Method of depositing films on aluminum alloys and films made by the method

Abstract

Method for depositing a metallic material on an aluminum alloy surface for galvanic displacement type deposition, electrodeposition or electroless deposition of a metallic film on the surface wherein the alloy surface is oxidized (e.g. anodized) to form aluminum oxide and the oxidized surface is etched to leave a partial thickness of a barier aluminum oxide on the alloy surface. The partial thickness of the barrier oxide is controlled by etching to form a porous, metallic particulate film for a thin barrier oxide, or a continuous metallic film for thicker barrier oxide. The metallic film then is electrodeposited or electroless deposited on the barrier film.

Description

This application is a continuation-in-part of U.S. Ser. No.11/201,766 filed Aug. 11, 2005, and claims benefits and priority of provisional application Serial No. 60/663,659 filed Mar. 21, 2005.

FIELD OF THE INVENTION

The invention relates to galvanic displacement type deposition, electroless deposition or electrodeposition of metallic films on treated aluminum alloys as well as to the deposited metallic films and components, such as capacitor electrodes, embodying the metallic films.

BACKGROUND OF THE INVENTION

The surface of aluminum metal is spontaneously oxidized in the ambient atmosphere. This oxidation creates a dielectric film of native aluminum oxide, which has an adverse effect on electrodeposition or electroless deposition of metals or alloys such as Ni, Ag, Au, and Cu and their alloys.

With respect to overcoming the problem of electrodeposition, the zincate process has been employed in industry for the deposition of adhesive metallic films on aluminum. The process consists of immersing the aluminum substrate in a strong alkaline zincate solution. The native aluminum oxide is dissolved, and zinc is deposited on the surface via galvanic displacement of aluminum. As a result, the zinc-coated aluminum surface becomes amenable for electrodeposition of adhesive layers of metals, including nickel and copper. Zincate surface activation of aluminum has proven to be a cost-effective process for nickel bumping of wafers prior to flip-chip assembly.

Since the zincate method is sensitive to many variables, there are incentives for developing alternative methods for the deposition of metals on aluminum. One alternative method has involved direct electrodeposition of copper on aluminum using several copper complexes. An electroplating procedure for nickel displacement of aluminum followed by electroless nickel deposition has also discussed. In addition, an organic solvent has been used to lay a seed layer of copper or palladium on aluminum substrates. Then, electroless deposition with a reducing agent was utilized to deposit substantially more copper.

In general, the galvanic displacement type deposition process proceeds via two concurrent electrochemical reactions, which involve the reduction of ions of metals and the oxidation of the substrate surface. The driving force for this process is determined by a difference in half-cell potentials (e.g. redox potentials for corresponding metal/metal ion and oxidized substrate/substrate pairs). The half-cell potential of the reduced species has to be more positive than that of the oxidized substrate. Chemical etching, which effectively removes the surface layer of oxide, precedes and/or takes place simultaneously with the deposition of a film of metal.

The present invention involves deposition of metallic films (either porous or continuous) on an aluminum alloy surface to offer the opportunity for fabrication of heat dissipation systems, energy conversion and storage devices. For example, double layer capacitors are energy storage devices that store electrical energy by sustaining an electrical charge in a thin double layer at the interface between an ionically conducting electrolyte and an electronically conducting electrode. Potential applications for double layer capacitors include memory protection in electronic circuitry, portable electronic, and communication devices. Double layer capacitors can be built as either a self-standing device or a part of integrated electronic system. Mesoporous carbon, carbon nanotubes and other carbonaceous materials have been extensively investigated for use in double layer capacitors because of their very high specific surface areas. In contrast to porous metallic films, limitations of double layer capacitors based upon carbonaceous materials are two-fold. First, capacitance of these materials typically degrades at frequencies higher than 10 Hz. The second limitation arises from problematic incorporation of these capacitors into technologically relevant materials such as silicon.

In addition, the present invention involves deposition of metallic films on an aluminum alloy surface to offer the opportunity for fabrication of optical devices for surface enhanced FT-IR spectroscopy, surface enhanced Raman scattering and metal-enhanced fluorescence. In addition, composite materials with noble metal particles may have useful photo-catalytic, anti-microbial properties and tunable surface plasmon resonances. In addition, the deposition of continuous metallic films on aluminum alloys may be used for metallization of aluminum, for providing a soldering surface on aluminum and, consequently, for packaging of electronic devices (zincate-free nickel bumping of wafers prior to flip-chip assembly).

SUMMARY OF THE INVENTION

One embodiment of the present invention provides a method of depositing a metallic material on a substrate wherein the method includes the steps of providing a substrate comprising an alloy of aluminum and an alloying element, oxidizing a surface of the alloy substrate to form aluminum oxide thereon, etching the oxidized surface to leave a partial thickness of a barrier aluminum oxide on the alloy surface, and depositing by galvanic displacement type deposition, electroless deposition or electrodeposition discrete metallic nanoparticles having a particle density of about 10⁴ to about 10¹² particles/cm² on the barrier oxide.

In an illustrative embodiment of the invention, the etching step is conducted to leave a partial thickness of the barrier oxide to increase nucleation of discrete metallic nanoparticles thereon. This embodiment of the present invention provides an aluminum alloy substrate having a partial thickness of the barrier oxide on the surface and a porous, three dimensional film structure of electrically interconnected, discrete metallic metal nanoparticles deposited by galvanic or electroless deposition or electrodeposition on the barrier oxide. In this case, the nucleation density is high enough so that the neighboring metallic particles form the electrical connections to each other. The film structure includes randomly packed, generally spherical, metallic nanoparticles having a distribution of particle sizes. An electrode, such as a double layer capacitor electrode, can comprise the alloy substrate having the barrier oxide and porous and metallic film structure thereon.

Still another embodiment of the present invention provides a method of depositing a metallic material on a substrate wherein the method includes the steps of providing a substrate comprising an alloy of aluminum and an alloying element, oxidizing a surface of the substrate to form aluminum oxide thereon, etching the oxidized surface to leave a partial thickness of the barrier oxide on the alloy surface, and depositing by electroless deposition or electrodeposition a continuous metallic film on the barrier oxide. As an illustration, the metallic film is composed of Ni. In an illustrative embodiment, the etching step is conducted to almost completely remove the barrier oxide in order to deposit a continuous metallic film.

Features and advantages of the invention will become more readily apparent from the following detailed description taken with the following drawings.

DESCRIPTION OF THE DRAWINGS

FIG. 1A is a SEM (scanning electron micrograph) collected after the electroless deposition of silver particles on anodized and chemically etched aluminum-copper alloy film for 24 hours. FIG. 1B is a high magnification SEM of the deposit cross-section.

