Table of contents

We have measured details of the quasi-one-dimensional (Q1D) Fermi-surface sections in the organic superconductor κ-(BEDT-TTF)₂Cu(NCS)₂ and its deuterated analogue using angle-dependent millimetre-wave techniques. There are differences in the corrugations of the Fermi surfaces in the deuterated and undeuterated salts. We suggest that this is important in understanding how deuteration affects the superconducting transition temperature. The data suggest that the 'nestability' of the Q1D Fermi sheets may be important in understanding the 'universal' phase diagram of κ-(BEDT-TTF)₂X, in agreement with other recent studies. The experiments also support models for superconductivity which invoke electron–electron interactions depending on the topological properties of the Fermi surface.

It is well known that bulk metallic glass formers have a tendency to show local icosahedral chemical ordering. We argue that the frustration between this short-range bond ordering and the long-range crystalline ordering controls the fragility and the glass-forming ability of these liquids. Our model suggests that a system having a stronger tendency to show local icosahedral ordering should be less fragile and a better glass former. This scenario also naturally explains the close relationship among the degree of local icosahedral ordering in liquid, glass formability, and quasicrystal formability.

Ion dynamics in the two-dimensional fluoride ion conductor KSn₂F₅ have been studied by ¹⁹F nuclear magnetic resonance (NMR) and impedance spectroscopy. The electrical relaxation behaviour of the investigated material is represented in the conductivity and electric modulus formalisms. The values of the dc conductivity and the hopping frequency of mobile ions and their respective activation energies are estimated from the analysis of the conductivity spectra using Almond–West formalism. The activation energies for conduction and ion hopping processes are almost identical, with values of 0.53 and 0.54 eV, respectively, suggesting that the concentration of charge carriers is independent of temperature. The Roling scaling approach to the conductivity spectra is applied in order to gain insight into the temperature dependence of the relaxation mechanism. The scaling law successfully collapses the conductivity spectra into a single curve, indicating a temperature-independent relaxation mechanism. Moreover, chemical exchange simulation of the ¹⁹F NMR spectra is introduced. The discrepancy between the values of the activation energies deduced from the electrical and NMR experiments are discussed according to Ngai's coupling model.

The temperature dependence of the resistivity and magnetoresistance (MR) of quasi-one-dimensional NbSe₃ was studied. The sharp increase of the positive MR in the lower charge-density wave (CDW) phase is consistent with previous reports, and the violation of Kohler's rule is obvious. We found that the MR data can be fitted very well with a modified two-band model, and the temperature dependence of the resulting parameters was discussed. The MR, the effect of magnetic field on the CDW gap, and the thermoelectric power of NbSe₃ can be coherently understood within this model.

A detailed study of the variation of the thermoelectric figure of merit of Si–Ge alloys with alloy composition, temperature and carrier density is the subject matter of this paper. Such a study is of particular interest at higher temperatures when minority carrier effects (MCEs) start to play a significant role in degrading the performance, thereby setting an upper limit of high temperature application. Alloys rich in silicon content are of greater interest and attention for the obvious reason that silicon, due to its large energy gap, is important in pushing the MCE to higher temperatures. The purpose of the present paper is to examine in depth all those processes that tend to degrade the material performance.

Results of the electrical resistivity, magnetoresistance and specific heat measurements are presented for Y_1−xGd_xCo₂ in the vicinity of the critical Gd concentration x_c ≈ 0.15 for the onset of a long-range magnetic order. Pronounced maxima of the coefficient γ of the T-linear specific heat, Debye temperature and residual resistivity as a function of concentration are found to exist at x = 0.1, i.e. just below x_c. These anomalies, together with the existence of a minimum on the temperature dependences of the electrical resistivity at 0 < x < 0.15, are ascribed to the substantial influence of localized spin density fluctuations induced by f–d exchange in the itinerant d-electron subsystem of Co. The maximal extra contribution to the residual resistivity Δρ_res^max in Y_1−xR_xCo₂ pseudobinaries is found to depend on the critical concentration as well as on the type of R ions, while the value of Δγ_max depends only on the critical concentration of R ions. The significant upturns on C_p/T versus T² dependences observed in Y_1−xGd_xCo₂ at 0 < x < 0.15 may be indicative of a non-Fermi-liquid behaviour of the d-electron subsystem caused by the partial substitution of Gd for Y.

