Method and Apparatus for Photodeposition of Films on Surfaces

This invention is in the field of laser initiated photodeposition of films on surfaces.

Laser initiated deposition of films can be conveniently categorized as patterned or unpatterned processes wherein a laser beam imparts photon or thermal energy to reactants causing the reactants to dissociate and form a patterned or unpatterned layer on a substrate. Patterned laser depositions may be formed either by "direct writing" or projection patterning. In projection patterning an image is focused through a mask; whereas in "direct writing", a focused laser spot is translated along a substrate to form the desired pattern.

U.S. Pat. No. 4,340,617 to Deutsch et al. shows a "direct writing" laser photodeposition process for the direct generation of patterns of materials, such as metals or semiconductors, on substrates without the intermediate photolithographic steps required in conventional microelectronic fabrication. In the Deutsch et al. system, precursors, or reactants, in the gaseous phase are introduced into a reaction chamber in which a substrate is disposed. A laser beam is focused on, or adjacent to, the surface of the substrate upon which the pattern of material is to be deposited. One or more of the reactants absorbs a portion of the incident energy at a predetermined laser frequency, thereby effecting photodecomposition, or photolysis, of one or more of the reactants. For example, gaseous trimethylaluminum or dimethyl cadmium are dissociated to form aluminum or cadmium metal patterns on a substrate surface using a 257 nm laser beam.

The Deutsch et al. laser-induced deposition process is based on local-area photolytic reactions of gaseous precursors. Hanabusa et al..sup.(1) describe an alternate primarily pyrolytic deposition method. This method involves local heating of the substrate with a focused laser beam.

Although more difficult in practice, substantially photolytic reactions are often preferred since they can be sustained at low temperatures and are insensitive to substrate topography, thermal conductivity, or reflectivity changes. In particular, applications on thermally fragile or transparent substrates or membranes often require such nonthermal methods.

However, because of the general inaccessibility of the dissociative transitions of most stable precursor molecules, photolytic deposition has heretofore been limited to reactions excited in the UV at wavelengths shorter than those of simple commercially available visible light emitting cw lasers. It is highly preferable to use a visible laser for many applications because of the practical simplicity and reliability of visible light emitting lasers. A low power visible light process for depositing opaque, adherent film patterns is particularly desirable for certain applications, such as the repair of photomasks for optical lithography.

Additionally, a need exists for a photolytic deposition process capable of inducing photochemical deposition of ferromagnetic material on substrates using low power visible light. For example, chromium oxide is a ferromagnetic material extensively utilized in the fabrication of Surface Acoustic Wave (SAW) devices. Thin films of this material undergo irreversible transformation to the non-magnetic oxides of chromium during conventional high temperature deposition processing..sup.(2)(3)

In accordance with the invention, a method and apparatus is provided for photodeposition of metal or metal-containing compound films on the surface of substrates located in a reaction chamber using visible light emitting sources as the energy source for initiation of the photodecomposition process. A precursor gaseous phase metal containing oxyhalide compound is introduced into the chamber, typically at a subatmospheric vapor pressure, and usually entrained in an inert gas diluent. Preferably, the precursor/reactant is a gaseous compound, such as CrO.sub.2 Cl.sub.2 or one of its analogues containing an opaque ferromagnetic material, for example, an oxide of chromium. A relatively low power (i.e., 1 mw to 2 w power) beam of visible light is focused at, or near, the surface upon which the metal film is to be deposited. The focused laser beam initiates depositions through a substantially photolytic, non-pyrolytic dissociation of the organic material (C1) from the metal containing oxyhalide. This results in the deposition of the metal-containing material (chromium oxide) on the substrate surface by condensation of the metal material molecules.

While the deposition process, described above, is a non-thermal photolytic process in the initial stages, a second pyrolytic stage may occur as the thickness of the deposited film increases. In this second stage, local heating occurs due to increased light absorption in the deposited film on the substrate. This increased absorption in turn enhances the rate of thermal elimination of undesirable reactant by-products at the substrate surface. This results in a synergistic enhancement of the rate of film growth until the film thickness is much larger than the absorption length of the deposited material.

Alternatively, if desired, this second stage can be avoided by lowering the laser power. The method and apparatus is particularly suitable for repair of photomasks wherein a low cost visible light emitting laser can be employed to deposit opaque (metal) film in a desired pattern on a photomask (substrate).

FIG. 1 is a description of the apparatus used to practice the method of the invention.

FIG. 2 is a plot of the temporal evolution of the yield and transmission characteristics of a 488-nm argon laser beam through a growing chromium oxide film being deposited on a fused silica substrate in accordance with the invention.

FIG. 3 represents a plot of a chromium oxide deposition thickness on Si(100) as a function of incident laser power.

FIG. 4 is a plot of chromium oxide deposition thickness at fixed incident laser power and various chromyl chloride pressures and laser dwell-times.

The method and apparatus of the present invention will now be described in detail in connection with the drawings. In the embodiments which follow the precursor reactant material of choice, is vapor phase chromyl chloride, CrO.sub.2 Cl.sub.2, because the deposited reaction film product is opaque and ferromagnetic chromium oxide, a highly desirable film material for use in a variety of applications. It is not intended, thereby, to limit the scope of applicability of this invention.

The apparatus shown in FIG. 1 is similar to FIG. 1 of the Deutsch et al. U.S. Pat. No. 4,340,617. The vapor phase chromyl chloride CrO.sub.2 Cl.sub.2, contained in gaseous supply assembly 19, is introduced into a reaction chamber 14. Chamber 14 has a stainless steel or quartz body 17 defining an enclosed volume. At least a portion of the body is formed of quartz or other similar material transparent to visible light.

A laser source 10, adapted to generate visible light, produces a beam 11 of energy in the visible light spectrum. The beam of energy is directed through an optical system 12 through a visible light transparent portion 17a of body 14 to a substrate body 16 disposed on jig 15. The gaseous compound 18 of CrO.sub.2 Cl.sub.2 vapors within chamber 14 is photolytically decomposed by photodissociation of the CrO.sub.2 Cl.sub.2 via one or two subsequent single-photon photolysis reactions, according to the following equation:

wherein the Cl.sub.2 molecule is dissociated by a first photon (hv.sub.1) to produce CrO.sub.2 Cl+Cl and another photon (hv.sub.2) dissociates the Cl atom from CrO.sub.2 Cl to yield CrO.sub.2 +Cl+Cl. The CrO.sub.2 condenses on the substrate forming a chromium oxide film and the Cl gas is swept away.

Either the laser 10, the optical system 12, or the cell or chamber 14 can be moved to effect relative movement of the focused laser beam relative to the substrate surface 27. In the illustrated embodiment of FIG. 1, a scanning system has a scanning assembly 28 secured to laser 10 for moving laser 10 in a direction normal to the optical axis of optical system 12. As the laser is moved, the focused position of the laser beam on, or adjacent to the surface 27 of substrate 16 translates and the focused position is therefore scanned across the surface of substrate 16, according to a predetermined pattern. In this manner, a line deposit can be formed or written on the surface of the substrate.

Alternatively, as shown in FIG. 1, the substrate jig 15 may be translated along the beam 11 by means of translator 40 to effectuate relative movement for patterning the substrate.

At least portions of the chamber end walls are selected to be transparent to the laser supplied energy. The substrate body 16 can be selected so that it is substantially, or at least partially, transparent to and passes the energy incident thereon. Under these circumstances, incident laser radiation can pass through chamber 14, except as attenuated, for example, by the gas, the layer deposited on substrate body 16, and the substrate body itself; and the exiting energy can be monitored by a photodetecting element 20. The photodetecting element 20, which can be inside the chamber 14 or outside as illustrated, intercepts the beam 21 exiting the chamber. The photodetecting element 20 has an electrical output signal which is received by an amplifier 22 and the amplifier 22 provides an output to a display device, such as a chart recorder 24. The illustrated recorder 24 provides a visual record 26 of the signal output from the photodetector Thus, as the deposited layer builds up (in thickness), and the intensity of the beam intercepted by the photodetector correspondingly decreases, the record 26 provides an immediate visual record of the physical process.

