We have measured the thermal conductivity κ and the thermal diffusivity a of a dense bulk ceramic polycrystalline sample of YBa2Cu4O8 (1:2:4) in the temperature range 30-300 K. We find κ≊10 W m-1 K-1 at 100 K, significantly higher than in ceramic YBa2Cu3O7-δ (1:2:3) and approaching the in-plane value for single-crystal 1:2:3, and decreasing to 7.6 W m-1 K-1 at 300 K. The data for this sample can be described by standard theories for phonon thermal conductivity of crystalline materials with boundary, phonon, and electron scattering. The higher κ in 1:2:4 as compared to 1:2:3 is, in this model, due to the smaller point defect scattering in the former. The fitted parameters for the three scattering mechanisms all agree with independent estimates based on simple models; inserting data for electric resistivity, grain size, carrier density, and lattice properties we can predict κ and its T dependence to within about 20%. We also discuss models for the phonon and electron thermal conductivities in some detail, including some second-order effects such as inelastic electron scattering and a T-dependent carrier density.
We investigate the magnetic properties of a range of low-dimensional ferromagnets using a combination of first-principles calculations and atomistic spin dynamics simulations. This approach allows us to evaluate the ground state and finite temperature properties of experimentally well characterized systems such as Co/Cu(111), Co/Cu(001), Fe/Cu(001) and Fe/W(110), for different thicknesses of the magnetic layer. We compare our calculated spin wave spectra with experimental data available in the literature, and find a good quantitative agreement. We also predict magnon spectra for systems for which no experimental data exist at the moment, and estimate the role of temperature effects
Local-density-functional-based ab initio calculations are used to investigate hydrogen and carbon-hydrogen defects in GaAs. The equilibrium structure for both the C-H and C-H- complexes are shown to be similar, with the hydrogen located at a C-Ga bond-centered site. The dissociation of these complexes is investigated and it is found that the energy barrier of 1.84 eV for the process C-H → C- + H+ is substantially lowered to 0.88 eV in the presence of an electron resonantly bound to the defect. This is in good quantitative agreement with recent experiments. Isolated interstitial hydrogen is found to lie at a Ga-As bond-centered site for both H+ and H0 and at an antibonding site relative to a Ga atom for H-. It is also found that the stable form of the hydrogen dimer is a H2 molecule, the dissociation energy of which is 1.64 eV, and that interstitial hydrogen is a negative-U defect. Finally, a mechanism for minority-carrier-induced device degradation is proposed.
Local vibrational modes of the H2* defect in crystalline germanium are identified by a combination of infrared-absorption spectroscopy, uniaxial stress measurements, and ab initio theory. Germanium crystals are implanted with protons and/or deuterons at 30 K, and subsequently annealed at room temperature. A number of local vibrational modes of hydrogen are revealed by infrared-absorption spectroscopy. In particular, modes at 765, 1499, 1774, and 1989 cm-1 originate from the same defect which has trigonal symmetry according to the uniaxial stress measurements. The 765-cm-1 mode is two dimensional, while the 1774- and 1989-cm-1 modes are one dimensional. Measurements on samples coimplanted with protons and deuterons show that the defect contains a pair of weakly coupled and inequivalent hydrogen atoms. The 765-, 1499-, 1774-, and 1989-cm-1 modes are ascribed to the H2* defect. The 765-cm-1 mode is a Ge-H bend mode with an overtone at 1499 cm-1 and the modes at 1774 and 1989 cm-1 are Ge-H stretch modes. An excellent fit to the stretch frequencies is obtained with a simple model based on two coupled Morse-potential oscillators. In addition, the model gives intensity ratios in fair agreement with those observed. The structure, the local-mode frequencies, and the isotope shifts of H2* are calculated with ab initio local-density-functional cluster theory. The theoretical frequencies are consistently 5-10 % too high, as expected from the theory which often leads to overbinding. The isotope shifts, however, are in fair agreement with observations. These results provide additional support for our assignments, and show that the 765- and 1774-cm-1 modes primarily involve the hydrogen at the antibonding site, while the 1989-cm-1 mode is related mainly to vibration of the hydrogen near the bond-center site.