FIG. 2A are cyclic voltammograms obtained for silver particles/aluminum-copper alloy film electrode, and FIG. 2B shows charging current dependence upon the scan rate. While the solid line shows the fit over the whole range of scan rates, two dotted lines show the fit over two regions of slow and fast scan rates.

FIG. 3A is a Bode plot (data as symbols, modeling as lines) of EIS results, FIG. 3B is a Nyquist plot. FIG. 3C shows the equivalent circuit.

FIG. 4 shows polarization curves of the anodized and etched Al-Cu alloy film electrode where curve (a) is before and curve (b) is after addition of AgNO₃ to the final concentration of 3.0 mM.

FIG. 5 is a SEM micrograph collected after galvanic displacement type deposition of silver for 40 min.

FIG. 6 is a SEM micrograph collected after electrodeposition of silver at—1.3 V for 40 min.

FIG. 7 is a Bode plot of EIS data collected at OCP after galvanic displacement type deposition of silver by galvanic displacement for 40 min. Experimental data are symbols and results of modeling are solid lines. FIG. 7A shows the equivalent circuit is shown as an inset.

FIG. 8 is a Bode plot of EIS data collected at OCP after electrodeposition of silver at—1.3 V for 40 min. Experimental data are symbols and results of modeling are solid lines.

FIG. 9 shown the equivalent circuit used for modeling of the electrodeposited porous film of silver (FIG. 8).

FIGS. 10A-10F are diagrams of electrodeposition on patterned Al-Cu alloy film substrates where FIG. 10A shows photolithographic patterning; FIG. 10B shows anodization at 90 V for 2-3 min to form a dense layer of Al₂O₃; FIG. 10C shows photoresist removal to expose the underlying Al-Cu alloy layer or film; FIG. 10D shows anodization at 50 V for 20 min to form porous Al₂O₃; FIG. 10E shows chemical etching of porous A1₂0₃; and FIG. 10F shows electrodeposition of silver.

FIG. 11A is a SEM micrograph of patterned features, showing four circular regions of electrodeposited silver and FIG. 11B is a SEM micrograph of a patterned feature with electrodeposited silver.

DETAILED DESCRIPTION THE INVENTON

The invention provides a method for the galvanic displacement type deposition, electroless deposition or electrodeposition of a metallic film on a treated surface of an alloy of aluminum and an alloying element. The aluminum alloy can comprise an alloy of aluminum and one or more alloying elements to provide a binary, ternary, quaternary, etc. aluminum alloy. For purposes illustration and not limitation, the alloying element can include, but is not limited to, one or more of Au, Cu, Cr, Mn, Mo, Ni, Si, Ta, Ti, or Zn, or combinations thereof.

The invention can be practiced to deposit a variety of metallic layers or films on the aluminum alloy surface. For purposes of illustrating and not limiting the invention, the invention is useful to deposit a metallic layer or film comprising Au, Ag, Pd, Cu, Ni, Pb, Cr, Fe, W, Mo, or Co wherein the term metallic film includes a layer or film comprising a metal or an alloy of these metals one with another or with another different metal, or mixture of two or more metals, deposited concurrently or sequentially to provide a metallic film on the surface whereby the deposited metallic film comprises a binary alloy deposit, ternary alloy deposit, quaternary alloy deposit and so on. The metallic film can have a thickness in the range of 1 nm to 10 microns for purposes of illustration and not limitation; however, practice of the invention is not limited to any particular thickness of the metallic film since any suitable metallic film thickness can be deposited.

The method envisions providing an aluminum alloy surface that is treated in a manner effective to render the alloy surface amenable to galvanic displacement type deposition, electrodeposition or electroless deposition of a metallic film thereon. Galvanic deposition and electroless deposition both can occur with no external electrical power requirement such that galvanic deposition is considered by some to be a form of electroless deposition. Galvanic deposition generally is a deposition process in which the substrate comprises a less noble element that acts as a reducing agent for a metal cation dissolved in the deposition solution to effect deposition of the metal. Electroless deposition involves providing a reducing agent in the deposition solution containing the metal cation to be deposited to effect its deposition on the substrate, which may not be less noble than the metal to be deposited. The alloy surface to be treated pursuant to the invention can include, but is not limited to, any type of aluminum alloy substrate, layer, film, or other surface on which the metallic film comprising a metal or alloy is to be deposited by electrodeposition or electroless deposition.

The method of the invention involves treating the aluminum alloy surface by oxidizing the surface to form aluminum oxide thereon wherein the aluminum oxide comprises an outer porous aluminum oxide and an inner barrier aluminum oxide adjacent the alloy surface. The anodization results in enrichment of the alloying element present in the aluminum alloy, such as for example Cu, under and/or in the barrier aluminum oxide.

The invention can be practiced using anodizing to oxidize the surface to form aluminum oxide thereon. However, practice of the invention is not limited to any particular anodizing process. For example, the anodizing process can vary with particular type of surface to be treated. Any conventional anodizing process can be used with the type of electrolyte and parameters of anodizing, such as anodization voltage, electrical current density, temperature and electrolyte acidity being selected as desired. For example, the anodizing process can be conducted in any conventional aqueous electrolyte that includes, but is not limited to, solutions of oxalic acid, sulfuric acid, phosphoric acid, chromic acid, and mixtures of two or more of these acids. The invention also can be practiced using other oxidizing processes to form aluminum oxide on the surface. For purposes of illustration and not limitation, alternative oxidizing treatments to anodization include polishing, alkaline etching, acid pickling, electropolishing, heating up to 700° C. in an oxygen bearing atmosphere such as air, and any other treatment, which results in oxidation of the aluminum alloy surface and formation of aluminum oxide on the surface.

The aluminum oxide then is etched in a manner to remove the porous outer aluminum oxide and partially remove the barrier aluminum oxide to leave a portion of its original thickness on the alloy surface. The etching step is conducted for a time in a selected etchant to leave a controlled partial thickness of the barrier aluminum oxide remaining on the alloy surface in dependence upon the type of metallic film to be subsequently deposited.

For example, in an illustrative embodiment of the invention, the etching step is conducted to leave a partial thickness of the barrier aluminum oxide that is effective to enhance nucleation of discrete metallic nanoparticles on the barrier oxide during subsequent electroless or electrodeposition so as to form a porous, metallic particulate film. For purposes of illustration and not limitation, in this embodiment, the thickness of the partial thickness of the barrier oxide remaining on the alloy surface can be about 50 nm or less depending on deposition parameters, such as overpotential for electrodeposition, and presence of additives in the solution.

