Hollow defects

Derivation of growth mechanism of nano-defects in GaN from TEM data

Transmission electron microscopy (TEM) has been used to describe 'hollow' defects, e.g. nanotubes and pinholes (empty inside) formed in GaN heteroand homoepitaxial layers. These defects are believed to be formed due to the presence of impurities and dopants. These defects appear to be even more deleterious than dislocations for the application of GaN in devices because their formation interrupts two-dimensional growth of quantum wells, and therefore will influence electrical and optical properties of the material.

Experimental observations show that both these defects start from V-shaped indentations delineated by (10ll) planes, and their density increases with higher concentrations of impurities or dopants. These 'hollow' defects are often found to be attached to dislocations, but the frequently observed termination of these defects inside the crystal eliminates dislocations as the source of these defects. Theoretical predictions also do not suggest that dislocations would have large-diameter empty cores like these of nanotubes and pinholes.

.....for epitaxial growth, the lack of lattice-matched substrates and differences in thermal expansion coefficient between the nitrides and the substrates hampers good structural quality of nitride films. Large high-quality crystals of ni-V nitrides have not yet been obtained. Bulk GaN crystals, grown from a Ga melt at a temperature range 1500-1800 K under high (-15 kbars) nitrogen hydrostatic pressure have been grown in the form of platelets with a diameter close to 1 cm, but have not yet been used in devices. Most epitaxially grown GaN films use substrates which have lattice and thermal mismatch (-14% lattice mismatched for sapphire and -3.5% for SiC). As a consequence, a high density of defects is observed in heteroepitaxially grown crystals. These heteroepitaxial films start to grow three dimensionally.

.....However, for long-lifetime lasers, a lateral overgrowth technique has to be applied, which results in lower defect density. This is because higher dislocation density decreases carrier mobility. Besides dislocations, there are other defects, e.g. inversion domains (IDs), nanotubes and pinholes, which do not appear to be caused by lattice or thermal mismatch as they are also observed in homoepitaxial layers. However, they are probably more harmful than dislocations. It will be shown that these defects are caused by impurities or dopants which need to be introduced into crystals to obtain specific optical (or electrical) properties.

....Because the shape of pinholes and nanotubes is similar in plan-view configuration, therefore, some defects that have been assumed to be 'nanotubes' may not be. In order to give these defects proper interpretation one needs to understand when and where these defects are formed. The presence of both these defects is not limited to a particular growth method or substrate. They are present in all heteroepitaxial and homoepitaxial layers, but are not observed in bulk crystals grown from a Ga melt or laterally overgrown layers except at the place where two overgrown layers meet. A high density of nailtype defects, e.g. nanotubes, is observed in GaN grown on sapphire just above the buffer layer. Usually they are attached to the dislocation half-loops formed there.

.....This inversion domain and the dislocations were independently originated at the substrate interface, but with subsequent growth the dislocations terminated on the facet of the pinhole. The dislocations then tend to bend toward the hole during further layer growth to shorten their length and, therefore, reduce the total free energy of the system. Diffraction contrast used for the characterization of these dislocations showed that they were of different types, the lower dislocation closer to the inversion domain had a near screw character, while the upper one was lying at a larger angle to its Burgers vector. This observation clearly shows that the dislocations did not originate this pinhole but were later attracted to it. However, if this pinhole had been studied in plan-view configuration, especially if a thin TEM foil was prepared from the upper pan of the sample, the results would show that dislocations were attached to the pinhole. In this way a wrong conclusion could be drawn that dislocations initiated formation of the pinhole, which was not the case. This is a very important observation because in heteroepitaxial layers growth always starts threedimensionally and the number of dislocations is high. Any dislocation near the island edge has an image force which may cause the dislocation to bend.

Center for advance materials, 2000, oxfordjournals.org




Point defects govern the electrochemical properties

Point defects largely govern the electrochemical properties of oxides: at low defect concentrations, conductivity increases with concentration; however, at higher concentrations, defect–defect interactions start to dominate1, 2. Thus, in searching for electrochemically active materials for fuel cell anodes, high defect concentration is generally avoided. Here we describe an oxide anode formed from lanthanum-substituted strontium titanate (La-SrTiO3) in which we control the oxygen stoichiometry in order to break down the extended defect intergrowth regions and create phases with considerable disordered oxygen defects. We substitute Ti in these phases with Ga and Mn to induce redox activity and allow more flexible coordination. The material demonstrates impressive fuel cell performance using wet hydrogen at 950 °C....

Disruption of extended defects in solid oxide fuel cell anodes for methane oxidation, Nature 439, 568-571, 2006