Intentional ρ-type doping by carbon in metalorganic MBE of GaAs

We report on the intentional ρ-type doping of GaAs layers grown in an UHV system from molecular beams of arsine (AsH₃) and mixtures of frimethyl gallium (TMG) and friethyl gallium (TEG). The entire doping range between 10¹⁴ cm^-3 (growth from pure TEG) and 10²⁰ cm^-3 (growth from pure TMG) can be covered by using mixtures of TMG and TEG. As revealed by SIMS and photoluminescence (PL) carbon is the dominant acceptor in the layers. Comparison of the Hall mobility and of the PL spectra shows that the quality of our films equals that of the best LPE and MBE grown ρ-type GaAs layers.     

High purity GaAs grown by the hydride vpe process

Epitaxial growth of high purity undoped GaAs by the hydride VPE process with electrical properties comparable to the best AsC₃ VPE results have been achieved. Background carrier concentration n₇₇ = 2 × 10¹³/cc and mobility μ₇₇ = 201,000 cm²/Vsec was achieved by careful control of the growth process. The effects of HC1 partial pressure over the Ga source, arsine and secondary HC1 partial pressures were Investigated. Use of secondary HC1 was found to result in compensated layers and hence should be avoided for high purity layers. Aging of the liquid HC1 source in its stainless steel container over extended time period was also found to result in degradation of epilayer purity due to increased acceptor contamination. Low temperature photoluninescence and far infra-red photoconductivity measurements were used to determine the residual acceptors and donors respectively. The major acceptors found were carbon and zinc. The major donor was sulfur. Key works: GaAs, hydride VPE growth, high purity layers, impurity identification     

MO-CVD growth of InGaAs using Me3Ga,AsMe3, AsH3 and Me3In or Et3 in and analyses of adducts formed during the growth process

High quality InGaAs films have been grown using the adduct Me₃In.AsMe₃ to block the room temperature reaction between Me₃In and AsH₃ and by using a cover piece to prevent the preferential evaporation of phosphorus from the InP substrate during the warm up. Infrared spectroscopy shows that Me₃Ga + AsMe₃ and Me₃ Ga + AsH₃. form stable adducts at room temperature, Me₃In + AsMe₃, probably form a stable adduct, while Et₃In + AsMe₃ probaoly do not form a stable Lewis acid/base adduct. Poorer quality films are grown with Et₃In than with Me₃ In because the AsMe₃ is unable to prevent by adduct formation the room temperature reaction between Et₃In and AsH₃. Pyrolysis studies show that the individual alkyls are stable to^‒400×C and mixtures are stable to^‒350×C. The problems associated with the lower vapor pressure of the adducts, adduct dissociation at room temperature and the growth temperature, and using an adduct as the starting material are discussed.     

Growth of device quality GaAs by chemical beam epitaxy

The growth of high purity GaAS with excellent uniformity and very low defect density by chemical beam epitaxy using triethylgallium and arsine is described. The residual background impurity is mostly carbon. A mobility of 518 cm²/Vs with a hole density of 3.6 x 10¹⁴ cm⁻³ has been obtained for a growth temperature of 500° C. The electrical quality is further evaluated by fabricating a Si doped epilayer into MESFET device using 1 μm gate length. A transconductance of 177 mS/mm has been measured. The results indicate that chemical beam epitaxy is a very attractive growth technique for GaAs integrated circuits.     

Growth of (GaAs)1_x(Ge2)x by metalorganic chemical vapor deposition

We report deposition of (GaAs)_{1_x}(Ge₂)_x on GaAs substrates over the entire alloy range. Growth was performed by metalorganic chemical vapor deposition at temperatures of 675 to 750°C, at 50 and 760 Torr, using trimethylgallium, arsine, and germane at rates of 2–10 μ/h. Extrinsic doping was achieved using silane and dimethylzinc in hydrogen. Characterization methods include double-crystal x-ray rocking curve analysis, Auger electron spectroscopy, 5K photoluminescence, optical transmission spectra, Hall-effect, and Polaron profiling. Results achieved include an x-ray rocking curve full-width at half maximum as narrow as 12 arc-s, Auger compositions spanning the alloy range from x = 0.03 to x = 0.94, specular surface morphologies, and 5K photoluminescence to wavelengths as long as 1620 nm. Undoped films are n type, with n ≈ 1 × 10¹⁷ cm⁻³. Extrinsic doping with silane and dimethylzinc have resulted in films which are n type (10¹⁷ to 10¹⁸ cnr⁻³) or p type (5 × 10¹⁸ to 1 × 10²⁰ cm⁻³). Mobilities are generally ≈ 50 cm²/V-s and 500 cm²/V-s, for p and n films, respectively.     

