New Techniques in Triplet State Phosphorescence Spectroscopy: Application to the Emission of 2,3-Dichloroquinoxaline

Abstract

Until recently, the observed characteristics of the phosphorescence emission have been those for a population‐weighted average of the emission characteristics of the three zero‐field (zf) levels of the lowest triplet state. However, at sufficiently low temperatures (∼ 1.4°K) where the spin–lattice relaxation processes between the zf levels are slower than the emission processes, the characteristics of the emission become the sum of that for the three individual zf levels. Furthermore, the steady‐state populations of the three zf levels become generally unequal. Since the zf splittings are smaller than the inhomogeneous bandwidth of the phosphorescence, the emission characteristics of each zf level cannot be determined directly from the spectrum. The application of different perturbations—e.g., temperature changes, application of a magnetic field, and saturation of the zf transitions with microwave radiation of resonance frequencies—changes the relative population of the different zf levels and thus changes the relative intensity, the polarization, and the decay characteristics of the different vibronic bands in the spectrum. From the observed changes, a unique assignment of the originating zf level for each vibronic band in the spectrum can be made. A novel new technique is introduced which is designed to obtain an amplitude modulated phosphorescence‐microwave double‐resonance (am‐PMDR) spectrum. This is carried out by amplitude modulation of the microwave radiation, which is of sufficient peak power to saturate one of the three zf transitions, while detecting the optical spectrum at the modulation frequency. The corresponding changes of the phosphorescence intensity of the bands originating from either of the two zf levels being saturated are thus detected with opposite signs. Bands originating from the third zf level would not change in intensity and will not appear in the am‐PMDR spectrum. The assignment of the zf origin of the different vibronic bands is essential for the determination of the spin–orbit–vibronic schemes responsible for the phosphorescence process in these molecules. The above perturbational techniques have been applied to observe the changes in the spectra, in the polarization, and in the decay of the main vibronic bands of the phosphorescence of 2,3‐dichloroquinoxaline in different hosts. Unique conclusions concerning the phosphorescence mechanisms of this molecule are obtained and are found to be in good general agreement with theoretical predictions involving first‐order spin–orbit and second‐order spin–orbit–vibronic perturbation mechanisms. The important conclusions are: (1) Only two of the three zf levels (${\tau }_z$ and ${\tau }_y$ ) are found to be strongly radiative. (2) Three different types (I, II, and III) of bands are found in the spectra: a. The emission of type I bands originates solely from the top zf level ${\mathrm{(\tau }}_z)$ and requires only spin–orbit (SO), but not vibronic, perturbation for their appearance. b. In addition to SO, type II bands require vibronic perturbation involving mainly the 262‐cm⁻¹, a₂‐type vibration. Type II bands originate solely from the middle, ${\tau }_y$ , zf level. Type III bands also require vibronic perturbation in addition to SO, but originate from both the ${\tau }_z$ and ${\tau }_y$ zf levels and thus have mixed polarization whose characteristics change with temperature and microwave saturation. The most intense type III bands require vibronic perturbation involving $b_1$ type vibrations with a frequency of 490 cm⁻¹. (3) The agreement between theory and experiment is further improved if the only theoretical schemes considered involve spin‐orbit interaction between singlet and triplet states with opposite symmetry properties to reflection in the molecular plane (e.g., $\mathrm{n,\; \pi *}$ and $\mathrm{\pi ,\; \pi *}$ or $\mathrm{\sigma ,\; \pi *}$ and $\mathrm{\pi ,\; \pi *}$ ). (4) The intramolecular heavy‐atom effects on the radiative and nonradiative properties of the individual zf levels of quinoxaline are quantitatively determined. Whereas the over‐all lifetime at high temperature decreases only by a factor of 0.7, the radiative lifetime of the ${\tau }_y$ zf level is found to decrease by a factor of 0.04 upon chlorine substitution. This is the level responsible for Subspectrum II that appears upon halogen substitution. (5) The important intersystem crossing route in 2,3‐dichloroquinoxaline is found to be different from that in quinoxaline, in agreement with previous predictions concerning the intersystem crossing process in N‐heterocyclics [M. A. El‐Sayed, J. Chem. Phys. 38, 2834 (1963)].