Hironori Nishino

Direct growth of CdTe on (100), (211), and (111) Si by metalorganic chemical vapor deposition

CdTe epilayers were grown directly on (100), (211), and (111) silicon substrates by metalorganic chemical vapor deposition (MOCVD). The crystallinity and the growth orientation of the CdTe film were dependent on the surface treatment of the Si substrate. The surface treatment consisted of exposure of the Si surface to diethyltelluride (DETe) at temperatures over 600°C prior to CdTe growth. Direct growth of CdTe on (100) Si produced polycrystalline films whereas (lll)B single crystals grew when Si was exposed to DETe prior to CdTe growth. On (211) Si, single crystal films with (133)A orientation was obtained when grown directly; but produced films with (211)A orientation when the Si surface was exposed to DETe. On the other hand, only (lll)A CdTe films were possible on (111) Si, both with and without Te source exposure, although twinning was increased after exposure. The results indicate that the exposure to a Te-source changes the initial growth stage significantly, except for the growth on (111) Si. We propose a model in which a Te atom replaces a Si atom that is bound to two Si atoms.

Growth of (111) HgCdTe on (100) Si by MOVPE using metalorganic tellurium adsorption and annealing

Abstract(lll)B CdTe layers free of antiphase domains and twins were directly grown on (100) Si 4°-misoriented toward<011> substrates, using a metalorganic tellurium (Te) adsorption and annealing technique. Direct growth of (lll)B CdTe on (100) Si has three major problems: the etching of Si by Te, antiphase domains, and twinning. Te adsorption at low temperature avoids the etching effect and annealing at a high temperature grows single domain CdTe layers. Te atoms on the Si surface are arranged in two stable positions, depending on annealing temperatures. We evaluated the characteristics of (lll)B CdTe and (lll)B HgCdTe layers. The full width at half maximum (FWHM) of the x-ray double crystal rocking curve (DCRC) showed 146 arc sec at the 8 |im thick CdTe layers. In Hg1−xCdxJe (x = 0.22 to 0.24) layers, the FWHMs of the DCRCs were 127 arc sec for a 7 (im thick layer and 119 arc sec for a 17 (im thick layer. The etch pit densities of the HgCdTe were 2.3 x 106 cm2 at 7 ^m and 1.5 x 106 cm-2 at 17 um.

VIII ratio dependence of surface macrodefects in CdTe/ZnTe/GaAs(100) growth by metalorganic vapor phase epitaxy

We studied the surface morphology of CdTe(100) layers on GaAs(100) by metalorganic vapor phase epitaxy (MOVPE). When CdTe(100) layers were obtained using thin ZnTe nucleation layers, we observed high-density pyramidal macrodefects, known as hillocks, in the epilayer surface. We found the density of hillocks to be strongly dependent on the VIII ratio during both ZnTe and CdTe growth processes. We optimized both the TeZn and TeCd ratios to obtain a minimum hillock density of 1 × 102 cm−2. These results show that hillocks were nucleated at both the epilayer/substrate and CdTeZnTe interfaces. By X-ray diffraction measurement, we confirmed that the quality of the crystal structural was also good under this condition. We also found the initial nucleation conditions to be more important for the structural quality. In (100)HgCdTe/CdTe/GaAs growth, pyramidal hillocks on the CdTe buffer surface caused substantial (> 100 μm) macrodefects in HgCdTe layers, which were fatal for infrared devices. Their shape was enlarged especially in one direction. To achieve a low density of surface macrodefects in HgCdTe(100) or CdTe(100) layers on GaAs(100) substrates, we need to precisely control the VIII ratio.

