Infrared absorption bands of a- and -pinenesin the 8-14-gm atmospheric window region

H. William Wilson

There are bands in the 8 -14-Am atmospheric window region of the gas phase IR spectra of a- and -pinenesthat are intense enough to allow theoretically the naturally occurring terpenes to be determined in the 100-ppbv level at 1-km path lengths. Because of the broadness of the bands, however, and the relatively intenseatmospheric background at their frequencies of 786.5 cm-' and 880 cm-1 , respectively, practical concentra-tion limits for the use of the bands are probably not better than about 1-10 ppmv. Both type C bands have4-cm-1 Q branches which fall at frequencies which are essentially free of any major atmospheric fine struc-ture bands.

IntroductionIt has been shown by a number of workers" 2 that

daytime photooxidation processes quickly and effi-ciently reduce atmospheric terpene concentrations tolow ppmv or ppbv levels even in the immediate vicinityof relatively dense foliage. At night, however, whenphotochemical stationary state concentrations aregreatly reduced, terpene emission continues, albeit alsoon a diminished scale, and there is some evidence thatthe hydrocarbon levels rise appreciably at least on alocal basis.

At levels of 10,000 ft (3048 m) and above, the odor offir forests can often be detected throughout the hoursof darkness on 14,410-ft (4392-m) Mt. Rainier, Wash-ington, even though trees are up to 2 miles (3 km) belowthe observer at many locations. The effect is intensifiedif convective disturbances arising from pools of coldglacial cirque air are active. If pinene odors can reach15,000 ft (4572 m) at night, it may be possible that thehydrocarbons could reach the tropopause in transientconcentrations of 101o-1012 molecules cm-3 or more atnight.

As far as we know, no one has attempted to measurepinene concentrations at any level at night. Althoughthe most sensitive means for carrying out such deter-minations would likely involve FID-GLC, the techniqueis fraught with possible difficulties, primarily thoseconnected with heterogeneous reactions occurring in the

The author is with Western Washington University, ChemistryDepartment, Bellingham, Washington 98225.

Received 16 September 1978.0000:3-69:35/79/203434-04$00.50/0.,c. 1979 Optical Society of America.

trapped samples of the reactive hydrocarbons and otheratmospheric species. The use of lunar IR interferom-etry or perhaps long path length tuned laser spectro-photometry might offer a more rapid in situ and non-destructive means of monitoring nighttime levels of a-and 3-pinene.

We have studied both the IR and Raman spectra ofa- and -pinenes,3 and we have been able to identify gasphase IR bands in both terpenes that could conceivablybe used to identify specifically and measure the com-pounds in the 8-14-,um atmospheric window region.

Experimental Techniques and ResultsA portion of the gas phase IR spectrum of a-pinene,

2,6,6-trimethyl bicyclo[3.1.1]hept-2-ene, is shown in Fig.1 along with a sketch of the structure of the moleculeand the approximate location of the three principalrotational axes of the compound. All spectra were re-corded on a Perkin-Elmer model 521 IR spectropho-tometer at 1200C in a hot cell assembly that has beendescribed previously.45 The main sample cell had an8-cm path length, and at various temperatures rangingup to about 1200C, pressures within the cell could bevaried from about 10 to a few hundred Torr with anuncertainty of 1 Torr or less. Vapor pressures for thepinenes as a function of temperature are available in theliterature. 6

The moments of inertia were computed using variousstructural parameters available in the literature forcycloheptanes and cycloheptenes. They must be con-sidered to be approximate. Efforts to obtain a micro-wave spectrum of the molecule were not successful,possibly because of the small dipole moments that resultfrom the nearly saturated structure of the compoundand the lack of significant asymmetry around the doublebond.

3434 APPLIED OPTICS / Vol. 18, No. 20 / 15 October 1979

7 80 7 90 C -

IIJ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ C -

Fig.l1. Infrared gas phase spectrum of a-pinene equivalent ata 1-km path length and aconcentration of aboutl10ppmv. Nearby atmosphericbands are superimposed on the band.

As seen in Fig. 1, the largest moment of inertia occursaround the axis which is roughly perpendicular to anearly planar portion of the molecule containing thedouble bond. Resultant dipolar changes that occuralong this axis as the result of vibrational motions willgive rise to unresolved rotation-vibration band enve-lopes that are characterized by strong collected Qbranches and relatively nonprominent or slope shoul-dered P and R branches. The C, symmetry of a-pineneallows intermodal coupling so the types A, B, and Cband envelopes will hybridize. However, in spite of thefact that the asymmetry factor K = -0.96 in our cal-culations, the spectrum does display distinct type ABhybrids and type C band envelopes. We have com-puted the expected band envelopes for a harmonic, rigidrotor and find that a type C band should have about 20%of its intensity in the Q branch. Type AB hybrids, re-sulting from vibrational dipole changes along the A andB axes, will display much weaker Q branches, andtherefore the P and R branches will appear to be muchmore prominent.

