US 20040154925 A1
Several advantages result from the incorporation of nanoparticulates into the plating bath when electrodepositing metal composite microstructures or metal alloy composite microstructures. Adding even relatively small quantities of nanoparticulates can enhance the rate of alloy deposition, cause the deposition of smoother structures, and produce composites having superior properties as compared to otherwise similar alloys lacking the particulates—for example, composites having greater hardness. These techniques are useful for plating composites into deep recesses of a microstructure.
1. A process for making a metal composite microstructure comprising at least one metal selected from the group consisting of nickel, copper, iron, tungsten, cobalt, molybdenum, gold, zinc, silver, chromium, tin, and lead; said process comprising the steps of:
(a) supplying into a mold an aqueous solution comprising a salt of the metal and comprising suspended particulates; wherein the shape of the mold is complementary to the shape of the microstructure to be made, and wherein the mold includes one or more deep recesses having a depth of at least about 10 μm, and having an aspect ratio greater than about 1; and wherein the particulates have a number-average diameter less than about 500 nm; and
(b) depositing the metal from the aqueous solution into the mold by pulsed electroplating; wherein the electroplating pulses have an on time that is sufficiently short, and a duty cycle that is sufficiently low, that the products of the electroplating reactions and the products of any side reactions do not accumulate in concentrations that are sufficiently high to substantially interfere with the electroplating of the desired metal into the deep recesses; and wherein said depositing step is continued until a metal composite microstructure has been formed having a height of at least about 10 μm, and having an aspect ratio greater than about 1; wherein the composite microstructure comprises an intermixture of metal and particulates; wherein the concentration of the particulates is such that the rate at which the metal is deposited is faster than the metal would be deposited in an otherwise identical process lacking the particulates; and wherein the depositing of at least one of the metals is under mass transport control.
2. A process as recited in
3. A process as recited in
4. A process as recited in
5. A process as recited in
6. A process as recited in
7. A process as recited in
8. A process as recited in
9. A process as recited in
10. A process as recited in
11. A process as recited in
12. A process as recited in
13. A process as recited in
14. A process as recited in
15. A microstructure produced by the process of
16. A microstructure produced by the process of
17. A microstructure produced by the process of
18. A microstructure produced by the process of
19. A microstructure produced by the process of
20. A microstructure produced by the process of
21. A microstructure produced by the process of
22. A microstructure produced by the process of
23. A microstructure produced by the process of
24. A microstructure produced by the process of
25. A microstructure produced by the process of
26. A microstructure produced by the process of
27. A microstructure produced by the process of
28. A microstructure produced by the process of
29. A composite comprising a nickel-copper alloy intermixed with alumina particles, wherein said alumina particles have a number-average diameter less than about 500 nm.
30. A microstructure comprising at least one feature comprising a composite as recited in
31. A microstructure comprising at least one feature comprising a composite; wherein said feature has a height between about 10 μm and about 2000 μm; wherein said feature has an aspect ratio greater than about 1; wherein said composite comprises a metal alloy intermixed with particulates having a number-average diameter less than about 500 nm; wherein said particulates have a composition that differs from the composition of said alloy; and wherein the composition of said feature varies as a function of height.
32. A microstructure as recited in
33. A microstructure as recited in
 The development of this invention was partially funded by the Government under grant number ECS 9984775 awarded by the National Science Foundation. The Government has certain rights in this invention.
 Microdevices and micro-electro-mechanical systems (MEMS) are an important emerging technology. Novel devices on the micron scale may be used in a variety of applications. One step in the fabrication of micron-scale devices typically involves electrodeposition of metals into deep recesses. The cavities used in the electrodeposition step are typically produced by x-ray or UV lithography, depending on the depth of the cavity and the number of repetitive processing steps tolerated in a particular process. Nickel has often been the metal used in electrodeposition, because it is relatively easy to plate. However, the properties of unalloyed nickel are less than ideal for many applications; for example, it is relatively soft. Alternatively, metal composites or alloys may be used to impart a desired property. For example, SiC has been electrodeposited with Ni from a sulfamate electrolyte for MEMS applications.
 The production of composite materials and alloys on a micron scale often requires techniques that differ from alloy fabrication methods that have traditionally been employed at the macro level.
 S.-H. Yeh et al., “Microcomposite electroforming for MEMS technology,” pp. 82-90 in S.-C. Chang et al. (Eds.), Proc. SPIE, Micromachining and Microfabrication Process Technology III, vol. 3223 (Austin, Tex., Sep. 29-30, 1997) reported the electrodeposition of a composite micromold for a spinneret for molding microfibers, where the deposited composite contained SiC particles in Ni.
 T. Wang, Fabrication and characterization of LIGA nickel-alumina composite high-aspect-ratio microstructures, MS Thesis (Louisiana State University, Baton Rouge 2000) reported the electrodeposition of alumina particles in nickel from a conventional sulfamate electrolyte into 500 micron recesses using constant DC plating techniques. It was reported that the deposition rate decreased as compared to the theoretical. It was reported that microhardness improved with the incorporation of particles, but that ductility decreased.
 S. Steinhäuser et al., “Electroplated composite coatings with nanoscaled particles,” Abstract from 197th Meeting of the Electrochemical Society (Toronto, 2000) reported “nano-dispersion coatings” that incorporated nanoparticular TiO2 or Al2O3 into a matrix metal layer.
 E. Podlaha, “Selective electrodeposition of nanoparticulates into metal matrices,” Nano Letters, vol. 1, pp. 413-416 (2001) discloses the use of pulse-reverse electrodeposition to co-deposit γ-Al2O3 nanometric size particles, ≦30 nm in diameter, into a copper matrix thin film, from electrolytes with a disparity of particle sizes, including particle diameters exceeding 150 nm. The selectivity from the pulse reversal provided a means to control particle size distribution. The pulse reversal method employs long, steady-state pulses several seconds to minutes long. The metals and particles are deposited during the cathodic cycle. During the anodic cycle, part of the metal is removed, and thus concentrates the co-deposited particles. The pulse reverse method captures particles that are the same order or smaller than the net thickness of the deposit per cycle, releasing larger particles during the anodic dissolution cycle. See also E. Podlaha et al., “Pulse-reverse plating of nanocomposite thin films,” J. Electrochem. Soc., vol. 144, pp. L200-L202 (1997).
 W. Ehrfeld et al., “Materials of LIGA Technology,” Microsystem Technologies, vol. 5, pp. 105-112 (1999) mentioned, among others, the following composites as being under testing and development on the microscale, without giving details as to how the composites were being made, nor giving results of the work: Ni—SiO2, Ni—Al2O3, Ni-diamond, and Ni—TiN.
 Reviews and other papers concerning composite electrodeposition techniques that have previously been used to plate unrecessed substrates on the macro scale include A. Hovestad et al., “Electrochemical codeposition of inert particles in a metallic matrix,” J. Appl. Electrochem., vol. 25, pp. 519-527 (1995); J. Fransaer et al., “Mechanisms of Composite Electroplating,” Metal Finishing, vol. 91, pp. 97-100 (1993); J. Fransaer et al., “New Insights into the Mechanism of Composite Plating,” Galvanotechnik, vol. 92, pp. 1544-1550 (2001); J. Celis et al., “Electrolytic and Electroless Composite Coatings,” pp. 1-41 in R. Parkins (Ed.), “Reviews in Coatings and Corrosion,” vol. 5, No. 1-4 (1982); J. P. Celis et al., “Properties of Electrodeposited Copper-Alumina Coatings,” Trans. Inst. Metal Finishing, vol. 56, pp. 41-45 (1978); R. Narayan et al., “Electrodeposited Composite Metal Coatings,” pp. 113-153 in J. Yahalom (Ed.), Reviews on Coatings and Corrosion, vol. IV, No. 1 (1979); M. Musiani, “Electrodeposition of composites: an expanding subject in electrochemical materials science,” Electrochimica Acta, vol. 45, pp. 3397-3402 (2000); and J. Dini, “Copper composites,” Part E, pp. 134-137 in M. Schlesinger et al., Modern Electroplating (2000).
 The presence of alumina particles in macro-scale composite coatings has been reported to significantly enhance wear resistance and hardness. See Y-S. Chang et al., “Wear resistant nickel composite coating from bright nickel baths with suspended very low concentration alumina,” Materials Chemistry and Physics, vol. 20, pp. 309-321 (1988); V. Greco, “A review of fabrication and properties of electrocomposites,” Plating and surface finishing, pp. 68-72 (October 1989); V. Greco et al., “Electrodeposition of Ni—Al2O3; Ni—TiO2 and Cr—TiO2 Dispersion Hardened Alloys,” Plating, vol. 55, pp. 250-257 (1968); H. Ferkel et al., “Electrodeposition of particle-strengthened nickel films,” Mat. Sci. Eng., vol. A234-236, pp. 474-476 (1997); and C. Kerr et al., “The electrodeposition of composite coatings based on metal matrix-included particle deposits,” Trans. IMF, vol. 78, no. 5, pp. 171-178 (2000).
 Ultrafine particle sizes (below about a micron) have recently been suggested as a possible means to improve material properties by providing increased particle-to-metal uniformity, increased surface contact area for wear resistance, and improved hardness for metal matrices. For prior examples of composite electrodeposition carried out on unrecessed substrates, see X. Ding et al., “Mechanical behavior of metal-matrix composite deposits,” J. Mat. Sci., vol. 33, pp. 803-809 (1998); and P. Webb et al., “Electrolytic codeposition of Ni-γAl2O3 thin films,” J. Electrochem. Soc., vol. 141, pp. 669-673 (1994).
 J. Roos et al., “Electrochemical study of alloy and composite plating,” pp. 177-215 in Proc. Intl. Conf. Appl. Polarization Measurements in the Control of Metal Deposition (Victoria, Canada, May 4-7, 1982) describes the electrodeposition of CuNi alloy coatings from citrate baths on the macroscale, as well as the electrodeposition of Cu-alumina and Au-alumina composite coatings.
 P. Webb et al. (1994) reported that the rate of nickel deposition was inhibited by low concentrations of submicron alumina particles in the electrolyte (on the order of 10 g/L).
 J. Stojak et al., “Investigation of electrocodeposition using a rotating cylinder electrode,” J. Electrochem. Soc., vol. 146, pp. 4504-4513 (1999) reported that in electrodeposition of copper in the presence of alumina particles, alpha alumina particle loadings at or below 120 g/L had little or no effect on the mass-transfer limiting conditions. However, particle loadings of 158 g/L were reported to increase the limiting current density by as much as 32%.
 J. Stojak et al., “Effect of particles on polarization during electrodeposition using a rotating cylinder electrode,” J. Appl. Electrochem., vol. 31, pp. 559-564 (2001) reported experiments on the influence of suspended submicrometer alumina particles on polarization during electrodeposition with copper at a rotating cylinder electrode, in both kinetic control and mass transfer control regimes. The authors concluded that in the kinetic control regime, the suspended particles led to a decrease in current; that for mass transfer limiting conditions, alumina particle concentrations below 120 g/L had no effect; and that for mass transfer limiting conditions, an alumina particle concentration of 158 g/L increased the limiting conditions by as much as 32%.
 On a macroscale, pulsed electroplating techniques have been used to make deposited films of metal brighter and smoother, to alter morphology and grain size. Reviews of the state of the art in pulsed electroplating techniques include J. Dini, section 10 in “Current Modulation Techniques,” M. Schlesinger et al., Modern Electroplating, pp. 81-83 and 99-101 (John Wiley & Sons, 4th Ed., 2000); and F. Mueller, “Pulse Plating & Other Myths,” Plating & Surface Finishing, pp. 54-55 (April 2000).
 B. Löchel et al., “Ultraviolet depth lithography and galvanoforming for micromachining,” J. Electrochem. Soc., vol. 143, pp. 237-244 (1996) reported the pulsed plating of microdevices with a NiFe alloy containing 15% Fe, as well as pulsed plating with gold. The pulse mode for gold was conducted at 222 Hz (corresponding to a cycle time of about 4.5 ps) and a duty cycle of 0.44.
 B. Löchel et al., “Fabrication of magnetic microstructures by using thick layer resists,” Microelectronic Eng., vol. 21, pp. 463-466 (1993) reported the pulsed plating of microdevices with a NiFe alloy containing a maximum of 55% iron. The authors described an optimized deposition with a duty ratio of 0.4, a frequency of 250 Hz, and a pulse current density of about 20 mA/cm2. Resist patterns with structure sizes down to 7 μm were reported in a thickness of about 40 μm, or an aspect ratio of nearly 6.
 B. Löchel et al., “Galvanoplated 3D structures of Micro Systems,” Microelectronic Eng., vol. 23, pp. 455-459 (1993) reported the pulsed electrodeposition of Ni/Fe alloys having aspect ratios up to 10 (6 μm feature in a 60 μm resist layer). The authors described an optimized deposition with a duty ratio of 0.5, a frequency of 100 Hz, and a pulse current density of about 60 mA/cm2.
 S. Leith et al., “In-situ fabrication of sacrificial layers in electrodeposited NiFe microstructures,” J. Micromech. Microeng., vol. 9, pp. 97-104 (1999) reported electrodeposition with pulsed electrolyte agitation, followed by etching, to produce a range of layered NiFe magnetic alloys having differing Fe composition (e.g., 20, 48, 55, and 68 mol %), but apparently having low aspect ratios.
 See also P. Datta et al., “Design and Fabrication of Thermomechanical Microactuatorfor High Temperature Applications,” poster presented at HARMST (High Aspect Ratio Microsystem Technology) 2001 Conference in Baden-Baden, Germany (Jun. 17-19, 2001); L. Namburi et al., “Electrodeposition of Ni—W alloys into deep recesses,” Abstract, Joint 200th Meeting of the Electrochemical Society and the 52nd annual meeting of the International Society of Electrochemistry (San Francisco, Calif., Sep. 3, 2001); L. Namburi, “Electrodeposition of nickel/tungsten alloys into deep recesses,” M.S. Thesis (Louisiana State University, Baton Rouge, La., submitted October 2001); and P. Datta et al., “A Microfabricated Recurve Bimetallic Actuator,” abstract for Micromachining and Microfabrication Process Technology VIII Conference (January 2003).
 Incorporating particles such as alumina particles into alloys at the micron scale could provide significantly improved materials for microstructure fabrication. However, there is a need for increased metal deposition rates in such processes, as existing techniques can be quite slow.
 We have discovered a method for the improved electrodeposition of composite metal or composite metal alloy microstructures, by incorporating nanoparticulates into the plating bath, and electrodepositing the metal or metal alloy composites via pulsed plating. Adding even relatively small quantities of nanoparticulates substantially enhances the rate of metal deposition, promotes the deposition of smoother structures, and produces microcomposites having superior properties as compared to otherwise similar metals or metal alloys lacking the particulates—for example, greater hardness. The new techniques are particularly applicable for metal species under mass transport control. The new techniques are especially useful for plating alloys and composites into deep recesses of a microstructure. Prior techniques have not been particularly effective at plating alloys into deep recesses.
 The nanoparticulates increase the rate of electrodeposition for metal ions under mass transport control. As used in the specification and claims, “mass transport control” of electrodeposition refers to conditions in which mass transport is the rate-determining step for the electrodeposition of at least one metal species. As those of skill in the art will appreciate, to say that mass transport is the rate-determining step does not preclude other influences on the reaction rate. The rate-determining step is the slowest step in the reaction, a “bottleneck.” The overall reaction rate cannot be faster than that of the rate-determining step. An incremental change in the rate of the rate determining step will have a substantially greater effect on the overall reaction rate than will a proportionate incremental change in the rate of other steps in the reaction.
 The mass transport rate is determined primarily by diffusion, convection, and migration. “Migration,” in this context, is mass transport of charged species resulting from the application of an electric field. Its effects are generally small in electroplating, because the ionic strength of plating solutions tends to be high to reduce ohmic resistance. By contrast, in deep recesses, convection is low at the bottom of the recess, so diffusion dominates.
 A practical test for whether the electrodeposition of a given metal is under mass transport control under given circumstances is to measure the rate of deposition under otherwise identical conditions, but varying the rate of stirring, or varying the rate of convection. If the process is under mass transport control, then the rate of deposition will be strongly dependent on the rate of stirring or the rate of convection, at least over some range of the stirring rate or the convection rate. If the process is not under mass transport control, then the deposition rate will not be strongly dependent on the rate of stirring or the rate of convection.
 We have discovered that the addition of nanoparticles substantially enhances the rate of electrodeposition in deep recesses, where the electrodeposition is under mass transport control. By contrast, where electrodeposition is under kinetic control (i.e., some reaction step other than mass transport is rate-determining), we have found that adding nanoparticulates can sometimes increase the electrodeposition rate, sometimes decrease it, and sometimes leave it unaffected.
 Unless context clearly indicates otherwise, as used in the specification and Claims a “deep recess” means one having a depth of about 10 μm or greater, preferably about 50-100 μm or greater, for example 200-500 μm, up to about 2000 μm or even greater; and having an aspect ratio greater than about 1, preferably greater than about 5, most preferably greater than about 10.
 To produce an enhanced rate of electrodeposition, it is not necessary to use particle concentrations as high as those that have typically been used in the past for unrecessed geometries. In an unrecessed electrodeposition with particulates, the electrolyte has typically contained particle concentrations on the order of tens or hundreds of grams per liter. While such high concentrations may also be used in practicing the present invention, we have found that the presence of particles enhances the deposition rate at considerably lower concentrations, typically on the order of 1-10 g/L.
 Without wishing to be bound by this theory, it is believed that one reason why the ultrafine particles used in this technique (on the order of 1 μm or smaller) improve the properties of the deposited composite is that the surface area for contact between the particles and the metal is increased. Because the dimensions of microdevices are generally on the order of micrometers, the particles used to make composite microdevices should be at least an order of magnitude smaller, i.e., on the order of hundreds of nanometers or smaller, preferably on the order of tens of nanometers.
 The electrodeposition of a composite material in accordance with this invention may be carried out by suspending particles in the electroplating bath. The particles form part of the resulting composite material. Surprisingly, we have also discovered that the particles actually increase the mass transport rate and hence the metal deposition rate inside deep recesses. This last finding is particularly significant, because electrodeposition inside deep recesses can be very slow, sometimes taking on the order of days to complete. We have found that adding even a small concentration of nanoparticles to the electroplating bath (e.g., ˜3 g/L) can increase the deposition rate by a factor of ˜2-4, as compared to the deposition rate for an otherwise identical process in which the nanoparticles are absent. Furthermore, we observed improved deposit morphology with the incorporation of nanoparticles.
 Without wishing to be bound by this theory, it is hypothesized that another reason why the particles assist in the formation of metal composites or metal alloy composites is that an increase in metal ion concentration in the recesses results from micro-convection induced by the motion of the particles. Particle motion may be attributable to diffusion, convection, diffusiomigration, electromigration, and perhaps to a lesser extent gravity.
 The new techniques are especially useful for forming metal composites and metal alloy composites into deep recesses of a microstructure. Prior techniques have not been particularly effective at forming composites in deep recesses. We believe that we have discovered one of the principal obstacles to forming composites in deep recesses—an obstacle that had not previously been recognized. We have also discovered a pulsed electroplating technique for overcoming this previously unrecognized obstacle.
 The electrodeposition of metals creates local zones of high pH. In those conventional, macroscale, electrodeposition systems in which hydroxyl ions are generated as a side reaction or as a reaction product, convection eddies and other fluid movement tend to sweep away the hydroxyl ions, so that the production of hydroxyl does not adversely affect electrodeposition. We have discovered, however, that in the deep recesses that are typical of high aspect ratio microstructures (e.g., 200 μm, 500 μm, or even deeper), these eddies and fluid movements may be insufficient to remove hydroxyl anions or other reaction or side reaction products. The build-up of hydroxide ions or other products interferes with the electrodeposition. High pH can, for example, cause precipitation from the electrolyte, and the resulting precipitate layer can effectively stop further electrodeposition. We have discovered that proper combinations of electrolyte and pulsed electrodeposition (relatively short current or “on” times, relatively long relaxation or “off” times) can prevent the build-up of excess hydroxyl, even in deep recesses, allowing electrodeposition of much higher structures than is otherwise possible. The relaxation times used in our pulse plating are typically an order of magnitude or more longer than the relaxation times that have been used in conventional pulsed electroplating techniques. In conventional pulsed electroplating on a macroscale, pulsed current promotes small, scattered sites of nucleation, and thus smoother films, or it permits the more efficient transport of reactants, or both. Although transport of reactants is also a consequence of the pulsed electrodeposition used in the present invention, the principal reason for its use is the transport of unwanted reaction products such as hydroxyl ions away from the site of electrodeposition, thereby allowing electrodeposition to continue longer and more efficiently than it otherwise would.
 In practicing the novel process, the duty cycle is preferably less than about 0.2, more preferably less than about 0.125, most preferably less than about 0.1. The on time is preferably greater than about 0.1 second, most preferably greater than about 1 second. Depending upon the particular setup, the on time will often preferably be less than about 60 seconds, as that is often approximately the time that it takes to reach steady state. See generally commonly-assigned U.S. patent application Ser. No. 10/196,764, filed Jul. 17, 2002.
FIG. 1 depicts a typical SEM micrograph of an array of Ni—Cu-γ-Al2O3 alloy microposts electroformed by pulsed deposition.
FIG. 2 depicts a representative analysis of the copper content (mass ratio) of a micro-post cross-section as a function of height, both when nanoparticles were present and when they were absent.
FIG. 3 depicts a typical mass concentration of alumina in a composite micropost, as a function of height.
FIG. 4 depicts nickel deposition rate as a function of electric potential, both with and without particles.
 FIGS. 5(a) and (b) depict photomicrographs of deposits of Ni—Cu alloy and Ni—Cu-γ-Al2O3 composite, respectively.
 FIGS. 6(a) and 6(b) depict the effect of alumina particles on mass transport-limited copper partial current, in acidic (pH 4.0) and basic pH (pH 8.0) electrolytes, respectively.
 FIGS. 7(a) and (b) depict a comparison of the Ni partial current density with and without particles, for the acidic and basic electrolytes, respectively.
FIG. 8 depicts the effect of nanoparticle concentration on deposited copper concentration as a function of height.
 The corrosion resistance of nickel-copper alloys is far superior to that of unalloyed nickel alone. To test the effect of nanoparticulates on alloy deposition rate, we conducted experiments using a bimetallic electrolyte containing Cu(II) and Ni(II). In initial experiments, Ni(II) was present in excess as compared to Cu(II), so that only the Cu(II) reaction was under mass transport control. The mass-transport, diffusion-limited rate represents the effective maximum rate of reaction. Thus the copper species acted as a “probe” to measure mass transport and changes in mass transport. Put differently, the deposition of Ni was kinetically controlled, while the deposition of Cu was mass transport controlled.
 These Ni-rich NiCu alloys were electrodeposited into 500 micron deep recesses in a pattern comprising 187 μm×187 μm squares. The electrolyte contained 1.0 M nickel sulfate (NiSO4.6H2O), 0.04 M cupric sulfate (CuSO4.5H2O), and 0.3 M sodium citrate (Na3C6H5O7.2H2O), with pH adjusted to 8.0 with ammonium hydroxide (NH4OH). The alloy was electrodeposited both in the presence and the absence of 3.125 g/L γ-Al2O3 nanoparticles (a relatively low particle loading), with particles having an average diameter of 32 nm and a specific surface area of 30-60 m2/g, as reported by the manufacturer (Nanophase Technologies Corporation, Inc., Romeoville, Ill.). Current was applied in pulses. The applied current density was 10 mA/cm2, with an on time of 10 seconds and off time of 70 seconds. The applied current density was chosen so that the applied current density at the bottom of the recess was ten times larger than the steady-state copper limiting current density, and one-half of the nickel limiting current density.
FIG. 1 depicts a typical SEM micrograph of an array of Ni—Cu-γ-Al2O3 alloy microposts electroformed by pulsed deposition. FIG. 2 depicts a representative analysis of the copper content (mass ratio) of a micro-post cross-section as a function of height, both when nanoparticles were present and when they were absent. Cu content was enhanced when the nanoparticles were present, particularly at heights above about 300 μm. Without wishing to be bound by this theory, it is believed that as the metal diffusion boundary layer decreased due to motion of the particles, the rate of deposition increased, resulting in an increase in the concentration of copper as height increased. (The bottom of the recess, the substrate, is designated as the 0 μm point, that least accessible to the plating solution. The recess opening near the bulk solution was 500 μm.) Note that Cu is more noble than Ni, so it is plated preferentially, even when Ni is present in large excess. The concentration of Cu in the electrolyte was lowered to create mass transport control for Cu, thereby limiting its electrodeposition rate relative to that of Ni.
 Without wishing to be bound by this theory, it is believed that the difference in behavior above and below about 300 μm was due to diffusion control of the particles' motion. As the particle boundary layer shrinks, its flux increases, and the solution is mixed more.
FIG. 3 depicts a typical mass concentration of alumina in the composite micropost, as a function of height. The weight percentage of alumina was below 10%, usually below 5%, with the highest concentration near the opening of the recess at 500 microns.
 We conducted another set of experiments to rule out the possibility that the observed increase in Cu concentration in the micro-post deposited with particles might have resulted from inhibiting Ni deposition, rather than enhancing deposition of Cu. Electroplating of Ni alone (without copper), in both the presence and the absence of alumina nanoparticles, was conducted in macro-size, rotating cylinder electrodes. This experimental setup ensured that the electrodeposition of Ni was under kinetic control in the recesses. Thus the experiments on the macro rotating recess electrodes could be used to capture the same kinetic control that existed in the micro recesses without rotation. As shown in FIG. 4, at 1000 rpm, with a 1.2 cm long electrode having a diameter of 1 cm, there was no significant difference between the Ni deposition rate with particles and that without particles (neglecting an instability observed at −1.05 V vs SCE). Thus we concluded that the Cu enhancement observed in the micro-deposits described earlier was due to the enhanced reaction rate of the mass transport-controlled Cu, not to any inhibition of plating of Ni.
 We therefore concluded that nanoparticles may be used to enhance mass transport controlled electrodeposition rates.
 We have also observed that a higher particle concentration in the solution (6.25 g/L) further enhances the mass transport rate, and results in a higher particle concentration in the deposit. However, if too many particles are incorporated in the deposit (greater, for example, than about 20 mass % alumina), then internal stress can crack the deposit. In addition, too many particles in the solution may block the electrode surface, thereby inhibiting or terminating deposition.
 Any type of conductive or nonconductive nanoparticle may be used in practicing this invention—e.g., Al2O3, SiC, TiO2, Si3N4, antimony tin oxide, cerium oxide, copper oxide, indium tin oxide, iron oxide, yttrium oxide, zinc oxide, diamond, Cr, Cu, Au, other oxides and ceramics, other metals, etc. Nanoparticulates may be prepared through methods known in the art, or purchased from a commercial source such as Nanophase Technologies Corporation (Romeoville, Ill.).