Optimization of undoped gaas by low-pressure OMVPE using trimethylgallium

Optimization of the electrical characteristics of undoped GaAs grown by low-pressure OMVPE using trimethylgallium is presented. The use of a lower growth pressure was found to reduce both then- andp-type background impurity incorporation. Both electrical and optical measurements revealed that the arsine partial pressure controls the background impurity level during low-pressure growth. Variation of this parameter allowed the attainment of both high mobility (maximum of 136,000 cm²/V-sec at 77 K in this study) and high resistivity layers suitable for field-effect transistor buffers. AlGaAs/ GaAs two-dimensional electron gas structures and GaAs MESFETs using the high resistivity buffer layers showed excellent electrical characteristics. Similar trends were obtained using different AsH₃ sources despite variations in purity.     

The growth of GaAs,AIGaAs, InP and InGaAs by chemical beam epitaxy using group III and V alkyls

The growth kinetics of chemical beam epitaxy (CBE) were investigated with the growth of GaAs, AIGaAs, InP, and InGaAs. Results obtained with epilayers grown by using trimethylarsine (TMAs) and triethylphosphine (TEP) instead of arsine (AsH3) and phosphine (PH₃) were reviewed with some additional results. The CBE grown epilayers have similar optical quality to those grown by molecular beam epitaxy (MBE). Superlattices of GaAs/AlGaAs with abrupt interfaces have been prepared. Since trimethylindium (TMIn) and triethylgallium (TEGa) used in the growth of InGaAs emerged as a single mixed beam, spatial composition uniformity was automatically achieved without the need of substrate rotation in the InGaAs epilayers grown. Lattice-mismatch Δα/α< 1 x 10^-3 have been reproducibly obtained. For epilayers grown with high purity TMAs source, room-temperature electron mobility as high as 9000 cm²/V sec and concentrations of ˜7 x 10¹⁵ cm^-3 were produced. In general, the electron mobilities were as good as those obtained from low-pressure metalorganic chemical vapor deposition. (MO-CVD). Unlike MBE, since the In and Ga were derived by the pyrolysis of TMIn and TEGa molecules at the heated substrate surface, respectively, oval defects observed in MBE grown epilayers due to Ga splitting from Ga melt were not present in CBE grown epilayers. This is important for integrated circuit applications. Unlike MO-CVD, the beam nature of CBE allows for selective area growth of epilayers with well-defined smooth edges using mask shadowing techniques. Typically, growth rates of 2-5μm/h for InP, 2-6μm/h for GaAs and AIGaAs, and 2-5μm/h for InGaAs were used.     

Rapid capless annealing of28Si,64Zn,and9Be implants in GaAs

We report the use of tungsten-halogen lamps for rapid (−10 s) thermal annealing of ion-implanted (100) GaAs under AsH₃/Ar and N₂ atmospheres. Annealing under flowing AsH₃/Ar was carried out without wafer encapsulation. Rapid capless annealing activated implants in GaAs with good mobility and surface morphology. Typical mobilities were 3700–4500 cm²/V-s for n-layers with about 2×10¹⁷cm⁻³ carrier concentration and 50–150 cm²/v-s for 0.1–5xl0¹⁹ cm⁻³ doped p-layers. Rapid thermal annealing was performed in a vertical quartz tube where different gases (N₂, AsH₃/H₂, AsH₃/Ar) can be introduced. Samples were encapsulated with SiO when N₂ was used. Tungsten-halogen lamps of 600 or 1000 W were utilized for annealing GaAs wafers ranging from 1 to 10 cm² in area and 0.025 to 0.040 cm in thickness. The transient temperature at the wafer position was monitored using a fine thermocouple. We carried out experiments for energies of 30 to 200 keV, doses of 2×10¹² to 1×10¹⁵ cm⁻², and peak temperatures ranging from 600 to 1000‡C. Most results quoted are in the 700 to 870‡C temperature range. Data on implant conditions, optimum anneal conditions, electrical characteristics, carrier concentration profiles, and atomic profiles of the implanted layers are described. Presented at the 25th Electronic Materials Conference, Burlington, VT, June 22, 1983.     

