Paul M. Amirtharaj

Spectroscopic ellipsometry determination of the properties of the thin underlying strained Si layer and the roughness at SiO2/Si interface

The existence of both the strain and microroughness at the interface of thermally grown SiO2 films on Si was ascertained unambiguously for the first time by high accuracy spectroscopic ellipsometry. The dielectric function of the interface was determined by a comprehensive data analysis procedure. By carefully examining the dielectric function obtained by our model, the strain was seen to cause a red shift of 0.042 eV of the interband critical point E1 compared with the bulk silicon value. The thickness of the interface region was found to be 2.2 nm of which a significant part is due to the strain.

Photoreflectance study of Hg0.7Cd0.3Te and Cd1−xZnxTe: E1 transition

We present results of the first photoreflectance (PR) study in Hg0.7Cd0.3Te and MBE and bulk Cd1−xZnxTe and the first detailed measurement of the variation of E1 optical transition energy with x in Cd1−xZnxTe at 77 K. Photoreflectance is a completely contactless form of electroreflectance. The E1 optical transition line shapes were measured and analyzed using currently available models. The MBE Cd1−xZnxTe samples exhibited E1 features that could not be fit satisfactorily, which may be due to the presence of an electron‐hole interaction that was not included. The results indicate that the dependence of E1 with x may be expressed as follows: E1(x)=3.523 eV−0.022 eV ⋅ x+0.267 eV ⋅ x2. The linewidths of the E1 feature measured at 77 K in all the samples were much smaller than those at 300 K owing to the reduction in temperature broadening. This suggests that, in addition to being contactless, the characterization accuracy available with PR may be significantly better than that available in techniques such as ...

Self-aligned molecular beam epitaxy of CdZnTe for IR focal plane arrays

The crystallographic orientation of Cd1-xZnxTe (x approximately equals 0.045) grown by molecular beam epitaxy (MBE) on a clean (planar) (100) GaAs surface can be controlled by the proper choice of the GaAs surface stoichiometry. An As-stabilized surface initiates (100) oriented growth, while the Ga-stabilized surface yields (111) oriented growth. Cd1-xZnxTe (x approximately equals 0.045) MBE layers grown in recesses of shadow masked patterned (100) GaAs substrates were found to be in the (100) orientation regardless of whether precursor surfaces were stabilized with Ga or As. The epitaxial layers orientation and optical properties were determined by backscattered electron channeling and low temperature photoluminescence measurements, respectively. CdZnTe layers grown in recesses showed improved optical features as compared to the layers grown on planar substrates.

Interface roughness of short‐period AlAs/GaAs superlattices studied by spectroscopic ellipsometry

Spectroscopic ellipsometry (SE) has been used to study the effects of interface roughness on the optical properties of ultrathin short‐period 3×3 GaAs/AlAs superlattices grown by molecular‐beam epitaxy (MBE). The complex dielectric function and thickness of the whole superlattice and the thickness of the native oxide overlayer were simultaneously determined by an inversion technique from data in the 1.5–5.0 eV region. The main optical critical points E0, E0+Δ0, E1, E1+Δ1, and E2 were deduced by line‐shape fitting of the second derivative of the complex dielectric function of the superlattice to the analytical line‐shape expression. The interface roughness is found to shift the optical transitions, except E2, to higher energy and broaden their line shapes. A simple interpretation of the shift and broadening is given. The interface roughness and layer thicknesses obtained by SE are found to be consistent with the results of x‐ray diffraction and Raman scattering studies previously reported. The results in t...

Interface sharpness in low-order III–V superlattices

Abstract Superlattices composed of aluminum, gallium and indium are currently employed in a variety of device-related applications. Among these are edge-emitting GRINSCH lasers and vertical cavity surface emitting laser diodes. As the individual layer thickness is reduced, the role of interface sharpness becomes more critical in ensuring good two-dimensional growth. This work addresses the relationship between interface roughness and superlattice crystallinity for short-period AlAs/GaAs superlattices. Thin short-period superlattices with active layer thicknesses of 30 nm or less were also investigated to help determine the interface sharpness in the initial stages of growth. X-ray diffraction was used to assess interface roughness and to calculate superlattice periodicity. These results are compared with those obtained by reflection high energy electron diffraction (RHEED), Raman spectroscopy, and spectroscopic ellipsometry. The results indicate that interface roughness is promoted by a reduced arsenic flux growth condition at normal growth temperatures for short-period superlattices. The results also suggest that, for thin superlattices, a 10 nm buffer layer enhances interface roughness in the initial stages of growth and compromises the subsequent epilayer crystallinity. An analysis of these results in the light of structural, dynamical, and optical data is presented.

