† Corresponding author. E-mail:
Project supported by the Australian Research Council (Grant Nos. FT130101708, DP170104562, LP170100088, and LE170100233), Universities Australia-DAAD German Research Cooperation Scheme (Grant No. 2014-2015), and a Research Collaboration Award from The University of Western Australia. Facilities used in this work are supported by the WA node of Australian National Fabrication Facility (ANFF).
We review our recent efforts on developing HgCdSe infrared materials on GaSb substrates via molecular beam epitaxy (MBE) for fabricating next generation infrared detectors with features of lower production cost and larger focal plane array format size. In order to achieve high-quality HgCdSe epilayers, ZnTe buffer layers are grown before growing HgCdSe, and the study of misfit strain in ZnTe buffer layers shows that the thickness of ZnTe buffer layer needs to be below 300 nm in order to minimize the generation of misfit dislocations. The cut-off wavelength/alloy composition of HgCdSe materials can be varied in a wide range by varying the ratio of Se/Cd beam equivalent pressure during the HgCdSe growth. Growth temperature presents significant impact on the material quality of HgCdSe, and lower growth temperature leads to higher material quality for HgCdSe. Typically, long-wave infrared HgCdSe (x=0.18, cut-off wavelength of
The infrared band is a portion of electromagnetic spectrum, the wavelength of which ranges from
![]() | Fig. 1. Transmission spectrum of atmosphere.[1] |
In the past few decades, HgCdTe infrared detectors have developed to be the core technology for the applications in high-performance infrared imaging and sensing. As a ternary alloy semiconductor, the cut-off wavelength of Hg1−xCdxTe can be tuned widely to cover the whole infrared region by simply tuning the Cd composition (x value). More importantly, Hg1−xCdxTe alloys have high electron mobility and long minority carrier lifetime, resulting in high device performance such as high quantum efficiency (
Apart from developing alternative substrates for growing high-quality HgCdTe, another approach to achieve infrared detectors with lower cost and larger array format size is to research new infrared materials which can be lattice-matched grown on substrates with features of larger wafer size and lower wafer cost. Ternary semiconductor alloy HgCdSe provides an ideal candidate for achieving these:[8] (i) tuneable composition and thus cut-off wavelength. The bandgap of Hg1−xCdxSe, analogues to Hg1−xCdxTe, can be tuned from 0 to 1.7 eV through adjusting the composition ratio of Hg/Cd; (ii) physical properties favourable for detector application. HgCdSe materials present physical properties similar to HgCdTe, such as electron mobility and minority carrier lifetime, which can lead to infrared detectors with high device performance; (iii) nearly lattice-matched III–V GaSb substrate. Figure
![]() | Fig. 2. Energy bandgap of semiconductors vs. lattice constants. Taken from Ref. [9] |
All of the HgCdSe and ZnTe samples in this work were grown on epi-ready GaSb (211)B substrates in a Riber 32P MBE system. Hg (7N), Cd (6N), Se (5N5), Zn (6N), and Te (6N) were used as the source materials, and standard effusion cells were used to evaporate the source materials without cracking. Note that (211)-orientated GaSb substrates were used due to the fact that Hg atoms have a larger sticking coefficient on the (211)-orientated surface. The epitaxial growth procedure was as follows: following thermal desorption of the native oxide from the GaSb substrate surface at 580 °C for 5 min without Sb background flux protection, the substrate temperature was reduced to 320 °C for growing a ZnTe buffer layer. After that, the substrate temperature was then reduced to the required temperature (70 °C–120 °C) for growing the HgCdSe epitaxial layers. After the growth of HgCdSe, the samples were cooled to room temperature without any background Hg flux. During the growth of ZnTe buffer layers, a typical beam equivalent pressure (BEP) of mid-10−7 Torr was used for Zn, while low-10−6 Torr was used for Te. During the growth of HgCdSe, a large Hg BEP (∼low–10−4 Torr) was used due to the very low sticking coefficient of Hg, while much lower Se BEP (∼low-10−6 Torr) and Cd BEP (∼mid-10−7 Torr) were used, which are comparable to those used for the MBE growth of HgCdTe.