We have investigated the effect of different cap layers on the photoluminescence (PL) of self-assembled InAs/GaAs quantum dots (QDs). Based upon different cap layers, the wavelength of InAs QDs can be tuned to the range from 1.3 to 1.5 µm. An InAlAs and InGaAs combination layer can enlarge the energy separation between the ground and first excited radiative transition. GaAs/InAs short period superlattices (SLs) make the emission wavelength shift to 1.53 µm. The PL intensity of InAs QDs capped with GaAs/InAs SLs shows an anomalous increase with increasing temperature. We attribute this to the transfer of carriers between different QDs.

A high efficiency of thermoelectric conversion can be achieved by using materials with a maximum figure of merit Z, i.e. materials with a high Seebeck coefficient S, high electrical conductivity σ and low thermal conductivity K. Investigations of transport phenomena in alloys of the (TlBiS₂)_1−x(2PbS)_x system have shown that in solid solutions of type (A³B⁵C ₂⁶) _1−x (2A⁴B⁶)_x cation substitution according to the scheme 2A⁴⁽⁺²⁾ → A³⁽⁺¹⁾ + B⁵⁽⁺³⁾ leads to a strong decrease in lattice thermal conductivity K_L. In an area (X = 0.50) of the K_L = f(C) curve the lattice part of the thermal conductivity of (TlBiS₂)_1−x(2PbS)_x alloys decreases to 0.26 W m⁻¹ K⁻¹, which is approaching the theoretical minimum. As a result, the dependence of the thermoelectric figure of merit on composition Z = f(C) of alloys of the (TlBiS₂)_1−x(2PbS)_x system passes through a maximum in the area X = 0.50, exceeding by ∼25% the respective value of Z for lead sulfide at room temperature.

New directions in glass structure and dynamics

The papers in this special section are contributed by the fourteen invited speakers of a symposium entitled 'New Directions in Glass Structure and Dynamics'. The symposium was held as part of the Glass and Optical Materials Division (GOMD) autumn meeting in Pittsburgh, Pennsylvania in October 2002. The division is a subset of the American Ceramic Society (AcerS). The symposium was organized by Dr Jacqueline A Johnson of Argonne National Laboratory (ANL) and Professor John Kieffer from the University of Michigan. They would like to thank all the speakers who participated so enthusiastically in the symposium and the sponsors who kindly contributed to its success: NSF (National Science Foundation) and AcerS.

To study the ion dynamics of several inorganic glasses, including single- and mixed-cation glasses, we have determined conductivity spectra over wide ranges in frequency. In the case of the single-cation glasses, these spectra extend from a few hertz up to the terahertz regime. The spectra show a transition from their dc values to a dispersive regime where the conductivity increases continuously with frequency, tending towards a linear frequency dependence at sufficiently low temperatures. At high frequencies the dynamic conductivity is governed by vibrational contributions, whereas ionic hopping sequences determine the low-frequency part of the spectra. In an intermediate-frequency regime, both hopping and vibrational contributions contribute to the dynamic conductivity. The shape of the high-frequency conductivity spectra is discussed for various glasses. The low-frequency spectra are discussed in the framework of the concept of mismatch and relaxation.

For the mixed-cation glasses where spectra have been taken by impedance spectroscopy, we report on a new kind of mixed-alkali effect. In contrast to conductivity spectra of single-cation glasses which follow the time–temperature superposition principle, featuring a temperature-invariant shape, the shapes of the conductivity spectra of the mixed-alkali glasses studied here are found to change with temperature. To explain this effect, we suggest differently activated mobilities of the two different ionic species.

We report on our work on determining the structure of oxide glasses using laser ionization time of flight mass spectrometry. This technique is especially suitable for investigation of the intermediate-range order in glasses, where it can measure the presence and relative abundance of mesounits composed of 7–20 or more atoms. After introducing the experimental aspects and caveats of this new spectroscopy, we discuss the application of the instrument to a variety of heavy-metal oxide glass families. We separate our networks into those that have superstructural networks and those that show highly fragmented ones, including our studies of lead and bismuth borates, lead silicates, lead vanadates, and bismuth gallates. We also analyse our spectra for evidence of mixing in lead borosilicates, and to clarify the sharing of cations in sodium-doped lead borosilicates.

Here we show that the longitudinal (L) and transverse (T) phonon densities of states can be extracted from linearly and circularly polarized Raman spectra. A test of the method for a-Si yields results consistent with the L and T densities of states inferred by analogy with the crystalline state. For a-SiO₂, for which the density of states is not known, both the L and T densities of states are obtained. The results show that the broad peak at around 70 cm⁻¹, commonly called the boson peak, is almost purely T in nature.

Five Nd–aluminosilicate glasses along the 2NdAlO₃–3SiO₂ join were synthesized using conventional drop-quench techniques. A sixth glass, with the end-member NdAlO₃ composition, required synthesis by containerless liquid-phase processing methods to avoid crystallization. Enthalpies of drop solution (ΔH_ds) and formation (ΔH_f) for the Nd–aluminosilicate glasses and the NdAlO₃-composition end-member glass were measured in molten 2PbO–B₂O₃ at 1078 K in a twin Calvet type calorimeter. Values for ΔH_ds for the Nd–aluminosilicate glasses increase with decreasing silica content from 130.7 ± 1.5 to 149.6 ± 0.6 kJ mol⁻¹. Similarly, values of ΔH_f increase with decreasing silica content from 41.0 ± 2.0 to 59.0 ± 1.6 kJ mol⁻¹. Values of ΔH_ds and ΔH_f for NdAlO₃-composition glass were measured as 99.3 ± 0.9 and 139.2 ± 2.1 kJ mol⁻¹, respectively. Using transposed temperature drop calorimetry, the enthalpy of vitrification for NdAlO₃-composition glass was measured as 69.5 ± 0.9 kJ mol⁻¹ relative to the stable crystalline neodymium aluminium perovskite (NdAlO₃) phase. Enthalpies of mixing were calculated based on amorphous end members; the strongly negative values support the absence of immiscibility in this system. Differential scanning calorimetry was used to determine glass transition (T_g) and crystallization (T_x) temperatures, as well as values for the configurational heat capacity (ΔC_P(T_g)) and the temperature range of the supercooled liquid interval (ΔT(SCL)). The NdAlO₃-composition glass showed no evidence of a glass transition prior to crystallization; only a single exotherm was observed, the onset of which occurred at 1045 K. For the Nd–aluminosilicates, values of T_g and ΔT(SCL) increase with increasing silica content, from 1128 to 1139 K and from ∼95 to ∼175 K, respectively. Values of (ΔC_P(T_g)) increase with decreasing silica content, from ∼27 to ∼75 J/g fw ^∗K, reflecting the increasing fragility and decreasing stability of the liquids as the end member composition, NdAlO₃, is approached.

The local atomic structures of Fe-and Al-based amorphous metallic glasses have been investigated by the pair density function (PDF) analysis of neutron diffraction data. In the Fe system, chemical substitution at the transition metal ion sites induces significant changes of the local structural environment. With the substitution of Nb for Zr and Mn for Fe, a pre-peak in the structure function is observed that is indicative of short-range chemical rearrangement. This is accompanied by a reduction of volume in real space and a contraction of interatomic pair correlations that could in turn be associated with stronger glass forming ability. In the Al–Ni–Nd system, using the isotope difference neutron PDF technique, the local environments for Ni and Nd were separately determined. The Ni ions tend to form strong interactions with Al with significant shortening of the bond length and weaker interactions with Nd. The local Ni–Al cluster is almost spherical with a coordination close to twelve nearest neighbours, while Ni–Ni chemical short-range ordering is quite strong as indicated by the change of amplitude of the pre-peak with the isotopic substitution. On the other hand, the local Nd environment is more disordered with two separate Nd–Al coordination environments and weaker clustering.

This paper introduces the concept of biomolecular characterization of inorganic surfaces. The choice of biomolecule is discussed followed by techniques that can be used to analyse the quantity of bound species, strength of binding, the nature of binding sites, conformational changes and the layer morphology. The prospects of modelling this data using a combination of molecular dynamics simulation and protein structural modelling and the correlation to measured data are outlined. The studies described in this paper are directed toward assessing the feasibility of biomolecular characterization, however, the data collected in the process are designed to also help elucidate our understanding of the interaction between biomolecular species and inorganic materials interfaces.

