Bulletin of the American Physical Society
2007 APS March Meeting
Volume 52, Number 1
Monday–Friday, March 5–9, 2007; Denver, Colorado
Session U40: Semiconductors: Defects and Impurities |
Hide Abstracts |
Sponsoring Units: FIAP Chair: Chris G. Van de Walle, University of California, Santa Barbara Room: Colorado Convention Center 503 |
Thursday, March 8, 2007 8:00AM - 8:12AM |
U40.00001: Substitutional hydrogen as a cause of the unintentional n-type conductivity in ZnO. Anderson Janotti, Chris G. Van de Walle ZnO is a wide-band-gap semiconductor that is very promising for optoelectronic device applications. Serious challenges remain, however, and controlling the electrical conductivity is the most fundamental among these. As-grown ZnO is n-type and the cause of this n-type conductivity is still widely debated. Oxygen vacancies have long been invoked as the source of the n-type conductivity. This assignment is based on the assumption that oxygen vacancies are shallow donors, and on the fact that the conductivity changes inversely with the oxygen partial pressure. Contrary to the conventional wisdom, we have recently established that oxygen vacancies are not shallow but deep donors and have high formation energies in n-type ZnO. Therefore, they cannot cause conductivity. Based on first-principles calculations we show that SUBSTITUTIONAL hydrogen is a likely cause of the unintentional conductivity in ZnO. Hydrogen on an oxygen site forms a multicenter bond. Substitutional hydrogen is a shallow donor, and has a migration barrier of 2.5 eV, explaining recent observations of a hydrogen-related donor that is stable up to 500 $^{o}$C. [Preview Abstract] |
Thursday, March 8, 2007 8:12AM - 8:24AM |
U40.00002: Accuracy of defect formation energies from first-principles calculations: A study of vacancies in ZnO Yiyang Sun Formation energy of point defects determines the concentration of the defects at thermal equilibrium. Accurately calculated formation energies can be used to pinpoint the type of the dominant defects at specific experimental conditions. First-principles calculations based on the density-functional theory and the supercell modeling of the point defects are the state-of-the-art approach to obtaining the formation energies. In this study, taking the O and Zn vacancies in ZnO as examples, we investigate the effects of supercell size on the accuracy of calculated formation energies. It is found that the 5x5x3 supercell is sufficient to reach the accuracy of about 0.06 eV in the formation energies, which means that the defect concentration can be evaluated with an accuracy in the magnitude of 10, considering k$_{B}$T is about 0.026 eV at room temperature. For charged defects, the shift of valence band maximum (VBM) is an additional factor which affects the accuracy of calculated formation energies. It is found that the 5x5x3 supercell is able to obtain the VBM shift at an accuracy of about 0.02 eV. [Preview Abstract] |
Thursday, March 8, 2007 8:24AM - 8:36AM |
U40.00003: The shallow donor wavefunction in Si: Corrections to the KL effective mass theory (EMT) Theodore Castner The ENDOR data of Hale \& Mieher$^{1}$ (HM) provides detailed information on $\psi^{*}$$\psi$(R$_{nnm}$) at nearly 25 lattice sites for P, As, \& Sb. Ivey \& Mieher$^{2}$(IM) have given the most comprehensive calculation of $\psi$(r)= $\Sigma$A(k)u$_{k}$(r)e$^{ik.r}$ featuring a complex A(k) (and u$_{k}$(r)) and higher conduction bands. IM could identify most of the sites and reduce the rms error between theory values and experimental results from 60\% to 11\%. However, the IM results are poor for the (1,1,1) site [shell E] and don't provide clear evidence for subsidiary minima$^{3}$ (L$_{1}$,$\Gamma$$_{2}$$^ {'}$) from their region IV in the BZ. A reliable calculation of matrix elements $<$L$_{1}$[U(r)]$\Delta$$_{1,min}$$>$ is difficult because of the complicated core potential in the central cell. Using the equidistant matched lattice pair data [(3,3,3) \& (1,1,5); shells C and Q] provides a good estimate of the \% admixture from the L$_{1}$ minima, somewhat smaller than in [3]. The IM ImA(k) and the L$_{1}$ minima both provide corrections to the uniaxial strain i$_{d}$ parameter$^ {4}$. A data analysis for the odd lattice sites improves the agreement between theory and experiment. Important remaining theoretical issues will be discussed and new ENDOR experiments will be proposed. The corrections to EMT are important, but are smaller than implied in IM. 1) E.B. Hale \& R.M. Mieher, Phys.Rev.184, 739, 751 (1969). 2) J.L. Ivey \& R.M. Mieher, Phys.Rev.B11, 822 (1975). 3) T.G. Castner, Phys.Rev.B2, 4911 (1970). 4) E.B. Hale \& T.G. Castner, Phys.Rev.B1, 4763 (1970). [Preview Abstract] |
