Vapor Phase Epitaxy Of GaAs Research Articles

Heavy carbon doping in low-pressure metalorganic vapor phase epitaxy of GaAs using trimethylarsenic — a comparison between the carrier gases N 2 and H 2

The carbon doping behavior of trimethylarsenic (TMAs) in low-pressure metalorganic vapor phase epitaxy (LP-MOVPE) of GaAs was studied in the temperature range of 550 to 700°C and at various trimethylarsenic/trimethylgallium (TMAs/TMGa) ratios using the carrier gas N 2 in comparison to the standard H 2. At a growth temperature as high as 600°C, perfectly specular layers with a hole concentration equaling the carbon concentration of 1.7 × 10 20 cm -3 and a mobility of 55 cm 2/V·s were achieved for the carrier gas nitrogen. This is the highest fully active carbon concentration ever reported for MOVPE material, to our knowledge, and more than double the hole concentration obtained for the carrier hydrogen under the same growth conditions. This result indicates that the carrier N 2 is better suited for heavy carbon doping with TMAs at such a high growth temperature as 600°C than the carrier H 2, and is therefore more compatible to growth requirements for (AlGa)As/GaAs heterostructures than H 2. A method is introduced with which the acceptors are activated without post-growth anneal.

Metalorganic vapor phase epitaxial growth of GaAs on ZnSe; on the flow sequence of source precursors at the interface

Metalorganic vapor phase epitaxy of GaAs onto ZnSe was studied. To optimize the flow sequence of Ga and As source precursors at the beginning of GaAs growth, they were separately supplied to the ZnSe surfaces at 450–550°C. The Ga precursor dissociates into Ga metal, which deposits three-dimensionally on ZnSe, while the As precursor drastically etches the ZnSe layers. The flow sequence was determined according to these results as not to expose the ZnSe surface to the Ga or As precursors alone for a long time. The growth rate of GaAs on ZnSe is slower than that of a thick GaAs layer at the very initial stage and recovers gradually as the growth continues.

Behavior and mechanism of wide terrace formation during metalorganic vapor phase epitaxy of GaAs and related materials

We have investigated the behavior and mechanism of surface terrace formation during the annealing of GaAs in the presence of column-V gases and during the growth of GaAs, Al 0.35Ga 0.65As and AlAs by metalorganic vapor phase epitaxy (MOVPE) using atomic force microscopy. The ordered terrace arrays with monolayer-high steps are formed during annealing in arsine prior to growth. The width of the terraces formed during growth depends on the substrate misorientation angle. The critical terrace widths for two-dimensional nucleation for GaAs and AlGaAs are found to be similar, but those for AlAs are narrower. The widths for GaAs and AlGaAs grown by MOVPE are about one or two orders of magnitude larger than the widths for GaAs and Al 0.3Ga 0.7As grown by molecular beam epitaxy. However, a wide GaAs terrace is not formed when trimethylarsine (TMAs) is used as a column-V source. To explain wide terrace formation in MOVPE, we propose a hydrogen-radical terminated surface model and obtain results supporting this model. Hydrogen radicals generated from the decomposition of arsine enhance the surface of column-III species on H-terminated As atoms.

Electrical characterization of semi-insulating metalorganic vapor phase epitaxy GaAs grown by controlled oxygen incorporation

Semi-insulating GaAs layers were produced by low pressure metalorganic vapor phase epitaxy (MOVPE) through the intentional incorporation of oxygen. A commercially available oxygen precursor, diethyl aluminum ethoxide (DEALO, (C 2H 5) 2AlOC 2H 5), was used as the oxygen source. Bulk resistivity of the oxygen-doped GaAs (GaAs:O) was determined by current-voltage ( I-V) measurement on n-i-n mesa structure. Resistivities greater than 6 X 10 9 Ω cm at 297.5 K have been achieved. Constant capacitance values obtained from steady state capacitance-voltage ( C-V) measurements on the n-i-n structure indicate a very low net free carrier concentration in GaAs:O. Activation energy deduced from temperature-dependent resistivity measurement was found to be concentration-dependent. A second-order power law dependence of resistivity on DEALO mole fraction was observed. These observations coincide with the highly non-linear oxygen incorporation behavior using DEALO as the oxygen precursor.