The method of the invention, as heretofore described, was experimentally practiced on a variety of substrates, both transparent and non-transparent. In the experiments, the photon or laser light source was a line-tuneable 520 to 450 nm cw Ar-ion laser, which, for high-resolution deposition, was focused to a 5-micron-diameter spot on various substrates in the presence of a chromyl chloride vapor. An ArF UV excimer laser was also used to study the reaction mechanism. For transmission and mass-spectrometer studies of the reaction mechanism, a spot size of about 2 mm was used. The laser Ar-ion powers were 1 mW to several watts (i.e., power density of 10.sup.3 to 10.sup.7 watts/cm.sup.2). Much higher powers, exceeding 10.sup.3 watts were used with the ArF laser. Transmission measurements were performed by monitoring the attenuation of the deposition beam by the deposited films with appropriate imaging optics. Mass spectrometer studies employed a cryo-pumped mass analyzer. The chromyl chloride was purified by distillation and used in a flowing gas stainless steel reaction chamber systems. Analysis of the deposited materials was by X-ray photoelectron spectroscopy (XPS) and Auger-spectroscopic techniques.

A chromium oxide deposit was obtained on various substrates held at room-temperature in a variety of experimental conditions, as detailed below.

In Example 1, lines of 5 micron width were drawn on a Si (100) substrate in a direct writing geometry wherein the substrate was scanned in the focus of an Ar-ion laser, the latter impinging at normal incidence to the substrate surface. The laser-deposited films in this example were produced at a scanning speed of 20 microns/s (dwell time 0.25 s) with a thickness of about 5 to 50 nm, depending upon the laser power. The material of the lines was bright and specular, metallic in appearance. XPS studies show these films to be mixed Cr.sub.2 O.sub.3 and CrO.sub.2. Auger studies indicate that the CrO.sub.2 content increases at lower powers. In these experiments, excellent adherence was obtained at Si, GaAs, SiO.sub.2 and glass substrates. These films were strongly attracted by a magnet, unambiguously indicating the presence of CrO.sub.2, the only ferromagnetic oxide of chromium.

Using the same laser source, a 100 micron crystalline structure was obtained on a fused silica substrate, by translating the focusing microscope objective away from the substrate at a rate comparable to the rate of growth. Vertical growth rates of about 3 micron/s and crystal heights exceeding 1 mm were readily achieved. X-ray tests confirmed the monocrystalline nature of the grown structure. Such large single growths of compound crystal structures are only obtained in the upper range of laser powers of the experiments. The most favorable conditions for such growth appear to be at low-pressure (<0.1 Torr)

To investigate the surface reactions, experiments, as shown in FIG. 2, were performed by irradiation of SiO.sub.2 substrates within the mass spectrometer. FIG. 2 is a plot of the temporal evolution of the mass-spectrometric yield and transmission of 488-nm argon laser beam through the growing chromium oxide thin film being deposited on a fused silica substrate. An abrupt change is observed by both measurements at t 45 s, corresponding to a transition from photolytic to pyrolytic mechanisms.

FIG. 2 reveals the existence of two distinct reaction regimes characteristic of intermediate to high laser power, namely an initial photochemical stage, followed by a pyrolytic one. In the first stage, the only detected change due to irradiation is a constant increase in the concentration of chlorine atoms generated in the CrO.sub.2 Cl.sub.2 flow. A slight depression in the concentration of the parent molecule and its principal fragments (generated in the mass spectrometer ionizer) is also seen. In the same period, the transmission of the laser through the film slowly decays with time, indicating near-constant film growth. Following this stage, at about 45 s in FIG. 2, the Cl flux increases rapidly, accompanied by a rise of Cl.sub.2 and O.sub.2 signals as well. Simultaneously, the laser transmission drops rapidly, indicating rapid deposition.

