Application-Driven PVD Research

Much of the PVD research at IPML has been motivated by industry's need to improve its material synthesis abilities either through the development of new processing technology or through the improvement of its current technologies based upon accurate, insightful computer simulations of its processing systems. The most important application-driven PVD research conducted by IPML researchers to date is described below:

Thermal Barrier Coatings

Giant Magneto Resistance Materials

Continuous Fiber Reinforced Metal Matrix Composites

Polysilicon deposition for Thin Film Transistors

Thermal Barrier Coatings (TBCs)

Thermal barrier coatings (TBC’s) are used for the thermal protection of turbine engine components to increase the durability and efficiency of gas turbine engines. These two-layer coatings are composed of a porous, insulating yttria stabilized zirconia (YSZ) top layer and an underlying aluminide (nickel or platinum) bond layer. Here, an emerging electron beam-directed vapor deposition (EB-DVD) approach is explored as a method for producing TBC’s. The EB-DVD technique combines low vacuum (10-3 - 10 Torr) electron beam evaporation with a carrier gas jet to rapidly create and efficiently transport evaporant to a deposition surface. This approach has several attributes which makes it a potentially useful technique for the deposition of TBC layers. These include its high deposition efficiency, its ability to reactively form ceramic layers at high rate, its use a carrier gas jet to facilitate vapor phase mixing when depositing materials with large vapor pressure differences from multiple sources and its relatively low equipment costs. Studies of the deposition of yttria partially stabilized zirconia (YSZ) and reactively formed zirconia have been performed. The microstructure, phase content, pore morphology and texture of deposited layers have been determined. Results indicate an ability to form YSZ and zirconia coatings which have the columnar microstructure necessary for high performance gas turbine applications and that the operating deposition conditions have a strong effect on the column diameter, porosity, pore morphology and growth orientation of the layers. A further understanding of the chamber operating conditions, however, is necessary to determine the critical deposition conditions for depositing useful TBC layers using EB-DVD. For this purpose, vapor transport modeling using a Direct Simulation Monte Carlo technique will be used to investigate the carrier gas flow conditions, adatom deposition energy, adatom angle of incidence and the deposition efficiency. In addition, the thermal properties of the resulting TBC layers can be expected to be effected by observed microstructural changes and therefore the thermal conductivity of the deposited films will be measured.

Giant Magneto Resistance (GMR) Materials

Continuous Fiber Reinforced (CFR) Metal Matrix Composites (MMCs)

The primary motivating material system behind development of the University of Virginia's Directed Vapor Deposition (DVD) technology has been the coating of continuous fibers to be used as reinforcement in metal matrix composites. High temperature light weight load bearing aerospace structures such as compressor fan blades for the front section of aircraft engines represent one of the primary areas of application for these difficult to synthesize materials. While companies such as 3M and GE Aircraft Engines have employed techniques such as conventional e-beam evaporation and plasma spraying to create CFR MMCs, each of these methods has drawbacks. Conventional e-beam evaporation suffers from low materials utilization efficiency, with significant quantities of the expensive metal matrix material ending up on the walls of the processing chamber instead of the fibers. To capture as much vapor as possible during its conventional e-beam processing, 3M uses intricate fiber manipulation systems. These fiber manipulation systems frequently break fibers, forcing the coating run to be terminated. Plasma spray technology creates a rough surfaced, fiber containing plasma spray foil. During the hot isostatic pressing cycle which presses these layers together, significant fiber fracture often degrades the properties of the resulting material.

To explore the ability of DVD to create smoothly coated fibers efficiently, an array of 142 mm diameter SCS-6 fibers (supplied by Textron Specialty Materials, Lowell, MA) was set up in a square aluminum frame with an inside edge dimension of 5.08 cm. 25 colinear fibers were mounted vertically with a nominal center-to-center fiber spacing of 2 mm. During each run, the fiber frame itself was covered with aluminum foil to ensure that metal deposition on it was not included in the weight change measurements. For the fiber spacing used, 7.1% of the cross-sectional area of the frame interior was occupied by fibers. Thus, if a uniform flux were incident upon the fibers and deposition occurred only on surfaces that were insight of the flux, the average deposition efficiency (i.e. the efficiency achieved in a conventional e-beam system) would be 7.1%.

The observed chamber pressure/Mach number/deposition efficiency trends were similar to those recorded for flat substrate coating. The first figure below plots the relationship between Mach number, chamber pressure, and relative fiber coating efficiency (compared to that achieved by a line-of-sight deposition process). It should be noted that the deposition efficiency measured in this study is an average value over the entire surface area of the fibers. As a result, maximum deposition rates reached 50 - 100 mm per minute in certain portions of the fiber array while being limited to 1 - 2 mm per minute in other regions.

For many of the experimental conditions investigated, deposition efficiencies were well above line-of-sight levels (i.e. the area fraction occupied by fibers).