Wide terrace formation during metalorganic vapor phase epitaxy of GaAs, AlAs, and AlGaAs

GaAs, Al0.35Ga0.65As, and AlAs surface terrace structures formed during metalorganic vapor phase epitaxy (MOVPE) on vicinal (001) surfaces are investigated. The minimal terrace width for two-dimensional (2-D) nucleation for AlGaAs is almost equal to that for GaAs. In contrast, for AlAs it is reduced and the growth mode becomes three dimensional at temperatures below 580 °C. Monolayer step-flow growth without 2-D islands and terraces about 1 μm wide are obtained above 630 °C for all materials. These results are very different from those by molecular beam epitaxy and suggest that MOVPE is superior to grow wide uniform heterointerfaces for GaAs/AlAs and GaAs/AlGaAs.

Mg-doping transients during metalorgic vapor phase epitaxy of GaAs and AlGaInP

We studied magnesium-doping transients during metalorgic vapor phase epitaxy of GaAs and (Al x Ga 1- x ) 0.5 In 0.5P (0 ⪯ x ⪯ 0.7). We examined the transient of Mg concentration depth profile through epitaxial layers when Mg precursors are initially injected into the reactor (doping delay). We found that increasing the Al composition of epitaxial layers, i.e., increasing the mole fraction of Al precursors in the reactor, significantly reduces the Mg-doping delay. We obtained this result for both trimethylaluminum (TMAl) and triethylaluminum (TEAl). We quantitatively modeled this phenomenon based on the competitive adsorption of Mg and Al precursors on the internal surface of the reactor. Our model also explains that the Mg concentration in epitaxial layers increases either linearly or superlinearly with the Mg precursor input, depending on the length of the doping delay.

Dominant Traps in Liquid Phase Epitaxial GaAs Studied by Controlled Doping with Indium and Antimony

AbstractPhotocapacitance measurements in the photon energy range of 0.64 to 1.27 eV are done on GaAs layers grown by liquid phase epitaxy (LPE) and doped with varying amounts of indium (In) and antimony (Sb). Two hole traps with optical ionization energies of 0.65 and 0.79 eV are revealed in the material which are assumed to be two levels associated with the hole trap B reported in the literature. In addition, the hole trap A, usually found in LPE GaAs is also detected here by optical deep level transient spectroscopy (ODLTS) technique. The results of trap density measurements, as functions of isoelectronic doping concentration, indicate that VGa is an active constituent of the hole traps in LPE GaAs. The concentration of the 0.79 eV trap is found to be reduced in the material with high concentration of Sb. The analysis of photocapitance data using the existing theoretical models indicates the presence of an electron trap with activation energy of 0.75 eV in the heavily Sb‐doped material. This trap may be identified with the second charge state of the SbGa electron trap, obtained previously in bulk and metalorganic vapor phase epitaxial GaAs doped with Sb. The sharp reduction of the 0.79 eV hole trap density in Sb‐doped materials is explained by assuming the complex GaAsV, as its source.

Multiple deep levels in metalorganic vapor phase epitaxy GaAs grown by controlled oxygen incorporation

Oxygen intentionally introduced into GaAs, grown through the metalorganic vapor phase epitaxy process, is shown to introduce a controlled series of deep levels. A novel, commercially available, oxygen precursor, (C2H5)2AlOC2H5, was used as the oxygen source. Capacitance measurements have revealed compensation of both donors and acceptors in these materials when using this source. The determination of the deep level structure of these films has been carried out through the use of deep level transient spectroscopy (DLTS). The DLTS investigation, carried out on GaAs p+-n homojunctions, indicates that unlike the case of bulk oxygen-doped GaAs (GaAs:O), several deep levels are introduced and directly associated with the intentional oxygen introduction. Two principal traps are found to be located at 0.95 and 0.75 eV below the conduction-band edge, with several other minor traps being observed. These intentionally introduced deep levels have been used to form highly resistive GaAs, providing an analog to the currently employed low-temperature semi-insulating nonstoichiometric GaAs grown by molecular beam epitaxy.

