GaAs Melt Research Articles

Bouyant and thermocapillary flow in liquid encapsulated floating zone

Buoyant and thermocapillary flow in axisymmetric liquid bridges of GaAs melt and its encapsulant B2O3 are studied numerically. The liquid bridges are heated either at the top or at the bottom (half zone). The results show that, when the liquid bridges are heated at the bottom, buoyant flow enforces and dominates over Marangoni flow. On the contrary, when the bridges are heated at the top, flow in the melt is reduced by the opposing buoyancy effect. The motion in the encapsulant B2O3 is essentially unaffected due to its high viscosity and remains interfacial tension driven flow. In the presence of gravity, the deformation of the melt-encapsulant interface is caused mainly by hydrostatic pressure at small temperature differences. However, with increased temperature difference, the capillary pressure effect can become so strong as to totally reverse the deformation of the interface when the bridges are heated at the top. The presence of encapsulant B2O3 significantly decreases the thermocapillary convection in GaAs melt and increases the capillary stability of the liquid bridges.

Excimer laser melting of GaAs: Real-time optical study

In situ real-time reflectivity measurements have been made in GaAs under ArF excimer laser (λ=193 nm) irradiation. The results obtained provide a reflectivity value for the solid material at the melting temperature of 0.44 and for the liquid of 0.63, both at 633 nm. The reflectivity values obtained for fluences just above the melting threshold (E=225 mJ/cm2) show that melting proceeds inhomogeneously, the near-surface region being formed by a mixture of solid and liquid phases. The comparison of these results to those obtained previously for irradiation of single-crystalline Si and Ge shows that inhomogeneous melting can be a general phenomenon at least in semiconductors. Higher fluences (E≳300 mJ/cm2) are necessary to induce an homogeneous melted layer on the surface of the irradiated material.

The finite-element analysis of slider-induced convection in two-phase lpe growth system

In this study, the finite-element method is used to simulate the steady-state slider-induced convection of the confined melt cavity. The flux distribution within the melt cavity is analyzed and the relative momentum transfers on the poly GaAs melt and GaAs-melt interface are calculated for several kinds of cavity geometries. It is found that the amount of momentum transfer is related to the skew angle of the cavity, and therefore, proper choice of the cavity angle is needed.

Ultrafast electronic disordering during femtosecond laser melting of GaAs.

We have observed an ultrarapid electronic phase transformation to a centrosymmetric electronic state during laser excitation of GaAs with intense femtosecond pulses. Reflection second-harmonic intensity from the upper 90 atomic layers vanishes within 100 fs; reflectivity rises within 0.5 ps to a steady value characteristic of a metallic molten phase, long before phonon emission can heat the lattice to the melting temperature.

Silicon incorporation anomaly in LEC grown GaAs

The growth of Si doped LEC GaAs shows a significant deviation in the Si incorporation from the prediction of a classic segregation model. As a rule, it is expected that for a given impurity segregation coefficient, the dopant incorporation throughout the crystal may be calculated with fairly good accuracy. In fact, Te-doped GaAs LEC doping profiles show a close approximation to this classic model. However, in Si doped GaAs LEC growth, the doping distribution in the crystals deviates significantly from the segregation model. In some cases, this deviation is of several orders of magnitude. The degree of deviation was found to be dependent on the growth conditions. This work was performed to understand the source of the deviation from the model and to allow for an accurate account of the Si incorporation. In this work, Si doped GaAs crystals were grown by low pressure LEC. These crystals were targeted for doping in the 1016 to low 1018 cm-3 range. The actual dopant incorporation was significantly lower than the segregation model predicted, and the axial doping variation was much too flat. Moreover, in spite of an intentional Si doping, a significant number of crystals were semi-insulating. Analysis of the experimental data has shown that this doping behavior cannot be explained by the assumption of an unknown compensating impurity. A different model has been developed that is consistent with all of our observations and that provides an accurate account of the Si incorporation. Understanding the Si incorporating anomaly has allowed us to make successful process changes and improvements. This model explains the observed doping anomaly by taking account of the diffusion of Si between the GaAs melt and the B2O3 encapsulant, as well as some permanent Si trapping in the B2O3. When the interaction of the Si with B2O3 is taken into account, the Si incorporation follows the segregation model and the anomaly vanishes.