Probing the mechanisms of growth of gallium arsenide by metalorganic vapor phase epitaxy using experimental and theoretical studies of designed precursors

Decomposition of o-CH₃C₆H₄AsD₂ in the gas phase at 600–1000°C produces toluenes with 0-3 D atoms in the methyl group. It is shown that this cannot be accounted for by conventional mechanisms involving initial As-C bond cleavage or reductive elimination, but rather that the first step is As-D bond cleavage and this is followed by reductive elimination of o-CH₃C₆H₄D or H atom transfer to give o-HDAsC₆H₄CH₂· which abstracts D from an intact o-tolylarsine to give o-CH₂DC₆H₄AsHD. Repetition of these steps can lead to multiple D incorporation. The free energies of activation for reductive elimination or multiple D incorporation are found to be very similar. Theoretical studies on the decomposition of^tBuAsH₂ show that the first step for decomposition can be As-H bond cleavage to give^tBuAsH· or loss of H₂ to^tBuAs.^tBuAsH· decomposes to^tBu· which abstracts H· from^tBuAsH₂ to give 2-methylpropane or by β-H abstraction to give 2-methylpropene.^tBuAs, on the other hand only gives 2-methylpropene,via a β-H abstraction mechanism. Measured effects of total reactor pressure on product distribution are modeled qualitatively. Hex-5-enylarsine also decomposesvia initial As-H bond cleavage followed by reductive elimination of 1-hexene. However, it reacts in the liquid or solution phase with Me₃Ga to give the adduct. [Me₃Ga.AsH₂hex]. On heating, this loses methane to give first [Me₂Ga.AsHhex]₃ then [MeGa.Ashex]_n. Finally, GaAs is produced with the formation of methane and methylenecyclopentane. The last product indicates a free radical mechanism involving cleavage of the As-hex bond for the last step. In the gas phase at 600°C, GaAs is again formed but the major C₆ product is 1-hexene. This is interpreted as meaning that the adduct, [Me₃Ga.AsH₂hex] is not formed in the gas phase under growth conditions.     

Passivation of carbon acceptors during growth of carbon-doped GaAs,InGaAs, and HBTs by MOCVD

Carbon dopedp-type GaAs and In_0.53Ga_0.47As epitaxial layers have been grown by low-pressure metalorganic chemical vapor deposition using CC1₄ as the carbon source. Low-temperature post-growth annealing resulted in a significant increase in the hole concentration for both GaAs and In_0.53Ga_0.47As, especially at high doping levels. The most heavily doped GaAs sample had a hole concentration of 3.6 × 10²⁰ cm⁻³ after a 5 minute anneal at ≈400° C in N₂, while the hole concentration in In_0.53Ga_0.47As reached 1.6 × 10¹⁹ cm⁻³ after annealing. This annealing behavior is attributed to hydrogen passivation of carbon acceptors. Post-growth cool-down in an AsH₃/H₂ ambient was found to be the most important factor affecting the degree of passivation for single, uncapped GaAs layers. No evidence of passivation is observed in the base region of InGaP/GaAs HBTs grown at ≈625° C. The effect ofn-type cap layers and cool-down sequence on passivation of C-doped InGaAs grown at ≈525° C shows that hydrogen can come from AsH₃, PH₃, or H₂, and can be incorporated during growth and during the post-growth cool-down. In the case of InP/InGaAs HBTs, significant passivation was found to occur in the C-doped base region.     

Differences in Si doping efficiency in tertiarybutylarsine,monoethylarsine and arsine for GaAs and AlGaAs grown by MOVPE

We investigate differences in Si doping of GaAs and AIGaAs between group-V sources. Si₂H₆ and SiH₄ doping dependence on growth temperature, V/III ratio, total flow rate, growth rate, and off angle of substrate orientation was examined using tertiarybu-tylarsine (TBAs), monoethylarsine (EtAs), and arsine with a horizontal atmospheric pressure reactor. With either dopant source, Si incorporation for films grown using TBAs or EtAs was always higher than that using arsine. Using silane, dependence of Si incorporation on growth temperature and total gas flow velocity is different between group-V sources. Using disilane, dependence on V/III ratio and total gas flow velocity is different between group-V sources. These results imply that gas phase reactions play an important role. From the kinetic simulation of the decomposition of group-V sources, we verified that the concentrations of AsH₃, AsH₂, and AsH in vapor near the substrate are quite different among group-V sources. AsH₂ is dominant reactant when using TBAs. We propose that H₂AsSiH₃ (silylarsine) is formed by the reaction between AsH₂ radical and SiH₄ and silylarsine should contribute Si incorporation reactions, resulting in high Si incorporation efficiency with TBAs and EtAs. We also suggest that AsH₃ inhibits Si incorporation.     