We have studied the low-speed fracture regime for different glassy materials with variable but controlled length scales of heterogeneity in a carefully controlled surrounding atmosphere. By using optical and atomic force microscopy techniques, we tracked, in real-time, the crack tip propagation at the nanometre scale over a wide velocity range (10⁻³–10⁻¹²m s⁻¹ and below). The influence of the heterogeneities on this velocity is presented and discussed. Our experiments reveal also—for the first time—that the crack progresses through nucleation, growth and coalescence of nanometric damage cavities within the amorphous phase. This may explain the large fluctuations observed in the crack tip velocities for the smallest values. This behaviour is very similar to that involved, at the micrometric scale, in ductile fracture. The only difference is very probably due to the related length scales (nanometric instead of micrometric). The consequences of such a nano-ductile fracture mode observed at a temperature far below the glass transition temperature, T_g, in glass is also discussed.

Ab initio quantum chemistry calculations and comparisons with experimental ¹⁷O solid-state nuclear magnetic resonance (NMR) investigations were used to determine the dependence of the ¹⁷O quadrupolar coupling constant and asymmetry parameter on the first-coordination-sphere structure around bridging oxygen. The quadrupolar asymmetry parameter was found to be dependent on the Si–O–Si angle, in agreement with previous studies, and independent of the Si–O distance. In contrast, the quadrupolar coupling constant was found to have a strong dependence on Si–O distance as well as Si–O–Si angle. Analytical expressions describing these dependences were proposed and used to develop an approach for relating measured ¹⁷O quadrupolar coupling constant and asymmetry parameter values for bridging oxygen to their Si–O–Si angle and average Si–O distance. Examples of this approach were given using ¹⁷O NMR results from the crystalline silica polymorphs, coesite, α-quartz, cristobalite, and ferrierite.

Ternary (Ge₂X₃)_x(As₂X₃)_1−x glasses with X = S or Se are of interest because they span a mean coordination number r in the 2.40 < r < 2.8 range that is characteristic of stressed–rigid glasses. We have examined X = S glasses in Raman scattering and T-modulated differential scanning calorimetry measurements over the 0 < x < 1.0 range. Glass transition temperatures, T_g(x), increase monotonically in the 0 < x < 0.8 range and decrease thereafter (0.8 < x < 1) to display a global maximum near x = 0.8. Raman scattering provides evidence of sharp modes of As₄S₄ and As₄S₃ monomers, with scattering strength of these modes showing a global maximum near x = 0.3 and 0.5 respectively. The results suggest that at low x (0 < x < 1/2), addition of Ge₂S₃ to the As₂S₃ base glass results in insertion of Ge(S_1/2)₄ tetrahedra in the As(S_1/2)₃-based backbone as compensating As-rich monomers segregate from the backbone to deliver the requisite S. At higher x (0.4 < x < 0.8), the Ge₂S₃ additive continues to enter the glass in a majority (As₂S₃)(GeS₂) backbone and several minority nanophases including an ethane-like Ge₂(S_1/2)₆ and a distorted rock-salt-like GeS. In the 0.8 < x < 1 range, the nanophases grow qualitatively at the expense of the backbone as T_g values decrease and the end-member composition (x = 1) is realized. Heterogeneity of glasses near x = 1/2 or mean coordination, r = 2.60 derives intrinsically from the presence of several minority nanophases and a majority backbone showing that stressed–rigid networks usually phase separate on a nanoscale.

