We present an approximate model for laser beam-induced current (LBIC) imaging of arrays of semiconductor devices based on homogenization. LBIC is a non-destructive technique useful for the characterization and quality control of semiconductor focal plane arrays, a key component in modern imaging systems. The model provides not only an efficient alternative for LBIC image simulation of large uniform arrays, but also an effective method for detection of non-uniformities among arrays. Numerical examples are presented to illustrate the effectiveness of this method in detecting array non-uniformities due to variations in dislocation/size or doping level among p–n junction diodes.

This paper presents an interpretation of the physical mechanisms involved in the generation of laser beam-induced current (LBIC) in semiconductor p-n junction diodes. LBIC is a nondestructive semiconductor characterization technique that has been used in a qualitative manner for a number of years and is especially useful for examining individual photodiodes within large two-dimensional arrays of devices. The main thrust of this work is the analysis of LBIC in terms of nonzero steady-state circulatory current flow within the device and, hence, the interpretation of LBIC line profiles to diagnose the patterns of current flow within the structure. This provides an important basis for future studies seeking to relate LBIC to indicators of p-n junction performance and integrity such as dark current components and reverse bias saturation current. In particular, this paper examines the ideal cases of a single isolated p-n junction diode structure, and also considers an array of such devices in close proximity to each other. Modifications to the idealized theory that are required to account for localized junction leakage and surface recombination are presented, and the effect of Schottky contacts is discussed. Numerical simulations based on the HgCdTe family of semiconductors are presented to support the theory.

One of the main limitations in the operability of modern infrared focal plane arrays of p-n junction diodes formed on molecular-beam epitaxy (MBE)-grown HgCdTe is the effect of localized defects. Such defects, including voids, triangles and microvoids, are a feature of the MBE growth regime and can compromise the performance of devices fabricated within the vicinity of electrically active defects. While such defects can often be identified visually, not all defects are electrically active such that they provide a current leakage path shunting the p-n junction of the individual photodiode. In this paper, the use of laser beam-induced current is proposed as a nondestructive characterization technique, and quantitative aspects of its use in the study of electrically active defects in photodiode arrays are examined.

A reduced model is developed that has significant advantages over the full drift-diffusion model for the simulation of laser beam-induced current (LBIC) signals in the presence of heterojunctions. The model determines the contribution to the LBIC signal that would occur from photogeneration at any position within the semiconductor, and is particularly useful for heterostructures where judicious choice of illumination wavelength can result in photogeneration at different depths within the device structure. The reduced model is used to examine the basic features of LBIC as applied to two types of planar P-n HgCdTe heterojunction photodiode structures. In particular, the question of correctly identifying erroneous device structures formed during the fabrication process is addressed, and experimental measurements are presented to support the simulation results.

Laser-beam-induced current (LBIC) is being investigated as an alternative to electrical measurements of individual photodiodes in a two-dimensional array. This is possible because LBIC only requires two electrical contacts to an array and the two-dimensional scanning of a focused laser beam across the array to image the entire array. The measured LBIC profiles, obtained from linear arrays of HgCdTe photodiodes, will be used to study the uniformity of photodiodes in the array and to extract the R0A of the photodiodes. It will be shown that the shape of the LBIC signal is correlated to the electrical performance of the photodiode, with R0A related to the spreading length of the photodiodes. Linear arrays of n-on-p, mid-wavelength infrared (MWIR) and long wave-length infrared (LWIR) devices were formed in liquid-phase epitaxy HgCdTe epilayers using a plasma junction-formation technique. The LBIC profiles were measured on each of the devices at various temperatures. For the MWIR devices, the extracted spreading length shows no correlation with R0A. However, the LBIC signal does detect nonuniform devices within the array. For the case of the LWIR devices, the spreading length is extracted as a function of temperature, with the R0A subsequently calculated from the spreading length. The calculated R0A, obtained without requiring contact to each photodiode in the array, agrees well with electrical measurements. Asymmetry of the LBIC signals for certain devices in the arrays is shown to be a result of localized leakage at the photodiode junction or from the contact pads through the passivation layers. These results are confirmed by numerical modeling of the device structures.

