YY1: Defects in Semiconductors I
Session Chairs
Tuesday PM, April 26, 2011
Room 2014 (Moscone West)
9:30 AM - **YY1.1
Electronic Structure of O-vacancy in High-κ Dielectrics and Oxide Semiconductors.
Kee Joo Chang 1 , Byungki Ryu 1 , Hyeon-Kyun Noh 1 , Junhyeok Bang 2 , Eun-Ae Choi 3
1 Physics, Korea Advanced Institute of Science and Technology, Daejeon Korea (the Republic of), 2 Physics, Rensselaer Polytechnic Institute, Troy, New York, United States, 3 Memory Division, Samsung El. Co., Hwasung Korea (the Republic of)
Show AbstractHafnia (HfO2) and related alloys have been considered as promising high-κ dielectric materials for replacing SiO2 gate oxide in complementary metal-oxide-semiconductor (CMOS) devices with the spatial scale of sub-0.1 μm, because gate leakage current can be reduced by keeping the same effective oxide thickness. However, these materials still suffer from a high density of defects, especially O-vacancy, which cause degradation of devices, such as flat band voltage shifts, threshold voltage instability, and low carrier mobility. On the other hand, amorphous oxide semiconductors (AOSs) such as ZnSnO (ZTO) and InGaZnO (IGZO) have attracted much attention because of promising active channel materials for flexible transparent thin-film transistors (TFTs). AOSs have superior material properties, such as transparency for visible light, flexibility, and low-temperature deposition over large areas. Although the field-effect mobility was reported to be higher by an order of magnitude than that of commercially used hydrogenated amorphous Si, the device instability is still a challenging issue. Among native defects, O-vacancy is considered as a major cause of the degradation of devices. We perform first-principles calculations to investigate the defect properties of O-vacancies in high-κ gate oxides, amorphous oxide semiconductors, and at interfaces between AOSs and high-κ dielectrics. The role of O-vacancy in device performance is discussed by comparing the results of LDA+U, quasiparticle energy (GW), and hybrid density functional calculations.
10:00 AM - YY1.2
First Principle Studies of the Doping Properties of BiVO4.
Wanjian Yin 1 , Su-Huai Wei 1 , Mowafak Al-Jassim 1 , John Turner 1 , Yanfa Yan 1
1 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show Abstract Metal oxides are considered as potential candidates for PEC water splitting because of their resistance to oxidization and possible stability in aqueous solutions. Therefore searching for metal oxide semiconductors with appropriate bandgaps has been one of the research focus in the field of PEC water splitting . Among the large number of metal oxides proposed, BiVO4 has shown very promising performance. The monoclinic BiVO4 has a direct band gap of 2.4-2.5 eV, good optical absorption properties, and reasonable band edge alignment with respect to H/H2O and O/H2O redox potentials. For optimal solar to hydrogen conversion, the optical absorber layers (the utilized semiconductors) should have moderate p-type or n-type conductivities in order to minimize photo-generated carrier recombination. In this case, the doping properties are crucial for the application of oxides. The intrinsic and extrinsic doping properties of BiVO4, i.e., the transition energies and formation energies of defects and Fermi level pining positions, have been studied systematically by first-principle density functional theory. We find that for doping caused by intrinsic defects, O vacancies are shallow donors and Bi vacancies are shallow acceptors. However, these defects compensate each other and can only lead to moderate n-type and p-type conductivities at Bi-rich and O-rich growth conditions, respectively. To obtain BiVO4 with high n-type and p-type conductivities, which are required for forming Ohmic contacts, extrinsic doping using foreign impurities is necessary. Our results reveal that Sr, Ca, Na, and K atoms on Bi sites are very shallow acceptors and have rather low formation energies. The calculated Fermi level pinning positions predict that doping of these impurities under oxygen rich growth conditions should result in outstanding p-type conductivity. Substitutional Mo and W atoms on V sites are very shallow donors and have very low formation energies. Furthermore, the calculated Fermi level pinning positions suggest that the doping of Mo and W under oxygen poor growth conditions can produce excellent n-type conductivity. The trends of calculated formation energies and transition energies of the defects are discussed based on the atomic sizes and atomic chemical potentials of the dopants and host elements.
10:15 AM - YY1.3
Stable Complexes of Vacancies and Interstitial Group V Dopants in ZnO.
Brian Puchala 1 , Dane Morgan 1
1 Materials Science and Engineering, University of Wisconsin - Madison, Madison, Wisconsin, United States
Show AbstractWhile there have been increasing numbers of reports of ZnO light-emitting devices, the lack of reliable, reproducible p-type ZnO is still the major factor preventing widespread fabrication of ZnO optoelectronic devices. We report ab initio calculations of new stable configurations of group V dopant containing defect complexes in ZnO. We show that negatively charged Zn vacancies (VZn) in group V dopant containing complexes tend to repel each other due to Coulomb forces, as would be expected. However, we have found that in some cases, when VZn are first nearest-neighbors with each other, it becomes possible for a dopant to shift to an interstitial position between the VZn and significantly decrease the formation energy of the complex. Limiting the formation of these new vacancy-interstitial complexes may be important for reliably fabricating p-type ZnO because they are predicted to be deeper acceptors than the previously known complexes. More broadly, the identification and control of defect and dopant complexes in oxides is of fundamental importance in applications ranging from sensors to transparent conductors to optoelectronic devices. Due to the large computational expense of ab initio calculations and the great number of configurations possible with even a few defects comprising a complex, it is difficult for ab initio based searches for stable defect complexes to be exhaustive. The vacancy-interstitial complexes we have identified can be generalized to other compositions, dopants, and structures, and may help guide searches for stable defect complexes in other materials.
10:30 AM - YY1.4
Alkali Metal Doping in ZnO Studied by Hybrid Functional Calculations.
Yiyang Sun 1 , Shengbai Zhang 1
1 Department of Physics, Applied Physics and Astronomy, Rensselaer Polytechnic Institute, Troy, New York, United States
Show AbstractDoping ZnO by alkali metal elements, e.g., Li and Na, represents the first experimental attempts to make p-type ZnO, which has enormous applications in short-wavelength light-emitting devices. In these experiments, Li substitution for Zn (denoted by LiZn) was expected to produce a shallow acceptor. Unfortunately, it was found that the trapped hole at the LiZn defect had a binding energy of about 0.8 eV indicating that LiZn can only serve as a compensating defect in n-type samples, but cannot convert the samples from n-type to p-type. Contrary to the experimental findings, density functional theory (DFT) calculations based on the local density approximation (LDA) or generalized gradient approximations (GGA) suggest that LiZn is a shallow acceptor. This failure has been intuitively attributed to the self-interaction error (SIE) in the DFT calculations. The SIE makes the LDA or GGA favor delocalized defect states and consequently an undistorted defect structure. Experimental measurements, however, suggest that the LiZn defect undergo a significant distortion with the Li atom moving away from one of its four neighboring oxygen atoms (with an elongation of the corresponding Li-O bond length by 40%). Recently, hybrid functional method has been implemented for periodic systems. In the hybrid functional calculations, a portion of Hartree-Fock (HF) exchange energy is mixed into the DFT exchange energy. It has been demonstrated that the band gaps of semiconductor materials can now be accurately calculated by the hybrid functional method. Also, because of the inclusion of HF exchange, it is expected that the SIE will be reduced in the hybrid functional calculations. In this work, we study alkali metal doping in ZnO by hybrid functional calculations based on the PBE0 functional. This study is aimed at gaining two-fold understandings on this system: 1) the origin of the distortion observed in experiments, which is responsible for making the LiZn a deep level; 2) the possibility of doping ZnO to p-type by heavy alkali metal element doping (such as RbZn and CsZn), which is motivated by the experimental finding that NaZn produces a shallower acceptor than LiZn by about 0.2 eV.
10:45 AM - YY1.5
Origin of Device Instability in Amorphous Indium-gallium-zinc Oxide Thin Film Transistors.
Hyeon-Kyun Noh 1 , Byungki Ryu 1 , Eun-Ai Choi 2 , Kee Joo Chang 1
1 Department of Physics, Korea Advanced Institute of Science and Technology, Daejeon Korea (the Republic of), 2 Memory Division, Samsung El. Co, Hwasung Korea (the Republic of)
Show AbstractAmorphous oxide semiconductors such as amorphous indium-gallium-zinc oxide (a-IGZO) have attracted much attention because of their use as an active channel in transparent thin film transistors (TFTs). Recent experiments have shown that oxide TFTs exhibit very high field-effect mobilities, as compared to amorphous Si TFTs. Moreover, oxide devices have advantages such as transparency, flexibility, and large area uniformity at low growth temperature. However, device stability is sensitive to bias/current stress, temperature, light illumination, and environment contamination. Especially, instabilities of oxide TFTs are significant under negative bias illumination stress (NBIS), with negative shifts of the threshold voltage up to 18 V. Although the NBIS instability was suggested to result from the charge trapping of photoinduced holes, the details of charge traps are not clearly understood yet. In this work we perform first-principles density functional calculations to investigate the electronic properties of O-vacancy (VO) defects in a-IGZO and their role in the NBIS instability of oxide TFTs. We find that most of VO defects are deep donors, while some of them act as shallow donors due to severe outward relaxations of the neighboring metal ions. The VO defect behaves as a negative-U defect, similar to that of crystalline IGZO. From the band alignments between a-IGZO and various gate oxides such as HfO2, SiO2, and Al2O3, we find the valence band offsets to be less than 2 eV. Photoexcited holes drift easily toward the channel/dielectric interface due to small potential barriers and can be captured by VO defects in the gate oxide. It is suggested that VO acts as a hole trap and causes a negative shift of the threshold voltage in oxide TFTs.
11:30 AM - **YY1.6
Advances in Electronic Structure Methods for Defects and Impurities.
Chris Van de Walle 1 , John Lyons 1
1 , University of California, Santa Barbara, California, United States
Show AbstractFirst-principles calculations have made important contributions to the understanding of defects and impurities in semiconductors and insulators. Incorporation of impurities (“doping”) is essential for controlling electronic properties. Point defects often interfere; they have been suggested to lead to unintentional doping, and can act as compensating centers. Computations based on density functional theory (DFT) have been successful in elucidating experimental results, and in providing explanations and guidance in many cases where doping problems occur. Recently, a number of limitations of the approach have also been overcome. In particular, the “band-gap problem” inherent in DFT has traditionally made it difficult to interpret results relating to defect levels in the band gap; indeed, even the energetics of defects with occupied states in the gap can be affected. This problem becomes particularly severe for materials with small band gaps, such as Ge or InN, where the DFT band structure is metallic; as well as for materials with wide gaps, such as GaN, AlN, or ZnO, where the large energy range can lead to large errors. In this talk I will give an overview of the various approaches that have been applied to overcome the problems inherent in DFT. Hybrid functional have proven particularly successful. The approach will be illustrated with examples for ZnO [1], Ga2O3 [2], TiO2 [3], SiC [4], and GaN [5].Work performed in collaboration with C. Freysoldt, A. Janotti, G. Kresse, J. Lyons, J. Neugebauer, P. Rinke, M. Scheffler, J. Varley, and J. Weber.[1] J. L. Lyons, A. Janotti, and C. G. Van de Walle, Appl. Phys. Lett. 95, 252105 (2009). [2] J. B. Varley, J. R. Weber, A. Janotti, and C. G. Van de Walle, Appl. Phys. Lett. 97, 142106 (2010). [3] A. Janotti, J. B. Varley, P. Rinke, N. Umezawa, G. Kresse, and C. G. Van de Walle, Phys. Rev. B 81, 085212 (2010).[4] J. R. Weber, W. F. Koehl, J. B. Varley, A. Janotti, B. B. Buckley, C. G. Van de Walle, and D. D. Awschalom, Proc. Nat. Acad. Sci. 107, 8513 (2010).[5] J. L. Lyons, A. Janotti, and C. G. Van de Walle, Appl. Phys. Lett. 97, 152108 (2010).