As with flat substrate coating, low chamber pressures (low carrier gas fluxes) did not entrain the vapor and transport it towards the fiber array, resulting in low deposition efficiencies. Instead, the main vapor stream crossed the primary carrier gas flow and passed between the e-beam gun and the top of the fiber array. Once the carrier gas flow was increased so that the vapor passed through the frame, deposition efficiencies increased dramatically. Increasing carrier gas flows led to a peak in deposition efficiency (at about 70 Pa) followed by a significant decrease as chamber pressure was raised. This trend suggests that the presence of the observed microwall jet (See the next figure below.) results in vapor atom redirection around the fibers. At pressures near 70 Pa this redirection led to coating around the entire fiber circumference. At higher pressures the microwall jet prevented vapor atoms from reaching the fiber surface, leading to decreased deposition efficiency in much the same manner as the larger wall shock and wall jet deflected vapor away from flat substrates.

At the point of intersection between the vapor-laden carrier gas stream and the array of fibers, a wall shock is observed, demonstrating how the presence of the fibers affects the flowfield.

The fiber coating experiments revealed a peak deposition efficiency over twice the 7.1% expected for line-of-sight deposition. Scanning electron microscopy (See the following figure.) of these samples revealed that, for conditions of maximum efficiency, vapor deposited not only on the surface of the fiber facing the incoming vapor but also on the fiber's sides and its back. The backside fiber coating phenomenon appeared to be a manifestation of atomic scattering from the flow as it passes the fiber, a phenomenon discussed previously by Hill. A demonstration program created by G.A. Bird helps illustrate how backside fiber coating occurs.

Vapor deposition was evident upon the sides (~15 mm thick) and back (~5 mm thick) of this fiber held stationary during the deposition process (M = 1.50, chamber pressure = 66.7 Pa).

Based upon the fiber coating results presented above, DVD appears to have some promise as a fiber coating system since under certain process conditions the system presents a narrow region through which rotated fibers can be passed for uniform, high rate (50 - 100 mm/min.), high efficiency vapor deposition. These results suggest that it may be possible to accomplish industrially acceptable high rate fiber coating using a fairly small, low-cost e-beam gun system, perhaps even a system the size of the experimental setup developed at the University of Virginia.

To assess more fully the ability of DVD to coat fibers in an industrial setting, research should be conducted which will demonstrate the technology's ability to deposit the correct alloy compositions desired for the metal matrix composite fiber application (e.g., Ti-6Al-4V). Single crucible evaporation of this alloy system using DVD will almost certainly experience the same stoichiometry problems observed with conventional e-beam systems. However, use of two in-line vapor targets and the gas jet could quite conceivably overcome this troublesome problem by mixing the two vapor plumes for efficient, compositionally correct deposition (See the next figure below.). Incorporation of a high rate scanning system into the current DVD configuration would allow the beam power to be precisely split between the two targets for stoichiometrically-correct evaporation. Additional research will need to be conducted to investigate the deposition rate around fibers for different elements. It should be noted that the non-line-of-sight coating efficiency of low vacuum systems is a function of the mass of the chamber gas and the vapor atoms. No data currently exists which explains how to obtain a stoichiometrically-correct, industrially-desirable Ti-6-4 deposit in a DVD system. As a result DVD research at the University of Virginia will probably need to demonstrate that the technology can create a large quantity of stoichiometrically-correct clean coating on fibers during a single processing run before industry will seriously consider the technology a viable alternative. To reach this level of processing reliability, the lifetime of the e-beam gun's filament must be extended from the current ten hour limit. In an industrial setting, a tungsten filament lifetime of 100 - 250 hours will almost certainly be required.

Stoichiometrically-correct deposition using this DVD system configuration could be much faster and more efficient than that achieved in a conventional multicrucible e-beam evaporation system. Only mixed material in region (AB) can create a compositionally correct alloy.

Polycrystalline silicon (polysilicon) for Thin Film Transistors (TFTs)

The ability of DVD technology to deposit useful amorphous and polycrystalline silicon for thin film transistor (TFT) applications has been investigated. Silicon was evaporated and deposited upon glass substrates at different substrate temperatures. The morphological and crystalline characteristics of the deposited films were then studied via scanning electron microscopy and optical transmission and reflection techniques. For these experiments, a 1.27 cm diameter phosphorous-doped (0.1 W-cm) single-crystal silicon rod contained within a water-cooled crucible was used as the evaporation target, and the carrier gas jet consisting of helium with a 5% argon content. For each 10 minute deposition run, a gas jet pressure ratio (mixing chamber/processing chamber) of ~22 was generated using a 0.85 cm diameter nozzle and a chamber pressure of 0.14 Torr (~20 Pa). The selection of the substrate temperature range for the experiments considered the following:

Substrate temperatures used in other low temperature polycrystalline silicon deposition processes (400 - 580°C),
Softening temperature of the (Corning 7059) glass substrates used (593°C), and
Minimum temperature required to reduce argon incorporation into the film (~300°C).

