We have used electrical and optical methods to characterize and map semi-insulating GaAs single crystals grown by a liquid encapsulated vertical zone melting (VZM) method, including zone refining (ZR) by one or multiple passes of a molten zone. For (100)-oriented crystals, we monitored uniformity over the cross-section, and along the length from seed (bottom) to top. Hall and resistivity measurements for both VZM and ZR crystals grown without deliberate doping reveal near-intrinsic conditions [n0(300) < 7 × 106 cm-3] for much of crystal's length. The CAs acceptor concentration detected by local vibrational mode absorption decreases as growth proceeds, as expected for a carbon segregation coefficient exceeding unity. Near-infrared mapping shows an EL2 defect concentration N0 ∽ 1016 cm-3 for both VZM and ZR crystals, with no variations of crystallographic significance across a wafer's area. This is consistent with the small dislocation density (ND<5000 cm-2), and absence of any stress birefringence.

Electrical and optical measurements are reported for samples from two undoped semi-insulating GaAs crystals grown by a vertical molten zone method. The electrical data, taken over the range 290–420 K, included results for samples from both crystals that were so close to intrinsic as to require an ambipolar correction in determining the electron concentration. The compensation balance in this material is controlled by the EL2 midgap defect, of which the fraction ionized depends on trace presence of CAs shallow acceptors, and of shallow donors, probably including SiGa . An increase of the latter towards the tail (top) end of one crystal led to a reduction of the EL2 ionized fraction, and a lowered resistivity—but one still within the conventional semi-insulating range. Carbon was measured from the strength of its local vibrational mode absorption, while near-infrared measurements showed that EL2 was present in a concentration ∼1016 cm−3, with relatively small variation across a wafer.

A new method for growing single crystal GaAs has been developed. This new method utilizes a liquid encapsulant (boric oxide), and a pyrolytic boron nitride (PBN) crucible in a vertical configuration. A special furnace with a thermal spike creates a molten zone in a bar of GaAs and furnaces on either end of the spike adjust the length of the molten zone and control the thermal gradient at the growth interface.

Both optical and electrical methods were used to provide separate evaluations of the fraction Pi of EL2 defect donors which have lost an electron. The semi-insulating GaAs which permitted these assessments had a large enough fraction of the EL2 compensated (by CAs acceptors) to make this a sensitive test. Room-temperature Hall effect data, analyzed with an ambipolar correction, indicated that Pi ≂0.34 in material from a vertical-zone grown crystal, while near-infrared optical absorption could be modeled on a superposition of EL2 photoionization and photoneutralization to yield Pi ≂0.35 in the same crystal. The presence of compensating CAs acceptors was verified by local mode vibrational absorption.

Thin films of YBa2Cu3Ox have been grown by spraying nitrate solutions of Y, Ba, and Cu onto a heated substrate. The new orthorhombic phase of the deposited film has unit cell parameters of a = 4.701 Å, b = 5.747 Å, and c = 16.245 Å. Annealing the film at 950°C causes the crystal structure to become the superconducting orthorhombic 1:2:3 phase with a = 3.860 Å, b = 3.877 Å and c = 11.663 Å. Results are presented for SEM and X-ray diffraction measurements.

Photoluminescence (PL) and photoluminescence excitation (PLE) spectroscopy have been performed on the 0.635-eV emission band in GaAs. Recently, this band has become of increased importance as a consequence of having been associated with the main electron trap (EL2) in GaAs. All samples studied exhibit a strong peak in the PLE at 1.5 eV, whose characteristics indicate that this emission band arises from deep donor-acceptor pair recombination. We also observe a related PL band at 0.68 eV which does not exhibit the 1.5-eV peak in the PLE and is attributed to recombination between a deep level and a hole in the valence band. The PLE spectra for both of these emission bands exhibit oscillations resulting from the resonance cooling of hot electrons by the emission of LOΓ phonons. This phenomenon, which has apparently never before been reported for such a deep level, suggests that an intermediate shallow donor state is involved in the recombination processes for both of these PL bands. We will discuss the relationship between our measurements and various microscopic models which have been proposed to describe the EL2 defect center.

