Deep level transient spectroscopy has been used to investigate hydrogen passivated p-type GaAs. Annealing studies have shown reactivation of the passivated CuB and Fe2+/3+ defects under differing bias conditions. The dissociation energies and dissociation attempt frequencies of these defects have been determined. It has been found that Fe–H has dissociation energies of 1.37 eV under zero bias annealing and 0.83 eV under reverse bias annealing. The CuB–H defect only displays measurable reactivation under reverse bias conditions, and has a dissociation energy of 1.57 eV. The passivated CuA defect does not show reactivation, even after extended periods of annealing at high temperatures.

Epitaxial p-type GaAs layers were exposed to a hydrogen plasma. Some of the passivated samples were then annealed under reverse bias in order to determine the energy for thermal dissociation of the passivated deep level-hydrogen complex. In particular, the deep levels attributed to iron and copper were examined. Deep Level Transient Spectroscopy (DLTS) has been used to determine the extent of passivation and reactivation of these deep levels. Although passivation of the copper levels has previously been reported, this is the first time that passivation of the iron level has been observed. A dissociation energy of 0.83 eV was determined for the Fe-H complex.

The EL2 defect in GaAs is passivated by exposure to a hydrogen plasma forming an electrically inactive EL2-H complex. The electrically active defect is restored to approximately 30 percent of its original concentration by annealing at 384 K for 2 hours in the presence of an electric field. The passivation of the EL2 is therefore concluded to be less stable than previously reported. The thermal dissociation of the EL2-H complex is characterised by two first-order processes, one significantly more rapid than the other.

Hydrogen passivates shallow dopants as well as the EL2 defect in n-type GaAs. The passivation of the EL2 defect is shown to be thermally less stable than the shallow dopant passivation. Thermal dissociation of the EL2-H complex under an applied reverse bias occurs more rapidly than under zero-bias conditions. It is shown that the reverse bias dissociation is characterized by two first-order processes, one considerably more rapid than the other. The rapid process has an activation energy of 1.33 eV and a dissociation attempt frequency of 1.0 X 10(14)s-1.

Undopedn --GaAs epitaxial layers were grown by OMVPE onn + (1 × 1018 cm3) and semi-insulating (SI) GaAs substrates. The as-grown epitaxial layers grown onn + substrates contained several deep level defects whereas those grown on SI substrates were, apart from the EL2, virtually “defect free”. Upon Cu diffusion, deep levels which may reduce hole and electron diffusion lengths and lifetimes, were formed. Optical deep level transient spectrocopy (ODLTS) has been used to identify such levels atE v + 0,41 eV andE c-0,31 eV respectively. The EO1 (EL2) trap concentration reduced after Cu had been diffused into the epitaxial layers. The magnitude of this reduction was approximately equal to the concentration of the trap found atE c - 0,31 eV which suggests that the two may be related. Activation energies and capture cross-section values are presented for the deep levels detected in these epitaxial layers.

A technique is presented for the analysis of digitally recorded capacitance transients in order to determine the energy and capture cross section of deep levels. The method is essentially a digital simulation of the measurement carried out by a lock-in amplifier. This technique in principle allows trap parameters to be determined in a single temperature scan. The more efficient use of data brought about by the method permits a detailed study of the Arrhenius plot and in particular gives insight into cases where the capacitance transient is nonexponential. The method may be adapted to allow more efficient data usage in conventional lock-in amplifier deep level transient spectroscopy.

The technique of cross-sectional transmission electron microscopy has been applied to study the nature of the radiation damage in GaAs bombarded with 300 keV protons to total doses of 1015 and 5×1015 protons cm−2. The results indicate that upon annealing at temperatures of 500 °C and above, the precipitated damage becomes visible in the form of hydrogen platelets (i.e., hydrogen-filled edge vacancy loops) on the {110} cleavage planes of GaAs.