The resulting porous, metallic particulate film deposited on the Al alloy substrate comprises the partial thickness of the barrier oxide on the alloy surface and a porous, three dimensional film structure of electrically interconnected metallic metal nanoparticles deposited by galvanic or electroless deposition or electrodeposition on the barrier oxide. The film structure includes randomly packed, generally spherical metallic nanoparticles having a distribution of particle sizes and high particle density in the range of about 104 to about 10¹² particles/cm². For example, the metallic nanoparticles can have a particle diameter in the range of about 20 nm to about 1000 nm. An electrode, such as a double layer capacitor electrode, can embody such alloy substrate having the partial thickness of the barrier oxide and this film structure thereon.

In another illustrative embodiment of the invention, the etching step is conducted to leave a partial thickness of the barrier aluminum oxide that is thin enough to yield a continuous (non-porous) metallic film on the barrier oxide during subsequent electroless or electrodeposition. For purposes of illustration and not limitation, in this embodiment, the thickness of the partial thickness of the barrier oxide remaining on the alloy surface can be about 5 nm or less depending on deposition parameters, such as overpotential for electrodeposition, and presence of additives in the solution.

Practice of the invention is not limited to any particular etching process. For example, the etching process can vary with the particular type of aluminum alloy surface to be treated. Any conventional etching process can be used with the type of etchant and time of etching being selected empirically to achieve a desired etched barrier aluminum oxide of controlled partial thickness described above depending upon the metallic film to be deposited. For example, the etching process can be conducted in any conventional acid etchant that includes, but is not limited to, an acidic aqueous solution (phosphoric acid, oxalic acid, sulfuric acid) or a mixture of an acid and an inhibitor of aluminum oxidation, such as chromic acid. Other inhibitors can be used as an alternative to chromic acid. Etching also can be performed in an alkaline solution of sodium hydroxide, or any other hydroxide.

Although the Examples set forth below involve anodizing using an aqueous oxalic acid solution and certain anodizing parameters and acid etching using an aqueous solution of phosphoric acid and chromic acid, these are offered merely for purposes of illustrating and not limiting the invention. Similarly, although the Examples are described with respect to a surface of a thin film or layer of an alloy of Al and Cu where Al and Cu are present in respective amounts of 99.5 weight % and 0.5 weight % of the alloy, the Examples are offered merely for purposes of illustrating and not limiting the invention.

Example 1 describes a method pursuant to an illustrative embodiment of the invention wherein an aluminum-copper alloy is anodized and then chemically etched followed by electroless deposition of silver on the treated alloy surface.

EXAMPLE 1

Galvanic Displacement Type Deposition

In particular, aluminum-copper alloy film covered wafers used in this Example were fabricated as follows: First, a 600-nm layer of SiO₂ was thermally grown by steam oxidation of each silicon wafer. Second, a 3 -μ m thick layer Al-Cu alloy (99.5 weight % aluminum and 0.5 weight % copper) was deposited on the layer of Sio₂ by physical vapor deposition (PVD). Third, each wafer having the Al-Cu alloy layer was anodized in an electrochemical cell at 50 V dc for 20 min in 3% by weight oxalic acid aqueous solution at 0° C. The electrical contact was made to the top metallic layer outside the electrochemical cell. The steady state current density, established after 5 minutes of anodization, was approximately 1.4 mA/cm . Preliminary experiments revealed that the entire 3 μ m thick metallic layer was anodized in approximately 80-85 min. Thus, anodization for 20 min consumed about 0.75 μ m of the metallic layer. Following anodization, the porous and barrier layers of aluminum oxide were etched in a mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids at 60° C. and for approximately 1 hour to remove the outer porous aluminum oxide and partially remove the thickness of the barrier aluminum oxide. Chromic acid is known to be an inhibitor for corrosion of aluminum and was used to decelerate the dissolution of the remaining metallic layer. Galvanic displacement of Al-Cu by silver (1.5 mM AgNO₃) was performed in a mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids, at 60° C. with no stirring and for approximately 48 hours. Anodization of Al-Cu alloy films was carried out with a Pt mesh counter electrode. All electrochemical measurements were carried out using a three-electrode cell with the Pt mesh counter electrode and a reference electrode (either a Pt wire or a Ag/AgCl electrode). Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were performed with an IM6-e impedance measurement unit (Zahner), in the same mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids, at 22° C. EIS data were acquired at open circuit potential (OCP) over a frequency range between 0.1 Hz and 100 kHz and with a potential amplitude of 5 mV. The CV and EIS data were normalized to the geometric electrode area (1.4 cm 2). During the course of electroless deposition, the depletion of silver and accumulation of copper in the deposition solution were monitored with an inductively coupled plasma (ICP) atomic absorption Perkin-Elmer spectrophotometer. The surface morphology of deposited silver particles was evaluated by a Hitachi (S-5200) scanning electron microscope operated at 2-3 kV.

Galvanic displacement proceeds via two concurrent electrochemical reactions: the reduction of metal ions and oxidation of the substrate surface. If a substrate has a layer of surface oxide, electroless deposition follows and/or coincides with chemical etching of the oxide layer. Little is known about the displacement deposition of silver on aluminum films containing copper in acidic media. Therefore, it is important to address the issue of the substrate pretreatment for the electroless deposition of silver on aluminum/copper films.

Three main observations can be made from work related to this Example. First, under the conditions of this Example, no electroless deposition of silver was observed with the pure aluminum substrate (99.997 % pure aluminum foil), which was processed in the same way as alloy films containing 99.5 wt % Al and 0.5 wt % Cu. Second, electroless deposition of silver did not take place on either of these two substrates if anodization and etching steps were omitted. Third, the Al-Cu alloy films are made amenable for electroless deposition of silver by anodization followed by complete chemical etching of porous aluminum oxide and partial chemical etching of barrier aluminum oxide. Upon completion of chemical etching, the thickness of barrier aluminum oxide remaining on the treated alloy surface is approximately 1.5 nm. A combination of anodization and chemical etching results in the copper enrichment in and underneath the thin layer of barrier aluminum oxide. This enrichment enables the charge transfer between silver cations and metallic substrate. In contrast to pure aluminum substrates, the reduction of silver cations on anodized and etched Al-Cu alloy substrates becomes possible. Thus, silver particles are deposited by galvanic displacement.

With respect to the species that are oxidized during the electroless deposition, after approximately 48 hours of electroless deposition, the depletion of silver in the electrochemical cell corresponded to the deposition of 4.0 mg (37 μ mole) of silver. In contrast, only 0.45 mg (7 μ mole) of copper was determined to accumulate in the deposition solution. If copper were the only reducing species, the molar ratio between amounts of depleted silver and accumulated copper would be 2 to 1. The high molar ratio of approximately 5 suggests that both copper and aluminum are oxidized during the electroless deposition. One can expect that aluminum is oxidized because aluminum alloys containing copper are known to have lower corrosion resistance and are more susceptible for an attack by an oxidizing agent (silver cations in our case) than pure aluminum.

