Electrochemical Factors Controlling the Patterning of Metals on SAM

Jul 19, 2007 - ... reduction, making samples more corrosion resistant and improving the overall patterning process that we call electrochemical printi...
0 downloads 0 Views 453KB Size
Download PDF
Langmuir 2007, 23, 9661-9666

9661

Electrochemical Factors Controlling the Patterning of Metals on SAM-Coated Substrates Jeffrey B. Nelson and Daniel T. Schwartz* Electrochemical Materials and Interfaces Laboratory, Department of Chemical Engineering, UniVersity of Washington, Seattle, Washington 98195 ReceiVed April 6, 2007. In Final Form: June 7, 2007 Alkanethiol self-assembled monolayers (SAMs) have been used in electrochemical microfabrication processes. The reductive desorption potential of alkanethiol SAMs, Edes, can be comparable to, greater than, or less than the metal reduction potential during electrodeposition, Emet. As a result, the SAM layer can passivate the surface or desorb simultaneously with metal deposition. We show that these electrochemical traits can be combined with a rastering microjet electrode to pattern SAMs directly and create patterned metal films without lithography steps. For the case of copper deposition on 1-octanethiol (OT)- and 1-dodecanethiol (DT)-coated substrates, Edes is significantly negative of Emet, resulting in high-resolution metal patterns with poor nucleation and poor adhesion to the substrate. However, nickel patterns deposited on 1-butanethiol (BT), OT, and DT have traits similar to bare gold (excellent nucleation and adhesion) because Edes is positive of Emet. Substrates with SAMs also suppress adventitious chemistries that occur distant from the rastering microjet electrode, such as oxygen reduction, making samples more corrosion resistant and improving the overall patterning process that we call electrochemical printing.

Introduction The combination of electrochemistry and patterned alkanethiol self-assembled monolayers (SAMs) has been used in several microfabrication processes. Soft lithography-based patterning of the SAM followed by electrodeposition or etching in the unprotected areas is the most common approach,1-6 although alternative SAM patterning methods such as physical abrasion,7-9 pen writing,10,11 STM tip methods,12,13 and electron,14-17 ion,18,19 and photon beams20 are also suitable for masking the substrate. In typical applications, the SAM mask is a passive element in the subsequent electrochemical processing of the substrate. (1) Kumar, A.; Whitesides, G. M. Appl. Phys. Lett. 1993, 63, 2002-2004. (2) Moffat, T. P.; Yang, H. J. Electrochem. Soc. 1995, 142, L220-222. (3) Pesika, N. S.; Fan, F.; Searson, P. C.; Stebe, K. J. J. Am. Chem. Soc. 2005, 127, 11960-11962. (4) Pesika, N. S.; Radisic, A.; Stebe, K. J.; Searson, P. C. Nano Lett. 2006, 6, 1023-1026. (5) Xu, Q.; Tonks, I.; Fuerstman, M. J.; Love, J. C.; Whitesides, G. M. Nano Lett. 2004, 4, 2509-2511. (6) Xia, Y.; Venkateswaran, N.; Qin, D.; Tien, J.; Whitesides, G. M. Langmuir 1998, 14, 363-371. (7) Abbott, N. L.; Folkers, J. P.; Whitesides, G. M. Science 1992, 257, 13801382. (8) Abbott, N. L.; Rolison, D. R.; Whitesides, G. M. Langmuir 1994, 10, 2672-2682. (9) Seo, K.; Borguet, E. Langmuir 2006, 22, 1388-1391. (10) Kumar, A.; Biebuyck, H. A.; Abbott, N. L.; Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 9188-9189. (11) Lopez, G. P.; Biebuyck, H. A.; Frisbie, C. D.; Whitesides, G. M. Science 1993, 260, 647-649. (12) Ross, C. B.; Sun, L.; Crooks, R. M. Langmuir 1993, 9, 632-636. (13) Schoer, J. K.; Ross, C. B.; Crooks, R. M. Langmuir 1994, 10, 615-618. (14) Felgenhauer, T.; Yan, C.; Geyer, W.; Rong, H. T.; Golzhauser, A. Appl. Phys. Lett. 2001, 79, 3323-3325. (15) Kaltenpoth, G.; Volkel, B.; Nottbohm, C. T.; Golzhauser, A. J. Vacuum Sci. Technol., B 2002, 20, 2734-2738. (16) Sondag-Huethorst, J. A. M.; Helleputte, H. R. J. v.; Fokkink, L. G. J. Appl. Phys. Lett. 1994, 64, 285-287. (17) Volkel, B.; Koltenploth, G.; Handrea, M.; Sahre, M.; Nottbohm, C. T.; Kuller, A.; Paul, A.; Kavtek, A.; Eck, W.; Golzhauser, A. Surf. Sci. 2005, 597, 32-41. (18) Dulcey, C. S.; Georger, J. H., Jr.; Chen, M.-S.; McElvany, W. W.; O’Ferrall, E. O.; Benezra, V. I.; Calvert, J. M. Langmuir 1996, 12, 1638-1650. (19) Takehara, K.; Yamada, S.; Yasushi, I. J. Electroanal. Chem. 1992, 333, 339-344. (20) Gillen, G.; Wight, S.; Bennett, J.; Tarlov, M. J. Appl. Phys. Lett. 1994, 65, 534-536.

