Theoretical study of the Cox-Symons model for normal muonium in silicon

A theoretical study of the Cox-Symons model for normal muonium in Si is presented. The calculations are performed using polarized basis set ab-initio Hartree Fock calculations followed by corrections for electron correlation. It is shown that, if lattice relaxations are included, the antibonding site becomes a minimum of the potential energy surface (PES) for neutral interstitial hydrogen. The energy at this minimum is lower than that at the undistorted tetrahedral interstitial site. The spin density changes from being almost entirely on the muon (for Mu at the T site) to being almost entirely on a three-fold coordinated Si atom (for Mu in the AB configuration). The mechanism required to explain the isotropy and magnitude of the observed hyperfine tensor of Mu in c-Si is complicated. Large displacements of some host atoms are needed, and the system must be dynamic. However, this model is the first able to produce a minimum of the PES together with an isotropic hyperfine interaction and a delocalized spin density.     

Lattice relaxation for normal muonium in diamond

The effects of symmetric lattice relaxation around the tetrahedral (T) and the hexagonal (H) interstitial sites are calculated for normal muonium (Mu) in diamond using the non-semiempirical method of Partial Retention of Diatomic Differential Overlap (PRDDO). We use clusters containing 3 and 4 host atom shells around the T and the H sites (C₂₀H₃₂ and C₃₀H₄₀). The only significant relaxations are the outward displacement of the 6 nearest neighbors (NN) to the H site and, when the muon is at the T site, the simultaneous outward displacement of the first and second NN to the T site. In the case of diamond, the effects of these relaxations on the energy profiles and spin densities are small.     

Electronic structure and hyperfine interaction of muonium in semi-conductors

The Unrestricted Hartree-Fock self-consistent field cluster procedure is being utilized for first-principle investigations of the electronic structures and hyperfine interactions in normal and anomalous muonium states in semi-conductors. Our results for the total energy for the normal muonium state for a twenty-seven atom cluster in diamond, including the muonium and its neighboring atoms, show a minimum at the tetrahedral site and a maximum at the hexagonal site indicating that normal muonium is located in the tetrahedral region and avoids the hexagonal region. Using the calculated spin-density as a function of the position of muonium and carrying out averaging over the vibrational motion of the muon governed by the total energy curve obtained from our work, we have derived a muon hyperfine constant which is about 75% of that in free muonium, in good agreement with experiment. The natures of the total energy and spindensity curves permit us to draw conclusions regarding the origin of the observed trend in the hyperfine constants for normal muonium in diamond, silicon and germanium. The UHF cluster procedure is also applied to study a model of a muon in a positively charged environment for the anomalous muonium center in diamond. This model leads to a hyperfine interaction tensor with the observed feature of strong anisotropy but significantly weaker than experiment. The results obtained for this model indicate the importance for the anomalous muonium state with its relatively weak hyperfine interaction, of exchange polarization effects inherent in the UHF procedure.     

Hartree-Fock cluster investigation of location and hyperfine properties of normal muonium in silicon-trend with respect to diamond

Using the first-principles Hartree-Fock Cluster procedure employed earlier for normal muonium (Mu) in diamond, the total energy and hyperfine field at the muon site in silicon have been studied as a function of muon position along the <111> direction. The muon was found to be localized in the tetrahedral interstitial region, although the potential was significantly shallower as compared to diamond. The vibrationally averaged hyperfine constant for the muon shows a correct trend compared to diamond but is somewhat larger than experiment, possible reasons for which will be discussed. Results for the superhyperfine constants in silicon will be presented and compared with those for diamond.     

Electronic structure of bond-centred hydrogen and muonium in III–V compound semiconductors: Ab initio cluster calculations

The molecular orbital model for bond-centred hydrogen or muonium in the III–V compound semiconductors is developed with the help of ab initio cluster calculations. The influence of the loss of symmetry in going from the elemental (group IV) to the compound (III–V) materials on the electronic structure is studied. The equilibrium configurations, potential energy surfaces and electronic structures of hydrogen or muonium near the bond-centred site in GaAs, GaP and InP are calculated at the ab initio HF level in the clusters Ga₄As₄H₁₈ and the corresponding one for GaP and InP using a split-valence basis set and ab initio pseudopotentials for the core orbitals. First results of the calculations using a large Ga₂₂As₂₂H₄₂ cluster are discussed. Preliminary results for InP indicate that ionization of the bond-centre defect may be considerably easier than in the other III–V compounds. which would explain why μSR-signals corresponding to the neutral Mu^* state have not been detected in this material.     

Self-consistent field cluster study of hyperfine properties of normal and anomalous muonium in elemental semiconductors

Using the Unrestricted Hartree Fock (UHF) Cluster Procedure, it is shown that for the normal muonium (Mu) center, the tetrahedral site is the most favorable in the two systems diamond and silicon investigated, while for the anomalous muonium (Mu^*) center, a site displaced in the <111> direction with respect to a vacancy in a double-positively charged environment is the appropriate one for all three elemental semiconductors. Using our calculated electronic wave-functions, one is able to explain all features of the observed hyperfine properties of both centers and, in a number of cases, obtain good quantitative agreement with experiment.     

Theoretical Investigation of Muon and Muonium Trapping and Hyperfine Interactions in the Chemical Ferromagnet p-NPNN (β-phase)

The trapping sites for muon and muonium in β-phase ferromagnetic p-NPNN have been determined by the first-principles Unrestricted Hartree–Fock procedure. Four trapping sites are found for the muon near the two nitrogen and two oxygen atoms of the two NO groups. For the singlet state of trapped muonium, two trapping sites are found near the two oxygens of two NO groups and for the triplet state two trapping sites are found near the two oxygens of the NO₂ group. The observed μSR signal at zero field with frequency 2.1 MHz is assigned to the singlet muonium sites near the two oxygens of the two NO groups and the high frequency signal ascribed to an isotropic hyperfine constant of 400 MHz is assigned to the two trapped muon sites near the two nitrogen atoms of the two NO groups. This revised version was published online in September 2006 with corrections to the Cover Date.     

Zero field isotropic Muonium Kubo — Toyabe relaxation functions

Static zero field Gaussian Kubo — Toyabe relaxation functions for muons in isotropic muonium atoms are presented. That is, as with diamagnetic muons, an average of the spin dynamics of a muon in an isolated isotropic ground state muonium atom is taken over an isotropic Gaussian continuous classical local random magnetic field distribution. This motion approximates the exact quantal spin dynamics generated by the dipole-dipole interactions between the muonium atom and the surrounding nuclear spins associated with the site at which the muonium atom has stopped. Expressions are derived for triplet muonium only since, in general, singlet muonium is not observed. For normal nuclear spins and ground state muonium, the resulting relaxation functions are identical to the standard diamagnetic function (except for a shift in the time scale).     

Quantum state of muon and proton in crystalline silicon

Quantum state and dynamics of muon and proton in crystalline silicon have been studied by solving one‐particle Schrödinger equations for the impurities. The ground state wavefunctions and energies are determined as a function of local distortion of the host Si lattice, where the Si–H interaction we used is that parameterized by Ramírez and Herrero following the results of earlier pseudopotential‐density‐functional calculation. It is shown that quantum zero‐point motion of muon induces an effective potential barrier between the tetrahedral‐symmetry (T) site and the bond‐center (BC) site, which ensures bistability of muonium states observed in experiments. It is also shown that, if we fully consider the quantum effect with the present model potential, the BC site becomes less stable than the T site on the contrary to the experiments and former theoretical calculations.     

