Several different insulating substrates were coated with various metal films and ion beam irradiated using either Xe or Kr ions. These were then examined by Rutherford backscattering spectroscopy and scanning electron microscopy to determine if interfacial mixing had taken place. These results were compared with the sign of the reaction enthalpy of the metal and substrate to test the proposition that metals mix on insulators if the reaction enthalpy is negative and do not mix if it is positive. The enthalpy rule is in general valid. Two exceptions were found: Cr on Si02 and Zr on Al203; however, these exceptions contain ambiguous features.

Irradiation with a light, reactive ion was found to produce no mixing.

A new geometrical theory of sputtering is extended to investigate the behaviour of interface mixing as a function of the angle of ion incidence. It is found that the amount of mixing peaks with the angle of incidence at different angles depending upon the depth of the interface. When the interface is at the surface the mixing parameter behaves as the sputtering yield and maximises around 70°, whilst deeper interfaces, around the damage range, have maximum mixing at normal incidence.

The use of rapid thermal processing is reported for simultaneous formation of TiSi2 from Ti deposited layers and activation of As or Sb implanted profiles in Si. Properties of the silicide and the doped Si are reported with emphasis on impurity redistribution and defect removal.

Ion-beam mixing studies were performed with arsenic implantation up to a dose of 1015/ cm2 through Ti on n+-Si contacts. Lateral silicide growth of many microns over the contact edges was observed for furnace annealed samples. The lateral growth is greatly enhanced by ion-beam mixing and can be suppressed with silicon nitride capping or using rapid thermal annealing (RTA) with a flash-lamp system. For RTA annealed samples. the implanted arsenic profile measured by RBS showed no snowplowing during the silicide growth process. Hligh contact resistivity was measured in the high-temperature annealed samples. This observation is consistent with the lower dopant concentration at the silicide/Si interface, due to silicide growth consuming the n+-Si layer.

We have determined the dominant moving species during the formation of CrSi2, Ni2Si, and Pt2Si by irradiating with 600 keV Xe ions thin bilayers of evaporated metal on silicon and silicon on metal above room temperature. Two very thin markers were placed at the metal-silicon interface and at the bottom interface with the Si02 substrate. The separation of the two markers as a function of the amount of silicide formed by either ion mixing or thermal annealing was monitored by backscattering spectrometry. This experiment identifies the dominant moving species. The results establish that the moving species for ion mixing and for thermal annealing are the same in CrSi2, but not in Ni2Si and Pt2Si.

The linearity of mixing rates of ion-beam mixing of metals on silicon has been found to depend critically upon the method by which the mixed depth is calculated. For nonstoichiometric mixing, several methods of calculating the mixed depth may be used, namely: integrated area, halfheight, moment, error function, and 10%-90%. When applied to the same data, these methods can yield divergent results, from linear to square-root dependences of the mixed depth upon the mixing dose. For stoichiometric mixing, the calculation of mixed depth is more straightforward, but different methods of calculation still yield widely differing results. Possible causes for these discrepancies are discussed.

- P+ self-annnealing implantations through evaporated Ti is performed to simultaneously form the silicide and the junction underneath, avoiding the need for further post-implantation thermal treatments. The formation of the TiSi2, phase and the good quality of the silicon-silicide interface are observed, when operating at power densities ≤ 16 W/cm2. Deep phosphorus distributions extending below the silicide down to 480÷600 nm are obtained.

TiN films prepared by implantation onto evaporated Ti layers are tested as diffusion barriers in contact with a thick Al overlayer. Both the TiN/Al and TiSi2/TiN/Al contact structures are evaluated, after thermal treatments up to 600 C, on shallow junction diodes and four terminal resistor test patterns for contact resistance measurements. It is shown that, upon annealing at 600 C, the TiSi2/TiN/Al contact system still exhibits excellent electrical performances. The degradation is found to depend on TiSi2 thickness and contact area.

We have explored the origin of the ion-induced conductivity in polymer films and the distinction between low energy ion-implantation and high-energy ion irradiation. In experiments involving irradiation of polymer and carbon films with ions of energy from 200 keV to 25 MeV we have established that with both low and high energy ions the polymers undergo carbonization. However, the saturation resistivity obtained with low energy implantation is four to six orders of magnitude larger than those obtained by high energy ion irradiation. In experiments on irradiation of carbon films low energy ions caused a two orders of magnitude increase in the resistivity while high energy ion caused a two orders of magnitude decrease. This implies that the structure of the carbonized polymer is different for the low and the high energy ion irradiation. While in the former case there may be no crystalline order in the films; in the latter case, a microcrystalline graphitic structure is obtained (with four orders of magnitude larger conductivity than in the former case). The formation of graphitic crystalline order with increasing high energy ion dose was verified by electron energy loss spectroscopy. This is an interesting example of crystallization induced by electronic excitation alone with no macroscopic thermal effect.

Under exposures of high energy radiation in vacuum, decarboxylation takes place efficiently on polymer surfaces having carboxylic acid groups. Low energy electron and ion beams are particularly efficient in these chemical reactions on the surfaces because their energy deposition is limited to very shallow depths. The surface areas exposed to the radiation have properties similar to hydrocarbons after decarboxylation, resulting in non-reactions in gas phase silylation, while unexposed areas undergo silylation readily due to the presence of reactive acid groups. The non-silylated areas etch faster in oxygen reactive ion etchings, giving rise to positive tone polymer images. Polymers like poly(methacrylic acid), poly(acrylic acid), and their copolymers have been used. Other polymers like poly(styrene), poly(chloromethylstyrene), and Novolac resins can also be used, because these polymers readily undergo photo-oxidation under uv-irradiation in air, generating carboxylic acid groups on their surfaces.

Outgassing of HUNT HPR 204 photoresist during ion implantation has been characterized by means of quadrupole mass analysis. The influence of the various parameters (implantation eneray, beam current, ion species) has been studied in order to point out the evolutions, in terms of time or implanted dose, of the target chamber pressure and the peaks at masses 26 and 28, characteristic of the photoresist outqassing. The pressure integral in the target chamber which is easily related to the integrated outgassing flux, has been found essentially dependent on the total energy deposited in the photoresist layer and on the ion penetration depth.

We have observed significant improvements in the efficiencies of dendritic web and edge-supported-pulling silicon sheet solar cells after hydrogen ion beam passivation for a period of ten minutes or less. We have obtained electron-beam-induced current data that show the hydrogen passivation of dislocations as well as grain boundaries in edge-supportedpulling silicon sheet solar cells. We have studied the effects of the hydrogen ion beam treatment with respect to silicon material damage, silicon sputter rate, introduction of impurities, and changes in reflectance.

Recoil ion implantation of In2O3 into soda glass substrates has been investigated. Increased adhesion results, while the optical and electrical properties are altered. Large energy deposition rates can lead to the reduction of the oxide, thus decreasing visible transmission. Thermal annealing in air results in the recovery of optical and electrical properties. The influence of ion energy and dose on the modification of the glass has been examined.

Cooper and platinum films were deposited by evaporation and sputtering techniques onto prepared substrates of alumina, sapphire (cut along an a or c axis) and yttria stabilized zirconia. The films are then bombarded with ions of H, He, Ne and Kr at energies within and outside the ion beam mixing regime. Ion beam induced modification in adhesion and its thermal stability were measured by three techniques - a scratch test, a pull test and a peel test. Adhesive energies of these solid-solid systems were determined by contact angle measurements using scanning electron microscopy. The resistance of the films to chemical attack is modified by ion bombardment and will be shown to correlate with adhesion alterations.

Recent advances of focused ion beam systems and their applications are presented. The applications include maskless ion implantation and various maskless patterning techniques which make use of ion induced chemical effects. These are ion beam assisted etching, deposition and ion beam modification techniques and are promising to improve patterning speed and extend applications of focused ion beams.

A focused ion beam of gallium was used to implant into the top layer of a bi-layer structure on a silicon wafer. The structures studied were spin-on glass/organic and electron beam deposited Si/organic. Plasma etch conditions were found which gave substantial etch rate differences between the implanted and non-implanted material. The material was then patterned by dry processing methods. The plasma etch produces a vertical wall 2.0 µm high with a rectangular profile free of debris. The variation in line width along the line is a small fraction of the line width. Or, the line width is constant along the line and is approximately equal to 0.4µm, the spot size of the focused ion beam on the wafer. The variation in line width as a function of implantation dose was determined and shown capable of yielding a reproducible line width. The results are shown to be superior to single resist exposure. Potential applications are presented.

A theoretical model of liquid metal ion source (LMIS) operation has been developed by Kingham and Swanson. In this paper we consider beams from LMIS on the basis of this model. In particular we consider properties such as angular intensity, energy spread and relative abundance of differently charged species of the ion beam, and the dependence of these properties on source current and elemental composition. The conclusion is that the brightest focussed beam for a given probe size is attainable at the lowest possible source current as previously stated by Swanson. LMIS sources have an onset current of typically 1-2[A and will not operate stably below this current, thus limiting the maximum focussed ion beam brightness. The physical reason for this is discussed. The relevance of these properties to fine focussed ion beam applications, particularly semiconductor processing, is discussed. Useful, and in some cases unique, device manufacturing techniques can be postulated using one or more of the momentum, energy or atomic addition properties inherant tothis type of system. Advanced research tools are discussed, together with some examples of the use of microfocussed ion beams with probe sizes down to less than 50nm. Immediate applications include: high resolution ion imaging and SIMS microanalysis; ion beam machining and microfabrication; ion beam resist exposure and ion beam mask repair.

Thermal-wave measurement techniques have recently been developed for measuring ion implant dose over the range 1E10 to 1E16 ions/cm2. The spatial resolution obtained with this method is 1 µm, an improvement of approximately 3000X over that of other dose monitors. This resolution capability together with the fact that the thermal-wave measurements are noncontact and nondestructive, permit dose measurements to be made on implanted areas (device features) on patterned wafers, as well as on test wafers.

A new potential application of the thermal-wave technique is the measurement of dose implanted by focused ion beam (FIB). Since the FIB technique is a high-resolution, direct-write process and since current FIB systems are limited to scan fields of less than 0.5 mm2, a high-resolution probe to monitor dose and uniformity is necessary.

In this paper, we report the measurement of ion dose implanted by focused beams into areas of ‹10−2 mm2. Even single FIB-implanted lines of 0.20 µm width have been detected, and programmed lateral dose variations measured. Experimental results demonstrating the potential of this technique are presented.

An ion-induced decomposition of thin metal-compound films on insulating surfaces is described. 100-keV ion beams of Ar+, Kr+, and Xe+ was used to decompose thin (~10 nm) films of a palladium compound on polyimide and alumina. Changes in the stoichiometry have been observed with Rutherford Backscattering Spectrometry. The compound was also decomposed in a pattern using a mask, and the nondecomposed film could easily be removed with a suitable solvent. The decomposed palladium compound was used as a catalyst for a direct, electroless metal deposition of Cu. The plating speed was 30 nm/min, and a 4-5 µQcm copper film could be obtained after 3-4 min plating. Plated metal patterns are demonstrated on both polyimide and alumina, with a resolution in the µm range. On polyimide plating of a three-dimensional structure has been demonstrated.

Progress in applying rapid thermal annealing for activating implants is surveyed. Advantages to be gained through curtailing substrate semiinsulating loss, enhancing heavy n-type activation, and in reducing the diffusive redistribution of p-type implants are discussed in detail. Consideration is given to encapsulation requirements and the role of stoichiometry in activation. While the emphasis is on GaAs, work on InP and InGaAs has also been included.