J. E. Baker

Carbon diffusion in undoped, n‐type, and p‐type GaAs

The effects of background doping, surface encapsulation, and As4 overpressure on carbon diffusion have been studied by annealing samples with 1000 A p‐type carbon doping spikes grown within 1 μm layers of undoped (n−), Se‐doped (n+), and Mg‐doped (p+) GaAs. The layers were grown by low‐pressure metalorganic chemical vapor deposition using CCl4 as the carbon doping source. Two different As4 overpressure conditions were investigated: (1) the equilibrium pAs4 over GaAs (no excess As), and (2) pAs4 ∼2.5 atm. For each As4 overpressure condition, both capless and Si3N4‐capped samples of the n−‐, n+‐, and p+‐GaAs crystals were annealed simultaneously (825 °C, 24 h). Secondary‐ion mass spectroscopy was used to measure the atomic carbon depth profiles. The carbon diffusion coefficient is always low, but depends on the background doping, being highest in Mg‐doped (p+) GaAs and lowest in Se‐doped (n+) GaAs. The influence of surface encapsulation (Si3N4) and pAs4 on carbon diffusion is minimal.

Heavy carbon doping of metalorganic chemical vapor deposition grown GaAs using carbon tetrachloride

A mixture of 500 ppm CCl4 in H2 has been used to grow heavily doped p‐type GaAs by low‐pressure metalorganic chemical vapor deposition with TMGa and AsH3 as the group III and V sources, respectively. Carbon acceptor concentrations between 1×1016 and 1×1019 cm−3 were obtained. In addition, abrupt carbon‐doping profiles were achieved with no noticeable memory effects. Carrier concentration was studied as a function of CCl4 flow, V/III ratio, growth temperature, and growth rate using electrochemical capacitance‐voltage profiling. Carbon incorporation was found to depend on CCl4 flow, V/III ratio, and growth temperature. Carbon incorporation had no dependence on the growth rate.

Background doping dependence of silicon diffusion in p‐type GaAs

Junction depth measurements via scanning electron microscopy and secondary ion mass spectroscopy are used to characterize silicon diffusion in GaAs crystals that contain varying amounts of zinc background doping. The zinc concentration is found to control the silicon diffusion process. A reason for this is suggested based on the shift in Fermi level with increased p‐type doping. Also, the electric field due to the p‐n junction formed at the silicon diffusion front is shown to have a large effect on the zinc background doping profile.

native oxide stabilization of AlAs-GaAs heterostructures

Data are presented on the stabilization of AlAs‐GaAs heterostructures against atmospheric (destructive) hydrolysis using the native oxide that can be formed (N2+H2O, 400 °C, 3 h) on the AlAs layer. The ∼0.1‐μm‐thick native oxide formed from the AlAs layer is shown to be stable with aging (∼100 days), while unoxidized samples degrade through the AlAs (0.1 μm) down into the GaAs as deep as ∼1 μm. Relative to oxides formed (∼25 °C) on AlAs (or AlxGa1−xAs, x ≳ 0.7) under atmospheric conditions (hydrolysis), oxides formed (via N2 +H2O) at higher temperatures (≳400 °C) are much more stable and seal the underlying crystal (e.g., GaAs).

Impurity diffusion and layer interdiffusion in AlxGa1−xAs‐GaAs heterostructures

Data are presented and a model describing the diffusion of the donor Si in GaAs from grown‐in dopant sources. In addition, the effects of background impurities on Si diffusion and layer interdiffusion in AlxGa1−xAs‐GaAs superlattices are described. These results are obtained on epitaxial GaAs samples with alternating doped and undoped layers and on AlxGa1−xAs‐GaAs superlattices with doped (Si or Mg) layers. The layer‐doped GaAs and the AlxGa1−xAs‐GaAs superlattices have been grown using metalorganic chemical vapor deposition and are characterized using secondary ion mass spectroscopy and transmission electron microscopy. Different annealing conditions are used to study the interaction between the grown‐in impurities and the native defects of the crystal controlling the diffusion processes. The model describing the impurity diffusion and layer (Al‐Ga) interdiffusion is based on the behavior of column III vacancies, VIII, and column III interstitials, IIII, and the control of their concentration by the posi...

IMPLANTATION AND ION BEAM MIXING IN THIN FILM ANALYSIS.

Abstract In thin film analyses obtained using sputtering techniques, primary ion implantation and ion beam mixing effects frequently produce significant alteration of the substrate before it can be sampled by the sputtering front. The present state of qualitative and quantitative understanding of these effects is discussed, with particular reference to: ion yield matrix effects and interface transients in secondary ion mass spectrometry, reduction of overlayer sputter rates due to ion beam mixing, the role of ion beam mixing in preferential sputtering, and the use of ion beam mixing to allow quantitative analysis of interfacial layers.