However, SAMs can be an active, switchable part of the patterning and assembly of microfabricated systems. For example, electrochemical methods have been used to adsorb/desorb SAMs from the substrate and to oxidize their terminal group, creating patterned hydrophilic and hydrophobic regions for subsequent processing.21-24 Though not comprehensively reviewed here, SAM patterning and electrochemical processing have always occurred in separate steps, to the best of our knowledge. Nonetheless, there is no fundamental requirement for separating these steps. The alkanethiol SAM reductive desorption potential, Edes, can be comparable to, greater than, or less than the metal reduction potential during electrodeposition, Emet, depending on the alkanethiol chain length, electrolyte pH, and nobility of the metal to be deposited. For example, Zhong and Porter showed that Edes changed from roughly -800 to -1l00 mV versus Ag/AgCl for C4 to C16 alkanethiols on gold when desorbed into strong base.25 Edes also shifts by 59 mV per pH unit. Noble metal deposition is well positive of Edes whereas base metal deposition is well negative of these potentials, especially when plating occurs at high overpotentials (high rates). The relationship between Edes and Emet should determine whether the alkanethiol remains on the surface as a passivation mask during electrodeposition or desorbs simultaneously with metal electrodeposition. In this article, we first demonstrate that a SAM-coated substrate can be electrochemically patterned using a method that we call electrochemical printing (EcP). EcP is a software-reconfigurable direct-write micropatterning process that involves localized electrochemistry (usually metal deposition) on conductive substrates via a rastering microjet electrode.26-28 SAM-coated (21) Xiong, X.; Hanein, Y.; Fang, J.; Wang, Y.; Wang, W.; Schwartz, D. T.; Bo¨hringer, K. F. J. Microelectromech. Syst. 2003, 12, 117-127. (22) Hoeppener, S.; Maoz, R.; Sagiv, J. AdV. Mater. 2006, 18, 1286-1290. (23) Maoz, R.; Frydman, E.; Cohen, S. R.; Sagiv, J. AdV. Mater. 2000, 12, 424-429. (24) Maoz, R. et al. AdVanced Materials 2000, 12 (10), 725-731. (25) Zhong, C.; Porter, M. D. Fine structure in the voltamemetric desorption curves of alkanethiolate monolayers chemisorbed at gold. J. Electroanal. Chem. 1997, 425, 147-153. (26) Whitaker, J. D. Electrochemical Printing. In Chemical Engineering University of Washington: Seattle, 2003.

