- Email: [email protected]
A surface acoustic resonator with template-patterned interdigitated fingers
Sensors and Actuators A 248 (2016) 73–77
Contents lists available at ScienceDirect
Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna
Short Communication
A surface acoustic resonator with template-patterned interdigitated fingers David A. Rolfe a,b,∗ , Kristen L. Dorsey a , Jim C. Cheng c , Albert P. Pisano d a
Department of Mechanical and Aerospace Engineering, University of California, San Diego, La Jolla, CA 92093, USA Department of Mechanical Engineering, University of California, Berkeley 94720, USA Maxim Integrated, San Jose, CA 95134, USA d Jacobs School of Engineering, University of California, San Diego, La Jolla, CA 92039, USA b c
a r t i c l e
i n f o
Article history: Received 5 August 2015 Received in revised form 19 April 2016 Accepted 1 July 2016 Available online 19 July 2016 Keywords: SAW resonator Microfluidic printing Printed electronics Silver nanoparticles Nanofabrication
a b s t r a c t Printed electronics techniques show promise as high-throughput, low-cost methods for fabricating micro-scale device features. In this work, we demonstrate a SAW resonator fabricated through a printing process. Silver nanoparticle ink is introduced to a vapor-permeable template that is patterned with the desired SAW resonator metallization. After the solvent in the ink evaporates, the nanoparticles within the patterned template tightly pack to form the device metallization. Interdigitated resonator electrodes with widths and spacing ranging from 500 nm to 1 m were patterned on lithium niobate, corresponding to center frequencies of 2.1–2.4 GHz. The maximum quality factor was 13. Limitations and potential process improvements are discussed to improve resonator quality factor in future work. © 2016 Published by Elsevier B.V.
1. Introduction Nanoimprint lithography [1] and printed electronics processes [2] have recently attracted interest as potential methods for microand nanoscale metallization because they may reduce process complexity and cost over traditional fabrication methods. One such method, Advective Micromolding in vapor-Permeable Templates (AMPT) [3–6], deposits metal nanoparticles on a desired substrate using microfluidic channels to form a device metallization layer. Previously, the AMPT process has fabricated features as small as 350 nm in width [6]. In this work, we demonstrate the AMPT process’s potential towards surface acoustic wave (SAW) resonator fabrication though additive patterning of nanoparticle ink. SAW resonators are an excellent candidate for printed electronics techniques such as AMPT because the resonant frequency is defined by the pitch of patterned electrodes and only one metal layer is required. In lithium niobate, the minimum feature size to reach the 2.4 GHz industrial, scientific, and medical radio band is 420 nm [7,8], which
∗ Corresponding author at: 403 Cory Hall #1774 University of California, Berkeley, CA 94720-1774, USA. E-mail addresses: [email protected] (D.A. Rolfe), [email protected] (K.L. Dorsey), [email protected] (J.C. Cheng), [email protected] (A.P. Pisano). http://dx.doi.org/10.1016/j.sna.2016.07.001 0924-4247/© 2016 Published by Elsevier B.V.
is within the theoretical capabilities of AMPT fabrication. Four SAW resonators with electrode width of 500 nm, 600 nm, 800 nm, and 1 m are fabricated and the resonant behavior is characterized.
2. Design and fabrication A critical part of the AMPT process is the vapor-permeable template: a polymer film that is patterned with the desired device features, into which the nanoparticle ink is drawn to form the metallization [3]. The template is patterned from a master fabricated through photolithography. The master contains raised features in a layout identical to that of the desired nanoparticle pattern. In this work, the resonator electrode features (width of 1 m or less) are patterned lithographically (Fig. 1a) onto a thermally-grown SiO2 layer and etched using deep reactive ion etching (DRIE) (Fig. 1b). The wider fill channels that lead to the electrodes are lithographically patterned on an SU-8 layer that is spun onto the smaller SiO2 features. The template is composed of a poly(4-methyl-2-pentyne) (PMP) film. PMP is highly rigid (1.6 GPa [6]) and extremely permeable to organic solvents, with a permeability for ethanol of 185,000 Barrer. The combination of rigidity and permeability insures that the nanoparticle ink solvent can rapidly evaporate through the film while the patterned channels retain their shape during deposition and formation of the nanoparticle features.
74
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
contained arrays of grounded electrodes to reflect the SAW wave at the Bragg condition interval, constraining the wave within the IDT area. The grid features that connected to the IDTs and fill ports served two purposes. The large surface area served as the contact pad to test the resonator operation using a ground-source-ground (GSG) probe. These contact pads were designed as grids with the same feature width as the IDTs, rather than open areas, because a grid of channels is less prone to collapse under the induced pressure of evaporating solvent. To test the ability of the AMPT process to fabricate SAW resonators with various center frequencies, a template with four resonators was fabricated. The SAW resonators were patterned on a Y-Z cut lithium niobate (LiNbO3 ) substrate, which was chosen for its high speed of sound and coupling efficiency. The model for the center frequency of a SAW was used to determine the appropriate electrode pitch [9], fc =
v 2(w + d)
(1)
where v is the longitudinal velocity of wave propagation in the substrate, and w and d are electrode width and spacing, respectively. The target frequencies of the resonators were 860 MHz, 1.07 GHz, 1.43 GHz, and 1.72 GHz, corresponding to electrode width and spacing of 1 m, 800 nm, 600 nm, and 500 nm, respectively. Fig. 1. A schematic of the AMPT process showing (a) lithography, (b) etching of the master wafer, (c) the template on the substrate, (d) introduction of nanoparticle ink, (e) drying and removal of the template.