Transport restriction effect for gaseous components on the carbon content of LEC GaAs

Gas convection above B 2O 3 was restricted using a quartz cap over the crucible to examine the gas transport effect on the carbon content of LEC GaAs crystals. It has been found that the carbon content of crystals grown with a cap closing the crucible is much larger than that of crystals grown without a cap. Also, a BN cylinder was introduced in a puller to reduce the interface area between GaAs melt and B 2O 3, and between B 2O 3 and the ambient gas. Estimates of carbon extraction rate from the melt were made by gas analysis in the puller which was operated in a hypothetical pulling process with different cylinder diameters. In addition, the dependence of the carbon concentration in polycrystalline GaAs remaining in the crucible on the pass time from the start of heating was investigated. The results suggest not only that the rate of carbon extraction is faster than incorporation, but also that the initial carbon content of the starting charge affects the carbon content of the grown crystal. It has been concluded that the extraction rate of carbon from the melt as well as the initial carbon content of the starting charge are significant factors in estimating the actual distribution of carbon concentration along the growth direction of LEC GaAs.

Growth of GaAs single crystals by the floating zone technique under microgravity

The experimental conception of the floating zone (FZ) technique for the growth of GaAs crystals with larger diameters under microgravity, including the use of a stoichiometry control by a controlled As-vapor source, is explained in this paper. Also first results of preparatory experiments are given, which are obtained from ballistic rocket flights in the frame of the German TEXUS program and from terrestrial crystal growth. In the second part of the paper we discuss the very important problem of the thermocapillary convection (Marangoni effect) occurring in a GaAs melt zone. The results of a three-dimensional time-dependent numerical simulation of the thermocapillary convection show a very good agreement to existing experimental results in the literature and of our group. It follows from these calculations for liquid GaAs and from the experimental results of the terrestrial FZ growth of GaAs single crystals (6 mm diameter), that the limit for the transition from steady to unsteady thermocapillary convection, i.e. the formation of doping striations, is relatively low. Therefore, means and methods to stabilize this type of convection, i. e. to increase the critical Marangoni number for the transition to unsteady convection were studied. It is especially discussed to what extent static magnetic fields can be used to solve this problem for the case of the FZ method for GaAs. A value of the induction B 0 = 0.08 T was calculated which is necessary to avoid striations in the GaAs crystals (20 mm diameter) to be grown during the German D2-mission by the FZ technique.

Laser melting of GaAs covered with thin metal layers

GaAs melting simulations with a pulsed ruby laser are reported. The presence of a thin metal layer deposited on the GaAs surface gives rise to a reduction in the melting threshold and to an increase of melted depths when compared with nude GaAs surfaces. Melting thresholds around 0.3 J/cm2 for nude GaAs surface and slightly below 0.25 J/cm2 for GaAs covered with a 120 A tin layer are predicted in reasonable agreement with experimental results.

Surface tension of GaAs melt

Surface tension of GaAs melt

Femtosecond laser-induced melting of GaAs probed by optical second-harmonic generation

Primary steps of melting (disordering) of GaAs surface layers after femtosecond pulse excitation have been measured through second-harmonic generation in reflection. A numerical fit of the experimental data yield a characteristic time delay between excitation and lattice melting of about 75 fs.

GaAs single crystal for 3 inch diameter wafers grown by horizontal zone melt technique

We have developed a new horizontal zone melter (HZM) for boat growth of 3 inch diameter GaAs wafers. A silicon carbide (SiC) ring heater was used in the GaAs melting zone to concentrate the heating power in the small zone, and to minimize the variation in the longitudinal carrier concentration in the crystal. In order to maintain a flat, or a slightly convex, solid-liquid interface in the direction of crystal growth, the axial temperature gradient of the GaAs melt was minimized and the melt convection currents were reduced only in a small part of the GaAs melt near the solid-liquid interface. A Si-doped GaAs single crystal some 10 kg in weight and 600 mm in length was grown with the HZM. The length of the melting zone was 200 mm long. The longitudinal carrier concentration was constant from the seed (g = 0.1) to the middle (g = 0.65). The average dislocation densities of the 3 inch diameter wafers were about 2000 cm -2 both from near the seed and the tail.

Pulsed laser annealing of GaAs implanted with Se and Si

Carrier activation and mobility were studied by Raman spectroscopy and the Hall effect in pulsed laser annealed samples of GaAs implanted with doses of Si and Se from 2.2x 1012 to 6.Ox 1014 cm2. The samples were annealed using a pulsed XeCl excimer laser (X = 308 nm) and a pulsed dye laser (A = 728 nm) with energy densities from 0.06 to 0.9 J/cm2 and pulse durations of about 10 ns. Very high carrier concentrations of 3x1019 and 1.5x 1019 cm3 were obtained for the best n-type GaAs samples annealed with the dye and excimer lasers, respectively. The dye laser consistently produced higher activation than excimer laser annealing. A transient reflectivity signal was used to identify the GaAs melt threshold and the melt phase dynamics of the GaAs under the nitride cap. The threshold energies for cap damage were 0.34 and 0.12 J/cm2 for the excimer and dye lasers, respectively. Raman spectroscopy was used to identify the threshold energies for the GaAs implant layer recrystallization and for optimum carrier activation. Four major peaks were observed on the photoinduced current transient spectra in the temperature range of 60 to 400 K, corresponding to traps with activation energies between 0.06 to 0.80 eV.