10.1021/la701014u CCC: $37.00 © 2007 American Chemical Society Published on Web 07/19/2007

9662 Langmuir, Vol. 23, No. 19, 2007

Nelson and Schwartz

substrates are not required to perform EcP patterning of metals, but one-step SAM desorption and electrodeposition should improve the EcP process. Here we explore the trade-offs exhibited by systematically varying the relationships between Edes and Emet for copper and nickel deposition on 1-butanethiol (BT)-, 1-octanethiol (OT)-, and 1-dodecanethiol (DT)-coated gold substrates. We show that print resolution can be improved, for certain conditions, by focusing the current distribution under the microjet. Undesirable adventitious chemistries (such as oxygen reduction) occurring on the large unpatterned regions of the substrate are also squelched, reducing degradation of the patterned metal by corrosion (as one example). Adhesion of the pattern to the substrate is also dependent on the relationship between these key potentials. Experimental Section Substrate Preparation. Substrates were glass microscope slides (25 × 76 mm2) with gold atop a titanium adhesion layer. Glass slides were first cleaned in piranha solution (60 wt % H2SO4, 10 wt % H2O2, and 30 wt % H2O) for several minutes, rinsed with water, and dried with air. Slides were then placed in an e-beam evaporator where 10 nm of titanium and 100 nm of gold were deposited. Alkanethiol SAMs were prepared by cleaning the gold-coated glass slides in piranha solution for several minutes, followed by a water rinse. Substrates were then placed in a 2 mM solution of the appropriate alkanethiol in 200 proof ethanol, purged with argon, and stored at room temperature for at least 24 h prior to use. Some experiments used bare gold substrates without a SAM and were electrochemically cleaned by repeatedly sweeping the potential from -780 to 1300 mV in a 1.0 M H2SO4 electrolyte versus a mercury sulfate (MSE) reference electrode and a platinum counter electrode. All substrates were then assembled as shown in Figure 1a. EcP Hardware. Detailed EcP procedures and equipment are described in refs 26 and 27. Figure 1a shows how the substrates were placed between two pieces of Mylar adhesive backing. Figure 1b shows a schematic of the EcP print head consisting of a 65 µm i.d. glass capillary nozzle and platinum counter electrode. The glass nozzle was fabricated by pulling a 1/8 in. o.d. glass capillary tube in a Bunsen burner flame, breaking it in two, and reclosing it in the flame. The desired diameter was obtained by carefully sanding back the tip with fine grit sand paper. A MAXNC three-axis mini-mill with 3 µm spatial resolution was used to position the substrate beneath the nozzle, and a Keithley 2400 source meter was used as the galvanostatic power supply. Patterning a SAM Mask with EcP. A 0.0010 M H2SO4 (pH 2.7) electrolyte was used to pattern a DT SAM with EcP. Copper was then electrodeposited on regions of the substrate not covered by the SAM in a 0.100 M CuSO4, 0.100 M H2SO4 (pH 1.4) roomtemperature aqueous electrolyte. EcP Deposition on SAM-Coated Substrates. Copper deposition in EcP was done with a 0.100 M CuSO4, 0.0010 M H2SO4 (pH 2.8) electrolyte. The nickel plating bath was a 0.300 M NiSO4, 0.100 M NaCH3COO, and 0.020 CH3COOH (pH 4.8) electrolyte. Solutions were prepared from reagent-grade salts (CuSO4, NiSO4, NaCH3COO), 96.2 wt % sulfuric acid, and glacial acetic acid, and all electrodeposition occurred at room temperature. Metrology. Pattern topography was obtained with a Wyco NT 3300 noncontact optical profiler with a 5.0× objective at a 0.5× field of view (2.5 mm × 1.9 mm). An Olympus BX51 optical microscope with an Olympus QColor3 digital camera through a 5.0× objective was used to take optical micrographs of each sample. An Oakton model 510 pH/conductivity meter was used to measure the pH. Experimental Design. Figure 2 shows the equilibrium potentials eq eq for copper and nickel (ECu 2+/Cu and ENi2+/Ni, respectively) and (27) Whitaker, J. D.; Nelson, J. B.; Schwartz, D. T. J. Micromech. Microeng. 2005, 15, 1498. (28) Nelson, J. B.; Schwartz, D. T. J. Micromech. Microeng. 2005, 15, 24792484.

Figure 1. Substrate assembly (a) and EcP print head schematic (b, not to scale). Shown in part a is a piece of Mylar (1, adhesive side down) with a 5 mm hole (2) placed atop the substrate (3), a piece of aluminum tape for electrical contact (4), and a second piece of Mylar with the adhesive side up (5). The print head consists of a 65-µm-diameter glass capillary nozzle and a platinum counter electrode. A three-axis mini-mill controls the fly height (h) and x-y position of the electrode. The red region represents a metal deposit beneath the microjet print head (not to scale).

