Electrostatically actuated dip pen nanolithography ... - Semantic Scholar
Sensors and Actuators A 125 (2006) 504–511
Electrostatically actuated dip pen nanolithography probe arrays David Bullen, Chang Liu ∗ Micro Actuators, Sensors, and Systems Group, Micro and Nanotechnology Laboratory, University of Illinois, Urbana, IL 61801, USA Received 13 December 2004; received in revised form 18 August 2005; accepted 1 September 2005 Available online 10 October 2005
Abstract Dip pen nanolithography (DPN) is a method of creating nanoscale chemical patterns on surfaces using an atomic force microscope (AFM) probe. Until now, efforts to increase the process throughput have focused on passive multi-probe arrays and active arrays based on thermal bimetallic actuation. This paper describes the first use of electrostatic actuation to create an active DPN probe array. Electrostatic actuation offers the benefit of actuation without the probe heating required for thermal bimetallic actuation. Actuator cross talk between neighboring probes is also reduced, permitting more densely spaced probe arrays. The array presented here consists of 10 cantilever probes, where each is 120 m long and 20 m wide. Each cantilever probe is actuated by the electrostatic force between the probe and a built-in counter electrode with a 20–25 m gap. The tip-to-tip probe spacing, also called the array pitch, is 30 m. Patterns of 1-octadecanethiol were created on gold surfaces to demonstrate single-probe actuation, simultaneous multi-probe actuation, and overlap of patterns from adjacent probes. The minimum line width was 25 nm with an average line width of 30–40 nm. © 2005 Elsevier B.V. All rights reserved. Keywords: Dip pen nanolithography (DPN); Electrostatic actuator; Atomic force microscope; Nanolithography
1. Introduction Dip pen nanolithography (DPN) [1] is a process for directly depositing chemical patterns on surfaces with nanoscale dimensions and precision. To perform DPN, an atomic force microscope (AFM) probe is coated with a chemical and slowly translated over the surface in contact mode. The chemical diffuses from the tip to the surface to create the pattern. Researchers have used DPN to deposit a variety of organic, inorganic, and biological inks on metal, semiconductor, and oxide surfaces. This is in contrast to other scanning probe methods such as thermomechanical data storage [2], or local field oxidation of the surface [3], although some aspects of the probe technology are related. DPN is becoming increasingly popular as a unique tool for nanoscale surface engineering. We refer interested readers to [4] for an overview of DPN research and applications. Despite its usefulness in research, single-probe DPN is a slow process and not suitable for industrialization. One way to increase the patterning throughput is to perform the process in a parallel fashion with multiple probes. Passive probe
∗
Corresponding author. Tel.: +217 333 4051; fax: +217 244 6375. E-mail address: [email protected] (C. Liu).
0924-4247/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2005.09.001
arrays are one way to address this need [5]. These are arrays of identical, closely spaced AFM probes affixed to a common holder chip. When placed in contact with the surface, each probe writes the same pattern simultaneously. Passive DPN arrays have been produced using silicon nitride and silicon [6], permalloy, polyimide, SU-8 photoepoxy [7], and polydimethylsilizane [8]. An additional level of functionality is achieved by including a Z-axis actuator on each probe to create an active probe array. The actuator allows individual probes to be lifted from the surface independently of the array to halt the deposition process. As a result, different patterns can be created by each probe even though they all follow the same path on the surface. There are several possible actuation methods, including thermal bimetallic, piezoelectric, and electrostatic actuation (sometimes called capacitive actuation). Each has advantages and disadvantages for a given patterning task. The only mechanism that has been studied in detail for DPN applications is thermal bimetallic actuation [9,10]. In thermally actuated DPN (TA-DPN) arrays, the probe is a rectangular cantilever formed from layers of high and low thermal expansion coefficient material, typically gold and silicon or silicon nitride. The gold film is patterned to form an ohmic heater at the base of the probe. Passing current through the heater warms the entire
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
probe, causing it to curl away from the substrate, suspending deposition. When the probe is cooled, the tip returns to the surface and deposition begins again as it does in a passive array. Thermal bimetallic actuation is robust, simple, and effective for many, but not all, DPN applications. Its main limitation is the potential effect of the temperature excursion on the chemical ink. TA-DPN arrays typically warm 10–40 ◦ C above ambient during actuation [11]. Some ink compounds (proteins for example) may be adversely affected by this elevated temperature. Another problem with thermal bimetallic actuation is thermal cross talk between adjacent probes. Thermal cross talk refers to unwanted probe deflection caused by heat transfer between probes. In TA-DPN arrays, thermal cross talk is mainly due to heat conduction and convection through the air, rather than conduction through the substrate or radiation. These become more pronounced as the array pitch is reduced. To date, the smallest pitch achieved with a manageable cross talk is 100 m [11]. In these devices, an actuated-probe deflection of 16.3 m results in 3.7 m (23%) of deflection in the first adjacent probes and 1.9 m (12%) of deflection in the second adjacent probes. Previous DPN experience suggests that the maximum allowable cross talk value is about 30–40% before its impact on usability becomes unacceptable. Since the array pitch is an important metric (it relates directly to pattern throughput and overlap capability), a smaller value is highly desirable, but difficult to obtain with thermal actuation. An alternative to thermal bimetallic actuation is electrostatic actuation. Electrostatic actuation offers a way to deposit temperature sensitive chemicals without damage. Cantilever electrostatic actuators scale well to small sizes and can generate large tip deflections relative to the cantilever length. Electrostatic actuation has already been used to produce actuated AFM probes [12], although the design was not suitable for multi-probe arrays where each probe is individually addressable. A previous attempt to adapt the method for DPN [13] yielded devices that had an excessively large array pitch, produced inadequate tip deflection, and could not be demonstrated with multi-probe lithography. In this paper, we present a new active DPN array concept based on cantilever electrostatic actuation. The concept presented here is effective, scaleable, and proven in DPN patterning tasks.
505
Fig. 1. The operating concept of an EA-DPN probe. When the counter electrode is grounded, the probe tip is in contact with the surface and performing DPN. To actuate the probe and suspend DPN, a potential is applied to the counter electrode to pull the tip off the surface.
and surface. Since all the probes are fabricated from the same film, this configuration also eliminates the need to individually wire bond probes to ground pads on the holder. This allows the design to be scaled to a larger number of probes without making the assembly process more difficult. Fig. 2 shows an SEM image of the probes and counter electrodes. Each probe is 20 m wide, 120 m long, and 0.65 m thick. The spring constant is approximately 0.11 N/m. The array pitch is 30 m, which is much smaller than the 100 m achieved with thermal bimetallic actuation. The probe tips are 7–8 m tall and have an apex radius of less than 100 nm. The probe cantilevers are fabricated from a film stack that consists of silicon nitride, chromium, and gold. As shown later in the fabrication section, this stack is oriented so that the gold film faces the holder and the silicon nitride faces the writing surface. The silicon nitride layer consists of a zero-stress component and a tensile component. They are oriented to cause the probe to deflect away from the holder after release, as shown in Fig. 2. This increases the clearance between the probe cantilevers and the writing surface during lithography. At the base, the gap between the probe and the counter electrode is 20 m. At the tip, the gap is approximately 25 m. Fig. 3 shows an SEM image of a packaged array. The probes are fabricated as part of a 2 mm × 2 mm film, which is separated from the glass holder by the insulator. The film is electrically
2. Description Fig. 1 illustrates the operating concept of an electrostatically actuated DPN (EA-DPN) probe. In this design, the probe acts as an electrode and is separated from its counter electrode by an insulator. There is one counter electrode per probe in the 10probe array. All the probes are fabricated from the same film and are attached to a holder with a common insulator. The holder is made of glass and the insulator is SU-8 photoepoxy. The actuation potentials are applied to the counter electrodes on the holder. The probes are grounded by connecting them to a ground trace with conducting paste. Applying a voltage to the counter electrode generates an attractive force that pulls the tip off the surface. If the lithography surface is grounded, this arrangement eliminates any electric field or actuator force between the tip
Fig. 2. An SEM image of an EA-DPN probe array. The cantilevers and tips are fabricated as part of a common thin film that is attached to a holder by an SU-8 insulator. The inset shows a single-probe tip.
