There is growing interest in assessing the clinical value of ultrasound in the prediction and management of osteoporosis. However, the mechanism of ultrasound propagation in cancellous bone is not well understood. The Biot theory is one approach to modelling the interaction of sound waves with cancellous structure, and porosity is one of its input parameters. In this paper we report the relationship between broadband ultrasonic attenuation (BUA) corrected for specimen thickness (nBUA) and porosity in a porous Perspex cancellous bone mimic, a stereolithography cancellous bone mimic and in natural human and bovine tissue. nBUA and porosity have a non-linear parabolic relationship. The maximum nBUA value (nBUAmax) occurs at approximately 30% porosity in the Perspex mimic, approximately 70% in the stereolithography mimic and approximately 75% in natural cancellous bone. We discuss the effect of structure on the form of the nBUA-porosity relationship.

No abstract

There has been considerable debate on the relative dependence of broadband ultrasound attenuation (nBUA, dB MHz(-1) cm(-1)) upon the density and structure of cancellous bone. A nonlinear relationship between nBUA and porosity has recently been demonstrated using stereolithography models, indicating a high structural dependence for nBUA. We report here on the measurement of trabecular perimeter and fractal dimension on the two-dimensional images used to create the stereolithography models. Adjusted coefficients of determination (R2) with nBUA were 94.4% (p < 0.0001) and 98.4% (p < 0.0001) for trabecular perimeter and fractal dimension respectively. The feature of fractal dimension representing both the porosity and connectivity of a given structure is most exciting. Further work is required to determine the relationship between broadband ultrasound attenuation and fractal dimension in complex three-dimensional cancellous bone structures.

PACS 68.37. Hk, 78.55.Cr, 78.66.Fd Using wavelength dispersive X-ray spectrometers on an Electron Probe Micro-Analyser we have accurately quantified the elemental composition of a series of homogeneous AlInGaN epitaxial layers. The thickness of the quaternary layer (~100 nm) necessitates the combination of data measured at a number of different electron beam energies and an analytical model based on a layered structure. The samples studied have aluminium fractions in the range 0.03-0.12 and indium fractions in the region of 0.01. Photoluminescence data from the samples are used to plot the dependency of the luminescence energy, linewidth and intensity on the composition. WDX mapping was employed to investigate spatial variations in the elemental compositions and the films were found to be uniform with no evidence for clustering of In or Al on a >100 nm scale. IntroductionThe ternary nitride semiconductors InGaN and AlGaN have proved to be highly successful active layers in GaN-based devices, including light emitters and transistors. However, lattice mismatch in ternary heterostructures imposes serious limitations for the design and operation of certain devices. In addition the efficiency of InGaN light emitters is severely degraded by the incorporation of too much, or too little, indium. Use of quaternary AlInGaN layers offers potential for the fabrication of lattice matched III-N heterostructures and improvement of the quantum efficiency of light emitters [1,2], particularly in the UV region with its host of important applications. The growth of high quality AlInGaN layers requires a balance between the high temperatures suited for Al-containing layers and the low temperatures necessary for incorporation of indium. Measuring the composition of AlInGaN layers using standard X-ray diffraction is complicated by the lack of a unique solution for the quaternary system and further by overlapping diffraction peaks in the case of lattice-match to the underlying GaN. In this paper we describe the compositional analysis of a series of AlInGaN layers using an Electron Probe Micro-Analyser (EPMA) equipped with wavelength dispersive X-ray (WDX) spectrometers, allowing the elemental composition to be determined with high accuracy and sub-micron spatial resolution. The compositional data are then correlated with the light emitting properties of the layers.

In the absence of the use of prevention methods, AlGaN layers on GaN are known to form an array of cracks if a critical thickness is exceeded. In this study growth of AlGaN-GaN structures was carried out by metalorganic vapour phase epitaxy using sapphire substrates. Under optical and atomic force microscopy two distinct crack populations have been obseved. In thin, highly strained films an initial high density population of microcracks are found propagating from threading dislocations. This crack array extends as the thickness increases, before contracting with the onset of large cracks whick extend over many mm across the wafer. The formation of the microcrack array is very dependent on the stain in the epilayer and does not follow the classic critical thickness dependence with lattice mismatch. Avoiding stresses high enough to prevent the microcracking phenomenon may be critical in the use of highly tensile strained layers in nitride devices.1 Introduction AlGaN epilayers are critical to the formation of many III-nitride devices. These include laser diodes (LDs), heterojunction field effect transistors (HFETs), ultra-violet emitters, solar blind photodetectors and also intersubband devices. However, the lattice mismatch between AlN and GaN means that AlGaN layers grown on GaN experience a high degree of tensile stress which can lead to the formation of an array cracks over the sample surface [1][2][3]. These cracks, observable in an optical microscope on a ~ mm scale, have been reported to have a potentially dual role in the strain relaxation, through the formation of misfit dislocation on the basal plane [2] in addition to the relaxation caused by the crack itself. Such crack networks can be avoided through the use of a low temperature interlayer of AlN at the AlGaN-GaN interface for example [4], but these methods are not suitable in all cases and can cause an increased threading dislocation density.The above studies have discussed the propagation and relaxation induced by cracks, but none have reported on their nucleation or on the microscopic structural properties for layers below which the gross cracks have been observed. In a previous report [5] we have suggested that the gross cracks may be formed as a result of "microcracks" observed at much lower thicknesses. These microcracks are short, sub-µm in length an hence would not be observable by optical microscopy for example. In this paper evidence is shown that this microcracking phenomenon in thin highly strained AlGaN layers grown on GaN can occur in very thin layers indeed, and that their density rapidly decreases with the onset of "gross" crack formation.