12:00 PM - YY1.7
Time-dependent Density Functional Study on the Excitation Spectrum of Point Defects in Semiconductors.
Adam Gali 1 2
1 Research Institute for Solid State Physics and Optics, Hungarian Academy of Sciences, Budapest Hungary, 2 Department of Atomic Physics, Budapest University of Technology and Economics, Budapest Hungary
Show AbstractA common fingerprint of the electrically active point defects in semiconductors is the transition between their localized defect states upon excitation which may result in characteristic absorption or photoluminescence spectrum. While density functional calculations have been very successful in exploring the ground state propertieslike formation energies or hyperfine tensors the density functional theory, in principle, is not capable of providing reliable excitation spectrum. Time-dependent density functional theory, however, addressesthis issue which makes possible to study the properties of point defects associated with their excited states. In this paper we apply the time-dependent density functional theory on two characteristic examples: the well-known nitrogen-vacancy defect in diamond and theless known divacancy in silicon carbide. The former defect is a leading candidate in solid state quantum bit applications where detailed knowledge about the excitation spectrum is extremely important. The excitation property of divacancy will be also studied and its relevance in different applications will be discussed.
12:15 PM - YY1.8
Quantum Computing with Defects.
Joel Varley 1 , Justin Weber 1 , William Koehl 1 , Bob Buckley 1 , Anderson Janotti 2 , Chris Van de Walle 2 , David Awschalom 1
1 Physics, University of California, Santa Barbara, Santa Barbara, California, United States, 2 Materials, University of California, Santa Barbara, Santa Barbara, California, United States
Show AbstractA critical step in the development of a functional quantum computer is identifying and designing physical systems for use as qubits, the basic units of quantum information. Among the possibilities in the solid state, a defect in diamond known as the nitrogen-vacancy (NV-1) center stands out for its robustness, since it is an individually-addressable quantum system that can be initialized, manipulated, and measured with high fidelity at room temperature. While the great promise of the NV-1 stems largely from the defect’s nature as a deep center (a point defect with highly-localized electronic bound states confined to a region on the scale of a single lattice constant), no systematic effort has been made to identify other deep centers that might behave in a similar way. Recently we have developed guidelines for systematically identifying other deep center defect systems with similar quantum-mechanical properties [1]. This includes a list of physical criteria that these centers and their hosts should meet. Here we illustrate how these requirements can be used in conjunction with electronic structure theory to intelligently sort through candidate systems. To elucidate these points, we compare electronic structure calculations of the NV-1 center in diamond with those of several deep centers in 4H silicon carbide (SiC). Using novel hybrid functionals that do not suffer from the band-gap problem of conventional density-functional theory, we report defect formation energies, configuration-coordinate diagrams, and defect-level diagrams to compare and contrast the properties of these defects. We find that the direct structural analog of the NV center in diamond, the NC-VSi-1 center in SiC, may be a suitable center with very different optical transition energies. Our criteria also allow us to conclude that the silicon vacancy in SiC, in the absence of complexing with any impurity, could be a promising spin center as well, if it can be stabilized in either the -2 or the neutral charge state. The proposed criteria can be translated into guidelines to discover NV analogs in other tetrahedrally-coordinated materials as well.[1] J. R. Weber, W. F. Koehl, J. B. Varley, A. Janotti, B. B. Buckley, C. G. Van de Walle, and D. D. Awschalom, Proc. Nat. Acad. Sci. 107, 8513 (2010).This work was supported by ARO, AFOSR, and NSF.
12:30 PM - YY1.9
Oxygen Diffusion in SrTiO3.
Amit Samanta 1
1 Applied and Computational Mathematics, Princeton University, Princeton, New Jersey, United States
Show AbstractAnionic defects in strontium titanate (SrTiO3) play an important role in determining itsstructural and electronic properties. Using ab-initio calculations, we report the mechanism ofdiffusion of oxygen interstitials and vacancies in their different charge states in SrTiO3.Point defect concentrations in SrTiO3 are estimated from the laws of mass action of thegoverning equilibria. By studying the dominant defect reactions the equilibrium Fermi energyof the system is determined as a function of temperature and oxygen partial pressure. From theequilibrium defect concentrations a phase diagram of dominant defect species is developed. Thestability of the lattice at high defect concentrations is also discussed.
YY2: Method of Electronic Structure Calculations
Session Chairs
Tuesday PM, April 26, 2011
Room 2014 (Moscone West)
2:30 PM - **YY2.1
Calculating Semiconductor Properties Using Empirical and Ab Initio Computational Approaches.
Marvin Cohen 1
1 Physics, U of Calif, Berkeley, CA, California, United States
Show AbstractAfter some background discussion, I’ll focus on a few recent developments in the areas of theoretical studies of semiconductor electronic structure, photovoltaics, semiconducting boron nitride nanotubes, and the search for doped semiconductors with higher superconducting transition temperatures.
3:00 PM - YY2.2
Strain Effects and First-principles Determination of Deformation Potentials in Group-III Nitrides.
Qimin Yan 1 , Patrick Rinke 1 2 , Matthias Scheffler 1 2 , Chris Van de Walle 1
1 Materials Department, University of California, Santa Barbara, Santa Barbara, California, United States, 2 , Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin Germany
Show AbstractThe group-III-nitrides (AlN, GaN and InN) have already been commercialized as light emitting and laser diodes. Due to the large differences in lattice parameters between the substrates and the semiconductor epilayers, and between different ternary alloys (such as InGaN and AlGaN), strain is always present in optoelectronic devices. Strain plays a major role in the design of nitride-based devices, since it influences not only the energy of optical transitions but also the band structures near the valence-band maximum (VBM) and the conduction-band minimum (CBM). The effects of strain on the band structure of wurtzite semiconductors are characterized by deformation potentials. The experimental determination of deformation potentials is quite difficult, partly explaining why available experimental data for GaN are scattered over a very large range. For other nitrides experimental data is scarce. Previous theoretical studies have also produced widely differing values. In our work we employ advanced first-principles calculations based on density functional theory. The Heyd-Scuseria-Ernzerhof (HSE) hybrid functional and exact-exchange based G0W0 calculations are used to produce accurate band structures. Our computational studies show that the HSE method produces accurate lattice parameters and band gaps in good agreement with experimental data. Comparison with G0W0 results reveals that HSE also gives a good description of strain effects on band structures, while LDA and GGA underestimate the change of band gap under strain. We also observe a pronounced nonlinear dependence of band-energy differences on strain. We have generated a complete set of deformation potentials which can serve as essential and accurate input for device modeling using semiempirical methods such as k.p theory. This work was supported by the UCSB Solid State Lighting and Energy Center and by the Center for Energy Effcient Materials, an Energy Frontier Research Center funded by the U.S. DOE-BES under Award No. DE-SC0001009.
3:15 PM - YY2.3
Novel Method to Calculate Polarization Induced Interfacial Charges in GaN-AlN Heterostructures.
Rohan Mishra 1 , Oscar Restrepo 1 , Siddharth Rajan 2 1 , Wolfgang Windl 1
1 Materials Science and Engineering, The Ohio State University, Columbus, Ohio, United States, 2 Electrical and Computer Engineering, The Ohio State University, Columbus, Ohio, United States
Show AbstractInterfacial charges formed due to piezoelectric and spontaneous polarization in thin film heterostructures are traditionally calculated using bulk polarization constants of their individual layers. This naturally leads to the important question regarding the validity of this approximation when approaching “non-bulk” dimensions. As an alternative, we propose a new method to calculate polarization induced interfacial charges in semiconductor heterostructures using classical electrostatics applied to real-space band diagrams from first principles calculations. We use GaN/AlN heterostructures with ultrathin AlN layers (2-4 monolayers) as a test system. We show that the calculated values of electric fields and interfacial charges are independent of the choice of the exchange-correlation functionals used (LDA, PBE-GGA and HSE06) and are in excellent agreement with available experiments. We show that such a method is more convenient and reliable for the determination of interface charges and electric fields compared to a traditional atomic-charge analysis. For the system studied the calculated values of electric field and interfacial charges are in excellent agreement with those calculated using the bulk polarization constants thus rendering support to the use of bulk constants even for very thin films.
3:30 PM - **YY2.4
Correlation Effects in Transition Metal, Rare Earth, and Actinide Based Small gap Semiconductors: From GW to DMFT.
Gabriel Kotliar 1
1 Physics, Rutgers University, Piscataway, New Jersey, United States
Show AbstractStrong correlation effects are usually associated with metallic systems. However, there many instances in which small gap insulating materials containing transition metals, rare earths require methodologies based on many body theory. In this talk we will illustrate the problem of correlated insulators and their treatment with realistic electronic structure methods using several examples of useful materials.We will stress how correlation effects impact qualitatively different material properties and how to capture these effects in realistic electronic structure calculations. Ce2O3 is used in catalysis and pigments, and we have applied a combination of hybrid density functional theory and DMFT to describe their properties [1]. Plutonium chalcogenides, PuTe, PuSe, and PuS are small gap semiconductors, and we will present LDA+DMFT studies of their photoemission spectra[2]. Finally, correlation effects are important in thermoelectric materials such as SbFe AsFe and FeSi, and we will describe how we treat correlation effects in this systems, and how correlations affect the thermoelectric properties[3]. References:1- Combining the hybrid functional method with dynamicalmean-field theory. D. Jacob(a), K. Haule and G. KotliarEPL, 84 (2008) 57009 2- Valence fluctuations and quasiparticle multiplets in plutonium chalcogenides and pnictidesC. Yee G. Kotliar and K. Haule , PRB 81, 035105 _2010_3-Thermopower of correlated semiconductors: Application to FeAs2 and FeSb2 Jan M. Tomczak, K. Haule, T. Miyake, A. Georges, and G. Kotliar. Phys. Rev. B 82, 085104 (2010)
4:30 PM - **YY2.5
The Semiconductor Energy Gap : Temperature Dependence and Many-body Corrections.
Xavier Gonze 1
1 ETSF / IMCN, UCLouvain, Louvain-la-Neuve Belgium
Show AbstractThe energy bands of semiconductors exhibit significant shifts with temperature at constant volume. Formulas derived by Allen, Heine and Cardona [1-3] in a semi-empirical context cannot be transposed to Density Functional Theory or to Many-Body Perturbation Theory [4] without critical reexamination. Indeed, the rigid-ion approximation was appropriate in this semi-empirical context, while the complete formulation includes an extra term, the non-site-diagonal Debye-Waller term. The importance of this extra term was examined for diatomic molecules and found to be of a size similar to the standard Debye-Waller and Fan terms. This might explain the discrepancy found between previous theory and experiment for solids. Furthermore, the slow convergence of the sum over unoccupied states of the Allen-Heine-Cardona approach can be avoided in a new formalism proposed here, based on Density-Functional Perturbation Theory, leading to a dramatic decrease of calculation times.In the second part of the presentation, I will focus on GW many-body perturbation theory, for the accurate calculation of energy gaps. The extrapolar technique [5] will be shown to speed up dramatically such heavily CPU consuming calculations, for a variety of technological materials, including impurity-containing amorphous silica, phosphors for white-LED or transparent conducting oxydes.[1] P.B. Allen, V. Heine, J. Phys. C 9, 2305 (1976).[2] P.B. Allen, M. Cardona, Phys. Rev. B 24, 7479 (1981).[3] P.B. Allen, M. Cardona, Phys. Rev. B 27, 4760 (1983).[4] A. Marini, Phys. Rev. Lett. 101, 106405 (2008).[5] F. Bruneval and X. Gonze, Phys. Rev. B 78, 085125 (2008).
5:00 PM - **YY2.6
Theoretical Spectroscopy of Metal-insulator Transitions in Correlated Electron Materials.