Gettering of donor impurities by V in GaAs and the growth of semi-insulating crystals

Vanadium added to the GaAs melt getters shallow donor impurities (Si and S) and decreases their concentration in the grown crystals. This gettering is driven by chemical reactions in the melt rather than in the solid. Employing V gettering, we were able to grow reproducibly semi-insulating GaAs by horizontal Bridgman and liquid-encapsulated Czochralski techniques, although V did not introduce any midgap energy levels. The compensation mechanism in these crystals was controlled by the balance between the native midgap donor EL2 and residual shallow acceptors. Vanadium gettering contributed to the reduction of the concentration of shallow donors below the concentration of acceptors. The present findings clarify the long-standing controversy on the role of V in achieving semi-insulating GaAs.

Defect-induced Schottky barrier height modification by pulsed laser melting of GaAs

A pulsed XeCl excimer laser (λ=308 nm) is used to melt n- and p-type GaAs samples. Melt-induced defects shift the surface Fermi level to a new pinning position at 0.58±0.04 eV below the conduction-band minimum for both n- and p-type samples. The Schottky barrier height of Au, deposited on the GaAs after laser irradiation, is increased by 0.38 eV (from 0.43 to 0.81 eV) for p-type, and decreased by 0.30 eV (from 0.84 to 0.54 eV) for n-type samples. In the post-melted GaAs near-surface region, four deep levels are found using deep level transient spectroscopy. The observation of minority-carrier traps in the Schottky diode structures suggests the existence of minority-carrier source. We speculate a compensated region forms near the GaAs surface. A bulk Fermi level stabilization model is used to explain the changes observed in the Schottky barrier heights.

Arsenic-rich melt effect on threshold voltage scattering for Si-implanted GaAs metal-semiconductor field-effect transistor

The microscopic electrical activation efficiency of implanted Si atoms in undoped, liquid-encapsulated Czochralski-grown GaAs substrates was investigated by means of the threshold voltage (Vth) of implanted metal-semiconductor field-effect transistors as a function of stoichiometry and impurities, especially of carbon as determined by Fourier transform infrared (FT-IR) spectroscopy (≤2.5×1016 cm−3). The Vth scattering, which is related to the electrical activation efficiency dispersion, was small for an As-rich substrate grown from an As-rich melt, whether or not there existed a horizontal 3400-Oe magnetic field. The smallest amplitude of the Vth scattering occurred for the As-rich melt and did not vary with magnetic field. The Vth scattering with magnetic field was independent of carbon concentration (≤2.5×1016 cm−3). However, without magnetic field, an effect clearly appeared at low carbon concentration (nearly nondetectable). It seems that the Vth scattering or the electrical activation efficiency dispersion is dominated by the dispersion of stoichiometry, that is, composition, of the GaAs substrate because the electrical activation efficiency depends on the ratio of the Ga vacancy concentration and As vacancy concentration and the magnitude of their concentrations at the annealing temperature for the Si-implanted GaAs substrate. Therefore small Vth scattering means that the dispersion of the composition (Ga and As) is small. The composition dispersion for the most As-rich melt is small because the solidification is controlled by the straight As-rich side solidus line of the Ga:As composition-temperature phase diagram, and thus temperature variations of the GaAs melt in the pyrolytic boron nitride crucible are not as important. Also, there was no correlation between Vth scattering and dislocation density.

Growth of semi-insulating GaAs crystals with low carbon concentration using pyrolytic boron nitride coated graphite

It has been determined that the incorporation of carbon into GaAs crystals grown by the liquid encapsulated Czochralski method occurs as the result of a reaction between the GaAs melt and gaseous oxides of carbon, rather than carbon particles. The incorporation occurs during growth as well as during synthesis. The incorporation of carbon was particularly large when a commercially available large puller was used because of the greater number of carbon components in the hot zone which act as sources. Crystals with a low carbon concentration of 1×1015 cm−3 have been successfully grown by employing pyrolytic boron nitride coated graphite components in the large puller.