[10] It should be noted that no Sb background flux protection was introduced during the thermal desorption of native oxide from GaSb substrates because Sb is an effective dopant for HgCdSe, and thus can cause contamination to the II–VI MBE growth chamber and unintentional doping to the HgCdSe grown. In order to study the relaxation of lattice mismatch and misfit strain in the materials, the thickness of the ZnTe buffer layer was varied from 150 nm to
During the growth of HgCdSe and ZnTe layers, in-situ reflection high-energy electron diffraction (RHEED) was applied to undertake real-time monitoring of the surface crystalline quality. The interface structure and composition of the grown HgCdSe and ZnTe samples were characterized with scanning electron microscopy (SEM, Hitachi TM3030) and energy dispersive x-ray spectroscopy (EDX, Hitachi Quantax 70) which was integrated on the Hitachi SEM equipment. The crystal quality was characterized with double crystal x-ray diffraction (XRD) rocking curve measurements (Panalytical Empyrean XRD), while the strain relaxation and dislocation generation in the ZnTe buffer layer were analyzed by XRD reciprocal space mapping (RSM) measurements. The surface morphology and roughness of the HgCdeSe epilayers were measured with atomic force microscopy (AFM, Keysight 5500). The cut-off wavelength of the HgCdSe films was determined by analyzing the transmission spectra measured by Fourier transmission infrared spectroscopy (FTIR, Bruker V70). The alloy composition (x-value) of the Hg1−xCdxSe layers was determined through Rutherford backscattering spectrometry (RBS) undertaken by EAG Laboratory. The carrier mobility and carrier concentrations of the HgCdSe epitaxial layers were determined via Hall measurements followed by mobility spectrum analysis (MSA) using a variable magnetic field from 0 to 2 T. Minority carrier lifetimes were measured using the photoconductive decay technique under low-level carrier injection conditions.
Because ZnTe is nearly lattice-matched to GaSb and HgCdSe, it provides an excellent buffer layer for the subsequent growth of II–VI HgCdSe infrared materials. In addition, the ZnTe buffer layer will also be helpful for electrically isolating HgCdSe from conductive GaSb substrate, as well as preventing possible Ga atom diffusion into HgCdSe.[8] Therefore, the growth of ZnTe buffer layer is essential to the subsequent growth of high-quality HgCdSe epitaxial layer. However, ZnTe is slightly lattice-mismatched with GaSb (0.13%) and thus the thickness of the ZnTe buffer layers must be optimized in order to achieve the best relaxation of misfit strain and thus less dislocation generation in the subsequent HgCdSe layers. At UWA, ZnTe buffer layers with different thicknesses (150 nm, 300 nm, 400 nm, 600 nm, and 1000 nm) were grown without growing the subsequent HgCdSe layers, and the lattice mismatch/misfit strain in these samples was analyzed with XRD RSM measurements in order to find the optimum thickness for the ZnTe buffer layer. The RSM measurements were undertaken by doing multiple ω/2θ scans with varying ω values. The data of ω/2θ scans in real space can be converted into the data in reciprocal space with coordinates qx and qy by using the following equations:[15]
![]() |
![]() | Fig. 3. RSM data for ZnTe buffer layers grown on GaSb substrates with a thickness of (a) 150 nm, (b) 300 nm, (c) 400 nm, (d) 600 nm, and (e) 1000 nm. |
For an RSM with a symmetrical scan, the perpendicular mismatch between two diffraction peaks in the RSM indicates the lattice mismatch and strain status. Because the lattice constant of ZnTe (6.105 Å) is larger than that of GaSb (6.095 Å), ZnTe layer grown on GaSb substrate suffers a compressive strain in the plane, but a tensile strain along the growth direction (vertically), which are sketched in Fig.