Figure 2. Equilibrium and desorption potentials. The top potential scale shows the estimated equilibrium potential versus a standard hydrogen electrode (SHE) for a 0.100 M CuSO4/H2SO4 electrolyte (pH 2.8) and the estimated desorption potentials for 1-butanethiol des des (Edes BT ), 1-octanethiol, (EOT), and 1-dodecanethiol (EDT) at pH 2.8. The bottom potential scale shows the estimated equilibrium potential for a 0.300 M NiSO4, 0.100 M NaCH3COO, 0.020 M CH3COOH electrolyte (pH 4.8) and the estimated desorption potentials for the same alkanethiols at pH 4.8. des des desorption potentials for BT (Edes BT ), OT (EOT), and DT (EDT) at two 25,29-33 The pH values correspond to the copper different pH values. and nickel plating baths. The difficulty in implementing a reference electrode in the microgeometry of EcP means that all electrodeposition is controlled galvanostatically. As a result, we can show only estimated values of Emet during high rate EcP deposition, typically

(29) Healy, J. P.; Pletcher, D. J. Electroanal. Chem. 1992, 338, 155-165. (30) Lachenwitzer, A.; Magnussen, O. M. J. Phys. Chem. B 2000, 104, 74247430. (31) Moffat, T. P. Electrochem. Solid State Lett. 2001, 4, C26-C29. (32) Saraby-Reintjes, A.; Fleischmann, M. Electrochim. Acta 1984, 29, 557566. (33) Zech, N.; Podlaha, E. J.; Landolt, D. J. Electrochem. Soc. 1999, 146, 2886-2891.

Patterning of Metals on SAM-Coated Substrates

Langmuir, Vol. 23, No. 19, 2007 9663

Figure 3. Patterning a 1-dodecanethiol (DT) SAM with EcP. Shown in part a is a schematic of the EcP print head and substrate covered with a DT SAM (not to scale). The SAM is locally desorbed beneath the microjet electrode as shown in part b. A conventional electrodeposition cell (c) is then used to deposit metal in regions not covered by the SAM. The DT SAM was patterned with the 5 × 5 dot array shown in part d, where each black square represent an active pixel and the white regions are inactive. The SAM was patterned with a 0.0010 M H2SO4 electrolyte at a 15 µm fly height with a 0.2 mL/min flow rate according to the currents and charges shown in part d. Copper was electrodeposited on regions not covered by the SAM in a 0.100 M CuSO4, 0.100 M H2SO4 (pH 1.4) electrolyte at -125 mV vs SCE. The resulting copper microdot array is shown in an optical micrograph (e) and the corresponding tilted topography map (f). >1 A/cm2. Figure 2 provides these estimates as red and gray bars on the potential axis for copper and nickel, respectively. For nickel, it is clear that Edes is positive of Emet Ni for all cases, but for copper, Emet Cu is likely to be comparable to or more positive than Edes, depending on the alkanethiol chain length. Given these different relationships among potentials, EcP-printed copper and nickel patterns (at identical current densities and charge) should be a good indicator of how adhesion and pattern resolution are impacted by Edes versus Emet. Also shown is the equilibrium potential for hydrogen evolution, EHeq+/H2, in each electrolyte. This shows that we should be able to pattern the SAM directly at any pH without simultaneous metal deposition.

