Journal of Molecular Structure 612 (2002) 207±211
www.elsevier.com/locate/molstruc
Detection of the bending vibration of the CO±orthoN2 complex q L.A. Surin a,b,*, H.S.P. MuÈller a, E.V. Alieva b, B.S. Dumesh b, G. Winnewisser a, I. Pak a,b a
I. Physikalisches Institut, UniversitaÈt zu KoÈln, ZuÈlpicher Str. 77, D-50937 Cologne, Germany b Institute of Spectroscopy, Russian Academy of Sciences, 142190 Troitsk, Moscow, Russia Received 8 June 2001; accepted 9 July 2001
Abstract The bending vibration of the CO±N2 complex has been investigated in the millimeter wave range from 130 to 155 GHz using an intracavity OROTRON jet spectrometer. Six transitions, P(2), P(1), R(0), R(1), R(2), and R(3) from the K 0 ground state to the K 0 bending state of the orthoN2 spin modi®cation were measured and analyzed. Nuclear quadrupole structure due to the presence of two equivalent 14N nuclei was partly resolved and analyzed to give information about the angular anisotropy of the interaction potential. The frequency of the bending vibration was determined to be 139892.459(35) MHz. The nuclear quadrupole coupling constant for the K 0 bending state of CO±orthoN2 was obtained for the ®rst time to be xaa 20:768
43 MHz: The different value and sign of this constant from the one in the K 0 ground state,
xaa 10:19641 MHz suggests that the orientation and motion of the N2 subunit are very different in these two states. q 2002 Elsevier Science B.V. All rights reserved. Keywords: Millimeter wave; van der Waals complexes; Bending vibration; OROTRON
1. Introduction The CO±N2 complex is one of the most fundamental and simplest among the van der Waals (vdW) complexes. This system, containing N2, the main atmospheric component, and CO, a signi®cant atmospheric pollutant, is of great interest because of its direct relevance to the earth's atmosphere. Spectroscopically, CO±N2 is closely related to some extensively studied CO-containing complexes, such as Ar± CO [1,2] Ne±CO [3±6], He±CO [7±9], CO±H2 [10,11] and CO±D2 [12,13]. The K 1±0 spectra of these complexes due to CO motion were measured by q
This paper is dedicated to Professor Paolo G. Favero and Professor Helmut Dreizler in appreciation of their signi®cant contributions to the ®eld of microwave spectroscopy. * Corresponding author. Address: I. Physikalisches Institut, UniversitaÈt zu KoÈln, ZuÈlpicher Str. 77, D-50937 Cologne, Germany. Tel.: 149-221-470-35-60; fax: 149-221-470-51-62. E-mail address:
[email protected] (L.A. Surin).
our group. They display the effects of large-amplitude internal motion and consequently cannot be described with a simple semirigid asymmetric rotor model. Spectroscopic study of CO±N2 is also important, because of its close relationship to the CO dimer, which has been the subject of considerable progress in the past few years [14±16]. Theoretical work on the CO±N2 intermolecular potential comprises only the semiempirical result reported by Franken and Dykstra [17]. More extensive and detailed potentials have been developed for the CO±CO [18,19] and N2 ±N2 [20,21] dimers. Since these molecules are isoelectronic, it may be expected that the potential for CO± N2 will be something like an appropriate combination of the potentials of these two dimers. The ®rst experimental studies of CO±N2 were made in the 4.7 mm infrared (IR) region of the CO stretching vibration by Kawashima and Nishizawa [22] using a pulsed molecular beam and by Xu and McKellar [23] using a continuous slit-jet nozzle
0022-2860/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. PII: S 0022-286 0(02)00091-1
208
L.A. Surin et al. / Journal of Molecular Structure 612 (2002) 207±211
observed, which involved an excited bending state with an energy of about 4.7 cm21. In the present paper, we report the ®rst observation of this bending vibration in the millimeter wave region using the OROTRON spectrometer [2]. Six transitions, P(2), P(1), R(0), R(1), R(2), and R(3), associated with the ground and bending state K 0 levels of the orthoN2 spin modi®cation were measured and analyzed. Nuclear quadrupole splitting due to the two equivalent 14N nuclei was partly resolved yielding additional information about the angular orientation and dynamics of the N2 subunit within the complex. The quadrupole splitting also facilitated the assignments. 2. Experimental Fig. 1. A recording of the R(0) transition together with a simulation of the spectrum obtained from the ®nal set of spectroscopic constants.