506
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
Fig. 3. An SEM image of a packaged EA-DPN probe array including the actuator and ground traces on the holder surface.
Fig. 5. A packaged EA-DPN probe assembly. The glass holder carries actuation leads from an external connection to the counter electrodes under each probe.
connected to a ground pad by conducting paste. The use of a large film minimizes the risk that the paste will wet around the edge of the device and short the actuator traces when it is applied. The paste is carbon-based and becomes electrically conducting after the solvent evaporates. For this work the paste was applied manually. More efficient and controllable methods for applying the paste and providing electrical conduction pathways will be developed in the future. Many works in the past have successfully dealt with problem of high density connections to one- and two-dimensional arrays [14]. There are 10 actuator traces on the holder surface, one for each probe. Five traces approach the array from each side and pass under the insulator on their way to the counter electrodes. As shown in Fig. 4, there is a ground trace between each actuator trace. This ground reduces the electrostatic interaction between the actuator traces. Without it, when one probe is charged, the electric field around its actuator trace will induce a charge in the adjacent traces, resulting in actuation of the adjacent probes. The grounds reduce this effect significantly at the cost of a small increase in system capacitance.
A photograph of the complete probe assembly is shown in Fig. 5. The probe holder is a glass slide that is 35 mm long, 6 mm wide, and 1 mm thick. The actuator and ground traces extend from the probe array to a region where they match up to a 20-conductor, 0.5 mm pitch cable. Using glass allows the probes to remain visible through the holder during lithography. The fact that the grounded metal film on each probe faces the counter electrode is a potentially catastrophic problem if a cantilever were to snap-in during actuation. Three design and operating features mitigate this problem. First, the probes are always operated at less than their experimentally estimated snapin voltage. Second, as shown in Fig. 6, the counter electrode is shorter than the probe to prevent them from touching if the probe pulls in. The shorter counter electrode also allows the backside of the probe tip to be visible through the glass holder during lithography. Finally, there is a 200 M resistance inserted in each actuator’s electrical pathway. This limits the steady-state short circuit current to less than 1.5 A in the event that the probe touches its counter electrode. Extensive use during DPN has demonstrated the adequacy of these measures.
Fig. 4. Actuator and ground trace layout. Ground traces reduce the magnitude of the electric field interaction between the actuator traces. They extend from the external power supply to the edge of the probe array.
Fig. 6. An SEM image of a snapped-in probe. The probe is longer than the counter electrode to prevent it from touching the counter electrode if it snaps-in. This tip is from an earlier array that was larger, but otherwise identical, to those pictured in Fig. 2. Unlike the smaller probes, these devices were capable of being charged to the snap-in condition by the SEMs current.
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
507
(a) Fabrication of the probe array starts with an oxidized (1 0 0) silicon wafer. Square openings in the oxide layer are etched where each tip will be located. Pyramid trenches are then anisotropically etched in the silicon substrate and the remaining oxide is removed. (b) The probe array film stack materials are then deposited in ˚ zinc oxide, 800 A ˚ tensile silicon the following order: 1000 A ˚ ˚ chromium, nitride, 5000 A stress-free silicon nitride, 50 A ˚ and 200 A gold. All the layers are patterned with the same mask to form the probe array film. A 20 m thick layer of SU-8 25 is then spin coated and patterned to form the insulator. The wafer is then diced into individual dies. (c) The holder fabrication begins by trimming a glass micro˚ chromium scope slide to 35 mm × 6 mm × 1 mm. A 150 A ˚ adhesion layer and 2000 A gold electrode layer are then deposited and patterned to form the ground and counter electrode traces. The entire surface is then coated with a 2–4 m thick layer of SU-8 5 and soft baked. (d) The probe array is assembled by bonding the probe chip to the glass holder. This is done by placing the two in contact with the probes and counter electrodes aligned. The probe chip is heated to 70–80 ◦ C to allow the uncured SU-8 on the glass holder to reflow and wet the cured SU-8 insulator on the probe array. This creates a temporary joint when the probe chip is cooled. The joint is made permanent by exposing the reflow region to UV light through the glass holder and heating it to crosslink the uncured SU-8. The remaining unexposed SU-8 is then removed with SU-8 developer. (e) The probe array is released from the silicon chip by etching the zinc oxide sacrificial layer between the array and the silicon die. After release, the gold layer on the probe array film is connected to the ground electrode at the rear of the array by manually applying conducting carbon paste. 4. Experimental Fig. 7. Process flow for producing an EA-DPN probe assembly.
3. Fabrication The fabrication procedure is based on a mould-and-transfer process [7] and is illustrated in Fig. 7. The probe array and holder are fabricated separately then bonded together at the end of the process. The deposition and etching methods for each film are listed in Table 1: Table 1 Deposition and etching method used for each film in the fabrication process Material
Deposition
Etchant
Si SiO2 ZnO Si3 N4 Cr Au SU-8
Substrate Thermal oxidation RF magnetron sputtering PECVD Thermal evaporation Thermal evaporation Spin on
35% KOH Buffered HF 0.5% HCl (step b), 38% HCl (step e) Tetrafluoromethane RIE Cr etchant (CEP-200, Microchrome) Au etchant (TFA, Transene) SU-8 developer (MicroChem)
4.1. Deflection voltage In order to verify the effectiveness of electrostatic actuation in DPN applications, several experiments were carried out. The first was to determine a suitable operating voltage for the device to avoid a snap-in condition [15]. Although parallel-plate electrostatic actuators produce most of their deflection near the snap-in voltage, it is preferable to avoid the snap-in condition for two reasons. First, near this point the cantilever position is very sensitive to nuances in the mechanical and electrostatic state of the system. This makes tip displacement from probe to probe very non-uniform. Second, the repeated snap-in events may create particles that could interfere with the writing results. Fig. 8 shows the average displacement versus voltage relationship of several probes in the array. To illustrate the impact of the electric field on tip deflection, two probes, located at the edge and center of the array, are plotted as well. These two probes are numbered 1 and 5, respectively. As Fig. 8 shows, the displacement within an array becomes more non-uniform as voltage increases. At counter electrode voltages above 260 V, probe snap-in events are common. Based on this data, a safe
508
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
Fig. 8. EA-DPN probe deflection vs. counter electrode voltage. Circles represent the average deflection of all 10 probes. The solid line is a polynomial fit to the average deflection. The deflection of probe 1 (at the edge of the array) and probe 5 (in the center of the array) is included for comparison. Error bars represent the measurement error.
