The high resolution infrared spectroscopy of cyanogen di-n-oxide [onccno]

The high resolution infrared spectroscopy of cyanogen di-N-oxide
(ONCCNO)
Bujin GuoCentre for Molecular Beams and Laser Chemistry, Department of Chemistry, University of Waterloo,Waterloo, Ontario N2L 3G1, Canada Tibor Pasinszki and Nicholas P. C. WestwoodGuelph-Waterloo Centre for Graduate Work in Chemistry, Department of Chemistry and Biochemistry,University of Guelph, Guelph, Ontario N1G 2W1, Canada Peter F. BernathCentre for Molecular Beams and Laser Chemistry, Department of Chemistry, University of Waterloo,Waterloo, Ontario N2L 3G1, Canada and Department of Chemistry, University of Arizona,Tucson, Arizona 85721 ͑Received 4 May 1995; accepted 16 May 1995͒ The high-resolution infrared absorption spectrum of the oxalodinitrile di-N-oxide ͑ONCCNO͒molecule has been recorded in the gas phase with a Fourier transform spectrometer at a resolutionof 0.003 cmϪ1. No previous high-resolution spectra have been recorded for this semistablepalindromic molecule. On the basis of the 2:1 intensity alternation in the rotational lines caused bynitrogen nuclear spin statistics, the ONCCNO molecule appears to be linear. A quasilinear structure,however, cannot be ruled out at this stage of the analysis. The ␯4 and ␯5 fundamental modes at2246.040 55͑23͒ cmϪ1 and 1258.475 30͑11͒ cmϪ1 have been analyzed to give ground staterotational constants of B ϭ 0.042 202 10͑96͒ cmϪ1 and D0 8.77͑70͒ϫ10Ϫ10 cmϪ1. By fixing the CN and NO bond lengths to 1.1923 and 1.1730 Å, respectively, the C–C bond length wasdetermined to be 1.3329 Å using the B0 value. This short C–C bond length is thus similar to thatobserved for a carbon–carbon double bond. 1995 American Institute of Physics. INTRODUCTION
freshly prepared solution of ONCCNO in n-hexane beforepolymerization occurs to make polyfuroxan.3 Since dilute so- The cyanogen di-N-oxide ͑oxalodinitrile di-N-oxide, lutions are stable at 0 °C for several hours, ONCCNO has ONCCNO͒ molecule was first prepared at the beginning of been widely used in organic chemistry for 1,3-dipolar cy- this century.1 However, its chemical formula and isolation were not achieved until the early 1960s by Grundmann.2,3 The ONCCNO molecule is a candidate for astrophysical The ONCCNO molecule was prepared in organic solutions observation since it contains only the relatively abundant el- by HCl elimination from the stable dichloroglyoxime͑ ements C, N, and O. The symmetric CNO dimer structure is HONvC͑Cl͒–͑Cl͒CvNOH͒ precursor. Two strong infrared also of spectroscopic and structural interest since there is the absorption peaks at 2190 and 1235 cmϪ1 were observed in possibility of quasilinear behavior as found in HCNO.5 The CCl4 solution. The ultraviolet spectrum, which contains symmetric linear CNO dimer ONCCNO has no dipole mo- maxima at 312, 295, and 262 nm, can be obtained with a ment so that pure rotational spectra will be very weak andinfrared or ultraviolet observations will be necessary.
u CNO antisymmetric stretching mode of the ONCCNO molecule. Note the lines due to the HNCO molecule near2270 cmϪ1. ͑b͒ An expanded portion of the R branch 2:1 intensity alterna- FIG. 2. Portions of R and P branches of the ␯4 mode of the ONCCNO tion caused by the nuclear spin statistics of two equivalent nitrogen nuclei in molecule. The J quantum numbers are for the fundamental band and the perturbed R͑96͒ and P͑98͒ lines are marked by arrows.
J. Chem. Phys. 103 (9), 1 September 1995
Guo et al.: Infrared spectroscopy of ONCCNO TABLE I. The line list for the ␯4 vibrational mode of ONCCNO ͑cmϪ1͒.
J. Chem. Phys., Vol. 103, No. 9, 1 September 1995 Guo et al.: Infrared spectroscopy of ONCCNO aObserved minus calculated line positions.