expansion, both combined with a diode laser spectrometer. The latter authors [23] performed a more detailed analysis, assigning four connected subbands to K 0, 1, and 2 states, as well as one separate subband to a K 1 state. The four connected subbands were attributed to transitions involving the rotationless jN2 0 states, associated with orthoN2 with total nuclear spin I I1 1 I2 0 and 2, where I1 and I2 are the nuclear spins of the two identical 14N nuclei. The separate subband was tentatively assigned to transitions involving jN2 1 states, associated with paraN2 with total nuclear spin I 1: The observed spectra revealed a quite regular structure, similar to those of rare gas±CO complexes. This simplicity arose from the fact that all observed states correlated with jN2 0; where the N2 unit is in a pure rotationless state, or with the state jN2 1 and jCO 0; where the CO unit is in a pure rotationless state. Using the prediction from these IR works [22,23], the microwave observations [24±26] of K 0, 1 states for both spin modi®cations as well as millimeter wave study [25] of K 1±0 transitions for CO±orthoN2 have now been made. In the very recent work of Xia, McKellar and Xu [27], the IR spectrum of the CO±N2 complex has been studied further in detail, extending the previous tentative assignment of just one subband in CO± paraN2 to include over 10 linked subbands. In the same work, two new subbands in CO±orthoN2 were
The millimeter wave measurements of the bending vibration of CO±orthoN2 were made using an intracavity OROTRON jet spectrometer. A detailed description of the spectrometer is given in Ref. [2]. Brie¯y, a millimeter wave generator OROTRON and a supersonic jet apparatus were placed in a vacuum chamber, which was evacuated by a diffusion pump with 1000 l/s capacity. The molecular beam was produced by a pulsed solenoid valve with opening diameter of 1.0 mm and pulse duration of 500 ms, operating at a repetition rate of 5±10 Hz. The molecular beam was injected into the OROTRON cavity perpendicular to the cavity axis. The high Q-factor of the cavity resulted in 100 effective passes of the radiation through the jet. The absorption in the cavity was detected very sensitively by measuring the corresponding change of the OROTRON collector current. In the process of measurements, the frequency of the OROTRON was modulated at 25 kHz by a sine-wave. The collector signal was demodulated by a digital lock-in ampli®er operated in 2f mode with a time constant of 160 ms. The output of the lock-in ampli®er was then processed by a boxcar integrator with a time constant of 1 s. For frequency measurements, a small part of the radiation was coupled out of the cavity and mixed with the radiation of a MW synthesizer. The typical sample mixture contained 1% CO and 1% N2 in Ne with a backing pressure of about 3 atm. For most of the absorption measurements, the full linewidth at half height (FWHH) is about 300 kHz and the accuracy is estimated to be about 50 kHz.
L.A. Surin et al. / Journal of Molecular Structure 612 (2002) 207±211
209
Table 1 Observed transitions (MHz) for the K 00 K 0 0 bending vibration of CO±orthoN2, uncertainties (kHz) and residuals o2c (kHz) J 0 ±J 00
I 0 ,F 0 ±I 00 ,F 00
Frequency
unc.