Additional evidence that aluminum along with copper is oxidized during the electroless deposition of silver comes from examining the steady state mixed potential, E_mp, established during the electroless deposition of silver. Based upon the mixed potential theory, E_mp is located between the formal potentials of the reduced and oxidized species. Under our experimental conditions, E_mp (−0.7/−0.5 V vs. Ag/AgCl) was more negative than the formal potential of Cu²⁺/Cu. Such a negative value can be rationalized if the partial reaction of oxidation involves a species with a sufficiently negative formal potential (in this case aluminum).

FIG. 1A shows that galvanic displacement deposition for 48 hours proceeds via the formation of a network of spherical and randomly packed particles of silver with a broad distribution of particle diameters. The particle diameter varies from 50 to 600 nm and the mean particle diameter is about 200 nm. Particles overlap with the neighboring particles and are electrically interconnected to each other. The interconnection of particles enables the charge transfer between the aluminum-copper alloy substrate and silver cations in the solution. As a result, the growth of particles does not stop upon the deposition of the first layer of silver particles adjacent to the aluminum-copper alloy surface. Rather, galvanic displacement deposition proceeds with the formation of a 1-2 μ m thick multi-layer structure, where the average particle diameter (the thickness of a single layer) is smaller than the overall thickness of the porous layer of silver particles. FIG. 1A shows that the particle density in a single layer is about 10⁹ particles cm⁻². The inter-particle space allows for the electrolyte access among silver particles. Thus, by analyzing FIG. 1A, one can conclude that the multi-layer and porous structure composed of interconnected silver particles is expected to have a high ratio between electrolyte accessible and geometric surface areas.

Careful examination of individual silver particles (FIG. 1B) reveals that silver particles are slightly rough and interconnected to each other by other particles. Shown as white spots on the surface of 200 nm particles are silver nano-particles, which tend to grow into 200 nm particles over the time of galvanic displacement deposition. Analysis of FIG. 1A, 1B confirms that nucleation of new silver particles is a progressive process, which coincides in time with growth of existing silver particles.

In order to characterize the dielectric properties of the silver particles/electrolyte interface, a series of cyclic voltammograms was collected at scan rates of 25, 50, 100 and 200 mV/sec (FIG. 2A). While the negative potential limit is determined by water electrolysis, the positive potential limit is defined by copper oxidation. Applied potentials (FIG. 2A) are more negative than potentials, which result in significant oxidation of silver. FIG. 2A demonstrates that the current magnitude does not significantly change during either cathodic or anodic scan in the potential window of about 0.6 V between −0.5 and 0.1 V vs. Ag/AgCl. Thus, at the reported range of scan rates the faradaic current due to any possible oxygen and/or proton reduction is small. Moreover, the observed current is linearly proportional to the scan rate (FIG. 2B), which indicates its capacitive origin. This capacitance, (C_area), normalized to the electrode geometric area and determined over the shown range of scan rates, reaches a value of 1.7±0.2 mF/cm^2. We note that the capacitance is larger at slow scan rates (25 and 50 mV/s) than at fast scan rates (100 and 200 mV/s) as indicated by the slopes of dotted lines (FIG. 2B). This observation is consistent with previously reported observations that the double-layer capacitance of a porous electrode depends upon the time scale of measurements. Measurements over a long time scale result in the deep penetration of potential perturbation into the porous structure and a large sampled surface area. Thus, it is not surprising that the linear regression analysis of cyclic voltammograms with slow scan rates (25 and 50 mV/s) produces larger values of the double layer capacitance (2.0 mF/cm²) than that obtained by analyzing the whole range of scan rates.

Electroless deposition of silver and nickel has been previously shown to increase the double layer capacitance. A high capacitance of the metal/electrolyte interface has been explained by an increased surface roughness (a ratio between real and geometric surface areas). However, the capacitances reported in literature for electroless deposition of metal particles have been only one order of magnitude higher than ones typical for the metal / electrolyte interface (20-40 μ F/cm²). In contrast, C_area of the electroless deposited silver particles pursuant to this Example of the invention exceeds these values almost by two orders of magnitude.

It would be worthwhile to estimate the electrolyte accessible surface area of silver particles per gram of silver (S_a, m²/g) and compare this value with the specific surface area of spherical particles of silver per gram of silver (S_g, m²/g). Equation (1) connects S_a with the gravimetric capacitance of silver particles/electrolyte interface per gram of deposited silver (C_mass; F/g) and specific capacitance of smooth silver/electrolyte interface (C_spec=20×10⁶ F/cm²).
S_a=C_mass/C_spec=(C_area×A)/(m×C_spec)=3.5m²/g   (1)

In Equation (1), A is the geometric electrode area (1.4 cm²) and m is the mass of deposited silver (4.0 mg). Thus, one can calculate that C_mass is equal to 0.70 F/g and S_a is equal to 3.5×10⁴ cm^2/g (3.5 m²/g). For spheres of silver, S_g can be calculated according to Eq. (2), where ρis silver density (11×10⁶ g/m³) and D is the mean diameter of spheres (200 nm), as determined from FIGS. 1A, 1B.
S_g×6/(ρ×D)×2.7 m²/g   (2)
Comparison of these two values indicates that S_a is larger than S_g most likely due to slight roughness of silver particles (FIG. 1B). Thus, the surface area of silver particles is completely (except interconnected areas) utilized to increase the double layer capacitance.

S_g of interconnected silver particles is lower than S_g reported for activated carbonaceous materials (˜1000 m2/g). The difference between these specific surface areas is compensated for by one order of magnitude, when one considers that the atomic weights of carbon and silver are, respectively, 12 and 109 g/mol. Moreover, S_g of carbonaceous materials is usually determined by gas adsorption methods. Thus, the electrolyte accessible area, which is available for charged species, may be appreciably smaller than the specific surface area due to possible hydrophobicity. As shown in the previous paragraph, the opposite trend is observed for interconnected silver particles. To alleviate the difference in S_g for two materials, S_g of silver particles/electrolyte interface may be increased with the deposition of smaller particles, which have a higher inner surface to volume ratio than one obtainable with 200 nm particles.