operating voltage of 190 V was chosen for lithography operations. The actuator cross talk was also measured in this experiment. In TA-DPN arrays, experience has shown that thermal cross talk between probes is one of the greatest impediments to reducing the pitch of the array. Electrostatic actuators also suffer from cross talk due to fringe electric fields. To determine the cross talk magnitude, individual EA-DPN probes were actuated while measuring the deflection of the adjacent, unactuated probes. In these single-probe experiments, an actuated-probe deflection of 9 m was accompanied by 3 m (33%) of deflection in the first adjacent probes and 1 m (11%) of deflection in the second adjacent probes. Probes further away did not move by a detectable amount and the results were the same for probes at the edge and center of the array. On a percentage basis, this cross talk magnitude is approximately equal to the cross talk in thermally actuated arrays (∼20%), but with a significantly reduced array pitch. Once the actuator performance was known, the arrays were demonstrated by writing 1-octadecanethiol (ODT) patterns on an electrically grounded gold surface. ODT was chosen because it is chemically robust, can be coated onto probes in a variety of ways, can be imaged using lateral force microscopy (LFM), and anchors to the surface using the same gold–sulfur bond that anchors many other DPN chemistries. The writing surface was a polished silicon substrate with a 5 nm chromium layer and 20 nm gold layer deposited by thermal evaporation. The first experiment was to determine if the electric field affects the deposition process. In this experiment, the probe array was coated using a vapor coating method [16]. In this method, solid ODT is liquefied by heating it to 70 ◦ C. Hot ODT vapor from the source then condenses on the cool probe surface for 20 min to create a uniform coating. The coated array was installed in the AFM and lowered until the probe tips made contact with the gold surface. It was then depressed an additional 3 m to prevent the tips from releasing under an actuation voltage of 175 V. A serpentine pattern was then drawn at 0.2 m/s
Fig. 9. DPN writing while alternating the counter electrode potential. If the tip is prevented from releasing from the surface, changes in counter electrode potential have no visible effect on the deposition process. The line becomes narrower as the ink concentration at the tip decreases.
while the probes were intermittently actuated. The temperature and relative humidity of the system was 25 ◦ C and 33%. The resulting DPN pattern is shown in Fig. 9. In the first leg, the regions where the counter electrode was charged and grounded are indicated. The line in this region shows no difference between these two states. During the second leg, the probe began releasing from the surface during actuation due to the decreasing adhesion force caused by decreasing ODT ink concentration (i.e. decreasing amount of ink on the probe tip). In the LFM image, the result of the probe releasing is clearly visible. When this was observed, the actuation voltage was lowered to 150 V to prevent tip release for the remainder of the pattern. This experiment shows that the electric field is not visibly affecting the ODT deposition process. When an actuation voltage is applied, there is no change in the appearance of the ODT line unless the probe actually releases from the surface. The average line width of the first leg is 150 nm. A second experiment was conducted to demonstrate controlled probe actuation. For this experiment, the array was immersion coated instead of vapor coated. Since vapor coating involves heating, the vapor may contain a larger percentage of low molecular weight contaminants than the source material. It is possible that this contamination may play a role in the formation of the “halo” that surrounds the line in Fig. 9 and has been observed in other studies [11,17]. If so, immersion coating may reduce the small-molecule contamination since the ink source is not fractionated by heating. It also helps to demonstrate that the array can be wet released since immersion coating is used to coat probes with water-solvated inks. In the immersion coating method [18], the ODT ink is dissolved in ethanol. The probe array is then dipped in the solution, removed, and dried with nitrogen to leave an ODT residue on the probe surface. The EA-DPN probes are stiff enough that the nitrogen air stream provides enough agitation to keep them from adhering to the holder as they dry. The first part of the experiment demonstrates single-probe actuation. In this trial, the array was drawn across the surface at 0.05 m/s while actuating a probe at 50 mHz and 190 V. The temperature and relative humidity were 24 ◦ C and 49%. The resulting dashed line is shown in a seven-image LFM composite in Fig. 10a. This image shows that the probe releases cleanly and
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
509
Fig. 10. Single- and multi-probe patterns made by an EA-DPN array. (a) The dashed line is 30–40 nm wide and was created by periodically actuating a probe while translating over the surface. (b)–(e) Differential actuation demonstrated by simultaneously writing four different patterns with four probes in the array. Average line width is 30–40 nm. (f) The narrowest line written with this probe design is 25 nm wide. This compares favorably with the narrowest lines made by commercial probes.
completely from the surface. The average line width is 30–40 nm and there is no halo. The second part of the experiment demonstrates differential actuation. In this trial, a pattern consisting of four parallel lines was followed by each probe in the array. During the run, different probes were actuated at different times, causing them to skip writing one or more of the lines. Fig. 10b–e shows the results from four of the probes, illustrating the four combinations that were written simultaneously. Actuator cross talk was mitigated by depressing the array against the surface approximately 2 m beyond the point of first contact. The writing conditions were the same as those in the previous experiment and the line width was also in the 30–40 nm range. The final image in this set, Fig. 10f, shows the narrowest line drawn by this probe design. It is continuous and 25 nm
wide. This value is comparable to the best results obtainable with commercial silicon nitride, contact mode cantilevers. As discussed in the introduction, electrostatic actuation was pursued, in part, to reduce the pitch of the probe array. As the DPN process is improved and feature sizes are reduced, greater resolution and accuracy will be required from the AFM scanner. Since the digital controller’s resolution is generally fixed, this is obtained by limiting the scanner’s range of motion. When the scanner range is so limited, the array pitch must also be reduced if we wish to connect adjacent probe patterns. A second reason to reduce the array pitch is to speed multiink patterning tasks. Previous multi-ink efforts required several probe chips, each coated with a different ink, and a time consuming process of sequentially installing and aligning each probe with the existing pattern [16]. Multi-probe arrays can speed this
510
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
permits arrays with a smaller pitch than thermal actuation and allows temperature sensitive inks to be deposited without damage. This paper describes a 10-probe array with a 30 m pitch and an operating voltage of 190 V. Individual probe operation, simultaneous multi-probe operation, and multi-probe pattern overlap were demonstrated by depositing ODT on a gold surface. In most cases, the average line width was 30–40 nm and the thinnest line created was 25 nm. Acknowledgements Fig. 11. If the array pitch is smaller than the maximum scanner deflection, some surface regions are reachable by more than one probe. If different probes are coated with different inks, then multi-ink patterns can be written in these overlap regions without the need for supplemental alignment processes.