Matteo Gatti 1
1 NanoBio Spectroscopy group, ETSF - UPV, San Sebastian Spain
Show AbstractMetal-insulator transitions (MIT) in strongly correlated materials are one of the key open issues in solid-state physics. MIT are due to the interplay between different interactions with comparable energy scales in such a way that strongly correlated materials are characterised by a great sensitivity to external parameters, like temperature, pressure, doping, etc. Solids in which electrons are strongly correlated display a broad range of very interesting properties, which make the possible technological applications from these materials very promising. On the other side, the ability to devise novel functionalities of devices for specific applications crucially depends on the microscopic understanding of the elementary electronic excitations.Here we use parameter-free approaches of many-body perturbation theory, namely the GW approximation [1] and the Bethe-Salpeter equation [2], to understand the effects of electronic correlation across the MIT in prototype transition-metal oxides. We simulate complementary electron spectroscopies: photoemission, inelastic X-ray scattering, and optics. We demonstrate that: i) our ab initio approaches provide a consistent interpretation of the MIT, not only for electronic states close to the Fermi energy, but also for states at higher binding energies; ii) excitonic and crystal local-field effects are important for the interpretation of optical spectra.[1] L. Hedin, Phys. Rev. 139, A796 (1965).[2] G. Onida, L. Reining, and A. Rubio, Rev. Mod. Phys. 74, 601 (2002).
5:30 PM - YY2.7
The Absorption of Diamondoids from Time-dependent Hybrid Density Functional Calculations.
Márton Voros 1 , Tamas Demjen 2 , Adam Gali 2 1
1 Atomic Physics, Budapest University of Technology and Economics, Budapest Hungary, 2 , Hungarian Academy of Sciences, Research Institute for Solid State Physics and Optics, Budapest Hungary
Show AbstractRecent technological developments allow the size and shape selectedpreparation of small diamond nanocrystals,i.e. diamondoids [1], that are the smallest building blocks ofdiamond. Since the characteristic size of thesediamondoids is in the range of nanometer, they possess severalinteresting properties that do not show up inthe case of bulk diamond. Recently, Lasse et al. measured theabsorption spectrum of several pieces of thediamondoid series [2]. We show that all-electron time-dependent DFT(TD-DFT) calculations includinghybrid functional in the TD-DFT kernel are able to providequantitatively accurate results [3]. Our calculationsdemonstrate that Rydberg transitions are characteristic even forrelatively large nanodiamonds resulting inlow optical gaps. Continuing our work on pristine diamondoids[3], we also investigate how impurities may affectthe optical properties of small diamondoids.[1] J. E. Dahl, S. G. Liu, and R. M. K. Carlson, Science 299, 96 (2003),[2] L. Landt et al., Phys. Rev. Lett. 103, 047402 (2009),[3] Márton Vörös and Adam Gali, Physical Review B 80, 161411(R) (2009).
5:45 PM - YY2.8
Second-harmonic Generation Spectroscopy from Time-dependent Density-functional Theory.
Eleonora Luppi 2 1 3 , Hannes Huebener 2 3 , Matteo Bertocchi 4 , Elena Degoli 4 , Stefano Ossicini 4 , Valerie Veniard 2 3
2 , Laboratoire des Solides Irradiés, Ecole Polytechnique , Palaiseau France, 1 Department of Chemistry, University of California, Berkeley, Berkeley, California, United States, 3 , European Theoretical Spectroscopy Facility , Palaiseau France, 4 Dipartimento di Fisica, Universita’ degli studi di Modena e Reggio Emilia, Modena Italy
Show AbstractNonlinear Optics is one of the most active fields for fundamental and applied research in physics. The interest for nonlinear optical phenomena is becoming extremely strong, because of their versatile and innovative properties and technological applications.Very recently we have developed a first-principles theory [1,2], based on the Time-Dependent Density-Functional Theory approach, for the calculation of the second-order susceptibility χ2. We find a general expression for χ2 valid for any fields, containing the ab initio relation between the microscopic and macroscopic formulation of the second-order responses. We consider the long wavelength limit and we develop our theory in the Time-Dependent Density-Functional Theory framework. This allows us to includestraightforwardly many-body effects such as crystal local-field and excitonic effects. We apply this formalism to the calculation of the Second-Harmonic Generation spectra for different type of materials: cubic semiconductors[1,2], hexagonal SiC polytypes finding good agreement with experiments. We are also exploring with this formalism the effects of anysotrophies for more complex systems like CaF2/Si multi quantum-well and silicon surfaces. [1] Eleonora Luppi, Hannes Hübener and Valérie Véniard, J. Chem.Phys. 132, 241104 (2010)[2] Eleonora Luppi, Hannes Hübener and Valérie Véniard, in press. Phys. Rev. B (2010)
YY3: Defects in Semiconductors II
Session Chairs
Wednesday AM, April 27, 2011
Room 2014 (Moscone West)
9:30 AM - **YY3.1
Defects in Complex Materials: Transparent Conducting Oxides and Chalcogenides.
Risto Nieminen 1
1 COMP/Applied Physics, Aalto University, Espoo Finland
Show AbstractI present computational results from density-functional-theory (DFT) calculations, using hybrid-functional methods, for atomic-scale defects in transparent conducting oxides (TCO) as well as solar-cell chalcogenides (CIGS). The methodology provides a robust description of the band-gap region of these materials, and is thus able to predict the positions of the ionisation levels of charged defects correctly. The results are used to discuss the origin of n-type conductivity in TCOs, their doping limits, as well as the luminescence spectra of CIGS.
10:00 AM - YY3.2
Hybrid Functional Studies of Oxygen-related Defects in Hafnia and Zirconia.
John Lyons 1 , Anderson Janotti 1 , Chris Van de Walle 1
1 Materials, University of California, Santa Barbara, Santa Barbara, California, United States
Show AbstractHafnia and zirconia are two leading candidates to replace the SiO2 gate dielectric in metal-oxide semiconductor (MOS) devices. Introduction of these new materials into the gate stack has proven challenging, however, as high-κ dielectrics have proven to be more susceptible to fixed-charge and charge-trapping problems than SiO2. Defects in the oxide are also known to lead to current leakage. Based on experimental evidence, and on the fact that these oxides are often deposited under O-rich conditions, oxygen-related defects are suspected to be the most likely culprits. In this work, we investigate native defects in these oxides with first-principles calculations based on density-functional theory, using the hybrid functional approach of Heyd, Scuseria, and Ernzerhof (HSE). By including a portion of exact exchange, the functional corrects the band gap of semiconductors and insulators, allowing for quantitative predictions of transition levels and formation energies of defects. Using this method, we have investigated the oxygen-related native defects in zirconia and hafnia, as well as their interactions with hydrogen. The calculations are performed for each of the phases of these materials, i.e., cubic, tetragonal, and monoclinic. We have also estimated band offsets between zirconia, hafnia, and silicon by using the universal alignment of the hydrogen interstitial level. The band alignment allows us to examine the charge states and transition levels of these oxides relative to the silicon band edges, enabling a determination of which defects are likely to act as traps or sources of fixed charge. This work was supported by the SRC under Task # 2009-VJ-1867.
10:15 AM - YY3.3
Intrinsic Point Defects in Cadmium Telluride Studied Using Hybrid Density-functional Theory Calculations.
Paul Erhart 1 , Daniel Aberg 1 , Vincenzo Lordi 1
1 Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, California, United States
Show AbstractThe electronic and structural properties of point defects in cadmium telluride and cadmium zinc telluride are investigated using density-functional theory. The calculations employ a hybrid exchange-correlation functional that yields a band gap in very good agreement with experiment. For CdTe, Cd interstitials and Te antisites are found to be the dominant defects under Cd and Te-rich conditions, respectively. Vacancies have larger formation energies and therefore occur in much smaller concentrations. They do, however, display a stronger propensity to induce deep levels that have a detrimental effect on carrier lifetimes. In general, our hybrid functional calculations predict defect levels to be deeper and more localized than suggested by earlier conventional DFT calculations. We will discuss optical and equilibrium transition levels in comparison with experiment and previous calculations. Finally, we employ our data to compute free charge carrier concentrations as a function of temperature and doping/impurity conditions, and relate this information to applications in room temperature radiation detection.
10:30 AM - YY3.4
First Principles Predictions of Intrinsic Defects in Aluminum Arsenide, AlAs.
Peter Schultz 1
1 Multiscale Dynamic Materials Modeling 1435, Sandia National Laboratories, Albuquerque, New Mexico, United States
Show AbstractProcess models and modeling of radiation effects require accurate values of defect reaction energies and defect energy levels to predict evolution of defect chemistry and electrical response of devices. Using density functional theory (DFT) calculations within a computational model that incorporates explicit and rigorous treatment of charged boundary conditions, I predict energies and electronic transitions for a comprehensive set of intrinsic defects in AlAs, identifying stable charge states and computing the associated defect energy levels. These simulations use conventional local density and generalized gradient approximations, and are shown to be converged with respect to computational model parameters (such as supercell size). The predicted defect levels, computed as total energy differences, do not suffer from a ‘band gap problem’, spanning a defect band gap that matches the experimental band gap, despite a Kohn-Sham eigenvalue band gap that is much smaller than experiment. As in GaAs, [P.A. Schultz and O.A. von Lilienfeld, Modeling Simul. Mater. Sci. Eng. 17, 084007 (2009)], defects in AlAs exhibit a surprising complexity---with greater range of charge states, bistabilities, and multiple negative U systems---that would not possible to resolve fully in experiment. This demonstrates that DFT simulations have achieved sufficient accuracy to serve semiconductor device engineering needs, and can be used to populate defect physics models in III-V semiconductors reliably in those cases where experimental data is either lacking or ambiguous.Sandia National Laboratories is a multi-program laboratory operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin company, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
10:45 AM - YY3.5
Atomic-scale Modeling of Transition-metal Doping of Semiconductor Quantum Dots.
Tejinder Singh 1 , Triantafillos Mountziaris 1 , Dimitrios Maroudas 1
1 Chemical Engineering, University of Massachusetts, Amherst, Massachusetts, United States
Show AbstractDoping of bulk semiconductors allows for precise control of their properties and forms the basis for the development of electronic and photovoltaic devices. Recent reports on the successful synthesis of doped semiconductor nanocrystals (quantum dots) are particularly promising for applications in solar cells and spintronics. For example, nanocrystals of ZnSe (with zinc-blende lattice structure) and CdSe and ZnO (with wurtzite lattice structure) have been doped successfully with transition-metal (TM) elements (Mn, Co, or Ni). However, the underlying mechanisms of doping of colloidal nanocrystals are not well understood. In this presentation, we report a comprehensive theoretical analysis toward a fundamental kinetic and thermodynamic understanding of doping in ZnO, CdSe, and ZnSe quantum dots based on first-principles density functional theory (DFT) calculations. Our theoretical predictions are consistent with experimental measurements and provide interpretations for the experimental observations.The mechanisms of doping of colloidal ZnO nanocrystals with the TM elements Mn, Co, and Ni is investigated. The dopant atoms are found to have high binding energies for adsorption onto the Zn-vacancy site of the (0001) basal surface and the O-vacancy site of the (0001) basal surface of ZnO nanocrystals; therefore, these surface vacancies provide viable sites for substitutional doping, which is consistent with experimental measurements. However, the doping efficiencies are affected by the strong tendencies of the TM dopants to segregate at the nanocrystal surface facets. Furthermore, using the Mn doping of CdSe as a case study, the effect of nanocrystal size on doping efficiency is explored. We show that Mn adsorption onto small clusters of CdSe is characterized by high binding energies, which explains, in conjunction with the characteristics of Mn surface segregation on CdSe nanocrystals, experimental reports of high doping efficiency for small-size CdSe clusters.In addition, we present a systematic analysis of TM doping of ZnSe nanocrystals. The analysis focuses on the adsorption and surface segregation of Mn dopants on ZnSe nanocrystal surface facets, as well as dopant-induced nanocrystal morphological transitions. Both surface kinetics (dopant adsorption onto the nanocrystal surface facets) and thermodynamics (dopant surface segregation) are found to have a significant effect on the doping efficiencies of ZnSe quantum dots. The analysis also elucidates the important role in determining the doping efficiency of ZnSe nanocrystals played by the chemical potentials of the growth precursor species, which determine the surface structure and morphology of the nanocrystals.