In addition to the electrolyte accessible area and gravimetric capacitance, another point to evaluate technological utility of electroless deposition of silver particles for the fabrication of double layer capacitors is the frequency response. This variable is obtained using EIS, electrochemical impedance spectroscopy. Although EIS results are frequently presented in literature dealing with super-capacitors, the experimental data are rarely analyzed with an equivalent circuit. The difficulty in modeling of porous electrodes results from the distributed nature of the double layer capacitance along the pore length in the direction perpendicular to the electrode surface. The charge transfer resistance and Warburg impedance have to be considered in the presence of a faradaic reaction. Due to the distributed character of the interfacial impedance, the impedance of the porous electrode (including the multi-layer and porous structure composed of interconnected silver particles) is properly described by the transmission line model described in R. de Levie, Electrochem Acta 8, page 751 (1963). The model can be applied to either straight or tortuous pores.

The EIS results obtained for the silver particles/aluminum-copper electrode are summarized in FIG. 3A, 3B. While FIG. 3A shows the Bode plot (both the magnitude and phase), FIG. 3B shows the Nyquist plot. The equivalent circuit used for modeling of EIS data is shown in FIG. 3C. According to the previously developed transmission line model (www.zehner.de Application Note 01 (1997), the impedance of the porous layer consists of three elements: R_pore, the ionic resistance of pores filled with the electrolyte, R_silver, the electronic resistance of the solid layer (interconnected silver particles) and Z_q, the impedance of the interior interface between silver and electrolyte. Z_q is modeled as a serial connection of two constant phase element/resistor combinations. The constant phase element (CPE) is often used instead of a pure capacitance to describe interfacial dielectric properties. EIS were performed at OCP while the electroless deposition of silver was taking place. Therefore, resistors (R₁ and R₂) model the charge transfer across the silver particles/electrolyte interface due to reduction of silver cations and concurrent oxidation of the aluminum/copper substrate. It is worthwhile to note that the EIS data between 0.1 and 100 Hz can be adequately modeled with R_pore, R_silver and Z_q containing only a single (CPE₁ R₁) combination. However, satisfactory description of the frequency response between 100 Hz and 100 kHz requires that the second (CPE₂ R₂) combination be introduced. The exact origin of the second (CPE₂ R₂) combination is undefined. In addition to the impedance of the porous layer, R_s is used in the equivalent circuit to model the bulk electrolyte resistance. The underlying silicon substrate does not appear in the analysis of EIS data because the electrical connection was made to the top metallic layer.

The magnitude of CPE₁, which describes the silver particles/electrolyte capacitance, is determined to be 1.4±0.1 mF×s^α−1/cm², a value that is slightly smaller than one obtained by CV (assuming that α₁≃1). The smaller magnitude of CPE₁, results from the fact that EIS is performed over a shorter time scale than CV. Other parameters of the equivalent circuit, which is used to fit the EIS data to the transmission line model, are calculated as follows: R_s =27±1Ω×cm², R_pore=370±60Ω×cm², R_silver=31±2Ω=cm², α₁=0.98±0.01, R₁=3.9±1.6 kΩ×cm², CPE₂32 200±25 μ F×s^α−/cm², α₂=0.87±0.02, R₂=1.9±0.2 Ω×cm². These values agree with the qualitative analysis of EIS data. For example, the total cell impedance at frequencies higher than 10 kHz is indicative of a pure resistor of 58 Ω×cm² (FIG. 3B). Due to negligible impedance of capacitors, this value is a sum of the bulk electrolyte resistance, R_s, and the electronic resistance of interconnected silver particles, R_silver. At frequencies between 200 Hz and 10 kHz, a combination of the ionic resistance of pores, R_pore, and increased impedance of CPE₁ produces a depressed semi-circle (FIG. 3B). At frequencies between 10 and 200 Hz, the impedance of the interior interface between silver particles and electrolyte, Z_q, becomes dominant, which results in a straight line on the Nyquist plot (FIG. 3B).

The Nyquist plot for porous electrodes is typically divided into two regions by the “knee” frequency. As discussed in the previous paragraph, the impedance in the high frequency region (≧200 Hz) is due the porous electrode structure, the impedance in the low frequency region (≦200 Hz) is dominated by the whole interior interface between silver particles and electrolyte. Examination of FIG. 3A reveals that the “knee” frequency is located at about 200 Hz. This value suggests that the electrical energy can be stored in the double-layer capacitor at frequencies up to 200 Hz. In contrast, the majority of carbon-based super-capacitors, with a few exceptions, show the “knee” frequency around a few Hz. Therefore, their capacitive behavior deteriorates between 10 and 100 Hz. The superior performance (the high “knee” frequency) achievable with the capacitor described here is believed to result from the suitable porosity (the inter-particle distance of about 20-40 nm) and easy access of the electrolyte among the deposited silver particles. The high “knee” frequency of 200 Hz is an important advantage of the capacitors reported in this paper.

A critical issue in the design of high power density super-capacitors is the low electronic resistivity of the porous electrode structure. For example, it ahs been reported that the contact resistance between the elements (particles or fibers) in the electrode must be very small. Among metals, silver has the lowest electrical resistivity (1.6 μΩ× cm). However, the electronic resistance of interconnected silver particles, R_silver, is comparatively high. This value is determined by tiny interconnected areas among silver particles as well as the contact resistance of silver particles to the aluminum substrate. One can speculate that the interconnected areas can be increased and, accordingly, R_silver can be decreased with the deposition of smaller particles. Future study will aim at the deposition of 30-60 nm particles of silver, which are expected to have a higher inner surface to volume ratio and larger interconnected areas than those reported in this paper.

To summarize this Example, the utilization of electroless deposition of silver by the galvanic displacement mechanism on aluminum-copper alloy films has been demonstrated in order to fabricate a multi-layer and porous network composed of electrically interconnected silver particles. This structure has a high ratio between electrolyte accessible and geometric surface areas (≃100). Two electrochemical techniques (CV and EIS) independently suggest that the capacitance of the silver particles/electrolyte interface normalized to the electrode geometric area (1.7 mF/cm²) exceeds those typical for the smooth silver/electrolyte interface (20 μF/cm²) by two orders of magnitude. Evaluation of electrochemical data and SEM micrographs suggests that the surface area of silver particles is completely accessible to the electrolyte (except interconnected areas). The gravimetric capacitance of silver particles/electrolyte interface per gram of silver is 0.70 F/g and the useful DC potential range is approximately 0.6 V. Analysis of the Nyquist plot shows that the network of silver particles is capable of storing electrical energy at frequencies up to 200 Hz. This value is higher than those typically reported for carbon based double layer capacitors. In addition, use of silicon wafers with aluminum/copper alloy films is attractive because these wafers are frequently employed in standard micro-fabrication lines. Given these two advantages, the described capacitor could find applications for special electronic circuits where a high frequency response is required.