o2c
1±2
2,2±2,2; 2,2±2,3; 0,1±2,2; 2,2±2,1; 2,2±0,2; 0,1±2,1; 0,1±0,2 2,3±2,3; 2,3±2,4 2,1±2,1; 2,1±0,2; 2,1±2,0
130711.097 130711.463 130711.945 135437.600 144071.583 144071.908 144072.336 147974.897 147975.414 151603.925 151604.481 154962.595 154962.845 154963.257
50 50 50 500 50 50 50 50 50 70 50 100 100 50
5 28 26 2310 12 5 6 211 211 28 18 26 6 23
0±1 1±0 2±1 3±2 4±3
2,2±2,2; 0; 1±0,0; 0,1±2,2 2,3±2,2 2,1±0,0; 2,1±2,2 2,2±2,2; 2,2±0,1; 2,3±2,2; 2,3±2,3 2,4±2,3; 0,2±2,2; 0,2±0,1; 2,1±2,1; 0,2±2,1 2,3±2,2; 2,3±2,3; 2,4±2,3 2,2±2,1; 2,5±2,4; 0,3±0,2 2,4±2,3 2,5±2,4 2,3±2,2; 2,6±2,5; 0,4±0,3; 2,2±2,1
3. Assignment and analysis Six transitions in the frequency range from 130 to 155 GHz were measured and assigned on the basis of the IR prediction [27]. These are 4 R-branch and 2 Pbranch parallel transitions from the K 0 ground state to the ®rst excited bending state of CO±orthoN2. The differences between the observed and predicted frequencies were less than 10 MHz. Assignment of R(0) and P(2) lines was further con®rmed through the combination difference from precise MW data [25]. Most of the observed lines display a partly resolved 14N hyper®ne structure. As an example, the R(0) transition is shown in Fig. 1 together with a simulation of the spectrum obtained from the ®nal set of spectroscopic constants. The measured transitions frequencies, assignments, uncertainties, and residuals are listed in Table 1. The line frequencies were used to ®t simultaneously the band origin s , the rotational and quartic centrifugal distortion constant B and D, respectively, along with the nuclear quadrupole coupling constant x aa employing an exact diagonalization program [28]. Each line consisting of overlapping hyper®ne components was treated as intensity weighted average of the individual components. Contributions of components weaker than the strongest one by more than a factor of 10 were not considered. Because of the overlap and the proximity of the hyper®ne components, simulations were made to determine the differences between the intensity weighted average line positions and the
apparent peak positions. Second derivatives line shapes were calculated in which each hyper®ne component was represented by a Gaussian line pro®le of 290 kHz FWHH. The corrections applied were between 10 and 40 kHz. The spectroscopic constants of the n bend;CO 0 state [25] were kept ®xed in the analysis. Other ®ts were also tried in which only the K 0 FTMW or all nCO 0 data were included in the ®ts. The effects on the nbend;CO 1 state parameters were well within the uncertainties. The hyper®ne analysis was fairly straightforward because the patterns for the P(2) and R(0) transitions depend largely on the ground and excited state x aa constants. In fact, it was found that for the nbend;CO 1 state this constant was suf®cient to describe the hyper®ne structure well within the uncertainties. Contributions from x bb, nuclear spin±rotation coupling, and from centrifugal distortion effects on the quadrupole coupling were found to be negligible. In addition, the sextic distortion constant H was insigni®cantly determined as 20.044 (140) £ 10 23 MHz, and it was omitted from the ®nal ®t. The resulting spectroscopic constants are given in Table 2 together with previous values from an infrared study. The present and the previous values for the band origin s and the rotational constant B are in good agreement. The values for D are similar. A comparison of CO±N2 nuclear quadrupole coupling constants is presented in Table 3. The absolute standard deviation excluding P(1) is 8.9 kHz. This as well as the reduced standard deviation of 0.234 suggest the experimental uncertainties of
210
L.A. Surin et al. / Journal of Molecular Structure 612 (2002) 207±211
species, indicating that the CO internal rotation is relatively free. Another strong indication that the CO±N2 complex cannot be considered as a semirigid rotor, is that the coupling constant for the ground state
x aa 0:19641 MHz [25] and the bending state
xaa 20:768 are not only very different in magnitude, but they also have opposite signs. The ®rst determination of the quadrupole coupling constant for the bending state is the most important result here, because it gives information about the angular orientation and dynamics of the N2 subunit within the complex. For the ground state of CO±orthoN2, it was established [25] that the complex has an approximate T-shaped geometry with N2 forming the top and CO the leg, and that the oxygen atom is on average closer to the N2 subunit than the carbon atom. For the bending state studied here, the quite large and negative x aa value indicates that the N2 subunit is remarkably localized in parallel direction with respect to the intermolecular axis. The present study con®rms the previous IR [22,23,27] and MW [24±26] investigations, that the CO±N2 complex is a rather complicated molecule, which is not expected to be close to any limiting case or to have simple energy level patterns or good quantum numbers. The simplicity of our analysis and the similarity to the case of the triatomic CO±rare complexes re¯ect only the fact that all observed states correlate with the jN2 0 rotationless states of the complex, i.e. orthoN2 modi®cation. The N2 unit may be attributed to a rare gas atom having a mass of 28 amu. However, the observation and analysis of the other states with jN2 . 0 will be more complicated.
