Fabrication of a Nanoscale Electric Field Sensor Yun Zheng,a* Todd King,b Daniel Stewart,b Stephanie Gettyb a Detectors Systems Branch, NASA Goddard Space Flight Center, Greenbelt, MD 20771 b Materials Engineering Branch, NASA Goddard Space Flight Center, Greenbelt, MD 20771 ABSTRACT A new nanoscale electric field sensor was developed for studying triboelectric charging in terrestrial and Martian dust devils. The sensor was fabricated using MEMS techniques, integrated at the system level, and deployed during a dust devil field campaign. The two-terminal piezoresistive sensor consists of a micron-scale network of suspended singlewalled carbon nanotubes (SWCNTs) that are mechanically coupled to a free-standing electrically conductor. Electrostatic coupling of the conductor to the electric field is expected to produce a deflection of the conductor and a corresponding change in nanotube device resistance, based on the known piezoresistive properties of SWCNTs. The projected device performance will allow measurement of the large electric fields for large dust devils without saturation. With dimensions on the 100 μm scale and power consumption of only tens of nW, the sensor features dramatically reduced mass, power, and footprint. Recent field testing of the sensor demonstrated the robustness of suspended SWCNT devices to temperature fluctuations, mechanical shock, dust, and other environmental factors. Keywords: Nanotechnology, MEMS, Single-walled carbon nanotube, Piezoresistance, Electric field sensor *[email protected]; phone 1 301-286-3606

1. INTRODUCTION Terrestrial dust devils are organized dry cyclones originating from temperature inversion, such as a hot surface and cooler air column, that lift dust grains into circulation. They have been recognized to support substantial electric fields (E-Fields) due to grain-to-grain contact electrification [1-4]. Likewise, Mars is known to have a dynamic atmosphere with dust devils and global dust storms, and understanding the magnitude of electric fields associated with these dust devils can provide substantial insight into the interplay of atmospheric chemistry and weather near the Martian surface. Figure 1 shows a dust devil observed by NASA’s Sprit Rover on Mars. Comparison to their terrestrial cousins indicates the presence of large tribocharged electric fields within their structure.

Figure 1. Dust devils, like this one observed by the Spirit rover in early 2007, likely produce locally large electric fields that may have implications for atmospheric and soil chemistry on Mars.

Micro- and Nanotechnology Sensors, Systems, and Applications, edited by Thomas George, M. Saif Islam, Achyut K. Dutta, Proc. of SPIE Vol. 7318, 731815 · © 2009 SPIE · CCC code: 0277-786X/09/$18 · doi: 10.1117/12.818958

Proc. of SPIE Vol. 7318 731815-1

The electric fields in terrestrial dust devils can be of order 104 to 105V/m, large enough to saturate commonly used electric field mills. Historically, measurements of dust devil electric field have been made either at a significant distance from the dust devil center [1] or by instruments that saturated prior to direct interception with the dust devil [2], but recent work by Farrell,et al., [3] developed an E-field measurement device which can measure large electric field without saturation. The objective of the work presented here was to fabricate a complementary E-field sensor with a miniaturized footprint and high saturation field. In landed planetary missions, mass is a critical consideration. The capacity to make environmental measurements using miniaturized tools will allow high scientific return within a limited payload envelope. This is the driver, therefore, for a compact, low mass, low power sensor that can detect the large electric fields expected for dust devils on Mars and throughout the Solar System. To minimize a device or a sensor, nanotechnology and microelectromechanical systems (MEMS) technology present promising routes to novel miniaturized devices. Single-walled carbon nanotubes, in particular, exhibit an unusually high degree of piezoresistance, or change in electronic properties with mechanical strain. This has been demonstrated by other groups in previous work, such as tensile stretching applied with an atomic force microscope [5]. Small band-gap semiconducting (or quasimetallic) nanotubes exhibit the largest resistance changes and piezoresistive gauge factors under axial strains [6]. Previous workers [5-9] have undertaken painstaking experiments to isolate a single nanotube for study. Here, we use an ensemble of SWCNTs, with some anticipated detriment to integrated piezoresistive response, to address manufacturability and robustness. (a)

(b)

(c)

'H

E-field

Figure 2. Notional representation of nano e-field sensor measurement of electric field. (a) The device is first voltage-biased to measure two-terminal conductance. (b) An applied field couples to the device through polarization of the gold deflector. (c) The biased substrate pulls on the needle, imparting strain to the SWCNTs, which causes a change in device conductance.

