Spectral shapes of Lennard-Jones chains
A. Cuccoli, V. Tognetti, A. A. Maradudin, A. R. McGurn, and R. Vaia
Phys. Rev. B 48, 7015 (1993) - Published 1 September, 1993
A. Cuccoli, V. Tognetti, A. A. Maradudin, A. R. McGurn, and R. Vaia
Phys. Rev. B 48, 7015 (1993) - Published 1 September, 1993
We show how the results of classical molecular-dynamics simulations can be used to improve the calculation of quantum spectral densities of anharmonic crystals. In particular we show that the spectral density C(k,ω) of the displacement correlation function for a quantum chain of N atoms interacting through a nearest-neighbor Lennard-Jones potential, can be calculated accurately in the following way. C(k,ω) is expressed in the form of a continued fraction, whose coefficients are given in terms of its even-frequency moments. The latter, up through the sixth, are calculated from an effective potential that includes the effects of quantum fluctuations. The continued fraction is then terminated by the use of a termination parameter that is determined from a fit of the same continued fraction to the spectral density calculated by means of a classical molecular-dynamics simulation.
S. G. Fedorenko, A. I. Burshtein, and A. A. Kipriyanov
Phys. Rev. B 48, 7020 (1993) - Published 1 September, 1993
The impurity quenching of incoherent excitation accelerated by resonance migration in solid solutions is studied. It is shown that a lower bound for the survival probability N(t) is established by so-called encounter theory, which is linear in the quencher concentration and predicts an exponential decay at the end of the migration-accelerated stage following a short static decay. Another lower bound, which is the probability of not leaving the site where the excitation was initially created, decays much more slowly at the very end of a process. The crossover point of these lower bounds establishes an upper time boundary for binary approximation. The real quenching initially well described by binary-encounter theory is completed by the ‘‘fluctuation asymptotics’’ of quenching, which is multiparticle in principle. In the hopping quenching limit, the continuous-time random-walk equation is used to find the starting point of the fluctuation stage and the share of excitations that have survived at that time N(). When migration is fast enough the latter is shown to be small, and the binary description of preceding stages proves to be good. However, with a decrease of the donor concentration, an intermediate migration-accelerated stage is ousted by multiparticle static quenching from one side and the fluctuation asymptotics from another.
Onuttom Narayan and Daniel S. Fisher
Phys. Rev. B 48, 7030 (1993) - Published 1 September, 1993
The dynamics of driven interfaces in random media are analyzed, focusing on the critical behavior near the depinning threshold. The roughening exponent in the critical region is shown to be independent of the type of disorder, in contrast to the equilibrium static behavior where there are two different universality classes corresponding to random-bond and random-field disorder. This critical dynamic roughening exponent is argued to be equal to its equilibrium static random-field value: =ε/3 to all orders in ε in 5-ε dimensions. All other critical exponents are obtained in terms of z, the dynamic exponent, which is calculated to O(ε) to be z=2-2ε/9+O(). The results agree fairly well with recent numerical simulations. For random-field disorder, the same results have been obtained earlier by Nattermann et al. [J. Phys. (France) II 2, 1483 (1992)] to O(ε). The results above threshold are used, together with scaling laws, to yield conjectures on the critical behavior as threshold is approached from below. In particular, the probability that the diameter of an ‘‘avalanche’’ exceeds l decays as just below threshold with κ=d-3+ζ.
Bo E. Sernelius
Phys. Rev. B 48, 7043 (1993) - Published 1 September, 1993
I present exact results for the free-polaron absorption up to order . The results are obtained from the current-current correlation function. All contributions involving one and two phonons are included. The calculation is valid for zero temperature and low electron concentration. I have managed to get the results in closed analytical form. I also separate the contributions into self-energy terms, vertex terms, and others and find that they all are of great importance. The results are compared to those from earlier calculations.
A. P. Mills, Jr. and N. Karl
Phys. Rev. B 48, 7050 (1993) - Published 1 September, 1993
Based on Doppler-shift measurements we have determined the positron drift velocity u in a high-purity monoclinic α-perylene single crystal as a function of applied electric field F and temperature. The electric field is applied as a triangular wave with a maximum field . At low fields the drift velocity displays a linear field dependence, while it assumes a sublinear field dependence above a characteristic velocity =50 km/s and finally tends to saturate at 110 km/s, presumably due to optical-phonon generation above a certain threshold kinetic energy. Unlike in the case of diamond, is much greater than the longitudinal sound velocity in the solid. By fitting the observed nonlinear electric-field dependence of u to a Shockley expression for acoustic deformation potential scattering of ‘‘warm’’ charge carriers we extract the zero-field limit of the positron mobility along the crystallographic c’ axis (c’∥a×b). At 297 K =(136±3±14) , where the first error is statistical and the second is an estimated ±10% calibration uncertainty. Over the temperature range 100–350 K the mobility exhibits a temperature dependence with n=-1.04±0.03, showing a clear departure from the dependence one might expect. Below 100 K still increases with decreasing temperature, but at a given temperature its value decreases as the maximum applied field increases, possibly indicating interference caused by the presence of a field-enhanced accumulation of trapped carriers that cause scattering at low temperatures.
Below 50 K the limit of as →0 does not further increase but reaches a maximum value of ≊1000 , an upper limit which is presumably set by the presence of residual impurities. These data will be compared with positron mobility results obtained for anthracene.
Robert N. Schwartz and Barry A. Wechsler
Phys. Rev. B 48, 7057 (1993) - Published 1 September, 1993
We have studied the electron-paramagnetic-resonance (EPR) powder spectra of transition-metal-doped processed under oxidizing and reducing conditions. From the EPR data a relative ionization-energy-level scale for the dopant species in was established. These results along with a thermodynamic analysis that models the effects of temperature and oxygen partial pressure on the average charge states of dopants was used to provide a framework for understanding the role of material processing on the Fermi-level position.
David G. Clarke, Thierry Giamarchi, and Boris I. Shraiman
Phys. Rev. B 48, 7070 (1993) - Published 1 September, 1993
We investigate a number of cases of an impurity spin S coupled to a D-dimensional antiferromagnetic (AFM) background. The qualitative behavior of the impurity susceptibility depends upon the sublattice symmetry of the coupling, whether S is integer or half-integer, and the spectral density of the low-lying AFM excitations. When the AFM background is ordered we show that the appropriate model to consider is a non-Abelian spin-boson model. Within this model, we argue that the symmetrically coupled half-integer spin impurity retains the Curie divergence of its susceptibility for D≥2. In the case of a D=1 spin-1/2 AFM chain the sublattice symmetric impurity behaves in the same way as the Kondo spin in the two-channel Kondo model and thus has a non-Curie ln(/T) divergent susceptibility. This latter case is investigated via bosonization at the Emery-Kivelson strong-coupling line.
R. Boscaino, F. M. Gelardi, and J. P. Korb
Phys. Rev. B 48, 7077 (1993) - Published 1 September, 1993
The decay of transient nutations has been experimentally investigated in S=1/2 spin systems at microwave frequency: E’ centers in silica and [ centers in quartz have been studied. We have found that the damping is well described by a single exponential decay function, as expected from a - model (Bloch model). However, the agreement is only qualitative. In fact the measured decay rate Γ is faster than expected and depends on the driving-field amplitude: it tends to the Bloch value =1/2 in the low-power limit and becomes faster and faster on increasing the input power. In all the cases examined the power dependence of the decay rate is fit well by a simple linear dependence of Γ on the induced Rabi frequency χ. The observed power dependence of Γ cannot be ascribed to the inhomogeneity of χ over the sample volume nor to the radiation damping, since both effects are negligible in our experiments. Other mechanisms, which can, in principle, yield a χ dependence of Γ, e.g., the direct interaction of the driving field with structural two-level systems or the spreading of the spin-field coupling constant, are not compatible with the experimental conditions. So, our results suggest that the homogeneous dephasing time of each isochromat contains an intrinsic term and a χ-dependent one. The latter may originate in a field-induced enhancement of the hyperfine or dipolar interaction; however, neither of these mechanisms completely fits the experimental features. The relationship with the decay properties of other coherent regimes is also discussed.
M. Hayashi
Phys. Rev. B 48, 7086 (1993) - Published 1 September, 1993
Exact formulas for the Mössbauer-absorption spectra are derived on the basis of the dispersion theory, for solid absorbers as well as for systems of interacting fine particles. The formulas give the spectra in a form of a superposition of a great many Lorentzian lines of the natural width. These formulas can be written in an approximate form, according to which the spectra are expressed as a superposition of several spectral components that are related with different categories of vibrational modes. From the exact and the approximate formulas, information about the detailed structure of the spectra is derived, including the intensities, linewidths, line shifts, and line shapes of the spectral components. In particular, for the systems of interacting fine particles, we demonstrate the existence of broad lines due to the particle-oscillation modes, observable with Mössbauer spectroscopy. The discussion is extended to multigrade-structured systems of interacting particles. Classical formulas are also derived and compared with the quantum-mechanical ones, leading to a criterion for the validity of the classical theory. Comparison with the experimental data shows that the present theory satisfactorily explains the experiments.
T. Valet and A. Fert
Phys. Rev. B 48, 7099 (1993) - Published 1 September, 1993
By starting with the Boltzmann equation, we calculate the transport properties of magnetic multilayers for currents perpendicular to the layers. Our model takes into account both volume and interface spin-dependent scattering. We show that the macroscopic equations already used by Johnson et al. or van Son et al. are justified if the spin-diffusion length is much longer than the mean free path, even for individual layer thicknesses of the order of the mean free path. But, second, we show that Johnson’s assumption of additive effects from independent interfaces in multilayers is incorrect and we obtain different results by taking into account the interplay between successive interfaces. The simple expressions derived for individual thicknesses much shorter than the spin-diffusion length are in agreement with the analysis of experimental results already proposed. It turns out that data on the perpendicular magnetoresistance can be used to separate clearly the volume and interface contributions to the spin-dependent scattering.
P. Gawiec and D. R. Grempel
Phys. Rev. B 48, 7114 (1993) - Published 1 September, 1993
We study the zero-temperature spin dynamics of a random-exchange two-dimensional XY model using both exact numerical methods and an approach based upon the coherent potential approximation (CPA). The model, which presents a mixed-phase to spin-glass phase transition, consists of a ferromagnetic host with nearest-neighbor bonds J to which one substitutes impurity bonds of strength -λJ at concentration x. In both phases, the long-wavelength magnetic excitations of this system are spin waves with a linear spectrum. The x and λ dependence of the spin-wave velocity is determined numerically by calculating the stiffness constant of the system with a new transfer-matrix algorithm. The stiffness constant and the spin-wave velocity decrease rapidly with increasing x. The two quantities are smooth across the phase boundary, and they saturate in the spin-glass phase. At shorter wavelengths, the q dependence of the energies and lifetimes of the spin waves in the ferromagnetic and spin-glass states differs qualitatively, reflecting the morphological differences that exist between the spin configurations characteristic of the two phases. Line shapes and linewidths depend strongly on the polarization of the excitations. Whereas out-of-plane modes stay propagative at all concentrations, in-plane ones are inhomogeneously broadened and are overdamped at long wavelengths. There is good overall agreement between the exact numerical results and those obtained using the CPA.
A. J. van der Sijs
Phys. Rev. B 48, 7125 (1993) - Published 1 September, 1993
The Heisenberg model, a quantum-mechanical analogue of the Ising model, has a large ground-state degeneracy, due to the symmetry generated by the total spin. This symmetry is also responsible for degeneracies in the rest of the spectrum. We discuss the global structure of the spectrum of Heisenberg models with arbitrary couplings, using group-theoretical methods. The Hilbert space breaks up in blocks characterized by the quantum numbers of the total spin, S and M, and each block is shown to constitute the representation space of an explicitly given irreducible representation of the symmetric group , consisting of permutations of the N spins in the system. In the second part of the paper we consider, as a concrete application, the model where each spin is coupled to all the other spins with equal strength. Its partition function is written as a single integral, elucidating its N dependence. This provides a useful framework for studying finite-size effects. We give explicit results for the heat capacity, revealing interesting behavior just around the phase transition.
P. Crespo, A. Hernando, R. Yavari, O. Drbohlav, A. Garci´a Escorial, J. M. Barandiara´n, and I. Oru´e
Phys. Rev. B 48, 7134 (1993) - Published 1 September, 1993
A ferromagnetic and supersaturated fcc solid solution has been obtained by mechanical alloying. After subsequent thermal treatments the fcc phase undergoes a spinodal decomposition which finally, at 780 K, yields a mixture of fcc and bcc phases. In this work, a systematic magnetic study is carried out on samples at different decomposition states in order to determine the process of transformation into the stable phases. We observe a 20% maximum diminution on the magnetic moment with increasing temperatures of the thermal treatment. The Mo¨ssbauer spectrum taken at 8 K shows that 20% of the Fe atoms are in a nonferromagnetic state. On the other hand, upon heating up to 723 K the room-temperature coercive field increases dramatically to 640 Oe, and after cooling down to 10 K it decreases to 270 Oe. Deviations from the law in the temperature dependence of the magnetization have been observed. This behavior is explained by fluctuations in composition due to the spinodal decomposition, which lead to fluctuations of the magnetic order parameters, i.e., magnetic moment and Curie temperature.
Xin-Zhong Yan
Phys. Rev. B 48, 7140 (1993) - Published 1 September, 1993
The square-lattice extended Hubbard model at half filling is studied by a perturbation scheme. By taking into account the ground state of the mean-field theory as a reference state, the scheme treats the collective fluctuations beyond the mean field as the predominant perturbation. The system at ground state can be a spin-density wave (SDW) or a charge-density wave (CDW) depending on the relative strengths of the on-site electron interaction and the nearest-neighbor interaction. The excitation modes above the ground state are analyzed through diagonalizing an effective Hamiltonian. The energy spectra and the wave functions are self-consistently given by the formalism. By comparing the ground-state energies, we investigate the SDW-CDW transition.
Helena M. Petrilli and Sonia Frota-Pessôa
Phys. Rev. B 48, 7148 (1993) - Published 1 September, 1993
We have applied the recently developed first-principles real-space muffin-tin-orbital scheme to study the electronic structure around Fe impurities in bcc Mo and Nb hosts. These local-spin-density calculations give values for the magnetic moment, hyperfine field, and isomer shift at the impurity site, which are in good agreement with experimental results, when available. We note that, in contrast with reported Korringa-Kohn-Rostoker Green’s-function (KKR-GF) results for the same systems, the magnetic behavior at the impurity is well described when the angular momentum of the wave functions is cut off at =2. Our local impurity moments, calculated using a minimum basis of s, p, and d orbitals, are compatible with those obtained via the KKR-GF formalism, when a cutoff =3 is used.
Tian-shi Liu and Guo-zhu Wei
Phys. Rev. B 48, 7154 (1993) - Published 1 September, 1993
We consider a layered ferromagnetic superlattice with spins and . The exchange energies are and between layers, which are arranged in a quasiperiodic Fibonacci sequence, and J in each layer. When ==J we use a rescaling approach to obtain an exact decimation transformation for local magnon of layers α,β,(γ), and δ(τ) in the ferromagnets. Iteration of the transformation provides numerical results for the local density of states (LDOS) and the magnetization. We found that the bandwidths of the LDOS of layers are the functions of η=/.