Table 2 Spectroscopic constants (numbers in parentheses are one standard deviation in units of the least signi®cant ®gures) (MHz) of the K 00 K 0 0 bending vibration of CO±orthoN2
s B D £ 10 3 H £ 10 3 x aa a
Present
Previously a
139892.459 (35) 2089.7495 (106) 32.90 (50)
139904 (6) 2089.12 (44) 7.5 (69) 20.205 (29)
20.768 (43)
Ref. [27].
the transition frequencies, and thus of the spectroscopic constants, to be rather conservative.
4. Discussion The new information presented here concerns the lowest bending state of CO±orthoN2, which was precisely found to lie above the ground state by 139892.477(35) MHz or 4.6663038(12) cm 21. This value may be compared with values for the CO±rare gas complexes. The bending energy for He±CO is 5.390357(3) cm 21 [8,9], for Ne±CO this value is 8.5805(4) cm 21 [4], whereas for Ar±CO the appropriate energy has been determined to 12.0142927(17) cm 21 [29]. The bending frequencies for CO±paraH2 and CO±orthoD2, which have also relevance to CO±rare gas complexes, are 7.0794 cm 21 and 6.9781 cm 21, respectively [10,12]. The value for CO±N2 is thus lowest among these Table 3 Comparison of CO±N2 nuclear quadrupole coupling constants (MHz) Modi®cation
n bend
K 0, x aa
K2 1
x aa Ortho Ortho Para Para a b c d
0 1 0 1
0.196 20.77 b
a
25.112
d
Ref. [25]. This work. Ref. [26]. Ref. [26]; a value of 8.44 MHz was obtained for x bb.
21.039
K1 1
x bb a
0.524 c
0.063
x aa
x bb
0.535 c
23.002 c
a
20.013 c
L.A. Surin et al. / Journal of Molecular Structure 612 (2002) 207±211
During the search for CO±orthoN2 bending vibration, many unassigned millimeter wave transitions of different intensity and different hyper®ne structure have been observed. We suppose, that these transitions occur between the states of the paraN2 modi®cation, namely jCO 1±0; jN2 1: The lowest state of CO±paraN2 corresponding to jCO 0; jN2 1 was well characterized very recently in MW study by Xu and JaÈger [26]. This state consists of the three closely spaced stacks with a projection of total angular moment J on the intermolecular axis K 0,1. The higher state of CO±paraN2 with excited rotation of the CO unit, i.e. jCO 1 and jN2 1; is more complicated. This state includes nine stacks with K 0, 1, 2, which are located 4±6 cm 21 above the ground state of CO±paraN2 [27]. The detection and assignment of these not yet observed stacks will be a subject of future millimeter wave studies. In summary, this work presents the ®rst millimeter wave observation of the bending vibration of CO± orthoN2 providing precise information about the vibrational frequency and the rotational constants for this state. The determination of the 14N nuclear quadrupole coupling constant of the bending state makes an important step in the characterization of the structure and internal dynamics of the weakly bound complex CO±N2. Acknowledgements The work was supported by the DFG via Research Grant SFB-494 and by the Ministry of Science and Technology of the Land NRW. The work of B.S.D. and E.V.A. at Cologne was made possible by the DFG through a grant aimed to support Eastern and Central European Countries and Republics of the FSU. L.A.S. thanks the Alexander von Humboldt Foundation for a research fellowship. The support through grants of the Russian Foundation for Basic Research (RFBR: 0002-17525 and 00-02-17606) and the Russian Ministry of Science (`Physics of Microwaves') is gratefully acknowledged by B.S.D.
