Misfit Dislocation Structure Research Articles

Descriptions of low energy misfit dislocation structures using the parabolic interaction potential

The need for and development of the parabolic models as a power series approximation of the sinusoidal Frenkel-Kontorowa model, used to study low energy misfit dislocation (MD) structures in epicrystals, is briefly discussed. Its application to monolayers on thick substrates, thickening epilayers, superlattices and structural ledges is outlined with special emphasis on critical misfit and thickness in coherent to incoherent transitions. The “nucleation” of an MD by climb from the free surface is calculated. The critical misfits obtained from (a) the continuum (approximation) and exact solutions, using the parabolic model, and (b) the continuum solutions of the sinusoidal and parabolic models are compared and the differences between the various approaches assessed.

Misfit dislocations in In0.15Ga0.85As/GaAs strained-layer superlattices

The misfit dislocation structure in In0.15Ga0.85As/GaAs strained-layer superlattices has been characterized by transmission electron microscopy. Most of the dislocations are 60° dislocations confined to the buffer/strained-layer superlattice interface. This interface also contains sources that emit glissile dislocations into the {111} planes. These dislocations develop into the buffer layer and the substrate but hardly into the strained-layer superlattice. The sources are connected to the interactions between misfit dislocations in the buffer/superlattice interface.

Structural and Electronic Properties of GaAs/InGaAs/GaAs Heterostructures

AbstractWe have made a study of GaAs/InGaAs/GaAs (001) strained layer heterostructures using Transmission Electron Microscopy (TEM) as a structural tool to determine the misfit dislocation structure and density as a function of Indium concentration. The average misfit dislocation spacing varies from > 10 µm for x < 0.3, to a few microns at x = 0.3, and drops to a few hundred Angstroms at x = 0.5. We did in-situ annealing experiments in order to study the strain relaxation process, measuring the temperature at which the structure begins to relax, and the dislocation velocities. Dislocation velocities are a few microns per second at the growth temperature of 450 ° C, and tens of microns per second at 690 ° C. In addition to interfacial dislocations in the usual <110> directions, in samples where x ≥ 0.4, we observed dislocations running in <100> directions. A study of the electrical characteristics of the material was made in parallel with the structural measurements: the mobility of the InGaAs layer was measured, the material was processed into Heterojunction Bipolar Transistors (HBT’s) and the gain was measured. The electrical characteristics initially improved with the addition of In, peaking at x = 0.1 and dropping sharply for higher x.

Misfit Dislocation Structures at MBE-Grown Si1-xGex/Si Interfaces

Misfit dislocation structures introduced at Si1-xGex/Si (001) Interfaces were studied using transmission electron microscopy to understand the relaxation mechanism of lattice misfit in this material system. Misfit dislocations observed at the interfaces for x≤0.5 are inclined 60° dislocations introduced by glide processes. The density of the misfit dislocations increases with mole fraction or thickness of the Si1-xGex as a result of the Hagen-Strunk multiplication mechanism. The threading dislocations contained in the epilayers originate in transition dislocation segments appearing during the multiplication processes. In contrast with these results, the misfit dislocations observed at the Ge/Si interface are the 90° -type introduced by climb processes. The misfit dislocations were regularly distributed with an average spacing of ∼100Å, which is sufficient to relax the 4.2% misfit. Thus, the relaxation mechanism of the misfit strain in the Si1-xGex/Si system changes depending on the lattice misfit.

Boron Precipitation and misfit dislocation structure in Si(Ge,B) buried epitaxial layers

Heavily boron-doped silicon buried epitaxial layers are becoming increasingly important in the fabrication of thin membranes, three dimensional structures in silicon and latch-up free CMOS circuits. Si(Ge,B) co-doping has been utilized to compensate the B-induced lattice contraction in Si and hence buried high conducting layers which are strain-free and lattice matched to the Si substrate have been realized. The utilization of isoelectronic Ge also alters the point defect distribution in silicon resulting in reduced dopant diffusion which is an added advantage in realizing shallow junctions and reduced interfacial transition width in epitaxial layers. This contribution addresses the evolution of misfit dislocation structure and B precipitation behavior in heavily B-doped buried Si epitaxial layers. In addition, the effect of Ge co-doping on B solubility in Si will be discussed.Silicon epitaxial layers were grown at 1080°C by chemical vapor deposition on 4-inch diameter p-type (100) substrates (10 and 0.04 Ω-cm) employing the SiH2Cl2-B2H6-GeH4-H2 chemical system. Single 5 μm thick epilayers and 2 μm buried layers with 4 μm intrinsic cap layers (10 Ω-cm) were grown.

