MENISCUS-CONFINED THREE-DIMENSIONAL ELECTRODEPOSITION

Abstract

The invention relates to a process and a system for 3-dimentional (3D) fabrication of sub-micron structures and is established by local electrochemical deposition methods.

Claims

1. A meniscus-confined electrochemical deposition method, the method comprising dispensing through a deposition tool an amount of an electrolyte solution containing a reducible material onto a surface of a substrate, causing said reducible material undergo reduction, measuring a change in at least one parameter selected from a distance between the deposition tool and the surface and a change in a force applied on the deposition tool once the material is deposited, such that when a change in the distance or force is measured/detected, the position of the deposition tool and/or the substrate is modified with respect to the substrate or the deposition tool, and repeating the method one or more times to afford a deposited object on the surface; wherein the method is carried out under air or under an inert gas or wherein the method is carried out while the substrate is immersed completely or partially in an electrolyte bath.

2. (canceled)

3. (canceled)

4. The method accoridng to claim 1, the method comprisng (a) providing a deposition tool in a form of an electrolyte solution reservoir comprising a reduciable form of at least one material, the deposition tool having an end tip for dispensing an amount of the electrolyte solution, (b) positioning the tip at a distance from the surface of the substrate, (c) dispensing a first amount of the electrolyte solution onto the surface of the substrate, thereby forming a liquid bridge between the tip and the substrate's surface, and (d) causing reduction of the reducible form of the at least one material in the liquid bridge.

5. The method according to claim 4, further comprising measuring a change in the force applied on the tip, and modifying the tip-to-substrate distance by normal or lateral movement.

6. The method accoridg to claim 4, the method comprising (a) providing a deposition tool in a form of an electrolyte solution reservoir comprising a reducible form of at least one material, the deposition tool having an end tip for dispensing an amount of the electrolyte solution, (b) positioning the tip at a distnace from the surface of the substrate, (c) dispensing a first amount of the electrolyte solution onto the surface of the substrate, thereby forming a liquid bridge between the tip and the substrate's surface, (d) causing reduction of the reducible form of the at least one material in the liquid bridge, (e) measuring or detecting a change in the distance between the tip apex and the substrate/deposition front, (f) once a change in the distance is detected, modifying at least one positional parameter associated with the tip, while dispensing a further amount of the electrolyte solution onto the reduced material, thereby forming a liquid bridge between the tip and the reduced material, and (g) repeating steps (d)-(f) one or more times.

7. The method according to claim 6, wherein the at least one positional parameter associated with the tip is a tip-to-substrate distance and/or the tip lateral position.

8. The method according to claim 6, wherein the change in distance is determined by measuring the actual change in the distance or by measuring a change in the applied force on the tip.

9. The method according to claim 1, wherein the deposition tool has a dispensing tip in the form of a micropipette with a microscopic or nanoscopic opening, said opening being optionally between 40 nm and 5 μm.

10. The method according to claim 1, wherein the deposition tool comprises a plurality of reservoirs, each reservoir having different or independent dispensing tips.

11. The method according to claim 1, wherein the deposition tool comprises a plurality of reservoirs, at least a portion or all of said plurality of reservoirs being connected to a single dispensing end.

12. The method according to claim 9, wherein the micropipette is an AFM tip.

13. The method according to claim 9, wherein the micropipette is in a form of a hollow glass tube.

14-24. (canceled)

25. The method according to claim 1, for fabricating nanowires, high-density interconnects, sub-micron scale circuitry, conductive bridges and precise electrical connections, thermocouples, interposers, high-frequency terahertz antennas, probe arrays and precision sensors; for fabricating micro- or nano-electromechanical systems, batteries and fuel cells; or for repairing or modifying micro-sized or nano-sized features.

26-27. (canceled)

28. A printing system comprising a liquid deposition tool, a closed-loop feedback control, an environmental chamber, a source meter, and optionally a visualization system, wherein the closed-loop feedback control comprises a force/distance meter, wherein the deposition tool is in a form of an electrolyte solution reservoir having an end tip for dispensing an amount of the electrolyte solution; and wherein the force/distance meter is functionally associated with the deposition tool for measuring a change in the distance between the deposition tool and the surface or a change in a force applied on the deposition tool end tip.

29-33. (canceled)

34. The system according to claim 28, wherein the tip is in a form of a micropipette.

35. The system according to claim 34, wherein the micropipette is an AFM tip.

36-42. (canceled)

43. The system according to claim 28, wherein the force sensor is a tuning fork.

44. The system according to claim 28, wherein the force sensor is a deflection sensor.

45. The system according to claim 44, wherein the deflection sensor is an optical deflection sensor or a piezoresistive or piezoelectric deflection sensor.

46. A 3D printing system, the system comprising a closed-loop feedback control, a deposition tool, and means for modifying the position of the deposition tool with respect to a substrate or a feature on the surface of the substrate, the closed-loop feedback control comprising means for detecting a change in a force imposed on the deposition tool by the substrate or a feature formed on the substrate, or a change in the distance of the deposition tool from the substrate or from a feature formed on the substrate, such that upon detecting a change in the force or distance, the position of the deposition tool is modified vertically or laterally with respect to the substrate or the feature formed on the substrate.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0197] In order to better understand the subject matter that is disclosed herein and to exemplify how it may be carried out in practice, embodiments will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:

[0198] FIG. 1 is an illustration of an exemplary system used in accordance with the invention. The AFM is positioned inside an acoustic chamber, which serves as an environment chamber. The relative humidity percentage is regulated by passing air through a flask containing water directly to the chamber. The humidity inside the chamber is monitored with a thermohygrometer. The environment chamber is positioned on an anti-vibration table. An optical objective lens with a CCD camera is mounted above the tip in order to monitor its position in a continuous manner. The substrate is positioned on top of the piezo driver, which retracts the tip according to the feedback loop. A course stepper motor brings the tip to an initial connection with the substrate. A platinum wire is inserted to the back of the tip and is connected together with the conductive substrate to a SourceMeter. The inset shows a zoom-in of the pipette apex in proximity to the substrate, the liquid meniscus between them, and the reduction of the cations taking place.

[0199] FIGS. 2A-B show: FIG. 2A shows three types of hollow AFM tips: [1] Illustration of a beam-bounce tip glued to a magnetic holder. A gold coating is applied by chemical vapor deposition (CVD) from the cantilever nearly all the way to the tip end. A platinum wire is inserted through the backside. [2] Illustration of a lateral tuning fork (LTF) in which the pipette is attached to the tuning fork on its side. [3] An image of a normal tuning fork (NTF) in which the pipette is attached to the bottom cantilever of the tuning fork. FIG. 2B shows a scanning electron microscope (SEM) image of a 1.5 μm in diameter copper pillar with an illustration of the tip apex and the liquid meniscus connected to it. The electrochemical setup contains the AFM tip, filled with the ionic solution (CuSO.sub.4), a polished copper foil that acts as the working electrode, and a platinum wire inserted to the back of the tip acting as the counter electrode. Owing to the fact that the maximum currents applied in these experiments were around a few nano amperes (nA), the polarization of the platinum wire could be considered negligible, and it was referred to as a pseudo-reference electrode. The printing begins when the tip approaches the substrate while current is applied. When a connection is recognized and a liquid meniscus is formed, the tip is stopped by the AFM closed-loop control. At this point, a closed electrical circuit is established (as evident from a sharp change in potential/current), and reduction of the metal ions is initiated, confined to the meniscus area. While deposition continues in the vertical direction, the force applied on the tip changes. As a result of a change in the tuning fork, signal is recognized, and an in situ correction of tip-substrate distance takes place by the AFM closed-loop control. As 3D printing continues, the tip-to-substrate distance changes at the same rate, maintaining a constant meniscus shape and a uniform cross-section, as shown in FIG. 2B. The liquid bridge (i.e., meniscus) formed by dispensing an electrolyte solution from the micropipette tip onto the surface has a diameter that substantially equals the inner diameter of the tip, and a height defined by the distance between the substrate surface or a site of further deposition (in a sequential off-substrate deposition step) and the tip. The structure of the deposited feature as well as the resolution are mostly dependent on the tip apex shape and size. Using tips with smaller inner diameters will result in pillars with smaller cross-sections.

[0200] FIGS. 3A-H show: (FIG. 3A) Potential transient during galvanostatic printing (I=−5 nA) of a Cu pillar from a micropipette with an inner diameter of 1 μm. (FIG. 3B) SEM side image of the self-supported straight pillar from (A) on a polished Cu foil substrate. Note the uniformity and outstanding surface finish, an important advantage of electrochemical 3D printing over conventional AM technologies such as PBF and DED. (FIG. 3C) SEM top view of the same pillar. Note the high density and lack of evident defects, an important advantage of electrochemical 3D printing over conventional AM technologies such as PBF and DED. (FIG. 3D-E) SEM images of Cu pillar printed at I=−1.5 nA from a micropipette with an inner diameter of 400 nm. (FIG. 3F) SEM image of two Cu pillars printed at I=−400 pA from a micropipette with an inner diameter of 100 nm. The inset shows a zoom-in image of the upper part of one pillar. (FIG. 3G) SEM image of an overhang pillar printed with a 1 μm tip and a constant lateral movement speed of 400 nm/s, resulting in an overhang angle of 80°. Lateral motion started after 1 min of vertical printing. One of the advantages of electrochemical 3D printing over conventional AM technologies could be the ability to manufacture items with overhang angles higher than 45° from the vertical, which is usually regarded as the threshold value, above which supports are needed in PBF processes. (FIG. 3H) Automatic printing of a 3×3 matrix of 1.5 μm pillars. Deposition current: −5 nA. Note that the high-aspect-ratio pillars in FIG. 3 neither deflect nor buckle under their own weight (i.e., the pillars are relatively stiff compared with their own weight) because, on small length scales—body forces are relatively ineffective at deforming structures compared to surface forces.

[0201] FIGS. 4A-G present characterization of the chemical composition, microstructure, surface and cross-section of printed copper pillars. (FIG. 4A) SEM image of the cross-section extracted by focused ion beam (FIB)-SEM from of a 1.5 μm printed Cu pillar. (FIG. 4B) SEM image of the cross-section extracted by FIB-SEM from of a 500 nm printed Cu pillar. (FIG. 4C) The outer circumference of the pillar in (FIG. 4A). Comparing the surface finish of the laid pillar to that of the pure copper substrate polished down to 40 nm colloidal silica illustrates one of the advantages of electrochemical 3D printing—outstanding surface finish that cannot be achieved by conventional AM technologies such as PBF and DED. (FIG. 4D) Chemical analysis by energy dispersive X-ray spectroscopy (EDS), showing that the printed pillar is made of pure copper. (FIG. 4E) Scanning transmission electron microscope (STEM) image of a 500 nm in diameter pillar. (FIG. 4F) STEM electron diffraction pattern of (FIG. 4E). The four diffraction rings match, from inside outside, the (111), (200), (220) and (311) lattice planes of face-centered cubic (FCC) copper. No preferred orientation is evident in the z-direction. (FIG. 4G) Transmission electron backscatter diffraction (t-EBSD) inverse pole figure, z-direction coloring scheme, from the transverse cross-section of a 500 nm in diameter copper pillar. No preferred orientation is evident.

[0202] FIG. 5 shows a longitudinal cross-section from a 500 nm in diameter, 20 μm high printed copper pillar. On the right is a STEM image of the pillar, divided to four parts in order to avoid collapse. Corresponding SEM images of the four parts are shown on the left. It is evident that the material is fully dense, and that there is some change in the grain size from the central axis of the pillar (where it is smaller) to the surface (where it is larger). The second characteristic is probably the result of recrystallization at the pillar's surface over time, which is typical of nanocrystalline copper structures at ambient conditions as a result of the high energy stored. FIG. 5 reveals another advantage of electrochemical 3D printing over conventional AM technologies such as PBF and DED—materials printed with these conventional technologies are often not uniform in the z-direction (mainly, due to thermal effects).

[0203] FIGS. 6A-B show the copper grain size distributions based on t-EBSD analysis of: (FIG. 6A) transverse cross-section of a 500 nm printed copper pillar (see FIG. 4B), and (FIG. 6B) longitudinal cross-section of the same pillar (see FIG. 5). The difference in the average grain size may indicate a non-equiaxed grain structure. Using the Hall-Petch relationship, the yield strength of this copper pillar is estimated at 537 MPa, much higher than typical values for wrought copper C11000, either in the annealed or in the extra hard (H06) conditions. The high strength of the as-printed copper pillars may be attributed to their sub-micron grain size.

[0204] FIGS. 7A-B show the four-point probe test assembly for electrical resistivity measurements and typical characteristics of the as-printed pillars. (FIG. 7A) SEM image of the four-point probe test assembly prepared in the FIB microscope. [1] A silicon oxide substrate. [2] A 500 nm in diameter, ˜20 μm long copper pillar positioned on the substrate. [3] The pillar is connected to four gold pads via platinum wires. After positioning the four probes, potential is applied while measuring current and potential through the outer and inner probe pairs, respectively. The average resistivity is calculated based on five measurements and the distance between the two inner platinum probes. (FIG. 7B) Four-point probe plot drawing the applied and measured potentials (left and right ordinates, black and blue graphs, respectively) versus the measured current. A linear, Ohmic behavior is evident. For the measured potential: slope=4.01Ω, R.sup.2=0.99964. For the applied potential: slope=4.54Ω, R.sup.2=0.99998.

[0205] FIG. 8 provides a Table summarizing the effects of the micropipette inner diameter on the printed copper pillar diameter, deposition current, deposition current density, and average deposition rate. As the pipette orifice diameter is reduced, both the printed pillar diameter and the associated deposition current decrease, whereas the current density and the related deposition rate increase. For a smaller-sized meniscus it is found that a higher current density can be applied while avoiding hydrogen evolution, probably as a result of a higher evaporation rate from the meniscus surface that causes higher mass transport within it. The maximum 3D printing rate monitored so far has been 300 nm/s (1,080 μm/h). This deposition rate is sufficiently high for many industrial applications.

DETAILED DESCRIPTION OF EMBODIMENTS

[0206] In the work leading to the development of the technology disclosed herein, the inventors developed a 3D electrochemical deposition system having a sub-micron resolution and based on an AFM together with the MCED method. Taking advantage of the AFM closed-loop control, this invention cancels the need for trial-and-error experiments, as the feedback loop recognizes deposition height and changes the distance as the experiment takes place. The printing method uses a special AFM tip made from hollow borosilicate glass (or quartz), which is used as the solution provider as well as part of the mechanical sensing system. Using this scheme, the inventors successfully demonstrated printing of different sized straight and overhang copper self-supported pillars on a polished copper substrate.

[0207] A system according to the invention is constructed around either a commercialized or a homemade AFM, may include a vibration isolation table/plate, an environmental chamber, a sensitive SourceMeter, and a visualization system, as seen in FIG. 1. In order to achieve successful 3D printing, the system should have high sensitivity to forces applied on the tip, fast response, and a rigid structure. Thus, the tip cantilever does not deflect by the weight of the solution or the surface tension force (snap on/off). A hollow glass AFM tip connected to the bottom cantilever of a tuning fork (see FIG. 2A [3]) and working in tapping mode with phase feedback was found effective in meeting all operational requirements, resulting in uniform deposition.

[0208] In order to deposit 3D copper features, the AFM tip was filled with an acidic (pH=1) aqueous electrolyte solution consisted of 50 mM copper (II) sulfate pentahydrate (CuSO.sub.4.5H.sub.2O) and 50 mM sulfuric acid (H.sub.2SO.sub.4). Before inserting to the pipette, the solution was filtered through a 0.2 μm syringe filter, in order to prevent clogging by large particles. When using micropipettes with an orifice smaller than 500 nm in diameter, the solution was filtered through a 100 kDa centrifugal filter. A two-electrode configuration was used. The counter electrode was either a 25 μm or 50 μm in diameter 99.99% pure platinum wire. Because the highest applied currents were of the order of only few nano-amperes, the polarization of the platinum wire can be assumed negligible, and it can be regarded as a pseudo reference electrode. The substrate (working electrode) was 675 μm thick, 99.9% pure copper foil. The foil was first ground with SiC papers, from 240 P down to 2400 P. Next, it was polished with a 1 μm diamond suspension, followed by 40 nm colloidal silica suspension. Then, it was rinsed with water, placed in a chemical glass with deionized water, sonicated for 5 min, rinsed again with water, rinsed with ethanol, and dried with cold blowing air. After electrochemical printing, the substrate was cleaned with a droplet of ethanol and dried again. A current (or potential) was applied between a platinum wire and a copper foil substrate using a SourceMeter, which was controlled via a freeware software. A potential of ca. −1.0 V (vs. Pt) was found to give good deposition rate while avoiding hydrogen evolution, which might cause instability of the meniscus. The relative humidity (RH) was maintained constant between 60% and 70%. The RH affects the wettability properties of the surfaces as well as the evaporation rate of water from the liquid bridge, which both influence the quality of the printed item. Using hollow borosilicate glass pipettes with orifice diameters of 1 μm, 400 nm, and 100 nm, printed pillars with diameters of 1.5 μm, 500 nm, and 250 nm, respectively, as can be seen in FIGS. 3 and 4. The printed pillars had uniform and exceptionally smooth surfaces, and reached a height-to-diameter ratio (i.e., aspect ratio) of more than 100.

[0209] In order to study the capability to print overhang structures, a designated feature in the AFM software was exploited. 3D printing of overhang pillars took place by keeping the closed-loop control “on” and adding a horizontal movement in a predefined speed and path. FIG. 3G presents a structure printed in one continuous motion at a constant lateral movement speed of 400 nm/s, resulting in an overhang angle of 80°. Lateral motion started after 1 min of vertical printing. As the lateral speed was decreased from 400 to 100 nm/s, the overhang angle decreased from 80° to 40°. Lateral speeds higher than 400 nm/s caused the meniscus to break. Owing to the meniscus shape, it is more difficult to maintain a steady connection while moving horizontally, especially when there is no closed-loop control that corrects the tip position.

[0210] Examining the printed structures, it is clear that they possess smooth and uniform surfaces (FIGS. 3B & 4C) attributed to the high and constant current density in addition to the closed-loop controlled motion that maintains a constant distance between tip and substrate. The composition as well as the inner structure of the pillars were characterized. Transverse and longitudinal cross-sections were prepared with a FIB. Examining the different cross-sections (FIGS. 3C,E, 4A,B,E and 5), it can be seen that the electrochemical 3D printing led to dense features, with almost no porosity detected. Samples for EDS were prepared by FIB microscope down to a thickness of several microns, and were kept in a desiccator until characterization. EDS analysis (FIG. 4D) indicated chemical composition of pure copper (Pt peaks are also evident, but they are related to sample preparation).

[0211] Thin samples (˜70 nm thick) were prepared by FIB, were placed on a TEM copper grid, and were used for t-EBSD (also known as transmission Kikuchi diffraction, TKD) characterization inside the SEM as well as for STEM characterization, FIG. 4A,B,E.

[0212] Electron diffraction conducted in the STEM (FIG. 4F) supports the EDS results, showing a rings pattern corresponding to the lattice planes of the FCC copper structure. t-EBSD measurements showed that the inner structure of the 500 nm pillar had no preferred orientation, and that the average grain size was 62±24 nm (n=45). Measuring the average grain size manually on microscope images, using the Heyn lineal intercept procedure, gave a similar, yet more accurate, value of 49±3 nm. In comparison, the corresponding value for the larger (1.5 μm in diameter) cross-section was much higher: 158±13 nm. The larger grain size can be attributed to the lower deposition current density (282 vs. 500 mA/cm.sup.2). Substituting σ.sub.0=40 MPa and k=0.11 MPa m.sup.1/2 into the Hall-Petch relationship, we estimate the yield strength of copper in the 1.5 μm and 500 nm pillars to be 317 MPa and 537 MPa, respectively.

