M. Harsdorff

Heterogeneous nucleation and growth of thin films

Abstract A review is given of the current discussion concerning heterogenous nucleation limited to the Volmer-Weber mechanism. We consider the theoretical but stress the experimental aspects. Various coating techniques, e.g. evaporation under ultrahigh vacuum conditions and r.f. sputtering, along with modern analysis and data-processing methods, such as transmission electron microscopy, the scanning electron microscopy combined with X-ray fluorescence analysis and digital quantitative image analysis, are described. The efficiency of existing theoretical models is discussed in terms of selected examples of experimentally investigated systems such as noble metals on alkali halides, mica and carbon. Furthermore, the extension of the kinetic nucleation model to the condensation of alloys, condensation on disturbed surfaces and oriented growth will be considered.

Epitaxial nucleation of Au on KBr (100) surfaces

Abstract In order to study the epitaxy of gold on KBr (100) surfaces, the degree of epitaxial order was measured and its dependence on the experimental parameters evaporation time, flux rate and substrate temperature was determined. The specimens were produced in UHV, fixed in situ , and examined under an electron microscope. The micrigraphs taken in the electron microscope were analysed quantitatively with an electronic image analysing computer, the area covered, the cluster density and the cluster size distribution being investigated. The results confirm a theory which differentiates between spontaneous orientation and cluster alignment.

Nucleation and growth of binary alloys on substrates

Abstract The nucleation and growth of silver-gold alloys on rock-salt cleavage surfaces was investigated systematically by evaporating the two metals from different sources so that the alloy was formed on the substrate surface. Measurements were carried out on the time dependences of the crystallite populations and size distributions and of the composition of the condensed material. It was found that the composition varies strongly with crystallite diameter and with the deposition time. The results were interpreted in terms of an extended kinetic nucleation model from the difference between the single atom adsorption energy and the activation energy for surface diffusion of the components. The adsorption energy for gold on rock-salt was found to be about 0.08 eV higher than for silver, whereas the activation energy for surface diffusion was about 0.02 eV lower for gold than for silver. These results were compared with data from other workers and show some discrepancies which may be due to different surface conditions.

Nucleation and growth kinetics of gold films deposited onto rock salt single-crystal (100) surfaces by r.f. sputtering with helium

Abstract The results of a systematic study of the nucleation and growth of gold clusters deposited by r.f. sputtering onto NaCl(100) substrates are presented in this paper. In contrast with evaporation experiments the cleaved surface is exposed to helium during deposition. The variations in the island density and the maximum cluster diameter with the deposition time, deposition rate and substrate temperature were studied. The results, e.g. nucleation rate and growth kinetics, are compared with existing data for evaporation experiments. The growth kinetics in the present study and in evaporation experiments are found to be identical although the data for the other aspects are very different. In order to explain the results obtained for the nucleation rate it has been necessary to assume that some of the sputtered atoms can produce defect sites on the substrate which affect the nucleation process. By extending the kinetic nucleation model it is possible to determine a value for the activation energy of a gold atom leaving a defect site and the fraction of atoms that can produce defects.

Improvement of experimental work on nucleation kinetics by means of a “quantimet 720” image analysing computer

Abstract Extensive experimental work on nucleation kinetics usually leads to the very time-consuming evaluation of many electron micrographs. The possibility of overcoming this problem by the use of an image analysing computer is discussed. The main features of such an instrument are described with reference to the measurements which can be made. It is found that those quantities which are commonly measured, such as the number of clusters, may be obtained with a great saving of time. The electronic method also makes possible other measurements, especially information about cluster dimensions, which could scarcely be obtained otherwise.

Mobility of small gold crystallites on the cleavage planes of alkali halides

Abstract The mobility of small gold crystallites on KBr cleavage planes was investigated by measuring the distances between the crystallites and calculating the spatial distribution function P ( r ). There is a pronounced maximum in P ( r ) due to the existence of repulsive forces between adjacent crystallites. Spatial distribution functions for various diameter intervals were calculated for the reference crystallites in order to obtain information about the effect of the diameter. Although the distances between the centres of gravity of adjacent crystallites vary, the distances between their edges remain constant. In the Au/KBr system the range of the repulsive forces is independent of the crystallite diameter. No effect of the shape and the preferred orientation of the crystallites on the P ( r ) functions was detected.

Extension of the kinetic nucleation model to binary alloys

Abstract The kinetic nucleation model in its simplest form, neglecting coalescence effects and the mobility of clusters, was extended to systems with two components deposited simultaneously from different evaporation sources. By this extension of the model, the composition at the very beginning of the condensation process can be predicted as a function of the evaporation rates of the components, the substrate temperature and the differences in their adsorption and activation energies for surface diffusion. Calculations were made for the AuAg system on alkali halides using energy values given by several authors. The results demonstrated that it is possible to determine the differences in the energy parameters with some accuracy by the experiments described.

The influence of charged point defects and contamination of substrate surfaces on nucleation

Abstract Point defects are created in many different ways such as the cleavage of alkali halides to obtain uncontaminated substrate surfaces. It is also possible to create point defects by heating a substrate such as mica to high temperatures. Moreover, the bombardment of the substrate surface with charged or neutral particles arising from an evaporation or sputtering source can generate point defects in sensitive substrates. In contrast with steps which can be made visible by decoration, point defects are not immediately evident. Their existence is recognizable on account of characteristic changes they cause in the nucleation process. Examples will be presented of all the above-mentioned processes yielding point defects, and their various effects on the nucleation process will be discussed. The influence of surface contamination is much more difficult to determine especially if more than one contaminant is present as must be assumed for experiments carried out in a high vacuum. It is possible to obtain an overall picture of the influence of contaminants in nucleation experiments by letting gas into ultrahigh vacuum and thus producing a known contaminated layer on which nucleation experiments are then carried out. Some examples will be discussed.

Scattering experiments with gold on rock-salt and potassium bromide (100) surfaces

Abstract We describe a set of experiments which was carried out in order to test basic assumptions and predictions of the kinetic nucleation theory developed by Frenkel, Zinsmeister and others. Measurements of the flight time and angular distribution of scattered atoms confirmed that they are accommodated on the surface. The surface lifetimes were found to be as predicted on KBr ((5±2) × 10 -5 s) but not on NaCl ( -5 s). An explanation for this discrepancy based on the determination of the fraction of gold dimers in the adatom beam is given.

Experimental problems of studying thin film nucleation

Investigations of heterogeneous nucleation on substrates have been carried out for several decades and modern theoretical concepts of the problem are more than fifty years old. Nevertheless the full understanding of the mechanisms needed for an interpretation of all experimental results is still lacking. This is particularly evident in the case of oriented overgrowth. In many cases experiments were performed in a high vacuum environment with pressures of about 10 -5 Torr without any control of the composition of the residual gases. Now we know that successful investigations of highly surface-dependent processes are not possible under such conditions. In the recent past, some progress has been achieved towards a better understanding of the most important processes involved in the condensation of materials on substrates. Theoretical efforts have been concentrated on heterogeneous nucleation 1-3, atom migration and capture processes on substrate surfaces 4-6, and on the mobility of small crystallites 3,7-9. Experimentally, the introduction of ultrahigh vacuum (UHV) techniques was the decisive condition to obtain reproducible data. Furthermore UHV techniques permit the application of powerful surface analytical tools such as low energy electron diffraction (LEED) and Auger electron spectroscopy (AES). With these methods it is possible to control the structure, chemical composition and cleanness of substrate surfaces before any deposition of material. Furthermore the orientation and quantity of deposited material can be monitored with a sensitivity and accuracy to a fraction of a monolayer. Besides these new experimental methods, the classical high resolution transmission electron microscope plays an important role in giving accurate data about the morphology and structure of the crystallites forming the deposit. The electron microscope has two advantages over many other averaging analytical instruments: firstly a resolving power of a few ~ingstr6ms, and secondly the possibility of obtaining crystallographic data by selected area electron diffraction from areas of about 1/~m 2. In investigations of epitaxy it is possible to get quantitative information about the degree of orientation by using the instrument in a high resolution dark field mode 1°.