The template is fabricated by casting PMP dissolved in cyclohexane onto the master. The cyclohexane is evaporated slowly through a PDMS membrane over the course of three days. The slow evaporation into a high concentration cyclohexane environment ensures proper feature formation and minimal residual stress. When fully dry, the template is removed from the master by softening it with isopropanol and peeling gently. To pattern the desired features using the AMPT process, the template is placed channel-side-down on a substrate flooded with solvent (Fig. 1c). Nanoparticle ink is introduced to fill ports which are connected to features in the center of the template by means of fill channels (Fig. 1d). As solvent in the channels evaporates through the template, nanoparticle ink is drawn into the channels. The nanoparticles become increasingly concentrated and begin to densely pack as the ink solvent evaporates. After the features are fully packed and the excess solvent evaporates, the template is removed (Fig. 1e). The conductive features are patterned with DGP AP ink from Advanced Nano Products Co. Ltd., (Buyong-myeon, South Korea). This ink contains nanoparticles averaging 20 nm at about 40 wt% in ethanol. The ink was diluted to about 8 wt% during the patterning process. Features were patterned by heating the substrate to 45 ◦ C and lowering the template onto ethanol. Silver ink and ethanol were added to the fill ports over 30 min to evaporate all solvent. After solvent evaporation, the device and template were sintered in nitrogen at 250 ◦ C for 30 min. Finally, the template was removed by soaking the patterned device in acetone to swell the template and aid in delamination from the substrate. The resonator layout is shown in Fig. 2. The widths w and d are equal to each other and range from 500 nm to 1 m. The lengths la and ld range from 60 m and 5 m to 120 m and 10 m, respectively. The ratio of la to ld ’s length remains constant across all fabricated resonators. Each resonators has 40 pairs of interdigitated electrodes. A surface acoustic wave was only excited along the length marked a in Fig. 2. The remaining electrode length from the interdigitated transducer (IDT) to the fill ports (d in Fig. 2) was designed to dampen higher order modes. Gratings on either side of the IDT
3. Results and discussion The 1 m and 600 nm electrode width resonators are shown in Fig. 3. In both resonators, the fill ports are well defined, but the interdigitated electrodes are not completely patterned across the resonator. The circled areas indicate regions where the patterning is not complete. Lighter areas indicate places where silver nanoparticles have completely filled the template, whereas darker areas show uneven or incomplete patterning, or the existence of increased solvent residue. Both incomplete patterning and solvent residue increase the resistivity of the resonator features, reducing the quality factor of the resonator. These patterning errors are caused by limitation of the AMPT process. One such limitation is the use of polymers as the template materials. During patterning, the PMP and polydimethysiloxane (PDMS) backing material absorb the solvent in the nanoparticle ink and swell. Mismatch in PMP and PDMS solvent diffusion rate and solvent absorption distorts the template [10,11]. Distortion may also occur within the template. The solvent diffusion rates vary over the template due to feature size and the presence of corners [12]. The PMP along the channel walls swells more than the PMP far from the channel walls. The localized swelling causes the PMP closest to the ink channels to expand more than the rest of the template. The concentration gradient of solvent can also lead to buckling, in which solvent in the fill channels causes the bottom of the template to disproportionately swell and fold in on itself. Distortion is a concern for patterning fidelity and device functionality. The effects of swelling are visible in the pattern in Fig. 3a. The circled areas show where the swollen template buckled and delaminated from the template during patterning. As a result, the Bragg gratings are not completely formed. The circled area in Fig. 3b also shows incomplete nanoparticle ink patterning. In comparison to the IDT area in Fig. 3a, the interdigitated fingers are not clearly outlined. The real and imaginary components of each resonator’s reflection signal were measured using an Agilent Technologies E5071B network analyzer with a load termination of 50 . Each device showed a resonance between 1.1 and 1.5 GHz (Fig. 4). As predicted by the center frequency model in (1), the measured center frequency is inversely related to electrode width. A mismatch of 5%
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
75
Fig. 2. A one-port SAW resonator showing center IDT (red with green fill ports) and Bragg gratings (black with blue fill ports). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
(800 nm width) to 30% (500 nm width) exists between modeled and measured center frequency. The mismatch severity shows no dependence on resonator electrode width. In smaller channels, localized swelling widens the electrode spacing, lowering the center frequency of the smallest two resonators in Fig. 4. Conversely, the swelling of features surrounding the electrodes may have been responsible for the higher than expected center frequency in the larger resonators. The measured quality factors of the AMPT-patterned resonators range from 5.9 (500 nm) to 13.4 (1 m), which are several orders of magnitude lower than traditionally-fabricated SAW resonators [13]. The low quality factor and frequency model mismatch may be attributed to poor ink patterning and misalignment between the template and the substrate.