This work was supported primarily by the Nanoscale Science and Engineering Initiative of the National Science Foundation under NSF Award Number EEC-0118025. Additional support was provided by the US Department of Defense under contract number ARMY NU0650300F446646. References
Fig. 12. Simultaneously written line groups where each group consists of four lines written with four adjacent probes without using alignment markers. Each line is 3 m long and each LFM scan is 5 m wide. The numbers represent the probe number in the array as shown in Fig. 11.
process by placing several probe tips, each with a different ink, in close proximity to the pattern being constructed. If the array pitch is less than the scanner’s range of motion, multiple tips can be maneuvered into the same region without the need to replace probe chips or write alignment markers on the surface. This concept is illustrated in Fig. 11. The EA-DPN array has a 30 m pitch and is installed on a 100 m scanner. Thus, there is a 10 m wide region between each probe that can be reached by the two closest tips on either side. If each probe is coated with a different ink, then each group of four probes can work together to create a 4-ink pattern without the need for an intermediate alignment process. This concept was demonstrated by simultaneously writing seven blocks of multi-probe patterns as shown in Fig. 12. In each block, each of the four lines was written by a different probe. The targeted spacing between each line was 1 m. The actual spacing averaged 966 nm with a standard deviation of 225 nm. This was accomplished without using alignment markers and can be improved by optimizing the scanner X–Y feedback loops. The write speed was 0.05 m/s and the probes were freshly coated with ink, resulting in an average line width of 277 nm. 5. Conclusion This paper presents the first successful use of electrostatic actuation to create a DPN probe array. Electrostatic actuation
[1] R.D. Piner, J. Zhu, F. Xu, S. Hong, C.A. Mirkin, Dip-pen nanolithography, Science 283 (1999) 661–663. [2] W.P. King, T.W. Kenny, K.E. Goodson, G. Cross, M. Despont, U. Durig, M.I. Lutwyche, H. Rothuizen, G.K. Binnig, P. Vettiger, Design of AFM cantilevers for combined thermomechanical data writing and reading, in: Presented at Solid-State Sensor and Actuator Workshop, Hilton Head Island, SC, USA, 2000. [3] E.S. Snow, P.M. Campbell, F.K. Perkins, Nanofabrication with proximal probes, in: Proceedings of the IEEE, vol. 85, 1997, pp. 601–611. [4] D.S. Ginger, H. Zhang, C.A. Mirkin, The evolution of dip-pen nanolithography, Angew. Chem. Int. Ed. 43 (2004) 30–45. [5] M. Zhang, D. Bullen, K.S. Ryu, C. Liu, Passive and active probe arrays for dip-pen nanolithography, in: Presented at Proceedings of the 2001 First IEEE Conference on Nanotechnology (IEEE-NANO, 2001), Maui, Hawaii, 2001. [6] M. Zhang, D. Bullen, S.-W. Chung, S. Hong, K.S. Ryu, Z. Fan, C.A. Mirkin, C. Liu, A MEMS nanoplotter with high-density parallel dip-pen nanolithography probe arrays, Nanotechnology 13 (2002) 212–217. [7] J. Zou, X. Wang, D.A. Bullen, K.S. Ryu, C. Liu, C.A. Mirkin, A mouldand-transfer technology for fabricating scanning probe microscopy probes, J. Micromech. Microeng. 14 (2004) 204–211. [8] X. Wang, K.S. Ryu, D.A. Bullen, J. Zou, H. Zhang, C.A. Mirkin, C. Liu, Scanning probe contact printing, Langmuir 19 (2003) 8951– 8955. [9] D. Bullen, X. Wang, J. Zou, S.-W. Chung, C. Liu, C.A. Mirkin, Development of parallel dip pen nanolithography probe arrays for high throughput nanolithography, in: Presented at Rapid Prototyping Technologies, Boston, MA, USA, 2003. [10] D. Bullen, S.-W. Chung, X. Wang, J. Zou, C.A. Mirkin, C. Liu, Parallel dip-pen nanolithography with arrays of individually addressable cantilevers, Appl. Phys. Lett. 84 (2004) 789–791. [11] D. Bullen, X. Wang, J. Zou, S.-W. Chung, C.A. Mirkin, C. Liu, Design, fabrication, and characterization of thermally actuated probe arrays for dip pen nanolithography, J. Microelectromech. Syst. 13 (2004) 594– 602. [12] J. Brugger, N. Blamf, P. Renaudb, N.F. de Rooija, Microlever with combined integrated sensor/actuator functions for scanning force microscopy, Sens. Actuators A 43 (1994) 339–345. [13] M. Zhang, Development of micromachined probe arrays for nanolithography applications, M.S. Thesis, University of Illinois at Urbana–Champaign, Urbana, IL, 2001, 64 pp. [14] E.M. Chow, H.T. Soh, H.C. Lee, J.D. Adams, S.C. Minne, G. Yaralioglu, A. Atalar, C.F. Quate, T.W. Kenny, Integration of through-wafer interconnects with a two-dimensional cantilever array, Sens. Actuators A 83 (2000) 118–123.
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511 [15] G.J. O’Brien, D.J. Monk, L. Lin, MEMS cantilever beam electrostatic pull-in model, in: Presented at Design, Characterization, and Packaging for MEMS and Microelectronics II, Adelaide, Australia, 2001. [16] S. Hong, J. Zhu, C.A. Mirkin, Multiple ink nanolithography: toward a multiple-pen nano-plotter, Science 286 (1999) 523–525. [17] P.E. Sheehan, L.J. Whitman, Thiol diffusion and the role of humidity in dip pen nanolithography, Phys. Rev. Lett. 88 (2002) 156104/ 1-4. [18] L.M. Demers, C.A. Mirkin, Combinatorial templates generated by dippen nanolithography for the formation of two-dimensional particle arrays, Angew. Chem. Int. Ed. 40 (2001) 3069–3071.
Biographies David Bullen received his BS degree in mechanical engineering at Colorado State University in 2000, and his PhD in mechanical engineering from the University of Illinois at Urbana-Champaign in 2004. His interests lie in the development and optimization of multiscale systems with coupled heat transfer, thermodynamic, and material constraints. He has written or coauthored 25 papers and holds two patents in the area of scanning probe lithography. In 2004, he received the Nanoscale Science and Engineering Center Outstanding Research Award for his contributions to the field
511
of nanoscience. He is currently a Senior Engineer at Bettis Atomic Power Laboratory. Chang Liu received his MS and PhD degree from Caltech in 1991 and 1995, respectively. His PhD thesis was titled “Micromachined sensors and actuators for fluid mechanics applications”. In January 1996, he joined the Microelectronics Laboratory of the University of Illinois as a postdoctoral researcher. In January 1997, he became an assistant professor with major appointment in the Electrical and Computer Engineering department and minor appointment in the Mechanical and Industrial Engineering Department. In 2003, he was promoted to Associate Professor with tenure. His research interests cover micro sensors, microfluidic lab-on-a-chip systems, and applications of MEMS for nanotechnology. He has 13 years of research experience in the MEMS area and has published 100 technical papers. Prof. Liu received the NSF CAREER award in 1998 and is currently an Associate Editor of the IEEE Sensors Journal. He teaches undergraduate and graduate courses covering the areas of MEMS, solid state electronics, and heat transfer. He won a campus “Incomplete list of teachers ranked as excellent” honor in 2001. Prof. Liu is currently a senior member of the IEEE. His work has been cited in popular media many times. Dr. Liu is a co-founder and a member of technical advisor board of NanoInk Corporation. He has consulted for several major MEMS companies. In 2002, he has been elected to the “Inventor Wall of Fame” by the Office of Technology Management of the University of Illinois.