Figure
To understand the growth mechanism and achieve high-quality HgCdSe materials, HgCdSe materials were grown on ZnTe/GaSb with various HgCdSe compositions (x=0.18–0.37) and various HgCdSe thicknesses (
For MBE growth, growth temperature presents significant influence on the material quality of the epitaxial layers, including surface morphology and crystal quality.[9] Figure
![]() | Fig. 6. Microscope images of the surface of HgCdSe epilayers grown under a growth temperature of (a) 70 °C, (b) 80 °C, (c) 90 °C, (d) 100 °C, (e) 110 °C, and (f) 120 °C.[12] |
Apart from surface defects, the growth temperature also presents a significant influence on the surface roughness of the HgCdSe epilayers. Figure
![]() | Fig. 7. Three-dimensional AFM images and lateral height profiles of HgCdSe epilayers grown under growth temperatures ranging from 70 °C to 120 °C.[12] |
Because microscope and AFM imaging can only provide information about the surface quality of the samples, XRD rocking curve measurements were undertaken on HgCdSe epilayers in order to examine the crystal quality of the whole epilayers. Figure
![]() | Fig. 8. (a) XRD FWHM of HgCdSe epilayers grown under various growth temperatures ranging from 70 °C to 120 °C; (b) representative SEM images of the HgCdSe surface after the 30 s EPD etch.[11] Inset of (a) shows the representative XRD rocking curve for HgCdSe grown under a growth temperature of 70 °C. |
Apart from structural characterization, at UWA significant effort has been devoted to studying the optical properties of HgCdSe epilayers, mainly the cut-off wavelength of HgCdSe with various alloy compositions and under various temperatures. Figure
To further explore the bandgap behavior of HgxCd1−xSe, temperature-dependent FTIR measurements were also carried out on the Hg1−xCdxSe samples with x values of 0.37, 0.21, and 0.19. Figure
For infrared detector applications, the two main physical parameters which directly impact on the device performance of infrared detectors are the carrierʼs mobility and the minority carrier lifetime of the infrared materials. Because GaSb substrates are conductive, standard Hall analysis cannot separate the Hall signal of HgCdSe layers from that of GaSb. At UWA, in order to determine the carrier mobility in HgCdSe, MSA measurements were undertaken on the HgCdSe samples, which allow the conductivity contribution of the HgCdSe epitaxial layer to be separated from that of the conductive GaSb substrate.[22] Note for this MSA measurement, standard Van der Pauw Hall measurements were first performed on HgCdSe materials, and then MSA technique was used to fit the raw Hall data to identify the conductivity contribution of Hg1−xCdxSe sample and that of GaSb substrate.[22] Figure
![]() | Fig. 10. Electron mobility spectrum at 132 K of (a) GaSb substrate only and (b) HgCdSe (x=0.18) epilayer grown at 70 °C on GaSb substrate.[11] |
To better understand the electrical properties of HgCdSe, temperature dependent MSA measurements were also undertaken on the HgCdSe sample with x=0.18. Figure
![]() | Fig. 11. Temperature dependent (a) electron mobility and (b) electron concentration of as-grown HgCdSe (x=0.18) grown at 70 °C on GaSb substrate.[11] |
Apart from carrier mobility, minority carrier lifetime is another important physical parameter impacting the performance of infrared detectors. At UWA, photoconductive decay measurements were used to study the minority carrier lifetime in the HgCdSe epitaxial layers grown on GaSb. Figure
![]() |
![](cpb_28_1_018103/cpb_28_1_018103_ieqn31.gif)
![](cpb_28_1_018103/cpb_28_1_018103_ieqn32.gif)
![](cpb_28_1_018103/cpb_28_1_018103_ieqn33.gif)
![](cpb_28_1_018103/cpb_28_1_018103_ieqn34.gif)
![](cpb_28_1_018103/cpb_28_1_018103_ieqn35.gif)
![](cpb_28_1_018103/cpb_28_1_018103_ieqn36.gif)
![]() | Fig. 12. (a) The 80 K photoconductive decay measurement and (b) temperature dependent minority carrier lifetime of HgCdSe (x=0.18) grown at 70 °C on GaSb substrate.[11] |
As discussed in Subsection 4.1, the growth temperature presents a significant influence on the structural properties of HgCdSe. Similarly, the growth temperature can also have a significant influence on the electrical properties of HgCdSe. To study the impact of growth temperature on electrical properties, MSA and photocurrent decay measurements were also undertaken on HgCdSe samples with similar x values (∼0.18) but grown at different growth temperatures ranging from 70 °C to 120 °C. Table
![]() | Table 1.