A current limitation of the AMPT process is patterning fidelity, especially at smaller device features. If the template distorts from solvent-induced swelling, nanoparticle ink may leak under the template and form a conductive path between the IDTs. As the electrode spacings decrease from 1 m to 500 nm, a conductive path is more likely to form between IDTs. The conductive path’s effect on Q and fc is modeled using a series RLC lumped model to represent the resonator with a parasitic resistance RP in parallel with the resonator. The R, L, and C values and the parasitic resistance RP were fit to measured data in the 1 m resonator. The real component of the RP model is shown for the 1 m resistor as a short dotted line (Fig. 4a). The fitted value of RP is approximately 50 for all patterned resonators.
Fig. 3. Resonator micrographs with (a) 1 m wide electrodes and (b) 600 nm wide electrodes.
76
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
Fig. 4. The real and imaginary components of the reflected signal for each resonator.
Two other process limitations have implications for AMPT’s use in SAW resonator fabrication. AMPT does not currently have a mechanism to align the patterning template to the substrate. The angle between the template and the piezoelectric substrate was not considered or recorded during device fabrication. This angle is critical to a high coupling coefficient: misalignment of only 10◦ results in a coupling coefficient decrease of 50% [14]. Misalignment of the patterned IDTs to the lithium niobate substrate could therefore be another source of performance loss. A fundamental limitation of the AMPT process is its use of nanoparticle ink for device metallization. The resistivity of the sintered nanoparticle IDTs was 32 nm, which is approximately twice the value of bulk silver. It is not likely that the higher ink resistivity degraded the device Q, but the resistivity could limit performance as the patterning fidelity and ink leakage issue are improved. 4. Conclusions A series of resonators with IDTs ranging from 500 nm to 1 m were fabricated from silver nanoparticles on a lithium niobate substrate using the AMPT process. Despite the low quality factor and deviation from modeled resonant frequency, these resonators demonstrate the ability of the AMPT process to pattern functional devices with less complexity than lift-off processes. From a single master fabricated with lithography or other processes, the templates for many devices may be cast. Once a template is cast, it can be used to pattern device features in less than 30 min without the need for photolithography, vacuum-processed metal deposition, or a clean room. Increased resonator performance will be achievable by primarily by limiting the parasitic resistance between the IDTs and by aligning the SAW device to improve the coupling coefficient. To reduce ink leakage between electrically isolated features, we propose adding a rigid structure to the PDMS template backing material to prevent it from swelling or deforming, and by increasing the pressure applied to the PMP template while patterning. In
future process iterations, we shall also use a thinner PMP template to reduce the effects of swelling on the deposited device pattern. To improve the electromechanical coupling and increase Q, we will incorporate an alignment mechanism into future AMPT process iterations. With these changes, the AMPT process could significantly reduce the complexity of fabricating SAW resonators for consumer electronics. Acknowledgments The authors would like to thank Fuji Electric Co., Ltd. for their financial support of this research. Fabrication was performed in the Biomolecular Nanotechnology Center at UC Berkeley and the Nano3 facility at UC San Diego. References [1] E. Forsén, D. Nilsson, W. Daniau, B. Franc¸ois, S. Ballandras, Fabrication of SAW resonators on quartz and langasite using nano-imprintlithography, in: Proceedings – IEEE Ultrasonics Symposium, 2010, pp. 2091–2094, http://dx. doi.org/10.1109/ULTSYM.2010.5936002. [2] A. Kamyshny, J. Steinke, S. Magdassi, Metal-based inkjet inks for printed electronics, Open Appl. Phys. J. 4 (2011) 19–36. [3] M.T. Demko, J.C. Cheng, A.P. Pisano, High-resolution direct patterning of gold nanoparticles by the microfluidic molding process, Langmuir 26 (22) (2010) 16710–16714, http://dx.doi.org/10.1021/la1022533 http://www.ncbi.nlm. nih.gov/pubmed/20886896. [4] M.T. Demko, S. Choi, T.I. Zohdi, A.P. Pisano, High resolution patterning of nanoparticles by evaporative self-assembly enabled by in situ creation and mechanical lift-off of a polymer template, Appl. Phys. Lett. 99 (2011) (2011) 1–4, http://dx.doi.org/10.1063/1.3671084. [5] M.T. Demko, T.P. Brackbill, A.P. Pisano, Simultaneous patterning of nanoparticles and polymers using an evaporation driven flow in a vapor permeable template, ACS J. Surf. Colloids 28 (25) (2012) 9857–9863, http://dx. doi.org/10.1021/la301587f http://www.ncbi.nlm.nih.gov/pubmed/22647075. [6] M.T. Demko, J.C. Cheng, A.P. Pisano, Rigid, vapor-permeable poly(4-methyl-2-pentyne) templates for high resolution patterning of nanoparticles and polymers, ACS Nano 6 (8) (2012) 6890–6896. [7] R. Weis, T. Gaylord, Lithium niobate: summary of physical properties and crystal structure R, Appl. Phys. A: Mater. Sci. Process. 37 (1985) 191–203, http://dx.doi.org/10.1007/BF00614817.
