T. H. Ghong

InAs critical-point energies at 22 K from spectroscopic ellipsometry

We report dielectric function data from 0.74 to 6.54 eV for InAs at 22 K, obtained by spectroscopic ellipsometry. Critical-point (CP) structures are blueshifted and significantly sharpened relative to those seen at room-temperature (RT). The E0′, E2Δ, E2, E0′+Δ0′, and E2′ features in the E2 energy range of 4.0 to 5.6 eV cannot be resolved at RT but are clearly separated at 22 K. The energies of the CPs giving rise to these structures are determined by line shape fitting to numerically calculated second energy derivatives, and their Brillouin-zone locations identified by band structure calculations using the linear augmented Slater-type orbital method.

Effect of overlayers on critical-point parameters in the analysis of ellipsometric spectra

As the complete removal of overlayers may not be possible in general, the authors investigate the effect of incomplete removal on critical-point parameters in the analysis of ellipsometric data. Using an approximate analytic expression, they show that energies and broadening parameters are much less affected by overlayers than amplitudes and phases. These conclusions are confirmed by false-data calculations for GaAs and overlayer-removal data for CdTe.

Dielectric response of AlSb from 0.7 to 5.0 eV determined by in situ ellipsometry

We present pseudodielectric function data ⟨e⟩=⟨e1⟩+i⟨e2⟩ from 0.7 to 5.0 eV of oxide-free AlSb that are the closest representation to date of the intrinsic bulk dielectric response e of the material. Measurements were done on a 1.5 μm thick film grown on (001) GaAs by molecular beam epitaxy. Data were obtained with the film in situ to avoid oxidation artifacts. Overlapping critical-point (CP) structures in the E2 energy region were identified by means of band-structure calculations done with the linear augmented Slater-type orbital method. Calculated CP energies agree well with those obtained from data, confirming the validity of the calculations.

Optical properties of InxAl1−xAs alloy films

Pseudodielectric functions ⟨e⟩ of InxAl1−xAs ternary alloy films were determined from 1.5to6.0eV by spectroscopic ellipsometry. We minimized overlayer effects by performing wet-chemical etching to more accurately determine intrinsic bulk dielectric responses. Energies of the E1, E1+Δ1, E0′, E2, E2+Δ2 and E2′ critical points (CPs) were identified by band structure calculations of the linear augmented Slater-type orbital method. These calculations also showed a crossing of the E0′ and E2 CP structures with increasing In composition and a new saddle point in the AlAs band structure.

Electrical properties of the Sm2Ti2O7 thin films for metal-insulator-metal capacitor applications

A homogeneous crystalline Sm2Ti2O7 (ST) phase was formed in films grown at temperatures ranging between 100 and 200 °C and subsequently annealed at 900 °C. The ST film had a large dielectric constant of 58, which is similar to that of ST ceramics. The leakage current density of the ST film was low and the Poole-Frenkel emission was suggested as being the leakage current mechanism. The ST film had a negative quadratic voltage coefficient of capacitance (VCC), possibly due to the dipolar relaxation. The 100-nm-thick ST film had a high capacitance density of 5.2 fF∕μm2 with a low leakage current density of 1.34 nA∕cm2 at 2 V. Its quadratic and linear VCCs were −99.5 ppm∕V2 and 11 ppm/V, respectively, with a low temperature coefficient of capacitance of 135 ppm∕°C at 100 kHz. These results confirmed the potential for the ST film to be used as a high performance metal-insulator-metal capacitor.

Imaging ellipsometry study on the Ni-mediated crystallization of a-Si

We have investigated by imaging ellipsometry (IE) the crystallization of amorphous silicon (a-Si) at low temperature by Ni-silicide-mediated crystallization on a glass substrate. We observed a significant difference in the ellipsometric image between polycrystalline silicon (poly-Si) and a-Si and confirmed that annealing for longer time leads to increasing area of poly-Si domain. The progress of crystallization is modeled as the growth of a disk in a two-dimensional surface since the crystallized region expanded as a disk shape. The rate of increasing radius of the disk and the average distance between initial nucleus points were obtained by fitting the change of the area of the disk measured by the IE with annealing time.

Investigation of effective-medium approximation, alloy, average-composition, and graded-composition models for interface analysis by spectroscopic ellipsometry

We critically test the capabilities of the effective-medium approximation (EMA) and alloy models to describe multilayer samples with gradual interfaces by analyzing spectroscopic ellipsometric (SE) data of two AlGaAs samples grown expressly for this purpose. The dielectric functions e of the interfaces are calculated in the EMA and alloy models, and the interfaces themselves simulated either as a single layer of Al0.5Ga0.5As or a stack of layers of AlxGa1−xAs with x increasing or decreasing between 0.1 and 0.9 in increments of 0.1. The EMA essentially fails completely for either interface representation. For the alloy model the stepwise-graded representation is significantly better, not only simulating the data more accurately but also yielding thicknesses in essential agreement with those obtained by cross-sectional transmission electron microscopy. The results highlight the types of errors that are encountered with the different models, and show that the analysis of SE data can provide information about...

Overlayer effects in the critical-point analysis of ellipsometric spectra: Application to InxGa1−xAs alloys

We investigate the effect of incomplete removal of semiconductor overlayers on critical-point (CP) parameters determined from the analysis of ellipsometric spectra. An approximate analytic expression shows that CP energies and broadening parameters should be relatively unaffected for isolated CPs if the dielectric response of the overlayer varies slowly with energy. The results are confirmed by model calculations for InAs, which show that the energies of the E1 and E1+Δ1 CP structures that are commonly used for compositional analysis of semiconductor alloys are relatively unaffected. We also analyze overlayer-removal data for a series of InxGa1−xAs alloy samples. Consistent with the above, the amplitudes and phases are affected significantly for all CPs, while the energies of the well-separated E1 and E1+Δ1 transitions are relatively invariant. The results show that accurate values of composition can be obtained from the analysis of the E1 and E1+Δ1 CP structures, even if complete removal of overlayers is...

Spectroscopic ellipsometric analysis of interfaces: Comparison of alloy and effective-medium-approximation approaches to a CdMgTe multilayer system

We discuss the accuracy and detectability of interface layers in the analysis of ellipsometric spectra in the CdxMg1−xTe system. Using parametric-alloy and effective-medium-approximation (EMA) representations to simulate interfaces in a single-quantum-well structure, we show that EMA analysis overestimates thicknesses of alloy interfaces by more than a factor of 3. While detailed results will clearly depend on the nature of the epitaxial materials involved, the results suggest that analyses of interfaces by the EMA should be done with caution.

Nondestructive analysis of coated periodic nanostructures from optical data

Optical data are essential for the accurate nondestructive determination of profiles of periodic structures in integrated-circuit technology. In rigorous coupled-wave analysis, the sample is generally modeled as layers consisting of a single material and the ambient. We extend present capabilities to the analysis of structures with overlayers and demonstrate our approach by determining quantitatively the thicknesses of top, sidewall, and bottom oxides of deliberately and naturally oxidized structures.