Misfit dislocation structure at a Si/SixGe1−x strained-layer interface

The misfit dislocation structure at a Si/Si0.75Ge0.25 strained-layer interface has been characterized by transmission electron microscopy. Through weak-beam imaging it is found that partial dislocation in the form of extended nodes exist in the misfit dislocation network. The density of nodes as observed by microscopy compares favorably with the estimate of the density of charged interface states derived from capacitance-voltage measurements.

Microstructure of MBE Grown Al/Gaas [100] schottky contacts

The Aℓ/GaAs contact system is used in microwave field effect transistors, but to date, little systematic TEM characterization of the contact microstructure has been performed. There are three possible epitaxial configurations of Aℓ on GaAs [100].[100]Aℓ; Aℓ [100] ||GaAs[100], Aℓ [010] ||GaAs[011] [110]Aℓ; Aℓ [110] ||GaAs[100], Aℓ [110]||GaAs[011] [110]R Aℓ; Aℓ[110] ||GaAs[100], Aℓ [001] ||GaAs [011]The interfacial atomic structure of Aℓ[100] ||GaAs[100] has been investigated previously using a CBED technique on plan view specimens. this paper we study the distribution of the various Aℓ orientations as a function of initial GaAs surface reconstruction and film thickness. The misfit dislocation structure of [100] Aℓ/[100] GaAs has been characterized by weak beam microscopy. Also, lattice images from cross-sectional samples are presented. The samples studied were MBE grown and subsequent electrical characterization showed them all to have Schottky barrier heights (SBH) of 0.77eV and excellent rectifying properties over a large temperature range.

Misfit and threading dislocations in HgCdTe epitaxy

A new theory relating substrate, misfit, and threading dislocations is presented for abrupt and graded heterojunctions based on the assumption of complete local strain relief. We have for the first time been able to treat two- and three-dimensional misfit dislocation structures in a unified way and experimentally verify the distinction in epitaxial HgCdTe. We confirm earlier findings that the areal density of interfacial dislocations nA in a cross section of the interface is proportional to the lattice constant gradient. If the interface width w is narrower than a theoretically predicted critical value w0, the misfit structure becomes two-dimensional and a linear dislocation density nL applies. The maximum of nA (attained at w=w0) is equal to n2L. Another new result of the theory is that the average misfit segment length is inversely proportional to the substrate dislocation density. The threading dislocation density is predicted to be low when the average segment length is comparable to the substrate lateral dimensions. For cases of low epitaxial dislocation content, the calculated substrate densities corresponding to such segments compare favorably with measured values. Compositional profiles and misfit dislocation densities were obtained on the same angle lapped HgCdTe epilayers by visible reflection spectroscopy and preferential etching, respectively. The former monitored the E1 interband transition as a function of position with a microspectrophotometer while the latter employed a modified Polisar No. 2 etchant. Both linear and areal distributions were observed and their occurrence was correctly predicted by the theory.

Characteristics of lath martensite: Part II. The martensite-austenite interface

This investigation, using an Fe-20 pct Ni-5 pct Mn (wt pct) alloy, deals with the nature of the lath martensite-austenite interface. For the first time the misfit dislocation structure associated with a martensite interface has been observed experimentally. The interface consists of a single set of parallel dislocations having Burgers vector α/2[l•1l]martensite = α/2[011]austenite. Relative to the austenite, the observed dislocation line direction is [0•57], and the dislocation line deviates about 10 and 15 deg from the pure screw orientation in the austenite and martensite, respectively. However, the dislocations are in screw orientation on an atomic scale, although the interface step structure causes them to deviate from the exact screw orientation macroscopically. The spacing of the interface dislocations varies from 26 to 63A. The observed interface dislocation array satisfies the requirements for a glissile interface, which suggests that the dislocations are misfit dislocations which accomplish the lattice invariant shear of the crystallographic theories.