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77 [8] J. Bell, R. Li, Surface-acoustic-wave resonators, Proc. IEEE 64 (5) (1976), http:// dx.doi.org/10.1109/PROC.1976.10200. [9] S. Achenbach, D. Klymyshyn, T. Mappes, A. Kachayev, V. Subramanian, G. Wells, J. Mohr, Submicron-scale surface acoustic wave resonators fabricated by high aspect ratio X-ray lithography and aluminum lift-off, Microsyst. Technol. 14 (9–11) (2008) 1715–1719, http://dx.doi.org/10.1007/s00542-0070498-7. [10] T. Honda, M. Miyazaki, H. Nakamura, H. Maeda, Controllable polymerization of N-carboxy anhydrides in a microreaction system, Lab Chip 5 (2005) 812–818, http://dx.doi.org/10.1039/b505137a. [11] A. Morisato, I. Pinnau, Synthesis and gas permeation properties of poly(4-methyl-2-pentyne), J. Membr. Sci. 121 (1996) 243–250. [12] D.A. Rolfe, K.L. Dorsey, J.C. Cheng, A.P. Pisano, A model to guide template-based nanoparticle printing development, in: ASME 2015 International Technical Conference and Exhibition on Packaging and Integration of Electronic and Photonic Microsystems collocated with the ASME 2015 13th International Conference on Nanochannels, Microchannels, and Minichannels, American Society of Mechanical Engineers, 2015, V003T03A006. [13] U. Knauer, J. Machui, C.C.W. Ruplpel, S. Member, Design, fabrication, and application of GHz SAW devices, in: IEEE Microwave Theory and Techniques Society, IEEE, Denver, CO, 1997, pp. 1821–1824. [14] K. Yamanouchi, K. Shibayama, Propagation and amplification of Rayleigh waves and piezoelectric leaky surface waves in LiNbO3 , J. Appl. Phys. 43 (3) (1972) 856–862, http://dx.doi.org/10.1063/1.1661294.
Biographies
David Rolfe is a researcher in the Berkeley Micromechanical Analysis and Devices (BMAD) Lab at the University of California, Berkeley (now the PRIME lab at University of California, San Diego), as a student of Albert P. Pisano. His research interests include nanoparticle patterning, microfluidics, laser transfer processes, rapid prototyping, nano-engineering ethics, and flame ionization sensors. He received his M.S. and Ph.D. from the department of Mechanical Engineering at University of California, Berkeley, and his B.S. from Harvey Mudd College in Claremont, CA.
77
Kristen Dorsey is an assistant professor of engineering in the Picker Engineering Program at Smith College. Her research interests include the modeling of reliability and stability issues in MEMS and sensor/device fabrication using printed and soft lithography processes. Dorsey completed her postdoctoral work at UC San Diego and UC Berkeley in Albert P. Pisano’s group and received her Ph.D. in Electrical and Computer Engineering at Carnegie Mellon University in Gary Fedder’s group. Her dissertation title was “Dielectric Charging in CMOS MEMS.” Her B.S. in Electrical and Computer Engineering is from Franklin W. Olin College of Engineering in Needham, MA. Jim Cheng is senior member of the technical staff at Maxim Integrated. He has been leading efforts focused primarily on nanoprinting platforms, but also encompassing energy storage solutions, biosensors, and passive cooling for electronics. Dr. Cheng has been involved in several start-ups and most recently founded ChillFlux, Inc., to commercialize his research into novel thermal management. Dr. Cheng received his B.Eng. in Electrical Engineering from McMaster University, Canada in 2004. He was formerly a research scientist at the University of California, Berkeley, where he had previously received his M.S. and Ph.D. degrees in Electrical Engineering in 2006 and 2009, respectively. Albert P. Pisano is the dean of the Jacobs School of Engineering and holds the Walter J. Zable Chair in Engineering at University of California, San Diego. In 2001, Pisano was elected to the National Academy of Engineering. Pisano earned his undergraduate (’76) and graduate degrees (’77, ’80, ’81) in mechanical engineering at Columbia University. Prior to 2013, Pisano served on the UC Berkeley faculty, and previously was the MEMS program manager for DARPA. He has held research positions with Xerox Palo Alto Research Center, Singer Sewing Machines Corporate R&D Center and General Motors Research Labs.
