Table of contents

The only difference between nanotechnology and many other fields of science or engineering is that of size. Control in manufacturing at the nanometre scale still requires accurate and traceable measurements whether one is attempting to machine optical quality glass or write one's company name in single atoms. A number of instruments have been developed at the National Physical Laboratory that address the measurement requirements of the nanotechnology community and provide traceability to the definition of the metre. The instruments discussed in this paper are an atomic force microscope and a surface texture measuring instrument with traceable metrology in all their operational axes, a combined optical and x-ray interferometer system that can be used to calibrate displacement transducers to subnanometre accuracy and a co-ordinate measuring machine with a working volume of (50 mm)³ and 50 nm volumetric accuracy.

An e-beam nanolithography process with metal lift-off is optimized to produce four-electrode co-planar metal-insulator-metal nanojunctions with an inter-electrode distance down to 10 nm. The e-beam exposure technique is simulated to find the best geometry of the four electrodes. The fabrication process is presented and compared in detail with the simulation. AFM images of four-electrode nanojunctions down to 10 nm are presented. These nanojunctions are well adapted to interconnect hybrid molecular electronic devices.

Gold nanoclusters, chemically passivated with decanethiol and dissolved in toluene, have been deposited from solution onto selected regions of oxidized silicon (100) surfaces patterned either with photoresist or an etched step. When the perimeter of a droplet crosses the boundary between the resist and the silicon surface, we observe transport of cluster solution along such discontinuities, outside of the droplet. Such guided flow can extend for over 600 µm across the surface, producing cluster chains as narrow as ~120 nm, once the toluene has evaporated. The same experiment with an etched step produces no transport of clusters, but rather selective deposition and growth around the discontinuity. These different responses are attributed to the step/boundary material - principally its interaction with the toluene solvent during evaporation.

We show that regular arrays of pillars and helices, with repeat distance below 100 nm, can be grown by oblique deposition onto a patterned rotating substrate. The size limits for such structures are discussed.

This paper deals with a new tracking control method for piezoelectric actuators. When actuating in an open-loop manner, in order to compensate for the creep effect of the piezoelectric transducer as well as hysteresis, a new concept of `voltage creep' is proposed, which is represented by the two parameters V₀ and γ_v. It is shown that these two parameters have the properties of wiping out and congruency, which make it possible to derive parameter values appropriate to control piezoelectric actuators using the Preisach model. Finally, a tracking control experiment of piezoelectric actuators for an arbitrary desired trajectory is performed giving greatly improved results compared to other open-loop actuating methods.

The thermal conductivity of several single-wall carbon nanotubes has been calculated over a temperature range of 100-500 K using molecular dynamics simulations with the Tersoff-Brenner potential for C-C interactions. In all cases, starting from similar values at 100 K, the thermal conductivities show a peaking behaviour before falling off at higher temperatures. The peak position shifts to higher temperatures for nanotubes with larger diameters and no significant dependence on the tube chirality is observed. It is shown that this phenomenon is due to the onset of Umklapp scattering, which shifts to higher temperatures for nanotubes with larger diameters.

Two bola-amphiphilic α,ω-diboronic acids separated by a (CH₂)₁₁ or (CH₂)₁₂ group were synthesized. Complexation with chiral diols readily gave new amphiphiles end-capped with the chiral substituents. Some of these acted as good gelators of organic solvents. Transmission electron microscope and scanning electron microscope observations established that a variety of super-structures are created in the organogels, depending on the solvents and the structure of the chiral end-cap groups. In most cases, the fibrous aggregates, the network structure which is the driving-force for gelation, showed a helical higher-order structure reflecting the chirality of the end-cap groups. The results indicate that the combinatorial approach utilizing boronic acid functions and diol compounds is useful in creating a variety of new super-structures in the gel phase.

Different types of InGaAs/GaAs deep-etched quantum wire (QWI) structure were successfully fabricated by high-energy electron beam lithography on GaAs(100) surfaces. A selective wet-chemical-etching technique, preceded by chemically assisted ion-beam etching, reduced the controlled lateral dimensions of the wires to ~10 nm due to strong under-etching. Various types of wire in the [01] and [011] crystallographic directions were prepared by the combined etching method. The side-walls of the wires were defined by the selectively etched low-index crystallographic planes. A molecular-beam-epitaxy-grown graded InGaAs/GaAs quantum well was realized at the narrow `neck' region of the wires, thus providing the strongest possible lateral confinement of the QWI structure. Consequently, similarly to the selective growth of self-narrowing ridge structures, selective wet-chemical etching induced a controlled self-narrowing of the wire structures. Scanning electron microscopy images of the QWI nanostructures showed smooth side-walls defined by the crystallographic planes. Low-excitation photoluminescence spectroscopy of the structures revealed extremely high quantum efficiency and a size-dependent blue shift as a result of the strong lateral confinement.

We present results to show the effects of device scaling on the relative performance of three device technologies, namely resonant tunnelling diodes (RTD), electronic quantum cellular automata (QCA) and complementary metal-oxide-semiconductor (CMOS) transistor technology. The minimum feature size (MOSFET or HFET gate lengths λ) is scaled from 250 to 50 nm, while QCA inter-dot separations of 20 to 2 nm are examined. The comparison is made using a standard digital circuit architecture, namely a memory-adder model. Our aim is to compare these device technologies on the system level rather than individual device level. The results show that the RTDs offer speed advantages over CMOS, but improvements in the circuit density are limited. The electronic QCAs will suffer both from effectively low packing density and low operating speeds in comparison to CMOS if conventional designs and a 2D architecture are used.

The equivalent electrical circuit of a single C₆₀ electromechanical transistor in a planar lay-out is presented using its experimental STM characteristics. This circuit is used to demonstrate that such a hybrid molecular electronic device can be used as a class A amplifier, a NOT or NOR gate and to implement an SRAM memory point. All the devices are simulated using the SPICE routine to find their optimum load resistance and cantilever grid size. The class A amplifier can operate with a cut-off frequency of a few gigahertz while the logic gate and memory are limited to a few tens of megahertz, but for a very small power design in the picowatt range.

Ultrasonic vibration can be nonlinearly detected by means of an atomic force microscopy cantilever when the tip is in contact with a sample surface owing to the so-called (sample-induced) ultrasonic force. The procedure has been developed as a novel technique, ultrasonic force microscopy (UFM), that provides information about the nanoscale elastic and adhesive properties of surfaces. Here, we compare differences in the UFM signal when ultrasound is excited from the back of the sample (sample UFM) and from the cantilever base (waveguide UFM). UFM relies on the nonlinear ultrasound-induced cantilever displacement (due to the aforementioned ultrasonic force), and does not monitor the linear high-frequency vibration of the cantilever. In this paper, we discuss the influence of a linear high-frequency cantilever response in the UFM measurements and provide experimental evidence of the feasibility of nonlinearly detecting the free ultrasonic cantilever vibration when the tip is out of contact with the sample surface using the typical laser-beam deflection method for monitoring cantilever displacements.

We proposed a method of implementing the Boltzmann machine neural network on electronic circuits by making use of the single-electron tunnelling phenomenon. The single-electron circuit shows stochastic behaviour in its operation because of the probabilistic nature of the electron tunnelling phenomenon. It can therefore be successfully used for implementing the stochastic neuron operation of the Boltzmann machine. The authors developed a single-electron neuron circuit that can produce the function required for the Boltzmann machine neuron. A method for constructing Boltzmann machine networks by combining the neuron circuits was also developed. The simulated-annealing operation can be performed easily by regulating an external control voltage for the network circuits. A sample network was designed that solves an instance of a combinatorial optimization problem. Computer simulation demonstrated that, through the simulated-annealing process, the sample network can converge to the global minimum energy state that represents the correct solution to the problem.

The phase diagrams of small particles (with diameters in the nanometre range) are studied theoretically. In the limit where thermodynamical arguments remain valid, it is deduced that the phase diagram of small particles is a function of their size. This is discussed for the cases of eutectics and regular solutions. The effects of surface segregation are also treated and lead to further modifications of the phase diagrams.

Photonic transfer through elongated optical structures of the submicrometre section microfabricated at the surface of dielectric or semiconductor samples can be enhanced by an appropriate structuration of the local refraction index (1998 Europhys. Lett.44 686-92). We show from computerized simulations, that both the light localization and the spectroscopic properties of such structures can be used to selectively excite, in coplanar geometry, individual molecules located in the near-field.

Various nanostructures of both convex and concave shapes have been successfully fabricated on newly discovered stable surfaces of NdBa₂Cu₃O_y (Nd123) single crystals and thin films by means of the tip of an ultrahigh-vacuum scanning tunnelling microscope (STM) at room temperature. It is also demonstrated that the materials on the surfaces of Nd123 thin films can be removed gradually using an STM tip milling technique. Possible mechanisms responsible for the fabrication processes are discussed. From these preliminary results, we suggest that many other microscopic or macroscopic modification and patterning processes may be performed on the as-prepared surfaces of Nd123 high-temperature superconductors using scanning probe microscopes.

While performing some calculations based on the results in my paper on nanotechnology implications of space systems, I discovered an error in table 3, about which I wish to alert readers. The first and third column of 'Cost per kg to reference orbit' are out of proportion.

The paper purports that the first column of 'Cost per kg to reference orbit', ranging from $14k to $59k, is based on a per mission cost estimate of $1000 per kg of dry, empty launch vehicle mass, but the actual numbers correspond to a cost estimate of $5000 per kg. Thus, the ratio between the first and third columns is overstated by a factor of five. If the numbers shown in the first column are reasonable estimates of achievable cost to orbit for a new, titanium-based fully reusable rocket system, then the costs for an MNT-based system using the same cost model ought to be a factor of five higher than shown in the third column. If one accepts that the costs in the first column could reasonably be a factor of five lower than shown, then the MNT-based third column is acceptable as is.

This error does not impact the second column of 'Cost per kg to reference orbit', nor any other aspect of the paper.