It has recently been shown that domain walls (DWs) in ferromagnets can be moved in the presence of thermal gradients. In this work we study the motion of narrow domain walls in low-dimensional systems when subjected to thermal gradients. The system chosen is a monolayer of Fe on W(110) which is known to exhibit a large anisotropy while having a soft exchange, resulting in a very narrow domain wall. The study is performed by means of atomistic spin dynamics simulations coupled to first-principles calculations. By subjecting this system to thermal gradients we observe a temperature-dependent movement of the domain wall. The thermal gradient always makes the domain wall move towards the hotter region of the sample with a velocity proportional to the gradient. Our material specific study is complemented by model simulations to discern the interplay between the thermal gradient, magnetic anisotropy, and the exchange interaction and shows that the larger DW velocities are found for materials with broader domain-wall width. The relatively slow DW motion of the Fe/W(110) system is hence primarily caused by its narrow domain-wall width, which results from its large magnetic anisotropy and soft exchange
In this work, the magnetization dynamics of clusters supported on nonmagnetic substrates is shown to exhibit a complex response when subjected to external magnetic fields. The field-driven magnetization reversal of small Co clusters deposited on a Cu(111) surface has been studied by means of first-principles calculations and atomistic spin dynamics simulations. For applied fields ranging from 1 to 10 Tesla, we observe a coherent magnetization reversal with switching times in the range of several tenths of picoseconds to several nanoseconds, depending on the field strength. We find a nonmonotonous dependence of the switching times with respect to the strength of the applied field, which we prove has its origin in the complex magnetic anisotropy landscape of these low-dimensional systems. This effect is shown to be stable for temperatures around 10 K, and is possible to realize over a range of exchange interactions and anisotropy landscapes. Possible experimental routes to achieve this unique switching behavior are discussed
We provide, by a detailed first-principles investigation, evidence for weak electronic correlations in SrRuO 3. The magnetism in SrRuO 3, in terms of the equilibrium magnetization and critical temperature, is well described by the generalized gradient approximation. Including Hubbard-type correlations results in worse agreement with experiment
The inter- and intralayer contributions to the layer-resolved complex optical conductivity tensor for semi-infinite layered systems are calculated in terms of the Luttinger formula within the spin-polarized relativistic screened Korringa-Kohn-Rostoker method. Ab initio Kerr angles are then obtained for arbitrary geometry and incidence via a 2x2 matrix technique including all multiple reflections and all optical interferences. Applied to in-plane single-domain magnetized bcc Ni/Ni(100), it is proven that the assumed appropriate formula of Kerr angles widely used to explain magneto-optical Kerr effect with rotating magnetic field measurements fully agrees with our ab initio Kerr data. From the experimental Kerr data of Tian [Phys. Rev. Lett. 94, 137210 (2005)], however, it cannot be concluded that the deduced magnetic properties apply for bulk Ni, since about 75% of the contributions to the Kerr rotation angle arise from the surface.
Hydrogen-related local vibrational modes at 2643, 2651, 2688, 2725, 2729, and 2775 cm-1 are thought to arise from C-H stretch modes from defects similar to the hydrogen-passivated carbon acceptor complex CAsH. These lines appear in samples that are grown using trimethylgallium metalorganic precursors, and it has been suggested that the 2688-, 2725-, 2729-, and 2775-cm-1 bands may be due to CAs dimers decorated with one or more H atoms. We present here the structures, energies, and vibrational modes of (CAs)2, (CAs)2H, and (CAs)2H2 complexes obtained from ab initio local-density-functional cluster calculations to investigate these assignments.