Thursday, March 8, 2007 8:36AM - 8:48AM |
U40.00004: Vacancy--interstitial interactions in crystalline Silicon Matthew J. Beck, L. Tsetseris, S.T. Pantelides Extensive experimental and theoretical investigations of fundamental defects in Si have led to the conclusion that both interstitials and vacancies diffuse athermally according to a carrier recombination-enhanced Bourgoin-Corbett mechanism. It is therefore widely accepted that Si vacancy–-interstitial pairs, or Frenkel pairs (FPs), either rapidly recombine or dissociate, even at cryogenic temperatures. This has recently been challenged by X-ray scattering experiments that are interpreted to show FPs in Si persisting to $> 100$ K (Partyka, P., et al., \emph{Phys. Rev. B} v. 64 art. no. 235207). Here we report first-principles density functional theory calculations of the properties of FPs in Si. A novel electronic interaction involving charge transfer from the interstitial to the vacancy in a FP is described. In a bound FP, this interaction suppresses the athermal diffusion mechanisms of both the interstitial and vacancy. This reconciles the existing experimental results by establishing that there are thermal barriers to the recombination or dissociation of bound FPs, but not affecting the previously described athermal diffusion of isolated interstitials and vacancies. This work supported by the AFOSR-MURI program. [Preview Abstract] |
Thursday, March 8, 2007 8:48AM - 9:00AM |
U40.00005: First-Principles Study of Er Location in Er-Si Systems with Oxygen Co-Dopants R.H. Pink, Junho Jeong, Dip N. Mahato, M.B. Huang, T.P. Das, R.H. Scheicher, Sitaram Byahut Using the Hartree-Fock cluster procedure, we are investigating possible models for the Er$^{3+}$ ion in silicon with oxygen co-dopants[1]. We are examining first the Hi (hexagonal interstitial) site with six O atoms in the intrabond regions of the six Si-Si bonds for this center[2], allowing for relaxation in positions of the O and Si atoms. The aim of this study is to see if the presence of the O atoms is indeed able to change the situation of a maximum in the potential surface for Er$^{3+}$ found from our recent investigations of the Er-Si system without co-dopants to a minimum in the co-doped Oxygen system. Results for the Er$^{3+}$ potential curve for the Hi center and the geometries of the Si and O atoms will be presented. [1] F. d'Acapito, S. Mobilio, S. Scalese, A. Terrasi, G. Franz\'{o} and F. Priolo, Phys. Rev. B 69, 153310 (2004) and references therein. [2] J.D. Carey, J. Phys. Condens. Matter 14, 8537 (2002) [Preview Abstract] |
Thursday, March 8, 2007 9:00AM - 9:12AM |
U40.00006: \textit{Ab-initio }Study of the Diffusion Mechanisms of Gallium in a Silicon Matrix Kevin Levasseur-Smith, Normand Mousseau We present the results of a study into the diffusion mechanisms of Ga defects in crystalline Si. The dominant neutral configurations for single and multi-atom defects are established by \textit{ab-initio} calculations using the density functional theory in the LDA approximation, with a LCAO basis as implemented in the SIESTA package. We find formation energies of 0.7 eV and 2.9 eV, respectively, for the substitutional and tetrahedral interstitial defects, while the diatomic substitutional-tetrahedral complex has a formation energy of 2.2 eV. Subsequent calculations using this same DFT package in conjunction with the activation relaxation technique (ART nouveau) allow us to determine possible diffusion pathways as well as their corresponding saddle points and energy barriers. [Preview Abstract] |
Thursday, March 8, 2007 9:12AM - 9:24AM |