11:30 AM - **YY3.6
Speculations about Making Nickelate High-temperature Superconductors.
Xiaoping Yang 1 , Ole Krogh Andersen 1 , P. Hansmann 2 , A. Toschi 2 , K. Held 2
1 , Max-Planck-Institut fuer Festerkoerperforschung, Stuttgart Germany, 2 , Vienna University of Technology, Vienna Austria
Show AbstractTo my knowledge, no new class of superconductors has been discovered by design, but by chance or by following empirical rules which the next discovery then showed to be of limited validity. The discoveries of the A15 compounds in the seventies, the cuprates in the eighties, MgB2 in 2001, and the iron pnictides and chalcogenides in 2008 are all examples of this. Despite 25 years’ intensive research, high-temperature superconductivity in the cuprates has not been understood, and no superconductor with Tc higher than 150 K has been found since 1993. Even in cases where the mechanism behind the superconductivity has been understood, as for MgB2, this has so far not helped in designing any better superconductor.With the new possibilities for building oxide heterostructures, it has been speculated that d7 = eg1 = (3z2-1, x2-y2)1 nickelates might be used instead of d9 = eg3 = (3z2-1)2 (x2-y2)1 cuprates to obtain even higher Tc by confining a single layer of LaNiO3 between layers of an insulating oxide. By using density-functional calculations followed by downfolding to a correlated, low-energy Hubbard Hamiltonian, and solving the latter in the dynamical mean-field approximation (DMFT), we have shown that it might be possible to empty the (3z2-1)-like band and enforce a single (x2-y2)-like conduction band with a Fermi-surface whose shape is like that of the cuprates with the highest Tc [1]. The electronic correlations help to achieve this, but we find that it is even more important to limit the nickelate to a single layer and to choose the confining material properly, i.e. by cation control. This also seems to be the way to suppress competing phases with charge and/or spin order, e.g. 2d7 → d6+d8, as our LDA+U calculations show. Experimental studies are under way.[2][1] P. Hansmann, Xiaoping Yang, A. Toschi, G. Khaliullin, O. K. Andersen, K. Held; Phys. Rev. Lett. 103, 016401 (2009).[2] J. Chakalian et al. (private communication), B. Keimer et al. (private communication), H.-U.Habermeier (private communication), Y. Tokura et al. (private communication).
12:00 PM - **YY3.7
Role of Intrinsic Defects in Doping Semiconductors with Rare Earth Elements and Molecules.
Peter Yu 1 2
1 Department of Physics, University of California at Berkeley, Berkeley, California, United States, 2 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley , California, United States
Show AbstractIn this presentation, I will talk about the roles played how intrinsic defects, such as vacancies, can affect the doping semiconductors with large magnetic ions or small molecules. Typically intrinsic defects can help to relax the local strain introduced by large magnetic ions replacing smaller host atoms. The resultant complex formed by intrinsic defects and the dopants can have profound effects on the host electronic, magnetic and optical properties. The other extreme is when semiconductors with large lattice constants are doped with smaller atoms such as hydrogen and oxygen. These dopants tend to form small molecules with strong covalent bonds. The large vacancies allow the dopant atoms to be incorporated as molecules with lower formation energies. I will use as examples the introduction of Gd into semiconductors like ZnTe and GaN and of oxygen into CdTe and CdSe. The possibility of developing new devices by taking advantages of doping semiconductors with these dopant-intrinsic defect complexes will also be discussed. *Research performed in collaboration with L. Liu, W. Cheng, Z. X. Ma and S. S. Mao.
12:30 PM - YY3.8
Investigating Dislocations in Silicon Using Molecular Dynamics Simulations.
Lucas Hale 1 , Xiaowang Zhou 3 , Neville Moody 4 , Roberto Ballarini 2 , William Gerberich 1
1 Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota, United States, 3 Mechanics of Materials, Sandia National Laboratories, Livermore, California, United States, 4 Hydrogen and Metallurgy Science, Sandia National Laboratories, Livermore, California, United States, 2 Civil Engineering, University of Minnesota, Minneapolis, Minnesota, United States
Show AbstractDislocations are believed to be a major influence on the yielding behavior of silicon nanostructures even at room temperature. With this in mind, molecular dynamics simulations using classical potentials have been performed to investigate dislocation interactions in silicon. A particular emphasis is placed on dislocation-interface interactions. The behavior of dislocations approaching a free surface and approaching an oxide layer are observed showing a marked difference between the two. The oxide layer was simulated with a modified Stillinger-Weber potential that allows for good interfacial behavior. The simulation results are compared with analytical models to determine how consistent they are with each other.
12:45 PM - YY3.9
Effect of Pressure and Stress on Blistering Induced by Hydrogen Implantation in Silicon.
Christophe Coupeau 1 , Eloi Dion 1 , Marie-Laure David 1 , Jerome Colin 1 , Jean Grilhe 1
1 , Institut P', Universite de Poitiers/CNRS/ENSMA, Futuroscope France
Show AbstractH-implantation induced blistering in silicon is a complex phenomenon involving various mechanisms, from the nucleation and growth of gas-vacancies complexes, at the nanometer scale, to the mechanical buckling of the materials at the microscopic scale. On one hand, the internal stresses induced by either hydrogen implantation, buried SiGe strain layers or overpressurized helium plates appear to play a key role in the platelets formation. On the other hand, the pressure due to the diffusion and absorption of hydrogen in the cavities during post-implantation thermal treatments has been identified as a relevant parameter for understanding the blistering phenomenon. Analytical models based on elasticity have also been proposed to evidence its influence on buckling. It has been found that the geometrical parameters of the blisters are directly related to the internal pressure. Recently, finite element simulations taking into account plastic deformations that may occur in circular buckles have been performed and the amount of gas accumulated in the cavities of implanted tungsten has been estimated. A number of issues still remain however unexplored, among which is the combined effect of pressure and stresses on blistering. In both analytical models and numerical simulations reported in the literature, the internal stresses induced by implantation in the upper layer of the material are neglected, even if they were experimentally determined to increase with the ion fluence. It is the purpose of this work to explore the coupling effect on blistering between the internal pressure in the cavities and the stresses in the implanted materials. Experimental observations of blisters induced by hydrogen implantation in silicon wafers are first reported. The effect of pressure and stress on buckling is then discussed in the framework of Föppl-von Karman theory of thin plates. The internal pressure in the cavities is finally determined from the model, with the help of the profiles of the observed buckling structures.
YY4: Low Dimensional Materials I
Session Chairs
Wednesday PM, April 27, 2011
Room 2014 (Moscone West)
2:30 PM - **YY4.1
Novel Electronic and Optical Properties of Graphene and Graphene Nanostructures.
Steven Louie 1 2
1 Department of Physics, University of California at Berkeley, Berkeley, California, United States, 2 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States
Show AbstractGraphene, a single atomic layer of carbon in a honeycomb structure, exhibits many fascinating properties of fundamental and practical interests. Its carriers behave like two-dimensional massless Dirac fermions with pseudospin character. We report here results from recent theoretical studies on the electronic, transport and optical properties of graphene and graphene-based nanostructures including nanoribbons, superlattices and bilayer structures. A number of phenomena unique to graphene are shown. The carrier dynamics in graphene exhibits anomalous anisotropy when subjected to an external periodic electrostatic potential of nanometer dimensions (called graphene superlattices). Under appropriate conditions, these graphene superlattices are predicted to be electron supercollimators and new generation of massless Dirac fermions may be created. The nanoribbions are semiconductors with novel electronic, magnetic and optical properties. Our latest studies have further shown that: i) Edge states persist on chiral graphene nanoribbons and are seen in scanning tunneling spectroscopy (STS) measurements. ii) The electronic properties of graphene under a 1D inhomogeneous magnetic field may be mapped into those of one under an electric field and vice versa. iii) Because of reduced dimensionality, electron-electron interaction effects are particularly important leading to strong excitonic effects in the optical response and to new features in the STS of these 2D systems. iv) Bilayer graphene has novel excitons that are tunable with applied electric fields. v) Quantum phases in graphene such as the Berry’s phase are directly extractable from angle-resolved photoemission spectroscopy.
3:00 PM - **YY4.2
Excitonic Effects on the Optical Absorption Spectra of Graphene.
Li Yang 1
1 Department of Physics, Washington University in St. Louis, Saint Louis, Missouri, United States
Show AbstractIn this talk, I will present our first-principles studies on excitonic effects and optical absorption spectra of graphene and few-layer graphene. In addition to resonant excitons from π and π* bands, the unique parallel σ and π* bands of graphene give rise to enhanced excitonic effects in such two-dimensional (2-D) semimetals; significant broadly resonant excitonic states reshape the absorption profile but without moving the absorption edge; more importantly, one prominent nearly bound exciton (an extremely narrow resonant state) is discovered below the absorption continuum with a large binding energy, 270 meV in graphene and 80 meV in bilayer graphene. This is a surprising result because enhanced excitonic effects and well defined electron-hole pairs are usually not expected in metallic systems due to the extremely strong metallic screening. Moreover, our first-principles calculations show that these nearly bound excitons can be prominent even under significantly doping conditions, which makes them detectable in experiment. If time is enough, I will reveal the interesting relationship between excitonic effects of graphene and those of bulk silicon because of their similar parallel band structures involved in optical absorption.
3:30 PM - YY4.3
First-principles Study of Gas Adsorption and Catalysis of Defect-engineered Graphenes and Nanotubes.
Yong-Hyun Kim 1 , Sang-Ouk Kim 2 , Duck Hyun Lee 2 , Heechol Choi 3 , Young Choon Park 3 , Yoon Sup Lee 3
1 Graduate School of Nanoscience and Technology, KAIST, Daejeon Korea (the Republic of), 2 Department of Materials Science and Engineering, KAIST, Daejeon Korea (the Republic of), 3 Department of Chemistry, KAIST, Daejeon Korea (the Republic of)
Show AbstractIn future energy materials applications, one has to control properly gaseous molecules such as hydrogen (H2), oxygen (O2), and carbon dioxide (CO2), respectively, in hydrogen storage and production, fuel-cell and battery catalysis, and CO2 sequestration and conversion. The first step of such controlling processes would be the favorable adsorption of the gas molecules on surface sites of nanostructured materials. Because of their intrinsic large surface areas, graphenes, carbon nanotubes, and boron nitride (BN) nanotubes have been considered as good potential materials for high-capacity gas adsorbents. However, the complete-shell sp2 honey-comb networks of the carbon and BN nanotubes are almost inert so that they do not allow any significant interaction with as-well complete-shelled gas molecules. One of the most important criteria of the gas adsorption for ambient-condition energy materials applications is that the adsorption energy should be greater than the standard free energies of gaseous molecules, which range from 0.4 to 0.7 eV at room temperature depending on the type of gas molecules. In order to create favorable gas adsorption sites, extrinsic defects as like B, Be [1], and transition metals [2,3] in the sp2 networks should be engineered, and enhanced hydrogen adsorption characteristics on the defect sites has been discussed based on results of first-principles density-functional theory calculations. Here, we report the favorable adsorption characteristics of O2 and CO2 on defect-engineered graphenes, and C and BN nanotubes, from results of DFT-based total energy calculations. The calculated adsorption energies of O2 and CO2 are 0.8-1.0 eV, large enough to operate for ambient-condition catalysis [4] and sequestration [5]. Particularly, a biomimetic defect center of the Fe-N4 porphyrin moiety incorporated into graphitic walls shows a good promise for efficient oxygen reduction reaction catalysis. Also, for BN nanotubes we have found that boron antisites, in which a boron atom sits at the nitrogen site surrounded by three boron atoms, can interact with a CO2 molecule strongly enough to break a C=O double bond. We will also discuss its implication to CO2 sequestration and fuel conversion. [1] Y.-H. Kim, Y. Zhao, A. Williamson, M. J. Heben, and S. B. Zhang, Phys. Rev. Lett. 96, 016102 (2006).[2] Y. Zhao, Y.-H. Kim, A.C. Dillon, M.J. Heben, and S.B. Zhang, Phys. Rev. Lett. 94, 155504 (2005).[3] W. I. Choi, S.-H. Jhi, K. Kim, and Y.-H. Kim, Phys. Rev. B 81, 085441 (2010).[4] D. H. Lee, W. J. Lee, W. J. Lee, S. O. Kim, and Y.-H. Kim, submitted (2010).[5] H. C. Choi, Y. C. Park, Y.-H. Kim, and Y. S. Lee, to be submitted (2010).
3:45 PM - YY4.4
Tensile Strain Effects on Absorption and Diffusion of H Atom on Graphene.
Ming Yang 1 2 , Chun Zhang 1 , Ariando Ariando 1 2
1 Physics Department, National University of Singapore, Singapore Singapore, 2 Nanocore, National University of Singapore, Singapore Singapore
Show AbstractThe interaction between adatom and substrate has been intensively studied, which can change related physical and chemical properties markedly. The understanding of how theadatom interacts with the substrate such as absorption and diffusion and how to control this process are of importance to explore novel applications. Recently, it is found that the absorption of hydrogen atoms on graphene can modify the electronic properties greatly, giving another possiblity to control the electronic properties. However, the study on how external tensile strain will influence the H atom absorbed and diffused on graphene, and how these modified absorption and diffusion will affect electronic properties of graphene is still limited.In this study, effects of tensile strain on absorption and diffusion of hydrogen atom in graphene have been studied by first-principles calculations. It is found that there is a barrier about 0.3 eV for H atom diffused from free space to graphene surface. The tensile strain induces increasing charge transfer between C and H atoms, which decreases absorption energy of hydrogen on graphene, and increases the barrier of in-plane diffusion. The tensile strain can also reduce bonding strength of C sp2 bond, making the barrier of out-plane diffusion lower. In particular, the out-plane diffusion barrier is affected more significantly by the strain along armchair direction, of which H atom maydiffuse through the center of C-C bond with 10% strain, but the H atom will be trapped at the center of C-C bond when the strain increases to 15%.
4:30 PM - **YY4.5
Excitonic Effects on the Optical Properties of SiC Sheet and Nanotubes.
Hung-Chung Hsueh 1 , Guang-Yu Guo 2 3 , Steven Louie 4 5
1 Department of Physics, Tamkang University, Taipei 25137 Taiwan, 2 Graduate Institute of Applied Physics, National Chengchi University, Taipei 11605 Taiwan, 3 Department of Physics, National Taiwan University, Taipei 10617 Taiwan, 4 Department of Physics, University of California, Berkeley, California, United States, 5 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States
Show AbstractSince the discovery of carbon nanotubes (CNTs) in 1991 by Iijima, carbon and other nanotubes have attracted considerable interest worldwide because of their unusual properties and also great potentials for technological applications. Apart from CNTs, inorganic tubukar materials, such as BN, AlN, GaN and SiC, have also been predicted and synthesized [1,2]. Though CNTs continue to attract great interest, other inorganic nanotubes may offer different opportunities that CNTs cannot provide. Therefore, we have previously carried out systematic computational studies of the optical and nonlinear optical properties of SiC nanotubes within the density functional theory (DFT) with the local density approximation (LDA) [3,4]. In this talk, we will present ab initio quasiparticle and excitonic calculations of the electronic and optical properties of both SiC sheet and nanotubes. The signicant band-gap correction of up to 1eV to the LDA of semiconducting SiC-NTs and SiC sheet is found to be mainly due to the many-body screening effect which is included in our GW approximation [5] for the valence electron self-energy. Furthermore, the excitons with a large binding energy a strong anisotropy are present in our GW+BSE (Bethe-Salpeter equation) calculations [6] of the low-dimensional SiC systems. The characters of these strongly bonded excitons in SiC-NTs is also analyzed in terms of the obtained excitonic wavefunctions.[1] X. Blase, A. Rubio, S. G. Louie, and M. L. Cohen, Europhys. Lett. 28, 335 (1994)[2] N. G. Chopra, R. J. Luyken, K. Cherrey, V. H. Crespi, M. L. Cohen, S. G. Louie, and A. Zettl, Science 28, 335 (1994)[3] I. J. Wu and G. Y. Guo, Physical Review B 76, 035343 (2007)[4] I. J. Wu and G. Y. Guo, Physical Review B 78, 035447 (2008)[5] M. S. Hybertsen and S. G. Louie, Phys. Rev. B 34, 5390 (1986)[6] M. Rohlfing and S. G. Louie, Phys. Rev. B 62, 4927 (2000)
5:00 PM - YY4.6
Molecular Dynamics Simulation Study of Silicon Nanowires Oxidation.
Byung-Hyun Kim 1 2 , Mina Park 1 , Mauludi Ariesto Pamungkas 1 3 , Gyubong Kim 1 , Kwang-Ryeol Lee 1 , Yong-Chae Chung 2
1 Computational Science Center, Korea Institute of Science and Technology, Seoul Korea (the Republic of), 2 Department of Materials Science and Engineering, Hanyang University, Seoul Korea (the Republic of), 3 Department of Nanomaterial Science and Technology, University of Science and Technology, Daejon Korea (the Republic of)
Show AbstractApplication of Si NW to various nano-electronic devices requires a systematic understanding of the oxidation behavior of Si NW in atomic scale, since the dielectric layer of the NW channel would be crucial for most nano-devices. In the cases of the oxidation of NWs or nanoclusters, it has been known that the self-limiting oxidation occurs due to the stress at the interface. On the other hand, a recent experimental work with very thin Si NW revealed that the oxidation occurs throughout the NW without retardation [1]. More investigation with thin NW is thus required for a complete understanding of the 2 dimensional oxidation. Computational approach of both molecular dynamics simulation and the first principle calculation would provide an alternative way to understand the atomistic details of the oxidation process. In the present work, we investigated the stress evolution during oxidation of very thin Si NWs using molecular dynamics simulation. To obtain more comprehensive understanding of oxidation process, we also performed pseudopotential density functional calculations with oxygen adatom in pressurized Si bulk structure. The density functional calculations provided the fundamental bonding characters between oxygen and Si in correlation with the applied pressures. Reactive force field was employed as the interatomic potential for the oxidation simulation. We focused on the difference in the stress evolution behavior during early stage of the oxidization of the Si NWs of 5 and 10 nm in diameter. Single crystal Si NWs ([311] along the axis) were oxidized at 1073K by the oxygen molecules filled in the simulation box. For the relaxed Si NW configuration before oxidation, core part of the NWs exhibited the compressive stress in both the radial and circumferential direction due to the curved surface. As the surface oxidation occurred, volume expansion of the surface layer released the compressive stress of the core. On the other hand, the oxidized surface layer had a compressive stress. We found that the surface compressive stress is larger in magnitude for the Si NW of diameter 10nm. The difference in the residual stress originated from the fact that the mechanical constraint of the core part against the volume expansion of the surface layer decrease as the diameter decreases. Based on the present results, we suggested a criterion of the self-limiting oxidation of Si NW in terms of the elastic energy of the system and the driving force of the oxidation. This scheme can explain recent experimental result where thin Si NW was fully oxidized while thicker Si NW exhibited typical self-limiting oxidation. [1] Ilsoo Kim and Heon-Jin Choi, unpublished work (2010).
5:15 PM - YY4.7
Carbon Nanotube with Square Cross-section: An Ab Initio Investigation.
Pedro Autreto 1 , Marcelo Flores 1 , Sergio Legoas 2 , Douglas Galvao 1
1 Applied Physics, State University of Campinas, Campinas, Sao Paulo, Brazil, 2 Physics, Federal University of Roraima, Boa Vista, Roraima, Brazil
Show AbstractThe study of the mechanical properties of nanoscale systems presents new theoretical and experimental challenges. The arrangements of atoms at nano- and macroscales can be quite different and affect electronic and mechanical properties. Of particular interest are the structures that do not exist at macroscale but can be formed (at least as metastable states) at nanoscale, specially when significant stress/strain is present. One example of these cases is the recent discovery of the smallest metal (silver) nanotube from high resolution transmission electron microscopy experiments [1]. These tubes are spontaneously formed during the elongation of silver nanocontacts. A natural question is whether similar carbon nanotubes (carbon nanotubes with square cross-section (CNTSCs)) can exist. From the topological point of view, CNTSC tubes would require carbon atoms arranged in multiple square-like configurations. Molecular motifs satisfying these conditions, the so-called cubanes, do exist and they are stable at room temperature. In this work we report ab initio (DMOL3 code) results for the structural, stability, and electronic properties for such hypothetical structures. Our results show [2] that stable (or at least metastable) CNTSC structures can exist. They also show that it is possible to convert SWNT(2,2) to CNTSC under radial compression. CNTSCs should share most of the general features exhibited by “standard” nanotubes. Although the CNTSCs have not yet been observed, we believe our results had proven their feasibility.[1] M. J. Lagos, F. Sato, J. Bettini, V. Rodrigues, D. S. Galvao, and D. Ugarte, Nature Nanotechnology v4, 149 (2009).[2] P. A. S. Autreto,S. B. Legoas, M. Z. S. Flores, and D. S. Galvão, J. Chem. Phys. v133, 124513 (2010).
YY8: Poster Session II
Session Chairs
Thursday PM, April 28, 2011
Salons 7-9 (Marriott)
1:00 AM - YY8: Poster II
YY8.12 Transferred to YY6.5
Show AbstractYY6: Semiconducting Energy Materials
Session Chairs
Thursday PM, April 28, 2011
Room 2014 (Moscone West)
10:00 AM - YY6.2
First-principles Study of III-V Semiconductor-water Interfaces for Photoelectrochemical Hydrogen Production.
Brandon Wood 1 , Tadashi Ogitsu 1 , Eric Schwegler 1
1 Quantum Simulations Group, LLNL, Livermore, California, United States
Show AbstractPhotoelectrochemical devices based on III-V semiconductors represent a promising pathway for sustainable hydrogen production using sunlight and water. However, combining high efficiency with stability under operating conditions is difficult. Device improvement requires an in-depth theoretical understanding of the complex relationship between stability and catalytic activity of III-V semiconductor surfaces in contact with water. Accordingly, we have performed extensive first-principles molecular dynamics simulations on model III-V semiconductor surfaces (InP, GaP) in realistic aqueous environments. Our results are used to investigate the structure, stability, and chemical activity of these surfaces, with the aim of understanding the reactive states precursory to photoexcitation and hydrogen evolution. Our results show that surface oxide nucleation is key to facilitating reactivity of III-V materials, and that the bonding arrangement around the surface oxygen is a significant determiner of the available pathways for water dissociation and corrosion. This points to the importance of III-V surface oxides as intermediates in the water-dissociation component of hydrogen evolution. Prepared by LLNL under Contract DE-AC52-07NA27344.
10:15 AM - YY6.3
Prediction of Semiconductor Band Edge Positions in Aqueous Environments from First Principles.
Yabi Wu 1 , Maria Chan 1 2 , Gerbrand Ceder 1
1 Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States, 2 Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois, United States
Show AbstractThe ability to predict a semiconductor's band edge positions in aqueous environment is important for the design of a water-splitting photocatalyst. In this talk, we introduce a novel first principles method to compute the conduction band minima of semiconductors relative to the water H2O/H2 level. The methods combines semilocal density functional theory with classical molecular dynamics. We test the method on six well known photocatalyst materials: TiO2, WO3, CdS, ZnSe, GaAs and GaP. The predicted band edge positions are within 0.34 eV compared to experimental data, with a mean absolute error of 0.19 eV.
10:30 AM - YY6.4
Electronic Structure Study of the Reduced Anatase TiO2-x Using the GGA+Ud+Up Method.