 Alumina was used in the prototype experiments reported here, primarily due to its high hardness. The particle size used depends, in part, on the geometry of the mold. If the dimensions of the mold are on the order of microns, then submicron particle sizes should be used. As a general rule, the smaller the size of the particles, the better, in order to promote more uniform deposition. Particle sizes on the order of 10-50 nm are generally preferred. However, the invention should also work with larger particle sizes, e.g., on the order of 50-500 nm. The particles should be less than about 500 nm in size.
 To the knowledge of the inventors, no one has previously reported a nickel-copper alloy/alumina composite material, whether as a microstructure or otherwise. As compared to a nickel-copper alloy lacking such particulates, the nickel-copper alloy/alumina composite is expected to demonstrate improved corrosion resistance and improved hardness.
 As those skilled in the art will appreciate, nanoparticle samples will generally not be uniform in size, but instead will typically have a range of sizes. As used in the specification and claims, unless context clearly indicates otherwise, the size of a sample of particles is determined as the number-average diameter of the particles in the sample. To refer to the “diameter” of a particle does not imply that the particle is spherical; rather, the diameter refers to the length of the largest chord passing through the particle. The number-average diameter of a sample of particles may be determined or estimated by a number of methods known in the art including, for example, transmission electron microscopy analysis.
 Other characteristics of the particles, such as composition of the particles, may be chosen depending on the desired properties of the finished composite microstructure. Kerr et al. (2000), and A. Hovestad et al. (1995) briefly summarize such properties for bulk results. No review on the microscale exists, but bulk results on the relation between particle composition and composite properties may be extrapolated in many cases. For example, incorporating SiC particles into Ni has been reported to enhance the wear resistance of Ni. Teflon™ particles (PTFE) have been embedded into Ni to enhance anti-stick and lubrication properties, as have other low-friction particles such as MoS2 and graphite. In general, incorporating oxide particles (such as Al2O3 or TiO2) into metal deposits improves the yield strength and hardness of the metals, and can also decrease corrosion rates.
 Additional experiments were conducted both in recessed microelectrodes, and in rotating cylinder electrodes (RCE).
 Except as otherwise clearly stated, experimental conditions and procedures in all Examples were as described in the present Example 3.
 Recessed electrodes were produced by exposing a 500 μm thick poly (methylmethacrylate) (PMMA) sheet, attached to a copper sheet 5.0 cm×5.0 cm×1 mm thick, to x-ray radiation at the LSU-CAMD synchrotron facility (Baton Rouge, La.). To promote adhesion to the PMMA sheet, the copper surface was first oxidized by placing it in a solution containing 54.26 g/L NaClO2, 68 g/L NaOH, 11.44 g/L Na2CO3 and 4.67 g/L NaCl at a temperature of 95° C. for 20 minutes, followed by rinsing with deionized water and air-drying. The adhesive comprised methyl methacrylate (MMA) (0.919 mass-fraction) and dimethyl aniline (DMA) (0.011 mass-fraction) mixed together first, followed by the addition of powdered PMMA (0.055 mass fraction), and benzoyl peroxide (BPO) (0.0164 mass fraction), stirred for 6-8 hours. The adhesive was used to join the copper to the PMMA, and pressure was applied immediately to the substrate, and the adhesive was allowed to cure overnight.
 The bonded PMMA/Cu sample was exposed to collimated X-rays (wavelength 4 Å) at the XRLM-2 and XRLM-3 beamlines at CAMD using a tantalum mask and a 16 μm thick Al filter. The mask was patterned with 187×187 μm squares, spaced 20 μm apart from one another. The exposed electrode regions were then developed and rinsed. The developing solution contained diethylene glycol butyl ether (60 volume %), morpholine (20 volume %), 2-aminoethanol (5 volume %) and deionized water (15 volume %). The rinsing solution contained diethylene glycol butyl ether (80 volume %) and deionized water (20 volume %). The developing and rinse procedure included four cycles of 10 minutes each in the developer, 2 minutes in a pre-rinse and 60 minutes in the rinse. (The pre-rinse consisted of diluted older rinse solution or deionized water. Immediately before electrodeposition, the copper oxide at the bottom of the recess was etched away with a solution containing 0.5 M KCl and 0.5 M HCl.
 The cylinder electrodes used in the RCE experiments were 1.2 cm long and 1 cm in diameter, made of 410-stainless steel plated with a 0.5 μm layer of gold from a commercial electrolyte (Technic, Inc., Techni Gold 25 E (Providence, R.I.)). A single-compartment cell was used, with a saturated calomel reference and a nickel sheet as the counter electrode.
 Current was controlled during deposition either with an EG&G Princeton Applied Research Potentiostat/Galvanostat (Model 363) coupled to an EG&G PARC Universal Programmer (Model 175), or with an Amel Instruments Potentiostat (Model 2051) coupled to a Wavetek DDS function generator (Model 29) for pulsed current and DC waveforms. Polarization curves were corrected for ohmic drop by impedance analysis with a Bas-Zahner IM6 system. Scanning Electron Microscopy (SEM) with energy dispersive X-ray spectroscopy (EDS), JEOL 840A, and wavelength dispersive X-ray spectrometers (WDS), JEOL JXA-733, were used to obtain topographic and chemical information for the deposits. Both EDS and WDS were used to analyze the composition of the alloy.