We have developed a prototype electric field sensor, with an operational schematic shown in Figure 2. The sensor consists of a gold needle supported by a network of SWCNTs that are suspended across a trench in a silicon substrate. An electric field will generate an electric dipole in the conductor, which is mechanically coupled to the SWCNTs. With the silicon substrate below the sensor held at a known voltage bias, an electrostatic deflection of the conductor occurs, leading to a deflection of the SWCNT network. Because the SWCNTs are piezoresistive, their electrical conductance is expected to be reduced by the application of strain. Deflection of the conductor with application of an electric field can

Proc. of SPIE Vol. 7318 731815-2

therefore be transduced into an electronic signature using this microscale sensor. The compact size and low operating power requirements may make this device attractive for distributed terrestrial sensing, as well as spaceflight applications.

2. FABRICATION The fabrication protocol of a SWCNT nano E-field sensor is shown in Figure 3. There are three major steps to the process. First, single-wall carbon nanotubes are grown on a substrate. Next, lithographic patterning is used to form a gold needle and electrodes. Finally, wet etching is used to form a trench underneath of the gold needle. E-field sensor fabrication was carried out at NASA Goddard Space Flight Center using carbon nanotube growth facilities in the Materials Engineering Branch and device fabrication tools in the Detector Development Laboratory, which is a class 10 clean room with MEMS fabrication facilities. Feedstock Flow

(a5N

950° C

Si02

Silicon

Au

SWCNTs

Au

Trench

Figure 3. The fabrication protocol (a) SWCNTs are grown using catalyst-assisted chemical vapor deposition. (b) Electrodes and a conductive deflector are formed using lithography and thin film deposition in the DDL. (c) A trench is formed to free the structure using wet chemical etching

2pm

Mag = 3.34 K X

H-H

WD =

8 nmnm

EHT = 2.00 kV Time :9:55:37

Signal A = InLens Photo No. = 8494

MEB 541

Figure 4. SEM picture of successful single-walled carbon nanotube growth shows a dense, interconnected network.

In detail, 400 nm of thermal oxide was grown onto a standard four inch silicon wafer. The wafer was then diced into 10 x 10 mm chips. Submonolayer iron catalyst was deposited onto the surface of the chip by indirect resistive evaporation in vacuum, as has been described elsewhere [10]. The chip was then put into a chemical vapor deposition (CVD) system

Proc. of SPIE Vol. 7318 731815-3

for single-walled carbon nanotube growth at a temperature of 950°C with a co-flow of CH4, C2H4, H2 and argon gas, as depicted in Figure 3(a). After growth, the chip was studied using a scanning electron microscope (SEM; Leo Supra 50VP) to check the quality of nanotube growth. Figure 4 shows a representative SEM micrograph of single-walled carbon nanotube growth on the silicon oxide surface. The length of the SWCNTs can be considerably more than 10 microns long, and they are randomly distributed to form a dense single-walled carbon nanotube net. The nanotube chip then was spin coated with MMA and PMMA electron-beam bilayer resist and baked on a hotplate at 180°C for 5 minutes and 160°C for 10 minutes, respectively. Using a SEM retrofitted with the Nanometer Pattern Generation System, the chip was patterned with the deflector and two electrodes, shown in Figure 3(b). After developing, a 10 nm Cr/100 nm Au bilayer was electron-beam evaporated onto the chip. Liftoff was conducted in heated acetone to remove the unwanted metal, resulting in four patterns per chip. Each pattern consists of a 2 μm x 80 μm deflector and two opposing Au electrodes that connect to large bonding pads. To form a trench in the substrate, the chip was spin coated with Shipley PR-1811 photoresist and baked on a hot plate at 110°C for two minutes. The trench area was defined to be an oval centered on the metal deflector. Importantly, the ends of the deflector were not exposed to prevent loss of the suspended metal during etching. After exposure and development, the chip was put into a 7:1 buffered HF solution to remove silicon oxide from beneath the SWCNT-deflector region. The exposed silicon was then wet etched in a 20% KOH solution to complete the trench. The trench depth between the gold needle and substrate was controlled to be about 17 micron. After gentle rinsing in deionized water, a methanol freeze dry process was used to minimize surface tension during drying. At this point, the deflector is supported at its ends by contact to the substrate. As a final step, the ends of the deflector are severed using localized heating of a laser dicing tool. Precise micron-scale milling is achieved, and the successful device is shown in Figure 5. The metal deflector is successfully suspended between two Au electrodes on a nanotube network above a trench. The electrodes serve to clamp the ends of the SWCNTs to the underlying substrate.