Electrostatically actuated dip pen nanolithography probe arrays David Bullen, Chang Liu ∗ Micro Actuators, Sensors, and Systems Group, Micro and Nanotechnology Laboratory, University of Illinois, Urbana, IL 61801, USA Received 13 December 2004; received in revised form 18 August 2005; accepted 1 September 2005 Available online 10 October 2005
Abstract Dip pen nanolithography (DPN) is a method of creating nanoscale chemical patterns on surfaces using an atomic force microscope (AFM) probe. Until now, efforts to increase the process throughput have focused on passive multi-probe arrays and active arrays based on thermal bimetallic actuation. This paper describes the first use of electrostatic actuation to create an active DPN probe array. Electrostatic actuation offers the benefit of actuation without the probe heating required for thermal bimetallic actuation. Actuator cross talk between neighboring probes is also reduced, permitting more densely spaced probe arrays. The array presented here consists of 10 cantilever probes, where each is 120 m long and 20 m wide. Each cantilever probe is actuated by the electrostatic force between the probe and a built-in counter electrode with a 20–25 m gap. The tip-to-tip probe spacing, also called the array pitch, is 30 m. Patterns of 1-octadecanethiol were created on gold surfaces to demonstrate single-probe actuation, simultaneous multi-probe actuation, and overlap of patterns from adjacent probes. The minimum line width was 25 nm with an average line width of 30–40 nm. © 2005 Elsevier B.V. All rights reserved. Keywords: Dip pen nanolithography (DPN); Electrostatic actuator; Atomic force microscope; Nanolithography
1. Introduction Dip pen nanolithography (DPN) [1] is a process for directly depositing chemical patterns on surfaces with nanoscale dimensions and precision. To perform DPN, an atomic force microscope (AFM) probe is coated with a chemical and slowly translated over the surface in contact mode. The chemical diffuses from the tip to the surface to create the pattern. Researchers have used DPN to deposit a variety of organic, inorganic, and biological inks on metal, semiconductor, and oxide surfaces. This is in contrast to other scanning probe methods such as thermomechanical data storage [2], or local field oxidation of the surface [3], although some aspects of the probe technology are related. DPN is becoming increasingly popular as a unique tool for nanoscale surface engineering. We refer interested readers to [4] for an overview of DPN research and applications. Despite its usefulness in research, single-probe DPN is a slow process and not suitable for industrialization. One way to increase the patterning throughput is to perform the process in a parallel fashion with multiple probes. Passive probe
∗
Corresponding author. Tel.: +217 333 4051; fax: +217 244 6375. E-mail address: [email protected] (C. Liu).
0924-4247/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2005.09.001
arrays are one way to address this need [5]. These are arrays of identical, closely spaced AFM probes affixed to a common holder chip. When placed in contact with the surface, each probe writes the same pattern simultaneously. Passive DPN arrays have been produced using silicon nitride and silicon [6], permalloy, polyimide, SU-8 photoepoxy [7], and polydimethylsilizane [8]. An additional level of functionality is achieved by including a Z-axis actuator on each probe to create an active probe array. The actuator allows individual probes to be lifted from the surface independently of the array to halt the deposition process. As a result, different patterns can be created by each probe even though they all follow the same path on the surface. There are several possible actuation methods, including thermal bimetallic, piezoelectric, and electrostatic actuation (sometimes called capacitive actuation). Each has advantages and disadvantages for a given patterning task. The only mechanism that has been studied in detail for DPN applications is thermal bimetallic actuation [9,10]. In thermally actuated DPN (TA-DPN) arrays, the probe is a rectangular cantilever formed from layers of high and low thermal expansion coefficient material, typically gold and silicon or silicon nitride. The gold film is patterned to form an ohmic heater at the base of the probe. Passing current through the heater warms the entire
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
probe, causing it to curl away from the substrate, suspending deposition. When the probe is cooled, the tip returns to the surface and deposition begins again as it does in a passive array. Thermal bimetallic actuation is robust, simple, and effective for many, but not all, DPN applications. Its main limitation is the potential effect of the temperature excursion on the chemical ink. TA-DPN arrays typically warm 10–40 ◦ C above ambient during actuation [11]. Some ink compounds (proteins for example) may be adversely affected by this elevated temperature. Another problem with thermal bimetallic actuation is thermal cross talk between adjacent probes. Thermal cross talk refers to unwanted probe deflection caused by heat transfer between probes. In TA-DPN arrays, thermal cross talk is mainly due to heat conduction and convection through the air, rather than conduction through the substrate or radiation. These become more pronounced as the array pitch is reduced. To date, the smallest pitch achieved with a manageable cross talk is 100 m [11]. In these devices, an actuated-probe deflection of 16.3 m results in 3.7 m (23%) of deflection in the first adjacent probes and 1.9 m (12%) of deflection in the second adjacent probes. Previous DPN experience suggests that the maximum allowable cross talk value is about 30–40% before its impact on usability becomes unacceptable. Since the array pitch is an important metric (it relates directly to pattern throughput and overlap capability), a smaller value is highly desirable, but difficult to obtain with thermal actuation. An alternative to thermal bimetallic actuation is electrostatic actuation. Electrostatic actuation offers a way to deposit temperature sensitive chemicals without damage. Cantilever electrostatic actuators scale well to small sizes and can generate large tip deflections relative to the cantilever length. Electrostatic actuation has already been used to produce actuated AFM probes [12], although the design was not suitable for multi-probe arrays where each probe is individually addressable. A previous attempt to adapt the method for DPN [13] yielded devices that had an excessively large array pitch, produced inadequate tip deflection, and could not be demonstrated with multi-probe lithography. In this paper, we present a new active DPN array concept based on cantilever electrostatic actuation. The concept presented here is effective, scaleable, and proven in DPN patterning tasks.
505
Fig. 1. The operating concept of an EA-DPN probe. When the counter electrode is grounded, the probe tip is in contact with the surface and performing DPN. To actuate the probe and suspend DPN, a potential is applied to the counter electrode to pull the tip off the surface.
and surface. Since all the probes are fabricated from the same film, this configuration also eliminates the need to individually wire bond probes to ground pads on the holder. This allows the design to be scaled to a larger number of probes without making the assembly process more difficult. Fig. 2 shows an SEM image of the probes and counter electrodes. Each probe is 20 m wide, 120 m long, and 0.65 m thick. The spring constant is approximately 0.11 N/m. The array pitch is 30 m, which is much smaller than the 100 m achieved with thermal bimetallic actuation. The probe tips are 7–8 m tall and have an apex radius of less than 100 nm. The probe cantilevers are fabricated from a film stack that consists of silicon nitride, chromium, and gold. As shown later in the fabrication section, this stack is oriented so that the gold film faces the holder and the silicon nitride faces the writing surface. The silicon nitride layer consists of a zero-stress component and a tensile component. They are oriented to cause the probe to deflect away from the holder after release, as shown in Fig. 2. This increases the clearance between the probe cantilevers and the writing surface during lithography. At the base, the gap between the probe and the counter electrode is 20 m. At the tip, the gap is approximately 25 m. Fig. 3 shows an SEM image of a packaged array. The probes are fabricated as part of a 2 mm × 2 mm film, which is separated from the glass holder by the insulator. The film is electrically
2. Description Fig. 1 illustrates the operating concept of an electrostatically actuated DPN (EA-DPN) probe. In this design, the probe acts as an electrode and is separated from its counter electrode by an insulator. There is one counter electrode per probe in the 10probe array. All the probes are fabricated from the same film and are attached to a holder with a common insulator. The holder is made of glass and the insulator is SU-8 photoepoxy. The actuation potentials are applied to the counter electrodes on the holder. The probes are grounded by connecting them to a ground trace with conducting paste. Applying a voltage to the counter electrode generates an attractive force that pulls the tip off the surface. If the lithography surface is grounded, this arrangement eliminates any electric field or actuator force between the tip
Fig. 2. An SEM image of an EA-DPN probe array. The cantilevers and tips are fabricated as part of a common thin film that is attached to a holder by an SU-8 insulator. The inset shows a single-probe tip.