Example 2 describes a method pursuant to another illustrative embodiment of the invention wherein an aluminum-copper alloy surface is anodized and then chemically etched followed by galvanic deposition or electrodepostion of silver on the treated alloy surface.

EXAMPLE 2

In particular, aluminum-copper alloy covered wafers used in this Example were fabricated as follows: First, silicon wafers with a 600 nm thick layer of SiO₂ overlayed with a 3 μ m thick layer of 99.5 wt % Al and 0.5 wt % Cu (deposited by physical vapor deposition) were used in all experiments. The Al-Cu alloy layer was anodized with a Pt mesh counter electrode at 50 V DC for 20 min in 3 wt % H₂C₂O₄ acid and at 0° C. The electrical contact was made to the top metallic layer outside the electrochemical cell. After anodization, the porous and barrier layers of aluminum oxide were etched in a mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids at 50 ° C. for approximately 90 minutes to remove the outer porous aluminum oxide and partially remove the thickness of the barrier aluminum oxide. Upon completion of etching the specific capacitance of barrier aluminum oxide was determined to be 5.8 μ F/cm². Assuming that the dielectric constant was 8.6, the layer of barrier aluminum oxide remaining on the treated alloy surface was estimated to be 1.3 nm thick. H₂CrO₄ was used to inhibit the dissolution of the remaining Al-Cu alloy film. Following anodization and chemical etching, either galvanic displacement type or potentiostatic electrodeposition of silver (3.0 mM AgNO₃) was performed in a mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids at 50° C., with no stirring and for 40 min. Potentiodynamic polarization measurements and electrochemical impedance spectroscopy (EIS) were performed with an IM6-e impedance measurement unit (Zahner). All electrochemical measurements were carried out using a three-electrode cell with a Pt mesh counter electrode and a reference electrode (either a Pt wire or a Hg/HgSO₄ electrode) in the mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids at 50 ° C. For polarization measurements, the potential step was 2 mV and the time delay to sample the steady state current was 1 s. EIS data were acquired at open circuit potential (OCP) over a frequency range between 0.5 Hz and 100 kHz, with a potential amplitude of 5 mV and were normalized to the geometric electrode area (1.4 cm²). The surface morphology of deposited silver particles was evaluated by a Hitachi (S-5200) scanning electron microscope operated at 2-3 kV.

According to the mixed-potential theory for galvanic displacement type deposition, the steady state mixed potential, E_mp, is determined by the partial currents of reduction and oxidation reactions, which are equal to each other in the magnitude and opposite in the sign. When the rate of partial cathodic reaction is increased by adding a suitable reducible species, E_mp shifts in the anodic direction. In this Example, the effect of addition of AgNO₃ on both the exchange current density and E_mp is further investigated with potentiodynamic polarization experiments (FIG. 4) performed prior to either galvanic displacement type or electrodeposition. Three features are notable by comparing the representative E-log (j) curves of stationary Al-Cu alloy electrodes, (a) after anodization and etching and (b) upon addition of AgNO₃. First, before addition of AgNO₃, OCP of −0.88 V is determined by partial reactions involving oxidation of the Al-Cu substrate and reduction of protons and/or residual oxygen. Upon addition of AgNO₃, E_mp shifts to ≃230 mV more positive indicating that reduction of silver cations becomes the dominant cathodic reaction. Second, the induced anodic shift of E_mp results in a larger overpotential for the oxidation reaction, which increases the partial anodic current density and accelerates oxidation and dissolution of the Al-Cu substrate. Third, as a result of higher rates of cathodic and anodic reactions, the exchange current density increases from (a) 2×10⁻⁷A/cm2 to (b) 7×10⁻⁶A/cm². FIG. 4 illustrates that both the exchange current density and E_mp are correlated with the rates of partial electrode reactions. Therefore, polarization experiments are informative to predict whether galvanic displacement type deposition proceeds with a reasonable deposition rate.

FIG. 5 shows the state of the surface of anodized and etched Al-Cu alloy substrate after deposition of silver by galvanic displacement for 40 min. The black scallops with white edges can be observed behind silver particles. These shallow scallops are formed on the surface as a result of anodization and subsequent chemical etching. The coverage of randomly distributed particles of silver is about 4×10⁹ particles cm⁻². A relatively large D_mean of 180±110 nm and S_g of 3 m²/g indicate that new particles do not nucleate as fast as existing particles grow.

FIG. 4 allows one to determine how the cathodic electrode polarization increases the rate of silver reduction. While the exchange current density at E_mp (−0.65 V) is 7×10⁻⁶ A/cm², the cathodic current density at −1.3 V (0.65 V more negative than E_mp) is 6×10⁻⁴ A/cm ². One can hypothesize that the cathodic electrode polarization would increase the rate of nucleation of new particles rather than the growth of existing particles. In order to investigate this hypothesis, the Al-Cu alloy electrode was polarized to −1.3 V vs. Hg/HgSO₄ for a time interval of 40 min, during which the cathodic current density gradually increased from 0.6 to 2 mA/cm². The faradaic efficiency for silver reduction at −1.3 V is about 40% due to reduction of protons and/or residual oxygen. This assumption is based upon the comparison of curves (a) and (b) (FIG. 4) and observing an inflexion point in the curve (b) around −1.1 V. The inflexion point indicates that the faradaic current due to residual faradaic reactions starts to exceed the faradaic current due the electrodeposition of silver. Comparison of curves (a) and (b) suggests that the current density due to residual faradaic reactions is higher in the latter case. This observation can be explained by a high catalytic activity and a high surface area of silver particles electrodeposited during a potentiodynamic scan in comparison with those of the anodized and etched Al-Cu film. Taking into account the faradaic efficiency for silver reduction, a mass of deposited silver (1.0 mg) is determined by integrating the current density over time because electrodeposition dominates over galvanic displacement at −1.3 V.

FIG. 6 shows a cauliflower type film composed of densely packed nanoparticles, which are produced by electrodeposition of silver on the anodized and etched Al-Cu alloy surface. The coverage of nanoparticles is about 10¹¹ particles cm⁻² and D_mean is 30±7 nm. Comparison of FIG. 5 and 6 indicates a dramatic difference in the surface morphology between two methods of deposition. Electrodeposition results in the formation of particulate silver films with D_mean one order of magnitude smaller and the particle coverage two orders of magnitude higher than those obtained with galvanic displacement type deposition. For spherical particles, S_g can be calculated according to Equation (1′), where p is the density of silver (11×10⁶ g/m³). Table 1 summarizes S_g obtained with both methods (galvanic displacement and electrodeposition).
S_g=6/(ρ×D_mean )   (1′)

Due to the small D_mean , the fabricated 3-dimensional network of electrodeposited silver nanoparticles has a higher S_g than that composed of silver particles deposited by galvanic displacement.