Laser beam induced current (LBIC)/nondestructive technique that has been used for a number of years to qualitatively examine large arrays of p-n junctions, especially in HgCdTe infrared focal plane arrays. In this paper, we quantitatively study the application of the LBIC imaging technique to semiconductor arrays Based on a previous mathematical model for LBIC applied to individual devices, we employ the homogenization method to derive approximations of the LBIC images of large arrays. Such approximations reduce the computational burden in simulations of these LBIC bouges. We then illustrate the application of our approximations for the purpose of recovering array parameters from the LBIC images.

We point out that the comparison in Fig. 1 of the recent publication by S. Kiatgamolchai et al. [Phys. Rev. E 66, 036705 (2002)] of the proposed maximum entropy-mobility spectrum analysis (ME-MSA) with our quantitative mobility spectrum analysis (QMSA) is misleading. Rather than comparing with the more recent "improved" version of QMSA [Vurgaftman et al., J. Appl. Phys. 84, 4966 (1998)], a preliminary version that was three years older and demonstrably inferior was employed. We show that ME-MSA and the improved QMSA give quite similar results.

To date, most studies of the use of laser beam induced current (LBIC) for non-destructive characterisation of photodiodes have been qualitative and/or empirical, due in part to the difficulty of isolating the influence of the large number of material and device parameters on which the LBIC signal is dependent. The development of methodologies whereby these parameters can be determined quantitatively from the LBIC measurements is important for the technique to gain wider acceptance. This work describes, for the first time, the specific experimental conditions under which some of the variable parameters can be eliminated, substantially reducing the complexity of the analysis. In particular, temperature dependence of the peak-to-peak LBIC measurements on p–n junctions is examined for the first time, revealing that a saturation state is reached at low temperatures. When measurements are performed under these saturation conditions, the peak-to-peak LBIC signal becomes independent of doping density and bulk recombination parameters, allowing other parameters including device geometry to be examined with less ambiguity. This concept represents a crucial step towards a quantitative procedure for extraction of p–n junction material and device parameters using LBIC.

Laser-beam-induced-current (LBIC) imaging is a nondestructive technique used for the characterization of the electrical structure within a semiconductor. In this paper a model is formulated for this technique using the standard drift-diffusion model, and, subsequently, an approximate version and its dual are derived for the study of the inverse problem. The formulation is then applied to a cross-sectional model for n -on- p devices of finite depth to study in detail the relation between the LBIC images and the device parameters. Numerical methods are developed for the simulation of the LBIC image of a diode as well as for the identification of parameters from the LBIC image by least-squares formulation. Numerical examples are presented to illustrate the success of identifying parameters such as junction depth, diffusion length, and equilibrium potential of an abrupt p - n junction diode from its LBIC image. The differentiability of the image with respect to the parameters also is established.

A new technique is described in which the series resistance of laser beam induced current measurements can be determined. Once obtained, this resistance can be equated to an analytic expression for the resistance that involves the depth of the p-n junction in the illuminated photodiode, and a value for the junction depth can be obtained, as has previously been demonstrated. The new measurement scheme involves both laser beam induced current and lateral photovoltage measurements on the same device using the same remote contacts. This avoids the need for a contact to both sides of the p-n junction and hence the technique can readily be applied to individual photodiodes within a large focal plane array.

