Table of contents

A perturbation detection technique is proposed to enhance the measurement resolution of electric force gradients in electrostatic force microscopy. The technique benefits from the perturbation in the probe tip position as it responds to interaction forces, thus eliminating the need to externally modulate the probe-circuit spacing and avoiding unwanted mechanical vibrations. Force gradients are detected by monitoring the resulting phase shifts in the probe vibration using a sensitive phase shift monitoring system. The constructed system enables the detection of electric force gradients at a probe-circuit spacing as high as 0.5 µm with a sensitivity of 20 µN m^-1 in 1 Hz measurement bandwidth. Scan images of electric force gradients above integrated lines structure are presented against typical force measurements and demonstrate a significant enhancement in measurement resolution.

We report on a preliminary investigation on the maskless implantation of Ge ions into Si for the production of Si_1-xGe_x microstructures. The technique employs a focused ion beam system using a liquid metal alloy ion source. Closely spaced simple structures down to about 1 µm in width, with well-defined boundaries, have been produced. On some of these structures, spreading resistance measurements were carried out.

This work involves an extensive study of the dependence on temperature (T) of the energy spread (ΔE_1/2) and potential energy deficit of the energy distribution of the Ge^{+ +} beam produced by a liquid metal alloy ion source. The liquid metal alloy is Au₇₇Ge₁₄Si₉. Differences between consecutive sets of measurements were found but the energy spread versus temperature measurements agree remarkably well with a recent model that predicts ΔE_1/2∝T^1/2. However, some questions involving the results either remain unresolved or are tentatively answered.

Electron cyclotron maser (ECM) oscillators are high power sources of microwave radiation and have applications in fusion plasma heating and diagnostics with potential for radar and telecommunications systems. The radiation comes from coherent bremsstrahlung emission by relativistic electrons gyrating in a magnetic field. It has been observed that the University of Strathclyde ECM oscillators fitted with high-current explosive electron emission cathodes exhibit a rapid modulation in the amplitude of the microwave output signal, whereas similar configurations using a thermionic emission cathode do not. This rapid communication describes preliminary experiments investigating this complex behaviour. In particular, it is of interest to establish whether the modulations were due to the cathode emission process or some dynamic cavity phenomena. We will present experimental results demonstrating a connection between the automodulation behaviour and the length of the interaction space in a 23 GHz ECM. The ECM had an ill-defined cavity consisting of a cylindrical copper waveguide of 1 m length. The magnetic field limited the length of the interaction space to <9 cm (the length of its central plateau), adjustable using waveguide cut-off reflectors. With an 8 cm long interaction space the output signal from the ECM demonstrated a full amplitude modulation with a period of ~4-7 ns, but reducing the length to 1.5 cm caused the modulation amplitude to reduce to ~20%, with a period of ~20 ns whilst simultaneously changing in nature from stochastic to cyclical.

Energetic ion bombardment of solid targets can lead to the production of atomic recoils and defects within the solid and the ejection or sputtering of atoms from the surface with the consequent erosion of the solid. The yield of sputtered atoms per ion depends on a number of ion and target parameters but, particularly, on the gradient of the surface with respect to the incident ion flux, the surface curvature and higher spatial derivatives of the height. As a result of these dependences of the local erosion rate, the morphology of a surface can be modified. But, in addition, surface atomic relaxation effects which may be mediated by the irradiation can occur and so the evolution of the surface may be complex. If the nature of these, often competing, processes is understood and can be suitably controlled by selection of experimental conditions, ion beam erosion can be employed to generate useful surface geometries.

This review briefly summarizes current understanding of the sputtering process and the origin of the above dependences and describes how, if only a surface gradient related mechanism dominates, the evolution of surface geometry can be accurately predicted. The higher-order and surface relaxation processes are then considered in both a deterministic approach and in a stochastic approach and these are shown to lead to fine spatial scale modifications to evolving surfaces. In both these areas, the physical models are supported by experimental observations. It is then shown how the lowest and higher orders and competing mechanisms can be selected in order to produce the desired surface morphologies in several application areas, including depth profiling of impurities in solids, ion milling and polishing, and the creation of repetitive surface structures.

We present experimental evidence of neck formation between an atomic force microscope (AFM) tip of Si₃N₄ and a flat Si(001) substrate. This neck formation occurs only for tips having sufficently large radius of curvature. When the AFM tip slides on the substrate the neck formed between them breaks which results in microstructural mound formation on the substrate and the apex of the tip. The contact between the Si₃N₄ tip and a Si(001) substrate could be well described by the Johnson-Kendall-Roberts (JKR) model which also predicts neck formation.

This paper describes the formation of cubic silicon carbide planar waveguides which use a region implanted with high doses of oxygen as the guiding layer. The material used for these experiments is cubic silicon carbide grown on-axis on {001} CZ silicon substrates. Oxygen implants were either conducted at 200 kV for singly charged ions (O⁺) or 400 kV for molecular O₂⁺ species. The results from the two types of implant are indistinguishable. To look at how the microstructure of the sample was influenced by implantation temperature a dose of 1.4×10¹⁸ cm^-2 ±10% was implanted into samples at temperatures ranging from ≈170 °C to 600 °C. It was found that implanting below 200 °C gives an amorphous layer, while at 600 °C the damage is comparable with the quality of the original single crystal starting material. The upper oxide/surface interface also sharpens with increasing temperature. The dose dependence of the microstructure was also examined. Doses ranging from 1×10¹⁷ to 1.8×10¹⁸ O cm^-2 were implanted while the wafer was maintained at 600 °C using a heated sample stage. At the lower doses ⩽8×10¹⁷ O cm^-2 carbon and silicon self-interstitials are produced by the dissociation of the silicon carbide host lattice, in a manner analogous to that observed for silicon implanted silicon carbide. At higher doses, ⩾1.4×10¹⁸ O cm^-2 chemical effects predominate and the growing SiO₂ layer causes the migration of the excess silicon and carbon interstitials towards the interfaces of the synthesized region. Increasing the implanted dose beyond 1.4×10¹⁸ O cm^-2 results in a significant increase in the level of damage in the surface region. The results also show that for doses of 1.4×10¹⁸ O cm^-2 and greater, a layer of SiO₂ is formed at the peak of the implanted distribution.

The lattice site of implanted tellurium in gallium nitride heteroepitaxial films grown on sapphire substrates has been studied. GaN samples were implanted with 400 keV ¹³⁰Te ions at 300 °C. The ion dose was 1.5×10¹⁵ cm^-2. The samples were analysed with Rutherford backscattering spectrometry channelling using 2.0 MeV ⁴He ions. The maximum Te concentration was 0.2 at.%. Based on measured channelling angular scans for Ga and Te along the ⟨0001⟩ and ⟨101⟩ axial directions, analysed by comparison with Monte Carlo simulations, it was concluded that in the as-implanted sample about 70% of Te atoms were on the Ga atom lattice sites, slightly displaced. The measured ⟨101⟩ scan showed no evidence of substitutional Te on N lattice site. The post-implantation annealing at 900 °C was found to decrease the substitutional fraction of Te while little recovery of the host lattice was observed. This indicates an interaction process between Te atoms and implantation-induced lattice defects. No migration of Te during annealing at 900 °C was observed. In addition, the effect of misfit dislocations in the host lattice on the channelling yield was visible in the ⟨101⟩ angular scan measured at a 47° tilt angle.}

\fnm{3}{Present address: Vaisala Oyj, FIN-00421 Helsinki, Finland.

Silicon nanocrystals have been prepared in thin-film form by a dc magnetron sputtering technique. The optical transmission, optical absorption, Raman effect and photoluminescence of the nanocrystals have been studied. The observed blueshift of the band gap due to a reduction in the crystallite size is correlated with the transmission electron microscopy observation. Raman spectra indicated the presence of a TO mode which showed redshift with the decrease in the size of the nanocrystallites.

The interface between a thin Au film deposited on an atomically clean GaN surface has been investigated by means of synchrotron radiation scanning photoelectron microscopy. The first stages of the interface formation at room temperature and the evolution of the interface after stepwise annealing up to 870 °C have been studied. Our results reveal the temperature dependence of the Schottky barrier heights, caused by structural rearrangements below 500 °C, and by the formation of an Au gallium alloy above 500 °C. The effect of the reaction-induced heterogeneity of the interface on the local Schottky barriers is examined.

Bilayer films of poly(ethyleneimine) and poly(ethylene-co-maleic acid) have been prepared using the layer-by-layer assembly technique. These films have been characterized using surface plasmon resonance (SPR). Exposure to aqueous solutions of metal acetate (metal = copper, nickel) resulted in a shift in the position of the SPR curve. This shift could be monitored by observing the intensity of the reflected laser beam at fixed angle and concentrations of copper acetate down to at least 10^-6 M could be detected.

Porous silicon was fabricated by using a Nd:YAG laser in a laser-induced etching process. Scanning electron microscopy was used to monitor changes in surface morphology produced during the etching process. Porous silicon samples were subjected to spectroscopic investigations using an argon-ion laser. The first-order Raman line asymmetry was found to decrease with decrease of the incident argon-ion laser excitation photon energy, while the peak position remained unchanged for a given etching time and power density. The photoluminescence spectra exhibits a red shift in peak position and a dramatic decrease in intensity as the incident photon energy was decreased. Both Raman and photoluminescence data were explained using appropriate quantum confinement models involving two-dimensional confinement and Gaussian size distributions of nanocrystallites constituting porous silicon samples. There is reasonable agreement between the results obtained from Raman and photoluminescence spectroscopic investigations of the PS samples.

Crystalline carbon nitride films have been synthesized on Si(100) substrates by a microwave plasma chemical vapour deposition technique, using a gas mixture containing nitrogen and methane at various ratios as precursors. Scanning electron microscopy shows that the films consisted of hexagonal crystalline rods, which are about 1-2 µm long and about 0.4 µm wide. X-ray diffraction and transition electron microscopy indicate that the films are mainly composed of α- and β-C₃N₄, and these results match more closely with the α-C₃N₄ than with the β-C₃N₄ phase. Fourier transform infrared absorption spectra of carbon nitride films support the existence of α- and β-C₃N₄.}

\fnm{3}{Author to whom correspondence should be addressed.

A calculation was performed of the energy volume density f_m in a magnetic field ^m that is caused by planar small-angle grain boundaries (GBs) in a polycrystalline ferromagnetic, such as silicon steel, with grains magnetized to saturation (M = M^S). Equations for the field components, H^m_ξi and energy, f_m, were obtained and solved, taking into account the µ*-effect near GBs. It is shown that deviations of the vectors such as _S from the easy directions, ⟨100⟩, under the action of fields ^m at GBs leads to an essential (in ten times) decrease of the (H^m_ξi)_max component and the ⟨(f_m)_max⟩ values. The influences of the orientation of the GB surfaces and µ* on the f_m distribution are discussed. The results of the computation are compared with the corresponding experimental data for samples of grain-oriented silicon steel sheets.

Magnetic and structural properties of RCo_12-xTi_x with R = Y and Sm and YFe_12-xTi_x compounds have been investigated by using x-ray diffraction and magnetic measurements. It was found that all compounds investigated are almost single phase with the ThMn₁₂-type structure. X-ray diffraction patterns of magnetically aligned powder samples indicate that the easy magnetization directions (EMD) at room temperature are in the basal plane for SmCo_12-xTi_x compounds while for YCo_12-xTi_x and YFe_12-xTi_x compounds the EMDs are along the c-axis. The lattice parameters increase monotonically with increasing Ti content for all the compounds investigated. The Curie temperature T_C decreases monotonically with increasing Ti content for both YCo_12-xTi_x and SmCo_12-xTi_x, while T_C of YFe_12-xTi_x is almost independent of Ti content. The saturation magnetization M_s decreases with increasing Ti content for all the three systems. The anisotropy field B_a of YCo_12-xTi_x and YFe_12-xTi_x compounds decreases monotonically with increasing Ti content, while the Ti content dependence of the anisotropy fields B_a for SmCo_12-xTi_x compounds at 1.5 K exhibits a maximum at about x = 1.7, which can be understood in terms of a two-sublattice model of anisotropy.

This paper presents an optimization procedure for the design of circular magnets of multipolar symmetry in the presence of an iron screen. The procedure is based on the combined use of field calculations and optimization techniques and can be applied when the field is generated either by a distribution of currents or of magnetic materials. The numerical optimization technique uses the positions and/or orientations of the field generating elements as parameters (currents or permanent magnets). The technique has proved to be very effective in obtaining a good uniformity of field in a large fraction of the magnet volume. It should reduce the subsequent shimming operations. The method is almost completely automatic and simpler than other standard methods normally used to design new magnet geometries which are based on field calculation programs and `manual' optimization procedures.}

\fnm{4}{Author to whom all correspondence should be addressed.

A two-dimensional time-dependent model of non-self-sustained dc and RF discharges is used for the analysis of recent experiments in optically-pumped non-equilibrium plasmas. The analysis shows that non-self-sustained dc and RF discharges can be successfully used for measurements of the electron production rate, electron recombination rate, and electron density in these plasmas. The inferred rate of electron production per unit volume by the associative ionization mechanism is S₀ = 1.0×10¹⁵ cm^-3 s^-1 and S₀ = 2.2×10¹⁴ cm^-3 s^-1 in CO/Ar = 2/100 and in CO/N₂ = 2/100 plasmas, respectively. The inferred electron-ion recombination rate coefficients are β>6.0×10^-6 cm³ s^-1 and β = (7.5±1.5)×10^-6 cm³ s^-1 in CO/Ar = 2/100 mixtures at P = 100 Torr and 20 Torr, respectively. In CO/Ar = 2/100 mixtures with a 0.05-0.1 Torr admixture of O₂ at P = 100 Torr and 20 Torr, the inferred recombination rate coefficients are β = (1.5±0.3)×10^-8 cm³ s^-1 and β = (5.1±2.9)×10^-8 cm³ s^-1, respectively. Finally, the inferred electron density in optically-pumped CO/Ar/O₂ plasmas at the laser beam axis is n_e = (1.7±0.2)×10¹¹ cm^-3 at P = 100 Torr and n_e = (6.2±0.8)×10¹⁰ cm^-3 at P = 20 Torr. These results demonstrate that the electron density in optically-pumped CO/Ar plasmas can be controlled and significantly increased by adding small amounts (up to ~0.1%) of species such as O₂ and NO, which result in the reduction of the electron-ion recombination rate.

A perturbation model has been developed to describe the evolution of an expanding plasma sheath around a cathode after a high-voltage negative pulse is applied to the cathode, simulating the conditions in devices such as those used for plasma source ion implantation. The set of governing equations consists of two coupled collisionless fluid equations for ions, and Poisson's equation and Boltzmann's assumption for electrons. The time-dependent, self-consistent expressions for the potential, ion density and ion flux are obtained and compared successfully with experimental and simulation results.

We present the results of a study on electrode floating potential formation in a hot cathode discharge plasma. The electron component of the plasma is composed of two populations. The high-temperature component develops from primary electrons and the cool component from secondary electrons born by ionization of cold neutral gas. A static, kinetic plasma-sheath model is used to calculate the pre-sheath potential and the floating potential of the electrode. For hot primary electrons a truncated Maxwellian distribution is assumed. The plasma system is also modelled numerically with a dynamic, electrostatic particle simulation. The plasma source injects temporally equal fluxes of ions and electrons with half-Maxwellian velocities. Again, the hot electron distribution is truncated in the high-velocity tail. The plasma parameters, such as ion temperature and mass, electron temperatures, discharge voltages, etc, correspond to experimental values. The experimental measurements of the electrode floating potential are performed in weakly magnetized plasmas produced with hot cathode discharges in argon gas. Theoretical, simulation and experimental results are compared and they agree very well.

We report lasing of the CVI Balmer-α line at 18.22 nm using the new technique of an induced MHD instability in an ablative capillary discharge. A large spike of this line during the second half-cycle of the discharge is observed. The spike is identified as amplified spontaneous emission (ASE), and enhancements are derived for capillaries of lengths between 15 and 50 mm. We explain the ASE as a result of charge exchange processes which occur between ions produced in the neck of the MHD instability and low temperature ions present in the same plasma column. To emphasize this instability, we implemented a new method by using a straight capillary with a waved structure imprinted inside. The results are encouraging and an exponential gain length is achieved with GL =4.5 for a 40 mm length capillary. Other evidence to support the observation is realised when a plane multilayer mirror with reflectivity <30% is introduced resulting in amplification.

Radiative efficiency of emission of resonance radiation from a dc glow discharge plasma is derived experimentally from spatial density profiles of atoms in resonant states. Density profiles are determined by tunable diode laser absorption spectroscopy. Resonance radiation transport is considered both by Lawler and Curry's analytical formula (Lawler J E and Curry J J 1998 J. Phys. D: Appl. Phys.31 3225) and by Monte Carlo simulations. Radial dependences of the escape factor are studied. Power loss by resonance radiation is found to be about 50% of electrical power input. The impact of resonance radiation on similarity laws in discharge physics is discussed.

The present work reports results obtained using combined laser shadowing and cross correlation techniques to investigate the metal transfer process. Experimental results of the droplet size and velocity of 0.9 mm and 1.2 mm wires from the globular region to the spray transfer region are presented. The results indicate that the droplet size is not continuous in the narrow region of mode transition, and also preferred bands of droplet size exist. Droplet velocity measurements show good agreement with a simple model based only on the electromagnetic pinch force.

Pulsed high-voltage glow discharge, a process that falls between conventional plasma ion implantation and plasma nitriding, is a potentially effective surface treatment technique. The electrical characteristics and initiation dynamics of the process are experimentally investigated in this work. The discharge behaviour is found to depend on the applied voltage, gas pressure, pulse duration, and pulsing frequency. The voltage-current characteristics basically obey the collisionless Child-Langmuir relationship, and the discharge current varies substantially with the applied voltage and gas pressure in the high-voltage, high-pressure domain. The seed plasma before each glow discharge pulse affects the initiation mechanism regardless of whether it is sustained by an external plasma source or it is simply the residual plasma of the afterglow after the voltage pulse has been turned off. Our results also disclose that the presence of an initial plasma before the glow discharge is ignited has a large impact on the initiation time of the glow discharge but has only a slight effect on the steady-state current.

A type-I walk-off compensation (WOC) arrangement has been made for the second harmonic generation (SHG) and the fourth harmonic generation (FOHG) of commercially available Q-switched Nd:YAG laser radiation, for the first time, using two identically cut Li₂B₄O₇ crystals. The enhancement of the conversion efficiencies has been realized in the WOC arrangement by a factor of 2.71 for the SHG and 2.65 for the FOHG relative to the double-crystal non-walk-off arrangement. The absolute values of the conversion efficiencies obtained for WOC SHG and FOHG are 3.52% and 3.56% respectively, applying input power densities for the fundamental and second harmonic of the Nd:YAG laser beam at only 28.3 MW cm^-2 and 11.7 MW cm^-2, respectively.

A high-energy-density laser beam-material interaction process has been simulated considering a self-evolving liquid-vapour interface profile. A mathematical scheme called the level-set technique has been adopted to capture the transient liquid-vapour interface. Inherent to this technique are: the ability to simulate merger and splitting of the liquid-vapour interface and the simultaneous updating of the surface normal and the curvature. Unsteady heat transfer and fluid flow phenomena are modelled, considering the thermo-capillary effect and the recoil pressure. A kinetic Knudsen layer has been considered to simulate evaporation phenomena at the liquid-vapour interface. Also, the homogeneous boiling phenomenon near the critical point is implemented. Energy distribution inside the vapour cavity is computed considering multiple reflection phenomena. The effect of laser power on the material removal mode, liquid layer thickness, surface temperature and the evaporation speed are presented and discussed.

A new method is suggested for measuring the refractive index profile of homogeneous highly-oriented fibres. This method depends on the principle of the variable wavelength interferometric technique. A numerical fitting is used to determine the fringe shift profile at a certain wavelength. The refraction of the light beam by the fibre is taken into consideration in the mathematical treatment. Poly(ethylene terephthalate) and poly(ethylene 2,6-naphthalene-dicarboxylate) fibres are used as examples of these highly-oriented fibres. The diffraction of a He-Ne laser beam is used to determine the diameter and the cross sectional shape of these fibres. Also, a two-beam interference microscope (Pluta microscope) is used for measuring the refractive index profile of these fibres, at the same wavelength, for light vibrating perpendicular to the fibre axes to confirm the results of the new method. Microinterferograms are given for illustration.

The relaxation times τ₁ and τ₂ for rotations of the flexible parts and the whole molecules of the isomers of anisidine and toluidine (j) in benzene (i) under a 9.945 GHz electric field are predicted from the slope and intercept of a linear equation of (χ_0ij-χ'_ij)/χ'_ij against χ''_ij/χ'_ij for different weight fractions w_j's of solute at 35 °C. χ'_ij and χ''_ij are the real and imaginary parts of the high-frequency complex orientational susceptibility χ*_ij and χ_0ij is the low-frequency susceptibility which is real. The τ's; dimensionless parameters b's involved with τ's and linear coefficients β's of χ'_ij-w_j curves are used to obtain dipole moments µ₁ and µ₂ of the flexible parts and the whole molecules. The theoretical weighted contributions c₁ and c₂ due to τ₁ and τ₂ towards dielectric dispersions from Fröhlich's equations are worked out to compare with experimental ones by graphical variations of χ'_ij/χ_0ij and χ''_ij/χ_0ij at w_j→0. The estimated τ's from the ratio of the individual slopes of χ''_ij and χ'_ij with w_j's at w_j→0 are compared with those of Murthy et al (1989 Indian J. Phys. B 63 491) and reported τ's. The symmetric and characteristic relaxation times τ_s and τ_cs by symmetric and asymmetric distribution parameters γ and δ are computed to suggest the symmetric relaxation behaviour of the molecules. The measured µ₁ and µ₂ from double relaxation phenomena and the reported µ's signify the fact that a part of the molecule is rotating under a gigahertz electric field. The slight disagreement between the theoretical dipole moment µ_theo from available bond angles and bond moments of substituent polar groups attached to the parent molecules and measured µ₁ demands the inductive and mesomeric effects suffered by the substituent polar groups, in addition to structural and associational aspects of such molecules.

Polycrystalline samples of modified lead germanate, Pb_4.9A_0.1Ge₂TiO₁₁ (A = Ca, Sr, Ba) were synthesized by a solid-state reaction technique. Studies of x-ray powder diffraction patterns and scanning electron microscopy images of these compounds at room temperature suggest that the compounds have been formed almost in a single phase with a trigonal crystal system and a uniform grain distribution throughout the surface of the samples. Detailed studies of dielectric properties of the compounds as a function of frequency (2×10² to 10⁶ Hz) at room temperature, and for temperatures from room temperature to 200 °C at 10, 100 and 1000 kHz show broadening (diffuseness) of the dielectric peaks and a change of the transition temperature with the ionic size of the dopants. The temperature dependence of the electrical conductivity, polarization and pyroelectricity for all these compounds provide some interesting results for device fabrication.

A procedure to optimize the collection of Brillouin spectra and to allow the simultaneous observation of phonons in two pairs of scattering geometries is proposed. The first geometry is used to observe the bulk phonons in right-angle- and back-scattering. This procedure reduces measurement time and requires the use of only one type of sample. The second geometry is used for simultaneous observation of bulk and surface phonons in transparent materials and can provide some valuable information on the relation between the bulk and surface elastic properties.

The alumina matrix of an electrolytically deposited solar absorber coating was etched away leaving nickel nanorods standing on an aluminium support. The oxidation kinetics of the nickel nanorods, with an approximate diameter of 30 nm and a height of approximately 300 nm, was determined by infrared (IR) spectroscopy. The IR absorption by longitudinal optical phonons in NiO was measured before and after exposure to pure oxygen at 260, 280, 300, 320 and 350 °C for different times. The absorptance peak height was determined and used as a measure of the degree of oxidation. The five different isotherms showed power-law behaviour with an exponent varying between 0.52 and 0.69. A fit to the homogeneous linear diffusion equation, derived for spherical geometry, gave parabolic rate constants, which are in agreement with data for larger nickel particles and bulk nickel. The deviations in the exponent from that of the parabolic law are discussed in terms of particle geometry and particle agglomeration. The apparent activation energy was determined, by the use of a master plot technique, to be 1.73 eV.

In this paper, mode mismatched thermal lens spectrometry is applied to determine the thermal and optical properties of polycarbonate as a function of temperature. The focus of the experiments was in the temperature range between 120 °C and 170 °C, where the sample glass transitions occurs. In order to validate and evaluate the sensitivity of the proposed method we have also carried out conventional differential scanning calorimetry measurements. The results showed that the temperature dependence of the signal amplitude of the thermal lens provided a better definition in locating the glass transition compared to the differential scanning calorimetry data. In addition, a sharp decrease in the thermal diffusivity values at the beginning of the sample glass transition was observed. According to the Debye model, this change was attributed to the variation of the sample elastic constant throughout the glass transition. Finally, we propose this procedure, called the differential thermal lens temperature scanning approach, as a new method for the investigation of phase transitions in transparent polymers.}

\fnm{3}{Author to whom correspondence should be addressed.

We consider the evaporation of material from a heated substrate into a vacuum, and the problem of determining the density, flow velocity and temperature of the vapour that streams off the surface. Treatments using the Boltzmann equation suggest that the vapour flows at the speed of sound, and with a temperature and density that depend on the substrate temperature. A simpler approach is to parametrize the velocity distribution function of vapour at the surface and then to use the conservation of mass, momentum and energy fluxes to characterize the flow. However, the mean velocity of the vapour is undetermined in this approach. We find, however, that by calculating the flux of Boltzmann's H function, we can exclude high mean velocities, since they correspond to the unphysical destruction of entropy in the evaporation process. Furthermore, it appears that the generation of vapour travelling at approximately the speed of sound corresponds to the maximum rate of entropy production. This lends support to this principle as a useful method for characterizing systems far from equilibrium.

Single-crystalline Bi nanowire arrays have been assembled into the nanochannels of anodic alumina membranes by electrodeposition. X-ray diffraction and transmission electron microscopy investigations revealed that the nanowires with diameter of either 50 nm or 20 nm are essentially single crystalline and highly oriented. The magnetotransport properties of the Bi nanowire arrays were measured, and positive magnetoresistance as high as 45% at 100 K was observed.}

\fnm{3}{Author to whom correspondence should be addressed.

In this paper, the power density (defined as the ratio of the power output to the maximum specific volume in the cycle) is taken as the objective for performance optimizations of an endoreversible closed Brayton cycle coupled to constant-temperature heat reservoirs in the viewpoint of finite-time thermodynamics (FTT) or entropy generation minimization (EGM). The optimum heat conductance distribution corresponding to the optimum power density of the hot- and cold-side heat exchangers for the fixed heat exchanger inventory is analysed using numerical examples. The influence of some design parameters on the optimum heat conductance distribution and the maximum power density and the optimum pressure ratio corresponding to the maximum power density are provided. The power plant design with optimization leads to higher efficiency and smaller size.}

\fnm{3}{Author to whom correspondence should be addressed.

A mid-infrared light-emitting diode (LED) operating at 3.6 µm with a pulsed output power as high as 2 mW at room temperature is reported. The device is based on a double heterostructure of InAs_0.42Sb_0.18P_0.40/ Ga_xIn_1-xAs_1-ySb_y/InAs_0.42Sb_0.18P_0.40 grown by liquid phase epitaxy. The active region composition was adjusted using different Ga content (x = 0.05-0.17) over which range the band gap is approximately constant. The electrical and optical operating characteristics are reported. The emission spectra measured with and without 100% formaldehyde have also been measured using a short-path optical cell to demonstrate the potential of these LEDs for use in solid state formaldehyde gas sensors.

To demonstrate the potential of gravity gradiometry in the detection of underground structures or voids, an experiment was carried out at an underground missile launch facility at Vandenberg Air Force Base, California. Gravity data were collected using a LaCoste-Romberg Model G gravimeter and gravity gradient measurements were made using a vertical spin-axis Lockheed Martin developed arms control verification gravity gradiometer. After performing all necessary corrections, the data were compared to modelled signatures of the launch facility. Peak gravity and gravity gradient anomalies of approximately -75 µGal (1 µGal = 10^-8 m s^-1) and 30 E (1 E = 10^-9 s^-2) respectively were measured over the centreline of the largest capsule. The presence of the buried launch facility was apparent in both the observed gravity and gravity gradient data. The gravity gradients were unambiguous in resolving the underground facility's signature and thus revealing its location. In contrast, identifying the launch complex solely from the gravity anomaly data would have proven difficult due to a strong regional trend across the target area.}

\fnm{4}{National Imagery and Mapping Agency, Patrick AFB, FL 32925-3438, USA.}

\fnm{5}{MITRE Corporation, Bedford, Massachusetts 01730-1420, USA.

Equation (9) is corrected to

Figure 19(b) is also corrected.

Figure 1. Geometry. The angles θ, ϕ, and ξ are defined, specifying the transformed magnetizations. The angles are functions of position, y.