G. B. Stringfellow

Solid phase immiscibility in GaInN

The large difference in interatomic spacing between GaN and InN is found to give rise to a solid phase miscibility gap. The temperature dependence of the binodal and spinodal lines in the Ga1−xInxN system was calculated using a modified valence‐force‐field model where the lattice is allowed to relax beyond the first nearest neighbor. The strain energy is found to decrease until approximately the sixth nearest neighbor, but this approximation is suitable only in the dilute limit. Assuming a symmetric, regular‐solutionlike composition dependence of the enthalpy of mixing yields an interaction parameter of 5.98 kcal/mole. At a typical growth temperature of 800 °C, the solubility of In in GaN is calculated to be less than 6%. The miscibility gap is expected to represent a significant problem for the epitaxial growth of these alloys.

Calculation of ternary and quaternary III–V phase diagrams

The models used to calculate binary and ternary III–V phase diagrams are reviewed. The simple solution model with Ω=a−bT is found to satisfactorily describe the liquid phase and the solid phase is adequately described using the regular solution model. The DLP (delta lattice parameter) model of the solid, based on the concept that the bonding in alloys is determined largely by the energy of the electrons, is found to be capable of predicting accurately the magnitudes of the solid interaction parameters knowing only the lattice parameters of the pure III–V compounds. The simple solution model with Ω=a−bT for the liquid and DLP model for the solid is then used in the calculation of the quaternary III–V phase diagrams. One outcome of the calculation is the establishment of a condition which can be used to check the consistency of the parameters Δ H F , T F and Ω 1 of the four solid III–V compounds bounding the solid phase field, against the solution model. The agreement of the calculated quaternary results with experiment in the Ga x In 1− x As y P 1− y system is satisfactory. In the Al x Ga 1− x P y As 1− y system, the consistency condition is not satisfied, and thus the calculation cannot agree with previous ternary calculations for the (Al,Ga)As, (Al,Ga)P and Ga(As,P) systems simultaneously. Parameters were chosen so that good agreement is established with all except the (Al,Ga)P ternary.

Miscibility gaps in quaternary III/V alloys

Abstract Thermodynamic concepts are developed for the calculation of miscibility gaps in III/V quaternary alloy systems of the type AxB1-xCyD1-y. Equations are developed using both the DLP model based on the dependence of free energy on lattice parameter and the regular solution model. An analytical expression is obtained for the critical temperature above which the solid is totally miscible. Solid-solid isotherms and tie lines are plotted for the system GaxIn1-xAsyP1-y for comparison with experimental data. Similar results are presented for the system AlxGa1-xAsySb1-y and GaxIn1-xAsySb1-y. Critical temperatures are calculated for all III/V quaternary alloys of the type AxB1-xCyD1-y involving Al, Ga, and In combined with P, As, and Sb.

The importance of lattice mismatch in the growth of GaxIn1−xP epitaxial crystals

The importance of lattice‐parameter mismatch between the GaAs substrate and the GaxIn1−xP epitaxial layer on crystal growth and properties has been investigated. Three major effects were observed: (i) The crystal morphology and substrate/epitaxial‐layer interface were found to be nearly perfect for a0 ≈ a0 GaAs and very poor including melt inclusions at the interface for other epitaxial‐layer compositions. (ii) The dislocation density was found to depend on a0 ‐ a0 GaAs varying from 105 to > 108 cm−2. (iii) The excess energy due to lattice‐parameter mismatch was found to perturb the solid composition from the chemical‐equilibrium composition toward the composition which minimizes mismatch; i.e., epitaxial layers with x=0.51±0.01 are grown from Ga–In–P liquids with chemical‐equilibrium solidus compositions ranging from 0.46 to 0.62. The result is that under the growth conditions used, highly perfect epitaxial layers could be grown only near x=0.51 with band gaps near 1.9 eV. Other compositions with EG>1.9 ...

Effect of mismatch strain on band gap in III‐V semiconductors

Interfacial elastic strain induced by the lattice parameter mismatch between epilayer and substrate results in significant energy–band‐gap shifts for III‐V alloys. The epilayers used in this study are GaxIn1−xAs on (100) InP and GaxIn1−xP on (100) GaAs prepared by organometallic vapor phase epitaxy. For layer thicknesses between 1 and 1.5 μm, and Δas.f./a0≤3.5×10−3 the misfit strain is assumed to be accommodated elastically. The energy–band‐gap shifts are determined by comparing the photoluminescence peak energies of the epilayers with the best experimental relation of band gap versus composition for unstrained layers. A calculation of the energy–band‐gap shift due to biaxial stress made for GaxIn1−xAs is found to agree with the photoluminescence measurements. In addition, a comparison of the energy–band‐gap shift for GaxIn1−xP shows a clearly different dependency for tensile and compressive strain, in good agreement with calculated results.

