Answers are grounded in this paper and cite their sources.
This work reports both experimental measurements and calculations of two-phonon Raman spin-lattice relaxation times for shallow donors in silicon. Saturation, line broadening, and adiabatic fast passage techniques were used to measure Tₛ (ΔMₛ=±1,ΔMI=0) between 2 and 30^∘{}K for P, As, Sb, and Bi at 3300 G. Tₓ (ΔMₛ=±1,ΔMI=∓1) was measured in the liquid He range for As, Sb, and Bi at 3300 G. The calculations of the spin-lattice relaxation rates 1Tₛ and 1Tₓ for Raman processes were made using only the $1S$ doublet and triplet valley-orbit excited states. The calculations were made for both the power-law temperature-dependent region and the exponential temperature-dependent region.1Tₛ for P and As showed a power-law temperature dependence Tⁿ with n closer to 9 than to 7 between 2.5 and 6^∘{}K for P and between 3 and 11^∘{}K for As. In this range there is no magnetic field-crystal axis orientation dependence and also no magnetic field dependence between 3300 and 8000 G as shown by Wilson and Feher. These facts suggest a spin-orbit spin-flip interaction which would have a "Van Vleck cancellation." No T¹³ Raman temperature-dependent 1Tₛ was observed experimentally. P, As, Sb, and Bi all show an exponential temperature-dependent 1Tₛ above 6, 11, 4, and 26^∘{}K, respectively. The rate constant E (1Tₛ-Ee^-ΔkT) gets larger as one goes to the higher Z donor impurities.The ratio TₛTₓ was measured between 2 and 5^∘{}K for As⁷⁵, the result showing TₛTₓ∝T^-2 above 2.2^∘{}K consistent with Tₛ∝-8.5 and Tₛ∝T^-6.5 near 5^∘{}K. Experimental values of Tₓ were obtained for Sb¹²¹ and Bi²⁰⁹ at 2.2 and 4.2^∘{}K, respectively.Use of the Hasegawa-Roth mechanism for the calculated Raman 1Tₛ for both the power-law and the exponential temperature-dependent regions gives a result nearly four orders of magnitude smaller than experiment. In the short-wavelength phonon region the $1S$ triplet can be equally as important as the $1S$ doublet via intervalley Umklapp processes. For As it is shown that intervalley Umklapp processes will predominate over intravalley processes. In the exponential case the calculated 1Tₛ rate can have a maximum value when the excited states have level widths equal 2gμBH.The calculated Raman 1Tₓ rates for the different donors show good agreement with the experimental values. The 1Tₓ rate can have an exponential temperature-dependent term but it will be much smaller than the T⁷ term for the temperature range considered here.Calculations have been made of the level widths of the P and As $1S$ excited states due to spontaneous phonon emission. It is shown for As that the dominant intervalley Umklapp spontaneous emission will be sharply peaked along the [100] axis.
No takes yet. Share an insight, caveat, or question.
T. G. Castner (1963) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: