MULTIPLE-INPUT MICROCANTILEVER SENSOR WITH CAPACITIVE

Presented at the 1999 Government Microcircuits Applications Conference (GOMAC) March, 1999 MULTIPLE-INPUT MICROCANTILEVER SENSOR WITH CAPACITIVE READOUT C. L. Britton, Jr., R. J. Warmack, S. F. Smith, P. I. Oden, R. L. Jones, T. Thundat, G. M. Brown, W. L. Bryan, J. C. DePriest, M. N. Ericson, M. S. Emery, M. R. Moore, G. W. Turner, A. L. Wintenberg, T. D. Threatt, Z. Hu, L. G. Clonts, J. M. Rochelle Oak Ridge National Laboratory, Oak Ridge, TN 37831 and The University of Tennessee, Knoxville, TN 37996

ABSTRACT A surface-micromachined MEMS process has been used to demonstrate multiple-input chemical sensing using selectively coated cantilever arrays. Combined hydrogen and mercury-vapor detection was achieved with a palm-sized, selfpowered module with spread-spectrum telemetry reporting. INTRODUCTION Microcantilevers, such as those used by atomic force microscopes, have been demonstrated as a universal platform for real-time, in-situ measurement of physical and biochemical properties. Sensitive detection of the cantilever deflection due to adsorption-induced forces and resonance frequency variation due to mass loading can result in part-per-billion (ppb) to part-per-trillion (ppt) sensitivity. A spectrum of physical, chemical, and biological sensors based upon the mi1-5 crocantilever platform has been tested . The typical dimensions of commercially available, micromachined, mass-produced microcantilevers are 50-200 µm long, 10-40 µm wide and 0.3-3 µm thick with mass in the range of a few nanograms. Cantilever motion is typically accomplished for AFM applications by laser-beam deflection or piezoelectric transduction. For a low-power, multiple-input device, an array of capacitively read cantilevers, each fabricated with a selected coating, appears to be ideal. ARRAYED SENSOR SYSTEM An initial demonstration of the concept of a universal platform is a one-dimensional, ten-element microcantilever array (MUMPS process of MCNC) selectively coated with gold for mercury 6 sensing and palladium for hydrogen sensing . The cantilevers, whose cross section is shown in Fig. 1, utilized the stress effect from exposing the coatings to a vapor of interest and measuring the resultant deflection as the vapor and coatings

Polysilicon beam

Beam anchor

Coating

Substrate

Bottom poly plate

Fig. 1. MUMPS cantilever cross-section

Cantilever under test

Sample and hold

Square-wave input To analog-digital converter Figure 2. Block diagram of the capacitive readout

interact. As the cantilever bends, the spacing between the beam itself and a polysilicon plate underneath the beam changes. This change is sensed as a change in capacitance. A simplified version of the circuit used is shown in Fig. 2. The readout was fabricated as an 8-channel chip in a 1.2-µm bulk CMOS process. The cantilever chip, it’s associated readout chip and interface support circuitry was all mounted as a single boardset onto a battery pack that utilized four AA batteries. The ‘palmtop’ package is shown in Fig. 3.

Presented at the 1999 Government Microcircuits Applications Conference (GOMAC) March, 1999

80 70

#3

60 #1

50 40 30 20

#7

10

#5

Jerome 2139

1872

1605

1337

1070

802

535

267

0.02

0

3.50 3.40 3.30 3.20 3.10 3.00 2.90 2.80 2.70 2.60

Readout Output (V)

This system for the detection of mercury does not self-regenerate, meaning that the mercury absorbed into the gold is not released to the atmosphere once the mercury vapor is removed and thus acts as a dosimeter.

Jerome Output (ppb)

MEASUREMENTS Measurements were performed using the palmtop system. Ultimately, we plan to coat each element with a different chemical coating, each selectively responding to components of a complex mixture. Key enabling factors of large arrays also include both redundancy and chemical specificity provided by an ensemble response of the array. Thus, selectivity need not be limited to that which can be accomplished by individual sensors.

Time (s)

Fig. 4. Multichannel mercury data

Noise in the output that was common to all the channels is also evident in Figure 4. Figure 5 shows that when an uncoated channel is subtracted from a channel that was coated this noise can be removed. From Figure 5 the mercury concentration at which the cantilever begins to respond is approximately 10ppb.

0.30

50 40

0.20

30

0.10

20 0.00

10 2030

1776

1523

1269

1016

762

-0.10 508

0

Time (s)

Fig. 5. Common-mode correction to mercury data

Hydrogen Sensing Results The results from a hydrogen-sensing test are shown in Figure 6. On this test chip cantilevers 1, 3, and 4 were coated with palladium. The coated cantilevers reacted to the introduction of hydrogen to the system by moving towards the bottom polysilicon plate, and there was no response to the hydrogen by the uncoated cantilevers. The graph also shows that the cantilevers returned to approximately the same levels as at the beginning of the test. The downward slope at the beginning of the test is due to regeneration from the previous test that was run. The cantilever was only allowed to stay at its baseline value for a short period of time because the location of the baseline was known and not required to be

Relative Output Change (V)

0.40

60

255

Mercury Sensing Results The results from a set of mercury tests are shown in Figure 4. Our reference was a Jerome sampling mercury detector. On this test chip, cantilevers 9 and 10 were unused. Cantilever numbers 1, 3, 5, and 7 were gold coated. All four of the coated cantilevers had a response to the mercury, as shown. The cantilevers moved toward the bottom plate resulting in an increased capacitance (away from the coating) thus the output voltage of the cantilevers readout channels moved in a positive direction. The change in cantilever capacitance ranged from 237fF on cantilever #1 down to 48fF on cantilever #3.

