Silicon-doped Gallium Research Articles

Photoluminescence study of acceptors in silicon-doped gallium arsenide grown by metalorganic chemical vapor deposition

Epitaxial silicon-doped GaAs layers were grown by metalorganic vapor deposition (MOCVD) using a silane-argon mixture as a dopant source. Two different dopant sources with silane mole fractions 2.4×10 −4 and 2.3×10 −5 were used. For both sources the electron concentration increased with the silane mole fraction in the gas flow, reached a maximum, and then fell abruptly. The maximum electron concentration was substantially higher (up to 2.1×10 19 cm −3) for layers grown with a more concentrated dopant source. The difference in silicon incorporation from different silane sources is thought to be caused by uncontrollable impurities in the diluting gas - argon. The low temperature (12 and 80 K) photoluminscence investigation revealed the presence of at least three types of acceptors in the layers: residual impurity zinc, group IV acceptors silicon versus carbon and the deep acceptor complex, silicon donor-gallium vacancy, responsible for the emission band with peak energy 1.24 eV. As a rule, this emission band was dominant in the most heavily doped layers with low electron mobilities, and its intensity was low for layers with good mobility values. Therefore the silicon donor-gallium vacancy complex is thought to be the main compensating acceptor in most heavily doped layers.

A new high purity doping source for CVD Si: Ga epitaxy

A new CVD growth chemistry has been developed for producing very pure and high crystalline-quality gallium-doped silicon epitaxy. Gallium trichloride, GaCl 3, has been employed as an improved precursor to the currently accepted organometallic gallium source, trimethyl gallium, (CH 3) 3Ga. In contrast to trimethyl gallium, this metal halide doping source does not cause carbon contamination of the epitaxial layer. Much more abrupt layer-to-layer transition regions have been demonstrated. Data on the electrical and crystalline characteristics of GaCl 3-grown Si:Ga layers are presented. The current doping limit of 4X10 17 cm -3 obtained with this chemistry under normal CVD growth conditions is well below the solubility limit and is indicative of the unusual surface chemistry involved in doping. Results of experiments to elucidate the doping mechanism are discussed.

Mechanism of compensation in heavily silicon-doped gallium arsenide grown by molecular beam epitaxy

Silicon-doped GaAs grown by molecular beam epitaxy has been characterized by Hall measurements, infrared local vibrational mode (LVM) absorption, secondary ion and laser source mass spectroscopy. Highly doped samples with [Si]∼3×1019 cm−3 show only a low carrier concentration of 8×1017 cm−3. LVM spectroscopy shows that SiGa donors are compensated predominantly by [Si-X] complexes, where X has been assigned previously to a gallium vacancy (VGa). Other compensating impurities are not present in the layers at significant concentrations.

Photoluminescence from heavily doped Si layers grown by liquid-phase epitaxy

Heavily phosphorus or gallium-doped silicon was grown by liquid-phase epitaxy and studied by photoluminescence. For the phosphorus-doped samples grown from In(P) solution, recombination of free electrons to compensating In acceptor levels was observed besides the free electron-free hole band-to-band emission. The gallium-doped samples showed a luminescence spectrum similar to the one observed in bulk-doped p-type material, indicating a good crystalline quality and low compensation in these samples. The band-gap shrinkage was found to be larger in heavily gallium-doped than in boron-doped silicon, indicating a dependence of this shrinkage on the chemical nature of the dopant atoms.

Determination of silicon in gallium arsenide by electrothermal atomisation atomic absorption spectrometry using the L'vov platform

Electrothermal atomisation atomic absorption spectrometry with a L'vov platform was investigated for the determination of silicon in gallium arsenide. The sample was decomposed with nitric acid-hydrochloric acid (3 M+ 1 M), calcium nitrate was added as a matrix modifier and the diluted solution was atomised from a tantalum carbide coated tube and platform. Ashing at 1650 °C and deuterium arc background correction allowed the non-specific absorbance to be eliminated. For a 40-mg sample the detection limit (3σ) is 4.0 µg g–1, equivalent to 5 × 1017 atoms cm–3. Results obtained for silicon-doped gallium arsenide samples are presented.

