A Passive Wireless Integrated Humidity Sensor
`
`Timothy J. Harpster, Brian Stark, and Khalil Najafi
`
`Center for Wireless Integrated Microsystems
`University of Michigan, Ann Arbor MI 48109-2122
`
`ABSTRACT
`
`This paper presents a single-chip integrated humidity sensor (IHS)
`capable of wireless operation through inductive coupling with a
`remote transmitter. The 1.5x0.5x8mm3 sensor chip consists of a
`planar electroplated copper coil (20pm thick, 30pm pitch, and 23
`turns) and a silicon substrate separated by a 4100-56OOA
`polyimide film. The resonant frequency of the IHS changes with
`humidity and the measured sensitivity ranges from 4-16kHz/%RH.
`Measurements show a hysteresis of 4.5%RH over a range of 30-
`70%RH for a 5600A-thick polyimide film IHS.
`
`INTRODUCTION
`
`Wireless operation is desirable in many emerging sensing
`applications, including those where sensors must operate in harsh
`environments and physical access to the sensor is not possible. In
`many of these applications, the sensing system should be powered
`remotely, occupy a small area, and provide sufficient resolution
`and sensitivity. One such application is in monitoring the
`hermeticity of implantable systems in-vivo [l]. Although dew
`point sensors, humidity sensors, and pressure sensors have been
`used to determine package hermeticity, these sensors require more
`difficult fabrication processes and are wired to an external power
`source andlor readout circuitry [l-31.
`
`The passive wireless link has been proven useful for in-vitro and
`in-vivo monitoring of an implantable micropackage. In MEMS
`2000 we reported a hybrid wireless humidity sensor consisting of a
`capacitive humidity sensor chip and a hybrid coil made up of
`copper wire wound around a ferrite core [4]. The hybrid sensor
`was utilized to test hermetic glass-silicon packages in accelerated
`testing environments and implanted in biological hosts. The
`development of a passive wireless integrated humidity sensor
`(IHS) combines a capacitive polyimide-based humidity sensor with
`a receiving coil onto a single substrate. The IHS presents
`significant improvements over the reported hybrid version of this
`sensor including simple processing, batch fabrication, repeatable
`sensor characteristics, comparable performance, and smaller size
`which is necessary for low profile packaging.
`
`DESIGN AND ANALYSIS
`
`Figure 1 shows an integrated humidity sensor and external
`transmitting antenna (a 3 turn layout is drawn for clarity). This
`system consists of a planar electroplated copper coil (20pm thick,
`30pm pitch, and 23 turns) and a silicon substrate separated by a
`4600-5600A Pyralin@ PI2613 polyimide film. The device makes
`use of the coil interwinding capacitance, C,,, and substrate
`distributed capacitance, C,, as shown in Figure 2. The IHS is
`modeled as an LC tank circuit where the electroplated copper coil
`forms both the inductor and the humidity sensitive capacitor. The
`natural resonant frequency of this system changes when the
`
`0-7803-5998-4/01/$10.00 (3200 1 IEEE
`
`553
`
`permittivity of the humidity sensitive polyimide under the coil
`changes thus altering the distributed capacitance. To remotely
`monitor the resonant frequency shift due to humidity changes, a
`1.0-1.5cm diameter loop antenna is used to stimulate the tank
`circuit, also shown in Figure 1.
`
`-cu
`
`Coil
`-Polyimide
`
`+Silicon
`Figure 1. Passive humidity monitoring system.
`
`Figure 2. Capacitive elements of the IHS.
`
`This resonant frequency shift is measured as a change in the load
`impedance reflected back to the antenna, as illustrated in the
`equivalent circuit shown in Figure 3. It will be shown that the
`reflected load impedance is a function of the humidity sensor
`capacitance. Thus one can extract the resonant frequency by
`monitoring the impedance change of the transmitting antenna.
`
`Equivalent
`Integrated
`External
`Circuit
`Humidity
`Antenna
`Sensor
`Figure 3. IHS equivalent circuit model.
`
`Abbott
`Exhibit 1020
`Page 001
`
`