Pure samples of ONCCNO are difficult to handle. Crys- The high-resolution absorption spectra were recorded talline, monomeric ONCCNO is stable at Ϫ78 °C but begins with a Bruker FTS 120 HR spectrometer at the University of to visibly decompose near Ϫ45 °C and explodes a few min- Waterloo with a resolution of 0.003 cmϪ1. The ␯4 mode, near utes later at that temperature.6 Maier and Teles7 have re- 2250 cmϪ1, was recorded with an InSb detector and 46 scans ported the formation of ONCCNO by flash vacuum pyrolysis were coadded in the 1800–2900 cmϪ1 region. A redpass fil- of dichloroglyoxime followed by condensation of the pyroly- ter with a cutoff at 2900 cmϪ1 set the upper wave number sis products diluted with argon on a cold ͑10 K͒ window, limit while the lower wave number limit was set by the band although they published no spectroscopic data. Until very gap of the InSb detector. A HgCdTe detector was used to recently no gas phase measurements have been made. Pasin- record the ␯5 mode at 1260 cmϪ1 and the spectrum was szki and Westwood8 have successfully studied gaseous ON- obtained by coadding 55 scans. Another redpass filter ͑cutoff CCNO by He I photoelectron spectroscopy, photoionization at 1672 cmϪ1͒ set the high wave number limit and the mass spectrometry, low-resolution midinfrared spectroscopy HgCdTe detector response set the lower limit. A KBr beam- as well as ab initio calculations. The strong IR absorption bands at 2226 and 1260 cmϪ1 correspond to the antisymmet-ric ͑␯ ␴ ͒ vibrations, respectively, which were reported earlier at 2190 ANALYSIS
cmϪ1 and 1235 cmϪ1 in CCl4 solution.2 Unfortunately, the abinitio calculations did not provide a solid conclusion as to The spectral analysis program PC-DECOMP, developed whether the ONCCNO molecule was linear or bent although by J. W. Brault, was used for the spectral line measurements.
a linear or quasilinear structure was preferred. High- Using this program, the line profiles were fitted with Voigt resolution gas phase spectroscopy can supply additional in- lineshape functions. The signal-to-noise ratio for the stron- formation on the molecular geometry.
gest lines in the spectrum was about 5:1 and the precision of In this work, we report on our high-resolution Fourier the line position measurement is better than Ϯ0.0006 cmϪ1 transform infrared spectra of the cyanogen di-N-oxide mol-ecule. The ␯4 and ␯5 vibrational modes have been recorded inthe gas phase and these two fundamental bands have beenrotationally analyzed.
EXPERIMENT
The ONCCNO molecule was generated in situ using the same method as described by Pasinszki and Westwood.8Briefly, the thermolysis of dichloroglyoxime in a quartz tube͑8 mm i.d. by 15 cm͒ heated to 550 °C gives a good yield ofONCCNO plus HCl with only trace amounts of the sideproducts, NO, CO, CO2, and HNCO. The experimental setupis typical for absorption spectroscopic work using a cell anda glower external to the spectrometer. The infrared glowerwas collimated by a parabolic mirror and passed through a20 cm long absorption cell equipped with KBr windows and entered the spectrometer through the emission port. The ther- FIG. 3. ͑a͒ An overview of the ␯5 u CNO symmetric stretching mode of theONCCNO molecule. ͑b͒ An expanded portion of the P branch of the ␯ molysis products were pumped slowly through the gas cell at mode. The marked lines are for the fundamental band and show the 2:1 intensity alternation caused by nuclear spin statistics.
J. Chem. Phys., Vol. 103, No. 9, 1 September 1995 Guo et al.: Infrared spectroscopy of ONCCNO TABLE II. The line list for the ␯5 vibrational mode of ONCCNO ͑cmϪ1͒.
J. Chem. Phys., Vol. 103, No. 9, 1 September 1995 Guo et al.: Infrared spectroscopy of ONCCNO aObserved minus calculated line positions.
for these lines. However, many of the weaker lines and the ceeded in the same way as for ␯4. The ␯5 mode also had a blended features were determined only to a precision of high line density but no perturbations were present to help the assignment. The rotational assignment of ␯5 was made by An interactive color Loomis–Wood computer program changing the relative assignment of the P and R branches was used to pick out the branches and helped to assign the until the ground state combination differences matched those spectra. The CO molecule was present in the cell as a side of ␯4. About 200 lines were assigned and the complete line product of the thermolysis. The measured spectral lines of the ␯4 vibrational mode were calibrated with the CO linesusing line positions taken from the literature.9 For the ␯5 The rotational constants
vibrational mode, we used the water absorption forcalibration.10 The line positions ͑Tables I and II͒ of the two fundamen- tal bands were fitted together in a global least-squares fit.