Rotational spacings in all types of bands will be lessthan 0.06 cm-' and therefore below the resolution limitsof most commercial interferometers. In addition, thereare several vibrational modes in the pinenes that occurat low frequencies, and hot bands must be expected,particularly in spectra such as ours where higher tem-peratures have been utilized. Together with anhar-monicities, nonrigid rotational characteristics, andpressure-broadening effects, the hot bands will have theeffect of diminishing actual spacings to values wellbelow those we have computed for the idealized mole-cule, if indeed they exist at all.

In examining the ca-pinene bands at a resolution ofabout 0.3-0.4 cm-', examples of both types AB and Cband envelopes can be found in the 8-14-Am region ofthe spectrum. The strongest of these is a 786.5 + 1-cm-' type C band which arises from a deformation ofthe molecule perpendicular to the double bond. Itsintensity must come from sp2 rather than sp3 MO dis-tortions. Other bands occur in this region at the fol-lowing frequencies with the band type, if it can be dis-

cerned, shown in brackets after the values: 1206(A),1126(A), 1090(A),, 1017(A), 953(A), 928(A), 886(C), and882(A). More details can be found in Ref. 3.

The strength and obvious type C character of the786.5-cm-1 band make it a prime choice for possible useas an analytical band. We have, therefore, measuredthe integrated intensity .of the band using a Wilson-Wells method7 on spectra obtained at various temper-atures in the hot cell assembly. Band areas were de-termined by Simpson's rule calculations with 0.5-cm-1intervals, and extrapolation to 0 concentration (pres-sure) yielded an apparent absolute intensity for theband of 43.5 atm1 cm 2.

Using these results, Fig. 1 was recorded showing theapproximate appearance of the 786.5-cm-1 band at a1-km path length and a concentration of about 10 ppmv.Under these conditions, a transmittance of about 80%might be expected in ideal circumstances, but unfor-tunately pinene levels can rarely be expected to reachthis range.

The atmospheric bands shown superimposed on thepinene band in Fig. 1 were taken from the literature 8 9

and represent absorption bands at 1-km or more pathlengths. They are scaled to arbitrary intensity unitsand not to the transmittance scale shown for the pineneband.

It is obvious from the diagram that, at 1-km pathlengths, the broad and rather weak structure of the pi-nene band will make it difficult to distinguish it fromthe atmospheric background even at the high levels ofpinene that were used in the calculations. The Qbranch, however, does occur at a location where it couldconceivably be discerned among weaker atmosphericbands, and it offers the most realistic possibility forpractical use.

We have not measured the intensities of the otherwindow bands of cy-pinene as they are considerablyweaker than the 786.5-cm-1 band. All the bands areas broad as the band in question, that is, the approxi-mately 4-cm- Q branch and about 22-cm-l half-bandwidths are typical of all the features.

It might be mentioned in passing that we have found

15 October 1979 / Vol. 18, No. 20 / APPLIED OPTICS 3435

the 786.5-cm-1 band to be a good spectrophotometricvehicle for the study of gas phase reactions involving thedouble bond of pinene. Known addition reactionsacross the double bond have yielded smooth repro-ducible data when monitoring was carried out with thespectrophotometer set on the band,10 and should theband become a feasible choice for the actual monitoringof atmospheric pinenes, their rate of degradation in theair could easily be studied.

Turning to the -compound, 6,6-dimethyl-2-meth-ylene bicyclo[3.1.1]heptane, there is an 880-cm-1 bandsystem which corresponds to the 786.5-cm-' band in/-pinene. The -compound is less symmetric thana-pinene with the now exo double bond and the con-version of the carbon in the three-position to tetrahedralsp3 hybrid angles, lowering the apparent asymmetryfactor to about -0.68, but the directions of the principalaxes are not drastically changed as is shown in thestructural sketch in Fig. 2. The lowered symmetry ofthe molecule will result in greater intermodal pertur-bations, but the computed band shapes of the types A,B, and C bands are not significantly altered for theharmonic rigid rotor model.

The 880-cm-1 band also results from a double bondring deformation mode, and it has characteristics similarto those described previously for the a-compound. Wehave measured the integrated intensity of the band andfind that it has a value of about 129 atm-1 cm-2.11Figure 2 shows the Q-branch region of the band for apath length of about 1 km and a concentration of 5ppmv. Again superimposed over the spectrum is a se-ries of atmospheric bands found in the region at path

WUz

i(0

*

875 880

lengths of 1 km or more. Although the pinene band isscaled to transmittance, the atmospheric bands arearbitrary and are not, therefore, correlated with thepinene intensity.