Additional observations on the gallium-X center in neutron-irradiated Si:Ga

In a previous paper we reported the formation of the gallium-X center by neutron irradiation of gallium-doped silicon. Since that report we have investigated a variety of other neutron-irradiated Si:Ga samples using float-zone crystals grown by three different vendors and irradiated at four different reactors. The behavior of Ga-X as a function of irradiation conditions, annealing temperatures (up to 800 °C) and various material properties has been studied by Fourier-transform infrared absorption spectroscopy and temperature dependent Hall-effect measurements. Ga-X was not observed in any of the as-grown samples but was definitely present in every irradiated sample. In all cases it first appears after a 400 °C anneal, reaches maximum concentration after a 600 °C anneal, and then decreases at higher anneal temperatures. The maximum Ga-X concentration observed in any sample appears to be dependent on the carbon concentration in that sample and not on the irradiation conditions. A more detailed Ga-X absorption spectrum was obtained showing lines not previously reported. A binding energy of 57.18±0.03 meV is deduced for the Ga-X ground state.

Infrared excitation spectrum of 40.4-meV acceptor level in neutron-irradiated gallium-doped silicon

An infrared absorption spectrum has been obtained for the shallow A2 acceptor level formed by neutron irradiation of Si:Ga. After a 600 °C anneal, absorption peaks were observed at 213.1, 244.9, and 286.9 cm.−1 These peaks appear to correspond to lines 1, 2, and 4 of a group III-like acceptor spectrum originating from a ground state with a 40.4-meV binding energy. Hall effect measurements confirm the presence of an acceptor level at this energy.

Focused ion beam microlithography using an etch-stop process in gallium-doped silicon

We have found that silicon, when implanted with doses of gallium in excess of 1013 ions/cm2, experiences little or no etching in aqueous caustic solutions. By exploiting a finely focused 40–50 keV gallium ion beam (0.05–0.1 μm diameter) in our scanning ion microscope, we have shown that it is possible to fabricate structures with submicrometer (0.1 μm) features. The silicon behaves as a negative resist with a sensitivity of about 1 μC/cm2. This etch-stop process is largely insensitive to crystallographic orientation, except for the highly insoluble (111) plane in which damage-promoted etching occurs.

Defects in electron-irradiated, gallium-doped silicon

Defect processes were studied in electron-irradiated, Ga-doped silicon using deep level transient spectroscopy. Evidence for the production of a Ga-related defect state is presented. Changes in minority-carrier lifetime are correlated with recovery and growth of defect states in annealing studies of Ga-doped silicon.

Autocompensation in molecular beam epitaxial gallium arsenide: The (110) orientation

Autocompensation has been studied on the (100) and (110) orientations of gallium arsenide. Both n- and p-type silicon-doped (110) gallium arsenide films have been grown by only varying the substrate temperature. Electrical and optical data show that n-type (110) films are heavily autocompensated with silicon.

Heat treatment of semi-insulating chromium-doped gallium arsenide substrates with converted surface removed prior to molecular beam expitaxial growth

Silicon-doped gallium arsenide molecular beam epitaxy (MBE) layers were grown on chromium-doped semi-insulating GaAs (Bridgeman grown) heat-treated (converted surface removed before growth) or nonheat-treated substrates. Secondary ion mass spectrometric (SIMS) measurements show a marked reduction of outdiffused manganese in layers grown on heat-treated substrates with converted surface removed prior to MBE growth. Secondary ion mass spectrometric studies and capacitance voltage free-carrier profiling indicate a redistribution of Si in layers grown on nonheated substrates not observed in epilayers on heat-treated substrates (converted surface removed prior to MBE growth).

A study of the effects of annealing, zinc diffusion and copper diffusion on the defect structure of silicon-doped gallium arsenide

Transmission electron microscopy has been employed to study as-received, annealed, zinc-diffused and copper-diffused silicon-doped GaAs. The results on the as-received material show that the defect structure is dependent on the silicon doping level, with small perfect and faulted loops being present at high doping levels. The results of the annealing study on heavily doped material show that the size and density of the loop structure is strongly dependent on the temperature and type of loop. This behaviour has been interpreted in terms of previous work on the variation of the point-defect populations and free-carrier concentration with annealing temperature. The effect of zinc diffusion on the defect structure of heavily Si-doped GaAs has been found to be very similar to that observed after equivalent diffusions into Te-doped GaAs and is explained with reference to these results. The differences between the two cases are thought to be mainly due to the amphoteric nature of silicon in GaAs. The diffusion of copper has been found to result in large increases in the perfect and faulted loop areas. This, it is suggested, is due to the mechanism of copper diffusion in GaAs and the effect on point-defect populations of in-diffusing a p-type dopant into n-type material.

