Stellar and 0ther High-Temperature Molecules

See allHide authors and affiliations

Science  13 Jan 1967:
Vol. 155, Issue 3759, pp. 155-164
DOI: 10.1126/science.155.3759.155


Optical spectroscopy of stellar molecules trapped at 4°K is particularly valuable when the data can be used to complement the corresponding gas data. The ground state is then directly established by measurement of the absorption spectrum at the low temperature, since all transitions beginning from excited states are eliminated. Isotopic substitution establishes the (0,0) bands of transitions to excited electronic states, and when these data are combined with infrared and fluorescence measurements at 4°K, most of the vibrational properties of the ground and excited states can be obtained. Of the many examples cited and discussed here, C3 is perhaps the most outstanding.

Because the various molecules trapped in matrices are usually identified through prior mass spectrometric work, the optical observations often lead to the discovery of band systems of molecules whose spectra have not previously been observed—for example, those of Si2C3, TaO2, and WO2. In these cases the location of the spectral regions in which molecular transitions appear may also serve to encourage the gas spectroscopist to further exploration with high-dispersion spectrographs.

I share Ramsay's view (4, p. 204) that further investigation of f-number determinations from matrix spectra should be encouraged, particularly because of the lack of such data for molecules in stars. The principal source of error probably lies in the estimation of the number of molecules per square centimeter in the matrix, but no real test of this has been made. The only existing f values from matrix spectra are those for the C3 (43, 44) and C2 (51) molecules, and these are not ideal for test purposes. Because of the anomalous nature of the matrix results discussed above, the rather good agreement between f values for the solid and gas phases of C2 (51) cannot be considered as support for the matrix determinations. What is needed is a matrix determination of several fv'v" values (that is, determinations for specific bands) for molecules such as CN and NO or, preferably, for a gas vaporized at high temperature, for which these f values are relatively well established in the gas phase.

In this connection the possibility of determining other molecular properties in matrices comes to mind. However, it has been clearly shown that the shape of the potential energy function in the electronic states of molecules can be affected when molecules are trapped in matrices. Brewer, Brabson, and Meyer, in work on S2 (55), and Schnepp and Dressler, in work on O2 (56), have examined the anharmonicity in matrices over many vibrational levels. Distortion of the gas potential energy curves occurs in the heavier matrices and sometimes at high vibrational levels in the light ones. Recently work has been begun on comparing the Franck-Condon factors connecting the ground state and two excited states of ScF trapped in a neon matrix (57) with factors calculated from the gas-spectrum data of R. F. Barrow et al. (58) (Δre, the change in interatomic distance upon excitation, has a relatively large value of ~ 0.1 Å in these systems). As is well known, such factors are particularly sensitive to the value of Δre, but differences in anharmonicity do not, however, have as significant an effect upon the Franck-Condon factors. Hence a comparison of the matrix and gas factors will lead to further information about matrix effects and will indicate whether Franck-Condon factors can be obtained from matrix spectra.

One of the important problems in the study of stellar molecules is the determination of the low-lying electronically excited states, similar to the 1Δ ↔ X3Δ difference (~ 580 cm-1) in TiO measured by Phillips. Most of the transition-metal oxides have such low-lying levels, and they must be taken into consideration in any calculation of thermodynamic effects at high temperature. It appears that the study of emission spectra in the infrared at 4°K may be one approach to this problem, and an attempt is now being made to confirm the TiO value in order to test the method.

Perhaps the greatest advantage of matrix-isolation is the fact that it allows study of these molecules—or of any molecules difficult to produce in a microwave cavity—by electron-paramagnetic-resonance spectroscopy. Study of molecules by this means can provide information about ground state wave functions that is not obtainable by optical spectroscopy, as illustrated by the investigation of ScO, YO, and LaO. Also, it seems likely that the preferential orientation of molecules in matrices, which is probably achievable in most cases, will be a valuable asset in the study of the magnetic properties of molecules by electron-paramagneticresonance spectroscopy, regardless of whether the molecules are "hot" or "cold" (60).