The ␯4 fundamental mode
Figure 1 shows the high-resolution spectrum of the CNO F͑J͒ϭ␯ ϩBJ͑Jϩ1͒ϪD͓J͑Jϩ1͔͒2ϩH͓J͑Jϩ1͔͒3 antisymmetric stretching mode ͑␯ ␴ ͒ HNCO molecule is present on the high wave number side of was used in the fit and the resulting rotational constants are the overview spectrum ͓Fig. 1͑a͔͒. Figure 1͑b͒ is an ex- panded portion of the R branch of the ␯4 fundamental. The 0.042 202 10 cmϪ1 ͑1.265 187 GHz͒, is in good agreement alternation in intensities is caused by the nuclear spin statis- with the ab initio prediction8 of 1.24ϳ1.25 GHz. The cen- tics of two equivalent ͑Iϭ1͒ nitrogen atoms. The symmetri- trifugal distortion constants of the excited ␯4 vibrational level cally located nitrogen atoms cause an intensity alternation of may be perturbed by interaction with other modes.
2:1 with the even JЉ values being stronger.
The main problem in the analysis was the presence of a DISCUSSION AND CONCLUSION
large number of hot bands which causes a line density ashigh as 100 lines/cmϪ1. It was impossible to locate the band The semistable cyanogen di-N-oxide ͑ONCCNO͒ mol- origin because of the overlapping hot bands. The final J as- ecule can be generated in situ with a good yield in the gas signments were achieved with the help of a small local per- phase from the precursor molecule dichloroglyoxime by turbation at JЈϭ97 in the upper level. Figure 2 shows the thermolysis. The high-resolution spectra of the ONCCNO perturbed R͑96͒ and P͑98͒ lines. The R͑96͒ line was shifted molecule have been obtained for the two strongest vibra- by 0.005 cmϪ1 to higher wave numbers while the P͑98͒ line tional modes ␯4 and ␯5. The rotational analysis of these two was shifted by the same amount. There are several other modes and the intensity alternation of the lines indicates that small local perturbations in this band which also confirm the assignments. In total, 230 rotational lines were assigned forthe ␯4 band and they are listed in Table I.
TABLE III. The rotational constants for ONCCNO ͑in cmϪ1͒. One standard The ␯5 fundamental mode
The high-resolution spectrum of the CNO symmetric portion of the P branch is shown in Fig. 3͑b͒, again demon- strating the 2:1 intensity alternation due to nitrogen nuclear spin statistics. The measurement and assignment of ␯5 pro- J. Chem. Phys., Vol. 103, No. 9, 1 September 1995 Guo et al.: Infrared spectroscopy of ONCCNO TABLE V. The C–C bond lengths ͑r ͒ brS bond lengths from HCNO, Ref. 11.
cSee text for details.
dBond length calculated using experimental rotational constant.
Indications from the ab initio calculations, and compari- sons with similar molecules, suggest that the important CC achieved. For example, the possibility of quasilinear behav- bond is quite short.8 The single B value obtained in this work ior cannot be ruled out at this stage of the analysis. We plan is insufficient to provide this parameter without some as- additional experiments to record spectra of the much weaker sumptions about the CN and NO bond lengths. One approach bending vibrational modes as well as combination bands.
to this problem is to use the known CN and NO bond lengths This work will help assign the numerous hot bands that we in the parent HCNO molecule,11 and assuming a linear struc- have measured and will provide information on the missing ture, extract the CC value from the rotational constant ͑B ͒ determined in this work. Table IV, structure I, shows thestructure of ONCCNO, predicted by this method. This, of ACKNOWLEDGMENTS
course, assumes that the CN and NO bond lengths are trans- We thank the Natural Science and Engineering Research Council of Canada ͑NSERC͒ for the support of this research.
A possible improvement on this method is to see how a T.P. thanks NSERC for the award of a NATO Science Fel- computational method ͓MP3͑full͒/6-31G*͔8 performs for lowship. Partial support was provided by the Petroleum Re- both HCNO and ONCCNO, and then ‘‘correct’’ the CN and search Fund and the NASA laboratory astrophysics program.