Microscopic examination of deposited silver particles was supplemented with macroscopic electrochemical measurements. In order to confirm that electrodeposition results in the formation of silver nanoparticles with a high inner-to-geometric surface area ratio, the electrolyte accessible surface area, S_a, was estimated by EIS. FIGS. 7 and 8 show the Bode representation of two EIS data sets collected after galvanic displacement type deposition and electrodeposition, respectively. Qualitative analysis of EIS spectra (FIGS. 7 and 8) around 100 Hz and 0.5 Hz results in two observations. The magnitude of the total cell impedance is lower and the phase of the total cell impedance is less negative for the electrode with electrodeposited particles of silver than those for the electrode with particles deposited by galvanic displacement. Both observations can be explained by the fact that a larger capacitance makes a smaller contribution to the total cell impedance in the former case in comparison with the latter case.

For quantitative analysis, EIS spectra are modeled with the equivalent circuits shown in FIG. 7A and FIG. 9. The equivalent circuit describing particles deposited by galvanic displacement for 40 min (FIG. 7A) does not require considering the 3-dimensional structure of deposited particles. On the contrary, the equivalent circuit used for modeling of electrodeposited particles (FIG. 9) includes the transmission line model developed for a porous electrode. This model has to be employed because the capacitance normalized to the geometric electrode area, C_area, obtained for the electrode with electrodeposited particles of silver exceeds the specific capacitance of a smooth silver/electrolyte interface, C_spec(20×10⁻⁶ F/cm²), by two orders of magnitude (Table 1). Both models were described in details in previous publications. C_area, and gravimetric capacitances normalized to the mass of deposited silver, C_mass, are shown in Table 1. The electrolyte accessible surface area of electrodeposited silver particles per gram of silver (S_a, m²/g) can be determined as the ratio of C_mass and C_specaccording to Equation (2′).
S_a=C_mass/C_spec=(C_area×A)/(m ×C_spec)   (2′)
In Equation (2′), A is the geometric electrode area (1.4 cm²) and m is the mass of deposited silver. As expected, S_a is approximately the same as S_g (Table 1). Thus, the surface area of electrodeposited particles of silver is completely accessible to the electrolyte. Based upon these observations, one can conclude that electrodeposition results in the deposition of porous silver films composed from electrically interconnected nanoparticles of silver. S_g of electrodeposited particles is one order of magnitude higher than that obtained with particles deposited by galvanic displacement.

It is important to note why the cathodic polarization of the Al-Cu electrode favors the deposition of silver particles with a higher nucleation density than that obtained with galvanic displacement. The nucleation density is known to exponentially increase with the applied overpotential. Therefore, it is not surprising that a high nucleation density is obtained when the Al-Cu alloy electrode is polarized to 0.6-0.7 V more negative than E_m (−0.65 V) established during galvanic displacement. The cathodic polarization increases the applied overpotential for reduction of silver cations, which translates in a high nucleation density.

Particles of silver remain adhesive to the Al-Cu alloy substrate during electrodeposition, electrochemical and microscopic examination. Mechanical strength of the fabricated porous structure can be explained by the fact that the D_mean of 30 nm allows for a plenty of interconnection points among silver particles per unit volume. The corrosion resistance of electrodeposited particles of silver can be further improved with co-deposition of refractory metals such as tungsten.

Electrodeposition of silver on anodized and etched Al-Cu alloy substrates can be compared with galvanic displacement. In both cases, the combination of anodization and chemical etching results in the copper enrichment in and underneath the thin layer of barrier aluminum oxide. The enrichment in copper during anodization enables subsequent electrodeposition of silver in this Example.

In the case of electrodeposition, cathodic polarization allows one to control the nucleation and growth of silver particles and, consequently, properties of deposited silver films. It is worthwhile to point out that the technologically important zincation of aluminum substrates is galvanic displacement type deposition.

In order to increase the technological utility, the procedure to activate the Al-Cu alloy film for electrodeposition must be compatible with standard photolithographic methodology. Previous results obtained in our laboratory indicated that photoresist did not exhibit acceptable adhesion to the Al-Cu alloy film during porous type anodization. This problem was overcome with the negative pattern transfer technique. The procedure for electrodeposition on patterned Al-Cu alloy films is summarized at FIGS. 10A-10F. Silicon wafers with Al-Cu films were patterned with photoresist, producing two circular patterns with diameters of 10 and 30 um (FIG. 10A-step 1). The photoresist mask was transferred to an approximately 100 nm thick layer of barrier aluminum oxide (FIG. 10B-step 2). This layer was formed in the regions not covered with the photoresist under the following conditions: voltage of 90 V, anodization time of 2-3 min, temperature of 3° C. and in 3 % w/v Na₂C₂0₄ (pH 4.5). After removal of photoresist (FIG. 10C-step 3), subsequent 20 min anodization (FIG. 10D-step 4) at voltage of 50 V, temperature of 3° C. and in 3 % w/v H₂C₂0₄ (pH 1.5) resulted in the formation of porous aluminum oxide only in those regions, which were not covered with the layer of barrier aluminum oxide and were initially covered with the photoresist. In order to enable electrodeposition, the whole layer of porous aluminum oxide and almost all of underlying barrier aluminum oxide (1.3 nm left) were chemically etched (FIG. 10E-step 5) in a mixture of 0.4 M H₃PO₄ and 0.2 M H₂CrO₄ acids at 50° C. for approximately 90 min as described in the experimental section for blanket Al-Cu films. Electrodeposition of silver (3.0 mM AgNO₃) was performed at -1.3 V vs. Hg/HgSO₄ (FIG. 10F-step 6).

FIGS. 11A, 11B demonstrate the state of the alloy surface after completion of all steps in the fabrication procedure (FIGS. 10A-10F). Silver is electrodeposited (FIG. 10F-step 6) only in the circular regions, which underwent porous type anodization (FIG. 10D-step 4). The space-selective deposition of silver provides indirect evidence that the copper enrichment, which occurs during porous typelanodization of Al-Cu alloy films, enables the electrodeposition of silver on anodized and chemically etched regions of Al-Cu alloy films. The silver deposits are porous and similar to those shown in the high magnifications micrograph (FIG. 6). FIGS. 11A, 11B confirm that the pretreatment method developed for activation of Al-Cu alloy films (anodization and chemical etching) is compatible with photolithographic techniques.