In this work, the effect of the reactive ion etching (RIE)-induced p-to-n type conversion process on the transport properties of HgCdTe is investigated. Magnetic-field-dependent differential Hall and resistivity measurements have been performed to determine the n-type doping profile and temperature-dependent transport properties of carriers resulting from the RIE-induced type conversion. The study examined Hg 1-x Cd x Te with x=0.23 and x=0.31 for both vacancy-doped and gold-doped p-type epilayers grown by liquid phase epitaxy (LPE) on lattice-matched CdZnTe, which were partly converted to n-type by a RIE process. Analysis using quantitative mobility spectrum analysis (QMSA) reveals that RIE type conversion results in a damaged surface layer characterized by moderate mobility electrons and a bulk n-type region exhibiting higher electron mobility. The surface and bulk electrons show a distinct temperature dependence. It can be observed that, generally, the concentration and mobility of the surface electrons are similar for all samples studied and are independent of temperature in the measured temperature range. In contrast, the bulk electrons exhibit classical behavior with characteristics that are strongly dependent on temperature and consistent with high-quality HgCdTe material. Differential Hall results indicate that the n-type dopants resulting from the RIE process are distributed into the sample via a diffusion process. The results suggest a p-to-n type conversion mechanism involving neutralization of the p-type dopants and diffusion of extrinsic n-type dopants from the surface.

This work reports on the application of laser beam induced current (LBIC) to the determination of the product of zero-bias dynamic resistance and area, R0A, of homojunction photodiodes. The technique involves using LBIC to measure temperature dependent values of the photocarrier spreading length of a particular device, and then fitting the theoretical temperature dependence. The photocarrier spreading length is a measure of the rate of decay of photoinduced forward bias of the p-n junction with distance from the location of electron-hole pair generation and is partly responsible for the shape of the LBIC profile. The LBIC magnitude is not required to be measured in any quantitative sense, although there is a requirement that the photocarrier spreading length be shorter than the device length in order for it to be measurable. The technique provides a non-contact method of determining the performance of individual devices within large two-dimensional focal plane arrays of photodiodes. Experimental results from Hg0.77Cd0.23Te infrared photodiodes are presented to demonstrate the procedure.

While differential Hall measurements are a standard approach to determination of junction depth in multi-layer semiconductors, significantly more information can be obtained from magnetic field dependent differential Hall measurements. When such measurements are treated using Quantitative Mobility Spectrum Analysis (QMSA), detailed depth resolved profiling of both carrier concentrations and mobilities can be achieved, giving important data directly related to potential device performance. The doping profile is obtained by performing a series of etch-back experiments with magnetic field dependent Hall measurements performed between the etching steps. This technique is illustrated on a number of vacancy and gold doped Hg1-xCdxTe p-type epilayers, which have been partly or wholly converted to n-type by a reactive ion etching (RIE) process. The QMSA analysis reveals that there are several electron species present in the layers as well as the original p-type carrier. The electron species have been identified as low mobility surface electrons, and high and low mobility electrons located at various depths through the epilayer. It also indicates that the p-to-n conversion depths range from less than 0.5micrometers for vacancy doped Hg0.7Cd0.3Te material, to more than 17micrometers for Au-doped Hg0.8Cd0.2Te for the same type conversion conditions.

The minority carrier diffusion length in p-HgCdTe is a key indicator of material quality and gives an indication of n-on-p diode performance when the zero bias resistance is diffusion limited. We present results of a temperature dependent study of diffusion length in p-HgCdTe using laser beam induced current (LBIC). Carriers are collected by a p-n junction formed using standard diode junction formation conditions, and thus not necessarily extending to the substrate. Two-dimensional modeling is used to examine the validity of results obtained using this geometry, as compared to the more standard diffusion length test structure geometries, which are harder to fabricate. The temperature dependence of the diffusion length can be compared with theoretical models to determine the dominant recombination mechanisms.