Nonlinear optical studies and CO2 laser-induced melting of Zn-doped GaAs

The absorption of CO2 laser radiation in p-type GaAs is dominated by direct free-hole transitions between states in the heavy- and light-hole bands. For laser intensities on the order of 10 MW/cm2, we report that the absorption associated with these transitions in moderately Zn-doped GaAs (∼1017 cm−3) begins to saturate in a manner predicted by an inhomogeneously broadened two-level model. At higher laser intensities surface melting occurs initially at localized sites in moderately doped material and more uniformly in heavily Zn-doped samples (≳1018 cm−3). As the energy density of the CO2 laser radiation is progressively increased further, the surface topography of the samples shows signs of ripple patterns, high local stress, vaporization of material, and exfoliation of solid GaAs fragments. Electron-induced x-ray emission data taken on the laser-melted samples show that there is a loss of As, compared to Ga, from the surface during the high-temperature cycling. By irradiating the samples in air, argon, and vacuum, we find that the vaporization rates are directly influenced by the ambient environment, particularly by the interaction of oxygen with the molten GaAs. Secondary ion mass spectrometry measurements are used to study the diffusion of oxygen from the native oxide and the incorporation of oxygen in the near-surface region of the GaAs samples that have been melted by a CO2 laser pulse. We find that oxygen incorporation does occur, and that the amount and depth of the oxygen incorporation depends on the laser energy density, number of laser shots, and ambient environment. For samples that are irradiated in argon or vacuum, we find that removal of the native oxide can be accomplished with CO2 laser pulses. Similar measurements are performed on Si-implanted GaAs, and results are reported for the redistribution of the implanted silicon atoms, the deviations from stoichiometry, and the incorporation of oxygen in the resolidified layer.

Pulsed Co2 Laser Induced Melting and Nonlinear Optical Studies of GaAs

ABSTRACTWe report new optical and structural properties of p-type GaAs that result from the absorption of high-intensity 10.6 μm radiation. Prior to the onset of surface melting, we find that the absorption coefficient decreases with increasing intensity in a manner predicted by an inhomogeneously broadened two-level model. As the energy density of the CO2 laser radiation is increased further, the surface topography shows signs of melting, formation of ripple patterns, and vaporization. Auger spectroscopy and electron-induced x-ray emission show that there is loss of As, compared to Ga, caused by the melting of the surface. Using plain-view TEM we find that Ga-rich islands are formed near the surface during the rapid solidification of the molten layer. Auger and SIMS measurements are used to study the incorporation of oxygen in the near-surface region, and the results show that oxygen incorporation can occur for GaAs samples that have been irradiated in air.

Secondary ion mass spectrometry analysis of impurity distributions across liquid-encapsulated Czochralski GaAs wafers

The impurity distributions in liquid-encapsulated Czochralski-grown GaAs were studied using secondary ion mass spectrometry along the diameter of circular wafers. Si and Al concentrations gradually decrease from the center to the edge of the wafer, while B has an inverse profile with respect to Si and Al. On the other hand, Cr concentration is almost constant across the wafer. To explain the impurity profiles across the wafer, we propose a model which shows that the nonuniform impurity distributions in the GaAs melt near the solid-liquid interface during crystal growth is due to the reducing reaction of the encapsulant B2O3.

The Case For Inelastic Reflection Of Photons In Laser Diodes

The wavelength of stimulated emission at 8175.25 A of GaAlAs broad-area, single-heterojunction laser diodes of the close-confinement type was experimentally found to increase as a function of emission angle. The shift was 0.42± 0.08 A/deg and approximately 0.01 A/deg in the lateral and transverse directions, respectively. The increase of wavelength is postulated to be due to momentum loss of the photons by inelastic reflections at the Fabry-Perot surfaces. An optical model was developed by which the changes in the following properties per photon per reflection were calculated for the optical cavity-air interface (no = 3.60, na = 1.00): wavelength, AX = 3.03 x 10-4 A; momentum, Ap = 3.00 x 10-30 erg.s/cm; momentum loss coefficient, n = 3.70 x 10-8; energy, DE = 2.50 x 10-20 erg; and increased angle of reflection, a = 2.04 x 10-4 deg. By determining the number of reflections per beam path in the cavity, the full width at half power for the lateral beam divergence was calculated to be 20.6°. The pressure and energy per pulse at the cavity surface of the diode were calculated from the momentum loss concept and are in reasonable agreement with the Knoop hardness and heat of fusion for the fracturing and melting of GaAs.