Alkoxide precursors for controlled oxygen incorporation during metalorganic vapor phase epitaxy GaAs and AlxGa1−xAs growth

Controlled oxygen incorporation in GaAs using Al-0 bonding based precursors, dimethyl aluminum methoxide (DMALO) and diethyl aluminum ethoxide (DEALO), is presented in this investigation. A comparison study of oxygen incorporation kinetics between nominally undoped AlxGa1−xAs using trimethyl aluminum and DMALO-doped GaAs suggests that DMALO is one of the most important oxygen-bearing agents responsible for unintentional oxygen incorporation in AlxGa1−xAs. Controlled oxygen doping using DEALO is reported for the first time. Oxygen incorporation behavior, especially on the effect of the V/III ratio, was found to be quite different from the case of DMALO, mainly due to the differences between methyl- and ethyl-based growth chemistries. Physical, electrical, and optical properties of these oxygen-doped GaAs are also reported.

Annealing behavior of deep-level defects in 1 MeV electron irradiated GaAs

Isochronal annealing behavior of deep-level defects in 1 MeV electron irradiated vapor phase epitaxy GaAs was studied through deep level transient spectroscopy measurements. As the annealing temperature was increased, the activation energy of the EL2-A trap remained at Ec−0.823 eV, whereas the EL2-B trap (Ec−0.843 eV) transformed into a new trap, ELN-1 (Ec−0.870 eV), and finally into another new trap, ELN-2 (Ec−0.891 eV) before returning to the single EL2-A level at a 270 °C annealing temperature. The EL6 trap (Ec−0.335 eV) varied similarly, transforming into Ec−0.357 eV (P1) before staying constant at Ec−0.396 eV (ELN-3) after a 270 °C annealing temperature. The capture cross sections of EL2-B and EL6 increased by an order of magnitude during the annealing. These results suggest that the EL2-B trap which was split from the EL2 center by the 1 MeV electrons could be related to EL6. Traps E1 and E2 remain at Ec−0.032 eV and Ec−0.129 eV, respectively, throughout the annealing, before annihilation at a 290 °C annealing temperature. This result indicates that the atomic structures of E1 and E2 defects could be related. The EL3 (Ec−0.420 eV) is transformed into a new trap, ELN-4 (Ec−0.456 eV), and then into Ec−0.50 eV (P2) during annealing. The free carrier concentration of the irradiated sample increases rapidly with annealing temperature and returns back to the starting free carrier concentration due to the rapid annealing rates of the electron induced defects.

X-ray-scattering analysis of surface structures produced by vapor-phase epitaxy of GaAs.

We have used in situ x-ray-scattering techniques to analyze surface reconstructions and ordered steps at various stages in the growth of GaAs by organometallic vapor-phase epitaxy. In contrast to an earlier report of a c(4\ifmmode\times\else\texttimes\fi{}4) reconstruction, our current measurements reveal a related structure with 1\ifmmode\times\else\texttimes\fi{}2 symmetry. The 1\ifmmode\times\else\texttimes\fi{}2 structure is a disordered form of the c(4\ifmmode\times\else\texttimes\fi{}4) reconstruction in which long-range correlations between dimer rows are preserved but dimer vacancies are located at random. Aside from reconstructions, ordered step structures were studied on samples with both high (2.2\ifmmode^\circ\else\textdegree\fi{}) and low (0.07\ifmmode^\circ\else\textdegree\fi{}) levels of miscut. Using Gaussian step-spacing distributions with widths of order the mean terrace size, a numerical model is used to reproduce measured truncation rod profiles in scans perpendicular to steps. We also use the model to follow the narrowing of the step-spacing distribution which occurs during growth. After the growth of several hundred angstroms of GaAs, truncation-rod profiles sharpen, accompanied by a shift in peak positions. The shift is due to an asymmetric weighting of the scattering from terraces of different widths: wide terraces make a greater contribution to the structure factor. For samples with a broadened distribution of step spacings, this effect results in a smaller peak splitting than that which would occur for an ideally miscut sample.