We have performed an extensive test of the ability of density functional theory within several approximations for the exchange-correlation functional, local density approximation + Hubbard U, and local density approximation + dynamic mean field theory to describe magnetic and electronic properties of SrRuO3. We focus on the ferromagnetic phase, illustrating differences between the orthorhombic low-temperature structure versus the cubic high-temperature structure. We assess how magnetism, spectral function, and cohesive properties are affected by methodology, onsite Hubbard U, and double-counting corrections. Further, we compare the impact of the impurity solver on the quasiparticle weight Z, which is in turn compared to experimental results. The spectral functions resulting from the different treatments are also compared to experimental data. Finally, the impact of spin-orbit coupling is studied, allowing us to determine the orbital moments. In the orthorhombic phase, the orbital moments are found to be tilted with respect to the spin moments, emphasizing the importance of taking into account the distortion of the oxygen octahedra
Carbon impurities implanted into single-crystalline germanium are studied with infrared absorption spectroscopy and ion channeling. After implantation of 12C+ at room temperature and subsequent annealing at 350 °C, a sharp infrared absorption line is observed at 531 cm-1. When 12C+ is substituted by 13C+, the line shifts down in frequency to 512 cm-1 and co-implantation of 12C+ and 13C+ does not give rise to additional lines. Therefore, the 531-cm-1 line represents a local vibrational mode of a defect containing a single carbon atom. Channeling measurements are carried out around the 〈100〉, 〈110〉, and 〈111〉 axes in 12C+-implanted samples annealed at 450 °C. The analysis of the data shows that 31±3 % of the carbon atoms are located at substitutional sites, while the remaining carbon atoms appear to be located randomly. The population of the substitutional site and the intensity of the 531-cm-1 mode have identical temperature dependencies. It is concluded that the 531-cm-1 mode is the three-dimensional T2 stretch mode of substitutional carbon. The effective charge of the mode is determined to be (3.4±0.5)e.mAb initio local density functional cluster theory is applied to calculate the structure and the local vibrational modes of substitutional carbon in germanium. The calculated frequencies and isotope shifts for the T2 stretch mode are in good agreement with the observations.
We study the two-dimensional XY model in a magnetic field from a phenomenological point of view by means of Monte Carlo simulations and a simple analytic approach. We consider the system as consisting of an ideal gas of vortex-antivortex pairs. Within this model we can calculate the density of vortex pairs, the average size of the pairs, the specific heat, and the depletion of the magnetization by the vortices. These quantities are compared with Monte Carlo results in which we have extracted the pure vortex contribution by means of a Lagrange-multiplier technique.
The anharmonicities of the C-H stretch modes in HCN and the passivated C acceptor in GaAs are investigated using ab initio local-density-functional cluster theory. The effective-mass parameter χ for the C-H stretch mode is shown to be less than unity in HCN, and greater than unity for the GaAs case. The calculated anharmonic parameter for the first defect is found to be 106 cm-1 and is in very good agreement with experiment. For the second defect, the anharmonicity is about 50% larger in agreement with empirical estimates. The frequencies of the fundamental transitions in both systems are shown to be very sensitive to the C-H length. This limits the accuracy of theoretical investigations of these high frequency H modes. Finally, the effects of electrical anharmonicity are considered and it is shown that they reduce the intensity of the overtone in the C-H complex in GaAs by about 70%.
The vacancy model for impurity vacancy defects in semiconductors assumes that the ground and low-energy excited states are derivable from the four sp3 hybrid orbitals on atoms bordering the vacancy. There are many cases where this model works but we describe here a counterexample concerning the lowest excited state of the [V-N3] defect in diamond. It is shown that a shallow electron trap, localized outside the vacancy, is involved in the first excited state and responsible for the N2 and N4 optical bands associated with the defect.