Dislocation profiles in HgCdTe(100) on GaAs(100) grown by metalorganic chemical vapor deposition

We studied dislocation etch pit density (EPD) profiles in HgCdTe(lOO) layers grown on GaAs(lOO) by metalorganic chemical vapor deposition. Dislocation profiles in HgCdTe(lll)B and HgCdTe(lOO) layers differ as follows: Misfit dislocations in HgCdTe(lll)B layers are concentrated near the HgCdTe/CdTe interfaces because of slip planes parallel to the interfaces. Away from the HgCdTe/CdTe interface, the HgCdTe(111)B dislocation density remains almost constant. In HgCdTe(lOO) layers, however, the dislocations propagate monotonically to the surface and the dislocation density decreases gradually as dislocations are incorporated with increasing HgCdTe(lOO) layer thicknesses. The dislocation reduction was small in HgCdTe(lOO) layers more than 10 μm from the HgCdTe/CdTe interface. The CdTe(lOO) buffer thickness and dislocation density were similarly related. Since dislocations glide to accommodate the lattice distortion and this movement increases the probability of dislocation incorporation, incorporation proceeds in limited regions from each interface where the lattice distortion and strain are sufficient. We obtained the minimum EPD in HgCdTe(100) of 1 to 3 x 106 cm-2 by growing both the epitaxial layers more than 8 μm thick.

CdTe(111)B/Si(100) structure grown by metalorganic vapor phase epitaxy with Te adsorption and annealing

We studied the crystal structure of CdTe(111)B layers directly grown on Si(100) by MOVPE using a new pre-growth process, which includes a metalorganic Te adsorption and an annealing process. In this paper, we discussed the CdTe structure from the three aspects of antiphase, twinning and tilt. We investigated the dependence of the antiphase content in CdTe(111)B on the anneal temperature and the Si misorientation angle. From the results, we assume that the origin of the antiphase formation is the difference in the arrangement of adsorbed Te atoms. Te arrangement leading to antiphase formation occurs on Si terraces away from steps at relatively low temperatures. We reduced most of the twinning in epilayers by optimizing the VI/II ratio. We think the remaining twinning was confined to near the interface and it nucleated from the Te arrangement on terraces. We found that the Si(100)-CdTe(111) tilt was much smaller than that expected from the well-known Nagai model. We propose that a negative tilt is induced to reduce the lateral mismatch. To adjust the lateral distance of unit cells, 30 CdTe lattices match to 31 Si lattices. CdTe(111)B planes are inclined to reduce the remaining mismatch between two lattices. This initial tilt also causes wider CdTe terraces. We modified Nagais tilting model for this reconstructed CdTe surface. The total tilt angle is defined by these two tilting mechanisms.

Effect of dislocations on 1/f noise of long wavelength infrared photodiodes fabricated with HgCdTe layers grown on GaAs by metalorganic vapor phase epitaxy

We studied the effect of dislocations on the 1/f noise current of long wavelength infrared photodiodes fabricated with HgCdTe layers grown on GaAs by metalorganic vapor phase epitaxy. N-on-p junctions were formed by boron ion implantation into Hg-vacancy doped epilayers. The 1/f noise dominated from 0.5 to 100 Hz, and shot noise caused by photocurrent (√2eIp) dominated at higher frequencies. We observed two types of 1/f noise. One is caused by the leakage current generated at dislocations, and the other is induced by the photocurrent. The 1/f noise current increased with the photon flux in the low-etch pit density (EPD) range independently of EPD. It increased with EPD in the high-EPD range. The 1/f noise current measured at zero field of view increased with EPD. This suggests that the 1/f noise generated by the photocurrent dominated in the low-EPD range, and that the 1/f noise current caused by dislocations dominated in the high-EPD range. In order to obtain a thermal image of a room-temperature object, the 1/f noise current induced by background photon flux is as high as that caused by dislocations of more than 107 cm−2. Therefore, the 1/f noise current induced by the photocurrent is dominant in photodiodes fabricated with HgCdTe layers on GaAs, since the EPD is less than 2 x 106 cm−2. We expect the detectivity to be as high as with LPE-layers. We fabricated 64 x 64 photodiode arrays, and obtained a thermal image.