While differential Hall measurements are a standard approach to determination of junction depth in multi-layer semiconductors, significantly more information can be obtained from magnetic field dependent differential Hall measurements. When such measurements are treated using Quantitative Mobility Spectrum Analysis (QMSA), detailed depth resolved profiling of both carrier concentrations and mobilities can be achieved, giving important data directly related to potential device performance. The doping profile is obtained by performing a series of etch-back experiments with magnetic field dependent Hall measurements performed between the etching steps. This technique is illustrated on a number of vacancy and gold doped Hg1-xCdxTe p-type epilayers, which have been partly or wholly converted to n-type by a reactive ion etching (RIE) process. The QMSA analysis reveals that there are several electron species present in the layers as well as the original p-type carrier. The electron species have been identified as low mobility surface electrons, and high and low mobility electrons located at various depths through the epilayer. It also indicates that the p-to-n conversion depths range from less than 0.5mum for vacancy doped Hg0.7Cd0.3Te material, to more than 17mum for Au-doped Hg0.8Cd0.2Te for the same type conversion conditions.

The minority carrier diffusion length is a key indicator of material quality and gives an indication of n-on-p diode performance when the zero bias resistance is diffusion limited. In this study diffusion length is measured using a laser beam induced current technique, applied to a shallow p-n junction formed using standard diode junction formation conditions. Two dimensional modelling is used to examine the validity of results obtained using this geometry, as compared to the more standard diffusion length test structure geometries, which are more difficult to fabricate.

A non-destructive optical characterisation technique is used for the investigation of Mercury Cadmium Telluride (HgCdTe) photovoltaic devices. The technique uses a scanning laser microscope to obtain Laser Beam Induced Current (LBIC) data from which it may be possible to extract information such as junction depth, array uniformity, and other material and device parameters. LBIC has been previously used only as a qualitative technique, but in this work the procedure is being developed into a quantitative tool. At present the only junction depth profiling techniques are destructive, while array uniformity can only be examined after bonding to readout circuits. In this paper we present both theoretical and experimental results which show that LBIC can be employed as a quantitative tool for device characterisation. The primary measure of performance of IR detectors is the zero bias dynamic resistance junction area product, R0A. LBIC measurements indicate that the peak LBIC signal varies by a factor of ≃2 for long wavelength infrared photodiodes for which the R0A varies between 70Ωcm2 (acceptable for operation and the lower bound of typical values) and 8Ωcm2 (unacceptable and typical for poor quality diodes).

The hydrogenation effects on HgCdTe diode performance are presented and the mechanism of hydrogenation is revealed. By the hydrogenation, R(0)A is increased by 30 times and photo-response is also improved. It is supposed that these are explained by the increased minority carrier lifetime by the hydrogenation. However, it is found from LBIC measurements that the minority carrier lifetime doesn't increase by the hydrogenation. An important clue that explains the hydrogenation effects is found from Hall measurements. It is found that, after the hydrogenation, the doping concentration of Hg-vacancy doped substrate decreases and the mobility increases. For the heavily hydrogenated bulk substrate, it is also found that the hydrogen passivates the whole Hg-vacancy and reveals the residual impurity and p-type doping concentration is exponentially graded. From these measurements, the diffusion current model of gradually doped diode is proposed. This model shows that the diffusion current of the graded junction diode is 2 orders of magnitude smaller than that of the abrupt junction diode, which clearly explains the R(0)A increase by the hydrogenation. Medici simulation to investigate the change of LBIC signal by the doping grading also coincides with the measurements. From these measurements and model, the hydrogenation effects are attributed to the grading of Hg-vacancy doped p-type substrate by the diffused hydrogen.

Magnetic field dependent differential Hall effect and resistivity measurements have been used to study the transport properties of Au-doped, Hg0.77Cd0.23Te which had undergone p to n type conversion by reactive ion etching. Analysing the data by Quantitative Mobility Spectrum Analysis shows that after two minutes of etching in a CH4/H-2 plasma at 0.4Wcm(-2) RF power density, the 17mum thick Hg0.77Cd0.23Te epilayer had been totally converted from p- to n-type down to the substrate. This is in stark contrast to results on Hg0.69Cd0.31Te which indicate a conversion depth of 2-3mum. The analysis indicates the presence of two electron species with peak conductivity at 2.3x10(5) cm(2)V(-1)s(-1) and 9.3x10(4)cm(2)V(-1)s(-1) respectively. A depth profile of these two carriers is also presented.