Results We first illustrate how a DT SAM mask is patterned using the EcP tool without metal deposition by performing SAM desorption from a pH 2.7 sulfuric acid electrolyte. Figure 3a is a schematic of the EcP print head (not to scale) with a SAM atop the substrate. When the switch closes and current begins to flow, some of the SAM is electrochemically desorbed beneath the microjet electrode as shown in Figure 3b. Water electrochemistry (or proton reduction) also occurs because, as shown in Figure 2, Edes DT and EHeq+/H2 are within 20 mV of each other. Once the SAM is patterned, conventional copper electrodeposition will occur in regions not covered by the SAM, as shown in Figure 3c. The effects on the applied current and charge on the amount of SAM desorbed were examined systematically by patterning a SAM according to the 5 × 5 drawing shown in Figure 3d. Each black square represents an active print region, or pixel, for a different current or charge as marked along each row and column. The microjet electrode fly height above the substrate and electrolyte flow rate were constant at 15 µm and 0.20 mL/min, respectively. The substrate was then placed in a conventional three-electrode plating cell where copper was electrodeposited at 116 mV versus SHE in an acid copper sulfate electrolyte (pH 1.4). At this potential and pH, the regions covered by DT remained chemisorbed to the substrate and served as a deposition mask because Edes DT (-100

met 25 mV at pH 1.4) was significantly negative of ECu/Cu 2+ (116 mV). An optical micrograph and a corresponding topography map (tilted view) of the resulting copper microdot array are shown in Figure 3e,f, respectively. Figure 3e shows that copper deposits only in regions that have been patterned by EcP. The resulting copper dot diameter increases from bottom to top and from right to left, following the same trend of increasing current and charge shown in Figure 3b and confirming that the amount of SAM desorbed depends on both. As the current increases, the region where the local surface potential is negative enough to desorb the SAM broadens, resulting in a larger “hole” in the SAM. Passing more charge for a given current also increases the size of the SAM desorbed area. Dark regions seen in Figure 3e are due to rough copper that has high optical scattering. As we discuss later in Figure 6, this type of copper growth is likely due to deposition through partially desorbed SAMs. The corresponding tilted topography map of Figure 3f shows dots with a more cylindrical shape than normally seen for EcP on uncoated substrates (reported elsewhere27). The process illustrated in Figure 3 is easily extended to larger and more complex patterns. Figure 4 shows a computer drawing of the UW Husky logo (part a) used to pattern a DT SAM under the conditions described in the Figure. Again, the black regions represent active pixels to be patterned, and the white regions represent inactive areas. A copper film was deposited through the SAM mask and passed a subsequent scotch tape adhesion test, indicating good nucleation on the substrate and a dense deposit. An optical micrograph of the copper on gold pattern is shown in part b, and a topography map is shown in part c. Figures 3 and 4 illustrate how a custom SAM mask is created with the EcP tool from a software drawing. EcP-patterned SAMs can be easily placed in a commercial plating cell to deposit the desired metal without the need for any lithography steps, in the same way any patterned SAM can be used. The build accuracy (i.e., replication of Figure 4a) could be improved by selecting a different current, reducing the nozzle fly height, or changing the dot spacing.27

9664 Langmuir, Vol. 23, No. 19, 2007

Nelson and Schwartz

Figure 4. Patterning a 1-dodecanethiol SAM in the shape of a UW Husky logo. Shown in part a is a computer drawing of the UW Husky logo used to pattern a 1-dodecanethiol SAM. The SAM was patterned using EcP at 30 µA current, 30 µC charge per pixel, 15 µm fly height, 25 µm pixel spacing, and 0.2 mL/min flow rate in a 0.0010 M H2SO4 electrolyte. A copper film was deposited in regions not covered by the SAM at -125 mV vs SCE (116 mV vs SHE) in a 0.100 M CuSO4, 0.100 M H2SO4 electrolyte (pH 1.4). An optical micrograph of the copper pattern is shown in part b, and the corresponding topography map is shown in part c.

Figure 6. Copper deposition with EcP on 1-octanethiol (a and c) and 1-dodecanethiol (b and d). In this case, the desorption potentials (Edes) are less than the copper deposition potential (Emet). The microdot arrays shown in parts a and b were deposited at the currents marked along each column and 100 µC charge per dot. Pattern c was deposited at 30 µA current and 30 µC charge per pixel, and pattern d was deposited at 50 µA current and 25 µC charge per pixel. All patterns were electrodeposited at a 15 µm fly height and 0.05 mL/min from a 0.100 M CuSO4, 0.0010 M H2SO4 electrolyte (pH 2.8). Figure 5. Copper deposition with EcP atop bare gold (a) and a 1-butanethiol SAM (b). Both patterns were left in the electrolyte at open circuit for equal time intervals of 15 min (c, d) and 45 min (e, f). Both patterns were electrodeposited under the same operating conditions of 30 µA, 30 µC charge per pixel, 15 µm fly height, 25 µm pixel spacing, and 0.200 mL/min flow rate from a 0.100 M CuSO4, 0.0010 M H2SO4 electrolyte (pH 2.8).

