The band-gap energy Eg of silicon has been reevaluated with high precision between 2 and 300 K by the following method: the derivative of the absorption coefficient, resulting from free-exciton absorption, has a well-defined singularity, which can be detected unambiguously by wavelength-modulation spectroscopy. The energy of this singularity yields the band gap. Our data deviated by more than 5 meV from the earlier results of MacFarlane et al. and Haynes et al. and fell between their Eg(T) curves. The approximation of Eg(T) = A + BT + CT2 gives A = 1.170 eV, B = 1.059×10−5 eV/K, and C = −6.05×10−7 eV/K2, for 0<T≤190 K, and A = 1.1785 eV, B = −9.025×10−5 eV/K, and C = −3.05×10−7 eV/K2, for 150≤T≤300 K, which fits our data within 0.2 meV. The major uncertainty of about 1 meV lies in the energies of exciton and TO phonon.

We report, for the first time in a biaxial laser material, on the distribution of fluorescence cross sections in the index ellipsoid. The largest effective fluorescence cross section of pure NdP5O14 at room temperature (2.0±0.1) ×10−19 cm2 for light at 1.0512 μ wavelength polarized parallel to the 2 axis which coincides with the crystallographic b axis. We predict from the cross sections the cw thresholds and polarizations of the four highest-gain laser transitions and verify the results quantitatively by operating one crystal in all major polarizations. We measured the major phase and group refractive indices and the thermal expansion coefficient of the optical lengths near 1.05 μ wavelength. We also discuss the twin domains which have different orientations of the crystallographic and optic axes. In imperfect crystals they can easily be created by thermal stresses which can drastically increase the laser threshold.

No abstract