A comparative study of muonium states in crystalline and amorphous hydrogenated silicon

A study of muons implanted into amorphous hydrogenated silicon (a-Si: H), using both transverse and longitudinal field μSR, is presented. Particular use is made of the muon repolarization curves in longitudinal fields. By comparison with the results of similar measurements on polycrystalline silicon, both the diamagnetic and Mu^* fractions are found to be substantially increased. We postulate that weak strained bonds in the amorphous structure are responsible. Little evidence has been found from longitudinal field measurements for isotropic muonium Mu', and a transverse field experiment on a-Si: D suggests that this state might not exist in the amorphous material.     

Equivalent states of muonium and implanted hydrogen in silicon (experiment)

Main experimental data on the hydrogen-like states with an anisotropic hyperfine structure forming in silicon single crystals in the implantation of high energy muons and protons are presented. The characteristics of the “anomalous” muonium (Mu^*) and hydrogen-containing silicon AA9 states studied by the muon spin rotation (μSR) and ESR techniques in silicon with a due inclusion of the isotope effect are shown to be similar, thus suggesting the existence of two equivalent structures in silicon, Mu^* and AA9, differing only in the mass of the paramagnetic center.     

Muonium states in silicon carbide

Implanted muons in samples of silicon carbide have been observed to form paramagnetic muonium centers (μ ⁺ e⁻). Muonium precession signals in low applied magnetic fields have been observed at 22 K in a granular sample of cubic β-SiC, however it was not possible to determine the hyperfine frequency. In a single crystal sample of hexagonal 6H-SiC, three apparently isotropic muonium states were observed at 20 K and two at 300 K, all with hyperfine frequencies intermediate between those of the isotropic muonium centers in diamond and silicon. No evidence was seen of an anisotropic muonium state analogous to the Mu^* state in diamond and silicon.     

Hyperfine interactions of muonium and hydrogen in semiconductors: Cluster calculations

The electronic structure of muonium (Mu) located at different interstitial sites of the silicon crystal is calculated by the complete neglect of differential overlap (CNDO) and intermediate neglect of differential overlap (INDO) methods. Calculations of the electronicg- and hyperfine interaction tensors of the impurity atom are performed. The results obtained are compared with the experimental properties of both “normal” (Mu′) and “anomalous” (Mu^*) muonium centers. It is shown that the most likely dynamic model for Mu′ is that in which neutral Mu diffuses rapidly in the silicon lattice, whereas for Mu^* it is the model wherein interstitial Mu is located at the bond-center site. A correlation is made between the characteristics of the hydrogen-bearing Si-AA9 center, recently observed by EPR in a silicon crystal, and those of Mu^*. The Si-AA9 center is shown to be a hydrogen-bearing paramagnetic analogue of the Mu^* center.     

Measurement of17O quadrupole interactions and identification of the diamagnetic fraction in ice by muon level crossing resonance

By means of Level Crossing Resonance in a sample of ice which is enriched in H₂ ¹⁷O, the final diamagnetic state of implanted positive muons is determined to be the muonium-substituted molecule HMuO, accommodated in the regular and fully relaxed Ih structure. The¹⁷O quadrupole coupling constant is measured to be 6.1 MHz at 200 K assuming an asymmetry parameter close to unity, a decrease of about 5% relative to that in normal ice Ih at 77 K. The isotope effect is attributed to a greater polarization in the vicinity of a muonium (as opposed to a normal hydrogen) bond. At 50 K, an additional resonance is observed which could correspond to a precursor state, so far not definitely identified. One possibility is a muon trapped at a Bjerrum L-defect, giving a {H₂O−Mu−OH₂}⁺ species with an,¹⁷O quadrupole coupling constant of 8.2 MHz and asymmetry parameter of 0.55. Above this temperature, the fall in the (Gaussian) line-width parameter is attributed to the increasing rate of proton or muon migration, the correlation time dropping from 4 μs at 80 K to 1 μs near the melting-point. The increase in the diamagnetic fraction with rise in temperature is attributed to the increasing proportion of trapping sites available for muon capture.     