Misfit stress relaxation mechanism in CdTe(100) and CdTe/ZnTe(100) on a GaAs(100) highly mismatched heteroepitaxial layer

We studied the misfit stress in CdTe(100)/GaAs(100). In general, this highly mismatched system initially forms a ‘‘coincidence interface,’’ where seven CdTe lattices match with eight GaAs lattices. In this system, the epilayer is elastically strained by the mismatch remaining between both groups of lattices. Afterward, the misfit stress is gradually relaxed with the generation of misfit dislocations. In this work, we derived a new model to describe this misfit relaxation mechanism and verified it with experimental observations. As introducing the effects of the interface modification, we found that the simple force balance between the substrate and epilayer governed the misfit stress. We assume that the relaxation process during growth is similar to that in CdTe/CdZnTe, since the residual misfit was fairly small. By extending the stress relaxation model of CdTe/GaAs, it is possible to explain the strain in CdTe/ZnTe/GaAs, which contains two highly mismatched interfaces. The threading dislocations in HgCdT...

Influence of step motion on hillock formation in CdTe(100) grown on GaAs(100) by metalorganic vapor phase epitaxy

We propose a new model for explaining hillock formation in CdTe(100) layers grown on GaAs(100) by metalorganic vapor phase epitaxy (MOVPE). Although many pyramidal surface macrodefects (hillocks) were observed in MOVPE-grown CdTe surface during our study, there was no structural defects attributable to as their origin. We found only line dislocations and no other defects such as the stacking faults or twinnings from the images of transmission electron microscopy (TEM). We thus suggest that the hillocks had formed due to a 2D nucleation epitaxial growth without resulting in any crystalline defects. When atoms in the vapor phase are properly incorporated into the CdTe crystal during stepped growth, step-flow growth dominates the epitaxial growth, and hillock formation is suppressed. We estimated the incorporation probability for CdTe surface adatoms during the advance of growth steps from hillock configuration, and found that incorporation probability depends on the VI/II ratio and this dependence has opposite effects on A- and B-step growths. When we grew CdTe(100) on GaAs(100) misoriented towards the nearest (110), we were able to obtain smooth surface using VI/II ratio of around 1. This observation demonstrated that at this VI/II ratio, adatoms are properly incorporated for both A- and B-step growths.

Compositional profile of HgCdTe in metalorganic chemical vapor deposition (MOCVD) system with multinozzles

We studied the dependence of the compositional profile on the nozzle configuration in Hg1−xCdxTe growth using a metalorganic chemical vapor deposition (MOCVD) system having multinozzle injectors and a rotating susceptor in a vertical reactor. In growth, the composition depends on the concentration of dimethyl-cadmium (DMCd). In the epilayer grown by injecting DMCd from one nozzle, the x-value was highest just beneath the nozzle. The x-value peak shifts towards the rotation center when the epilayer is grown from an off-center nozzle with rotation. The compositional profile of the epilayer grown by the injection of DMCd from two nozzles was the same as the summation of the two profiles produced with one, and then the other nozzle. In a simulation based on these results, we found that 8 nozzle injectors are needed to grow the epilayer on a 3-inch substrate within a compositional variation (Δx) of 0.002.

Metalorganic chemical vapor deposition growth of CdTe on GaAs in a vertical reactor with multi-nozzles

Abstract The metalorganic chemical vapor deposition (MOCVD) growth of CdTe films on 3 inch diameter (100) GaAs substrates using a vertical rotating-susceptor reactor with linearly alligned multi-nozzles was studied. Specular single ( 1 1 1 )B CdTe films over the entire 3 inch wafer were obtained by adjusting the VI/II ratio individually for each nozzle. The thickness variation for a 2.5 μm thick CdTe films was below 8%. The full-width at half-maximum (FWHM) of the (333) double-crystal X-ray rocking curve (DCRC) was about 250 arc sec. To evaluate its suitability as a substrate for HgCdTe epilayers, a Hg0.7Cd0.3Te layer 3 μm thick was grown on the CdTe/GaAs using the same MOCVD system. The HgCdTe layer also showed a specular surface. No macroscopic crystal defects such as cellular structures or antiphase domains were observed on the etched HgCdTe surface.