Substrate temperatures were generally held at 325°C or higher to reduce the argon content in the films (4) while the maximum substrate temperature examined (550°C) was less than the critical temperature of the glass substrates. For the various d eposition runs, the e-beam power was set between 300 and 600 W. Prior to silicon deposition, an inert gas flow was used to remove any particulates from the substrate surface.

For these experiments the deposition rate varied between 150 and 300 nm/min except for the baseline 70^oC run which had a slightly higher deposition rate due to its elevated e-beam power. Variation of deposition rate resulted in changes in the thickness of the deposi ted films. The variation in film thickness (total thickness ~1.5-3.0 mm except for the slightly thicker 70^oC experiment) was assumed to have a minimal effect on film microstructure. The results of this silicon study a re summarized in the table below . The amorphous or polycrystalline nature of the films was determined via examination of the films' optical transmission spectra. These optical transmission measurements were made in the 0.6 - 2.4 mm wavelength range using a Cary 5E two-beam spectrophotometer. The optical spectra analysis indicated that the 515^oC and 550^oC films were polycrystalline was reinforced by the visible change in film color, from grey for the lower temperature amorphous films to silver for the polycrystalline films.

DVD deposition of silicon

Substrate Temperature (^oC) E-beam Power (W) Resulting Microstructure

701 600 Amorphous. Generally smooth morphology, repeated domed regions ~0.35 mm in diameter, no clear separation between adjoining domes. Film cracking evident. Optical appearance = grey.

325 480 Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.

425 360 Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.

475 420 Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.

515 420 Polycrystalline. Larger diameter columns ~0.33-0.50 mm in diameter, less obvious separation, continued rough surface morphology. Optical appearance = silver.

550 300 Polycrystalline. Larger diameter columns ~0.33-0.50 mm in diameter, less obvious separation, continued rough surface morphology. Optical appearance = silver.

¹ Condensation of the silicon vapor increased the substrate temperature from room temperature to 70^oC.

² Deposition rate for this experiment exceeded the deposition rate used in other experiments.

DVD deposition of silicon
Substrate Temperature (^oC)	E-beam Power (W)	Resulting Microstructure
701	600	Amorphous. Generally smooth morphology, repeated domed regions ~0.35 mm in diameter, no clear separation between adjoining domes. Film cracking evident. Optical appearance = grey.
325	480	Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.
425	360	Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.
475	420	Amorphous. Distinct columns ~0.25 mm in diameter, clear void spaces between columns, rough surface morphology. Optical appearance = grey.
515	420	Polycrystalline. Larger diameter columns ~0.33-0.50 mm in diameter, less obvious separation, continued rough surface morphology. Optical appearance = silver.
550	300	Polycrystalline. Larger diameter columns ~0.33-0.50 mm in diameter, less obvious separation, continued rough surface morphology. Optical appearance = silver.
¹ Condensation of the silicon vapor increased the substrate temperature from room temperature to 70^oC.
² Deposition rate for this experiment exceeded the deposition rate used in other experiments.

The figures below show scanning electron micrographs of silicon films deposited at 475^oC and 515^oC respectively. The amorphous film shown in the first figure has small, tightly-packed columns approximately 0.25 mm in diameter which give rise to a rough surface morphology. In contrast, the polycrystalline silicon film deposited at 515°C shown in the second figure has larger, loosely-packed columns approximately 0.33-0.50 mm in diameter which give rise to a rough, porous morphology. It is interesting to note that the transition from amorphous to polycrystalline microstructure has been reported to occur at temperatures around 580°C for traditional low pressure chemical vapo r deposition processes, in contrast to the transition observed here in the range of 475 - 515^oC .

A general examination of all of the films revealed two major transitions in the film microstructure. The first occurred between 70°C and 325°C. While the films at these two temperature are both amorphous, examination of the top surfaces of the f ilms revealed a morphology change, from a smooth continuous surface at 70°C to a rough surface consisting of tightly packed columns at 325°C. The second morphology transition occurred between 475°C and 515°C. In this instance, the chan ge in morphology from distinct columns with clear intercolumnar porosity to larger diameter columns with less obvious separation was accompanied by a change from amorphous to polycrystalline silicon.

While these experiments demonstrate that DVD technology can produce polycrystalline silicon films in a single step, low temperature process that is compatible with glass substrates, the morphology of the current films (i.e., not a smooth polycrystalline f ilm) precludes their use in thin film transistor applications and further work must now identify the process conditions which produce smooth polycrystalline films needed for many electronic material applications.


Porous structure consisting of amorphous columns separated by voids. (T = 475^oC, Beam power = 420 W)	Porous structure composed of polycrystalline columns separated by voids. (T = 515^oC, Beam power = 420 W)