This two-stage behavior can be interpreted as follows. First, photodissociation of gas-phase or adsorbed-phase chromyl chloride molecules takes place, via photolysis, with the resultant CrO.sub.2 molecules condensing on the substrate. As the thickness of the deposited (CrO.sub.2) film increases, local heating due to increasing laser adsorption in the deposited film on the (initially) transparent substrate takes place. Increased adsorption, in turn, exponentially enhances the rate of thermal Cl-, Cl.sub.2 -, and O.sub.2 -elimination surface reactions. A positive feedback mechanism is thus generated between deposited thickness and surface temperature, until the film thickness is much larger than the absorption length of the oxide. From then on, a new constant deposition rate is established, reflecting an equilibrium between decomposition and desorption of adsorbed molecules at the new, elevated temperature. The O.sub.2 -elimination reaction: ##STR1## wherein n is an integer, results in a solid state conversion to the more stable oxide, and takes place at about 400.degree. C..sup.(4), a temperature only slightly above that required for appreciable pyrolysis. Under the experimental geometry on SiO.sub.2 substrates, this temperature should be achieved as soon as the film thickness is large enough for about 20 mW to be absorbed by it. This is consistent with the observation that only the initial linear photochemical phase occurs, and that no transition to the thermal phase is observed in transmission or mass spectrometer measurements at powers below about 30 mW. When the fused silica is replaced with an equally transparent sapphire substrate, the photochemical phase was observed as before, but the transition to the thermal stage (at equal laser power) occurs at about one decade larger film thicknesses or, at approximately equal thickness, but one decade higher laser power. This result is readily explained by the difference in thermal conductivity of the two substrates (about 0.34 W/cm.degree.K. for sapphire compared to 0.014 W/cm.degree.K. for SiO.sub.2). Since the onset of heating is rapid, it is not possible to distinguish temporally the turn-on of the solid-phase conversion reaction from the turn-on of pyrolysis of gas phase molecules.

In order to gain insight into the roles of gas-phase and adsorbed-phase reactions in the photochemical stage of chromium oxide deposition, a series of experiments were performed with Si substrates. Although Si absorbs in the visible, its high thermal conductivity, 1.5 W/cm.degree.K., precludes the onset of the pyrolytic stage under our experimental conditions (7 micron spot diameter, P.congruent.0.2 W). FIG. 3 represents the chromium oxide deposition thickness on Si(100) measured with a profilometer, as a function of incident laser power. A laser wavelength of 488 nm, chromyl chloride presence of 0.1 Torr and dwell-time of 0.35 s was employed in these experiments. The initial linear rise, followed by a rapidly reduced deposition rate (thickness), result from the two competing processes: photodecomposition of chromyl chloride, whose rate is linear with laser power; and thermally induced desorption of adsorbed molecules, at a rate which exhibits Arrhenius behavior with the laser-induced surface temperature rise. Applying a simple rate equation analysis to the data in FIG. 3, a binding energy E.sub.b =0.25 eV was calculated. This value is very close to the heat of liquification of chromyl chloride, E.sub.L =0.21 eV, as determined from separate vapor-pressure measurements.

While FIG. 3 demonstrates the importance of desorption, the data in FIG. 4 clarify the important role that adsorption plays in initiating the nucleation of, and promoting the growth of, the chromium oxide film. In FIG. 4, the film thickness on silicon substrates was measured with a profilometer at fixed incident power (P=18 mW), at various chromyl chloride pressures and laser dwelltimes. In FIG. 4, chromium oxide deposition rate (full triangles, left hand scale) and induction time (full circles, right hand scale) as a function of chromyl chloride pressure are plotted. The data were determined from thickness vs. t.sup.1/2 curves, where the dwell-time, t, was in the range of 0.35-15 s. The laser power was 18 mW at 488 nm. The open triangles (broken line) represent the adsorption isotherm of chromyl chloride, scaled so that at high pressure, the isotherm and deposition rate coincide.