A local-density-functional cluster method is used to calculate the structure and vibrational modes of interstitial oxygen in silicon. We find Si-O lengths and the Si-O-Si bond angle to be 1.59 Å and 172°, respectively. The asymmetric and symmetric stretch frequencies are 1104 and 554 cm-1 and are close to observed modes at 1136 and 518 cm-1. The effective charge of the upper mode is 3.5e. Isotopic shifts of the modes are also reported. We find that the symmetric stretch mode at 554 cm-1 is independent of the O-isotopic mass in agreement with observation.
We have carried out local-density functional cluster calculations on the SiGa- defect in GaAs. We find that the distortion proposed by Chadi and Chang involving a large Si movement along 〈111〉 breaking an Si-As bond has the same energy as a simple breathing distortion. The calculated local vibratory modes of one Chadi-Chang structure do not agree with those recently assigned to DX from an infrared-absorption experiment. On the other hand, the calculated triplet mode due to the other structure above is in reasonable agreement with these observations. The effective charge of the second type of defect is about three times that of the first. It is proposed that both defects coexist with the Chadi-Chang one being almost infrared inactive.
The structural and dynamic properties of carbon defects in aluminum arsenide are investigated using first-principles local-density-functional cluster theory. The method accounts satisfactorily for the structure and phonon modes of AlAs. The carbon acceptor and donor possess triplet modes in the band gap between the acoustic and optic branches as well as localized triplet modes. The local mode of the acceptor lies within 40 cm-1 of the observed mode. Four modes of a close-by acceptor pair lie within 40 cm-1 of the local mode of the isolated acceptor and this supports a previous assignment of four satellite lines seen in heavily doped material to this defect. The modes of a donor-acceptor pair are investigated but there is no evidence of their existence.
Carbon is an important impurity in metalorganic molecular-beam-epitaxy-grown GaAs. CAs is a single acceptor that can be passivated by H. We describe local-density-functional cluster calculations on the structure and dynamics of the impurities and passivated complexes. The lowest-energy structure of the passivated acceptor is a H atom located 1.1 Å from C in a bond-centered orientation. The H-Ga distance is 2.1 Å. The other three C-Ga lengths are 2.18 Å. The H stretch frequency is found to be 2605 cm-1 and is observed at 2635 cm-1. We have also calculated C-H bend modes that should be visible in Raman but not in infrared experiments. The activation energy for the reorientation of the complex is 0.67 eV. Also described are the local modes of the two substitutional C defects in addition to a C-C pair. Modes of the latter are found around 553 and 425 cm-1, respectively, and have effective charges of about 0.5.
Interstitial carbon, Ci, defects in Si exhibit a number of unexplained features. The Ci defect in the neutral charge state gives rise to two almost degenerate vibrational modes at 920 and 931 cm-1 whose 2:1 absorption intensity ratio naturally suggests a trigonal defect in conflict with uniaxial stress measurements. The dicarbon, Cs-Ci, defect is bistable, and the energy difference between its A and B forms is surprisingly small even though the bonding is very different. In the B form appropriate to the neutral charge state, a silicon interstitial is believed to be located near a bond-centered site between two Cs atoms. This must give rise to vibrational modes which involve the motion of both C atoms in apparent conflict with the results of photoluminescence experiments. We use an ab initio local density functional cluster method, AIMPRO, to calculate the structure and vibrational modes of these defects and find that the ratio of the absorption intensities of the local modes of Ci is in reasonable agreement with experiment even though the structure of the defect is not trigonal. We also show that modes in the vicinity of those detected by photoluminescence for the B form of the dicarbon center involve independent movements of the two C atoms. Finally, the trends in the relative energies of the A and B forms in three charge states are investigated.