U40.00007: Surface effects in Si interstitial formation energies Ann E. Mattsson, Ryan R. Wixom, Rickard Armiento We are calculating Si self-interstitial formation energies using Density Functional Theory and several different exchange-correlation energy functionals. We show that the difference in results obtained with the LDA, PBE, PW91, and AM05 [1] functionals can be explained by the functionals' different surface intrinsic errors. We explain why surface effects are important for formation energies of interstitials in semi-conductors. Surface effects have previously been studied for metal vacancy formation energies. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000. [1] R. Armiento and A. E. Mattsson, Phys. Rev. B {\bf 72}, 085108 (2005). [Preview Abstract] |
Thursday, March 8, 2007 9:24AM - 9:36AM |
U40.00008: Confining P Diffusion in Si by an As-doped Barrier Layer Lugang Bai, Guang-Hong Lu, Feng Liu, Qi Wang, Hamza Yilmaz We investigate the effect of As-doping on P diffusion in Si, using first-principles total-energy calculations. The formation of As-vacancy complex is found to be energetically favorable, which indicates the consumption of vacancy by As to prohibit vacancy-mediated P diffusion. In the vicinity of As-vacancy complex, the diffusion barrier of P via exchanging with Si vacancy is considerably increased, which further decreases the P mobility. These results qualitatively explain the experimental observations and provide useful direction for designing As-doped diffusion barriers for confining P diffusion in Si wafer processing and in MOSFET device fabrication. [Preview Abstract] |
Thursday, March 8, 2007 9:36AM - 9:48AM |
U40.00009: Genuine impurity states vs. perturbed host states in a supercell approach Yong Zhang, A. Mascarenhas, L.-W. Wang The spatial localization is the most basic feature of an impurity state(IS) in a semiconductor. However, in the often used supercell approach, this feature might not be practically reliable, especially when the IS's are in resonance with the host state. We have developed a systematic approach for identifying the genuine IS's amongst the large number of states given by the supercell calculation[1], with the help of a charge-patching method that allows us to perform the electronic structure calculation with an accuracy near that of a self-consistent DFT and the ability handling $>$10K atoms[2]. We have applied this approach for two prototype systems: (1) isoelectronic ``acceptor'' in GaP:N, for which we have shown the non-existence of any excited state of N impurity in the conduction band within at least a few hundred meV proximity of the N bound state[1], in contrast to the decades old speculation and recent claim of the existence of such excited states; (2) isoelectronic ``donor'' in GaAs:Bi, for which we have found a resonant IS generated by Bi within the valence band (VB) and the topmost VB state is that of GaAs perturbed by Bi[3], which reveals a strong enhancement in the spin-orbit splitting (confirmed experimentally[4]) and resolves the speculations whether or not Bi could have a bound state in GaAs. [1] Zhang et al.,PRB\textbf{74},R41201 (06). [2] Wang, PRL \textbf{88},256402 (02). [3] Zhang et al., PRB\textbf{71},155201 (2005). [4] Fluegel et al., PRL\textbf{97},67205 (06). DOE/BES [Preview Abstract] |
Thursday, March 8, 2007 9:48AM - 10:00AM |
U40.00010: Origins of transparent conductivity in SnO$_{2}$ Abhishek Singh, Anderson Janotti, Chris Van de Walle, Matthias Scheffler SnO$_{2}$ belongs to a small class of materials that can conduct electricity while remaining transparent to visible light. Along with ITO it is widely used for contacts in flat-panel displays, solar cells, light emitting diodes and other optoelectronic applications. The origin of the observed unintentional $n$-type conductivity in SnO$_{2}$ is generally attributed to oxygen vacancies, $V_{\rm O}$. Using density functional calculations along with the LDA+$U$ method we show that $V_{\rm O}$ is not a shallow donor, but has a deep $\epsilon$(2+/0) level at $\sim$2.0 eV below the conduction band. The Sn interstitial is a shallow donor; however, its formation energy is very high making it very unlikely to be incorporated. Instead of point defects we propose that impurities, and in particular hydrogen, are responsible for the observed conductivity. We find that the hydrogen interstitial H$_{\rm i}$ has low formation energy and acts as a shallow donor in SnO$_{2}$. However, its migration barriers are small and therefore it is not stable at elevated temperatures. {\it Substitutional} hydrogen (H$_{\rm O}$), also acts as a shallow donor, and is more stable. Unlike H$_{\rm i}$ the formation energy of H$_{\rm O}$ depends on the abundance of oxygen and hence explains the experimentally observed dependence of conductivity on oxygen partial pressure in SnO$_{2}$. [Preview Abstract] |