Sensitivity of Si diffusion in GaAs to column IV and VI donor species

Secondary ion mass spectroscopy and carrier concentration measurements are used to characterize Si diffusion into GaAs wafers containing two fundamentally different forms of donors, the column IV donors Si or Sn and the column VI donors Se or Te. A decrease in the Si diffusion rate is found in GaAs containing the column VI donors compared to the column IV donors. This trend is consistent with the model in which the Si diffuses as donor‐gallium‐vacancy complexes. The decrease in the Si diffusion coefficient is attributed to the greater binding energy of column VI donor‐gallium‐vacancy nearest‐neighbor complexes, thus reducing the concentration of free‐gallium vacancies available to complex with the Si.

Epitaxial growth of Si deposited on (100) Si

Epitaxial growth of deposited amorphous Si on chemically cleaned (100) Si has been found and layer‐by‐layer growth occurred at rates comparable to those in self‐ion‐implanted‐amorphous Si. There is no evidence for appreciable oxygen penetration into the deposited layer during storage in air. The critical factors in achieving epitaxial growth are fast (∼50 A/sec) deposition of Si onto a surface cleaned with a HF dip as a last rinse before loading into the vacuum system. Channeling and transmission electron microscopy measurements indicated that the epitaxial layers are essentially defect free. Secondary‐ion mass spectroscopic analysis showed about 1014 oxygen/cm2 at the amorphous/crystal interface. With either higher interfacial oxygen coverage or slow (∼2 A/sec) deposition, epitaxial growth rates are significantly slower.

Properties and use on In 0.5 (Al x Ga 1-x ) 0.5 P and Al x Ga 1-x As native oxides in heterostructure lasers

Data are presented demonstrating the formation of native oxides from high Al composition In0.5(AlxGa1-x)0.5P (x≳ 0.9) by simple annealing in a “wet” ambient. The oxidation occurs by reaction of the high Al composition crystal with H2O vapor (in a N2 carrier gas) at elevated temperatures (≥500° C) and results in stable transparent oxides. Secondary ion mass spectrometry (SIMS) as well as scanning and transmission electron microscopy (SEM and TEM) are employed to evaluate the oxide properties, composition, and oxide-semiconductor interface. The properties of native oxides of the In0.5(AlxGa1-x)0.5P system are compared to those of the AlxGa1-xAs system. Possible reaction mechanisms and oxidation kinetics are considered. The In0.5(AlxGa1-x)0.5P native oxide is shown to be of sufficient quality to be employed in the fabrication of stripe-geometry In0.5(AlxGa1-x)0.5P visible-spectrum laser diodes.

A comparison of TMGa and TEGa for low-temperature metalorganic chemical vapor deposition growth of CCl 4 -doped InGaAs

Factors which influence the alloy composition and doping level of CCl4-doped In0.53Ga04.7As grown at low temperatures (450°C < Tg < 560°C) by low-pressure metalorganic chemical vapor deposition (MOCVD) have been investigated. The composition is highly dependent on substrate temperature due to the preferential etching of In from the surface during growth and the temperature-dependent growth efficiency associated with the Ga source. The lower pyrolysis temperature of TEGa relative to TMGa allows the growth of CCl4-doped InGaAs at lower growth temperatures than can be achieved using TMGa, and results in improved uniformity. High p-type doping (p ∼ 7 × 1019 cm-8) has been achieved in C-doped InGaAs grown at T = 450°C. Secondary ion mass spectrometry analysis of a Cdoping spike in InGaAs before and after annealing at ∼670°C suggests that the diffusivity of C is significantly lower than for Zn in InGaAs. The hole mobilities and electron diffusion lengths in p+-InGaAs doped with C are also found to be comparable to those for Be and Zn-doped InGaAs, although it is also found that layers which are highly passivated by hydrogen suffer a degradation in hole mobility. InP/InGaAs heterojunction bipolar transistors (HBTs) with a C-doped base exhibit high-frequency performance (ft = 62 GHz, fmax=42 GHz) comparable to the best reported results for MOCVD-grown InP-based HBTs. These results demonstrate that in spite of the drawbacks related to compositional nonuniformity and hydrogen passivation in CCl4-doped InGaAs grown by MOCVD, the use of C as a stable p-type dopant and as an alternative to Be and Zn in InP/ InGaAs HBTs appears promising.