To summarize this Example, activation of technologically relevant Al-Cu alloy substrates for electrodeposition is achieved by anodization followed by chemical etching of aluminum oxide. Electrodeposition of silver on anodized and etched Al-Cu alloy substrates results in the fabrication of a porous film built from electrically interconnected nanoparticles of silver with D_mean of 30 nm. The coverage of electrodeposited particles of silver is 4×10¹¹ particles cm⁻². Microscopic examination by SEM is supplemented with macroscopic electrochemical measurements (EIS). EIS is shown to be a useful in-situ method to monitor the surface area of deposited particulate films. The frequency response of the porous network of electrodeposited silver nanoparticles is evaluated using the transmission line model. The capacitance normalized to the geometric electrode area is 2.9±0.1 mF/cm² and the capacitance normalized to the mass of deposited silver is 3.9±0.1 mF/g. The electrolyte accessible area of electrodeposited silver nanoparticles is 20 m²/g. The method developed for electrodeposition on anodized and etched Al-Cu alloy films is compatible with photolithographic techniques. Electrodeposition on patterned Al-Cu alloy films is accomplished by transferring the photoresist mask to a layer of barrier aluminum oxide. This layer acts as a mask for porous type anodization. Following anodization and chemical etching, electrodeposition of silver takes place only on anodized and etched areas. Table 1. The mean particle diameter, D_mean ; specific surface area, S_g; mass of deposited silver, m; capacitance normalized to the geometric electrode area; C_area, gravimetric capacitances; C_mass, and electrolyte accessible surface area; S_a, for silver particles deposited by either galvanic displacement type deposition or electrodeposition.

	Dmean (nm)	Sg (m2/g)	m	Carea	Cmass	Sa
Process	(SEM)	(SEM)	(mg)	(mF/cm2)	(F/g)	(m2/g)
Galvanic	180 ± 110	3.0 ± 1.8	N/A	0.058 ± 0.002	N/A	N/A
displacement
ElectroDeposition	30 ± 7	18 ± 4	1.0	2.9 ± 0.1	3.9 ± 0.2	19 ± 1

Although the invention has been described in connection with certain embodiments thereof, those skilled in the art will appreciate that the invention is not limited to these illustrative embodiments and that changes and modifications can be made thereto within the scope of the invention as set forth in the following claims.

Claims

1. Method of depositing a metallic material on a substrate, comprising the steps of providing a substrate comprising an alloy of aluminum and an alloying element, oxidizing a surface of the substrate to form aluminum oxide thereon, etching the oxidized surface to leave a partial thickness of a barrier aluminum oxide of said aluminum oxide on the surface, and depositing by galvanic displacement type deposition, electroless deposition or electrodeposition discrete metallic nanoparticles on the barrier oxide having a particle density of about 104 to about 1012 particle/cm2.

2. The method of claim 1 wherein said etching is conducted to leave a barrier oxide portion having a partial thickness to increase nucleation of the discrete nanoparticles thereon.

3. The method of claim 1 wherein the substrate comprises a film or layer of the alloy.

4. The method of claim 1 wherein the alloying element includes one or more of Au, Cu, Cr, Mn, Mo, Ni, Si, Ta, Ti, or Zn.

5. The method of claim 1 wherein the surface is oxidized by anodizing, polishing, alkaline etching, acid pickling, electropolishing, or heating in an oxygen bearing atmosphere.

6. The method of claim 1 wherein the surface is acid etched by contact with a mixture of phosphoric acid and an inhibitor for aluminum dissolution.

7. The method of claim 6 wherein the inhibitor comprises chromic acid.

8. The method of claim 1 wherein the metallic material comprises one of Au, Ag, Pd, Cu, Ni, Pb, Cr, Fe, W, Mo, or Co.

9. The method of claim 1 wherein the substrate is disposed on a silicon wafer.

10. A substrate comprising an alloy of aluminum and an alloying element, said substrate having a barrier oxide on a substrate surface and a porous, three dimensional structure of electrically interconnected, metallic nanoparticles deposited by electroless deposition or electrodeposition on the barrier oxide.

11. The substrate of claim 10 wherein the structure includes randomly packed, generally spherical metallic nanoparticles having a distribution of particle sizes.

12. The substrate of claim 10 wherein the nanoparticles have a particle diameter in the range of about 20 nm to about 1000 nm.

13. The substrate of claim 10 wherein the nanoparticles have an interparticle spacing sufficient to provide electrolyte access among the nanoparticles, providing a high surface area material.

14. The substrate of claim 10 wherein the structure has a ratio between electrolyte accessible area and geometric surface area of about 100 and above.

15. Electrode comprising a substrate comprising an alloy of aluminum and an alloying element, said substrate having a barrier oxide on a substrate surface and a porous, electrolyte accessible, three dimensional structure of electrically interconnected, generally spherical, randomly packed metallic nanoparticles deposited by galvanic displacement type deposition, electroless deposition or electrodeposition on the barrier aluminum oxide.

16. The electrode of claim 15 wherein the nanoparticles have a particle diameter in the range of about 20 nm to about 1000 nm.

17. The electrode of claim 15 wherein the structure has a ratio between electrolyte accessible area and geometric surface area of about 100 and above.

18. The electrode of claim 15 wherein the barrier oxide film has been chemically etched.

19. The electrode of claim 15 wherein the substrate is disposed on a silicon wafer.

20. Capacitor having an electrode in accordance with claim 15.

21. Method of depositing a metallic material on a substrate, comprising the steps of providing a substrate comprising an alloy of aluminum and an alloying element, oxidizing a surface of the substrate to form aluminum oxide thereon, etching the oxidized surface to leave a partial thickness of a barrier aluminum oxide of said aluminum oxide on the surface, and depositing by galvanic displacement type deposition, electroless deposition or electrodeposition a continuous metallic film on the barrier oxide.

22. The method of claim 21 wherein said etching is conducted to leave a barrier oxide portion having a partial thickness to deposit a continuous metallic film.

23. The method of claim 21 wherein the substrate comprises a film or layer of the alloy.

24. The method of claim 21 wherein the alloying element includes one or more of Au, Cu, Cr, Mn, Mo, Ni, Si, Ta, Ti, or Zn.

25. The method of claim 21 wherein the surface is oxidized by anodizing, polishing, alkaline etching, acid pickling, electropolishing, or heating in an oxygen bearing atmosphere.

26. The method of claim 21 wherein the surface is acid etched by contact with a mixture of phosphoric acid and an inhibitor for aluminum dissolution.

27. The method of claim 26 wherein the inhibitor comprises chromic acid.

28. The method of claim 21 wherein the metallic material comprises one of Au, Ag, Pd, Cu, Ni, Pb, Cr, Fe, W, Mo, or Co.

29. The method of claim 21 wherein the substrate is disposed on a silicon wafer.