Contents lists available at ScienceDirect
Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna
Short Communication
A surface acoustic resonator with template-patterned interdigitated fingers David A. Rolfe a,b,∗ , Kristen L. Dorsey a , Jim C. Cheng c , Albert P. Pisano d a
Department of Mechanical and Aerospace Engineering, University of California, San Diego, La Jolla, CA 92093, USA Department of Mechanical Engineering, University of California, Berkeley 94720, USA Maxim Integrated, San Jose, CA 95134, USA d Jacobs School of Engineering, University of California, San Diego, La Jolla, CA 92039, USA b c
a r t i c l e
i n f o
Article history: Received 5 August 2015 Received in revised form 19 April 2016 Accepted 1 July 2016 Available online 19 July 2016 Keywords: SAW resonator Microfluidic printing Printed electronics Silver nanoparticles Nanofabrication
a b s t r a c t Printed electronics techniques show promise as high-throughput, low-cost methods for fabricating micro-scale device features. In this work, we demonstrate a SAW resonator fabricated through a printing process. Silver nanoparticle ink is introduced to a vapor-permeable template that is patterned with the desired SAW resonator metallization. After the solvent in the ink evaporates, the nanoparticles within the patterned template tightly pack to form the device metallization. Interdigitated resonator electrodes with widths and spacing ranging from 500 nm to 1 m were patterned on lithium niobate, corresponding to center frequencies of 2.1–2.4 GHz. The maximum quality factor was 13. Limitations and potential process improvements are discussed to improve resonator quality factor in future work. © 2016 Published by Elsevier B.V.
1. Introduction Nanoimprint lithography [1] and printed electronics processes [2] have recently attracted interest as potential methods for microand nanoscale metallization because they may reduce process complexity and cost over traditional fabrication methods. One such method, Advective Micromolding in vapor-Permeable Templates (AMPT) [3–6], deposits metal nanoparticles on a desired substrate using microfluidic channels to form a device metallization layer. Previously, the AMPT process has fabricated features as small as 350 nm in width [6]. In this work, we demonstrate the AMPT process’s potential towards surface acoustic wave (SAW) resonator fabrication though additive patterning of nanoparticle ink. SAW resonators are an excellent candidate for printed electronics techniques such as AMPT because the resonant frequency is defined by the pitch of patterned electrodes and only one metal layer is required. In lithium niobate, the minimum feature size to reach the 2.4 GHz industrial, scientific, and medical radio band is 420 nm [7,8], which
∗ Corresponding author at: 403 Cory Hall #1774 University of California, Berkeley, CA 94720-1774, USA. E-mail addresses: [email protected] (D.A. Rolfe), [email protected] (K.L. Dorsey), [email protected] (J.C. Cheng), [email protected] (A.P. Pisano). http://dx.doi.org/10.1016/j.sna.2016.07.001 0924-4247/© 2016 Published by Elsevier B.V.
is within the theoretical capabilities of AMPT fabrication. Four SAW resonators with electrode width of 500 nm, 600 nm, 800 nm, and 1 m are fabricated and the resonant behavior is characterized.
2. Design and fabrication A critical part of the AMPT process is the vapor-permeable template: a polymer film that is patterned with the desired device features, into which the nanoparticle ink is drawn to form the metallization [3]. The template is patterned from a master fabricated through photolithography. The master contains raised features in a layout identical to that of the desired nanoparticle pattern. In this work, the resonator electrode features (width of 1 m or less) are patterned lithographically (Fig. 1a) onto a thermally-grown SiO2 layer and etched using deep reactive ion etching (DRIE) (Fig. 1b). The wider fill channels that lead to the electrodes are lithographically patterned on an SU-8 layer that is spun onto the smaller SiO2 features. The template is composed of a poly(4-methyl-2-pentyne) (PMP) film. PMP is highly rigid (1.6 GPa [6]) and extremely permeable to organic solvents, with a permeability for ethanol of 185,000 Barrer. The combination of rigidity and permeability insures that the nanoparticle ink solvent can rapidly evaporate through the film while the patterned channels retain their shape during deposition and formation of the nanoparticle features.
74
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
contained arrays of grounded electrodes to reflect the SAW wave at the Bragg condition interval, constraining the wave within the IDT area. The grid features that connected to the IDTs and fill ports served two purposes. The large surface area served as the contact pad to test the resonator operation using a ground-source-ground (GSG) probe. These contact pads were designed as grids with the same feature width as the IDTs, rather than open areas, because a grid of channels is less prone to collapse under the induced pressure of evaporating solvent. To test the ability of the AMPT process to fabricate SAW resonators with various center frequencies, a template with four resonators was fabricated. The SAW resonators were patterned on a Y-Z cut lithium niobate (LiNbO3 ) substrate, which was chosen for its high speed of sound and coupling efficiency. The model for the center frequency of a SAW was used to determine the appropriate electrode pitch [9], fc =
v 2(w + d)
(1)
where v is the longitudinal velocity of wave propagation in the substrate, and w and d are electrode width and spacing, respectively. The target frequencies of the resonators were 860 MHz, 1.07 GHz, 1.43 GHz, and 1.72 GHz, corresponding to electrode width and spacing of 1 m, 800 nm, 600 nm, and 500 nm, respectively. Fig. 1. A schematic of the AMPT process showing (a) lithography, (b) etching of the master wafer, (c) the template on the substrate, (d) introduction of nanoparticle ink, (e) drying and removal of the template.