Carbon ions with an energy of 150 keV were implanted in {111} surfaces of copper single crystals at doses of 1016 C+ cm−2 and 1017 C+ cm−2. The radiation damage below the implanted surfaces was examined by transmission electron microscopy of cross sections of the surface regions. The effects of annealing for 30 min at 600 and 750 °C were also studied. Damage consisted of a band of small defect clusters, dislocation loops, and dislocations in the band and extending from it. The depth of this band is approximately 40% deeper than that predicted from ion range theory, and the reason for this is not clear. The range increases with larger dose. The dislocation structures are consistent with the strains due to swelling being accommodated by dislocations.

The two-beam dynamical theory of electron diffraction has been used to calculate transmission electron microscope image profiles for small dislocation loops in GaAs. It is shown that the method of matching computed and experimental image profiles of loops is useful in the determination of their Burgers vectors. The results also indicate that the isotropic elasticity theory may be used to a good approximation for GaAs.

The damage and ion distributions resulting from room-temperature bombardment of (111) silicon single crystal surfaces with 300 keV boron ions to a dose of 1017 cm−2 are studied by cross sectional transmission electron microscopy (TEM) and Auger electron spectroscopy (AES). The peak of the damage distribution is found to coincide exactly with values predicted by the LSS theory. The evolution of the damage with annealing is studied to temperatures up to 1135 °C. Information on the development of dislocations from the damage zone is obtained and correlated with the boron depth distribution obtained by AES. Very convincing evidence of the redistribution of the boron atoms via pipediffusion follows from these studies. Das Damage und die Ionenverteilung, die von dem Zimmertemperatur-Bombardement von (111)- Siliziumeinkristalloberflächen mit 300 keV-Borionen einer Dosis von 1017 cm−2 herrühren, werden mit Querschnitts-Transmissionselektronenmikroskopie (TEM) und Auger-Elektronenspektroskopie (AES) untersucht. Es wird gefunden, daß das Maximum der Damageverteilung exakt mit dem von der LSS-Theorie vorhergesagten Wert übereinstimmt. Die Entwicklung des Strahlenschadens mit der Temperung wird bis zu Temperatnren von 1135 °C untersucht. Informationen iiber die Entwicklung von Versetzungen aus der Damagezone werden erhalten und mit der Bor-Tiefenver- teilung aus den AES-Messungen korreliert. Uberzeugende Hinweise für eine Umverteilung der Boratome iiber Kanaldiffusion folgen aus diesen Untersuchungen.

The theoretical distribution of implanted Boron in Silicon as computed by Gibbons et al. shows physically unrealistic negative values for energies higher than 100 keV. The distribution for B in Si has been measured by Auger Electron Spectroscopy for an implant energy of 300 keV. It was found that the third moment as computed by Gibbons is too high. A good curve fitting with the experimental data can be obtained by a simplified Edgeworth distribution with half the third moment value proposed by Gibbons. Implant distributions are predicted for higher energies.

The effect of thermal expansion on the temperature distributions around dislocations in III-V crystals during conduction parallel to their lines is considered. It is found that the thermal expansion has little effect on the temperature distribution close to the core of the dislocation. It is further shown that these temperature variations give rise to lateral variations in the diffusion constants of point defects and may hence play an important role in device degradation due to the climb of threading dislocations.

Lines of diffraction contrast have been observed near the interfaces of epitaxial layers of III-V compounds in transmission electron microscope samples cut perpendicular to the growth plane. The contrast occurs even for extremely small misfits. To explain the origin of these diffraction contours, a numerical stress analysis was carried out and the diffraction contrast was calculated using the dynamical theory of image contrast.

The energy expression for a general threading dislocation in a single-layer heteroepitaxial structure which bows out under the influence of the coherency strain is derived. It is found that, whereas some threading dislocations begin to bow out only once the epilayer has attained a certain critical thickness, others commence the bowing-out process as soon as the epilayer starts to grow. The results of this work are applied to the special cases of (001) and (211) oriented GaAs-Ga0.7Al0.3As single-layer structures. Der Energieausdruck für eine allgemeine, sich durch eine einschichtige heteroepitaktische Struktur ziehende Versetzung, die sich unter dem Einfluß der Kohärenzspannung verbiegt, wird abgeleitet. Es wird gefunden, daß, während einige der sich durchziehenden Versetzungen erst dann anfangen sich zu verbiegen, wenn die Epitaxieschicht eine gewisse kritische Dicke erreicht hat, die anderen den Verbiegungsprozeß beginnen, sobald die Epischicht zu wachsen beginnt. Die Ergebnisse dieser Arbeit werden auf den speziellen Fall von (001)- und (211)-orientierten GaAs-Ga0,7Al0,3As-Einschichichtstrukturen angewendet.