`

`The equivalent circuit is used to model the IHS and aid in
`optinlizing the sensor performance. The following assumptions
`are tnade to simplify the inductive coupling of the equivalent
`circuit model: 1) the antenna and IHS coils are aligned coaxially;
`and 2) the IHS cross sectional area is much smaller than the
`external antenna.
`These assumptions allow the following
`approximations to be made: 1) the magnetic field of the extemal
`antenna will be calculated only along the coaxial z-axis, thus the
`magnetic field is a function B(z); and 2) the magnetic field being
`coupled to the IHS will be assumed constant across the IHS cross
`secticin. Equation nomenclature is given in Table 1.
`
`R = d , - P
`wh
`
`The IHS capacitance, C, is not modeled in this discussion and the
`measured value is used throughout this paper. Note that
`conventional analysis of integrated coil capacitance cannot be used
`because the substrate is floating. The aspproximated mutual
`inductance [5], a ratio of flux linkage on the 1 HS circuit due to the
`current in the extemal antenna. is
`
`The impedance of the IHS is
`
`( A)
`
`Z , H s ( ~ ) = R +
`j w L - -
`
`The impedance seen at the extemal antenna is given by
`
`From Eqs. 9 & 10, the impedance of the extemal antenna at the
`resonance frequency of the IHS is
`
`Figure 4 shows k(w)l and LZ(w) for decreasing separation
`'distance, z = 0.6.0.3 and Ocm, as seen in the increase in phase dip
`magnitude. Table 2 lists the simulation paratneter values.
`
`Table 1. Circuit moa
`-- Antenna:
`series resistance
`h!,
`L,
`inductance
`r,,
`radius
`number of tums
`A!,
`I,,
`antenna current
`core permeability
`po
`
`CouDlinq:
`M mutual inductance
`z
`separation distance
`
`equation nomenclature.
`Integrated Humidity Sensor:
`R
`series resistance
`L .
`coil inductance
`C
`total capacitance
`permeability
`P
`coil line-width
`W
`line spacing
`S
`h
`coil thickness
`N
`number of tums
`coil length
`Lcoi/
`coil width
`Wcoi/
`winding length
`dL
`5
`polyimide thickness
`
`Using Biot-Savart's law for a circular loop with radius r,, number
`of windings N,, carrying current I,,
`the magnetic flux density
`through an extemal antenna along the z-axis is at a distance z is
`
`The extemal antenna inductance is
`
`The inductance [SI, L, and resistance, R, of a planar spiral
`rectan,ylar coil is given by
`Ltr = L,,, - Nw-(N -1b
`W, = W,,, - NW - ( N - 1)s
`,
`(Eq. 4)
`kl = L, Iog(Leff +JL,')
`(Eq. 5)
`
`(Eq. 3)
`
`L=9.21x10-'N2,u (Llr +W,)log
`
`[
`
`3
`
`1
`
`- k l - k 2
`
`(Eq. 7)
`
`.-
`
`(8Lcflweff
`
`554
`
`Abbott
`Exhibit 1020
`Page 002
`
`

`

`Table 2. Simulation values.
`
`1 .O cm
`- 40pF
`- 8 m m
`
`10 m
`- 20 m
`- 23 turns
`
`The resonant frequency. and phase dip minima, occurs at
`
`(Eq. 14)
`
`The impedance phase dip magnitude is the difference in LZ(wJ
`between a coupled (M = finite) and uncoupled (M = 0) antenna,
`thus
`
`For maximum I Apdjr I the last term of Eq. 12 is maximized and the
`imaginary part is minimized. This is accomplished by increasing
`fi pa, N, Lc ",,, and Wcoj, and decreasing 5. z, Nn, and R. The Q of the
`IHS is given by
`
`Minimizing R and maximizing L increases QIHS as well as
`increases
`the reflected
`impedance and
`thus .the phase dip
`magnitude. High Q sensors are desired because the phase dip
`minima is more readily identified at sharp phase peaks and the
`increased reflected
`impedance extends the maximum testing
`distance between the IHS and the external antenna.
`
`Note that the resonant frequency, w,, is independent of coupling
`magnitude and thus humidity monitoring is only based on the
`resonant frequency identified at the phase dip minima.
`
`I
`
`FABRICATION AND MEASUREMENT
`An SEM photograph of the IHS is shown in Figure 5. The
`fabrication procedure is as follows: 1) -5000A of Pyralin@ PI2613
`polyimide is spun on a silicon wafer, soft baked at 150°C for
`30min, and cured on a hotplate at 350°C for 30min, 2) a TilCu
`seed layer is sputtered onto the cured polyimide, 3) 20pm AZ9260
`resist is spun and patterned to form the Cu coil mold, 4) a 20pm
`Cu coil is electroplated and the resist and seed layers are removed.
`The completed l.SxOSx8mm' sensor chip consists of a planar
`electroplated copper coil (dimensions given in Table 2) and a
`silicon substrate separated by a 4100A-5600A humidity-sensitive
`polyimide film. The IHS is inductively coupled to an external loop
`antenna, (24awg tin wire, 4-8 tums, s1.0-1.5cm diameter). The
`impedance of the antenna is measured with high resolution using .
`an HP41 94A ImpedancdGain-Phase Analyzer. Experimental
`values for R", Lo, R, L and C are listed in Table 3. Figure 6 shows
`the measured kwl and LZ(R of a fabricated system.
`
`Figure 5. SEM of the integrated cupper coil on polyimide.
`
`I
`I
`I
`I
`I
`I
`I
`I
`1
`Figure 6. Impedance measurements k(fj and LZyl.
`RESOLUTION AND SENSITIVITY
`
`J
`
`The IHS is placed in an ESPEC temperature and humidity chamber
`to calibrate the device. The temperature of the chamber is
`maintained constant and the impedance of a loop antenna (coupled
`to the IHS inside the chamber) is measured to monitor the IHS
`resonant frequency as humidity is controlled from 30% to 70%RH
`at 10% intervals. The normalized sensitivity, S, is
`
`The results below are obtained from a test set of 9 devices - 3
`sensors with I,, = 5600A, 3 sensors with tp = 4600A. and 3 sensors
`with tp = 4100A. Calibration curves at 37°C for all nine devices
`are shown in Figure 7.
`
`555
`
`Abbott
`Exhibit 1020
`Page 003
`
`