The template is fabricated by casting PMP dissolved in cyclohexane onto the master. The cyclohexane is evaporated slowly through a PDMS membrane over the course of three days. The slow evaporation into a high concentration cyclohexane environment ensures proper feature formation and minimal residual stress. When fully dry, the template is removed from the master by softening it with isopropanol and peeling gently. To pattern the desired features using the AMPT process, the template is placed channel-side-down on a substrate flooded with solvent (Fig. 1c). Nanoparticle ink is introduced to fill ports which are connected to features in the center of the template by means of fill channels (Fig. 1d). As solvent in the channels evaporates through the template, nanoparticle ink is drawn into the channels. The nanoparticles become increasingly concentrated and begin to densely pack as the ink solvent evaporates. After the features are fully packed and the excess solvent evaporates, the template is removed (Fig. 1e). The conductive features are patterned with DGP AP ink from Advanced Nano Products Co. Ltd., (Buyong-myeon, South Korea). This ink contains nanoparticles averaging 20 nm at about 40 wt% in ethanol. The ink was diluted to about 8 wt% during the patterning process. Features were patterned by heating the substrate to 45 ◦ C and lowering the template onto ethanol. Silver ink and ethanol were added to the fill ports over 30 min to evaporate all solvent. After solvent evaporation, the device and template were sintered in nitrogen at 250 ◦ C for 30 min. Finally, the template was removed by soaking the patterned device in acetone to swell the template and aid in delamination from the substrate. The resonator layout is shown in Fig. 2. The widths w and d are equal to each other and range from 500 nm to 1 m. The lengths la and ld range from 60 m and 5 m to 120 m and 10 m, respectively. The ratio of la to ld ’s length remains constant across all fabricated resonators. Each resonators has 40 pairs of interdigitated electrodes. A surface acoustic wave was only excited along the length marked a in Fig. 2. The remaining electrode length from the interdigitated transducer (IDT) to the fill ports (d in Fig. 2) was designed to dampen higher order modes. Gratings on either side of the IDT
3. Results and discussion The 1 m and 600 nm electrode width resonators are shown in Fig. 3. In both resonators, the fill ports are well defined, but the interdigitated electrodes are not completely patterned across the resonator. The circled areas indicate regions where the patterning is not complete. Lighter areas indicate places where silver nanoparticles have completely filled the template, whereas darker areas show uneven or incomplete patterning, or the existence of increased solvent residue. Both incomplete patterning and solvent residue increase the resistivity of the resonator features, reducing the quality factor of the resonator. These patterning errors are caused by limitation of the AMPT process. One such limitation is the use of polymers as the template materials. During patterning, the PMP and polydimethysiloxane (PDMS) backing material absorb the solvent in the nanoparticle ink and swell. Mismatch in PMP and PDMS solvent diffusion rate and solvent absorption distorts the template [10,11]. Distortion may also occur within the template. The solvent diffusion rates vary over the template due to feature size and the presence of corners [12]. The PMP along the channel walls swells more than the PMP far from the channel walls. The localized swelling causes the PMP closest to the ink channels to expand more than the rest of the template. The concentration gradient of solvent can also lead to buckling, in which solvent in the fill channels causes the bottom of the template to disproportionately swell and fold in on itself. Distortion is a concern for patterning fidelity and device functionality. The effects of swelling are visible in the pattern in Fig. 3a. The circled areas show where the swollen template buckled and delaminated from the template during patterning. As a result, the Bragg gratings are not completely formed. The circled area in Fig. 3b also shows incomplete nanoparticle ink patterning. In comparison to the IDT area in Fig. 3a, the interdigitated fingers are not clearly outlined. The real and imaginary components of each resonator’s reflection signal were measured using an Agilent Technologies E5071B network analyzer with a load termination of 50 . Each device showed a resonance between 1.1 and 1.5 GHz (Fig. 4). As predicted by the center frequency model in (1), the measured center frequency is inversely related to electrode width. A mismatch of 5%
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
75
Fig. 2. A one-port SAW resonator showing center IDT (red with green fill ports) and Bragg gratings (black with blue fill ports). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
(800 nm width) to 30% (500 nm width) exists between modeled and measured center frequency. The mismatch severity shows no dependence on resonator electrode width. In smaller channels, localized swelling widens the electrode spacing, lowering the center frequency of the smallest two resonators in Fig. 4. Conversely, the swelling of features surrounding the electrodes may have been responsible for the higher than expected center frequency in the larger resonators. The measured quality factors of the AMPT-patterned resonators range from 5.9 (500 nm) to 13.4 (1 m), which are several orders of magnitude lower than traditionally-fabricated SAW resonators [13]. The low quality factor and frequency model mismatch may be attributed to poor ink patterning and misalignment between the template and the substrate.