Low temperature cleaning of Si by a H2/AsH3 plasma prior to heteroepitaxial growth of GaAs by metalorganic chemical vapor deposition (MOCVD)

A method using a H₂/AsH₃ plasma to clean the Si surface before GaAs heteroepitaxy was investigated and the dependence of the effectiveness of this treatment on arsine partial pressure was studied. Thin GaAs-on-Si films deposited on the plasma-cleaned Si were analyzed using plan-view TEM, HRXTEM and SIMS. Although not optimized, this method of Si cleaning makes heteroepitaxial deposition of GaAs possible. Some roughening of the Si surface was observed and a possible explanation is offered. Using the results of this study, thick (2.5–3.0μm) epitaxial GaAs films were then deposited and their quality was evaluated using RBS, XTEM and optical Nomarski observation. All Si surface cleaning and GaAs deposition were carried out at temperatures at or below 650°.     

Effect of high-temperature annealing on GaInP/GaAs HBT structures grown by LP-MOVPE

We have investigated the effect of high-temperature annealing on device performance of GaInP/GaAs HBTs using a wide range of MOVPE growth parameters for the C-doped base layer. Carbon doping was achieved either via TMG and AsH₃ only or by using an extrinsic carbon source. High-temperature annealing causes degradation of carbon-doped GaAs in terms of minority carrier properties even at doping levels of p=1 × 10¹⁹ cm⁻³. The measured reduction in electron lifetime and luminescence intensity correlates with HBT device results. It is shown that the critical temperature where material degradation starts is both a function of doping method and carbon concentration.     

Investigations on low temperature mo-cvd growth of GaAs

The growth conditions for the deposition at low temperatures of epitaxial layers of GaAs on (100) GaAs crystals using TMG and arsine are studied in detail. The films are grown at atmospheric pressure in a vertical reactor in which the arsine is fed in through the rf heated susceptor for precracking. The growth temperature was varied between 680°C and 450°C. In the whole temperature range epitaxial growth was obtained. The growth rate at temperatures below 600°C depends on the AsH₃ flow, suggesting that the availibility of As vapor species, not AsH₃ limits epitaxial growth in this temperature range. For a constant AsH₃ /TMG ratio of 8 the growth rate decreases by exp (-E/kT) with an activation energy of E = 1.5 eV. Growth rates as low as 0.5 um/h have been achieved. Unintentionally doped layers show semi-insulating behaviour at growth temperatures below 500° C, similar to the behaviour seen from MBE layers. However, n-type layers with reasonable mobilities can be grown in the low temperature range (450 ° C) using H₂ Se as the doping gas.     

Effects of trimethylindium on the purity of In0.5Al0.5P and In0.5Al0.5as epilayers grown by metalorganic chemical vapor deposition

In_0.5Al_0.5P lattice-matched to GaAs and In_0.5A1_0.5As lattice-matched to InP epilayers were grown by atmospheric pressure metalorganic chemical vapor deposition (AP-MOCVD). The effect of trimethylindium on the purity of the as-grown layers was systematically studied using secondary ion mass spectroscopy (SIMS), deep level transient spectroscopy (DLTS), and capacitance-voltage (C-V) measurements. The SIMS results showed that oxygen is the main impurity in all layers and the oxygen concentration in InAlP was approximately one to four orders of magnitude higher than the oxygen concentration found in InALAs when the same indium source was used, indicating that more oxygen was introduced by the phosphine source than by the arsine source. Two electron traps in the InAlP epilayers and four electron traps in the InALAs epilayers were observed in this study. When a high-purity indium source was used, the best InAlP epilayer showed only one deep electron trap at 0.50 eV while the best InALAs epilayer showed no deep levels measured by DLTS. In addition, we also found that a high concentration of oxygen is related to the high resistivity in both material systems; this suggests that semi-insulating (SI) materials can be achieved by oxygen doping and high quality conducting materials can only be obtained through the reduction of oxygen. The oxygen concentration measured by SIMS in the best InALAs epilayer was as low as 3 × 10¹⁷ cm⁻³.     

Low temperature Si2H6 Si epitaxy in-situ doped with AsH3/SiH4

The use of disilane (Si₂H₆) as a silicon source for epitaxial deposition was investigated for both very low pressure chemical vapor deposition (thermal CVD) and plasma enhanced chemical vapor deposition (PECVD) from 600 to 800° C. The growth rates observed for temperatures at or below 750° C were at least an order of magnitude higher than those observed for silane (SiH₄) using similar deposition conditions. An argon plasma was used to sputter clean the silicon surface, in-situ, immediately before the deposition. It was found that a low dc bias on the substrate during the argon sputter cleaning process helped remove carbon and oxide from the surface of the silicon substrate. A 16 min Ar sputter clean at 650° C, 2.5 W rf power, and •50 V dc bias resulted in a carbon and oxygen concentration at the epilayer-substrate of less than 4 × 10¹⁸/cm³ and 2 × 10¹⁸/cm³, respectively. In situ arsenic doping during disilane epitaxial growth was carried out by thermal CVD and PECVD using arsine (AsH₃) diluted in silane (SiH₄) at 800° C. The results were compared to similar experiments using only SiH₄ as the silicon source. Up to 500 ppm of arsine was diluted in the reactant gas and it was found that the Si₂H₆ growth rates were insensitive to the arsine concentraton in the gas phase.     