`

`18.5
`
`18
`
`.-.I IN :c :E
`
`b d
`:r,
`17.5
`
`17
`
`16
`
`%
`I!!
`UL
`*-
`I= 16.5
`(0
`1:
`p'e
`
`15.5
`20%
`
`30%
`
`40%
`
`60%
`
`70%
`
`80%
`
`50%
`%RH
`Figure 7. Calibration data for all 9 devices at 37°C.
`
`It is observed that sensitivity increases for increasing polyimide
`thickness, tT The increase in sensitivity with increasing polyimide
`thickness contradicts results from other capacitive humidity
`sensors and this phenomena is not fully understood.
`
`Temperature
`
`.
`
`Calibration curves at 25"C, 37°C and 50°C of two devices with a
`5600A polyimide layer are shown in Figure 8. The saturation
`density of water vapor increases with temperature thus it is
`believed that the increase in sensitivity with temperature is due to
`increased water vapor density, pw, absorbed into the polyimide at
`consiant %RH for increasing temperature. The increased water
`vapor density absorbed into the polyimide at higher temperatures
`increases the permittivity, E, of the polyimide thus an increased
`positive AC occurs for fixed A%RH at higher temperatures.
`1 A .
`
`20%
`
`30%
`
`40%
`
`60%
`
`70%
`
`80%
`
`50%
`%FlH
`-
`Figure 8. Calibration.results for two tp = 5600i sensors at 25"C,
`37"C, or 50°C.
`
`The average and standard deviation of the normalized sensitivities
`are extracted from the calibration curves from the set of 9 devices
`and are listed in Table 4.
`
`The resonant frequency measurement depicted in Figure 6 can be
`measured within +20kHz, giving a tp = 56DOA sensor @37"C a
`resolution of fl.7%RH excluding effects of hysteresis or drift
`(discussed in the following sections). Note that the resonant
`frequency measurement resolution of f20kHz may be improved
`with a higher QIHs as previously discussed.
`
`Detection range
`
`The IHS can be wirelessly interrogated as long as the coupling
`between the transmitter and IHS is large enough such that the
`phase dip magnitude can be detected. From Eqs. 10 & 15 the
`phase dip magnitude, I A%,,,
`I , can be incleased by modifying pa
`using ferrite in the external antenna. Matenal#61 femte is used in
`the extemal antenna and has a published permeability of 125
`[Amidon Associates]. The maximum testing distance is lcm using
`the ferrite core antenna whereas the maximurn testing range is only
`0.6cm with an air core antenna. Increasing the inductance of the
`integrated coil, decreasing the sensor capacitance, or reducing the
`sensor resistance can further increase the maximum testing
`distance (as depicted in Figure 4). The phase dip magnitude for
`both air core and femte core antennas is plotted versus separation
`distance, z, between the transmitter and IHS in Figure 9. The
`differences in theoretical and measured results are due to
`approximations made in the mutual inductance derivation which
`ignore fringing of the antenna magnetic field.
`r - ---
`I /
`E
`
`0 Measured - Alr Core
`
`Hysteresis
`The hysteresis of the sensor is measured at 25°C. The nine devices
`are placed in the humidity chamber and the humidity is increased
`from 30%RH to 70%RH then decreased back to 30%RH in
`lO%RH increments and measurements are taken in 1 hour
`increments to ensure full sorption of water vapor. Figure 10 shows
`the results of one of the tp = 5600A devices. The calculated
`hysteresis is 4.5%RH.
`
`556
`
`Abbott
`Exhibit 1020
`Page 004
`
`