D. L. Leslie-Pelecky and J. A. Cowen
Phys. Rev. B 48, 7158 (1993) - Published 1 September, 1993
The electron-spin-resonance (ESR) linewidth ΔH(T) has been measured in multilayered /Cu (x=0.07,0.11) as a function of spin-glass layer thickness (1 nm≤≤1000 nm). For temperatures well above the freezing temperature , ΔH(T)=A+BT, where A is the residual linewidth and B the thermal-broadening coefficient. The residual linewidth increases with decreasing , indicating that the paramagnetic Curie-Weiss temperature FTHETA and/or the inhomogeneous broadening depend on . Preliminary high-temperature-susceptibility measurements confirm that θ decreases with decreasing , but are not sensitive enough to determine quantitatively if this change is entirely responsible for the observed changes in A. The thermal-broadening coefficient B characterizes the strength of the ESR bottleneck. As decreases, B increases, indicating the breaking of the bottleneck via the presence of additional relaxation paths. Possible mechanisms for these paths are suggested. The increase in B with decreasing is similar to that observed with increasing concentration of a third element in bulk spin glasses. Possible links between decreasing and increasing anisotropy are considered. Data parametrized in terms of ε, where ε=[(∞)-()]/(∞), show that both A and B increase linearly with increasing ε over the entire range of ε studied.
Antoine Georges and Werner Krauth
Phys. Rev. B 48, 7167 (1993) - Published 1 September, 1993
A detailed quantitative study of the physical properties of the infinite-dimensional Hubbard model at half filling is presented. The method makes use of an exact mapping onto a single-impurity model supplemented by a self-consistency condition. This coupled problem is solved numerically. Results for thermodynamic quantities (specific heat, entropy, . . .), one-particle spectral properties, and magnetic properties (response to a uniform magnetic field) are presented and discussed. The nature of the Mott-Hubbard metal-insulator transition found in this model is investigated. A numerical solution of the mean-field equations inside the antiferromagnetic phase is also reported.
A. J. Millis
Phys. Rev. B 48, 7183 (1993) - Published 1 September, 1993
I reexamine the work of Hertz on quantum phase transitions in itinerant fermion systems. I determine when it is permissible to integrate out the fermions and analyze the critical phenomena via an effective bosonic theory in which only fluctuations of the ordering field are explicitly retained. By solving appropriate scaling equations I obtain the different regimes of behavior of the correlation length and free energy in the disordered phase of the effective bosonic theory. The results in many cases differ from those of Hertz, but make contact with more recent work on the dilute Bose gas. I briefly discuss the relevance of the results to heavy-fermion materials.
J. Zaanen and A. M. Oleś
Phys. Rev. B 48, 7197 (1993) - Published 1 September, 1993
We address the role played by orbital degeneracy in doped Mott-Hubbard insulators. We observe that in all but the simplest systems the carriers bind to d-d excitons because of Hund’s-rule interactions. These three-particle bound states have distinct spectroscopic signatures and at least in one case these seem already confirmed experimentally. If the crystal-field gaps become of the order of the kinetic energy of the carriers, doping might tend to stabilize phases characterized by a finite occupation of d-d excitons in the ground state. If the total spin of both the carrier state and the spin background are at maximum, the relevant excitons do not involve a change in spin. As a consequence, the orbital channel can be in the first instance considered independently from the spin channel and we find an exciton-carrier coupling that in essence interpolates between the carrier-spin-wave couplings of the t-J model and the conventional couplings to optical phonons. We work out in detail a case involving high-spin holes in a cupratelike system and we show that the exciton-carrier coupling tends to stabilize an orthorhombic type of orbital order. On the other hand, if either the carriers or the background are in a low-spin state, the relevant excitons also change total spin locally and more-exotic order parameters are possible. We analyze in detail the case of a nickelate close to the high-spin–low-spin transition where we show that doping will tend to stabilize an ordering related to superpositions of low-spin and high-spin states, characterized by an overall spin-rotational invariance. We argue that such a state might be realized in n-type .