Effect of Sb doping on point defect ensemble in MOVPE - GaAs

Metalorganic vapor phase epitaxial GaAs doped with Sb in concentration limits 4×10 17 – 5×10 18cm −9, corresponding to low defect density and high electrical parameters of the layers has been investigated. An effect of the isoelectronic dopant on GaAs point defect ensemble has been revealed. From low temperature photoluminescence studies it has been shown that Sb doping leads to reduction of the emission bands attributed to C and Si and to (V Ga-D) complexes.

Application of deep-level transient spectroscopy for monitoring point defects in III–V semiconductors

Deep-level transient spectroscopy has been employed for monitoring grown-in point defects in vapour phase epitaxial GaAs 0.6P 0.4:Te and GaP:N,S. It is shown that the concentrations of deep electron traps T1 (0.20 eV) and T2 (0.18 eV) in GaAs 0.6P 0.4:Te are dependent on the shallow donor concentration and these traps are tentatively identified as Te AsGa AsTe As and O iTe As complexes respectively. The concentrations of 0.24 eV and 0.44 eV electron traps detected in GaP:N,S are dependent on the growth temperature and these traps are attributed to P vacancies and the ((N P) 2V Ga) − complex respectively.

Characterization of deep-level defects in GaAs irradiated by 1 MeV electrons

Deep level transient spectroscopy has been employed to determine the defect energy levels, capture cross sections, and trap densities in Si-doped vapor phase epitaxy GaAs both before and after irradiation by 1 MeV electrons at room temperature for electron fluence ranging from 1.1×1014 to 5.0×1015 e cm−2. The results indicate that the irradiated samples have an electron trap at Ec-0.334 eV(EL6) in addition to the two electron traps at Ec-0.815 eV(EL2) and Ec-0.420 eV(EL3) which are present in the nonirradiated sample. The density of the EL6 trap increases monotonically with irradiation fluence from 6.7×1013 to 24.4×1013 cm−3 as electron fluence is increased from 1.1×1014 to 3.1×1014 e cm−2. In contrast, both the EL2 and EL3 trap densities were found to be only moderately affected by electron irradiation with trap densities slightly greater than the nonirradiated sample. These results, along with the fact that the EL6 trap was not observed in the nonirradiated sample, strongly suggest that this trap is created by the electron irradiation. In addition to creating the EL6 trap, electron irradiation results in a nonexponential transient for the EL2 deep level at Ec-0.815 eV which can be resolved into the sum of two exponential transients arising from two closely spaced deep levels at Ec-0.815 eV and Ec-0.843 eV.

Parameter evaluation in automated digital deep-level transient spectroscopy (DLTS)

A method of directly evaluating the activation energy Delta E, capture cross section sigma , and density N/sub T/, of deep-level traps from the pulsed reverse bias capacitance transient is described. The main advantages of this technique are that it requires only a single temperature scan, and it can resolve nonexponential transients due to closely-spaced energy levels. The test samples used for this paper consisted of Schottky diodes fabricated on nonirradiated and 1-MeV electron-irradiated n-type VPE (vapor-phase epitaxy) GaAs wafers. The well known EL2 trap was identified with Delta E of 0.81 eV, and sigma /sub n/ of 1.0*10/sup -13/ cm/sup 2/ for the nonirradiated sample. These values were found to be in good agreement with published data using established, conventional DLTS techniques. For the irradiated samples a nonexponential capacitance transient was found in the EL2 range of temperatures. The discussed technique was able to resolve two closely spaced deep levels lying at E/sub c/-0.81 eV and E/sub c/-0.84 eV, and with capture cross sections of 1.5*10/sup -13/ cm/sup 2/ and 2.5*10/sup -12/ cm/sup 2/, respectively.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A semi-empirical model for the complete orientation dependence of the growth rate for vapor phase epitaxy: chloride VPE of GaAs

