ELECTRON PARAMAGNETIC RESONANCE IN CADMIUM(1-X)MANGANESE(X)TELLURIDE, CADMIUM(1 X)MANGANESE(X)SELENIDE AND CADMIUM(1-X)MANGANESE(X)SULFIDE (DILUTED MAGNETIC, SEMICONDUCTORS, EXCHANGE NARROWING, SUPEREXCHANGE, EPR)

NITIN SAMARTH, Purdue University

Abstract

We present the results of a systematic investigation of EPR in the Cd based DMS, namely, Cd(,1-x)Mn(,x)Te, Cd(,1-x)Mn(,x)Se and Cd(,1-x)Mn(,x)S, over a wide range of temperatures (1.3K - 250K) and sample compo- sition. The EPR measurements were made on a 35GHz transmission spectrometer, using circular polarization and Faraday effect tech- niques. The Faraday technique is perfected by showing that the fits to the experimental lineshapes are considerably improved if we employ a mixture of Faraday rotation and ellipticity. The EPR behaviour in all the three compounds is qualitatively identical: the absorption lineshape is Lorentzian, and broadens with increasing Mn concentration and decreasing temperature. At very low temperatures, of samples with a high Mn concentration, the line- shape deviates from Lorentzian behaviour and the resonance shifts to lower fields. We also establish the dependence of EPR linewidth on the anion of the host lattice. Over a wide range of sample compo- sition and temperatures, for a fixed temperature and Mn concentra- tion, given a particular cation, the linewidth in the order: sulfide, selenide and telluride. The Lorentzian regime linewidth is analyzed using exchange narrowing theory. While it appears that the EPR in the sulfide may be dipolar broadened, the linewidth in the selenide and telluride cannot be explained by dipolar anisotropy alone. The dominant anisotropy in the latter cases must come from the anisotropic part of the exchange interaction, and we speculate the the spin-orbit coupling of the anion has to play a role. The Dzyaloshinski-Moriya interaction offers a promising solution to this problem. At high tem- peratures, the x dependence of the linewidth is much stronger than predicted by exchange narrowing--the linewidth almost goes as x('2), while theory yields approximately a SQRT.(x) behaviour. This discrep- ancy is attributed to temperature dependent factors which are neglected in the infinite temperature theory. In all the three systems studied, for x(, )> 0.20, the temperature variation of the linewidth is a universal function of reduced temperature T/T(,0). The variation of the scaling temperature T(,0) with x resembles that of the spin glass temperature T(,g). The non-Lorentzian lineshapes at low temperatures are exam- ined using the Kubo-Toyabe theory. We find that the theory cannot explain the lineshapes in DMS, and the problem remains unresolved.

Degree

Ph.D.

Subject Area

Condensation

COinS