[0213] A longitudinal cross-section of a 20 μm high, 500 nm in diameter, printed pillar was also characterized, FIG. 5. It is evident that the material is fully dense, and that there is some change in the grain size from the central axis of the pillar (where it is smaller) to the surface (where it is larger). The second characteristic could result either from higher ion concentration at the liquid/solid interphase, which causes larger grain growth, or from recrystallization at the pillars surface over time, which is typical of nanocrystalline copper structures at ambient conditions as a result of the high energy stored. Since no grain growth was observed near the circumference of the transverse cross-sections that were stored for shorter times before FIB-SEM characterization, it is believed that recrystallization is more likely the cause of this grain growth here. Both SEM and t-EBSD measurements yielded grain size of 85±25 nm (n=58), i.e., significantly larger than in the transverse cross-section. This may indicate a non-equiaxed grain structure. As shown in FIG. 3B,E,G, the base of the printed pillar is of larger diameter. The larger base size can be attributed to a small tip-to-substrate distance and approach speed, as well as to the hydrophilic nature of copper. The grains in this zone are also larger, which can be related to less cathodic deposition potential/current density (see FIG. 3A) and, hence, lower nucleation rates and higher grain growth rates.

[0214] Measurement of the electrical resistivity of the printed structures is of interest for many applications. In addition, high electrical conductivity of the as-printed feature could indicate on a fully dense material, with a low concentration of defects such as impurities and defected interfaces between printed layers. The assembly for four-point probe measurements is shown in FIG. 7A. The potential vs. current plot for a 500 nm in diameter, ˜20 μm long pillar (FIG. 7B) reveals an Ohmic behavior (i.e., metallic nature) with a slope of 4.01Ω. The average resistivity is 3.15×10.sup.−7 Ω.Math.m, an order of magnitude higher than pure bulk oxygen-free high copper in its annealed condition at 20° C. This, however, may be related to the Pt contact resistance (that generates heat) rather than to an inherent property of printed copper. Moreover, this resistivity value is similar to that reported for nano-grained copper pillars, and could thus be related to electron scattering by high density of grain boundaries. Pulsed electrodeposition (PED) or higher current densities can be used to obtain a structure of high-density nanoscale twins, which exhibits superior mechanical and electrical properties compared to coarse-grained and nano-grained structures.

Experimental Section

[0215] Microprinting System: As a proof of concept, the system was constructed around an AFM (Multiview 1000, Nanonics, Jerusalem, Israel). The AFM is controlled via a LabVIEW-based designated software (NWS, Nanonics, Jerusalem, Israel). The AFM piezo motor has a range of 80 μm in the x and y axes, and 65 μm in the z-axis. The AFM was positioned inside an acoustic chamber (PicolC, Molecular Imaging, Phoenix, Ariz., USA), which served as an environment chamber, maintaining constant relative humidity of 60-70%. The relative humidity percentage was regulated by passing air through a flask containing water directly to the chamber. The humidity inside the chamber was monitored with a thermohygrometer (608-H1, Testo, Lenzkirch, Germany), while the flow control was done manually. The environment chamber was positioned on an anti-vibration table (78-227-12R/CleanTop® II, TMC/Ametek, PA, USA). The printed pillars were too small for in situ optical monitoring. However, an optical objective lens (zoom 6000, Navitar, New York, USA) with a 3.2-megapixel digital camera (ColorView 2, Olympus, Tokyo, Japan) was mounted above the tip in order to monitor its position in a continuous manner Illustration of the microprinting system is given in FIG. 1.

[0216] Micropipettes: Specially designed AFM tips (Nanonics, Jerusalem, Israel) with pipette orifice diameters of 1 μm, 400 nm, and 100 nm were used. The origin borosilicate glass or quartz tubes were heated and pulled using a laser puller (P-2000, Sutter, Calif., USA). Beam-bounce tips were coated with gold by CVD and glued to a magnetic holder, while tips with tuning fork were connected through a UV glue both to the tuning fork and to a special adapter. When the tuning fork is connected perpendicular to the ground, the pipette is connected beneath it, facing down, and the tip is referred to as NTF. In contrast, when connected horizontally, the pipette is attached to it from the side, and the tip is referred to as a LTF tip (FIG. 2A). The final tips had a cantilever length of 100-1000 μm, a backside opening of 500 μm, a ratio of 2:1 between their external and inner diameters, a high spring constant (k>20 N/m), and a resonance frequency of around 34.2 kHz (with tuning fork feedback).