At any given pressure, no deposition takes place during an induction time, t.sub.o, after which the thickness grows as the square root of time. The growth rate and t.sub.o are plotted in FIG. 4 as a function of chromyl chloride pressure. This rate increases rapidly with pressure up to about 0.1 Torr and then exhibits saturation-like behavior. For comparison, the experimentally determined adsorption isotherm of chromyl chloride, measured with a quartz-crystal microbalance is also plotted in FIG. 4. The near-perfect coincidence between the deposition rate and the chromyl chloride isotherm strongly suggests that the adsorption dynamics are controlling the rate of the deposition process. The t.sup.1/2 behavior of the growth rate is consistent with a deposition rate limited by mass-transport effects into the reaction zone. The induction time in FIG. 4 may reflect the time necessary for starting the film nucleation process, and appears to be inversely related to the chromyl chloride adsorbed-layer coverage.

Kubota's work.sup.(4) indicates that five different chromium oxide phases exist under the ambient conditions set forth above, each in their own discrete temperature bands, with little overlap between phases, over the range of 200.degree.-400.degree. C. Above this range, CrO.sub.2 and the other oxides are irreversibly converted to the most stable form, Cr.sub.2 O.sub.3. Although one may readily distinguish the photochemical from the thermal aspects of the CrO.sub.2 Cl.sub.2 decomposition to chromium oxides, it appears that the production of a pure oxide of a particular stoichiometry requires either a purely photochemical reaction, or one whose thermal nature was tightly controlled. Another possibility is laser-initiation of more complex chemical reactions. We have explored the generation of a reducing ambient in order to controllably increase the Cr:O ratio. In initial studies of excimer-laser (193-nm) deposition of chromium oxide films on fused silica, the Cr:O ratio was increased from 2:3 with 0.1 Torr CrO.sub.2 Cl.sub.2 to 6:5 in the presence of 200 Torr HCl.

The visible-laser direct write deposition of metal containing oxides, in accordance with the invention, has been developed by reactions of both the photolytic and pyrolytic type. On transparent substrates, a sharp transition from the former to the latter can be observed. These reactions are notable in that they deposit stable films with useful magnetic and/or opacity properties using visible light on both absorbing and transparent substrates.

The method and apparatus is useful for repair of defective photomasks wherein it is desirable to fill in voids in the phototmask pattern with patterned opaque lines. In this case, the substrate body 16 would comprise the photomask and patterns of opaque chromium oxide or other suitable opaque metal material would be formed by translating the light beam 11 in the presence of the CrO.sub.2 Cl.sub.2 or other suitable vapor.

While the principles of the invention have been described in connection with the metal-containing oxyhalide CrO.sub.2 Cl.sub.2, the following equivalent precursor/reactant materials are contemplated as visible light absorbing prospects: CrO.sub.2 F.sub.2, CrO.sub.2 (NO.sub.3).sub.2, CrO.sub.2 Br.sub.2, CrO.sub.2 I.sub.2 and Co(CO).sub.3 NO. Analogues of the above with similar electronic configurations are MoO.sub.2 F.sub.2, MoO.sub.2 Cl.sub.2, WO.sub.2 F.sub.2, WO.sub.2 Cl.sub.2, MoOF.sub.4, MoOCl.sub.4, WOF.sub.4 and WOCl.sub.4. Oxyhalides of closely related transition elements include TaO.sub.n F.sub.m, NbO.sub.n F.sub.m, VO.sub.n F.sub.m, TaO.sub.n Cl.sub.m, NaO.sub.n Cl.sub.m and VO.sub.n Cl.sub.m ; where n and m are integers.

Equivalents

While the above described embodiments of the invention are preferred, other configuration will be apparent to those skilled in the art and thus the invention is only to be limited in scope by the language of the following claims and equivalents.

1. Hanabusa et al., Appl. Phys. Lett., 35 (8) pp 626-627, October 1979.

2. T. J. Swoboda, P. Arthur, N. L. Cox, J. N. Ingraham, A. L. Oppegard and M. S. Sadler, J. Appl. Phys. 32, 374 (1961).

3. L. Ben-Dor and Y. Shimony, J. Crystal Growth 24/25, 175 (1974). S. Ibashi, T. Namikawa and M. Satou, Magnetism and Magnetic Materials 1976, Am. Institute of Physics Conf. Proc. No. 34, (J. U. Becker and G. A. Lander, eds.) p. 43.

4. B. Kubota, J. Am. Ceram. Soc. 44, 239 (1961).