Diffusion of muonium and hydrogen in diamond

Jump rates of muonium and hydrogen in diamond are calculated by quantum transition-state theory, based on the path-integral centroid formalism. This technique allows us to study the influence of vibrational mode quantization on the effective free-energy barriers DeltaF for impurity diffusion, which are renormalized with respect to the zero-temperature classical calculation. For the transition from a tetrahedral (T) site to a bond-center (BC) position, DeltaF is larger for hydrogen than for muonium, and the opposite happens for the transition BC-->T. The calculated effective barriers decrease for rising temperature, except for the muonium transition from T to BC sites. The calculated jump rates are in good agreement with available muon spin rotation data.     

Study of condensed nitrogen by μSR method

The temperature dependences of parameters of the muon spin relaxation in liquid and crystalline nitrogen have been studied. It has been established that in condensed nitrogen there takes place a fast depolarization of muons. An anomalous behaviour of the amplitude and phase of muon precession is found in the vicinity of the orientation phase transition in solid nitrogen. It has been shown that muon spin relaxation parameters in nitrogen do not change at reduction of the oxygen impurity content from 0.7·10⁻⁴ to 10⁻⁶. The fast depolarization of muons in condensed nitrogen is apparently due to the formation of muonium atoms. To explain the phenomena observed, a model of the muonium chemical reaction is proposed. The initial phase of the muon precession has been measured as a function of the perpendicular magnetic field to determine the state of short-lived muonium in nitrogen. It has been determined that muonium in nitrogen is in an excited state. Consideration of the nuclear hyperfine interaction of muonium in condensed nitrogen makes it possible to give a qualitative explanation for the temperature dependence of the initial amplitude of the muon precession.     

Muon spin-lattice relaxation and muonium diffusion in ice

Longitudinal muon spin relaxation is measured in ice, using samples with and without enrichment in H₂ ¹⁷O, with a view to studying the mobility of the muonium fraction. A conventional analysis of the data, on the assumption that relaxation of the diamagnetic fraction is negligible, suggests that more than one mechanism of muonium relaxation is at work. A Bayesian analysis warns that separation of the diamagnetic and paramagnetic signals may not be so straightforward.     

Electric-field dependence of muonium formation in liquid helium

The influence of electric fields on the formation of muonium in liquid helium (⁴He,³He, and mixture of⁴He + 0.2%³He) has been studied. It was found that the relative distribution of muon-electron pairs is anisotropic. The maximum muon density is shifted with respect to the electrons in the direction of the initial muon momentum. Due to the anisotropy the muonium asymmetry in normal liquid helium is enhanced by a factor of 3 in an electric fieldE=1 kV/cm.     

Muonium states in semiconductor crystals: Quantum-chemical calculations

The electronic structure of muonium (Mu) located at the bond-centered sites of the silicon and diamond crystals is calculated by the intermediate neglect of differential overlap method. Calculations of the electronicg- and hyperfine interaction tensors of the impurity atom are performed. The results obtained are compared to the experimental properties of “anomalous” muonium Mu^*. It is shown that the properties of Mu located at the bond-centered sites of the Si and C lattices are in qualitative agreement with the observed properties of Mu^*.     

Comparative Theoretical Study of Hyperfine Interactions of Muonium in A- and B-Form DNA

Muon Spin Relaxation (μSR) experiments in A- and B-form DNA have shown evidence for an enhanced electron mobility in the more closely-packed A-form. Besides dynamic effects (electronic diffusion) that could cause the observed difference in muon spin relaxation, one should also carefully examine the difference in the strengths of the hyperfine interactions of the muon (μ ⁺) with the moving electron in the two forms of DNA, since this could contribute to the observed difference in the muon spin relaxation rates as well. We have therefore investigated the (static) trapping properties of muon and muonium (μ ⁺ e ⁻) in A-form and B-form DNA from first-principles with the aim to understand how the different structural geometries of A- and B-form DNA can influence the hyperfine interaction of trapped muonium.