Effect of the capture coefficient in deep-level transient spectroscopy measurements

Abstract

Underlying the conventional deep-level transient spectroscopy (DLTS) method is the assumption of an exponential capacitance transient to the equilibrium state as a result of the thermal emission rate of free carriers from a filled trap energy level. This exponential capacitance transient may prove to be a good approximation for specific cases but, in general, the transient capacitance decay to the equilibrium state following a capture pulse is nonexponential. In this study nonexponential capacitance transients are shown to be encouraged by the presence of the free-carrier tail which spills over abundant free mobile carriers into the space-charge region thus negating the abrupt junction depletion approximation and favoring both capture and emission of carriers. An upper limit for this effect is obtained here by assuming the carrier concentration in the relevant part of the space-charge region which one has in the neutral region. This reduces the thermal emission rate by several orders of magnitude from what one would find with the assumption of pure exponential transient and neglecting spillover, as in the normal DLTS method. A particular case is considered, where both capture and thermal emission processes occur simultaneously in the Shockley–Read–Hall kinetic equation for a single-trap energy level. The variation of carrier occupancy with respect to time leads to a nonexponential capacitance transient decay to the equilibrium state.