EcP can also be used to directly deposit metal patterns on SAM-coated substrates. Figure 5 shows optical micrographs of standard EcP copper patterns deposited on bare gold (Figure 5a) and on a BT SAM-coated substrate (Figure 5b). On the basis of met Figure 2, we expect Edes BT ≈ ECu/Cu2+ for this system. Both patterns were deposited under identical conditions described in the Figure. The pattern in part b shows more detail in the ears, eye, and back of the head than does the pattern in part a, suggesting that the presence of the BT SAM focuses the current distribution. Adhesion is also affected by the BT SAM because copper patterns deposited on gold consistently pass a scotch tape adhesion test, but similar tests with copper patterns deposited on BT SAMs are highly inconsistent, with most showing partial failure. After the patterns were complete, they were left in the electrolyte at open circuit and removed at identical time intervals (15 and 45 min) after deposition to observe the copper corrosion behavior versus time (shown in Figure 5c-f). The pattern in Figure 5a was completely corroded within 60 min whereas the

patterns in Figure 5d,f show only a slight discoloration, indicating that the presence of the BT SAM reduces adventitious oxygen reduction on the gold substrate surrounding the pattern and therefore significantly suppresses copper corrosion. It is also possible that there is some BT SAM adsorbed to the copper pattern, also suppressing copper oxidation. We previously demonstrated that cathodic protection can be used to control copper corrosion on gold substrates during EcP, and here we show how the use of a BT SAM is also an option.28 Different results are obtained with copper deposition on substrates coated with longer-chain SAMs because Edes shifts significantly negative of Emet. Figure 6 shows four optical micrographs of electrochemically printed copper atop OT (a, c) and DT (b, d). Before printing large patterns, we routinely print dot arrays to screen for desirable operating conditions.27 Here, we show two dot arrays (Figure 6a,b) patterned at constant charge, fly height, and electrolyte flow rate but at different currents marked in the Figure. Copper dots deposited on OT are clearly visible from 10 to 50 µA. For copper on DT, deposits are barely visible at or below 30 µA, indicating a higher current density is required to electrochemically desorb enough of the SAM to yield reasonable nucleation densities. As a result of this prescreening, the deposition current used to print on DT was 50 µA and on 30 µA for OT. The pattern in part c was deposited on OT and shows a continuous film over the entire pattern, indicating good

Patterning of Metals on SAM-Coated Substrates

Langmuir, Vol. 23, No. 19, 2007 9665

Figure 7. Scanning electron microscope (SEM) micrographs of copper films on bare gold (a), 1-octanethiol (b), and 1-dodecanethiol (c). The film in part a was deposited at a 0.2 mL/min flow rate (V), 30 µA current (Iapp), and 30 µC charge per pixel (q). Operating conditions for part b were V ) 0.05 mL/min, Iapp ) 30 µA, and q ) 30 µC. For part c, operating conditions were V ) 0.05 mL/min, Iapp ) 50 µA, and q ) 25 µC. All films were deposited at a 15 µm fly height from a 0.100 M CuSO4, 0.0010 M H2SO4 electrolyte (pH 2.8).

Figure 8. Nickel deposition with EcP atop bare gold (a), 1-butanethiol (b), 1-octanethiol (c), and a 1-dodecanethiol (d) SAM. Each pattern was printed at identical operating conditions of 70 µA current, 35 µC charge per pixel, 15 µm fly height, 25 µm pixel spacing, and 0.20 mL/min flow rate from a 0.300 M NiSO4, 0.020 CH3COOH, 0.100 M NaCH3COO electrolyte (pH 4.8). Topography maps are shown to the right of each optical micrograph.