Photoluminescence quenching in Si1−xGex/Si multiple quantum wells grown with atomic hydrogen

We compare the photoluminescence spectra from a series of Si1−xGex/Si (0.1⩽x⩽0.3) multiple quantum well (MQW) samples grown with atomic hydrogen to a series of similar samples grown without atomic hydrogen. All of the samples were grown at 710 °C. We observe intense quantum confined photoluminescence in the Si1−xGex/Si MQW samples grown without atomic hydrogen. No quantum confined photoluminescence was observed in the Si1−xGex/Si MQW samples grown with atomic hydrogen. This was unexpected, since quantum confined photoluminescence has been observed in the hydrogen-assisted growth of Si1−xGex/Si quantum well structures grown at lower temperatures by other researchers. We believe that this is caused by defects, introduced into the samples during growth with atomic hydrogen, which lead to efficient nonradiative centers that compete with the radiative centers. These defects lead to a reduction of the photoluminescence of the samples grown with atomic hydrogen. Since hydrogen is used as a surfactant to reduce t...

Interface sharpness during the initial stages of growth of thin, short‐period III–V superlattices

Superlattices composed of III–V heterostructures have established applications in high‐speed electronic and optoelectronic devices. As layer thicknesses are reduced, the role of heterostructure interface sharpness becomes more critical to ensuring high quality two‐dimensional growth. In this work, short‐period (less than 1 nm) superlattices with active layer thicknesses of 31 nm were investigated to assess interface roughness in the initial stages of growth. X‐ray diffraction was used to evaluate interface roughness and to calculate superlattice periodicity. Results suggest that surface roughening by islanding may be promoted by GaAs buffer layers that are 10–100 nm thick. Smoother interfaces were obtained in samples with buffer layers 250 nm and greater.

Study of phonons in semiconductor superlattices by Raman scattering spectroscopy and microscopic model calculation

Abstract Raman spectroscopy is used to study phonons in a series of thin (AlAs)m/(GaAs)n superlattices (SLs) grown by molecular beam epitaxy (MBE). The influence of buffer layer type on the interface roughness of heterostructures is carefully evaluated. The accuracy of optical phonons and the degree of peak sharpness of GaAs-like confined modes are examined via off-resonance Raman spectroscopy. Theoretical calculations of phonons in thin (AlAs)m/(GaAs)n superlattices (i.e. samples with m, n ≤ 12) are reported for various directions of propagation by using a rigid-ion model. Optical phonons acquire significant dispersive character when the wavevector q forming an angle θ with the growth axis of the superlattice is changed from θ = 0 to π/2, i.e. from [001] to [100]. The frequency gaps in the angular dispersions due to mode anti-crossing behavior observed recently by Zunke et al. using micro-Raman spectroscopy and studied by a continuum model are found to be in reasonably good agreement with our lattice dynamical model calculations.

Buffer layer-modulation-doped field-effect-transistor interactions in the Al0.33Ga0.67As/GaAs superlattice system

The correlation between the structural and transport properties for a series of high‐quality modulation‐doped field‐effect‐transistor (MODFET) structures was made for various growth temperatures. X‐ray reflectivity, x‐ray diffraction, and magnetotransport measurements were used to assess structural quality and transport parameters. Four samples with growth temperatures in the range 500–630 °C were examined. The results show a correlation exists between the measured electron mobility and the quality of the interface width, as measured from satellite peaks of the buffer layer. In addition, these results show, for the first time to the best of our knowledge, that a direct correlation can be made between x‐ray reflectivity structural measurements and the measured electron mobility of high‐quality gallium–arsenide‐based MODFETs. Both x‐ray and transport results suggest a higher‐quality structure was obtained at higher growth temperatures.

High-resolution, high-accuracy, mid-IR (450 cm−1⩽ω⩽4000 cm−1)refractive index measurements in silicon

The real and imaginary parts of the refractive index of silicon, n(ω) and k(ω), have been measured by using Fourier Transform Infrared (FTIR) transmission spectral data from a double-sided-polished grade Si wafer. An accurate mechanical measurement of the wafer thickness, t, was required and two FTIR spectra were used: a high-resolution (Δω=0.5 cm−1) yielding a typical channel spectrum dependent mainly on t and n(ω), and a low resolution (Δω=4.0 cm−1) yielding an absorption spectrum dependent mainly on t and k(ω). Independent analysis of each spectrum gave initial n(ω) and k(ω) estimates which were then used together as starting point values for an iterative fit of the high- and low-resolution spectra successively. The accuracy of n(ω) and k(ω) values determined using this procedure is dependent upon the measurement error in the sample thickness, the absolute transmission values obtained from a sample-in and sample-out method, and the modeling of the influence of wafer thickness nonuniformity and the degr...