NO values and apply them to ONCCNO. With these ‘‘re-fined’’ CN and NO bond lengths, the mutual effect of back- 1 W. Steinkopf and B. Ju¨rgens, J. Prakt. Chem. 83, 453 ͑1911͒.
to-back CNO groups should be taken into account, and, us- 2 Ch. Grundmann, Angew. Chem. 75, 450 ͑1963͒; Angew. Chem. Int. Ed.
ing the experimental rotational constant, a value for the CC Engl. 2, 260 ͑1963͒.
bond length is obtained. This assumes that the differences Ch. Grundmann, V. Mini, J. M. Dean, and H.-D. Frommeld, Justus Liebigs
Ann. Chem. 687, 191 ͑1965͒.
between the CN and NO bond lengths in HCNO and ONC- 4 ͑a͒ N. E. Alexandrou and D. N. Nicolaides, J. Chem. Soc. C, 2319 ͑1969͒; CNO are predicted correctly by the MP3 method. Table IV, ͑b͒ D. N. Nicolaides and T. A. Kouimtzis, Chem. Chron. 3, 63 ͑1974͒; ͑c͒
structure II, shows the revised values for ONCCNO. This A. Gu¨l, A. I. Okur, A. Cihan, N. Tan, and O structure which takes into account the interaction of the two ¨ . Bekaroglu, Z. Anorg. Allg. Chem. 496, 197
͑1983͒; ͑e͒ Y. Go¨k and O¨. Bekaroglu, Synth. React. Inorg. Met.-Org.
CNO groups, indicates that the NO bond length has slightly Chem. 11, 621 ͑1981͒; ͑f͒ V. Ahsen, E. Musluoglu, A. Gu¨rek, A. Gu¨l, O
decreased, while the CN bond length has slightly increased, Bekaroglu, and M. Zehnder, Helv. Chim. Acta 73, 174 ͑1990͒; ͑g͒ Y. Go¨k
in accord with expectations based on a more delocalized and S. Serin, Synth. React. Inorg. Met.-Org. Chem. 18, 975 ͑1988͒; ͑h͒ Y.
Go¨k and A. Demirbas, ibid. 19, 681 ͑1989͒.
5 B. P. Winnewisser, Molecular Spectroscopy: Modern Research, edited by Both the directly transferred CN and NO lengths, and the K. N. Rao ͑Academic, New York, 1985͒, Vol. 3.
corrected values, lead to the result that the CC bond length is 6 Ch. Grundmann and P. Gru¨nanger, The Nitrile Oxides ͑Springer-Verlag, about 1.333 Å, approximately the same as that in ethylene, 7 G. Maier and J. H. Teles, Angew. Chem. 99, 152 ͑1987͒; Angew. Chem.
H2CCH2, a molecule of known double bond character. Table Int. Ed. Engl. 26, 155 ͑1987͒.
V shows CC bond lengths for a range of molecules. An im- 8 T. Pasinszki and N. P. C. Westwood, J. Am. Chem. Soc. ͑to be published͒.
portant observation is that the experimental CC bond length 9 A. G. Maki and J. S. Wells, Wavenumber Calibration Tables from Hetero- in cyanogen, NC-CN is 1.389 Å,12 already much shorter than dyne Frequency Measurements ͑NIST Special Publication 821, Washing-ton DC, 1991͒.
a typical CC single bond and we anticipate that this will 10 R. A. Toth, J. Opt. Soc. Am. B 8, 2236 ͑1991͒.
decrease upon adding terminal oxygen atoms ͑compare 11 ͑a͒ H. K. Bodenseh and M. F. Winnewisser, Z. Naturforsch. A 24, 1973
͑1969͒; ͑b͒ B. P. Winnewisser, M. F. Winnewisser, and F. Winther, J. Mol.
MP3͑full͒/6-31G* calculations for NCCN predict CC to be Spectrosc. 51, 65 ͑1974͒.
12 A. G. Maki, J. Chem. Phys. 3, 3193 ͑1965͒.
1.393 Å, with similar calculations for ONCCNO predicting 13 W. J. Lafferty and E. K. Plyler, J. Chem. Phys. 37, 2688 ͑1962͒.
CC to be 1.361 Å.8 The inference from this is that the CC 14 C. C. Costain, J. Chem. Phys. 29, 864 ͑1958͒.
length in ONCCNO is at least 0.03 Å less than that in 15 H. K. Bodensen and K. Morgenstern, Z. Naturforsch 25a, 150 ͑1970͒.
A. A. Westenberg and E. B. Wilson, Jr., J. Am. Chem. Soc. 72, 199 ͑1950͒.
17 H. C. Allen, Jr. and E. K. Plyler, J. Am. Chem. Soc. 80, 2673 ͑1958͒.
Clearly, experimental and theoretical studies on this mol- 18 A. Baldacci, S. Ghersetti, S. C. Hurlock, and K. N. Rao, J. Mol. Spectrosc.
ecule have a long way to go before a full characterization is 59, 116 ͑1976͒.
J. Chem. Phys., Vol. 103, No. 9, 1 September 1995

Source: http://bernath.uwaterloo.ca/media/141.pdf