[0217] The Substrate and the Two-Electrode Electrochemical Cell Configuration: A two-electrode configuration was used. The counter electrode was either a 25 μm or 50 μm in diameter 99.99% pure platinum wire (GoodFellow, Huntingdon, England). The substrate (working electrode) was 675 μm thick, 99.9% pure copper foil (Alfa-Aesar, MA, USA). The foil was first ground with SiC papers, from 240 P down to 2400 P. Next, it was polished with a 1 μm diamond suspension, followed by 40 nm colloidal silica suspension. Then, it was rinsed with water, placed in a chemical glass with deionized water, sonicated for 5 min, rinsed again with water, rinsed with ethanol, and dried with cold blowing air. After electrochemical printing, the substrate was cleaned with a droplet of ethanol and dried again. A current (or potential) was applied between a platinum wire and a copper foil substrate using a SourceMeter (2450-EC Electrochemistry Lab System, Keithley, Beaverton, Oreg., USA). The SourceMeter was controlled via a freeware software (KickStart, Tektronix, Beaverton, Oreg., USA).

[0218] The Electrolyte Solution for Printing Pure Copper: The acidic (pH=1) aqueous electrolyte solution for electrochemical printing consisted of 50 mM CuSO.sub.4.5H.sub.2O (copper (II) sulfate pentahydrate, 99%, Alfa-Aesar, MA, USA) and 50 mM H.sub.2SO.sub.4 (sulfuric acid, 95.0-98.0%, Sigma-Aldrich, MO, USA). Before inserting to the pipette, the solution was filtered through a 0.2 μm syringe filter (Minisart, Sartorius, Gottingen, Germany), in order to prevent clogging by large particles. When using micropipettes with an orifice smaller than 500 nm in diameter, the solution was filtered through a 100 kDa centrifugal filter (Amicon Ultra-4, Merck, NJ, USA).

[0219] Characterization of the Printed Copper Pillars: High-resolution secondary and backscattred electrons images were acquired using a SEM (Quanta 200 FEG ESEM, FEI, MA, USA). The chemical composition of the printed pillars was determined by EDS (INCA detector, Oxford Instruments, Abington, UK) integrated in the SEM system. Samples for EDS were prepared by FIB microscope (Helios NanoLab 600 DualBeam, FEI, MA, USA) down to a thickness of several microns, and were kept in a desiccator until characterization. Samples prepared by FIB and placed on a TEM copper grid were used for t-EBSD characterization inside the SEM. Images were processed with a designated software (AZtecHKL, Oxford Instruments, Abington, UK). Grain size distribution was obtained from t-EBSD data, and was compared to manual calculation from SEM images, using the Heyn Lineal Intercept Procedure. Bright-field and dark-field images as well as electron diffraction patterns were acquired using a STEM (JEM 2010F, JEOL, Tokyo, Japan). Samples for STEM characterization were prepared in a FIB microscope, down to a thickness of 70 nm. The STEM images were processed with a DigitalMicrograph software (Gatan, Pleasanton, Calif., USA). Electrical resistivity measurements were conducted on pillars 500 nm in diameter and ˜20 μm long. A silicon wafer with a 600 nm silicon oxide (SiO.sub.2) layer was used as the substrate in four-point probe measurements. Four 50 nm thick gold pads were deposited onto the wafer by lithography. The Cu pillars were positioned on the substrate inside the FIB microscope, and were connected to the gold pads by platinum wires, as shown in FIG. 7A. Electrical measurements were conducted in ambient environment, using a four-probe station (Janis, Woburn, Mass., USA). In a dual-channel SourceMeter (2603B, Keithley, Cleveland, Ohio, USA), the two inner probes were connected to one SMU, while the outer probes to the second one. A potential was applied between the two outer pads, and the current was measured. Potential was measured on the inner connections, eliminating the effect of contact resistance. To prevent excessive heating due to the relatively high resistance of the Pt connections, each potential was measured separately, for a short period of time.