XRD FWHM, electron mobility, and minority carrier lifetime of HgCdSe (x=0.18) materials grown at different growth temperatures.[11] . |
In this work, we have reviewed our recent efforts on developing HgCdSe infrared materials on GaSb substrates for fabricating infrared detectors with new features of lower cost and larger array format size. At UWA, HgCdSe materials are grown via MBE on ZnTe/GaSb substrates, instead of ZnTe/Si substrates in order to leverage the benefit of nearly lattice-matched growth of ZnTe and HgCdSe on GaSb. The study on the strain/lattice mismatch relaxation in ZnTe buffer layers grown on GaSb shows that the thickness of the ZnTe buffer layer needs to be controlled below 300 nm in order to minimize the generation of misfit dislocations in the material system and thus achieve high-quality HgCdSe grown subsequently. The cut-off wavelength/alloy composition of HgCdSe materials can be tuned widely from SWIR, to MWIR, and to LWIR by varying the Se/Cd BEP ratio during the HgCdSe growth. Growth temperature presents significant impact on the material quality of HgCdSe materials including surface defects, surface roughness, and electrical properties, and lower growth temperature can lead to higher material quality for HgCdSe materials. With the optimizing growth temperature, HgCdSe materials grown at UWA present high material quality with electron mobility and minority carrier lifetime that are significantly improved over the values obtained in previous work, and comparable to those for counterpart HgCdTe materials grown on lattice-matched CdZnTe substrates. Typically, LWIR HgCdSe (x=0.18, 80 K cut-off wavelength of
Although significant progress has been made in developing HgCdSe infrared materials on GaSb substrates, it is still quite challenging to achieve their ultimate industry applications. One of the main challenges is the high level of background electron concentration in as-grown HgCdSe materials. Despite the fact that the background electron concentration in our as-grown HgCdSe epitaxial layers, e.g., 1.6×1016 cm−3 for LWIR HgCdSe (x=0.18, 80 K) is one order of magnitude lower than the corresponding value reported in previous studies,[26] this level of background electron concentration, however, is not suitable for making high performance advanced infrared detectors based on p–n junctions. The background carrier concentration needs to be controlled below the low-1015 cm−3 level to allow effective and controllable extrinsic doping in the 1016 cm−3 range or above, which is required for fabricating p–n junctions for high-performance detectors based on photovoltaic and/or other modern device structures. Therefore, it is essential to reduce the background carrier concentration in HgCdSe materials. There are several potential approaches to address this challenge: (i) post-growth thermal annealing in a Se environment. As the high level of background carrier concentration in the as-grown HgCdSe is mainly caused by Se vacancies in HgCdSe material, post-growth thermal annealing in a Se environment can significantly reduce the Se vacancies, and thus effectively reduce the background doping concentration. One order of magnitude reduction in background electron concentration has been demonstrated for HgCdSe after post-growth annealing in a Se environment.[8] With further optimizing the annealing process, the background electron concentration can be further reduced. (ii) Utilization of higher purity Se source material for growing HgCdSe. At UWA, only 5N5 purity Se source material is used for growing HgCdSe materials. It has been demonstrated that higher purity Se source material can also effectively reduce the level of impurities in the HgCdSe epilayer, and thus the background doping concentration.[8] If Se source material with 7N purity or above is commercially available, it can be expected that the background electron concentration can be significantly reduced. (iii) Guttering/filtering of Se vacancies and impurities. It has been demonstrated in HgCdTe material that defects and impurities can be significantly reduced with properly designed gettering/filtering structures,[3] which can also be applied to HgCdSe materials. Therefore, it can be expected that a background electron concentration level within the low-1015 cm−3 range or even lower is achievable for HgCdSe with implementing the above approaches.
In summary, the MBE growth of HgCdSe on GaSb substrates is only a recent effort, and even higher quality HgCdSe materials are expected with further effort on this. Although significant amount of work is required to control the background electron concentration to
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] | |
[24] | |
[25] | |
[26] | |
[27] |