Table of contents

The resistivity of copper films grown by varying the pressure, and hence the growth rate, in metalorganic chemical vapour deposition has been studied in the temperature range 4.2 K - 300 K. The films exhibit a fairly high (300 K) of 8 - 20 cm. Analysis of the temperature variation of shows that the high values are not just caused by elastic scattering from the impurities but the temperature dependence of is also very high, resulting in a large deviation from Matthiessen's rule (DMR) in these films. This strong dependence on temperature and DMR has been explained in a semi-quantitative manner as arising from grain boundary (GB) and surface scattering (SS). This is corroborated by STM studies on the films which show that films having a smooth surface and well connected grains have a lower as opposed to films with poor connectivity.

The field of nanoscale physics is now widely regarded both as a frontier of science and as a generic foundation for a wide range of new technologies in sectors as diverse as optics, electronics, chemicals, magnetics and sensors. The development of a body of good science in this field depends on the careful preparation and characterization of well-defined nanometre-scale structures. The development of genuine and viable nanotechnologies requires that scientists are alert to the connections between their model systems and the practical requirements of technological applications. This is the domain of applied physics, and it is fitting that in this issue Journal of Physics D should play host to a set of papers which reviews some of the frontiers of this exciting field.

The subject of the first paper, by Perez et al from Lyon, concerns a new method for the controlled fabrication of novel nanostructures which are rich in potential - cluster assembled materials. This `bottom-up' approach defines a series of nanoscale building blocks, atomic clusters of a specific size, and assembles a thin film by the deposition of the clusters onto a surface. The paper discusses both the nature of the growth process and some of the intriguing physical properties of the films so formed - structural, optical and magnetic. Structural characterization of these and other nanosystems is one of the foundation issues in the development of a science of nanostructures, and the second paper in the set, by Castle and Zhdan from Surrey, is concerned with two of the principal scientific tools which are employed, i.e. the scanning electron microscope (SEM) and the (whole array of) scanning probe microscopes (in this case, the scanning force microscope (SFM)). Interest in truly nanometre-scale structures is pushing the veritable old SEM to its limits, though it has the advantage over its younger cousins, illustrated by the SFM, to whom atomic resolution is a mere trifle, of a much wider scan range as well as chemical analysis. The paper presents a careful analysis of what these complementary techniques can offer, and includes parallel images of the same regions of the same samples - precisely the kind of investigation which is required to develop a reliable metrology on the nanometre scale. But nanostructures need not be static; indeed, there is already a whole industry, heterogeneous catalysis, which depends on the evolution of the structural and electronic properties of supported nanoscale particles. The third paper, by Leibsle et al from Liverpool, shows how the scanning tunnelling microscope (STM) provides some searching insights into the spatial evolution of nanostructures during a chemical reaction in model surface systems. The paper also explores the fascinating domain of spontaneous (`self-organized') nanostructures (in this case, at oxide surfaces), a topic which may come to assume considerable importance for the large-scale production of nanostructured systems.

It is to be hoped that this short series of articles will allow readers to take stock of a fast moving, if somewhat `hyped', field. Genuine advances are taking place as a result of imaginative and careful experiments, and a coherent level of understanding of the distinctive properties of nanometre-scale structures is taking shape.

R E Palmer Nanoscale Physics Research Laboratory School of Physics and Space Research University of Birmingham, UK

31 December 1996

The low-energy cluster beam deposition technique (LECBD) is applied to produce cluster assembled films with hitherto unknown nanostructured morphologies and properties. Neutral clusters having the very low energy gained in the supersonic expansion at the exit of the inert gas condensation-type source are deposited without fragmentation upon impact on the substrate. Depending on the deposition conditions (nature, size and flux of incident clusters, nature and temperature of the substrate, vacuum conditions), granular nanostructures resulting from the diffusion and coalescence of supported clusters are obtained with materials of any type (covalent or metallic). A critical size for coalescence limits the supported grain size and, finally, highly porous thick films growing by random stacking of nanoparticles are obtained. A recent model developed by combining several dynamical processes simultaneously occurring on the substrate (deposition - diffusion - aggregation, DDA) is used to simulate the cluster assembled film morphology in good agreement with the experimental observations. Examples of novel materials obtained by LECBD are presented to illustrate the interesting potentialities of the technique. In the case of covalent materials such as carbon and silicon, 'amorphon'-type disordered structures, different from the conventional amorphous structures (a-C and a-Si), are obtained with some unique properties. With transition metal (Fe, Co and Ni) cluster assembled films, a specific magnetic behaviour, resulting from the competition between the intrinsic properties of the grains (magnetocrystalline anisotropy) and the interactions between grains, is observed. Also, films of clusters embedded in various co-deposited matrices are produced in order to control the interactions between grains via the matrix materials (insulating, conducting ...). Interesting optical properties (from metallic clusters in ) or giant magnetoresistance effects (from Co clusters in silver) are reported for such systems, emphasizing the future role of LECBD in various fields of applications such as optical and optoelectronic nanostructures, magnetic and magneto-optic nanostructures and quantum devices.

An overview is given of the application of scanning electron microscopy (SEM) and scanning force microscopy (SFM) to the characterization of surface topography. The potential `blindness' of SEM to surface topographical variations, and the contribution of tip - sample convolution to the measurement process, which can result in the misinterpretation of SFM data, are reviewed. The experimental characterization of different materials by SEM, employing high- and low-energetic primary electrons, and by SFM (including parallel studies of the same regions) allowed direct comparison of SEM and SFM data and permitted the direct estimation of the surface sensitivity of SEM relative to SFM. Further progress, including combining SFM and SEM in one instrument, is discussed.

A review of surface reactions studied with scanning tunnelling microscopy (STM) is presented which shows how STM can give unprecedented insight into the mechanisms at an atomic level. This is illustrated using a number of examples from work at the Interdisciplinary Research Centre in Surface Science. In addition, a number of common trends which occur in different systems are highlighted. Finally, results on the characterization of oxide surfaces are presented because these materials act as supports for many of the metals in industrial catalyst systems.

Be diffusion mechanisms during post-growth annealing have been investigated in p-type InGaAs epitaxial layers. These Be-doped InGaAs layers were grown between two undoped InGaAs layers. Annealing cycles with time durations of 30 s to 3 min and temperatures in the range of 500 - 800 for doping concentrations of , and were performed.

It was found that there is no significant Be diffusion during post-growth rapid thermal annealing (RTA) for all doping levels if annealing is performed at 500 and 600 for a time under 1 min. The same observation was made for a Be concentration of with annealing at 700 for 30 s.

For a doping level of , significant Be diffusion occurs for temperatures and durations greater than 700 and 1 min respectively.

The shapes of the obtained curves represent a `double profile' giving rise to a concave region called a `kink'. It was deduced from the experiments that the Be effective diffusion coefficient is approximately constant for the high region of the concentration profiles and proportional to the square root of concentration for the low part of the experimental curves.

Assuming the latter coefficient dependence, an effective diffusivity was substituted into Fick's second law and the resulting differential equation was solved numerically using an explicit finite-difference method. Good agreement was obtained between our depth profiles and the corresponding simulated distributions.

a-C:H films were prepared by the DC - RF PECVD technique. The variation of the structure and hydrogen content of the films versus the applied DC bias voltage were studied by FTIR and Raman scattering spectroscopy. The results showed that the structure of the films tended to be more graphite-like when at either a higher or a lower applied DC bias than at -800 V. The hydrogen content (total amount of both bonded and unbonded) of the films was found to have a maximum value at about -800 V. The results are discussed and compared with those reported in the published literature.

The optical constants of RF planar magnetron sputtered films are reported for the 0.3 - 2.5 wavelength range. The data are interpreted in terms of Maxwell-Garnett and Bruggeman effective medium theories. The effect of atomic Ni dispersion in the matrix on the effective medium formalism is discussed.

Thermo-field emission current densities calculated by the commonly used Richardson - Dushman equation corrected for the Schottky effect are compared with those calculated using the more accurate treatment of Murphy and Good for a wide range of temperatures (1000 - 5000 K), electric field strengths () and work functions (2 - 5 eV) encountered in numerical modelling of the thermal arc-cathode interactions. Results show that the emission current densities predicted by the Richardson - Dushman equation are always lower than those predicted by the more accurate treatment. The disagreement between the two predictions is in the range 20 - 30% for the conditions encountered in numerical modelling of the attachment of a thermal arc to a hot tungsten cathode whereas it is much larger when a cold copper cathode is considered (>175%). It is suggested that the more accurate treatment of Murphy and Good should be used in order to increase the accuracy of prediction of the numerical models.

In this paper, the antifriction behaviours of (3:1) molecules and their crystal powder were evaluated by different methods. It was found that the crystal powder possessed hexagonal close packed (hcp) crystal structure with a = 10.1 Å and c = 16.55 Å, and a transformation of crystal structure from hcp to face centred cubic (fcc) occurred easily during friction (burnishing). It was confirmed that two kinds of process, breakage of powder coagulated by nanoscale single crystals and rearrangement of the molecules along the friction direction, had occurred under the friction force. The extreme pressure (EP) performance of as an additive in paraffin liquid was investigated on an SRV oscillating wear machine. It was found that the extreme pressure load (EP value) of paraffin liquid was increased by dispersion of powder, accompanied by a slight improvement in the antifriction behaviour. it was confirmed that the improvement in EP value and antifriction behaviour of oil was dependent on the crystal structure of powder, but independent of the spherical molecular structure of or . The burnishing experimental results also proved that the antifriction behaviour was determined by the crystal structure and had no relation to the molecular structure. It was also found that fullerenes possessed some physical properties similar to those of graphite. Since the formation of compact fullerenes with high shear strength during friction can be effectively prevented by some other lubricants, it is suggested that fullerenes should be mixed with other lubricants for tribological application.

We propose a new beam forming approach based on `asymmetric apodization', which was originally used in optics for single openings by Cheng and Siu. Asymmetric apodization is applied to linear aperture arrays, namely a series of discrete small wavelet elements. The method of `anti-phased peripheral belts' is similar to that used in optics. The criterion for the optimum discrete aperture functions or linear arrays of wavelet elements is proposed for determining essential parameters. Calculation results are presented for several arrays with different numbers of elements. This technique is suitable for improving the performance of antenna arrays or optical gratings.

UV radiation during gas tungsten arc (GTA) welding in argon/helium shielding gas mixtures versus gas composition and arc length has been investigated experimentally and theoretically. UV radiation was found to increase with an increase of argon concentration in the mixture. A physical model describing this phenomenon has been developed. The model predicts an increase in the total amount of ions in the mixture with an increase of argon concentration despite a decrease in plasma temperature. The result of calculations using the model is in satisfactory agreement with the experimental data.

A physical model of keyhole support and propagation during high-translation-speed laser welding is described. A numerical code for the simulation of the front keyhole wall behaviour is developed on the basis of a `hydrodynamic' physical model assuming that: (i) only the front part of the keyhole wall is exposed to the high-intensity laser beam; and (ii) recoil pressure exceeds surface tension and propagation of the keyhole wall inside the sample is due to melt expulsion similar to that in laser drilling. The front keyhole wall profile, distribution of absorbed laser intensity and phase velocity of the solid/liquid (liquid/vapour) boundary are calculated for various processing parameters. The calculations show that, depending on the processing conditions, the absolute value of the keyhole wall velocity component parallel to the translation velocity vector can be higher than, smaller than or equal to the beam translation speed. When the component of the keyhole velocity vector parallel to the sample surface was higher than the beam translation speed, the formation of the humps on the keyhole wall was observed numerically.

Monte Carlo methods have been used to produce accurate benchmark transport parameters for electron swarm transport in four model gases in the presence of crossed electric and magnetic fields, for a wide range of applied electric and magnetic field strengths. The results of the simulations are compared with those predicted by a multi-term (moment method) solution of Boltzmann's equation and by the flight-time integral method (FTI) of Ikuta and Sugai when possible. The comparison clearly validates the theoretical basis and numerical integrity of the moment method, while supporting concerns about the validity of the FTI approach and the definition of the transport coefficients used. Certain trends in the transport coefficients in crossed electric and magnetic fields are addressed using physical arguments.

In this paper measurements of the behaviour of single microdischarges between metal and glass electrodes for both negative and positive polarities of the metal (designated and transitions, respectively), as well as the overall discharge behaviour between glass electrodes (GG), will be presented for different nitrogen - oxygen and water vapour - air mixtures. Increasing the oxygen concentration in nitrogen decreases the transferred charge per microdischarge for both polarities ( and ), while the total transferred charge per cycle is increased. With increasing water content in air, more charge is transferred per microdischarge for the polarity , but no significant change in the amount of transferred charge was found for a microdischarge. Less total charge is transferred per cycle with increasing water content. This is also true for double-dielectric barrier discharges (GG). The results suggest that water vapour coats the dielectric, reducing the surface resistance and increasing the effective dielectric capacity. This mechanism is supported by analysis of the voltage versus charge plot (Lissajous figure) for the involved capacitances of a double-barrier discharge. The reduction of the surface resistance of the dielectric and the resulting increase in the effective dielectric capacitance, are shown by photographs of spatially isolated microdischarges in a metal - pin dielectric arrangement.

Previous thermophoretic force studies for a spherical particle suspended in a rarefied plasma have been improved upon on the basis of a more detailed analysis concerning the interaction between a charged sphere and ions and electrons. Revised analytical expressions for the thermophoretic force acting on a non-evaporating or evaporating spherical particle for the extreme case of the free-molecule regime and a thin plasma sheath are presented. Besides for the metallic limiting case, analytical expressions are also obtained for the non-metallic limiting cases of particulate materials (with zero electrical conductivity). It has been shown that the ion and electron components of the thermophoretic force are appreciably different from the previous results. The thermophoretic force is directly proportional to the square of the particle radius. Evaporation may greatly enhance the thermophoretic force acting on a particle due to the additional contribution to the thermophoretic force caused by the reaction force of non-uniform evaporated mass flux flowing out from the evaporating particle. This evaporation-added thermophoretic force is more significant for particulate materials with comparatively low latent heat of evaporation and at high plasma temperatures. For the case without evaporation, there is only a small difference between the metallic and non-metallic particles in their thermophoretic forces. However, marked difference may exist at high plasma temperatures between the metallic and non-metallic evaporating particles in terms of their total thermophoretic forces because the different floating potential distributions on the surfaces of the two types of particles cause different distributions of the local heat flux and evaporated-mass flux along the particle surface.

This paper details an investigation into the physics of the external plasma in a Kaufman-type electrostatic ion thruster operating with xenon gas. The external plasma, which is located outside the hollow cathode between the keeper and the anode electrodes, was investigated with a single Langmuir probe. Detailed scans give a peak temperature of eV at the orifice of the keeper decreasing to eV at the anode. The temperature is essentially constant over the external plasma volume. The radial electron density distribution peaks on the axis at decreasing radially to at the boundaries of the plasma; these values have not been corrected for any perturbation of the plasma caused by the presence of the probe. The velocity distribution of the electrons was studied using the second derivatives of the probe characteristics assuming a collisionless sheath. It was shown that the electrons have a Maxwellian distribution. This allowed essentially independent calculations of the electron temperature and plasma potential distribution to be made. These measurements were taken in the stable `spot mode' discharge. Observations in the less stable `plume mode' were also made. However, electrical noise due to plasma instabilities masked the probe signal rendering the analysis inaccurate.

Spatially and temporally resolved spectroscopic measurements of light emitted from positive streamers in transformer oil are presented. Analyses of the measurements performed with a DC needle - plane gap yield electron densities and indications of the atomic excitation temperatures in the streamers. The hydrogen emission reveals an electron density below during the main part of the streamer propagation time (80 - 90%). Later the light is also characterized by emission from a high-density plasma with electron densities in the range . The electron density during this time increases approximately linearly with distance from the initiation point and a density factor of four higher has been measured at the streamer tip than at the root. Measurements with high spectral resolution detect both high and low electron densities simultaneously. A tentative model of the interior of the streamer plasma, spatially resolved, is presented.

This paper is aimed at studying the effect of gas flow on ozone generation in point-to-plane and wire - duct reactors stressed by positive and negative square-pulse voltages (50 ns rise time and 2 ms duration). Two directions of gas flow were attempted; one normal to and the other parallel to the discharge electrode in the reactor. The concentration of the generated ozone increases with the discharge power and is higher when generated by positive pulses (0 - 165 ppm) than generated by negative pulses (0 - 80 ppm) with axial flow in the point - plane reactor. The ozone generated with axial flow is higher than that produced with normal flow irrespective of the polarity of the applied pulses. For positive applied pulses, the concentration of generated ozone in the point - plane reactor is in the range 0 - 80 ppm with axial flow against a range of 0 - 30 ppm with normal flow for the same range of discharge power. The ozone generated in the wire - duct reactor is several times higher than that generated in the point - plane reactor irrespective of the polarity of the applied pulses or the direction of the gas flow in the reactor.

The velocity distributions of ions and electrons are evaluated in the region intermediate between the plasma bulk and the so-called drift region. A largely analytical one dimensional model is provided in which the motion of particles is described by the Langevin equation taking into account the effect of the electric field. In the case of ions, this produces a non-stationary Maxwellian distribution of velocities. For the much lighter electrons, the effect of the electric field is here described using the Druyvesteyn velocity distribution. The non-stationary Maxwellian character of the ion distribution may explain the significant differences in temperature inferred from measurements, causing different spectral lines which may arise at different distances from the cathode.

Characteristics of the hollow-cathode discharge containing negative ions have been investigated focusing on a discharge in oxygen. Discharges with other gases which do not generate negative ions were measured for comparison. The electron energy distribution f(E) was measured by using the probe technique and the negative ion density by the laser photodetachment technique. Radial variations in the negative ion density, the discharge pattern and optical emissions were measured. The form of f(E) has a depletion in the low-energy part, suggesting that the effect of dissociative attachment occurs to generate negative ions.

In this paper, the principal pollution flashover models are reviewed and then an impedance criterion for arc propagation is proposed. This criterion differentiates the case in which the arc elongates until total flashover occurs from the case in which the arc stops before it reaches the end of the insulator. It is shown that the Hampton criterion is not a sufficient condition for initiation of the arc. Indeed, the latter can expand, under certain conditions, even if the Hampton criterion is not satisfied. An analytical model allowing one to determine the critical voltage that insulators can withstand is also established. This is based on an energetic balance, an equivalent electrical circuit and the physical characteristics of the arc. The critical voltage, the critical current and the critical arc length for polluted insulators are calculated using the elaborated model. The results so obtained are found to be in accordance with the experimental ones represented by known empirical relations.

Photoconductive thin-film cells made from a -conjugated polymer, poly-(p-phenylenevinylene) (PPV) can be used for diagnostics of femtosecond laser pulses utilizing the phenomenon of photocurrent autocorrelation. In general, the autocorrelation photocurrent signal can be due to the nonlinearity of the primary photogeneration step or to the influence of the existing photogenerated charge carriers on the efficiency of production of further amounts of charges. Experiments with 150 femtosecond pulses from a Ti - sapphire laser at 800 nm show that two-photon induced photocarrier generation is the nonlinear process leading to the autocorrelation in PPV photoconductive cells. There is no indication of the presence of any photogeneration intermediates with picosecond range lifetimes.

The question of the order of kinetics of the thermoluminescence (TL) of peak 5 in LiF:Mg,Ti has been subject to considerable controversy. Recently, we have shown that peak shape analysis of post-irradiation annealed LiF:Mg,Ti (TLD-100) at 165 leads to non-first-order kinetics, reaching a value as high as following a 165 30 min post-irradiation anneal. When no post-irradiation anneal is applied, peak shape analysis almost always yields first-order behaviour. Isothermal decay data are equally ambiguous. To begin to resolve this discrepancy (at least at the experimental level) we have studied the isothermal decay of peak 5 at 165 (presumably cleansed of the low-temperature peaks 2 - 4 by a 165 15 min post-irradiation anneal). The isothermal decay was initiated at four dose levels of 10, 100, 300 and 1000 Gy, and these decays were not even vaguely similar. The initial rate of decay at Gy was at least five times greater than at 10 Gy and the decays were clearly non-exponential. Our data are not consistent with first-order decay (b = 1) and lead to using general-order (GO) model analysis.

A micromechanical model of an assembly of elastically deforming spherical particles under confined uni-axial compression is described which is based upon percolating load-bearing strings of particles. This model leads to a prediction of the uni-axial compressive elastic modulus of the assembly and the ratio of the transmitted to the applied stress which is less than unity due to friction at the wall of the compression die. Experimental evidence is presented which supports the validity of the model. It is also shown that the approach considerably improves the predictive capability of models developed to calculate the tensile modulus of coherent particle assemblies.

In the context of liquid metal ion source operation, a recent paper suggested that secondary electrons released by ion-collecting surfaces might play a fundamental role in source behaviour, by causing emitter heating. The present note looks again at the role of secondary electrons. It confirms the long-established view that in all normal circumstances, the associated heating effects are not sufficient to modify ion source operating characteristics.

This comment responds to the several criticisms of the paper on the role of secondary electrons in liquid metal ion sources raised by Mair. In particular, it is pointed out that the experimental evidence does not contradict the conclusions of that paper. In addition to the secondary electron experimental data quoted by Mair some more recent results are reviewed.