506
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
Fig. 3. An SEM image of a packaged EA-DPN probe array including the actuator and ground traces on the holder surface.
Fig. 5. A packaged EA-DPN probe assembly. The glass holder carries actuation leads from an external connection to the counter electrodes under each probe.
connected to a ground pad by conducting paste. The use of a large film minimizes the risk that the paste will wet around the edge of the device and short the actuator traces when it is applied. The paste is carbon-based and becomes electrically conducting after the solvent evaporates. For this work the paste was applied manually. More efficient and controllable methods for applying the paste and providing electrical conduction pathways will be developed in the future. Many works in the past have successfully dealt with problem of high density connections to one- and two-dimensional arrays [14]. There are 10 actuator traces on the holder surface, one for each probe. Five traces approach the array from each side and pass under the insulator on their way to the counter electrodes. As shown in Fig. 4, there is a ground trace between each actuator trace. This ground reduces the electrostatic interaction between the actuator traces. Without it, when one probe is charged, the electric field around its actuator trace will induce a charge in the adjacent traces, resulting in actuation of the adjacent probes. The grounds reduce this effect significantly at the cost of a small increase in system capacitance.
A photograph of the complete probe assembly is shown in Fig. 5. The probe holder is a glass slide that is 35 mm long, 6 mm wide, and 1 mm thick. The actuator and ground traces extend from the probe array to a region where they match up to a 20-conductor, 0.5 mm pitch cable. Using glass allows the probes to remain visible through the holder during lithography. The fact that the grounded metal film on each probe faces the counter electrode is a potentially catastrophic problem if a cantilever were to snap-in during actuation. Three design and operating features mitigate this problem. First, the probes are always operated at less than their experimentally estimated snapin voltage. Second, as shown in Fig. 6, the counter electrode is shorter than the probe to prevent them from touching if the probe pulls in. The shorter counter electrode also allows the backside of the probe tip to be visible through the glass holder during lithography. Finally, there is a 200 M resistance inserted in each actuator’s electrical pathway. This limits the steady-state short circuit current to less than 1.5 A in the event that the probe touches its counter electrode. Extensive use during DPN has demonstrated the adequacy of these measures.
Fig. 4. Actuator and ground trace layout. Ground traces reduce the magnitude of the electric field interaction between the actuator traces. They extend from the external power supply to the edge of the probe array.
Fig. 6. An SEM image of a snapped-in probe. The probe is longer than the counter electrode to prevent it from touching the counter electrode if it snaps-in. This tip is from an earlier array that was larger, but otherwise identical, to those pictured in Fig. 2. Unlike the smaller probes, these devices were capable of being charged to the snap-in condition by the SEMs current.
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
507
(a) Fabrication of the probe array starts with an oxidized (1 0 0) silicon wafer. Square openings in the oxide layer are etched where each tip will be located. Pyramid trenches are then anisotropically etched in the silicon substrate and the remaining oxide is removed. (b) The probe array film stack materials are then deposited in ˚ zinc oxide, 800 A ˚ tensile silicon the following order: 1000 A ˚ ˚ chromium, nitride, 5000 A stress-free silicon nitride, 50 A ˚ and 200 A gold. All the layers are patterned with the same mask to form the probe array film. A 20 m thick layer of SU-8 25 is then spin coated and patterned to form the insulator. The wafer is then diced into individual dies. (c) The holder fabrication begins by trimming a glass micro˚ chromium scope slide to 35 mm × 6 mm × 1 mm. A 150 A ˚ adhesion layer and 2000 A gold electrode layer are then deposited and patterned to form the ground and counter electrode traces. The entire surface is then coated with a 2–4 m thick layer of SU-8 5 and soft baked. (d) The probe array is assembled by bonding the probe chip to the glass holder. This is done by placing the two in contact with the probes and counter electrodes aligned. The probe chip is heated to 70–80 ◦ C to allow the uncured SU-8 on the glass holder to reflow and wet the cured SU-8 insulator on the probe array. This creates a temporary joint when the probe chip is cooled. The joint is made permanent by exposing the reflow region to UV light through the glass holder and heating it to crosslink the uncured SU-8. The remaining unexposed SU-8 is then removed with SU-8 developer. (e) The probe array is released from the silicon chip by etching the zinc oxide sacrificial layer between the array and the silicon die. After release, the gold layer on the probe array film is connected to the ground electrode at the rear of the array by manually applying conducting carbon paste. 4. Experimental Fig. 7. Process flow for producing an EA-DPN probe assembly.
3. Fabrication The fabrication procedure is based on a mould-and-transfer process [7] and is illustrated in Fig. 7. The probe array and holder are fabricated separately then bonded together at the end of the process. The deposition and etching methods for each film are listed in Table 1: Table 1 Deposition and etching method used for each film in the fabrication process Material
Deposition
Etchant
Si SiO2 ZnO Si3 N4 Cr Au SU-8
Substrate Thermal oxidation RF magnetron sputtering PECVD Thermal evaporation Thermal evaporation Spin on
35% KOH Buffered HF 0.5% HCl (step b), 38% HCl (step e) Tetrafluoromethane RIE Cr etchant (CEP-200, Microchrome) Au etchant (TFA, Transene) SU-8 developer (MicroChem)
4.1. Deflection voltage In order to verify the effectiveness of electrostatic actuation in DPN applications, several experiments were carried out. The first was to determine a suitable operating voltage for the device to avoid a snap-in condition [15]. Although parallel-plate electrostatic actuators produce most of their deflection near the snap-in voltage, it is preferable to avoid the snap-in condition for two reasons. First, near this point the cantilever position is very sensitive to nuances in the mechanical and electrostatic state of the system. This makes tip displacement from probe to probe very non-uniform. Second, the repeated snap-in events may create particles that could interfere with the writing results. Fig. 8 shows the average displacement versus voltage relationship of several probes in the array. To illustrate the impact of the electric field on tip deflection, two probes, located at the edge and center of the array, are plotted as well. These two probes are numbered 1 and 5, respectively. As Fig. 8 shows, the displacement within an array becomes more non-uniform as voltage increases. At counter electrode voltages above 260 V, probe snap-in events are common. Based on this data, a safe
508
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
Fig. 8. EA-DPN probe deflection vs. counter electrode voltage. Circles represent the average deflection of all 10 probes. The solid line is a polynomial fit to the average deflection. The deflection of probe 1 (at the edge of the array) and probe 5 (in the center of the array) is included for comparison. Error bars represent the measurement error.