Spallation neutron diffraction and high-energy x-ray diffraction methods have been used to study CaO:Al₂O₃  glasses at the 64:36 mol% eutectic and 50:50 mol% compositions. The samples were produced by the containerless cooling of liquid droplets heated by a laser beam and suspended in an aerodynamic levitator. The results show aluminium on average to be surrounded by 4.0(1) oxygen atoms at a distance of 1.76(1) Å in CaAl₂O₄, which increases to 4.8(1) at the eutectic composition. The two techniques have also been combined to reveal the local structure of the calcium atoms in the glass. In CaAl₂O₄  the calcium is found to be surrounded, on average, by 5.6(2) oxygen atoms at a distance of 2.38 Å. The Ca coordination decreases to the unusually low value of 3.9(2) oxygen atoms at a distance of 2.40 Å at the eutectic composition. No additional Ca–O correlations are observed up to 2.7 Å, but longer bonds cannot be ruled out. The higher-r correlations are shown to be similar in the two glasses, suggesting that both Al and Ca may act as network formers. The results are compared to previous studies on splat-quenched glasses with compositions near the eutectic.

We propose an extension to the technique of fluctuation electron microscopy that quantitatively measures a medium-range order correlation length in amorphous materials. In both simulated images from computer-generated paracrystalline amorphous silicon models and experimental images of amorphous silicon, we find that the spatial autocorrelation function of dark-field transmission electron micrographs of amorphous materials exhibits a simple exponential decay. The decay length measures a nanometre-scale structural correlation length in the sample, although it also depends on the microscope resolution. We also propose a new interpretation of the fluctuation microscopy image variance in terms of fluctuations in local atomic pair distribution functions.

We present a theoretical framework based on a higher order density correlation function, analogous to that used to investigate spin glasses, to describe dynamical heterogeneities in simulated glass-forming liquids. These higher order correlation functions are a four-point, time-dependent density correlation function g₄(r,t) and a corresponding 'structure factor' S₄(q,t) which measure the spatial correlations between the local liquid density at two points in space, each at two different times. g₄(r,t) and S₄(q,t) were extensively studied via molecular dynamics simulations of a binary Lennard-Jones mixture approaching the mode coupling temperature from above in Franz et al (1999 Phil. Mag.  B 79  1827), Donati et al (2002 J. Non-Cryst. Solids  307  215), Glotzer et al (2000 J. Chem. Phys.  112  509), Lacević et al (2002 Phys. Rev.  E 66  030101), Lacević et al (2003 J. Chem. Phys.  submitted) and Lacević (2003 Dissertation  The Johns Hopkins University). Here, we examine the contribution to g₄(r,t), S₄(q,t) and the corresponding dynamical correlation length, as well as the corresponding order parameter Q(t) and generalized susceptibility χ₄(t), from localized particles. We show that the dynamical correlation length ξ₄^SS(t) of localized particles has a maximum as a function of time t, and the value of the maximum of ξ₄^SS(t) increases steadily in the temperature range approaching the mode coupling temperature from above.

Modified lines were written inside Corning 7940 fused silica with 130 fs laser pulses from an amplified Ti–sapphire laser operating at 800 nm at a repetition rate of 1 kHz. The sample was scanned at 20 µm s⁻¹ with laser pulse energies ranging from 1 to 35 µJ, resulting in modified lines with diameters ranging from 8 to 40 µm. Confocal fluorescence and Raman microscopy was used to probe for spatial variations in defect concentration and glass structure across the modified lines. The fluorescence intensity decreased with increasing distance from the line centre, whereas the Raman intensity increased. No significant variations in the concentration of three- and four-membered ring structures were observed.

²⁹Si and ¹¹⁹Sn nuclear magnetic resonance (NMR) spectroscopy has been performed on binary xSnO(100 − x)SiO₂ and ternary xR₂O(50 − x)SnO50SiO₂ glasses (R = Li,Na,K,Rb for the ²⁹Si NMR and R = Na for ¹¹⁹Sn NMR). The spectra obtained have been fitted to obtain the spectral parameters. The variation of these parameters as a function of composition has been compared with structural information from previous density, neutron diffraction and Mössbauer spectroscopy studies. Tin is found to be present as Sn(II) in the binary silicate glasses but some Sn(IV) is present in the ternary glasses. The fraction of tin in the Sn(IV) oxidation state increases with the concentration of Na₂O. The ²⁹Si chemical shift reflects the average oxygen coordination number in the glass and the electronegativity of the neighbouring species. The ¹¹⁹Sn chemical shift shows different behaviour with SnO concentration for the binary and ternary systems and the latter give significantly narrower lines, indicating more symmetric ¹¹⁹Sn sites.