Control and Growth of Middle Wave Infrared (MWIR) Hg^sub (1-x)^Cd^sub x^Te on Si by Molecular Beam Epitaxy

Journal of Electronic Materials, Jun 2005 by Vilela, M F, Buell, A A, Newton, M D, Venzor, G M, Et al

Middle wave infrared (MWIR) HgCdTe p-on-n double-layer heterojunctions (DLHJs) for infrared detector applications have been grown on 100-mm Si (112) substrates by molecular beam epitaxy (MBE) for large format 2,560 × 512 focal plane arrays (FPAs). In order to meet the performance requirements needed for these FPAs, cutoff and doping uniformity across the 100-mm wafer are crucial. Reflection high-energy electron diffraction (RHEED), secondary ion mass spectrometry (SIMS), Fourier transform infrared spectrometry (FTIR), x-ray, and etch pit density (EPD) were monitored to assess the reproducibility, uniformity, and quality of detector material grown. Material properties demonstrated include x-ray full width half maximum (FWHM) as low as 64 arc-sec, typical etch pit densities in mid-10^sup 6^ cm^sup -2^, cutoff uniformity below 5% across the full wafer, and typical density of macrodefects

Key words: Middle wave infrared (MWIR), HgCdTe, HgCdTe/Si, molecular beam epitaxy (MBE), infrared photodiodes, large-format focal plane arrays (FPAs)

INTRODUCTION

HgCdTe infrared focal plane arrays (FPAs) continue to grow in dimensions, which require the growth of high-quality HgCdTe material on large area substrates. Material for production needs to be reproducible, and to have a high yield and affordable cost. The molecular beam epitaxy (MBE) HgCdTe/Si technology has proven the ability to produce FPAs, which indicates that Si substrates offer a viable solution for cost-effective largearea HgCdTe epitaxy.1-3 Additionally, Si provides a robust substrate for processing and its thermalexpansion is matched to the Si readout integrated circuit. In this case, the thin CdTe buffer layer and HgCdTe active layers are constrained to the thick Si substrate.

This work summarizes recent progress on the growth of 4-in. HgCdTe/Si for MWIR FPAs and describes the quality, reproducibility, and uniformity of these materials. The techniques that lead to growth temperature and flux control are presented as well as the main material characteristics. The reflection high-energy electron diffraction (RHEED), secondary ion mass spectrometry (SIMS), fourier transform infrared spectroscopy (FTIR), x-ray, and etch pit density (EPD) were monitored to assess and control the uniformity and quality of grown detector material. In addition, it is demonstrated that despite the huge lattice mismatch of about 19% between epilayers and Si substrate, which induces a much higher dislocation density compared to growth of HgCdTe on CdZnTe substrates (~10^sup 6^ cm^sup -2^ instead of 10^sup 5^ cm^sup -2^), HgCdTe/Si MWIR detector performance is shown to be comparable to HgCdTe/ CdZnTe.

EXPERIMENT

All heteroepitaxial layers (ZnTe, CdTe, and HgCdTe) were deposited in a 125-mm Riber Epineat MBE system (Rueil, France) using CdTe, ZnTe, and Te solid sources and a Raytheon-built liquid Hg source, which can be refilled without breaking the main vacuum chamber. The 100-mm diameter (112) Si substrates are prepared for growth using a hydrofluoric acid-based process. After the wet etch, wafers are loaded in a ultra high vacuum (UHV) introduction chamber prior to growth. The ZnTe/CdTe buffer layer and HgCdTe detector structures are then grown without wafer removal from the MBE system. All growth is performed in a rotating manipulator. Each growth begins with a thick buffer layer consisting of about 1 µm of ZnTe and 8-9 µm of CdTe. Early studies4,5 have demonstrated that the ZnTe initiation layer is important to preserve the substrate orientation, in the present case the (112)B, before the CdTe layer, which serves to attenuate the propagating threading dislocation density. After the buffer layer, a p-on-n MWIR double-layer heterojunction (DLHJ) having a room-temperature (RT) cutoff wavelength (λ^sub co^) of about 4.7 µm is grown. This wavelength corresponds to a cadmium x-value of around 0.285 for Hg^sub (1-x)^Cd^sub x^Te. The n-type region is doped with indium and the p-type region is doped with arsenic. A more detailed report about the performance of some MWIR arrays described in this work can be found in Ref. 6.

RESULTS AND DISCUSSION

General Characteristics

Table I summarizes typical characteristics that describe the 4-in MWIR HgCdTe/Si devices (material and performance) used in this work. The x-value and RT λ^sub co^ are determined by FTIR transmission measurements, macrodefects density (voids microvoids) are determined from optical microscopy, and n- and p-type doping are determined by SIMS. The radial composition uniformity is determined from the cut off from the center to the edge of the wafer, and the uniformity is calculated from the ratio of maximum to minimum values.

A great deal of effort is spent to control key growth parameters in order to guarantee the uniformity, quality, and reproducibility of devices grown. Some of the most important parameters are growth temperature, source beam flux stability, doping level, x-value, defect density, EPD, and x-ray FWHM. All parameters are closely interrelated, but in the following sections, each will be discussed separately.