operating voltage of 190 V was chosen for lithography operations. The actuator cross talk was also measured in this experiment. In TA-DPN arrays, experience has shown that thermal cross talk between probes is one of the greatest impediments to reducing the pitch of the array. Electrostatic actuators also suffer from cross talk due to fringe electric fields. To determine the cross talk magnitude, individual EA-DPN probes were actuated while measuring the deflection of the adjacent, unactuated probes. In these single-probe experiments, an actuated-probe deflection of 9 m was accompanied by 3 m (33%) of deflection in the first adjacent probes and 1 m (11%) of deflection in the second adjacent probes. Probes further away did not move by a detectable amount and the results were the same for probes at the edge and center of the array. On a percentage basis, this cross talk magnitude is approximately equal to the cross talk in thermally actuated arrays (∼20%), but with a significantly reduced array pitch. Once the actuator performance was known, the arrays were demonstrated by writing 1-octadecanethiol (ODT) patterns on an electrically grounded gold surface. ODT was chosen because it is chemically robust, can be coated onto probes in a variety of ways, can be imaged using lateral force microscopy (LFM), and anchors to the surface using the same gold–sulfur bond that anchors many other DPN chemistries. The writing surface was a polished silicon substrate with a 5 nm chromium layer and 20 nm gold layer deposited by thermal evaporation. The first experiment was to determine if the electric field affects the deposition process. In this experiment, the probe array was coated using a vapor coating method [16]. In this method, solid ODT is liquefied by heating it to 70 ◦ C. Hot ODT vapor from the source then condenses on the cool probe surface for 20 min to create a uniform coating. The coated array was installed in the AFM and lowered until the probe tips made contact with the gold surface. It was then depressed an additional 3 m to prevent the tips from releasing under an actuation voltage of 175 V. A serpentine pattern was then drawn at 0.2 m/s
Fig. 9. DPN writing while alternating the counter electrode potential. If the tip is prevented from releasing from the surface, changes in counter electrode potential have no visible effect on the deposition process. The line becomes narrower as the ink concentration at the tip decreases.
while the probes were intermittently actuated. The temperature and relative humidity of the system was 25 ◦ C and 33%. The resulting DPN pattern is shown in Fig. 9. In the first leg, the regions where the counter electrode was charged and grounded are indicated. The line in this region shows no difference between these two states. During the second leg, the probe began releasing from the surface during actuation due to the decreasing adhesion force caused by decreasing ODT ink concentration (i.e. decreasing amount of ink on the probe tip). In the LFM image, the result of the probe releasing is clearly visible. When this was observed, the actuation voltage was lowered to 150 V to prevent tip release for the remainder of the pattern. This experiment shows that the electric field is not visibly affecting the ODT deposition process. When an actuation voltage is applied, there is no change in the appearance of the ODT line unless the probe actually releases from the surface. The average line width of the first leg is 150 nm. A second experiment was conducted to demonstrate controlled probe actuation. For this experiment, the array was immersion coated instead of vapor coated. Since vapor coating involves heating, the vapor may contain a larger percentage of low molecular weight contaminants than the source material. It is possible that this contamination may play a role in the formation of the “halo” that surrounds the line in Fig. 9 and has been observed in other studies [11,17]. If so, immersion coating may reduce the small-molecule contamination since the ink source is not fractionated by heating. It also helps to demonstrate that the array can be wet released since immersion coating is used to coat probes with water-solvated inks. In the immersion coating method [18], the ODT ink is dissolved in ethanol. The probe array is then dipped in the solution, removed, and dried with nitrogen to leave an ODT residue on the probe surface. The EA-DPN probes are stiff enough that the nitrogen air stream provides enough agitation to keep them from adhering to the holder as they dry. The first part of the experiment demonstrates single-probe actuation. In this trial, the array was drawn across the surface at 0.05 m/s while actuating a probe at 50 mHz and 190 V. The temperature and relative humidity were 24 ◦ C and 49%. The resulting dashed line is shown in a seven-image LFM composite in Fig. 10a. This image shows that the probe releases cleanly and
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
509
Fig. 10. Single- and multi-probe patterns made by an EA-DPN array. (a) The dashed line is 30–40 nm wide and was created by periodically actuating a probe while translating over the surface. (b)–(e) Differential actuation demonstrated by simultaneously writing four different patterns with four probes in the array. Average line width is 30–40 nm. (f) The narrowest line written with this probe design is 25 nm wide. This compares favorably with the narrowest lines made by commercial probes.
completely from the surface. The average line width is 30–40 nm and there is no halo. The second part of the experiment demonstrates differential actuation. In this trial, a pattern consisting of four parallel lines was followed by each probe in the array. During the run, different probes were actuated at different times, causing them to skip writing one or more of the lines. Fig. 10b–e shows the results from four of the probes, illustrating the four combinations that were written simultaneously. Actuator cross talk was mitigated by depressing the array against the surface approximately 2 m beyond the point of first contact. The writing conditions were the same as those in the previous experiment and the line width was also in the 30–40 nm range. The final image in this set, Fig. 10f, shows the narrowest line drawn by this probe design. It is continuous and 25 nm
wide. This value is comparable to the best results obtainable with commercial silicon nitride, contact mode cantilevers. As discussed in the introduction, electrostatic actuation was pursued, in part, to reduce the pitch of the probe array. As the DPN process is improved and feature sizes are reduced, greater resolution and accuracy will be required from the AFM scanner. Since the digital controller’s resolution is generally fixed, this is obtained by limiting the scanner’s range of motion. When the scanner range is so limited, the array pitch must also be reduced if we wish to connect adjacent probe patterns. A second reason to reduce the array pitch is to speed multiink patterning tasks. Previous multi-ink efforts required several probe chips, each coated with a different ink, and a time consuming process of sequentially installing and aligning each probe with the existing pattern [16]. Multi-probe arrays can speed this
510
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511
permits arrays with a smaller pitch than thermal actuation and allows temperature sensitive inks to be deposited without damage. This paper describes a 10-probe array with a 30 m pitch and an operating voltage of 190 V. Individual probe operation, simultaneous multi-probe operation, and multi-probe pattern overlap were demonstrated by depositing ODT on a gold surface. In most cases, the average line width was 30–40 nm and the thinnest line created was 25 nm. Acknowledgements Fig. 11. If the array pitch is smaller than the maximum scanner deflection, some surface regions are reachable by more than one probe. If different probes are coated with different inks, then multi-ink patterns can be written in these overlap regions without the need for supplemental alignment processes.