Comments on “The core structures of misfit dislocations in grain boundaries”

(1979). Comments on “The core structures of misfit dislocations in grain boundaries”. Philosophical Magazine A: Vol. 39, No. 6, pp. 821-823.

Reply to “Comments on “The core structures of misfit dislocations in grain boundaries””

Reply to “Comments on “The core structures of misfit dislocations in grain boundaries””

The interfacial structure of the broad faces of ferrite plates

An investigation of the interfacial structure at the broad faces of ferrite plates embedded in retained austenite has been undertaken in an Fe-0.62 wt% C-2.0 wt% Si alloy by means of weakbeam dark-field TEM. These boundaries have been shown to be partially coherent and to contain single arrays of parallel misfit dislocations spaced ca. 15 to 25 Å apart and arrays of parallel structural ledges spaced ca. 22 to 90 Å apart. The dislocations are mixed in character and have lattice-type Burgers vectors lying in the atomic habit planes of the interface. The structural ledges are triatomic in height, with spacings and directions varying with the rotation of {111} γ relative to {110} α , about a common orthogonal axis. Data on apparent habit planes and orientation relationships were obtained with stereographic projection and trace analysis techniques. Comparisons of these data with those obtained by means of the interface modeling techniques described in the companion paper show good agreement. The misfit dislocation structure is always sessile with respect to glide out of the interface, whereas the lattice invariant deformation portion of a martensitic f.c.c. → b.c.c. transformation requires a glissile structure; hence, it is mechanistically impossible for ferrite plates to thicken by a martensitic mechanism.

Misfit dislocations in heteroepitaxial systems with low dislocation density

The growth of heteroepitaxial layers causes stress across the interface which in a certain range of layer thickness may he relaxed by the formation of misfit dislocations at the interface. Systematic investigations of such misfit structures by transmission electron microscopy (TEM) have previously heen conducted only on systems with a relatively large misfit parameter f, i.e. with a high dislocation density. However, this case presents difficulties in the analysis because of the complexity of the dense structures.Interfaces containing a low density of misfit dislocations, e.g. heteroepitaxial systems with low misfit, should consequently be investigated. High voltage electron microscopy (HVEM) enables us to study specimens of several μm in thickness and offers decisive advantages over 100 kV TEM: i) generally, specimens can he prepared by thinning the substrate only, ii) thick foils are mechanically stable, which allows the preparation of specimens with an electron transparent area of several mm2.We report here new results on the initial formation of misfit dislocation structures in the heteroepitaxial system Ge on GaAs (f = 0.074%).

Electron microscopic observations on the structure of dislocations in interphase boundaries

The misfit dislocation structure of interface boundaries between α-brass (f.c.c. structure) and β-brass crystals (b.c.c. structure) was investigated by transmission electron microscopy to test the dislocation model of incoherent interfaces. It was observed that a variation of one degree or less of the misorientation between neighbouring α/β-brass crystals may result in the invisibility of the misfit dislocations. This observation seems to be at variance with the predictions of the misfit dislocation model and suggests a change of the misfit dislocation core structure in the sense that the width of the misfit dislocation cores increases so rapidly with increasing deviation from the low energy orientation relationship that the cores overlap and the misfit dislocation model may be of limited physical significance.

Lattice matching and dislocations in LPE in 1− xGa xP 1− zAs z—InP heterojunctions

The growth of lattice-matched In 1- x Ga x P 1- z As z-InP heterojunctions by constant-temperature liquid phase epitaxy has been examined. The LPE heterojunction growth is carried out at 640°C employing In-rich melts saturated at 650°C. The effect of lattice mismatch on heterojunction interfaces and on crystal surface morphology is reported. In particular, the characterization of the misfit dislocation structure resulting from lattice mismatch is described. Also described is the growth of multilayer In 1- x Ga x P 1- z As z heterojunction structures which permit further examination of the effects of lattice mismatch and disl ocation formation, and which demonstrate that complex lattice-matched In 1- x Ga x P 1- z As z heterojunction structures can be grown b y liquid phase epitaxy.