`

`A passive wireless integrated humidity sensor has been developed.
`A i, = 5600A sensor can be used to detect humidity changes within
`5
`f2%RH at 37OC. The sensor can be wirelessly monitored up to
`lcm from a stimulating external ferrite core loop antenna. Future
`work will implement this wireless sensor for testing hermeticity of
`implantable micropackages.
`The IHS will facilitate fully
`automated testing, another useful application for this technology.
`
`d 17.25
`
`20%
`
`30%
`
`40%
`6OX
`SO0?$
`Y&H Change
`Figure IO. Hysteresis of IHS @ 3PC.
`
`70%
`
`One device exhibits a significant amount of drift, corresponding to
`about 10%RH (15lkHz from hour 1 to hour 48) at 40%RH and
`about 20%RH (300kHz) at 80%RH. This source of this drift is
`unknown however it could be due to swelling of the polyimide
`film due to water vapor sorption [8].
`
`CONCLUSION
`
`80%
`
`ACKNOWLEDGEMENTS
`
`The authors wish to thank Dr. F. T. Hambrecht and Dr. W. J.
`Heetderks of the National Institute of Health for their guidance and
`encouragement. This research is supported by the NIH, under
`contract # NIH-NO1 -NS-8-2387.
`
`REFERENCES
`
`Drift is a common characteristic seen in polyimide sensors [3, 6,
`71. The nine devices are placed in the humidity chamber at 40°C
`and 8OoC and monitored for 48 hours. The results are plotted in
`
`Dr$i
`
`Figure 11. -
`
`18.0
`
`N
`
`1
`
`B. Ziaie,, et. al. “A hermetic glass silicon micropackage
`with high-density on-chip feedthroughs for sensors and
`actuators.” JMEMS, pp.166-179, Vol. 5, No. 3, Sept.
`1996.
`M. Waelti, et. al. “Package quality testing using
`integrated pressure sensor.” International Journal of
`Microcircuits and Electronic-Packaging, v01.22, no. 1;
`1999; 11.49-56.
`M. Dokmeci, et. al. “A high-sensitivity polyimide
`humidity sensor for monitoring hermetic micropackages”
`MEMS’99, pp. 279-284, Jan. 1999
`T. Harpster et. al. “A passive humidity monitoring
`system
`for
`in-situ
`remote wireless
`testing of
`micropackages”, MEMS 2000, Miyazaki, Japan, Jan 22-
`27,2000.
`J. Von Arx, “A single chip, fully integrated telemetry
`powered system for peripheral nerve stimulation”,
`University of Michigan Ph.D. thesis, 1998.
`A. R. K. Rafston et. al., “A model for the relative
`environmental stability of a series of polyimide
`capacitance humidity sensors”, Sensors and Actuators,
`vol.B34, no.1-3; Aug. 1996; p.343-8.
`M. Matsuguchi et. al., “Drift phenomenon of capacitive-
`type relative humidity sensors in a hot and humid
`atmosphere”, Journal of the Electrochemical Society,
`~01.147, no.7; July 2000; p.2796-9.
`R. Buchhold, et. al. “Reduction of mechanical stress in
`micromachined components caused by humidity-induced
`volume expansion of polymer layers”, Microsystem-
`Technologies. vo1.5, no.1; Oct. 1998; p.3-12.
`
`5
`
`0
`
`10
`
`30
`20
`Time [Hours]
`Figure 11. Dr$t ai 37°C for 40%RH.
`
`40
`
`LL. 16.5
`c,
`
`d 15.5
`
`*
`
`0
`C
`
`16.5
`i7*0*
`
`0
`
`10
`
`30
`20
`Time [Hours]
`Figure 12. Dr$t at 37°C for 40%RH and 80%RH.
`
`. 40
`
`50
`
`50
`
`557
`
`Abbott
`Exhibit 1020
`Page 005
`
`