Misfit dislocation generation by means of the nucleation of dislocation half-loops is considered. The theory developed for semicircular dislocation loops is extended to half-loops lying in the Peierls troughs of the crystal. Contour plots of dislocation energy versus dislocation dimensions are produced from which the non-equlibrium behaviour of the loops can be deduced. Es wird die Bildung von Misfit-Versetzungen durch die Keimbildung von Versetzungshalbschleifen untersucht. Die für halb-zirkulare Versetzungsschleifen entwickelte Theorie wird auf Halb-Ver-setzungen, die in den Peierlströgen des Kristalls liegen, angewendet. Die Konturbilder der Versetzungsenergie über der Versetzungsdimension werden aufgezeichnet, aus denen sich das Nicht-gleichgewichtsverhalten der Schleifen ableiten läßt.

The energy of a double layer heterojunction device consisting of a substrate and two epilayers is calculated. The expression for the total energy of the system includes the coherency strain energy and the strain energy of the dislocations. The densities of misfit dislocations at the two interfaces are given as functions of the thickness of the second epilayer. The theory is extended to a similar system consisting of a substrate and three epilayers. The general theory is applied to the special case of GaAs-Ga0.95Al0.05As compound heterojunction structures. It is shown that a double-layer system attains coherency for small values of the first epilayer thickness. In the case of three-layer systems it is shown that for certain values of first and second epilayer thicknesses coherency can be attained at the second interface by increasing the thickness of the third epilayer. Die Energie einer Doppelschicht-Heteroübergangsordnung, die aus einem Substrat und zwei Epitaxialschichten besteht, wird berechnet. Der Ausdruck für die Gesamtenergie des Systems schließt die Kohärenzspannungsenergie und die Spannungsenergie der Versetzungen ein. Die Dichte der Misfit-Versetzungen an den beiden Grenzflächen wird als Funktion der Dicke der zweiten Epitaxialschicht angegeben. Die Theorie wird auf ähnliche Systeme aus einem Substrat und drei Epitaxialschichten ausgedehnt. Die allgemeine Theorie wird auf den speziellen Fall von Ga As-Ga0,95Al0,05As-Heteroübergangsstrukturen angewendet. Es wird gezeigt, daß ein Doppelschicht system Kohärenz für kleine Werte der Dicke der ersten Epitaxialschicht erreicht. Für den Fall eines Drei-Schichtsystems wird gezeigt, daß für bestimmt Werte der Dicken der ersten und zweiten Epitaxialschicht Kohärenz an der zweiten Grenzfläche durch Vergrößerung der Dicke der dritten Epitaxialschicht erreicht werden kann.

: The results of studies of interface structure in heterojunctions having near (001) orientations are briefly reported. Observations were carried out by X-ray topography, stress-induced birefringence and transmission electron microscopy and all support glide mechanisms and sources for the origin of the dislocations which compensate the misfit of the epilayer. In order to explore the effects of choosing different orientations on the interface structure, similar studies have been carried out on heterojunctions with 331 and 112 orientation. Preliminary results are reported. (Author)

The energy of a straight mixed dislocation in a coherent epitaxial layer at a uniform distance t from the interface is calculated. It is shown that, in addition to terms associated with coherency strain and the dislocation energy, there is a term associated with the interaction of the coherency and dislocation strain fields which makes the total energy decrease as the dislocation approaches the interface, but this term is negligible when the dislocation is nearer the epilayer surface than the interface. The result is used to modify a calculation of Jesser and Matthews of the critical thickness of an epitaxial layer at which the introduction of misfit dislocations becomes energetically favourable. The present work reduces the calculated critical thickness by 30–50%.