Although this band is somewhat more intense thanits ca-pinene analog, it still does not offer a sensitivemeans of detecting the terpene. Its broad characteralso causes it to blend too well into the atmosphericbackground and essentially disappear as a recognizablefeature at -pinene concentrations much below the 5-10-ppmv level.

There are other weaker and even less analyticallyuseful bands in the window region. Among them arefeatures at the following frequencies3 again with theband type, if available, given in brackets after the val-ues: 1204(AB), 1143(AB), 1105(AB), 1056(AB), 1048,and 765(C).

ConclusionsTheoretically, type C bands at 786.5 cm-' and 880

cm-' in the gas phase IR spectra of a- and 0-pinenes,respectively, are intense enough to allow the pinenes tobe determined with a sensitivity of about 10-100 ppbvat 1-km path lengths. In practice, the broad bands withthe approximately 4-cm-1 wide Q branches will bedifficult to distinguish from the background atmo-spheric absorption in their respective frequency regions,and a determination of the compounds in the 8-14-gmwindow will be limited in practice to 1-10 ppmv or more.It is unlikely that increased resolution in commercialinterferometers will improve the situation to any greatextent.

885 890

Fig. 2. The gas phase IR spectrum of fl-pinene corresponding to a 1-km path length and 5-ppmv concentrations with atmospheric bandssuperimposed on the band.

3436 APPLIED OPTICS / Vol. 18, No. 20 / 15 October 1979

!'\ I� I' ' II I--I

a.I' ii III "II I' I

'I' I� IfI II i 'III II JIIl III, II I'I i IjI I IIif II'I'I'

fI

I

I

I

I I�l11

If local concentrations of the terpenes rise to theppmv range, the selection of the band to be used tomonitor their levels and perhaps determine the rate oftheir degradation will depend on the type of vegetationinvolved. It has been shown that various types ofplants, particularly trees, emit either a- or -pinenesor sometimes both.12,'3 Both could, of course, bemonitored simultaneously if the concentrations reachthe proper levels. In any event, the method would bemuch less sensitive than FID-GC techniques, providedthe inherent problems involved in taking grab samplesof chemically reactive compounds are not over-whelming.

In examining published atmospheric long path lengthinterferometric spectra,8 we have failed to locate anyevidence of either compound even at large air massestaken at high solar zenith angles. However, the physicalevidence, i.e., the odor, we have noticed occurs primarilyat night, and as yet we have no access to a spectral studydone during darkness.

References1. R. R. Arnts, R. L. Seila, R. L. Kuntz, F. L. Mowry, K. R. Knoerr,

and A. C. Dudgeon, in Conference Proceedings 4th Joint Con-ference on Sensing of Environmental Pollutants, New Orleans,1977 (American Chemical Society, Washington, D.C., 1978), pp.829-833.

2. E. P. Grimsrud, H. H. Westberg, and R. A. Rasmussen, in Pro-ceedings of the Symposium on Chemical Kinetics Data for theUpper and Lower Atmosphere, Warrenton, Va., 1974 (Wiley,New York, 1975), pp. 183-196.

3. H. W. Wilson, Appl. Spectrosc. 30(2), 209 (1976).4. H. W. Wilson, Appl. Spectrosc. 24, 6 (1970).5. H. W. Wilson, Spectrochim. Acta Part A: 30, 2141 (1974).6. J. E. Hawkins and T. T. Armstrong, J. Am. Chem. Soc. 76, 3756

(1954).7. D. A. Ramsay, J. Am. Chem. Soc. 74, 72 (1952).8. R. J. Nordstrom, J. H. Shaw, W. R. Skinner, J. G. Calvert, W. H.

Chan, and W. M. Uselman, "A Comparison of High ResolutionObserved and Computed Air Spectra Between 700 and 2300cm-'," Ohio State University Research Foundation TechnicalReport RF 4221 (1976).

9. T. G. Kyle, "Atlas of Computed Infrared Atmospheric AbsorptionSpectra," National Center for Atmospheric Research TechnicalNote NCAR-TN/STR-112 (1974).

10. H. W. Wilson, unpublished results.11. H. W. Wilson, Conference Proceedings 4th Joint Conference on

Sensing of Environmental Pollutants, New Orleans, 1977(American Chemical Society, Washington, D.C., 1978), pp.834-835.

12. R. A. Rasmussen, J. Air Pollut. Control Assoc. 22, 537 (1972).13. R. A. Rasmussen, Environ. Sci. Technol. 4, 667 (1970).

H. William Wilson of Western Washington University photographed by W. Mankin of NCAR during the 1978 OSA Topical Meeting on At-mospheric Spectroscopy.

15 October 1979 / Vol. 18, No. 20 / APPLIED OPTICS 3437