PACS 68.65+k, 78.55.Cr, 78.66.Fd By means of a vertical low-pressure metalorganic chemical vapor deposition, high quality Al x1 In y1 Ga 1-x1-y1 N: barrier/Al x2 In y2 Ga 1-x2-y2 N: well (x1 > x2 and y1 < y2) multiple quantum well structures (MQWs) with the emission wavelengths at less than or equal to 350 nm have been successfully grown on sapphire substrates. The photoluminescence (PL) intensity at room temperature is dramatically enhanced by about two orders of magnitude compared to AlGaN/GaM MQWs with a similar emission wavelength. Furthermore, the PL intensity approaches that of InGaN/GaN MQWs, where the stimulated emission has been easily observed under optical pumping at room temperature. This means that AlInGaN MQW has potential to be used as the active region for ultra-violet light-emitting diode (LED) or laser diode (LD) with an emission wavelength down to 350 nm, and could have a similar performance to InGaN/GaN-based LEDs or LDs. The temperature dependant PL measurement indicates that there exists a strong exciton-localization-effect in AlInGaN MQW, which could be attributed to the greatly enhanced PL intensity. In addition, the influence of growth condition on the optical properties of AlInGaN are discussed, focusing on the effect of the growth pressure. 1 Introduction Ultra-violet light-emitting diode (LED) at wavelengths around 350 nm is becoming more and more attractive, since it has wide applications in many fields, in particular, such as environmental protection, medical devices as well as illumination [1][2][3]. However, to date, the quantum efficiency of UV-LEDs based on the GaN/AlGaN system on sapphire substrates is very low due to the poor crystal quality, since the efficiency is sensitive to the dislocations in AlGaN/GaN system. This is in remarked contrast to InGaN/GaN LEDs, where the efficiency is insensitive to the dislocations [2]. Currently, the optical power of the GaN/AlGaN-based UV-LED on sapphire substrate is very low, and of the order of tens µW at an injection current of 20 mA. Based on the epitaxial-lateral-overgrowth (ELOG) technology, the crystal quality can be greatly improved. However, the optical power of UV-LED based on ELOG technology is still only 0.6 mW at an injection current of 50 mA [4]. The high expensive cost of the ELOG technology makes it unsuitable for mass-production. Using GaN as a substrate, the crystal quality can be further improved, giving the highest reported optical power for a UV-LED of 0.55 mW at 20 mA injection current [5]. Clearly, it become a more serious problem than the ELOG technology due to even higher expensive cost. Indium is generally believed to cause the non-uniform indium composition in an InGaN active layer for violet to green LEDs, resulting in the insensitiveness of efficiency to the dislocations and then allowing the production of more than 20 mW violet or blue LEDs. Obviously, AlInGaN quaternary system is potentially a good candidate for the fabrication of high power LEDs with UV emission. Recently, AlInGaN quaternary based UV-LED ...

The tensile strain in AlGaN layers on GaN is well established to lead to cracking if a critical thickness is reached, unless measures such as interlayers are applied to prevent their formation. However in devices, such as HFETs such an approach is impractical. Growth of AlGaN-GaN structures was carried out by MOVPE using a standard two stage process for the growth of the GaN on sapphire. The crack structures were examined by optical and atomic force microscopy. Studies on thin AlGaN layers on GaN close to the crack critical thickness show the stress centres from which the cracks propagate are threading dislocations with cracks often initially forming to link together these stress centres if they are in close proximity. These cracks then extend and "lock" into the generally observed 0 1 1 2 direction in more highly strained layers. A macroscopically uniform crack array is observed in these thin AlGaN samples.

In Eq. (1) and in the text after Eq. (3), Ukk should be replaced by Ukk for consistency with Part I. The values used in the calculations were those given for U~ and Upp in Part I. Eq. 5, part II: