Introduction:
Previous CMN issues have included basic discussions on sputter deposition of thin films for optical and electronic applications. This edition will explore the technique in greater detail. Sputtering has several advantages over thermal evaporation by electron-beam or resistance-heated sources. A greater variety of materials, including alloys, compounds, pure metals and mixtures, can be sputtered. Disadvantages include higher equipment and materials costs and lower deposition rates. The electronics and media storage industries are the big users of sputter deposition because of the advantages of thickness control, high density, composition reproducibility, and automation.

Physics of the Sputtering Process:
The general process of sputtering requires the displacement of target atoms through the transfer of momentum carried by ions possessing high kinetic energy. Typically, argon is the working gas. Argon⁺ ions can be created in an ion gun which then imparts kinetic energy and directs the ions toward the target to be sputtered, or in a plasma that contains Ar⁺ and electrons. The plasma glows because of reactions between the electrons and atoms and ions and is neutral in charge. The spectral content of the glow is indicative of the ion species present and can be used to control the composition of the deposited film.

In the case of plasma sputtering, the target, which is the source material, is made the cathode, and the chamber walls or some other electrode is the anode. A voltage is developed across these electrodes; a discharge plasma is developed which generates electrons and ions and imparts kinetic energy to the ionized working gas. Ar⁺ ions bombard the target freeing surface material. The interactions between electrodes and ionized species and electrons is complicated, and the variety of sputtering configurations existent emphasize specific aspects of the plasma physics that is involved. For example, in magnetron sputtering powerful permanent magnets behind the target contain electrons in their fields to increase the probability of collisions with atoms and metastable species and thereby increase the density of available ions. In all forms of plasma sputtering, a virtual electrode is created at the boundary between the plasma and a volume known as the Crook's dark space, where electronic and ionic interactions are absent. Ar⁺ ions are extracted from the plasma and accelerated across the dark space to impinge on the target. During the momentum transfer at the target surface, positive and negative ions and electrons as well as atoms, dimers, and trimers are released. The positive ions return to the target where they contribute to heating. In some arrangements, negative ions and electrons can strike the substrate located near the anode.

Sputter rate is determined by target voltage and current density, as well as chamber pressure. High voltage and current (power) releases more sputtered species; high pressure provides more ion density but simultaneously reduces the energies of the ions and atoms by scatter. Each sputter process must be optimized for the materials used.

While the target voltage might be several hundred volts, most of the sputtered atoms leave the target with energies less than 5 eV. Since the power density determines the sputter rate, heat dissipation is a key issue. Power densities can range from 5 to 100 W/cm2 for insulating and metal targets. As much as 70% of the power goes into heating the target, however. For this reason, the target material must be cooled efficiently across a heat transferring interface by circulating water through the plenum to which the target is mounted. The power limit before target melting, cracking, or release from its backing plenum depends on the conductivity of the target and the volume of cooling water. Targets of a large variety of materials soldered, brazed, or cemented to backing plates are available from CERAC.

Processes: The End Result:
The surface of the target is sputtered. The composition of the sputtered material and thus the deposited layer is dependent on the ions present. If metal deposition is desired, pure argon gas is used. If it is desired to deposit a compound, such as a metal oxide or nitride, a specific percentage of the reactive gas is introduced with the argon. Ionized reactive gas causes growth of the compound on the target surface. This compound layer is then sputtered and condenses on the substrate. Control of the layer composition and deposition rate is achieved by monitoring and controlling the power to the target and gas mixture in a closed loop to provide stable production capability.

Metals sputter at rates several times that of dielectric compounds. For example, the atom / ion yield for silicon at 1000 V target voltage is 0.6, while that for silicon dioxide is 0.13. Techniques such as D. C. (diode or triode) configurations are best suited for metal sputter deposition. Insulating compounds require R. F. techniques to prevent dielectric building on the target surface to the point where the sputter rate falls to a very small value and / or arcing occurs.

CMN V2 Issue 4 (Oct.-Dec. 1992) contained a partial list of film layer materials that can be reactively sputter deposited from targets of various materials using different gasses.

Layer densities equal to bulk densities can be obtained through the high arrival impact energies present. The growth microstructure can also be controlled so forms from crystalline, amorphous or epitaxial growth can be produced. For example, sputtering is used to manufacture high temperature superconductors using mixed depositions from multiple targets.

What can be done to reduce the greater stress often experienced in sputtered coatings?
Sputtered coatings often grow with high compressive stress. The origins of this stress are inclusion of the work gas (argon) within the film, among other factors. Sputtering offers the ability to control stress to a larger degree than evaporation. Two methods for stress control are: (a) to apply a negative bias to the substrate (easy for a metal; a self bias is generated on a dielectric substrate) or (b) modify the gas pressure or composition. For example, doping the target can improve film uniformity and reduce stress. Zirconium oxide is a case in point. Doping with up to 8 mol% of yttria or silica, disrupts the tendency to form altered crystal states having different refractive indices, stress, and growth patterns, etc. Film durability improvement is a direct benefit for tribological applications; optical path uniformity is important for critical applications. Layer microstructure and thus intrinsic stress can be controlled by manipulating parameters, such as argon gas pressure, substrate temperature, ion energy, etc., according to the Structure Zone Model [ref. 1.] as discussed in CMN V2, Issue 3 (July 1992).

Why is step coverage better?
Step coverage on substrates having complex topographies is possible with higher gas pressures, where scatter spreads the distribution of sputtered species trajectories.

Can you provide more information about what materials can be sputtered and how targets of these special materials are prepared?
The variety of materials from which sputtering targets can be made is nearly limitless. For example, alloys of materials having different evaporation pressures can be sputtered but not evaporated. Targets of single-element materials, such as metals, are generally the pure metal, while mixtures and doped composition targets are made by powder metallurgy. Powder mixtures are hot-pressed under appropriate atmosphere composition and may be sintered. Non-metal targets are made by ceramic technology. Multi-element (or compound) mixtures can be specially made.

Targets are bonded to a copper backing plate that is mounted to the cooling plenum when installed in the sputter system. Indium metal bonding is most frequently used for production sputtering, but special materials require epoxy bonding. Special adhesive layers are plated to the target back surface, and a layer of indium is melted to them. Large area targets are often made in sections and bonded to the backing plate. The sections are butt-joined, leaving a small gap. Besides cost advantages, this approach reduces the risk of target cracking due to heating. It is important that composition and dimensional tolerances be specified when procuring a target.

Deposition of transparent conducting films of ITO and other compounds is routinely done by sputtering. The conductive and transmissive properties of these materials depend strongly on the composition of the deposition atmosphere and starting material. For example, film resistivity for materials such as Ni-Cr, Ni-Cr-Si, etc, is controlled by the percentage oxygen introduced. Sputter deposition offers a sufficiently high degree of control to permit routine annual production of millions of square meters of recording media, solar heating control film, touch panel switches, LCD electrodes, etc.

What are some basic differences between D.C. and other types of sputtering techniques?
D. C. sputtering is practically restricted to materials that are conductive, such as metals, but slightly reduced oxides (ZnO, SnO, ITO, etc) have sufficient conductivity to be sputtered at acceptable rates and without arcing. Arcing and subsequent particulate ejection from the target are usually attributed to dielectric buildup on the target surface. When the layer thickness reaches a certain value, breakdown occurs. As mentioned above, some techniques are designed to avoid this problem by alternating the field on the target (R. F. or A. C.)

Summary:
The basic physics as well as applications and advantages of sputter deposition processes have been discussed briefly. Some specific topics responding to frequently asked questions were covered. We hope this communication will remove some of the mystery from sputtering as a production process and perhaps encourage it's wider application.

Reference:
General background reading on sputter technology can be found in the following references:

1. John A. Thornton and D. W. Hoffman, Thin Solid Films, 171, 5 (1989).
2. L. I. Maissel, Physics of Thin Films, V3
3. W. D. Westwood, Physics of Thin Films, V14
4. J. Vossen and W. Kern, eds., Thin Film Processes, Academic Press, N. Y., 1978.

Samuel Pellicori, Principal Contributor
Pellicori Optical Consulting
P.O. Box 60723 | Santa Barbara, CA 93160
Phone/FAX: 805-682-1922

All printed, graphic and pictorial materials made available on this website are owned by CERAC and protected by Federal Copyright laws. None of the materials, in whole or in part, may be reprinted and distributed or otherwise made available to others for any purpose without CERAC's prior written consent.

Phone:  414-289-9800 /  FAX: 414-289-9805  /   info@cerac.com