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

Abstract

A new method to analyze multiexponential transients for deeplevel transient spectroscopy Novel method to determine capture crosssection activation energies by deeplevel transient spectroscopy techniques Appl. Phys. Lett. 52, 660 (1988); 10.1063/1.99366The effects of the nonabrupt depletion edge on deeptrap profiles determined by deeplevel transient spectroscopy 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 fined 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 nonexponentiat 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 spacecharge 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.

Charge-based deep level transient spectroscopy of phosphorous-doped homoepitaxial diamond

“…͑6͒ does not hold anymore, and a more complex expression for QЈ(t) results due to nonexponential time rates of change of the trap occupancy. 37 In the present situation, retrapping is indeed likely to have occurred due to one or more of at least three mechanisms. Two of them are well known in the conventional DLTS technique.…”

“…These are ͑1͒ thermally activated capture cross section, 28 ͑2͒ field-enhanced thermal emission rates, 11,35 and ͑3͒ retrapping effects. 27,36,37 As seen from Eq. ͑3͒, process ͑1͒ would lead to a measured activation energy equal to the sum of the potential barrier to capture and the equilibrium depth of the bound state.…”

“…27 Process ͑i͒ arises from the presence of a free carrier tail at the edge of the depletion region. Landberg et al 37 showed that this effect could reduce the thermal emission rate by up to several orders of magnitude in some cases. Mechanism ͑ii͒ is closely related to the presence of leakage currents at the Schottky contacts, which leads to the injection of majority carriers into the depletion region at the same time as thermal emission takes place.…”

“…Process ͑3͒ has the potential to strongly affect DLTS spectra. 27,36,37 This can be understood in the following way. The rate equation for the electron occupancy N(t) of a trap of energy E a and concentration N t is given, for an n-type material, by 27 FIG.…”

“…Two of them are well known in the conventional DLTS technique. These are ͑i͒ nonzero recombination rates at the edge of the depletion region [35][36][37] and ͑ii͒ leakage-related electron capture. 27 Process ͑i͒ arises from the presence of a free carrier tail at the edge of the depletion region.…”