Thursday, March 8, 2007 10:00AM - 10:12AM |
U40.00011: Localized Vibrational Modes of $\mathrm{O_{Te}}$ and ($\mathrm{O_{Te}-V_{Cd}}$) Centers in CdTe: Fundamentals and Second Harmonics$^{*}$ Gang Chen, I. Miotkowski, S. Rodriguez, A. K. Ramdas In CdTe, grown with excess Cd vacancies ($\mathrm{V_{Cd}}$), oxygen replacing Te ($\mathrm{O_{Te}}$) displays a pair of fundamental localized vibrational modes (LVMs), $\nu_{1} = 1096.78$ cm$^{-1}$ and $\nu_{2} = 1108.35$ cm$^{-1}$. They are ascribed to the non-degenerate $\Gamma_{1}$ ($\nu_{1}$) and the doubly degenerate $\Gamma_{3}$ ($\nu_{2}$) LVMs of ($\mathrm{O_{Te}-V_{Cd}}$) centers with nearest neighbor Cd missing, having $C_{3v}$ symmetry and $\hat{\bf{c}}$ axis along $\langle 111\rangle$. In CdTe grown with conditions suppressing $\mathrm{V_{Cd}}$, $\mathrm{O_{Te}}$ occurs with all the four Cd nearest neighbors, and exhibits a triply degenerate $\Gamma_{5}$ LVM at $\nu_{0} = 349.79$ cm$^{-1}$ of $T_{d}$ symmetry.[1] The harmonics of ($\mathrm{O_{Te}-V_{Cd}}$), i.e., of $\nu_{1}$ and $\nu_{2}$ occur at $\nu_{4} = 2198.66$ cm$^{-1}$ and $\nu_{5} = 2210.5$ cm$^{-1}$. The temperature dependence of both ($\nu_{1}$, $\nu_{2}$) and ($\nu_{4}$, $\nu_{5}$) pairs display a remarkable behavior: $\nu_{1}$ and $\nu_{2}$ approach each other and coalesce at $T^{*} \sim 300$ K, as do $\nu_{4}$ and $\nu_{5}$; beyond $T^{*}$ they behave as a triply degenerate $\nu_{0}^{*}$ and $\nu_{s}^{*}$, respectively. The relative intensity of $\nu_{2}$ : $\nu_{1}$ approaches $2$ as $T \rightarrow T^{*}$ while that of $\nu_{5}$ : $\nu_{4}$ approaches $1/2$. These features find a convincing explanation on the basis of the dynamic switching of the ($\mathrm{O_{Te}-V_{Cd}}$) dangling bond among the four $\langle 111\rangle$ axes and, for $T \geq T^{*}$, these centers \textquotedblleft acquire\textquotedblright ~$T_{d}$ symmetry. With its $T_{d}$ symmetry, $\mathrm{O_{Te}}$ displays a single second harmonic $\nu_{s}$ at $695.72$ cm$^{-1}$. [1] Chen \textit{et al.}, Phys. Rev. Lett., \textbf{96}, 035508 (2006). $^{*}$Work supported by NSF (DMR 0405082) [Preview Abstract] |
Thursday, March 8, 2007 10:12AM - 10:24AM |
U40.00012: Theoretical study of the deep defect states in PbTe thin films Khang Hoang, S.D. Mahanti, P. Jena The nature of deep defect states (DDS) in bulk PbTe has been studied recently using density functional theory and a supercell model [1]. It was found that substitution of Pb by the trivalent impurities Ga, In, and Tl gave rise to hyperdeep defect states (HDS) below the valence band (VB) and DDS near the band gap region. Here we discuss how these states are affected in a (100) PbTe film using a supercell slab model. The HDS and DDS are preserved in the film geometry. As one goes from the bulk-like layers to subsurface and surface layers, the HDS tends to move closer to the bottom of the VB and becomes narrower; the DDS also gets modified. We also find that the defect formation energy E$_{f}$ as a function of the distance from the surface shows interesting features: all three impurities have lowest E$_{f}$ in the first layer but E$_{f}$ increases monotonically in the case of Ga, whereas there is a potential barrier in the second layer and a shallow potential ``valley'' between the second and the bulk-like layers in the case of In and Tl. This suggests that Ga impurities will be annealed out whereas the other two can be trapped in the subsurface region. [1] Salameh Ahmad, Khang Hoang, and S. D. Mahanti, Phys. Rev. Lett. 96, 056403 (2006). [Preview Abstract] |
Thursday, March 8, 2007 10:24AM - 10:36AM |