 A citrate-based electrolyte was used. The nickel concentration (1.0 M nickel sulfate) was in large excess as compared to the concentration of copper (0.004 M or 0.04 M cupric sulfate), in 0.3 M sodium citrate. Two pHs were studied: pH 4.0 (obtained by adjusting with H2SO4), and pH 8.0 (obtained by adjusting with NH4OH). Alumina; particles (1.625-12.5 g/L) having an average diameter of 32 nm, with a specific surface area of 30-60 m2/g as reported by the manufacturer (Nanophase Technologies Corporation, Inc., lot #A70927-01) were added to the electrolyte. The lower concentrations of alumina (1.625, 3.125, and 6.25 g/L) was used in depositions into the recessed electrodes. The higher concentration (12.5 g/L) was used in the rotating electrode experiments to exaggerate any enhancement or inhibition effects.
 Galvanostatic DC plating (5-25 mA/cm2) in the recessed electrode caused precipitation in the recess, blocking the surface and preventing growth of the deposit. On application of direct current, the potential rose substantially after a few seconds (data not shown). The resulting rise in surface pH caused precipitation. Additionally, at higher applied currents, gas evolution from the hydrogen side reaction also caused electrode blockage and a loss in efficiency. To overcome these problems, we applied current in pulses, with long relaxation times, thus minimizing the rise in surface pH, and allowing the products of side reactions to diffuse out of the recess. See generally commonly-assigned U.S. patent application Ser. No. 10/196,764, filed Jul. 17, 2002.
 Deposits of Ni—Cu alloy and Ni—Cu-γ-Al2O3 composite are shown in FIGS. 5(a) and (b), respectively. In all cases, the applied current density was 10 mA/cm2, with an on time of 10 seconds and off time of 70 seconds. FIG. 5(a) shows a Ni—Cu micropost deposited from an acidic plating bath (pH 4.0). FIG. 5(b) shows a micropost deposited when alumina was present in an otherwise identical electrolyte. At pH 4.0, the surface morphology of the deposits obtained with the nanoparticles was better, i.e., smoother and less nodular, than that obtained without nanoparticles. The nodules present on the surface of the Ni—Cu alloy were reduced by the presence of Al2O3 nanoparticles.
 X-ray maps of sections of the posts (not shown) were used to investigate the distributions of the components of the composite microposts as a function of height. The concentration of copper showed an increase along the height of the micropost, i.e., the copper concentration was higher at the top of the micropost, i.e., the-region more accessible to the plating solution, than at the bottom. Such a graded concentration was expected due to the decreasing boundary layer thickness as the deposit grew for the diffusion-limited copper. The nickel composition showed a corresponding drop, while the distribution of alumina particles appeared to be almost random, with distinct regions of agglomeration. Similar results were seen in a back-scattered electron image (not shown).
 We compared elemental compositions in the deposits, both with and without alumina on a rotating cylinder electrode. FIGS. 6(a) and 6(b) show the effect of alumina particles on the mass transport-limited copper partial current, obtained from the acidic (pH 4.0) and basic pH (pH 8.0) electrolytes, respectively. At the lower pH, the concentration of copper was enhanced at both the top and bottom of the recess by the nanoparticles (data not shown). At the higher pH, enhancement of copper concentration was seen primarily from the middle to the top of the recess; the concentration of copper in the bottom portion was not strongly affected by the particles (data not shown).
 The observed enhancement in copper concentration might have been the result either of an inhibited nickel deposition rate, or an enhanced copper diffusion rate. To help distinguish these two possibilities, experiments with rotating cylinder electrodes were conducted. FIGS. 7(a) and (b) show a comparison of the Ni partial current density with and without particles, for the acidic and basic electrolytes, respectively. In the acidic electrolyte, for a given applied current, the polarization curves were shifted to more negative potentials in the presence of alumina. Since the slopes of the two nickel partial current densities were similar, the inhibition mechanism is inferred to be the result of surface site blockage by the alumina particles, and not by an alteration of the nickel reaction mechanism. By contrast, in the basic electrolyte no such inhibition was seen. The copper enhancement observed in the bottom of the recess in the low pH bath can be explained by the inhibition of Ni deposition. In the basic electrolyte, which did not exhibit inhibition of the nickel, no such enhancement was seen at the bottom of the recess.
 In either case, it appears that the mass transport-controlled deposition of copper was enhanced by the particles, while the particles could inhibit kinetic-controlled deposition of nickel, depending on the pH. However, even Ni, if under conditions of mass transport control, should show an enhanced deposition rate at pH 4 in the presence of nanoparticles.
 Unlike the experiments carried out with the rotating electrode, the addition of a small concentration of particles to the electrolyte caused a dramatic increase in the Cu concentration in the recess. The principal difference between the two situations is believed to be the convective hindrance of the recess. It is believed that the Cu concentration in the recess was increased by micro-convection induced by the particles during both the on-time and the off-time of the pulses. The improved convection, i.e., the decreased boundary layer thickness, produced a larger copper limiting current density and an associated increase in the copper deposited in the composite. By contrast, similar behavior was not seen in the rotating cylinder electrodes because the forced macro-convection in the latter system swamped the micro-convection induced by the particles.
 The preferred pH depends on the particular metal or alloy being deposited, and which species are kinetically controlled and which are mass transport controlled. The preferred pH for a particular system may readily be determined through routine experimentation. For example, for NiCu alloys, our results gave a preferred pH range of about 8-10.
