Scanning spreading resistance microscopy current transport studies on doped III–V semiconductors

Abstract

Two-dimensional (2D) carrier concentration profiling using scanning spreading resistance microscopy (SSRM) has been carried out on molecular beam epitaxy-grown GaAs and InP dopant calibration samples. The current transport mechanisms between the diamond-coated SSRM tip and the III–V semiconductor cleaved surface (110) was investigated as a function of semiconductor dopant concentration via current–voltage (I–V) measurement. A positive or negative tip bias was applied while scanning over each dopant concentration region (1016–1019 cm−3). The results were compared to simulated I–V curves based on thermionic emission theory. The best fits to the data obtained under forward bias indicated that the contact barrier heights, φb, were much lower than expected from conventional large area planar contacts to GaAs or InP. The effect increases with increasing doping concentration, as a result of a combination of barrier height lowering mechanisms such as image forces, thermionic field emission and minority carrier injection. Under reverse bias, the thermionic emission theory only holds for highly doped semiconductors while for low doping, surface and bulk generation currents determine the overall detectable SSRM current.

Investigation of CdZnTe crystal defects using scanning probe microscopy

“…From the logarithmic plots of the I-V curves, it can be seen that the I-V relationship is very nearly exponential for higher biases. However, the ideality factors for these curves were found to be extremely high ͑ϳ18 for CZT1 and ϳ16 for CZT2, calculated for the section of the I-V curve between V = 1.2 and 1.5 V͒, which is nevertheless in agreement with previous observations of high nonideality factors in SSRM measurements on InP and GaAs samples 8 and can be expected for point contact systems. 9 The stationary probe current measured on CZT2 sample for a particular probe bias was found to be a few times higher than the scanning probe current for the same bias, most likely due to better contact formation.…”

“…This behavior is consistent with the I-V characteristics on p-type doped samples. 8 Indeed Hall measurements performed on CZT2 indicate that the material is p type and the background majority carrier concentration of holes was measured to be ϳ2 ϫ 10 14 cm −3 . The CZT1 sample was found to have a very high resistivity of ϳ3 ϫ 10 11 ⍀ cm and the carrier type could not be determined from Hall measurements.…”

“…The exponential behavior of the I-V characteristics can be explained by considering the thermionic emission mechanism for the current conduction, rather than simple resistive current flow, following the model proposed by Lu et al 8 In their model, the probe current is limited by the interface potential barrier ͑giving rise to an exponential behavior͒, rather than the local doping level in the semiconductor ͑as in pure SSRM, where a linear I-V behavior is expected͒. The thermionic emission current in the forward bias ͑please note that for p-type doped sample, forward bias is established for negative tip voltage͒ is given as 8…”

Investigation of CdZnTe crystal defects using scanning spreading resistance microscopy

“…This proposition might be valid based on the previous observations to III-V semiconductors. 11 The average magnitude of SSRM current for p-type region in CZT2 under forward bias is observed to be about 0.5 nA, at least one order higher compared to the regions with few Te precipitates. Consequently, it is most likely to be two factors dominating the I-V characteristics, the density of Te precipitates and the spatial variation of conduction type on the surface.…”

“…In this experiment, the probe was located at the single point (with zero scan size), and the current was measured while the probe bias was usually swept from -1.6 to +1.6 V. The voltage range was limited by the current instability and noise. The I-V relationship might be explained considering thermionic emission theory, following the model proposed by Lu et al 11 By applying the theory, the average SSRM current was modeled as a function of probe bias. The thermionic emission current under forward bias (please note that for p-type doped sample, the forward bias is established under negative probe bias, while for ntype doped sample, it follows the opposite way) is given by 11 Where A ** is the effective Richardson constant, S is the contact area 2 2 r π for a hemispherical contact with r=20 nm, as specified in the datasheet from Veeco datasheet.…”