This work was supported primarily by the Nanoscale Science and Engineering Initiative of the National Science Foundation under NSF Award Number EEC-0118025. Additional support was provided by the US Department of Defense under contract number ARMY NU0650300F446646. References
Fig. 12. Simultaneously written line groups where each group consists of four lines written with four adjacent probes without using alignment markers. Each line is 3 m long and each LFM scan is 5 m wide. The numbers represent the probe number in the array as shown in Fig. 11.
process by placing several probe tips, each with a different ink, in close proximity to the pattern being constructed. If the array pitch is less than the scanner’s range of motion, multiple tips can be maneuvered into the same region without the need to replace probe chips or write alignment markers on the surface. This concept is illustrated in Fig. 11. The EA-DPN array has a 30 m pitch and is installed on a 100 m scanner. Thus, there is a 10 m wide region between each probe that can be reached by the two closest tips on either side. If each probe is coated with a different ink, then each group of four probes can work together to create a 4-ink pattern without the need for an intermediate alignment process. This concept was demonstrated by simultaneously writing seven blocks of multi-probe patterns as shown in Fig. 12. In each block, each of the four lines was written by a different probe. The targeted spacing between each line was 1 m. The actual spacing averaged 966 nm with a standard deviation of 225 nm. This was accomplished without using alignment markers and can be improved by optimizing the scanner X–Y feedback loops. The write speed was 0.05 m/s and the probes were freshly coated with ink, resulting in an average line width of 277 nm. 5. Conclusion This paper presents the first successful use of electrostatic actuation to create a DPN probe array. Electrostatic actuation
[1] R.D. Piner, J. Zhu, F. Xu, S. Hong, C.A. Mirkin, Dip-pen nanolithography, Science 283 (1999) 661–663. [2] W.P. King, T.W. Kenny, K.E. Goodson, G. Cross, M. Despont, U. Durig, M.I. Lutwyche, H. Rothuizen, G.K. Binnig, P. Vettiger, Design of AFM cantilevers for combined thermomechanical data writing and reading, in: Presented at Solid-State Sensor and Actuator Workshop, Hilton Head Island, SC, USA, 2000. [3] E.S. Snow, P.M. Campbell, F.K. Perkins, Nanofabrication with proximal probes, in: Proceedings of the IEEE, vol. 85, 1997, pp. 601–611. [4] D.S. Ginger, H. Zhang, C.A. Mirkin, The evolution of dip-pen nanolithography, Angew. Chem. Int. Ed. 43 (2004) 30–45. [5] M. Zhang, D. Bullen, K.S. Ryu, C. Liu, Passive and active probe arrays for dip-pen nanolithography, in: Presented at Proceedings of the 2001 First IEEE Conference on Nanotechnology (IEEE-NANO, 2001), Maui, Hawaii, 2001. [6] M. Zhang, D. Bullen, S.-W. Chung, S. Hong, K.S. Ryu, Z. Fan, C.A. Mirkin, C. Liu, A MEMS nanoplotter with high-density parallel dip-pen nanolithography probe arrays, Nanotechnology 13 (2002) 212–217. [7] J. Zou, X. Wang, D.A. Bullen, K.S. Ryu, C. Liu, C.A. Mirkin, A mouldand-transfer technology for fabricating scanning probe microscopy probes, J. Micromech. Microeng. 14 (2004) 204–211. [8] X. Wang, K.S. Ryu, D.A. Bullen, J. Zou, H. Zhang, C.A. Mirkin, C. Liu, Scanning probe contact printing, Langmuir 19 (2003) 8951– 8955. [9] D. Bullen, X. Wang, J. Zou, S.-W. Chung, C. Liu, C.A. Mirkin, Development of parallel dip pen nanolithography probe arrays for high throughput nanolithography, in: Presented at Rapid Prototyping Technologies, Boston, MA, USA, 2003. [10] D. Bullen, S.-W. Chung, X. Wang, J. Zou, C.A. Mirkin, C. Liu, Parallel dip-pen nanolithography with arrays of individually addressable cantilevers, Appl. Phys. Lett. 84 (2004) 789–791. [11] D. Bullen, X. Wang, J. Zou, S.-W. Chung, C.A. Mirkin, C. Liu, Design, fabrication, and characterization of thermally actuated probe arrays for dip pen nanolithography, J. Microelectromech. Syst. 13 (2004) 594– 602. [12] J. Brugger, N. Blamf, P. Renaudb, N.F. de Rooija, Microlever with combined integrated sensor/actuator functions for scanning force microscopy, Sens. Actuators A 43 (1994) 339–345. [13] M. Zhang, Development of micromachined probe arrays for nanolithography applications, M.S. Thesis, University of Illinois at Urbana–Champaign, Urbana, IL, 2001, 64 pp. [14] E.M. Chow, H.T. Soh, H.C. Lee, J.D. Adams, S.C. Minne, G. Yaralioglu, A. Atalar, C.F. Quate, T.W. Kenny, Integration of through-wafer interconnects with a two-dimensional cantilever array, Sens. Actuators A 83 (2000) 118–123.
D. Bullen, C. Liu / Sensors and Actuators A 125 (2006) 504–511 [15] G.J. O’Brien, D.J. Monk, L. Lin, MEMS cantilever beam electrostatic pull-in model, in: Presented at Design, Characterization, and Packaging for MEMS and Microelectronics II, Adelaide, Australia, 2001. [16] S. Hong, J. Zhu, C.A. Mirkin, Multiple ink nanolithography: toward a multiple-pen nano-plotter, Science 286 (1999) 523–525. [17] P.E. Sheehan, L.J. Whitman, Thiol diffusion and the role of humidity in dip pen nanolithography, Phys. Rev. Lett. 88 (2002) 156104/ 1-4. [18] L.M. Demers, C.A. Mirkin, Combinatorial templates generated by dippen nanolithography for the formation of two-dimensional particle arrays, Angew. Chem. Int. Ed. 40 (2001) 3069–3071.
Biographies David Bullen received his BS degree in mechanical engineering at Colorado State University in 2000, and his PhD in mechanical engineering from the University of Illinois at Urbana-Champaign in 2004. His interests lie in the development and optimization of multiscale systems with coupled heat transfer, thermodynamic, and material constraints. He has written or coauthored 25 papers and holds two patents in the area of scanning probe lithography. In 2004, he received the Nanoscale Science and Engineering Center Outstanding Research Award for his contributions to the field
511
of nanoscience. He is currently a Senior Engineer at Bettis Atomic Power Laboratory. Chang Liu received his MS and PhD degree from Caltech in 1991 and 1995, respectively. His PhD thesis was titled “Micromachined sensors and actuators for fluid mechanics applications”. In January 1996, he joined the Microelectronics Laboratory of the University of Illinois as a postdoctoral researcher. In January 1997, he became an assistant professor with major appointment in the Electrical and Computer Engineering department and minor appointment in the Mechanical and Industrial Engineering Department. In 2003, he was promoted to Associate Professor with tenure. His research interests cover micro sensors, microfluidic lab-on-a-chip systems, and applications of MEMS for nanotechnology. He has 13 years of research experience in the MEMS area and has published 100 technical papers. Prof. Liu received the NSF CAREER award in 1998 and is currently an Associate Editor of the IEEE Sensors Journal. He teaches undergraduate and graduate courses covering the areas of MEMS, solid state electronics, and heat transfer. He won a campus “Incomplete list of teachers ranked as excellent” honor in 2001. Prof. Liu is currently a senior member of the IEEE. His work has been cited in popular media many times. Dr. Liu is a co-founder and a member of technical advisor board of NanoInk Corporation. He has consulted for several major MEMS companies. In 2002, he has been elected to the “Inventor Wall of Fame” by the Office of Technology Management of the University of Illinois.