The electronic structure of bounded intrinsic stacking faults in silicon is studied. Especially the influence of the stacking fault width on the electronic states in the band gap is investigated. The extended defect studied comprises an intrinsic stacking fault with two reconstructed 90° partials as boundaries. The atomic structure is determined by different valence force fields. These are the Keating potential, the bond-charge model, and an anharmonic version of the bond-charge model. The electronic structure is calculated by linear combinations of atomic orbitals. Ten Gaussian-type atomic orbitals of s, p, and d-type are used, and up to fourth nearest neighbor interactions are taken into account. The levels in the band gap are evaluated by the recursion method for nonorthogonal basis functions, and by a continued fraction representation of the local density of states
The interaction of vacancies with 30° and 90° partial dislocations in silicon is examined. In particular, the structures and binding energies are calculated using hydrogen-terminated clusters and local density-functional theory. Moreover the electronic structure is determined using supercells containing dislocation dipoles. Vacancies are found to have binding energies of approximately 2.0 eV and 0.9 eV to 90° and 30° partials, respectively. The elastic strain field of the partials makes the fourfold vacancy reconstruct, which essentially clears the fundamental gap
Using the self-consistent Hartree-Fock-Slater model we have calculated the electronic structure for various Cux(CO)y clusters symbolizing not only CO bound to "on-top" and "bridge" sites but also some lateral CO-CO interaction on a Cu(111) surface. By comparison with experimental photoemission data we are able to reproduce the observed energies of the occupied CO 4σ, 1π, and 5σ orbitals as well as the partly occupied 2π orbital. In our model we assume CO to be adsorbed on "top" sites for coverages less than ⊖=0.33 [(sqrt[3]×sqrt[3])R30°] and on both top and bridge sites for ⊖>0.33. The experimentally observed peak of intensity at the Fermi edge which increases with coverage above 0.33 is in our model explained by the occupation of CO orbitals of the b1 and b2 symmetry types, i.e., "π" orbitals, degenerate for top positions (C3v), split by the change to bridge positions (C2v). Our results further indicate that the broadened 5σ-1π intensity peak at high coverages is a result of CO bound to top and bridge sites.
The local structure of CAs acceptors in AlxGa1-xAs has been investigated by studying the nondegenerate localized vibrational modes of H-CAs pairs with A1 symmetry, rather than those of isolated CAs impurities. Infrared absorption and Raman scattering measurements have been made on AlxGa1-xAs: 12C epilayers that (a) had been exposed to a radio-frequency hydrogen (deuterium) plasma or (b) contained hydrogen incorporated during growth. Arguments are advanced that indicate that the H(D) atom should occupy a bond-centered site between CAs and Ga atoms rather than between CAs and Al atoms at low temperatures. An ab initio local-density-functional calculation indicates that the energy is then lowered by 0.24 eV. This analysis has led to the assignment of five antisymmetric stretch modes and five symmetric (X) modes to H-CAs pairs at sites where the carbon atom that was originally unpaired had zero, one, two, three, or four Al nearest neighbors
Several nitrogen-related centers have been introduced by ion implantation of nitrogen into germanium and studied by infrared-absorption spectroscopy. Two local vibrational modes at 825.3 and 658.6 cm-1 were especially prominent. Measurements on annealed samples implanted with either pure 14N or 15N, or implanted with both isotopes showed these modes to arise from a nitrogen pair defect analogous to one previously suggested to occur in silicon. Based on this and ab initio pseudopotential cluster theory, a model of the pair is proposed which is consistent with the observed infrared absorption. This model is similar to that of the nitrogen pair in silicon and offers an explanation of the low donor efficiency of nitrogen in germanium. Several other nitrogen-related local vibrational modes are also observed in the implanted material.
In a recent paper [Phys. Rev. B 48, 17 806 (1993)] Cunha, Canuto, and Fazzio reported ab initio Hartree-Fock calculations on nitrogen impurities in group-IV semiconductors. In their paper it is suggested that nitrogen pairs form from substitutional atoms on adjacent lattice sites. However, the experimentally observed configuration in silicon and germanium is different from this. The aim of this Comment is to clarify the situation of the nitrogen pair in these materials.