A simple source cell design for MBE

A simple, low-cost, and easily-manufactured evaporation furnace, suitable for MBE, is described. The essential feature of the design is a composite heater-shroud and evaporation cell, machined from high-purity graphite. Good-quality silicon-doped gallium arsenide has been grown using this cell design.

Luminescence of traps in electron-irradiated gallium-doped silicon

Photoluminescence measurements are presented for a series of lines observed in Ga-doped silicon. The defect responsible for the luminescence has recombination properties which indicate that the responsible center is a highly radiative neutral trap. Two dominate lines are observed from a spin-singlet, spin-triplet pair. A model based on the properties of a trapped interstitial is postulated to explain the observed structure.

Zero-phonon transition of free electrons to bound holes in heavily gallium-doped silicon

A radiative zero-phonon transition of free electrons to bound holes has been observed in heavily gallium-doped silicon at various temperatures. This electronic transition competes with excitonic transitions, and dominates at high temperatures.

Optical absorption of gallium-doped silicon

The spectral dependence of the infrared absorption of Ga centers in silicon has been measured. Dependence on doping and impurity concentration and temperature is found for the optical cross section. The splitoff valence band has to be taken into account to explain the spectral dependence. Excited bound states cause a temperature decrease of the optical cross section.

Infrared spectra of new acceptor levels in boron-doped and gallium-doped silicon

Infrared-absorption measurements have shown the existence of shallower acceptor levels associated with both boron and gallium in silicon. The new spectra correspond to effective-mass-like acceptors with optical ionization energies of 37.1±0.3 meV in boron-doped silicon and 57.0±0.3 meV in gallium-doped silicon. Similar centers have been previously identified in indium- and aluminum-doped silicon so the defect structure involved seems to be common to the group-IIIA acceptors in silicon. The presence of this defect explains some of the early discrepancies between the thermally and optically measured activation energies for boron and gallium in silicon.

Characteristics of the Indium-and Gallium-Doped Silicon Infrared Sensing MOSFETS 's (IRFET's)

The extrinsic silicon infrared sensing MOSFET (IRFET) is an integrated circuit element which has been designed for infrared imaging applications employing large scale integrated arrays, in the near, middle and far infrared wavelength regions. Detailed results have been previously published on the operation of the gold-doped devices in the near infrared wavelength regions. This paper specifically addresses the characterization of the indium-doped and gallium-doped infrared sensing MOSFET in the middle, 3 to 5 micrometer, and far, 8 to 14 micrometer, wavelength regions respectively. Quantum efficiencies in the range 1 to 10.0% and responsivities in the range 1×104 to 1×109 A/J or 1×102 to 1×106 A/W can be obtained.

The strength of the infrared absorption from silicon donors and silicon acceptors in gallium arsenide

Hall measurements have been made on Bridgman and liquid-phase epitaxial silicon-doped gallium arsenide. The samples were then rendered transparent by e--irradiation and the absorption due to the localized vibrations of SiGa donors and SiAs acceptors was measured. These data were used to obtain values of 1.99e and 1.75e respectively for the apparent charges of the two defect modes, with the assumption that there were no other donors or acceptors present. This calibration is very similar to an earlier one due to Spitzer and co-workers (1968-69), but is significantly different with respect to the interpretation of electrical data for LPE layers.

Silicon-doped gallium arsenide antimonide electroluminescent diodes emitting to 1.06 μm

We report efficient room-temperature emission to 1.06 μm from silicon-doped gallium arsenide antimonide electroluminescent diodes fabricated in GaAs1−xSbx layers grown on GaAs substrates in a single-step LPE growth from melts containing from 0.1 to 1.1 at.% Si. The spectral half-width of the emission is 42.5 nm at 0.96 μm and 79 nm at 1.06 μm.