Photoluminescence of InSb, InAs, and InAsSb grown by organometallic vapor phase epitaxy

Infrared photoluminescence (PL) from InSb, InAs, and InAs1−xSbx (x<0.3) epitaxial layers grown by atmospheric pressure organometallic vapor phase epitaxy has been investigated for the first time over an extended temperature range. The values of full width at half maximum of the PL peaks show that the epitaxial layer quality is comparable to that grown by molecular‐beam epitaxy. The observed small peak shift with temperature for most InAs1−xSbx epilayers may be explained by wave‐vector‐nonconserving transitions involved in the PL emission. For comparison, PL spectra from InSb/InSb and InAs/InAs show that the wave‐vector‐conserving mechanism is responsible for the PL emission. The temperature dependence of the energy band gaps, Eg, in InSb and InAs is shown to follow Varshni’s equation Eg(T)=Eg0−αT2/ (T+β). The empirical constants are calculated to be Eg0=235 meV, α=0.270 meV/K, and β=106 K for InSb and Eg0=415 meV, α=0.276 meV/K, and β=83 K for InAs.

Ordered structures in GaAs0.5Sb0.5 alloys grown by organometallic vapor phase epitaxy

Electron diffraction measurements on (100) GaAs1−xSbx layers with x≊0.5 grown by organometallic vapor phase epitaxy indicate that ordered phases are formed during growth. Two ordered phases are observed. The simple, tetragonal AuCu‐I type phase consists of alternating {100} oriented GaAs and GaSb layers. Only the two variants with the tetragonal c axes perpendicular to the growth direction are observed. At least two variants are observed for the chalcopyrite E11 structure with alternating {210} oriented GaAs and GaSb layers.

A critical appraisal of growth mechanisms in MOVPE

Abstract The metalorganic vapor phase epitaxial (MOVPE) growth process will be considered by first examining, individually, each of the four fundamental aspects of the overall reaction: (1) thermodynamics, (2) homogeneous gas phase reactions, (3) mass transport, and (4) surface kinetic processes. These will first be examined in general, and then an attempt will be made to understand the MOVPE process in terms of these fundamental concepts using data for specific III/V systems as examples. Thermodynamics yields information concerning the basic driving force for both the crystal growth process itself and undesirable parasitic reactions occurring homogeneously in the vapor phase and on the reactor walls upstream from the substrate. Thermodynamic factors often control the solid composition of III/V alloys. The primary homogeneous gas phase reactions, pyrolysis of the reactants and adduct formation, possibly accompanied by elimination reactions which lead to polymer formation on the walls, will be examined. The vast difference in behavior between systems using trimethyl-versus triethyl-group III sources will be explored for the growth of Ga and In containing III/V compounds and alloys. In addition, the pyrolysis of both MO source molecules and the group V hydrides will be examined, especially relative to the determination of solid composition. For many systems, at ordinary growth temperatures, the MOVPE growth rate is controlled by diffusion through the gas phase. In this case, for V/III ⪢ 1, the temperature dependence of growth rate and the group III distribution coefficients for mixing on the III sublattice (AxB17minus;xC) are particularly simple; thermodynamics controls solid composition, i.e., at the interface, the solid is nearly in equilibrium with the vapor. The results obtained for the alloy system GaAsSb with V/III ⪡ 1 will be explored in detail. In this regime, under certain conditions, the MOVPE growth process is found capable of producing metastable alloys. Surface kinetics may control the MOVPE process at relatively low growth temperatures. The major factors observed to date seem to relate to AsH3 and PH3 pyrolysis at the growing interface. The detailed surface reaction mechanisms are nearly totally unknown.

Calculation of ternary phase diagrams of III–V systems

Abstract A method has been developed for the calculation of III–V ternary phase diagrams. The parameters needed for the calculation are the temperatures and entropies of fusion of the pure III–V compounds and the electronegativities, energies of sublimation and molar volumes of the three constituent elements. The thermodynamic properties of the liquid are treated using the regular solution model and the thermodynamic properties of the solid are calculated from the Van Vechten and Phillips spectroscopic theory of chemical bonding.

Reaction mechanisms in the organometallic vapor phase epitaxial growth of GaAs

The decomposition mechanisms of AsH3, trimethylgallium (TMGa), and mixtures of the two have been studied in an atmospheric‐pressure flow system with the use of D2 to label the reaction products which are analyzed in a time‐of‐flight mass spectrometer. AsH3 decomposes entirely heterogeneously to give H2. TMGa decomposes by a series of gas‐phase steps, involving methyl radicals and D atoms to produce CH3D, CH4, C2H6, and HD. TMGa decomposition is accelerated by the presence of AsH3. When the two are mixed, as in the organometallic vapor phase epitaxial growth of GaAs, both compounds decompose in concert to produce only CH4. A likely model is that of a Lewis acid‐base adduct that forms and subsequently eliminates CH4.