nucleation. The pattern in part d does not display a dense continuous film owing to poor nucleation but it does show an increase in the pattern resolution when compared to Figure 6c, indicating the increased chain length of the SAM further focuses the current distribution due to its more negative desorption potential and better packing order. Moreover, replicates of each case in Figure 6c,d consistently failed scotch tape adhesion tests, unlike copper films on BT SAMs. The patterning of thin films that can be lifted whole from the substrate is sometimes an attribute,34 and if that is sought, EcP of Cu on OT is an excellent choice for lifting off a dense continuously patterned thin film. The effects of depositing copper on bare gold, OT, and DT were examined more carefully with scanning electron microscopy. Shown in Figure 7 are SEM micrographs of copper films deposited on bare gold (a), 1-octanethiol (b), and 1-dodecanethiol (c) under the conditions listed in the Figure. For copper on gold (a), the film is very dense with grains of less than 100 nm. The micrograph in b also shows a continuous copper film with grain size and (34) Azzaroni, O.; Schilardi, P. L.; Salvarezza, R. C. Electrochim. Acta 2003, 48, 3107-3114.

density comparable to that in a, even though it failed a scotch tape adhesion test. The film morphology suggests a high nucleation density, but because the film lifts off, it is likely that some residual OT remains at the substrate-deposit interface. However, in part c, copper deposited on DT results in a much lower and inconsistent nucleation density. In short, as one goes from Emet ≈ Edes to Emet positive of Edes the adhesion and nucleation densities decrease. The last case from Figure 2 is when Emet is significantly negative of Edes. This case was examined by depositing nickel on bare gold and SAM-coated substrates. Figure 8 shows four optical micrographs of nickel patterns on bare gold (a), BT (b), OT (c), and DT (d) under conditions described in the Figure. The regions of the substrates, labeled i, were masked during the whole EcP process and are shown to provide an optical comparison to the electrolyte-wetted regions marked ii. In Figure 8a,b, there is a high optical contrast between regions i and ii. This indicates that adventitious peripheral deposition chemistry occurs over the entire electrolyte-wetted electrode; significantly less background deposition is observed on OT and DT (cf. regions i and ii in Figure 8c,d). Thus, the longer-chain SAMs squelch the adventitious background chemistry. Figure 8a-d also shows corresponding topography maps of each pattern shown to the right of each optical micrograph, where the colored scale bar represents the z scale and the solid black scale bar indicates the x-y scale. Unlike the copper patterns in Figures 4 and 5, the nickel patterns shown here are quite similar except that patterns deposited on a SAM show a little more detail than bare gold. This indicates that the SAM-coated substrates focus the current distribution a bit when compared to a SAMfree substrate. All four patterns shown in Figure 8 passed subsequent scotch tape adhesions tests. In these cases, local SAM desorption continues during nickel deposition, providing plenty of nucleation sites for the nickel and allowing the formation of adherent dense films. In short, when Emet is significantly negative of Edes, the SAM is lifted from the substrate, allowing good nucleation and adhesion, slightly improved pattern resolution, and a squelching of the background chemistry.

Concluding Remarks We have shown how EcP can pattern a SAM mask from a software drawing and an acid electrolyte. This process is an alternative to lithographic patterning and is completely software reconfigurable. If one chooses a two-step process, then one can use commercially available electrodeposition recipes to fabricate metal patterns through resulting SAM masks in the same manner that any patterned SAM is used. We have also shown how the relative magnitudes of Edes and met E affect the deposit morphology and adhesion. Deposits with poor adhesion but improved resolution are obtained when Edes < Emet, but films pass a scotch tape adhesion test and show

9666 Langmuir, Vol. 23, No. 19, 2007

minimal increased resolution when Edes > Emet. We further showed how adventitious chemistries (e.g., copper corrosion and background deposition) were suppressed in the presence of some alkanethiols. These results provide a framework for the implementation of other materials and surface-modifying agents in EcP. Glossary Edes Emet

peak desorption potential of an alkanethiol self-assembled monolayer deposition potential of a specified metal

Nelson and Schwartz Eeq BT OT DT SHE

equilibrium potential predicted by the Nernst equation 1-butanethiol 1-octanethiol 1-dodecanethiol standard hydrogen electrode

Acknowledgment. This work was funded by the National Science Foundation (DMI 0323160). LA701014U