Electrode

Deflector Figure 5. SEM picture of a completed nano e-field sensor. The opposing electrodes serve to clamp the SWCNTs to the substrate. The central device, demarcated by the oval trench outline, is suspended from the substrate. The deflector is supported entirely by the SWCNT network.

Proc. of SPIE Vol. 7318 731815-4

3. TEST AND MEASUREMENT 3.1 Laboratory Testing We have constructed an electric field test chamber for calibration purposes and laboratory testing. The fabricated nano E-field chip was mounted onto a ceramic substrate. The electrodes were then wire bonded to connect to leads on the ceramic substrate. The ceramic mount was then positioned between two parallel copper plates inside of a vacuum chamber. High-voltage feedthroughs were connected to a power supply (SRS PS350). The distance between the two plates is inversely proportional to the achievable E-field, therefore the spacing is kept small in order to generate high field. For calibration, the nano E-field sensor could be placed in the center of the two copper plates, as shown in Figure 6. An electric field of 500kV/m can be reached in this test set up. Results of laboratory testing will be presented elsewhere.

Figure 6. Laboratory test fixture uses two parallel copper plates to apply high E-field in the region between. A nano E-field sensor is (left) shown mounted onto a ceramic holder and (right) placed in the center of the plates for calibration testing.

3.2 Packaging Packaging of the nano E-field sensor was designed to achieve low screening of the external electric field. At the same time, the packaging was required to protect the sensor during field studies of multiple dust devils over a five-day investigation. Wind, dust, vibration, and temperature fluctuations were expected and planned for in this packaging approach. The result was a low-profile aluminum sheet metal box, with a plastic window to expose the nano E-field sensor and a control electrometer. The seams of the box were sealed with Kapton tape. The control sensor was included to provide a correlated signal using simple charge induction on a blank silicon substrate of the same dimensions as the nano E-field sensor. Electrical leads were routed through BNC bulkhead connections around the box perimeter. Coaxial cables were used to interface the device leads to an external battery-powered voltage supply and amplification electronics boxes. The entire compact system is shown in Figure 7.

Proc. of SPIE Vol. 7318 731815-5

Figure 7. Intermediate stage of sensor packaging shows the nano E-field sensor (top chip) and control electrometer (bottom chip) mounted inside a low-profile aluminum case. BNC bulkhead connectors allow coaxial cabling to the batterypowered electronics box shown above.

Figure 8. Left: Nano e-field sensor box (left) and a control electrometer were mounted on top of the chase vehicle. Right: SUV with nano E-field sensor intercepts a dust devil in Nevada desert.

3.3 Field testing Field testing was carried out in a dry lake bed near Henderson, Nevada, where high dust devil activity is known to occur during the summer months. The packaged devices were mounted on top of a chase vehicle and connected to a computer with digital acquisition capability inside of the vehicle to record the test data. Figure 8 shows the field test set up (left) and a dust devil intercept event (right). The vehicle intercepted as many as ten dust devils in four days. The presence of the vehicle tended to compromise the organization of the dust devils, so only a few cyclones were sufficiently well organized to support multiple passes. Figure 9 shows the signal for the nano E-field sensor and control electrometer as a function of time for multiple passes of one such storm.

Proc. of SPIE Vol. 7318 731815-6

E4-control

NanoSensor

7.35

3 2

7.3

1 0

7.25

-1 7.2

-2 -3 -4 350

400

450

Nano E-field Sensor (kΩ)

Electrometer Signal (V)

4

7.15 500

Time (s) Figure 9. Field testing data is shown for the Nano E-field Sensor (red) and control electrometer (blue). Deviations in nano E-field sensor signal can be seen at time 370 s, 410 s, 445 s, and 480 s, showing some correlation with peaks in the control signal.

