistry A

Figure 3. A part of the experimental spectrum of AMN in the 167−173 GHz frequency range showing (a) the intense series of bR1,±1 transitions and (b−e) a detailed view of selected transitions exhibiting torsion-inversion splittings. For these transitions, the assignment according to irreducible representations of the G4 molecular symmetry group and the relative intensities are also indicated.

as ΔJ, and δJ centrifugal distortion terms. The frequency predictions calculated on the basis of the experimental parameters allowed the assignment of the next series of quartets with the same selection rules and Ka = 1 and 2 and Kc = J − 1. These series of transitions also exhibited doublet structure (Figure 3c) that was attributed to tunneling effects. In a similar manner, the series of transitions with Ka = 2 and 3 and Kc = J − 2 was assigned. This series, in contrast to the two previous ones, had a barely resolvable quartet structure that provided important information on symmetry assignment. The quartet consisted of two unresolved strong transitions in the center and two weaker transitions on each side. The relative intensities of strong and weak transitions 3 to 1 correspond to statistical weights of the G4 group irreducible representations. Thus, one can assign the strong lines as belonging to B symmetry species and weaker ones to A symmetry species. The fact that no A lines were resolved for Ka = 0,1 and Ka = 1,2 series may be explained by the inverse order of B and, correspondingly, A lines for Ka = 2,3 and other series (see the assignment of the lines on Figure 3a). The corresponding A symmetry species transitions for Ka = 0,1 and Ka = 1,2 are therefore mixed with B species transitions, and they are not distinguished under the Doppler-limited resolution of our experiment. The inversion in the order of B lines is an indicator of a perturbation, and an appropriate model should be used to treat correctly this problem. In the present study, we used the model based on the reduced axis system (RAS) approach proposed by Pickett.30 It is well-suited for molecules having one ore two LAMs with a double minimum-potential. In the case of AMN, the S conformer is considerably higher in energy that the A conformer. Thus, the C−N torsion in the A conformer may be considered as a LAM between only two equivalent minima at least for the ground vibrational state. As has been shown by Coudert,31 the RAS method is equivalent to the internal axis method developed and successfully used by Hougen and coworkers for molecules with several LAMs.22,32−34 The

In the study of the rotational spectrum of gauche-ethylamine, Fischer and Botskor23 used different symmetry labeling for a similar problem of two LAMs connecting four equivalent minima. Starting from geometrically equivalent localized total wave functions for each of the four minima, they combined them first to produce symmetrical (S) and antisymmetrical (A) wave functions with respect to the torsional motion. Then, to treat the inversion, they combined the wave functions of the same torsional symmetry to obtain even (e) and odd (o) functions for the inversion. From a group theory point of view, in ref 23, the produced delocalized wave functions having symmetry labeling se, so, ae, and ao belong correspondingly to A1, A2, B1, and B2 irreducible representations of the G4 molecular symmetry group introduced in this paper.

■

ASSIGNMENT AND RESULTS The assignment started with the search for the most intense lines. In the frequency range of the experiment, these were expected to be μb-type transitions with a ΔJ = ±1 selection rule (see Figure 3a). First, on the basis of the predictions obtained from ab initio rotational and centrifugal distortion constants, we searched for Ka =0 and 1 and Kc = J bR±1,1 transitions because they are less subjected to centrifugal distortion correction and, thus, they are usually better predicted. These transitions form a series of easily distinguishable doublets with splittings that reduce with J increasing. In the recorded spectra, we located and assigned such transitions with the values of the J quantum number ranging from 32 to 42. The predicted Ka = 0 ← 1 and Ka = 1 ← 0 splittings for these transitions were much smaller than the Doppler-limited resolution of the experiment; therefore, the doublet structure observed (see Figure 3b) was attributed to tunneling effects. By taking the central frequency of each doublet, we fitted the observed transitions to the Watson A-reduction Hamiltonian in the Ir coordinate representation, and we produced the first experimental set of rotational parameters containing B and C constants (the A rotational constant was kept fixed to its ab initio value), as well 1051