211
References [1] A.R.W McKellar, Mol. Phys. 98 (2000) 111 and references therein. [2] L.A. Surin, B.S. Dumesh, F. Lewen, D.A. Roth, V.P. Kostromin, F.S. Rusin, G. Winnewisser, I. Pak, Rev. Sci. Instrum. 72 (2001) 2535. [3] K.A. Walker, T. Ogata, W. JaÈger, M.C.L. Gerry, I. Ozier, J. Chem. Phys. 106 (1997) 7519. [4] A.R.W. McKellar, M.-C. Chan, Mol. Phys. 93 (1998) 253. [5] G. Winnewisser, B.S. Dumesh, I. Pak, L.A. Surin, F. Lewen, D.A. Roth, F.S. Rusin, J. Mol. Spectrosc. 192 (1998) 243. [6] D.A. Roth, I. Pak, L.A. Surin, B.S. Dumesh, G. Winnewisser, Z. Naturforsch. 55a (2000) 754. [7] A.R.W. McKellar, Y. Xu, W. JaÈger, C. Bissonnette, J. Chem. Phys. 110 (1999) 10766. [8] L.A. Surin, D.A. Roth, I. Pak, B.S. Dumesh, F. Lewen, G. Winnewisser, J. Chem. Phys. 112 (2000) 4064. [9] L.A. Surin, D.A. Roth, I. Pak, B.S. Dumesh, F. Lewen, G. Winnewisser, J. Chem. Phys. 112 (2000) 9190 Erratum. [10] A.R.W. McKellar, J. Chem. Phys. 108 (1998) 1811. [11] I. Pak, L.A. Surin, B.S. Dumesh, D.A. Roth, F. Lewen, G. Winnewisser, Chem. Phys. Lett. 304 (1999) 145. [12] A.R.W. McKellar, J. Chem. Phys. 112 (2000) 9282. [13] L.A. Surin, B.S. Dumesh, G. Winnewisser, I. Pak, J. Chem. Phys. 113 (2000) 9351. [14] M.D. Brookes, A.R.W. McKellar, J. Chem. Phys. 111 (1999) 7321. [15] D.A. Roth, L.A. Surin, B.S. Dumesh, G. Winnewisser, I. Pak, J. Chem. Phys. 113 (2000) 3034. [16] K.A. Walker, C. Xia, A.R.W. McKellar, J. Chem. Phys. 113 (2000) 6618. [17] K.A. Franken, C.E. Dykstra, J. Phys. Chem. 97 (1993) 11408. [18] M. Rode, J. Sadlej, R. Moszynski, P.E.S. Wormer, A. van der Avoird, Chem. Phys. Lett. 314 (1999) 326. [19] A.W. Meredith, A.J. Stone, J. Phys. Chem. A 102 (1998) 434. [20] J.R. Stallcop, H. Partridge, Chem. Phys. Lett. 281 (1997) 212. [21] A. Wada, H. Kanamori, S. Iwata, J. Chem. Phys. 109 (1998) 9434. [22] Y. Kawashima, K. Nishizawa, Chem. Phys. Lett. 249 (1996) 87. [23] Y. Xu, A.R.W. McKellar, J. Chem. Phys. 104 (1996) 2488. [24] Y. Kawashima, Y. Ohshima, Y. Endo, Chem. Phys. Lett. 315 (1999) 201. [25] Y. Xu, W. JaÈger, L.A. Surin, I. Pak, V.A. Pan®lov, G. Winnewisser, J. Chem. Phys. 111 (1999) 10476. [26] Y. Xu, W. JaÈger, J. Chem. Phys. 113 (2000) 514. [27] C. Xia, A.R.W. McKellar, Y. Xu, J. Chem. Phys. 113 (2000) 525. [28] H.M. Pickett, J. Mol. Spectrosc. 148 (1991) 371. [29] M. Hepp, R. Gendriesch, I. Pak, F. Lewen, G. Winnewisser, J. Mol. Spectrosc. 183 (1997) 295.