A detailed analysis of the microstructure in B 12 As 2 epitaxial layers grown by chemical-vapor deposition on ͑0001͒ 6H-SiC substrates is presented. Synchrotron white beam x-ray topography enabled macroscopic characterization of the substrate/epilayer ensembles and revealed the presence of a quite homogeneous solid solution of twin and matrix epilayer domains forming a submicron mosaic structure. The basic epitaxial relationship was found to be ͑0001͒ B 12 As 2 ͗1120͘ B 12 As 2 ʈ ͑0001͒ 6H-SiC ͗1120͘ 6H-SiC and the twin relationship comprised a 180°͑or equivalently 60°͒ rotation about ͓0001͔ B 12 As 2 in agreement with previous reports. Cross-sectional high resolution transmission electron microscopy revealed the presence of a ϳ200 nm thick disordered transition layer which was shown to be created by the coalescence of a mosaic of translationally and rotationally variant domains nucleated at various types of nucleation sites available on the ͑0001͒ 6H-SiC surface. In this transition layer, competition between the growth of the various domains is mediated in part by the energy of the boundaries created between them as they coalesce. Boundaries between translationally variant domains are shown to have unfavorable bonding configurations and hence high-energy. These high-energy boundaries can be eliminated during mutual overgrowth by the generation of a 1 / 3͓0001͔ B 12 As 2 Frank partial dislocation which effectively eliminates the translational variants. This leads to an overall improvement in film quality beyond thicknesses of ϳ200 nm as the translational variants grow out leaving only the twin variants. ͑0003͒ twin boundaries in the regions beyond 200 nm are shown to possess fault vectors such as 1 / 6͓1100͔ B 12 As 2 , which are shown to originate from the mutual shift between the nucleation sites of the respective domains.

No abstract

No abstract

No abstract

Development of accurate models of complex clinical time series data is critical for understanding the disease, its dynamics, and subsequently patient management and clinical decision making. Clinical time series differ from other time series applications mainly in that observations are often missing and made at irregular time intervals. In this work, we propose and test a new probabilistic approach for modeling clinical time series data that is optimized to handle irregularly sampled observations. Our model is defined by a sequence of Gaussian processes (GPs), each restricted to a window of a finite size, where dependencies among two consecutive Gaussian processes are represented using a linear dynamical system. We develop algorithms supporting both model learning and inference. Experiments on real-world clinical time series data show that our model is better for modeling clinical time series and that it outperforms or is close to alternative time series prediction models.