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

Article

The Journal of Physical Chemistry A Table 2. Spectroscopic Constants of the A Conformer of AMN rotation parameters A (MHz) B (MHz) C (MHz) ΔJ (kHz) ΔJK (kHz) ΔK (kHz) δJ (kHz) δK (kHz) ΦJ (Hz) ΦJK (Hz) ΦKJ (Hz) ΦK (Hz) ϕJ (Hz) ϕJK (Hz) ϕK (Hz) Nc σ (MHz)d σwe

torsion-inversion

ground state b

5742.99822(40) 2884.81743(23) 2037.38599(21) 1.39923(12) −6.65593(57) 21.9327(15) 0.552005(27) 1.41647(36) 0.005616(20) −0.032406(98) 0.01784(59) 0.1724(17) 0.0026531(48) −0.001611(84) 0.11947(36) 1405 0.027 0.68

theorya

parameters

B species

5731.39 2874.87 2035.42 1.418 −6.916 22.51 0.5589 1.355

E* (MHz) E*J (MHz) E*K (MHz) E*JK (kHz) E*KK (kHz) Fab (MHz) FabJ (kHz) FabK (kHz) Fac (MHz) FacK (kHz)

2981.563(35) 0.004855(13) 0.05513(20) −0.001643(74) 0.00339(24) 5.6258(34) 0.0759(30) −0.729(54) 17.7973(61) −0.5635(84)

A species 3362.1(65) 0.05597(46)

5.516(49)

17.808(13) −0.502(12)

a

MP2/aug-cc-pVTZ calculations. bNumbers in parentheses are two standard deviations in the same units as the last digit. cNumber of distinct frequency lines. dRoot-mean-square deviation of the fit. eWeighted root-mean-square deviation of the fit.

momentum operator and its components. One may also notice that in the present study, we fit a single set of pure rotational and centrifugal distortion parameters for all four tunneling substates. Indeed, such an approach seems to be reasonable because in the case of ethylamine, the average values of rotational constants for A and B states are almost identical.23 In total, in the rotational spectrum of AMN, we assigned about 1400 distinct frequency lines with 10 ≤ J ≤ 74 and 0 ≤ Ka ≤ 23 that correspond to about 2500 rotational transitions of the A conformation. About 80% of the assigned lines are of B symmetry species because their intensities are favored by spin statistics. Due to extreme density of the spectrum and lower intensities caused by statistical weights, the assignment of A species lines was more difficult. The correctness of the assignment of B symmetry perturbed transitions was confirmed by the identification of weak μa-type lines. As was mentioned before, the recorded rotational spectrum of AMN is very dense. Owing to their strong intensities the μb-type lines were thus easily assigned. The intensities of the μa-type transitions calculated using the ab initio value of the μa dipole moment component are much weaker and, in comparison with the experimental spectrum, are at the confusion limit of the experiment. The μa-type lines connect B1 and B2 levels; therefore, their frequencies depend on the energy difference ΔEB between these two levels. The ΔEB term is usually highly correlated with Coriolis coupling terms. Incorrect assignment of μb-type transitions would produce incorrect values of Coriolis coupling parameters, thus making impossible the assignment of transitions, whose frequencies depend on the energy difference ΔEB. One should note that the extreme density of the rotational spectrum of AMN does not allow simple recognition of a pattern for such a series of transitions, which would be possible in the case of less dense and more intense spectrum. Thus, the correct assignment of μa-type transitions was made only on the basis of correct frequency predictions obtained from the analysis of the perturbed μb-type transitions.

perturbation terms in the RAS Hamiltonian are equivalent in this case to off-diagonal Coriolis coupling terms. The choice of the RAS method is also determined by the fact that it is implemented in widely used SPFIT/SPCAT programs for fitting and predicting molecular spectra. We applied the RAS method to treat the perturbations connecting two B levels and connecting two A levels or in terms of the labeling scheme introduced by Fischer and Botskor for ethylamine, “between the levels of different torsional parity but belonging to the same inversion state”. As was discovered during the spectral analysis, no perturbation terms connecting A and B levels were needed to fit the observed spectra within experimental accuracy. The Hamiltonian used has the following form ⎛ H − HA ⎞ 0 0 HIA Δ ⎜ rot ⎟ A ⎜ HA ⎟ 0 0 Hrot + HΔ I ⎟ H=⎜ B B ⎜ ⎟ 0 0 Hrot − HΔ HI ⎜ ⎟ ⎜ B B⎟ 0 0 HI Hrot + HΔ ⎠ ⎝ (1)

where Hrot is the standard Watson A-reduction Hamiltonian in the Ir coordinate representation and HIA and H IB are perturbation terms respectively for A and B symmetry species. In addition, the pure rotational part of the Hamiltonian is defined in such a way that it allows one to fit averaged rotational and centrifugal distortion constants for both tunneling substates.35 This procedure seemed to be more robust to possible correlation problems between different rotational and Coriolis coupling parameters. Fitting averaged constants is achieved by introducing HAΔ and HBΔ terms, which may be defined as HΔ = E* + E*J J 2 + EK*Jz2 + E2*(J+2 + J−2 ) + ...

(2)

where E* is a half-energy difference between two A or B levels (ΔE = 2E*) and J, Jz, and J± = Jx ± iJy are the rotational angular 1052

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

The Journal of Physical Chemistry A

■

The whole data set of the assigned transitions was fitted within experimental accuracy using a RAS Hamiltonian consisting of 31 parameters. The frequencies of assigned rotational transitions are presented in Table S2 of the Supporting Information. The rotational parameters obtained as the result of the fit are listed in Table 2. The unique set of rotational constants allows more straightforward comparison with corresponding ab initio values that are also listed in Table 2. As follows from the analysis of the parameters in Table 2, in addition to the rotational constants, we determined the full set of quartic and sextic centrifugal distortion constants. All of them agree well with corresponding values from ab initio calculations. The final results of the fit allowed us to perform the correct assignment of the tunneling substates. For this purpose, we applied the energy splittings scheme made for ethylamine by ref 23 (see Figure 5 of ref 23), and we used the energy differences ΔEA and ΔEB from our fit. The assumption of the scheme validity for AMN is supported by the fact that the values of the barriers to torsion and inversion motions in ethylamine and AMN are relatively close.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +33(0) 230434490. Fax: +33(0)20337020. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The Centre National d’Études Spatiales (CNES), the Action sur Projets de l’INSU, “Physique et Chimie du Milieu Interstellaire”, and ANR-13-BS05-0008-02 IMOLABS are acknowledged for financial support.

