Twinning Structure of {113} Defects in High-Dose Oxygen Implanted Silicon-on-Insulator Material.
- PDF / 2,392,599 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 49 Downloads / 172 Views
TWINNING STRUCTURE OF {113} DEFECTS IN HIGH-DOSE OXYGEN IMPLANTED SILICON-ON-INSULATOR MATERIAL. S. Visitserngtrakul, J. Barry*, and S. Krause Department of Chemical, Bio and Material Engineering, Arizona State University, Tempe, AZ 85387 *Electron Microscope Centre, University of Queensland, St. Lucia, Brisbane, Queensland 4067, Australia.
ABSTRACT Conventional and high resolution electron microscopy (HREM) were used to study the structure of the ( 113) defects in high-dose oxygen implanted silicon. The defects are created with a density of 1011 cm- 2 below the buried oxide layer in the substrate region. The ( 113) defects are similar to the ribbon-like defects in bulk silicon. Our HREM observations show that two crystalline phases are present in the defect. Portions of the defects exhibit the original cubic diamond structure which is twinned across 1115) planes. The atomic model shows that the {115) interface is a coherent interface with alternating five- and seven-membered rings and no dangling bonds.
INTRODUCTION Silicon-on-insulator structures can be produced by implantation of a high dose of oxygen to form a buried oxide layer below a thin, top silicon layer -SIMOX (Separation by Ihplantation of QXygen). The implantation generates defects not only in the top silicon layer, but also in the silicon substrate beneath the buried oxide layer. It has been previously reported that the defects present in the substrate region of as-implanted SIMOX are ( 1131 defects and stacking faults [1-4]. The ( 113) defects are very similar in shape to the oxygen-induced ribbon-like defects (RLDs) in the bulk silicon which have been extensively studied in the last decade [5]. Based on computer simulations of high resolution electron microscopy (HREM) images, the ( 113) defects were initially identified as ribbon of coesite, a high pressure phase of silicon oxide [6,71. Recently, however, evidence was presented that the defects consist of hexagonal silicon and are thus formed by self-interstitial precipitation [5,8,9]. A purposed nucleation and growth mechanism for the RLD is a pure silicon-interstitial agglomeration as described by Salisbury and Loretto [10] or Tan et al. [11]. More recently, Pirouz et al. [12,13], who had observed the cubic-hexagonal transformation in hot-indented silicon, discussed that hexagonal silicon formation by aggregation of silicon self-interstitials in RLD is unlikely because the hexagonal phase in silicon is not thermodynamically favored. Also, the orientation relationship of (01 1)cu # 2 (000)he, [01 1]cu I [1 0]he is unexpected. It would be preferential to have the more natural orientation relationship of (11 )cu // (0001)he, [1 i0]cu //[I 120 ]he. In SIMOX, the (113) defects were first discussed by van Ommen et al [1]. They analyzed the diffraction patterns of the substrate region in SIMOX and suggested that the ( 113) defect is a platelike precipitate. of monoclinic silica phase coesite oriented with (010) coesite // (110) silicon. De Veirman et al. [2] also identified 1113) defects as coesite by
Data Loading...