0.50

70

1.41

Fig. 3. Cantilever palmtop sensor

Jerome Reference Mercury Concentration (ppb)

80

Presented at the 1999 Government Microcircuits Applications Conference (GOMAC) March, 1999 measured before the start of this particular test. The figure shows the cantilever reaction to 0.1% hydrogen in nitrogen (1000 ppm hydrogen concentration). The typical capacitance change for this test was about 20 fF. This cantilever system has also been used to detect hydrogen levels down to approximately 100 ppm. Lower levels are practical with thinner cantilevers and coatings.

0.12

Fig. 3. The chip was fabricated in a 0.5-µm bulk CMOS process.

3.50

0.1

3.48

0.08

3.46

0.06

3.44

0.04

3.42

0.02

Channel Output

Percent Hydrogen

Uncoated

3.40

Pd coated

10145

9957

10051

9863

9769

9676

9582

9488

9394

9301

9207

9113

9019

8926

8832

8738

8644

8550

8457

8363

8269

8175

8082

7988

7894

3.38 7800

0

Seconds From Start

Fig. 6. Multichannel hydrogen data RF TELEMETRY A wireless network for data reporting is needed to field arrays of distributed sensors. For this, we have developed an RF-telemetry chip with onchip spread-spectrum encoding and modulation circuitry to improve the robustness and security of sensor data in typical interference- and mul6,7 tipath-impaired environments . We have also provided for a selection of distinct spreading codes to serve groups of sensors in a common environment by the application of code-division multiple-access techniques. Our initial intended operation is for use in the 915-MHz Industrial, Scientific, and Medical (ISM) band. The ‘Wirtx1’ chip, shown in Fig. 7, is comprised of a 10-bit analog/digital converter (ADC) with four input multiplexer, a 63-bit digital spreadingcode generator, a state-machine controller that allows the chip to act as an unattended data acquisition system, and the radio-frequency modulator and transmitter. The entire chip operates on 3.3V and is mounted on a printed-circuit board on the bottom side of the battery pack shown in

Fig. 7. Wirtx1 wireless data acquisition chip SUMMARY We have developed a sensor-readout-telemetry system that is battery-operated, utilizes multiple microcantilever sensors that allow mixtures of vapors to be measured with a high sensitivity. This research was sponsored by the U. S. Dept. of Energy and performed at Oak Ridge National Laboratory, managed by Lockheed Martin Energy Research, Inc. for the U. S. Dept. of Energy under Contract No. DE-AC05-96OR22464. REFERENCES 1. G. Y. Chen, T. Thundat, E. A. Wachter, and R. J. Warmack, (1995). “Absorption-induced surface stress and its effect on resonance frequency of microcantilevers,” J. Appl. Phys. 77(8), 3618-22. 2. T. Thundat, P. I. Oden, and R. J. Warmack, (1997). “Microcantilever Sensors,” Microscale Thermophysical Engineering 1(3), 185-99. 3. T. Thundat, G. Y. Chen, R. J. Warmack, D. P. Allison, and E. A. Wachter, (1995). “Vapor Detection Using Resonating Microcantilevers,” Anal. Chem. 67(3), 519-21; T. Thundat, E. A. Wachter, S. L. Sharp, and R. J. Warmack,

Presented at the 1999 Government Microcircuits Applications Conference (GOMAC) March, 1999 (1995). “Detection of Mercury Vapor Using Resonating Cantilevers,” Appl. Phys. Lett. 66, 1695-7. 4. P. I. Oden, P. G. Datskos, T. Thundat, and R. J. Warmack, (1996). “Uncooled Infrared Imaging Using a Piezoresistive Microcantilever,” Appl. Phys. Lett. 69(21), 3277-79. 5. P. I. Oden, G. Y. Chen, R. A. Steele, R. J. Warmack, and T. Thundat, (1996). “Viscous Drag Measurements Utilizing Microfabricated Cantilevers,” Appl. Phys. Lett. 68(26), 3814-16. 6. C. L. Britton, Jr., R. J. Warmack S. F. Smith, P. I. Oden, R. L. Jones, T. Thundat, G. M. Brown, W. L. Bryan, J. C. DePriest, M. S. Emery, M. R. Moore, G. W. Turner, A. L. Wintenberg, T. D. Threatt, Z. Hu, J. M. Rochelle “Batterypowered, wireless MEMS sensors for highsensitivity chemical and biological sensing” to be presented at the Symposium on Advanced Research in VLSI, Atlanta, GA, March, 1999. 7. C. L. Britton, Jr., R. J. Warmack, S. F. Smith, P. I. Oden, G. M. Brown, W. L. Bryan, L. G. Clonts, M. G. Duncan, M. S. Emery, M. N. Ericson, Z. Hu, R. L. Jones, M. R. Moore, J. A. Moore, J. M. Rochelle, T. D. Threatt, T. Thundat, G. W. Turner, A. L. Wintenberg, “MEMS Sensors and Wireless Telemetry for Distributed Systems”, Proceedings of the SPIE, Vol. 3328, pp. 112-123.