■

REFERENCES

(1) Snyder, L. E.; Buhl, D. Observations of Radio Emission from Interstellar Hydrogen Cyanide. Astrophys. J., Lett. 1971, 163, L48. (2) Hanel, R.; et al. Infrared Observations of the Saturnian System from Voyager 1. Science 1981, 212, 192−200. (3) Schloerb, F. P.; Kinzel, W. M.; Swade, D. A.; Irvine, W. M. Observations of HCN in Comet P/Halley. Astron. Astrophys. 1987, 187, 475−480. (4) Oró, J. Mechanism of Synthesis of Adenine from Hydrogen Cyanide under Possible Primitive Earth Conditions. Nature 1961, 191, 1193−1194. (5) Hudson, J. S.; Eberle, J. F.; Vachhani, R. H.; Rogers, L. C.; Wade, J. H.; Krishnamurthy, R.; Springsteen, G. A Unified Mechanism for Abiotic Adenine and Purine Synthesis in Formamide. Angew. Chem. 2012, 124, 5224−5227. (6) Wang, J.; Gu, J.; Nguyen, M. T.; Springsteen, G.; Leszczynski, J. From Formamide to Adenine: A Self-Catalytic Mechanism for an Abiotic Approach. J. Phys. Chem. B 2013, 117, 14039−14045. (7) Ferris, J. P.; Orgel, L. Aminomalononitrile and 4-Amino-5cyanoimidazole in Hydrogen Cyanide Polymerization and Adenine Synthesis. J. Am. Chem. Soc. 1965, 87, 4976−4977. (8) Sanchez, R. A.; Ferbis, J. P.; Orgel, L. E. Studies in Prebiodc Synthesis: II. Synthesis of Purine Precursors and Amino Acids from Aqueous Hydrogen Cyanide. J. Mol. Biol. 1967, 30, 223−253. (9) Roy, D.; Najafian, K.; von Ragué Schleyer, P. Chemical Evolution: The Mechanism of the Formation of Adenine under Prebiotic Conditions. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 17272− 17277. (10) Zaleski, D. P.; et al. Detection of E-Cyanomethanimine toward Sagittarius B2(N) in the Green Bank Telescope PRIMOS Survey. Astrophys. J., Lett. 2013, 765, L10. (11) Ferris, J. P.; Orgel, L. Studies in Prebiotic Synthesis. I. Aminomalononitrile and 4-Amino-5-cyanoimidazole. J. Am. Chem. Soc. 1966, 88, 3829−3831. (12) Ferris, J.; Edelson, E. Chemical Evolution. 31. Mechanism of the Condensation of Cyanide to Hydrogen Cyanide Oligomers. J. Org. Chem. 1978, 43, 3989−3995. (13) Johnson, T. B.; Nicolet, B. H. Researches on Pyrimidines. LXVI. The Formation of Pyrimidines from Diethyl Aminomalonate and Amino Malonicnitrile. J. Am. Chem. Soc. 1914, 36, 345−355. (14) Rayner, B.; Tapiero, C.; Imbach, J. Synthetic Nucleosides. IX. Some α Ribofuranonucleosides. J. Carbohydr., Nucleosides, Nucleotides 1976, 3, 1−14. (15) Hosmane, R. S.; Lim, B. B.; Burnett, F. N. Rearrangements in Heterocyclic Synthesis: A Novel Translocation of an (N-Amino-Nmethylamino) methylene Group from a Heterocyclic N-Amino-Nmethylformamidine Side Chain to the Vinylogous Nitrile Function. J. Org. Chem. 1988, 53, 382−386. (16) Birkett, P. R.; Chapleo, C. B.; Mackenzie, G. Synthesis of Novel Imidazotriazepines and Their Conversion to 9-(2-Ethylaminoethyl) adenine and 9-(2-Ethylaminoethyl) hypoxanthine. Synthesis 1991, 1991, 822−824.

■

CONCLUSIONS The results of the present study have shown that two conformations exist for the AMN molecule. Quantum chemical calculations indicate that the A conformer is preferred over the S conformer by 6.7 kJ/mol. This assumption is in general agreement with the experimental spectrum, where only the A conformer was observed. Quantum chemical calculations and spectroscopic analysis also show evidence for two LAMs in the AMN molecule. The analysis of the calculated two-dimensional PES indicates that four equivalent configurations exist for the A conformation of the molecule. The torsional motion along the C−N bond, the amino group inversion, or a combination of two motions may interconnect these four configurations. From ab initio calculations, the barriers to torsion and inversion are estimated to be, respectively, 12.5 and 19.1 kJ/mol. Notwithstanding relatively high barriers, we observed the tunneling splittings in the millimeter wave rotational spectrum of the A conformation measured with Doppler-limited resolution. One should also note that for the S conformation of AMN, there are two equivalent configurations that may be interconnected via tunneling motions. Thus, each rotational energy level of the S conformation is split into two sublevels, and one could also observe tunneling splittings in the rotational spectrum. The analysis of the observed transitions performed using the RAS approach allowed us to fit all of the observed transitions within experimental accuracy. As a result, the set of Hamiltonian parameters may be used to accurately predict frequencies and intensities of AMN rotational lines beyond the frequency range of the experiment and for transitions with J ≤ 75 and Ka ≤ 23. Thus, the spectroscopic information obtained in this study provides a solid basis for the astrophysical detection of AMN.

■

Article

ASSOCIATED CONTENT

* Supporting Information S

Synthesis of AMN, calculated molecular structure and harmonic vibrational frequencies, rotational line assignments, measured frequencies, experimental uncertainties, and deviations from the final fit. This material is available free of charge via the Internet at http://pubs.acs.org. 1053

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054

Article

The Journal of Physical Chemistry A (17) Corelli, F.; Summa, V.; Brogi, A.; Monteagudo, E.; Botta, M. Chiral Azole Derivatives. 2. Synthesis of Enantiomerically Pure 1Alkylimidazoles. J. Org. Chem. 1995, 60, 2008−2015. (18) Brieden, W.; Fuchs, R. Process for the Production of 3-Alkoxy-5alkylpyrazin-2-amines. Eur. Pat. Appl. EP 820993 A2 19980128, 1998. (19) Shibata, K.; Yoshida, M.; Doi, T.; Takahashi, T. Derivatization of a Tris-oxazole Using Pd-Catalyzed Coupling Reactions of a 5Bromooxazole Moiety. Tetrahedron Lett. 2010, 51, 1674−1677. (20) Smith, I.; Talbi, D.; Herbst, E. The Production of HCN Dimer and More Complex Oligomers in Dense Interstellar Clouds. Astron. Astrophys. 2001, 369, 611−615. (21) Yim, M. K.; Choe, J. C. Dimerization of HCN in the Gas Phase: A Theoretical Mechanistic Study. Chem. Phys. Lett. 2012, 538, 24−28. (22) Ohashi, N.; Hougen, J. T. The Torsional-Wagging Tunneling Problem and the Torsional-Wagging-Rotational Problem in Methylamine. J. Mol. Spectrosc. 1987, 121, 474−501. (23) Fischer, E.; Botskor, I. The Microwave Spectrum of gaucheEthylamine. J. Mol. Spectrosc. 1984, 104, 226−247. (24) Bhatta, R. S.; Gao, A.; Perry, D. S. A Comparative Ab Initio Study of Torsion−Inversion Coupling in CH3NH2, CH3OH2+ and CH3CH2•. J. Mol. Struct.: THEOCHEM 2010, 941, 22−29. (25) Moser, R.; Claggett, A.; Matthews, C. Peptide Formation from Aminomalononitrile (HCN Trimer). Tetrahedron Lett. 1968, 9, 1605− 1608. (26) Alekseev, E.; Motiyenko, R.; Margulès, L. Millimeter- and Submillimeter-Wave Spectrometers on the Basis of Direct Digital Frequency Synthesizers. Radio Phys. Radio Astron. 2012, 3, 75−88. (27) Frisch, M. J. et al. Gaussian 09, revision D.01; Gaussian Inc.: Wallingford, CT, 2009. (28) Smeyers, Y. G.; Villa, M.; Senent, M.-L. Ab Initio Determination of the Torsion-Wagging and Wagging-Bending Infrared Band Structure Spectrum of Methylamine. J. Mol. Spectrosc. 1998, 191, 232−238. (29) Kim, H.-W.; Zeroka, D. Internal-Rotation and Inversion Potential Energy Surfaces for Methylamine and Methylphosphine. Int. J. Quantum Chem. 2008, 108, 974−982. (30) Pickett, H. M. Vibration−Rotation Interactions and the Choice of Rotating Axes for Polyatomic Molecules. J. Chem. Phys. 1972, 56, 1715−1723. (31) Christen, D.; Müller, H. S.; Coudert, L. The g′Ga Conformer of Ethylene Glycol: A Test Case for Tunneling-Rotation Approaches Developed for Non-Rigid Molecules. 57th International Symposium on Molecular Spectroscopy; Ohio State University, 2002; TB03. (32) Hougen, J. T. A Generalized Internal Axis Method for High Barrier Tunneling Problems, as Applied to the Water Dimer. J. Mol. Spectrosc. 1985, 114, 395−426. (33) Ohashi, N.; Hougen, J. T. The Torsional-Wagging Tunneling Problem and the Torsional-Wagging-Rotational Problem in Hydrazine. J. Mol. Spectrosc. 1985, 112, 384−400. (34) Coudert, L.; Hougen, J. Tunneling Splittings in the Water Dimer: Further Development of the Theory. J. Mol. Spectrosc. 1988, 130, 86−119. (35) Christen, D.; Müller, H. S. The Millimeter Wave Spectrum of aGg′ Ethylene Glycol: The Quest for Higher Precision. Phys. Chem. Chem. Phys. 2003, 5, 3600−3605.

1054

DOI: 10.1021/jp512625s J. Phys. Chem. A 2015, 119, 1048−1054