A semi-empirical model has been developed to determine the complete crystallographic orientation dependence of the growth rate for vapor phase epitaxy (VPE). Previous researchers have been able to determine this dependence for a limited range of orientations; however, our model yields relative growth rate information for any orientation. This model for diamond and zincblende structure materials is based on experimental growth rate data, gas phase diffusion, and surface reactions. Data for GaAs chloride VPE is used to illustrate the model. The resulting growth rate polar diagrams are used in conjunction with Wulff constructions to simulate epitaxial layer shapes as grown on patterned substrates. In general, this model can be applied to a variety of materials and vapor phase epitaxy systems.

Vapor etching of GaAs and AlGaAs by CH3I

With the objective of developing an improved process for in situ etching of GaAs-based materials in organometallic vapor phase epitaxy reactors, GaAs wafers and AlxGa1−xAs epilayers have been etched with CH3I vapor in a horizontal reactor operated at atmospheric pressure with H2 or He carrier gas. For a H2 flow rate of 2.1 s lpm, etching temperatures from 400 to 625 °C, and CH3I mol fractions (yCH3I)from 0.0012 to 0.015, the measured GaAs etch rate r (in Å min−1) is given by r = k0 y0.83CH3I exp[ − 45(kcal mol−1)/RT] with k0=3.2×1016 Å min−1. The value of k0 depends on the type of carrier gas, flow rate, total pressure, and reactor geometry. The etch rate appears to be controlled mainly by the decomposition of CH3I to CH3 and I, for which the activation energy has been reported to be 43.5 kcal mol−1. The etch rate of AlxGa1−xAs epilayers with x up to 0.7, which was measured at 480 °C with yCH3I= 0.015, does not depend on Al content. The surface morphology of etched GaAs wafers improves with decreasing temperature. Specular surfaces are maintained at temperatures below 500 °C for etch depths up to 5000 Å.

Difference of reaction between AsH 3 and arsenic vapor from As metal in vapor phase epitaxy of GaAs

GaAs was grown by mixing either AsH 3 or arsenic vapor (As 4 or As 2) from As metal with GaCl 3 supplied by He or H 2 gas. When the carrier gas of GaCl 3 was He, GaAs was grown when AsH 3 gas was used. Gas etching of the substrate occurred if As metal was used as the As source. When the carrier gas of GaCl 3 was H 2, however, GaAs was grown by using either AsH 3 or As metal, but dependence of the growth rate on the substrate position was different. These results indicate that AsH 3 is not completely decomposed into As 4 or As 2 but directly reacts with GaCl 3 and GaCl, while the arsenic vapor (As 4 or As 2) can react only with GaCl but not with GaCl 3.

ChemInform Abstract: Characterization of Residual Acceptors in Semi‐Insulating and VPE GaAs by Selective Pair Luminescence (SPL) and Photoluminescence (PL) Methods.

AbstractThe results of detailed PL and SPL measurements as a function of pump laser energy, excitation laser intensity, and sample temp. on various less‐than‐perfect, semi‐insulating GaAs substrates, as well as good quality VPE (vapor‐phase epitaxial) layers of GaAs show that in lesser quality substrates, the SPL technique is significantly superior to the PL method in identifying acceptor impurities.

Characterization of Residual Acceptors in Semi‐Insulating and VPE GaAs by Selective Pair Luminescence and Photoluminescence Methods

Both photoluminescence (PL) and selective pair luminescence (SPL) measurements have been made on various substrates and vapor‐phase epitaxial (VPE) layers of as a function of pump laser energy, excitation laser intensity, and sample temperature. Comparison of the results shows that in lesser quality substrates, the SPL technique is significantly superior to the PL method in that it can provide an unambiguous impurity identification far better than that available from the PL method. However, since the structure of SPL spectra strongly depends upon the laser excitation energy, the SPL signals should be optimized with a proper pump laser energy for particular acceptors. Various residual acceptors in semi‐insulating and VPE substrates such as C, Mg, Zn, Si,and Ge have been identified through the SPL method, and energy values of their excited states are reported here.