On the core structures of misfit dislocations in grain boundaries

Abstract The core widths of grain boundary misfit dislocations in [001] twist boundaries close to coincidence site lattice misorientations have been studied by transmission electron microscopy on thin-film bicrystals of gold. Matching of experimental dark-field weak-beam images with profiles computed using the dynamical theory of electron diffraction shows that the core width for the 36·9°, σ = 5 boundary is ∼ 5 nm and may be even greater for the 22·6°, σ = 13 and 28·1°, σ = 17 boundaries. The results suggest that the coincidence site lattice plus dislocation network model of high-angle grain-boundary structures is applicable only to small angular ranges of misorientation close to angles corresponding to coincidence site lattice boundaries with relatively low energies.

Transmission electron microscopy observations of misfit dislocations in GaAsP epitaxial films

Transmission electron microscopy observations have been made of misfit dislocation structures in GaAsP epitaxial films in foils both parallel to (1) the interface between the epitaxial film and the substrate and (2) the {1 1 1} glide planes. These observations support a near surface source mechanism of dislocation multiplication for relief of the epilayer misfit. It is also suggested that the recently observed surface reconstruction in the III-V compounds might allow for an easier nucleation of dislocations at the surface than hitherto thought. Furthermore, an efficient Lomer dislocation has been observed forming from two 60° glide dislocations thus supporting the hypothesis that all dislocations found in these foils, including the sessile Lomer type, originate from a glide process.

Growth mechanisms of AuCu II plates

The growth mechanisms of AuCu II plates are considered upon the basis of the measurements of their lengthening and thickening rates as reported by Pedraza and Kittl ( Acta Met. 24, 835, 1976). These rates are 3–4 orders of magnitude smaller than those estimated for incoherent AuCu II: disordered matrix (interphase) boundaries. During growth, the interphase boundaries are deduced to be coherent. At the edges of plates, coherency is maintained because the succession of ledges is too rapid to give the misfit dislocation structure time to develop. At the broad faces of plates, the closely spaced twins (initially, ca. 50 Å apart) in the AuCu II greatly diminish the driving force for introduction of misfit dislocations until substantial coarsening of the twinned structure has occurred. Ledge growth at the broad faces is inhibited by the requirement that such growth occur parallel to the twins; this effect appears responsible for the order of magnitude larger spacing between ledges at the broad faces than at the edges of the plates and for the growth of AuCu II crystals as plates rather than as idiomorphs. The principal features of AuCu II growth were shown to be consistent with a massive rather than a martensitic mode of transformation.

Kinetics of growth of platelike precipitates

In order to resolve existing difficulties in the interpretation of the lengthening and thickening kinetics of platelike precipitates, these kinetics have been measured on θ' and η plates in Al-4 % Cu alloy and Al-0.2% Au alloy respectively, in conjunction with extensive observations of the ledge and misfit dislocation structures at the broad faces of the plates. Thickening has been found, in agreement with a well established theory of morphology, to occur by a ledge mechanism, where the migration rates of the ledges were diffusion controlled. Surprisingly, the lengthening rates of coherent plates were found to be much greater than those of the same kinds of plates when semicoherent. In the latter condition, lengthening rates approximate those to be expected on the basis of volume diffusion control. This result is explained by the role of the broad face coherency strain field which accelerates the lengthening. Thus a previous suggestion that such accelerated kinetics are due to diffusional short circuit is shown to be invalid.

The role of intruder dislocations in modifying the misfit dislocation structures and growth kinetics of precipitate plates

Intruder dislocations formed at θ’ and η plates in Al-4 pet Cu and Al-0.2 pct Au alloys respectively by a small plastic strain partially compensate the misfit by single arrays ; the Burgers vector of the dislocations has a component normal to the plate interfaces. On subsequently aging such deformed microstructures, little change takes place in θ’ where the misfit is low. In η, on the other hand, the large misfit is sufficient to nucleate other compensatory arrays which interact with the intruder dislocations to form the lowest energy dislocation network and to annihilate the Burgers vector component directed normal to the plates. The lengthening kinetics of θ’ plates are unaffected by the intruder dislocations, but the thickening kinetics are briefly accelerated, probably by means of vacancy-enhanced diffusion associated with the plastic deformation. The thickening enhancement later falls off as the defects are annealed. In Al-Au, an interesting morphological instability develops and leads to the formation of elongated plates. These we believe are caused by a mechanism of sympathetic nucleation.