Recent results of studies on the mechanism of processes induced by the interaction of beams of (reactive) particles with surfaces in UHV, often combined with concurrent bombardment of the surface with low-energy ion or excimer laser beams, are reviewed. Angular-resolved mass spectrometry combined with time-of-flight studies on the desorbing products is used as a key diagnostic. A more complete picture is obtained using several other diagnostic tools to characterize the surface before, during or after the interaction. Interactions at Si and Cu surfaces will be emphasized. The review will deal successively with data for interaction with a single beam of Cl2, low-energy noble gas ions or excimer laser pulses, followed by data on the concurrent interaction of crossed beams of Cl2 and low-energy noble gas ions or excimer laser pulses with these surfaces. Some conclusions will be drawn.

The modification of film stress in evaporated tungsten was studied as a function of deposition environment. Using concurrent ion bombardment of the growing film, the stress was seen to vary systematically with ion energy, ion flux, and substrate temperature. The qualitative behavior fits the model of stress modification developed for niobium films. X-ray diffraction was used to study the structure of the films, and a clear correlation between crystallographic texture and film stress is found. The original structure/impurity model for film stress modification due to ion bombardment has been modified to account for the relationship between film stress and texture.

Ion-assisted deposition allows for independent control of ion energy, ion flux, and adatom flux incident on the growing film. We describe herein general trends in property and structure modification of thin film metals using this technique.

Fe films have been deposited by using dual ion beam sputtering apparatus under various conditions, and the dependence of their magnetic properties and morphology on preparation parameters such as film thickness, δt, and argon gas pressure, PAr, have been investigated in detail. The saturation magnetiza ion 4πMs of the specimen films did not change remarkably with 6t in the range of 50 ∼1000nm. However, with decrease of 6t below 50 nm, 4πMs decreased to less than 20 kG and coercivity Hc increased to more than 16 Oe. As PAr increased from 0.5 to 1.6 mTorr without ion bombardment, 4πMs decreased to less than 20 kG and Hc increased to about 20 Oe. The SEM micrographs of these films deposited at higher PAr showed the columnar structure. On the other hand, the films deposited at Yower PAr and ones bombarded by argon ions with proper kinetic energy during deposition did not present any texture and exhibited better soft magnetism. Such a morphology may be attributed to the difference in arrival energy of sputtered Fe particles to film surface and related closely to soft magnetism. It has been found that the dual ion beam sputtering method can control 4πMs and Hc with changing PAr and so prepare Fe films with superior soft magnetism by adjusting the kinetic energy of bombarding argon ions at lower PAr.

Using in situ, real-time reflection high energy electron diffraction (RHEED), we have measured the evolution of Ge (001) surface morphology during simultaneous molecular beam epitaxy and Ar ion beam bombardment. Surprisingly, low-energy Ar ions during growth tend to smoothen the surface. Bombardment by the ion beam without growth roughens the surface, but the surface can be reversibly smoothened by restoring the growth beam. We have measured the effect of such “ion beam growth smoothening” above and below the critical temperature for intrinsic growth roughening. At all measured growth temperatures the surface initially smoothens, but below the critical roughening temperature the final surface morphology is rough whereas above this temperature the final morphology is smooth.

The secondary ion energy distributions (SIED) emitted from Si under various conditions of targets (Si, SiO2) and primary ions (Ar+, O+) were measured and the thin SiO2 films were deposited by magnetron sputtering techniques under the corresponding conditions to the SIED experiments. The most probable energies of silicon oxygen cluster ions of SimOn+ (m, n=l, 2,…) are equal to those of Al+ thermal ions, while those of Siℓ+ (ℓ=l, 2,…) remain unchanged with the introduction of oxygen in chamber during Ar+ ion bombardment. The currentvoltage plots of SiO2 films are also measured and found to be influenced by the deposition conditions.

It is concluded that the differences in current-voltage characteristics of SiO2 films prepared under various sputtering conditions can be reasonably explained in terms of the changes in the most probable energy of the sputtered particles.

Thin films of transition metal silicides have been produced at low temperatures on Si substrates by direct ion beam deposition (IBD) of the metal ion. Using a mass-analyzed beam of metallic ions rastered over the target at energies on the order of 100 eV and at substrate temperatures near 500°C, stoichiometric silicide films of varying thicknesses up to 300 nm have been formed on both n- and p-type Si. The advantages of this technique over other methods for silicide formation include good control of thickness by current integration, high purity due to the mass analysis, and control of incident ion energy which permits formation of the disilicide phase at low temperatures, thereby minimizing the thermal budget and the associated dopant diffusion in the underlying substrate. Films were characterized by Rutherford backscattering, transmission electron microscopy, and electrical measurements. Co, Fe, Ni, Ti, and W silicides have been formed by this direct deposition process. The effectiveness of the technique has been found to be dependent upon the diffusion characteristics of the particular metal/Si couple involved, with systems in which Si is the dominant diffuser, such as Ti/Si, giving the best results. Stoichiometric TiSi2 films produced at 550°C by this process show low bulk-like resistivity (15 μΩ-cm) without subsequent high-temperature annealing. All of these characteristics make silicide formation by IBD attractive for integrated circuit fabrication and shallow junction technology.

The composition of titanium nitride films prepared by ion beam assisted deposition was studied as a function of the partial pressure of N2 gas in the deposition volume, and as a function of the impingement 'ratio of nitrogen ions (500 eV) to evaporated titanium atoms. The amount of nitrogen incorporated from the ambient gas was derived by subtraction of the fraction introduced by the ion beam. It is shown that the primary effects of ion bombardment are an increase in the sticking coefficient and a reduction in the number of active surface adsorption sites.

Optical thin films of nitrides, oxynitrides and oxides of aluminum and silicon were deposited using ion assisted deposition. Coatings were deposited by thermal evaporation of AlN and e-beam evaporation of Si with simultaneous bombardment with 300 eV ions of nitrogen, a mixture of nitrogen and oxygen or oxygen. The chemical composition and the index of refraction of the coating was varied by varying the gas mixture in the ion beam. Optical properties of and environmental stability of coatings were examined. Results indicated that coatings are stable even under severe conditions of humidity and temperature.

In order to better study and control the processes occurring in a reactive sputtering situation, a unique deposition method has been used in which the Ar-ion-beam sputtering of an elemental Nb target is combined with an auxiliary Ar/N2 plasma at the substrate. The ion source allows independent control of the sputtering parameters (ion flux, energy). The magnetically enhanced triode plasma provides a source of ionized and excited nitrogen at the film surface, and allows independent control of the substrate plasma parameters. A conductancelimiting enclosure surrounds the substrate, resulting in a pressure differential of nearly an order of magnitude between the substrate and target regions. This enables us to separate the substrate and target reaction regimes at low nitrogen flows. NbN has been investigated because of its technological importance and the fact that it is representative of transition metal nitrides.

With the substrate plasma off and N2 provided at the substrate, the cubic superconducting NbN (δ phase) is produced even at low N2 flows, when the target is in the unreacted, metallic state. Upon increasing the N2 flow, the nitrogen content of the films abruptly increases as the target reaction proceeds. The addition of the substrate plasma results in the nonsuperconducting hexagonal δ' phase, which to our knowledge has not previously been produced as a singlephase thin film. The electrical properties of the δ' phase are reported.

Titanium nitride films have been synthesized at room temperature by alternate deposition of titanium and bombardment by nitrogen ions with an energy of 40KeV. The component depth profiles and the structure of titanium nitride films were investigated by means of RBS, AES, TEM, XPS and X-ray diffraction. The results showed that titanium nitride films formed by ion beam enhanced deposition (IBED) had columnar structure and were mainly composed of TiN crystallites with random orientation. The oxygen contamination in titanium nitride films prepared by IBED was less than that of the deposited film without nitrogen ion bombardment. It was confirmed that a significant intermixed layer exists at the interface. The thickness of this layer was about 40 nm for the film prepared on iron plate. The mechanical properties of the film have been investigated. The films formed by IBED exhibited high hardness, improved wear resistance and low friction.

We have produced alloy films of B(1−x)Nx (x:0 to 0.5) via ion beam assisted deposition (IBAD). Rutherford backscattering spectroscopy (RBS) measurements show that the stoichiometric films are achieved with a nitrogen to boron atom arrival rate of N/B = 2.55. The film N/B ratio for all measured compositions is fit with a model that predicts that the nitrogen sticking coefficient is 0.36. Measurements of B(1−x)Nx film thicknesses with a profilometer confirm the value of this sticking coefficient. The BN film is transparent and has an refractive index that varies between 1.90 and 1.77 in the range 400 nm to 3100 nm. The refractive index data can be fit with the Sellmeier equation. Films 500 run thick with N/B less than 0.6 appear metallic; the index decreases linearly with increasing nitrogen content. The films have two broad absorption peaks, centered at 750 cm−1 and 1370 cm−1 that correspond to BN vibrations. The characteristics of the bands suggest that the films are amorphous or a microcrystalline hexagonal phase. The Knoop microhardness of the films generally decrease with increasing N content. Stoichiometric films have a microhardness of Hk = 20 GPa, measured with a 25 gm load.

The concentrations of the main constituents of hard i-BN films produced by the ion beam assisted deposition were measured using the nuclear resonance reactions 11B(p,γ) 12C at Ep = 163 keV, and 15N(p,αγ) 12C at Ep = 429 key, respectively. The hydrogen contamination of the samples was investigated using the forward recoil spectroscopy (FRES) technique with a 2 MeV He+ beam. Some complementary analyses of carbon and oxygen were performed using (d,p) and (d,α) -reactions.

Hyperstoichiometric boron concentration were found in almost all films. The relative concentrations of nitrogen and boron were also slightly dependent on the deposition conditions as well as the deposition temperature. Contrary to this, hydrogen contamination, that was generally at a low level with few exceptions, was more independent of these parameters.

A new apparatus for dynamical ion beam mixing has been developed in our laboratory. The system is based on deposition by ion beam sputtering concurrently with ion beam implantation. It operates under a vacuum of 10−5 Pa before deposition and 10−3 Pa during deposition.

A general scheme for the analysis of deposition from hyperthermal (10–5000 eV) species is presented. Successful deposition involves consideration of species range, maximum local concentration obtainable, trapping efficiency, radiation damage, and sputtering efficiency. Examples of in situ parametric investigations of carbon deposition performed with a controlled mass selected UHV ion beam facility are presented. A subplantation model for diamond film deposition is discussed. XRD evidence for epitaxial growth of diamond(111) on Si(111) is provided.

Mass analyzed and non-mass analyzed carbon ion beams were used to deposit coatings on polished silicon wafers. Extremely, hard dense coatings that are structurally similar to i-carbon as determined by Ramon spectroscopy were achieved with high current beams of methane and argon. Residual contaminants of oxygen and nitrogen were minimized.

The effects of ion beam bombardment during ionized cluster beam (ICB) deposition of metal films on Si(111) and Si(100) substrates have been discussed. In the case of Al deposition, films have been epitaxially deposited on Si(lll) and Si(100) substrates at near room temperature. On Si(111) substrates, nearly perfect Al single crystal films could be formed. On Si(100) substrates, Al bicrystals have been grown epitaxially. A remarkable fact concerning these results is that the epitaxial films could be formed at nearly room temperature and on a large lattice mismatch (25%) substrate surface. Atomic resolution TEM analysis suggests that the epitaxy of Al occurs not only on Si surfaces but also at Al/Al grain boundaries. These epitaxial films exhibit extremely high thermal stability and long electromigration life time. To understand the deposition features and film characteristics, the effects of ICB bombardment on the film growth at the initial stage of the deposition and the resultant film structure have been studied. The results show that the role of very low energy ion bombardment is especially important in forming epitaxial metal films. Depositions of Au and Cu on Si substrates have also been made to understand whether ICB deposition may improve the characteristics of other metal films. Preliminary results of these film depositions are also obtained.

Ionized Cluster Beam (ICB) deposition has received considerable attention in recent years because of its potential application to low temperature semiconductor and metal film growth. To better understand this deposition process, however, it is necessary to determine the properties of the cluster source and the cluster size distribution obtained. This paper discusses the design of a time-of-flight mass spectrometer with single atom resolution for the measurement of cluster size in a conventional ICB deposition system. Results have been obtained for silver and germanium deposition. In both cases most of the clusters have been found to consist of less than 35 atoms. With the high resolution spectrometer it has been possible to show that certain cluster sizes are more stable than others.

An ionized cluster beam (ICB) source was used to deposit Al onto SiO2 substrates. A 60 gtm diameter wire held at the substrate served as a mask. After Al deposition, the wire was removed and the masked area was examined by scanning electron microscopy (SEM) and by scanning Auger microprobe (SAM). The ICB source was operated at 0, 3, and 6 kV acceleration voltages. The substrate was held at 80°, 200°, and 400°C during Al depositions. The Al deposition rate averaged 240 A per min. The chamber pressure during deposition was 2×10-6 Torr. The diffusion distance of Al under the mask edge was determined from the SEM micrographs and SAM line scans. The maximum diffusion distance for all acceleration voltages occured at a substate temperature of 200°C. The maximum diffusion distance at 200°C was 29 μm at 6 kV acceleration voltage. The minimum diffusion distance was 12 μm at 400°C for an acceleration voltage of 6 kV.

Various techniques which have been applied to modeling low-energy (<< 1 keV) ion-solid interactions on an atomistic scale are described. In addition to their individual strengths, all such methods also have a number of drawbacks, both fundamental and practical. The range of validity, and the problems encountered external to this range, will be outlined for the different approaches. Finally, examples of molecular dynamics simulations of low-energy ion-solid interactions will be presented.