A current limitation of the AMPT process is patterning fidelity, especially at smaller device features. If the template distorts from solvent-induced swelling, nanoparticle ink may leak under the template and form a conductive path between the IDTs. As the electrode spacings decrease from 1 m to 500 nm, a conductive path is more likely to form between IDTs. The conductive path’s effect on Q and fc is modeled using a series RLC lumped model to represent the resonator with a parasitic resistance RP in parallel with the resonator. The R, L, and C values and the parasitic resistance RP were fit to measured data in the 1 m resonator. The real component of the RP model is shown for the 1 m resistor as a short dotted line (Fig. 4a). The fitted value of RP is approximately 50 for all patterned resonators.
Fig. 3. Resonator micrographs with (a) 1 m wide electrodes and (b) 600 nm wide electrodes.
76
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77
Fig. 4. The real and imaginary components of the reflected signal for each resonator.
Two other process limitations have implications for AMPT’s use in SAW resonator fabrication. AMPT does not currently have a mechanism to align the patterning template to the substrate. The angle between the template and the piezoelectric substrate was not considered or recorded during device fabrication. This angle is critical to a high coupling coefficient: misalignment of only 10◦ results in a coupling coefficient decrease of 50% [14]. Misalignment of the patterned IDTs to the lithium niobate substrate could therefore be another source of performance loss. A fundamental limitation of the AMPT process is its use of nanoparticle ink for device metallization. The resistivity of the sintered nanoparticle IDTs was 32 nm, which is approximately twice the value of bulk silver. It is not likely that the higher ink resistivity degraded the device Q, but the resistivity could limit performance as the patterning fidelity and ink leakage issue are improved. 4. Conclusions A series of resonators with IDTs ranging from 500 nm to 1 m were fabricated from silver nanoparticles on a lithium niobate substrate using the AMPT process. Despite the low quality factor and deviation from modeled resonant frequency, these resonators demonstrate the ability of the AMPT process to pattern functional devices with less complexity than lift-off processes. From a single master fabricated with lithography or other processes, the templates for many devices may be cast. Once a template is cast, it can be used to pattern device features in less than 30 min without the need for photolithography, vacuum-processed metal deposition, or a clean room. Increased resonator performance will be achievable by primarily by limiting the parasitic resistance between the IDTs and by aligning the SAW device to improve the coupling coefficient. To reduce ink leakage between electrically isolated features, we propose adding a rigid structure to the PDMS template backing material to prevent it from swelling or deforming, and by increasing the pressure applied to the PMP template while patterning. In
future process iterations, we shall also use a thinner PMP template to reduce the effects of swelling on the deposited device pattern. To improve the electromechanical coupling and increase Q, we will incorporate an alignment mechanism into future AMPT process iterations. With these changes, the AMPT process could significantly reduce the complexity of fabricating SAW resonators for consumer electronics. Acknowledgments The authors would like to thank Fuji Electric Co., Ltd. for their financial support of this research. Fabrication was performed in the Biomolecular Nanotechnology Center at UC Berkeley and the Nano3 facility at UC San Diego. References [1] E. Forsén, D. Nilsson, W. Daniau, B. Franc¸ois, S. Ballandras, Fabrication of SAW resonators on quartz and langasite using nano-imprintlithography, in: Proceedings – IEEE Ultrasonics Symposium, 2010, pp. 2091–2094, http://dx. doi.org/10.1109/ULTSYM.2010.5936002. [2] A. Kamyshny, J. Steinke, S. Magdassi, Metal-based inkjet inks for printed electronics, Open Appl. Phys. J. 4 (2011) 19–36. [3] M.T. Demko, J.C. Cheng, A.P. Pisano, High-resolution direct patterning of gold nanoparticles by the microfluidic molding process, Langmuir 26 (22) (2010) 16710–16714, http://dx.doi.org/10.1021/la1022533 http://www.ncbi.nlm. nih.gov/pubmed/20886896. [4] M.T. Demko, S. Choi, T.I. Zohdi, A.P. Pisano, High resolution patterning of nanoparticles by evaporative self-assembly enabled by in situ creation and mechanical lift-off of a polymer template, Appl. Phys. Lett. 99 (2011) (2011) 1–4, http://dx.doi.org/10.1063/1.3671084. [5] M.T. Demko, T.P. Brackbill, A.P. Pisano, Simultaneous patterning of nanoparticles and polymers using an evaporation driven flow in a vapor permeable template, ACS J. Surf. Colloids 28 (25) (2012) 9857–9863, http://dx. doi.org/10.1021/la301587f http://www.ncbi.nlm.nih.gov/pubmed/22647075. [6] M.T. Demko, J.C. Cheng, A.P. Pisano, Rigid, vapor-permeable poly(4-methyl-2-pentyne) templates for high resolution patterning of nanoparticles and polymers, ACS Nano 6 (8) (2012) 6890–6896. [7] R. Weis, T. Gaylord, Lithium niobate: summary of physical properties and crystal structure R, Appl. Phys. A: Mater. Sci. Process. 37 (1985) 191–203, http://dx.doi.org/10.1007/BF00614817.