U40.00013: \textit{Ab Initio }Study of Electronic Structure of Defects in SnTe and GeTe. Salameh Ahmad, S.D. Mahanti \textit{Ab} \textit{initio} electronic structure calculations have been carried out within density functional theory (DFT) in SnTe and GeTe, two well-known narrow band-gap semiconductors, to understand the nature of deep defect states (DDS) introduced by Cd and In impurities substituting for Sn/Ge. These results are compared with similar studies in PbTe$^{1}$. The calculations have been carried out using a 64 atom super-cell model containing one defect. The density of states near the top of the valence band (VB) and the bottom of the conduction band (CB) get significantly modified by the defects as found in PbTe. The DDS associated with Indium impurity near the top of the VB is resonant in SnTe and lies in the gap in GeTe; its energy increasing in the order Sn-Pb-Ge. Cadmium on the other hand gives resonance (GeTe) and bound states (on SnTe) near the bottom of the CB, the energy of the DDS increasing from Sn-Pb-Ge. The positions of these DDS can have significant impact on thermoelectric and other transport properties of these semiconductors. 1. Salameh Ahmad, Hoang Khang, and S.D. Mahanti, Phys. Rev. Lett. 96, 056403 (2006); Salameh Ahmad et. al. Phys Rev. B74, 155205 (2006). [Preview Abstract] |
Thursday, March 8, 2007 10:36AM - 10:48AM |
U40.00014: Interfacial segregation and electrodiffusion of dopants in AlN/GaN superlattices P. Boguslawski, J. Bernholc, N. Gonzalez Szwacki Properties of semiconductors and their devices entirely rely on accurate control of doping and stoichiometry, determined in turn by solubility of impurities and the presence of native defects. In fact, it has been experimentally shown that thermally activated segregation of impurities through heterointerfaces leads to concentration differences that reach two orders of magnitude. We develop a first-principles theory of interfacial segregation in heterostructures and apply it to hydrogen and typical dopants and defects in GaN/AlN superlattices [1]. We show that the equilibrium concentrations of a dopant at two sides of an interface may differ by up to a few orders of magnitude depending on its identity and charge state, and are not directly given by dopant solubilities of bulk constituents. In addition, the presence of strong electric fields in GaN/AlN systems induces field-driven electromigration and accumulation of hydrogen at the appropriate interfaces. [1] P. Boguslawski, J. Bernholc, and N. Gonzalez Szwacki, Phys. Rev. Lett. 96, 185501 (2006). [Preview Abstract] |
Thursday, March 8, 2007 10:48AM - 11:00AM |
U40.00015: Effect of Native Defects in InN: Metallic bonding and N$_2$ Formation Xiangmei Duan, Catherine Stampfl Despite intense investigations, development of indium nitride technology remains at the stage of seeking to improve the growth techniques and fabrication of device-quality material. Doping and impurity control is essential for the advancement of InN-based electronic and optoelectronic devices, yet even the role of native point defects in InN, and their effect on the physical and electronic properties, is still lacking [1]. We have thoroughly investigated the distribution and electronic properties of native point defects in wurtzite InN through first-principles density- functional theory calculations. We find that both the nitrogen and indium vacancies have a tendency to form ``clusters'' in the neutral and negative charge states. For nitrogen vacancies, this results in local indium- rich regions with metallic-like bonds, while we find molecular nitrogen formation occurs either by the clustering of indium vacancies, or interstitial nitrogen. The formation energies of the indium vacancy clusters are however rather high, but that of interstial nitrogen in the 3+ charge state has a low formation energy under N-rich conditions. Our results help to explain a number of hitherto puzzling experimental observations[1-3]. [1] T. Shubina et al. Phys. Stat. Sol. a \textbf{202}, 377 (2005). [2] Y. Davydov et al. Phys. Stat. Sol. B \textbf{230}, R4(2002); \textbf{240}, 425 (2003). [3] H. Timmers et al. J. Cryst. Growth \textbf{288}, 236 (2006). [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700