Kwanghee Cho 1 2 , Hyung Dong Lee 2 , Seonggeon Park 2 , Blanka Magyari-Koepe 2 , Yoshio Nishi 2 , Sungjoo Hong 1 , Sungwoong Chung 1 , Jinwon Park 1 , Jaeyun Yi 1
1 Nt&I, hynix, Icheon Korea (the Republic of), 2 Electronic Engineering, Stanford, Stanford, California, United States
Show AbstractThe Resistive Random Access Memory (RRAM) is one of the most promising candidates for future memory devices. TiO2, a semiconductor material, had been most extensively studied and used in bipolar switching RRAM devices, it was recently found to show also unipolar resistive switching. The TiO2 film for RRAM can have either the rutile or anatase phase during the operation time of the device, as reported by many research groups. Accurate understanding of the electronic structure of the reduced anatase structure is important, since it is currently believed to be responsible for an intermediate process in the resistive switching in TiO2-based RRAM. It has been recently found that amorphous TiO2 transforms to anatase during the forming. There has been a discrepancy between the experimental structural parameters for anatase TiO2, band-gap of 3.2 eV, bulk modulus 180~190GPa and lattice constant a=b=3.8A, c=9.5A, and earlier first-principles simulations, especially the band-gap was severely underestimated. In these studies, the simulations usually employed the generalized gradient approximation, (GGA) or the addition of on-site Coulomb correction on the d orbtital of Ti, (GGA+U). The Coulomb correction improves on the bulk modulus and lattice constants agreement with the experimental values, however the band-gap were still underestimated, around 2.5 ~2.6eV in those early theoretical calculations. In this study, we investigated the electronic structure of anatase TiO2 based on density functional theory and GGA+Ud+Up method, where we apply the Coulomb correction on the oxygen p orbital in addition to the titanium d orbitals. We find that all of parameters, band gap, bulk modulus and lattice constants can be improved. The equilibrium lattice structure depends on the choice of both Ud and Up parameters and we have established certain trends to be used as guidelines in choosing the adequate Ud and Up values. With increasing Ud, the band gap and lattice constant increases at the expense of smaller bulk modulus. When Ud is fixed, the band gap and bulk modulus are increased, while the lattice constants decrease with increasing Up. In this work, we show how to choose the optimum values to obtain more reliable band gap, bulk modulus and lattice constant close to experimental values. The position of the defects states and defect energetics had also been investigated in a supercell Ti72O144 structure containing 1 and 2 oxygen vacancies.
10:45 AM - YY6.5
From Atomistic to Device Scale Modeling of Organic Photovoltaics: Linking Coarse-grained Molecular Dynamics with Phase Field Modeling.
Olga Wodo 1 , R. Jaeger 1 , Monica Lamm 2 , Baskar Ganapathysubramanian 1
1 Mechanical Engineering, Iowa State University, Ames, Iowa, United States, 2 Chemical and Biological Engineering, Iowa State University, Ames, Iowa, United States
Show AbstractOrganic solar cells (OSCs) fabricated from polymer-fullerene blends offer a promising low-cost strategy for harnessing solar energy. In addition to low cost, the possibility of flexible, large-area fabrication makes these devices highly attractive for ubiquitous solar electric conversion. State of the art organic solar cells with efficiencies of ~ 8% have been achieved using conjugated polymers. These devices are fabricated by spin coating the active layer from a blend of p-type photoactive polymer and n-type derivatives of fullerenes in the so-called “blend-heterojunction” architecture.A key property determining the power conversion efficiency of these OSCs is the final morphological distribution of the polymer and fullerene phases after the evaporation of the solvent in the coating process. This distribution is affected by multiple interacting phenomena spanning multiple length scales—in particular, the chemo-physical properties of the solutes, solvent and polymer (molecular scale), the thermodynamic interactions during morphology evolution (mesoscale) and the fabrication conditions like evaporation rate, spinning rate, substrate patterning (macroscale). However, current state-of-the-art approaches to understanding and designing high efficiency OSCs are either limited to a single scale analysis or involve combinatorial, trial-and-error based experimental investigations. We develop a multi-scale computational analysis framework to model morphological evolution during the fabrication process of organic photovoltaic coatings. In this approach, the mesoscale (and thermodynamic) phenomena are modeled using a phase field approach. We model evaporation-induced phase separation in ternary systems, which consist of conjugated polymer, fullerene derivative and solvent. The model takes into account both thermodynamic (e.g. interaction parameters between components) and kinetic parameters (e.g. diffusion coefficient). These parameters are introduced into the continuum phase field model via coarse-grained molecular dynamics (CGMD) simulation. This modeling framework has the advantage of utilizing molecular scale properties (from CGMD) in the phase field simulations and thus, can serve as a tool for the rational selection of molecularly engineered polymer-fullerene combinations.
11:30 AM - **YY6.6
Theoretical Understanding of the Stability and Defect Properties of Cu2ZnSn(S,Se)4 Solar Cell Absorbers.
Xin-Gao Gong 1
1 , Fudan University, Shanghai China
Show AbstractCu2ZnSnS4 is one of the most promising absorber materials for thin-film solar cells, because it is a low-cost material with the optimal band gap 1.5 eV for single-junction solar cells and has a high adsorption coefficient. Although this semiconducting material has been studied for almost 50 years, due to the complicity of the quaternary compound, the properties are still not well understood, which has affected the improvement of the performance of Cu2ZnSnS4 based solar cell. In this talk, I will show that Cu2ZnSnS4 and Cu2ZnSnSe4 can be derived from the binary zinc-blende semiconductor through sequential cation cross-substitution, therefore, their structural and electronic properties can also be understood through the evolution from the binary, to ternary, and to quaternary chalcogenide compounds. We found that the low energy crystal structure obeys the octet rule, and predicted that the ground state structure of Cu2ZnSnS4 to be kesterite and the direct band gap of Cu2ZnSnSe4 to be around 1 eV, which is experimentally confirmed recently. Examination of the thermodynamic stability of these quaternary compounds reveals that the stable chemical potential region for the formation of stoichiometric compound is small. Under these conditions, the dominant p-type defect in CZTS will be CuZn antisite, which has a relatively deeper acceptor level than the Cu vacancy. We propose that to maximize the solar cell performance, growth of Cu2ZnSnS4 under Cu-poor/Zn-rich conditions will be optimal, if the precipitation of ZnS can be avoided by kinetic barriers. The properties of Cu2ZnSn(S, Se)4 alloys will also be discussed. We will explain why adding Se to Cu2ZnSnS4 is beneficial for solar cell performance. This work is in collaboration with S.Y. Chen, A. Walsh and S.-H. Wei
12:00 PM - YY6.7
First Principles Calculations of Defect Formation in In-free Photovoltaic Semiconductors Cu2ZnSnSe4, Cu2ZnGeSe4 and u2ZnSiSe4.
Tsuyoshi Maeda 1 , Satoshi Nakamura 1 , Takahiro Wada 1
1 Department of Materials Chemistry, Ryukoku University, Otsu, Shiga, Japan
Show AbstractThe efficiency of Cu(In,Ga)Se2 (CIGS) thin film solar cells has reached 20.3%. Recently, finding a substitute for indium and gallium has become an important issue because they are expensive rare metals. Cu2ZnSnSe4 (CZTSe) and Cu2ZnSnS4 (CZTS) are anticipated as indium- and gallium-free absorber materials. We studied the phase stability of kesterite-, stannite- and wurtz-stannite-type CZTSe by first principles calculations [1]. Then, we observed characteristic diffraction peaks of the kesterite-type CZTSe in the neutron powder diffraction pattern. Recently, we reported on the vacancy formation energies in CZTSe and CZTS to compare them with the result for CIS [2, 3]. We predicted that Cu vacancies are easily formed in CZTSe and CZTS, but not easily formed in comparison with Cu vacancy in CIS.Recently, we reported the phase stability and electronic structures of Cu2ZnGeSe4 and Cu2ZnSiSe4 [4]. We predicted that in these compounds, the KS phase is more stable than the ST and WST phases. For Cu2ZnIVSe4 [IV: Sn, Ge, Si], the valence band maximum (VBM) is an antibonding of Cu 3d and Se 4p orbitals, while the conduction band minimum (CBM) is an antibonding of IV ns and Se 4p orbitals. However, we did not report on the defect formation in Cu2ZnGeSe4 and Cu2ZnSiSe4. In this study, we calculated the vacancy formation energies in Cu2ZnGeSe4 and Cu2ZnSiSe4 to compare them with the results for CZTSe and CZTS.
We performed first principles calculations within density functional theory, using a plane-wave pseudopotential method. The formation energy of the vacancy was calculated from the difference of total energy between perfect crystal and imperfect crystal with vacancy. Calculations for imperfect crystal with vacancy were performed using a supercell with 64 atoms, which is four times greater than that of a kesterite-type unit cell.
The results obtained for CZTSe and CZTS were as follows. (1) Under the Cu-poor and Zn-rich condition, the formation energy of the Cu vacancy was generally smaller than those of the Zn, Sn and Se or S vacancies in CZTSe and CZTS. (2) Formation energies of Cu, Zn, Sn vacancies in CZTS were larger than those in CZTSe. On the other hand, the formation energy of the S vacancy was smaller than that of the Se vacancy in CZTSe. (3) Under the Cu-poor and Zn-rich condition, formation energies of the Cu vacancy in CZTS and CZTSe were much larger than that in CIS. These results indicate that in kesterite-type CZTS and CZTSe, Cu vacancy is easily formed under Cu-poor, Zn-rich and Se-rich conditions, but it is more difficult than in CIS.
[1] T. Maeda et al., Mat. Res. Symp. Proc. (2009) 1165-M04-03.
[2] T. Maeda et al., Thin solid Films, accepted.
[3] T. Maeda et al., Jpn. J. Appl. Phys, submitted.
[4] S. Nakamura et al., Jpn. J. Appl. Phys, accepted.
12:15 PM - YY6.8
Phase Stability and Electronic Structure of In-free Photovoltaic Materials Cu2ZnSnS4, Cu2ZnGeS4 and Cu2ZnSiS4 by First-principles Calculations.
Satoshi Nakamura 1 , Tsuyoshi Maeda 1 , Takahiro Wada 1
1 Department of Materials Chemistry, Ryukoku university, Otsu, Shiga, Japan
Show AbstractCu(In,Ga)Se2 (CIGS) has been anticipated as one of the most promising materials for thin film solar cells. Substituting abundant elements for indium and gallium in CIGS has become an important issue because they are expensive rare metals. Cu2ZnSnS4 (CZTS) and Cu2ZnSnSe4 (CZTSe) are indium-free photovoltaic materials. The efficiency of CZTS-based thin film solar cells has reached 9.6% [1]. Recently, new materials that have wide band gap (Eg >1.7 eV) and narrow band gap (Eg <1.0 eV) are required for thin film full spectrum solar cells. Quaternary I2-II-IV-VI4 compounds are candidates for wide and narrow band gap absorber materials. For the I2-II-IV-VI4 compounds, three kinds of crystal structures, i.e. kesterite (KS)-, stannite (ST)-, and wurtz-stannite (WST)-type, have been reported. The KS and ST phases cannot be easily distinguished by powder X-ray diffraction because they belong to similar space groups. KS-type structure was reported for CZTS. On the other hand, both ST-type and WST-type structures were reported for Cu2ZnGeS4. For Cu2ZnSiS4, WST-type structure was reported.
Most recently, we have reported the phase stability and electronic structure of In-free photovoltaic materials Cu2ZnIVSe4 [IV: Sn, Ge, Si] [2]. We predicted that in these compounds, the KS phase is more stable than the ST and WST phases. The VBM of Cu2ZnIVSe4 consists of Cu 3d and Se 4p, while the CBM consists of IV ns and Se 4p. The band gaps of KS-type Cu2ZnGeSe4 and Cu2ZnSiSe4 are wider than that of KS-type Cu2ZnSnSe4. However, theoretical studies on phase stability and electronic structure of Cu2ZnGeS4 and Cu2ZnSiS4 have not yet been reported.
In this study, we evaluate the phase stability and electronic structure of the KS-, ST- and WST-type Cu2ZnGeS4 and Cu2ZnSiS4 by first principles calculations and compare the results with those of CZTS.
To evaluate the phase stability of the KS-, ST-, and WST-types of CZTS, Cu2ZnGeS4 and Cu2ZnSiS4, the formation enthalpies (ΔH) of these phases were calculated. For CZTS, the ΔH of the KS phase (-338.3 kJ/mol) is a little smaller than that of the ST phase (-336.2 kJ/mol) and much smaller than that of the WST phase (-331.5 kJ/mol). This result indicates that the KS phase is more stable than the ST and WST phases. The theoretical band gap calculated with the generalized gradient approximation (GGA) functional of the KS phase is 0.06 eV, which is a little larger than those of the ST phase of 0.00 eV and WST phase of 0.02 eV. However, the band gap energies calculated with the GGA functional are greatly underestimated. In the previous study, we reported that the screened exchange-LDA (sX-LDA) calculation greatly improved the band gap of CIS [3]. The band gap energies of Cu2ZnGeS4 and Cu2ZnSiS4 calculated with the sX-LDA method are presented.
[1] T. K. Todorov et al., Adv. Mater. 22, E156 (2010).
[2] S. Nakamura et al., accepted in Jpn. J. Appl. Phys.
[3] T. Maeda et al., Jpn. J. Appl. Phys. 44, 04DP07 (2010).
12:30 PM - YY6.9
Can Hot Electron Transfer Contribute to Charge Generation in Organic Photovoltaic Materials?
James Kirkpatrick 1
1 , Oxford University, Oxford United Kingdom
Show AbstractIn an organic blend, charges are generated by splitting excitons at the interface between materials with different electron affinities. This difference in electronic affinities is the driving force for charge separation and is often large (>0.5eV). Once generated charges must escape each other’s Coulomb attraction. Could some of the driving force for charge separation be used by the charges to escape each other's Coulomb attraction? This process has been proposed in the literature, but the molecular basis for such phenomena has not been investigated. In this contribution we will discuss the microscopic mechanism that allow some of the exciton's excess energy can be used to separate charges. The model relies on the fact that the same nuclear vibrational modes that are coupled to exciton splitting are also coupled to charge transport (at least for the electron acceptor). A relationship between excess energy and thermalisation distance is thus established.
12:45 PM - YY6.10
Morphology Descriptors of Bulk Heterojunctions in Thin Film Organic Solar Cells.
Olga Wodo 1 , S. Tirthapura 2 , S. Chaudhary 2 , Baskar Ganapathysubramanian 1
1 Mechanical Engineering, Iowa State University, Ames, Iowa, United States, 2 Department of Electrical and Computer Engineering, Iowa State University, Ames, Iowa, United States
Show AbstractOrganic solar cells (OSCs) fabricated from polymer-fullerene blends offer a promising low-cost strategy for harnessing solar energy. In addition to low cost, the possibility of flexible, large-area fabrication makes these devices highly attractive for ubiquitous solar electric conversion. The morphological distribution (of the polymer and fullerene) in these thin-film devices significantly affects the power conversion efficiency of the device. The morphology (so called bulk heterojunction morphology) is affected by the processing conditions (like evaporation rate, substrate surface). Both the processing conditions and the morphology affect the property of the final device (like current generation density, final efficiency of the device). Obtaining detailed knowledge of the process–structure–property relationship is thus necessary for optimizing processing conditions to achieve high-efficiency OSCs. Considering the large number of process variables and system variables which affect the morphology, linking structure and property is done via defining and extracting physically meaningful morphology descriptors from morphological data. We describe a suite of morphology descriptors that encode the various physical processes that affect the total power conversion efficiency of a photovoltaic cell: (a) energy (photon) absorption by the active layer to generate an electron-hole pair (exciton),(b) dissociation of electron-hole pairs along interfaces to create separated charge carriers, and (c) charge transport through the bulk heterojunction to the electrodes.These processes pose competing requirements on the optimal morphology. For instance, to increase the density of dissociated charges (step b above) the length of the interface has to be maximized; however this may cause the pathways to appropriate electrodes to become longer (step c above) or even scattered and disjoint from the electrodes (increasing chance of recombination and poorer charge transport to the electrode). We categorize morphology descriptors into two groups: morphology descriptors quantifying exciton dissociation (feature size quantifying the distance to the interface and the length of the interface) and descriptors quantifying charge transport (structural features). In particular, we utilize ideas from computational homology (Betti numbers, percolation threshold) to develop the structural descriptors. We will illustrate the link between developed descriptors (power spectral density, ratio of interface area to volume, Betti numbers, rose of intersections, percolation features) and the power conversion efficiency for a set of experimental and computational bulk heterojunction (BHJ) based thin film photovoltaic devices. This provides valuable insight into link between BHJ structure and device property and helps to characterize an “ideal” morphology.
YY8: Poster Session II
Session Chairs
Friday AM, April 29, 2011
Salons 7-9 (Marriott)
9:00 PM - YY8.10
Simulation of 4H-SiC UV Photodetector by Finite Element Method.
Stephane Biondo 1 , Wilfried Vervisch 1 , Laurent Ottaviani 1 , Olivier Palais 1
1 , IM2NP, Marseille France
Show AbstractUltra violet (UV) photodetector is essential in many places for safety reasons. Due to many cautions, it is important to forecast all the hazardous phenomena, in particular in the nuclear area, or in space. Thus, the wide band-gap materials such as III-V nitrides and 4H-SiC are mainly used for the realization of UV photodetector components. The expected reliability and radiation hardness lead us to choose the 4H-SiC to realise UV-photodetector.The Maxwell’s equations are solved by the Finite Difference Time Domain (FDTD) method. This rigorous calculus method allows to obtain the optical intensity and the carrier generation according to the structure. The simulation of the electrical phenomenon picks up the optical part results such as optical generation. The carrier generation rate is introduced in the continuity equation. These continuity equations for the electrons and the holes are associated to the Poisson’s equation and the physical phenomena such as the SRH recombination, and the Okuto-Crowell ionization.This paper presents optical simulation results for 4H-SiC ultraviolet photodetector based on p-i-n structure. We also show the evolution of current density with material properties and the device geometry. A theoretical study has led us to consider the generation-recombination phenomenon as the relevant mechanism in the 4H-SiC UV-photodetector device. This mechanism leads to the generation-recombination current density occurring mainly in the space charge region (SCR). This influence of the SCR is also pointed out by the simulation tool. Under the dark, the current density in accordance with reverse bias voltage in the range between 0-500V, is no influence by the p-type doping concentration and p-type layer thickness. However under the UV-light, the UV-photodetector current density evolves with this generate in accordance with the structure. Among the different structure parameters, the p-type doping concentration and the p-type layer thickness decrease lead to the current density increase. Even if the absorption improvement with different structures, we have pointed out the relevant influence of the SCR geometry. Finally, the simulation results allow us to propose an optimized UV-Photodetector device.
9:00 PM - YY8.11
Spin Relaxation Time in Group-IV Materials from First-principles.
Oscar Restrepo 1 , Wolfgang Windl 1
1 Materials Science and Engineering, The Ohio State University, Columbus, Ohio, United States
Show AbstractSemiconductor device technology increasingly meets paradigm-shifting roadblocks in its pursuit of lower power and higher performance through the traditional scaling approach. This has motivated increased research on replacement technologies. Among those, spintronics devices use the spin of electrons (up or down) as a binary code for information, instead of the on-off states of charge based devices. In order to fabricate such a device, spin-polarized electrons need to be injected, transported, and detected, while making sure that the spin-state is controlled and maintained during the entire time. Thus, the ultimately achievable spin lifetime is a make or break question of spintronics technology. In this talk we will present a recently developed parameter-free first-principles method to determine the electronic spin relaxation times for conduction electrons in systems with inversion symmetry. For graphite, graphene and the indirect-gap semiconductors silicon and diamond, we concentrate on spin relaxation induced by momentum scattering off phonons, which defines the ultimate limit for the spin relaxation time. Favorable comparisons for the cases of silicon and graphite with available experimental values show that this method is able to predict the potential of a material to be a spin conductor and thus is useful for finding and designing the best possible spintronics materials. The predicted spin relaxation times for diamond and graphene are surprisingly short and long, respectively, but can be reasoned to be sensible under closer examination.
9:00 PM - YY8.13
Electron Position Jumping in Double Concentric Quantum Rings.
Igor Filikhin 1 , Sergei Matinyan 1 , James Nimmo 1 , Branislav Vlahovic 1
1 Physics, North Carolina Central University, Durham, North Carolina, United States
Show AbstractSemiconductor heterostructures as quantum dots or quantum rings (QR) demonstrate discreet atom-like energy level structure. In an atom the position of electron can be changed by electromagnetic field influence with accompaniment of quantum number change. In presented work we show that in the weak coupled Double Concentric Quantum Ring (DCQR) the electron position jumping can exist due to level crossing which can place, in contrast to atom. We study DCQR composed of GaAs in an Al0.70Ga0.30As substrate under influence of magnetic field. In our model the DCQR is considered within three dimensional single sub-band effective mass approache [1]. Magnetic field is applied in the z direction, perpendicular to the DCQR plane. In Fig.1 the results of the numerical calculation for DCQR are presented. The quantum numbers of single electron states are n (n=1,2,3…) and l (orbital quantum number, |l|=0,1,2…). The electron position in DCQR is defined by effective radius. It is radius of most probable localization of the single electron. For example, for DCQR (with inner ring width (8 nm), outer ring width (18 nm), H = 4 nm, and inner radius of 5 nm) the position of the electron can be changed by jumping in an increasing magnetic field at the value 9 T. For this “jump value” there is crossing of two levels (1,-1) and (2,-1).We also study the electron structure of QD located at center of QR. The electron position jumping effect is considered. We discuss possibility of experimental implementations for the effect in this composite object.This work is supported by the NSF (HRD-0833184) and NASA (NNX09AV07A).[1] I. Filikhin, V. M. Suslov and B. Vlahovic, Phys. Rev. B 73, 205332 (2006).
9:00 PM - YY8.2
Vacancy-mediated Diffusion in Biaxially Strained Si: A Cross-analysis of Ab Initio Calculations and Experimental Results.
Damien Caliste 1 2 , Konstantin Rushchanskii 1 , Pascal Pochet 1 2
1 INAC/L_Sim, CEA, Grenoble France, 2 , UJF, Grenoble France
Show AbstractUnderstanding the influence of stress on diffusivity in Si based alloys is a key challenge in modern semiconductor technology. For layered materials, like SiGe alloys used in electronic, the lattice mismatch between Si and Ge, makes the connection between inter-diffusion and stress of particular importance. In recent years, it has been shown that external strain can enhance or retard diffusion of impurities as well as self-diffusion of the alloys components [1-3]. The parameter which describes the effect of strain on the diffusivity is denoted Q' as the strain derivative of the activation energy Q. Experimentally, two measurement methods have been used in the literature [1, 2].We will summarize the different approaches to Q' calculation and clarify its analytical formulation. This new analysis of published experiments of silicon self-diffusion in SiGe layers under stress will highlight that Q' is different between tensile and compressive environments. While current experimental values are scattered between 10 and 100, we will show that our new analysis makes all these results consistent. We will also propose an approach of the strain dependency of Q' based on DFT calculations. It relies on the strong non-linear stress dependence of the vacancy barrier, as represented by the split vacancy. We also consider the entropic contribution of the other paths describing the vacancy jump. We attribute the Q'-strain dependency mainly to an electronic effect in the pairing around the split vacancy.[1] N. E. B. Cowern, P. C. Zalm, P. van der Sluis, D. J. Gravesteijn, and W. B. de Boer, Phys. Rev. Lett. 72, 2585 (1994).[2] A. Y. Kuznetsov, J. Cardenas, D. C. Schmidt, B. G. Svensson, J. L. Hansen, and A. N. Larsen, Phys. Rev. B 59, 7274 (1999).[3] N. R. Zangenberg, J. Lundsgaard Hansen, J. Fage-Pedersen, and A. Nylandsted Larsen, Phys. Rev. Lett. 87, 125901 (2001).