 A series of additional rotating cylinder electrode experiments was conducted, using the same electrolyte as described in Example 3. Some observations from these experiments were the following: The current efficiency depended on the applied current density. In general, current efficiency was lower at lower applied current density. Further, it was found that below about 20 mA/cm2, the addition of alumina particles lowered the current density somewhat; at higher current densities, the presence of particles did not seem to have a pronounced effect on current efficiency. At lower current densities, below about 20 mA/cm2, a copper-rich alloy was deposited, while at higher current densities, a nickel-rich alloy was deposited—this effect did not seem to be affected strongly by the presence or absence of particles, and was consistent with mass transport control of copper deposition and kinetic control of nickel deposition. Particle concentration in the composite did not vary strongly with applied current density, consistent with incorporation controlled by particle transport.
 The effect of nanoparticle concentration was examined in another series of experiments. Conditions were otherwise as described above for Example 1, except that the concentration of alumina nanoparticles was varied: 0, 1.625, 3.125, or 6.25 g/L. As shown in FIG. 8, the results were hardly distinguishable between the no alumina trial and that employing 1.625 g/L of alumina. By contrast, a substantial enhancement in the concentration of copper was seen with 3.125 g/L alumina, at heights above about 300 μm. Even greater enhancement of copper concentration was seen at 6.25 g/L alumina, at heights above about 225 μm. Thus there appears to be both a lower limit for effective enhancement, and an increase in enhancement with increasing particle concentration (at least up to a point). The lower limit for effective enhancement, and the height above which enhancement occurs for a given particle concentration, will depend on the specific configuration and conditions used, and may readily be determined through routine experimentation for any given configuration and set of conditions. In this particular example, the lower limit for copper enhancement lay somewhere between 1.625 and 3.125 g/L of nanoparticular alumina. (Different particle shapes might behave differently, a question that we have not yet tested. The particles used in these experiments were approximately spherical, as verified by TEM analysis.) As shown in FIG. 8, an increase in particle concentration (above the lower limit) resulted in an increased concentration of deposited copper. Although deposition rates were not measured directly in this experiment, we inferred that the increased concentration of deposited copper was the result of an increased rate of deposition induced by the particles.
 To the inventors' knowledge, no one has previously reported the electrodeposition of Ni—Cu alloys or Ni—Cu-γ-Al2O3 nanocomposites in a deep recess micro-geometry. To the inventors' knowledge, no one has previously reported the electrodeposition of metal composites or metal alloy composites in a deep recess micro-geometry using pulsed current. To the inventors' knowledge, no one has previously reported that the inclusion of nanoparticles in an electrolyte solution can substantially enhance the rate of electrodeposition of a metal ion under mass transport control.
 We obtained a compositionally-graded alloy and composite. The gradient of the diffusion-limited copper is believed to have resulted from the decreasing thickness of the boundary layer as the material grew over time. Including alumina particles in the electrolyte resulted in higher copper concentrations in the deposits. The increased copper concentration in the recess was attributed to enhanced diffusion, or to a decrease in the effective boundary layer thickness in the recess due to the presence of nanoparticles and an inhibition of the nickel deposition rate at lower pH. Our results suggested that both kinetics and mass transport play a role in the deposit of metals in the microposts. Additionally, the microposts we obtained in the prototype experiments had a smoother morphology when electroplating was conducted in the presence of nanoparticles.
 Although certain embodiments were described above, those of skill in the art will appreciate that the present invention is not limited to those embodiments. For example, the invention may be used to produce microstructures from a wide variety of metals or metal alloys, including, for example, the following metals and alloys of the following metals: nickel, copper, iron, tungsten, cobalt, molybdenum, gold, zinc, silver, chromium, tin, and lead.
 The complete disclosures of all references cited in this specification are hereby incorporated by reference. Also incorporated by reference are the following instances of the inventors' own work, none of which is prior art to this application: A. Panda et al., “Nanoparticles to improve mass transport inside deep recesses,” Manuscript submitted to Electrochemical and Solid State Letters (2003); A. Panda et al., “Electrodeposition of Ni—Cu-γ-Al2O3 Alloys into Deep Recesses,” talk given at 202nd Electrochemical Society Meeting (Salt Lake City, Utah, Oct. 24, 2002), proceedings paper expected to be published in 2003 after the filing date of the present invention; A. Panda et al., “Electrodeposition of Ni—Cu—Al2O3 Nanocomposites,” poster presentation at the Nanocomposite Workshop at Argonne National Laboratories (Chicago, Ill, Mar. 28-29, 2002); A. Panda et al., “Electrodeposition of Ni—Cu-Alumina Nanocomposites,” poster presentation at the Center for Advanced Microstructures and Devices (CAMD) Users Meeting (Baton Rouge, La., Apr. 19, 2002) (open only to users of CAMD facility, not to the public); and A. Panda et al., “Electrodeposition of Ni—Cu—Al2O3 Nanocomposites,” Abstract both of the poster presentation at Argonne National Laboratories previously cited, and of the poster presentation at CAMD previously cited. The last item cited, the Abstract, was first submitted on Feb. 15, 2002 to CAMD, and it is believed that the Abstract was available to the public shortly after the Feb. 15, 2002 CAMD submission date, and that the abstract from the Argonne meeting was only available as of the Mar. 28-29, 2002 meeting.
 In the event of an otherwise irreconcilable conflict, however, the present specification shall control.
 The inventors also note that they presented portions of this work orally at a Gordon Conference in New London, N.H. on Aug. 11, 2002, “Electrodeposition of Ni Alloys and Nanocomposites for MEMS.” No publications resulted from the Gordon Conference presentation. Presentations made at a Gordon Conference are understood to be private communications to those persons who are in attendance, and that the information presented is not for public use.