부산대학교 도서관

While several 2D thermometry techniques exist, there is a lack of 3D thermometry techniques that work for wide range of materials and offer good resolution in time, space and temperature. X-ray diffraction (XRD) and nuclear magnetic resonance (NMR) imaging can provide 3D temperature information. However, XRD is typically limited to crystalline materials while NMR is largely limited to liquids where the resonance lines are narrow. We investigate electron paramagnetic resonance (EPR) of n-type silicon and germanium for 3D thermometry. While in germanium the EPR linewidths are too broad, EPR linewidths in silicon are reasonably narrow and exhibit a strong temperature dependence. The temperature dependence of the spin-lattice relaxation rate (1/T1) of conduction electrons in n-type Si for low dopant concentrations follows a T^3 law due to phonon broadening. For heavily doped Si, which is desirable for good signal to noise ratio (SNR) for application in thermometry, impurity scattering is expected to decrease the temperature dependence of 1/T1. Our results show, in heavily doped n-type Si, spin-lattice relaxation induced by impurity scattering does not drastically decrease the temperature dependence of EPR linewidths. In P-doped Si with donor concentration of 7 x 10^18 /cm^3, the EPR linewidth has a T^(5/2) temperature dependence; the temperature dependence decreases to T^(3/2) when the donor concentration is 7 x 10^19 /cm^3. While the temperature dependence of linewidth decreases for heavier doping, EPR linewidth is still a sensitive thermometer. We define a figure of merit for SNR for thermometry from EPR linewidths of n-type Si and observe that increasing the doping results in a better SNR. Using effective medium theory, we show that EPR linewidth can be a sensitive thermometer for application in 3D thermometry with systems embedding microparticles of heavily doped n-type Si.