9:00 PM - YY8.3
Electronic Structure Calculations Using a Modified Thomas-Fermi Approximation.
Gregory Dente 1 , Michael Tilton 2 , Andrew Ongstad 3
1 , GCD Associates, Albuquerque, New Mexico, United States, 2 , Boeing, Kirtland AFB, New Mexico, United States, 3 , Air Force Research Laboratory, Kirtland AFB, New Mexico, United States
Show AbstractIn order to obtain a reasonably accurate and easily implemented approach to many-electron calculations, we will develop a new Density Functional Theory (DFT). Specifically, we derive an approximation to electron density, the first term of which is the Thomas-Fermi density, while the remaining terms substantially correct the density near the nucleus. As a first application, this new result allows us to accurately calculate the details of the self-consistent ion cores, as well as the ionization potentials for the outer s-orbital bound to the closed-shell ion core of the Group III, IV and V elements. Next, we demonstrate that the new DFT allows us to separate closed-shell core electron densities from valence electron densities. When we calculate the valence kinetic energy density, we show that it separates into two terms: the first exactly cancels the potential energy due to the ion core in the core region; the second represents the residual kinetic energy density resulting from the envelopes of the valence electron orbitals. This kinetic energy cancellation in the core region and the residual valence kinetic energy term allow us to write a functional for the total valence energy dependant only on the valence density. This equation provides the starting point for a large number of electronic structure calculations. Here, we use it to calculate the band structures resulting from the self-consistent valence density and potential on the zinc-blende and diamond lattices. We will show results for the Group III-V and Group IV semiconductors.
9:00 PM - YY8.4
Ab Initio Calculation of the Binding Energy of Impurities in Semiconductors: Application to Silicon Nanowires.
Yann-Michel Niquet 1 , Luigi Genovese 2 , Christophe Delerue 3 , Thierry Deutsch 1
1 INAC/SP2M/L_Sim, CEA, Grenoble France, 2 , ESRF, Grenoble France, 3 ISEN, IEMN, Lille France
Show AbstractThe binding energy Eb of donors and acceptors is a key quantity in semiconductor physics because it determines the doping efficiency. So far, the calculation of Eb in bulk semiconductors has been possible only with semi-empirical methods. However, calculations based on density functional theory (DFT) have become practicable in ultimate nanostructures with a smaller number of atoms. Recently, the case of donors in Si nanowires has been addressed with both semi-empirical methods and DFT, with contradictory results. Tight-binding and effective mass calculations [1,2], supported by experiments [3], indeed suggest that Eb increases as 1/R with decreasing wire radius R, due to the interaction of the bound electron with the surface polarization (or "image") charges of the impurity, resulting in a significant decrease of the doping efficiency in the R < 10 nm range. In contrast, DFT calculations predict that Eb decreases much faster than 1/R, and is about 3-4 times lower [4,5]. Here, we discuss the binding energy Eb of impurities in semiconductors within DFT and the GW approximation. Using the insight gained from many-body perturbation theory, we show that present DFT approaches (including hybrid functionals), involving the Kohn-Sham Hamiltonian of the neutral donor, cannot predict Eb correctly in bulk and nanostructures, because they miss most of the interactions of the carriers with the polarization charges of the impurity. We trace this deficiency back to the lack of screened exchange in the present functionals. We propose an alternative strategy based on the Kohn-Sham Hamiltonian of the ionized donor which circumvents this difficulty (even in the simplest local density approximation). We support these conclusions with DFT calculations on Si nanowires. This new approach [6] can be applied to other charged defect bound states in solids, where it provides a reasonable substitute for much more expensive GW calculations.[1] M. Diarra, Y. M. Niquet, C. Delerue and G. Allan, Phys. Rev. B 75, 045301 (2007).[2] B. Li, A. F. Slachmuylders, B. Partoens, W. Magnus and F. M. Peeters, Phys. Rev. B 77, 115335 (2008).[3] M. T. Björk, H. Schmid, J. Knoch, H. Riel and W. Riess, Nat. Nanotechnol. 4, 103 (2009).[4] R. Rurali, B. Aradi, T. Frauenheim and A. Gali, Phys. Rev. B 79, 115303 (2009).[5] C. R. Leao, A. Fazzio and A. J. R. da Silva, Nano Lett. 8, 1866 (2008).[6] Y. M. Niquet, L. Genovese, C. Delerue and T. Deutsch, Phys. Rev. B 81, 161301(R) (2010).
9:00 PM - YY8.5
Reaction of Bis–diethylaminosilane with –OH on Si (001) Surface.
Seung-Bin Baek 1 , Dae-Hee Kim 1 , Yeong-Cheol Kim 1
1 Materials Engineering, Korea University of Technology and Education, Cheonan, Chungnam, Korea (the Republic of)
Show AbstractThe reaction of bis–diethylaminosilane (SiH2[N(C2H5)2]2, BDEAS) with –OH on Si (001) surface was studied using density functional theory. When the N atom of Si–N bonds in BDEAS reacted with the H atom of –OH on the surface, an energy barrier of 0.62 eV was required to produce a diethylaminosilane (–SiH2[N(C2H5)2]) group and diethylamine (NH(C2H5)2). When the H atom of Si–H bonds in BDEAS reacted with the H atom of –OH on the surface, an energy barrier of 1.18 eV was required to produce bis–diethylaminosilane (SiH[N(C2H5)2]2) group and H2. Since bond energies of Si–N and Si–H bonds in BDEAS were 4.48 and 4.77 eV, respectively, the difference between the bond energies could not explain the difference between the energy barriers. Since there was a bond energy of 0.2 eV between NBDEAS●●●Hsurface due to lone pair electrons in the N atom, but no bond energy between HBDEAS●●●Hsurface, the further reduction of the energy barrier for the formation of DEAS group could be explained.
9:00 PM - YY8.6
Scattering Simulations Based on FEM Strain Field Calculations for Nanofocus X-ray Diffraction at SiGe/Si(001) Nanostructures.
Martin Dubslaff 1 , Michael Hanke 1 , Sebastian Schoeder 2 , Manfred Burghammer 2 , Torsten Boeck 3 , Jens Patommel 4
1 , Paul Drude Institute for Solid State Electronics, Berlin Germany, 2 , European Synchrotron Radiation Facility, Grenoble France, 3 , Leibniz Insitute for Crystal Growth, Berlin Germany, 4 , Dresden University of Technology, Dresden Germany
Show AbstractIn recent years highly brilliant hard X-ray beams with spot sizes focussed down to the range of several 100 nm became feasible at third generation synchrotron radiation sources. This upcoming technique provides novel methods for semiconductor nanostructure analysis like scanning transmission X-ray microscopy, microtomography or scanning X-ray nanodiffraction. Conventional high resolution X-ray diffraction (HRXRD) with μm-sized or even mm-sized beams at nanostructured materials yields reciprocal space maps containing only a statistical average of many objects whereas X-ray diffraction by means of nano-focussed beams enables us to probe and compare individual nano-scaled objects. HRXRD can provide information, i.e., about shape, positional ordering, strain and chemical composition of the examined nanostructures.A very essential tool to support or disprove interpretations of the diffusely scattered intensities in reciprocal space maps are three-dimensional strain field calculations by means of finite element method (FEM) in combination with scattering simulations since the diffracted intensities obtained from an experiment are influenced by a plenty of experimental and sample-inherent properties.We investigated individual nano-scaled objects and parts of 1-μm-sized structures, such as individual side facets of SiGe islands and small arrangements of several (two or up to six) SiGe dots, so-called 'dot molecules'. The SiGe islands are good test-objects for the rather new method of hard X-ray nanodiffraction since this sample system has already been studied intensively by conventional HRXRD. The orientation of the dot molecules and the positional correlation between the individual dots within a dot molecule has been successfully investigated using scattering simulations based on FEM models for each dot molecule configuration. In addition the influence of the spot size and shape on the diffraction simulations has been examined.
9:00 PM - YY8.7
Ab Initio Studies of the Structural Properties and of the Auger Recombination Rate of ZnMgO Alloys.
Markus Heinemann 1 , Marcel Giar 1 , Christian Heiliger 1
1 I. Physikalisches Institut, Justus Liebig University, Giessen Germany
Show AbstractWe present ab initio calculations of structural parameters as a function of composition. The a lattice constant is slightly increasing and c is slightly decreasing in such a way that the volume is fixed. The origin is that pure MgO has almost the same volume like pure ZnO. In addition, we investigate the Auger recombination rate as function of the Mg content. It turns out that there is a peak in the rate at about 40% Mg concentration due to an interband transition. We will discuss this observation together with the complicated band structure of the ZnMgO alloy which changes drastically with the composition.
9:00 PM - YY8.9
Two-photon Absorption of Organic Semiconductor Materials Predicted with Density Functional Theory.
Iffat Nayyar 1 2 , Ivan Mikhailov 1 , Artem Masunov 1 2 3
1 NanoScience Technology Center, University of Central Florida, Orlando, Florida, United States, 2 Department of Physics, University of Central Florida, Orlando, Florida, United States, 3 Department of Chemistry, University of Central Florida, Orlando, Florida, United States
Show AbstractThe probability of the simultaneous absorption of pair of photons (two-photon absorption, 2PA) is many orders of magnitude less than that of one photon absorption and requires high power laser beams to be observed. Organic semiconductor materials with large 2PA cross sections are of potential interest for various applications such as deep-tissue fluorescence microscopy, three-dimensional microfabrication, optical data storage, etc. Accurate theoretical methods are therefore helpful in the design of nonlinear optical materials (NLO) exhibiting large 2PA as an alternative to experimental trial and error. Sum over state (SOS) and coupled electronic oscillators (CEO) are two methods implemented by density functional theory (DFT) for 2PA prediction. Linear response (LR) yields the excitation energies and ground to excited state transition dipoles whereas quadratic response (QR) is required to obtain permanent dipoles of excited sates as well as state-to-state transition dipoles. These can be used in SOS formulation for calculation of 2PA cross sections. In QR-DFT, the single (SR) and double residues (DR) of the QR function at the resonant frequencies are used to determine 2PA matrix elements either via SOS or directly. Recently proposed a posteriori Tamm-Dancoff approximation (ATDA) [1] simplifies the calculation of 2PA cross sections. Here, the state to state transition dipoles are calculated without solving the equations of full QR-DFT. In this study, we investigate the ability of various density functional theory (DFT) formalisms to predict 2PA spectra in a series of large conjugated organic molecules of donor-π-donor type. We access the accuracy of ATDA in prediction of 2PA properties for the benchmark set [2] of molecules and compare ATDA with CEO, QRSR and QRDR results. We conclude that the accuracy of ATDA is close to the exact results obtained in CEO formalism whereas QR-DFT overestimates the cross sections for the molecules under study. We trace the reasons to the lack of double excitations in QR-DFT (unlike CEO and ATDA). As a result, the contribution of the double excitations to the response is distributed over the nearest single excitations. This misrepresents the individual contributions to 2PA response and distorts the qualitative picture. For this reason, QR-DFT cannot be recommended for the essential state analysis, while ATDA can be used for computationally inexpensive but accurate predictions of NLO properties of organic materials. We also studied the effect of molecular geometry and exchange correlation functionals on 2PA energies and cross sections. The higher fraction of exact exchange and range-separated hybrids yield both quantitavely and qualitatively inaccurate predictions. Adjusting fraction of exact exchange in these functionals results in the improved agreement with experiment. [1] Mikhailov, I. A.; Tafur, S.; Masunov, A. E., Phys. Rev. A 77, 2008, 012510 [2] Masunov, A. E.; Tretiak, S., J. Phys. Chem. B 108, 2003, 899.