An analytical evaluation of GaAs grown with commercial and repurified trimethylgallium

An analytical study of the impurities in trimethylgallium (TMGa) and subsequent correlation of the effect of these impurities on resulting GaAs films grown by metalorganic chemical vapor deposition (MOCVD) is presented. The effects of using fractional distillation techniques to improve the quality of TMGa and to help isolate and identify major source impurities in TMGa is detailed. Photothermal ionization data are presented which show the residual donor species present and their relative concentrations in the epitaxial layers. Correlations of the residual donor concentrations with TMGa preparation are made. It is demonstrated that high purity GaAs with μ₇₇ K ≈ 125,000 cm²/V-sec can be grown by MOCVD using repurified trimethylgallium and arsine source materials. Work supported in part by the U.S. Naval Research Laboratory on Contract No. N00173-80-C-0066.     

Investigation of carbon incorporation in GaAs using13C-enriched trimethylarsenic and13Ch4

In the metalorganic chemical vapor deposition of GaAs there is increasing interest in replacing arsine with a less toxic arsenic source. However, GaAs films grown with metalorganic arsenic reactants usually contain significantly higher levels of carbon than films grown with arsine. Using 50% isotopically enriched¹³C trimethylarsenic (TMAs), we report the first direct evidence that the methyl groups from TMAs are a major source of the carbon observed in the GaAs films. The measured¹³C concentration in these films was 5 x 10¹⁶ cm.³ Conversely, incorporation of¹³C was not detected when 99%¹³C-enriched methane was added to the source gases during growth of GaAs with arsine in place of the¹³C-TMAs.     

Low pressure and low temperature gallium arsenide homoepitaxy employing in-situ generated arsine

Removed from the deposition region, an upstream hydrogen microwave plasma etches a surface of solid arsenic located downstream to generate arsenic hydrides. The latter are used with trimethylgallium (TMGa) to achieve low temperature (400–490° C) and low pressure (750 mTorr) homoepitaxial GaAs films. No active or afterglow plasma exists in the growth region. The homoepitaxial growth activation energy of 62 kcal/mole is consistent with the heterogeneous decomposition of TMGa in the absence of arsine. Precursor V-III ratios as low as 0.25 are used to achieve homoepitaxial films, but with high levels of carbon impurities (10¹⁹ to mid 10²⁰ cm⁻³). Carbon incorporation increases at low V-III ratios (0.25 to 0.5) for increasing temperatures with an activation energy of 23 kcal/mole. As the V-III ratios are increased above 1.0, the carbon incorporation activation energy decreases slightly to 15 kcal/mole.     

Reaction mechanisms of tertiarybutylarsine on GaAs (001) surfaces and its relevance to atomic layer epitaxy and chemical beam epitaxy

We explore the use of tertiarybutylarsine (TBAs) as an alternative arsine source in atomic layer epitaxy (ALE) of GaAs. X-ray photoelectron spectroscopy (XPS), reflection high energy electron diffraction (RHEED), and reflectance difference spectroscopy (RDS) are used to characterize the surface reactions of TBAs on GaAs (001) Ga-rich surfaces. At a substrate temperature of 320° C and an exposure level of 90 L of TBAs, AsH_x (x = 1 or 2) is thought to be the adsorbed arsenic species. As the substrate temperature increases, As-rich surfaces are readily obtained with improved RHEED 2 x 4 patterns. No carbon related species are observed throughout the TBAs exposure experiments between 320° C and 540° C. It is suggested that AsH_x is the adsorbed species after TBAs decomposes on surface Ga atom. Interactions between AsH_x pairs form arsenic atoms by H₂ release. RDS allows anin-situ real time study of TBAs on GaAs (001) Ga-rich surfaces. It is found that the RDS results are consistent with those obtained from XPS and RHEED investigations and can provide information on the rates of reactions and the extent of surface reconstruction simultaneously. Implications for the growth of GaAs by atomic layer epitaxy and chemical beam epitaxy using TBAs are discussed.