D.A. Rolfe et al. / Sensors and Actuators A 248 (2016) 73–77 [8] J. Bell, R. Li, Surface-acoustic-wave resonators, Proc. IEEE 64 (5) (1976), http:// dx.doi.org/10.1109/PROC.1976.10200. [9] S. Achenbach, D. Klymyshyn, T. Mappes, A. Kachayev, V. Subramanian, G. Wells, J. Mohr, Submicron-scale surface acoustic wave resonators fabricated by high aspect ratio X-ray lithography and aluminum lift-off, Microsyst. Technol. 14 (9–11) (2008) 1715–1719, http://dx.doi.org/10.1007/s00542-0070498-7. [10] T. Honda, M. Miyazaki, H. Nakamura, H. Maeda, Controllable polymerization of N-carboxy anhydrides in a microreaction system, Lab Chip 5 (2005) 812–818, http://dx.doi.org/10.1039/b505137a. [11] A. Morisato, I. Pinnau, Synthesis and gas permeation properties of poly(4-methyl-2-pentyne), J. Membr. Sci. 121 (1996) 243–250. [12] D.A. Rolfe, K.L. Dorsey, J.C. Cheng, A.P. Pisano, A model to guide template-based nanoparticle printing development, in: ASME 2015 International Technical Conference and Exhibition on Packaging and Integration of Electronic and Photonic Microsystems collocated with the ASME 2015 13th International Conference on Nanochannels, Microchannels, and Minichannels, American Society of Mechanical Engineers, 2015, V003T03A006. [13] U. Knauer, J. Machui, C.C.W. Ruplpel, S. Member, Design, fabrication, and application of GHz SAW devices, in: IEEE Microwave Theory and Techniques Society, IEEE, Denver, CO, 1997, pp. 1821–1824. [14] K. Yamanouchi, K. Shibayama, Propagation and amplification of Rayleigh waves and piezoelectric leaky surface waves in LiNbO3 , J. Appl. Phys. 43 (3) (1972) 856–862, http://dx.doi.org/10.1063/1.1661294.
Biographies
David Rolfe is a researcher in the Berkeley Micromechanical Analysis and Devices (BMAD) Lab at the University of California, Berkeley (now the PRIME lab at University of California, San Diego), as a student of Albert P. Pisano. His research interests include nanoparticle patterning, microfluidics, laser transfer processes, rapid prototyping, nano-engineering ethics, and flame ionization sensors. He received his M.S. and Ph.D. from the department of Mechanical Engineering at University of California, Berkeley, and his B.S. from Harvey Mudd College in Claremont, CA.
77
Kristen Dorsey is an assistant professor of engineering in the Picker Engineering Program at Smith College. Her research interests include the modeling of reliability and stability issues in MEMS and sensor/device fabrication using printed and soft lithography processes. Dorsey completed her postdoctoral work at UC San Diego and UC Berkeley in Albert P. Pisano’s group and received her Ph.D. in Electrical and Computer Engineering at Carnegie Mellon University in Gary Fedder’s group. Her dissertation title was “Dielectric Charging in CMOS MEMS.” Her B.S. in Electrical and Computer Engineering is from Franklin W. Olin College of Engineering in Needham, MA. Jim Cheng is senior member of the technical staff at Maxim Integrated. He has been leading efforts focused primarily on nanoprinting platforms, but also encompassing energy storage solutions, biosensors, and passive cooling for electronics. Dr. Cheng has been involved in several start-ups and most recently founded ChillFlux, Inc., to commercialize his research into novel thermal management. Dr. Cheng received his B.Eng. in Electrical Engineering from McMaster University, Canada in 2004. He was formerly a research scientist at the University of California, Berkeley, where he had previously received his M.S. and Ph.D. degrees in Electrical Engineering in 2006 and 2009, respectively. Albert P. Pisano is the dean of the Jacobs School of Engineering and holds the Walter J. Zable Chair in Engineering at University of California, San Diego. In 2001, Pisano was elected to the National Academy of Engineering. Pisano earned his undergraduate (’76) and graduate degrees (’77, ’80, ’81) in mechanical engineering at Columbia University. Prior to 2013, Pisano served on the UC Berkeley faculty, and previously was the MEMS program manager for DARPA. He has held research positions with Xerox Palo Alto Research Center, Singer Sewing Machines Corporate R&D Center and General Motors Research Labs.