We have examined polycrystalline and single crystal, liquid encapsulated Czochralski (LEC) grown InP for evidence of boron contamination using secondary-ion mass spectrometry (SIMS) and photoluminescence. Precipitates of boron or a boron compound have been found in InP grown by the LEC method with boric oxide (B2O3) encapsulation and pyrolytic boron nitride (pBN) crucibles. The density of precipatates appears to increase towards the last-to-freeze end of the boules. Neither B2O3 or pBN appears to be solely responsible for the observed phenomenon. At present, the source of boron contamination most consistent with our observations is the result of interaction between B2O3 and pBN. A diffusion coefficient of D(750°C) < × 10-14 cm2 s-1 for boron in InP has been determined in this work.

Ion-implanted one micron gate length InP FET's are described with noise figures as low as 3.5 dB at 12 GHz. This is the lowest published noise figure for InP FET's. The InP FET microwave performance data are compared with those of equivalent geometry GaAs FET's with either ion implanted or epitaxial channels. Microwave results indicate a definite gain advantage of InP FET's over these GaAs counterparts. Noise figures of both types of FET's are comparable but InP substrate improvement and implant profile optimization are suggested as a means of further reduction of the noise figure of InP FET's.

Several samples with Fermi levels ranging from the valence to the conduction band show that the resonance previously attributed to ${\mathrm{Cr}}^{1+}$ is due to ${\mathrm{Cr}}^{4+}$, and no additional signal is observed under conditions where ${\mathrm{Cr}}^{1+}$ would be expected to exist. The double-acceptor property postulated for Cr in GaAs is thus questionable. Depending on the Fermi-level location, two different spectra of light-induced changes in the Cr charge states are observed.

The application of a dc homogeneous electric field in a GaAs crystal will induce a quadrupolar splitting of the nuclear-spin levels. This splitting has been observed for $^{71}\mathrm{Ga}$, $^{69}\mathrm{Ga}$, and $^{75}\mathrm{As}$ using a pulsed NMR spectrometer. A constant (${R}_{14}$) that relates the coupling between the induced field gradient and the applied electric field has been measured. The values found were ${R}_{14}(\mathrm{Ga})=2.85\ifmmode\times\else\texttimes\fi{}{10}^{10}$ ${\mathrm{cm}}^{\ensuremath{-}1}$ and ${R}_{14}(\mathrm{As})=3.16\ifmmode\times\else\texttimes\fi{}{10}^{10}$ ${\mathrm{cm}}^{\ensuremath{-}1}$. These results are in disagreement with those from a similar experiment reported by Gill and Bloembergen. The measured values of ${R}_{14}$ were found to be independent of the applied electric field strength. The two-electron bond-orbital model developed by Huang, Moriarty, Sher, and Breckenridge has been used to interpret the experimental measurements. In evaluating the theory, a Hartree-Fock atomic $s{p}^{3}$ hybrid wave-function basis set is used to find the induced field gradient caused by the electrostatic distortion of the bond. With no adjustable parameters, the value of ${R}_{14}(\mathrm{As})$ calculated from the theory is smaller than the experimental number by a factor of 1.4, while the calculated ${R}_{14}(\mathrm{Ga})$ is a factor of 4 too small.

Chromium can assume three different charge states in semi-insulating GaAs :Cr. An EPR-optical method is described which allows the Cr concentration in each of these states to be quantitatively determined. Typical results are given for a number of Cr-doped GaAs samples.

Bulk polycrystalline InP is synthesized from the elements via a gradient freeze process. Hall data for a typical boule are Nd-Na= 4.7 × 1015/cm3 and Μ77 = 28,000 cm2/V-sec. Photoluminescence data indicate that zinc is present as an acceptor impurity in the polycrystalline InP and in nominally undoped LEC single crystals grown using the synthesized InP as charge material. A series of doping experiments have determined the effective segregation coefficient to be 1.6 × 10−3 for Fe in InP. Semi-insulating InP crystals with resistivity > 107 ohm—cm have been grown consistently from melts doped with 150 ppm Fe.

: High purity GaAs has been compounded in pyrolytic boron nitride (PBN) ware. Semi-insulating GaAs single crystals have been grown by the liquid encapsulation technique. High purity InP has been compounded in PBN boats and single crystals of InP have been grown by the high pressure liquid encapsulation technique. Liquid phase epitaxial layers of GaAs have been grown on a semi-insulating substrate that has been etched with an in-situ gallium etch prior to growth. (Author)