Local density functional theory is used to show that both α and β dislocations in GaAs are reconstructed. This is done by relaxing large 158-atom H-terminated clusters of GaAs containing 90° partial dislocations. The reconstruction is strongly influenced by impurities: acceptor pairs destroy the reconstruction of β partials but strengthen it for α dislocations. Donors have opposite effects. The implication of these results for the pinning of dislocations in GaAs is discussed.
The structures of straight 90° glide partial dislocations in SiC are calculated using an ab initio local density functional cluster method. Si partials containing core Si atoms are found to be strongly reconstructed with a Si-Si bond of comparable length to that in bulk silicon. The C partial possessing core C atoms is more weakly reconstructed with a bond length 16% longer than that in bulk diamond. The formation and migration energies of kinks on the partials are calculated and indicate that the C partial is the more mobile. The calculations also predict that n-type doping leads to an increase in the mobility of C partials whereas p-type doping increases the mobility of Si partials.
We present a combined experimental and theoretical study on electronic and magnetic properties of the Fe(001)-p(1X1)O surface. The ordered p(1X1) surface is investigated with spin-polarized scanning tunneling microscopy and spectroscopy accompanied by first-principles calculations. The atomic registry of the Fe(001)-p(1X1)O surface was confirmed in real space from the atomically resolved images. Tunneling spectroscopy reveals two oxygen induced features in the local density of states, around -0.7 eV and at the Fermi level, the origin of which is discussed based on first-principles calculations. Due to the hybridization of oxygen p(z) states with the Fe states near the Fermi level, the spin polarization in tunneling experiments is inverted upon oxygen adsorption.
We study the linear resistance at the Kosterlitz-Thouless transition by Monte Carlo simulation of vortex dynamics. Finite-size scaling analysis of our data shows excellent agreement with scaling properties of the Kosterlitz-Thouless transition. We also compare our results for the linear resistance with experiments. By adjusting the vortex chemical potential to an optimum value, the resistance at temperatures above the transition temperature agrees well with experiments over many decades.
We use Monte Carlo simulations of a layered XY model to study phase fluctuations in high-Tc superconductors. A vortex-antivortex interaction dominated by a term linear in the vortex separation is found in the low-temperature regime. This is in agreement with a zero-temperature variational calculation. At temperature just above the two-dimensional (2D) vortex-unbinding temperature, the linear term vanishes and an ordinary 2D vortex behavior is found. This explains the finding that high-Tc superconductors show 2D properties in the vortex fluctuations responsible for the resistivity transition close to the critical temperature.
We use Monte Carlo simulations of a two-dimensional XY model in a magnetic field to study a self-consistent mean-field theory for the three-dimensional anisotropic XY model. The relation between the critical temperature Tc and the interplane coupling J⊥ is determined. The magnetization exponent β is discussed and results for the specific heat cv are presented.
We have studied the nonlinear current-voltage characteristic of a two-dimensional lattice Coulomb gas by Monte Carlo simulation. We present three different determinations of the power-law exponent a(T) of the nonlinear current-voltage characteristic, V∼Ia(T)+1. The determinations rely on both equilibrium and nonequilibrium simulations. We find good agreement between the different determinations, and our results also agree closely with experimental results for Hg-Xe thin-film superconductors and for certain single crystal thin-film high-temperature superconductors.
The single kink formation and migration energies Fk and Wm of 90° glide partial dislocations in Si and GaAs are calculated using an ab initio local density-functional cluster method. Kink migration occurs via a concerted exchange of an atom at a dislocation core with one of its glide plane nearest neighbors. By constraining these atoms to sit in high-energy positions and relaxing a surrounding cluster of atoms, sufficient points in configuration space can be sampled for the energy barrier for the first step in kink pair formation to be estimated. By including an estimate of the elastic energy of the interaction of kink pairs, the single kink formation energy is calculated. It is found that Fk and Wm for Si are 0.1 and 1.8 eV, respectively. For the 90° α glide partial in GaAs, these quantities are 0.07 and 0.7 eV, respectively, and 0.3 and 1.1 eV for β partials