4. RESULTS As seen in Figure 9, deviations in the Nano E-field Sensor device resistance were found to be correlated with peaks in the electrometer signal. Based on the control electrometer calibration, the E-field is estimated to be of order several tens of kV/m, but more careful calibration is required in future work to ascertain precise quantitative results. During transportation of the nano E-field sensor to the test site in Henderson, NV, and during testing, the suspended SWCNT device experienced a number of harsh conditions that are worthy of note: broad-band vibrations during aircraft take-off, flight, and landing, vibration and shock during dust devil chases, temperature fluctuations from nighttime to daytime in desert conditions, up to 60 mph winds, and even a brief rain squall. Surviving some of these conditions (rain and wind) indicates the quality of the packaging approach. That the device successfully operated after the various vibrational conditions and instances of mechanical shock is recognized as evidence of the robustness of MEMS systems to vibration [11]. We can consider this further by evaluating the spectral distribution of vibrations for aircraft and automobiles. According to [12] and references therein, the power spectrum for an aircraft at cruise speed exhibits the largest amplitude at low frequencies and decays rapidly within a few tens of Hz. The spectrum for an automobile, on the other hand, extends to higher frequencies but also decays with increasing vibrational frequencies within a few hundred Hz [13]. MEMS and NEMS (nanoelectromechanical systems) devices, in contrast, have resonance frequencies that can be well above 1 MHz [14], and damage to these devices is unlikely.

5. CONCLUSIONS We have demonstrated fabrication, packaging, and field testing of a nanoscale electric field sensor. The Nano E-field Sensor features reduced mass, footprint, and power consumption (around 50 nW), relative to competing technology, for the study of environmental systems with large electric field, like terrestrial thunderstorms, and terrestrial and Martian dust storms. Results from field testing suggest that the nano E-field sensor responds to large electric fields, but more careful calibration in the laboratory is needed to validate and quantify these data. Notably, we have successfully conducted field testing of the MEMS device to find that the sensor is robust to vibration, mechanical shock, and temperature fluctuations.

Proc. of SPIE Vol. 7318 731815-7

ACKNOWLEDGEMENTS This work was supported by the NASA Goddard Space Flight Center Internal Research and Development Program and the GSFC Summer Internship Program. The authors are grateful to W. Farrell, T. Jackson, and H. Aguilar for their support during the preparation and execution of the field testing experiment. The authors would further like to acknowledge the contributions of Dewey Dove, Bruno Munoz, Charles He (NASA GSFC), and summer internship participants, Jonathon Brame and Nathan Woods from Brigham Young University.

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

[12] [13] [14]

Freier, G.D., “The electric field of a large dust devil,” J. Geophys. Res. 5, 3504, (1960). Farrell, W.M., et al., “Electric and magnetic signatures of dust devils from the 2000-2001 MATADOR desert tests,” J. Geophys. Res. 109, E03004 (2004). Jackson T.L., Farrell, W., “Electrostatic fields in dust devils: an analog to mars,” IEEE transaction on Geoscience and remote sensing 44, 2942 (2006). Renno, N. O. and Kok, J. F., “Electrical Activity and Dust Lifting on Earth, Mars, and Beyond,” Space Science Reviews 137, 419 (2008). Paulson, M.R. et al., “In situ resistance measurements of strained carbon nanotubes,” Applied Physics Letters 75, 2936 (1999). Cao, J., et al., “Electromechanical properties of metallic, quasimetallic, and semiconducting carbon nanotubes under stretching,” Physical review letters 90, 157601 (2003). T. Tombler, C Zhou, L Alexseyev, J Kong, H Dai, Nature 405, 769 (2000). A. Maiti, A. Svizhenko, and M. P. Anantram, “Electronic Transport through Carbon Nanotubes: Effects of Structural Deformation and Tube Chirality,” Phys. Rev. Lett. 88, 126805 (2002). C. Stampfer, A Jungen, C Hierold - Sensors, “Fabrication of Single-Walled Carbon-Nanotube-Based Pressure Sensors,” Nano Letters 6, 233 (2006). Brame, J., Getty, S., Goodsell, J., and Allred, D. D., “Strain-based Electrical Properties of Systems of Carbon Nanotubes Embedded in Parylene,” Materials Research Society Symposium Proceedings 963, Q20-18 (2007). Tanner, D. M., Walraven, J. A., Helgesen, K. S., Irwin, L. W., Gregory, D. L, Stake, J. R., and Smith, N. F., “MEMS reliability in a vibration environment,” IEEE 38th Annual International Reliability Physics Symposium 139 (2000). Stephens, D. G. “Developments in ride quality criteria,” in Human Factors in Engineering and Design Sanders, M. S. and McCormick, E. J., Eds.; New York, McGraw-Hill (1992). Kim, S. H., Lee, J. M., and Sung, M. H., “Structural-Acoustic Modal Coupling Analysis and Application to Noise Reduction in a Vehicle Passenger Compartment,” Journal of Sound and Vibration 225, 989 (1999). Craighead, H. G., “Nanoelectromechanical Systems,” Science 290, 1532 (2000).

Proc. of SPIE Vol. 7318 731815-8