Terrestrial Applications of Nano Accelerometer by mikesanye


									                        Terrestrial Applications of a Nano-g Accelerometer
                                                       Frank T. Hartley
                                                  Jet Propulsion Laboratory
                                              California Institute of Technology
                                         4800 Oak Grove Drive. Pasadena, CA. 91 109

Abstract:                                                           gravitation field by electrostatic levitation. The voltage to
     The ultra-sensitive accelerometer, developed for NASA          accomplish this is afunctionof the accelerationimposed
to monitor the microgravity environments of Space Shuttle,          upon the proof mass. This results in the elimination of one
fke orbitors and Space Station, needed to measure                   of the most serious difficulties in static and dynamic earth
accelerations up to 10 mg with an absolute accuracy of 10           calibrations of micro-g accelerometers. The same levitation
nano-g (IO" g) for at leasttwo orbits ( lo4 seconds) to             feature permits theaccelerometer to be nulledinawide
resolve accelerationsassociated with orbital drag. Also, the        varietyof conditions. The force actuation can provide an
accelerometersneeded to have less than los9 F.S. off-axi~           alternating excitation of the sensor to dynamically calibrate
sensitivity; to be thermally and magnetically inert; to be          it over the frequency range of interest. In addition to both
immune to quiescent shock, and to have an in-situ                   ground and in flight calibrations, this featlrre permits health
calibration capability.                                             monitoring, coefficient correction and sensor characterization
     Multi-axis compact seismometers designs that have              over long term space flights.
twelve decades of dynamic range will be described. Density               A fiather important design issue for accurate tri-axial
profilometers, precision gradiometers, gyros and vibration         accelerationmeasurement is the minimization of off-axis
isolation designs and applications will be discussed.              sensitivity. This was accomplishedby insuring that the
Finally, examples of transformations of theaccelerometer           proof mass and spring design was symmetric, weak in the
into sensitive anemometers and imaging spectrometers will          compliant direction, and operated with the tunneling tip in
be presented.                                                      close proximity to the unperturbed proof mass. As a
                                                                   consequence of a zero &flection flexure, thermal sensitivity
NANO-G ACCELEROMETER                                               is reduced.Thermalmismatched stress is also eliminated
     Under a     NASA advanced technology development              by fabricating the entire die out of mono-crystalline silicon
contract a nano-g accelerometer      was
                                      developed               in   and bonding these die together directly (i.e.no interface
collaboration with Northeastern University". The intended          material). Figure 1 shows a not to scale cross section of the
use of the accelerometer was the tri-axial measurement c f         accelerometer indicating the important featuresofthe 6 r       u
orbital drag on the Shuttle and Space Station which                dice structure. The tip die (the top die in the figure) has an
required an acceleration range of 10z-lO" G over a f k p n c y     approximately 3.75 p n high tunneling tip at its center.
range of 0.0001 - 25 Hertz.                                        Two identical proof mass dice are rotated by 180" and
     Silicon micromachined devices reported by others have         bonded together to form the 'proof mass'. The net weight
not achieved the necessary sensitivity. These devices have         of the proof  mass     is 0.18 gm which is held to the
been based on either piemresistive or capacitive position          surrounding fiameby a set of springs ref& to as 'crab
sensing elements. The resolution of accelerometers is              legs' and located in the exact center of the structure with the
directly proportional to the position detection capability         tunneling tip designed to just touch the unperturbed proof
and the square of the timdamental hquency of             the       mass.
mechanical structure.Our accelerometerdesign is motivated                Figure 2 shows a portion of the top view of the proof
by the need for small size and high sensitivity. For a low         massand springs and Figure 3 shows atop view ofthe
mass system this dictates ultra sensitive position detection       force plate die wherethemetalplaten         is covered with an
such as that of an electron tunneling ti? with an extreme          oxide layer (0.5 pm) to prevent an electrical          contact
spatial resolution of less than 0.00 1B H .'.
                                       Lz                          between the proof mass and the force plate when the proof
     Three distinctively different dice are fabricated    and      mass is being electrostatically clamped. Figure 4 shows the
subsequently assembled using a 'zero' thickness bonding            tip die complete with access ports for external connection.
technique. Hermetic sealing and             connections
                                    electrical                     Zero thickness referenced bonds essential to maintain the
between the diEerent dice are also accommodated during             tight spatial tolerancingandthermal insensitivity required
this bonding operation. The accelerometer is controlled by         of this accelerometer. The low temperature (< 400' C)
electrostatic force plates above and below the proof mass.         eutectic bonding, sealing and inter-connection       procedure
The lower  electrode has a dual role. In operation, it             wasdeveloped where etched channels werecreatedinthe
provides anecessary control electrode, andwhennot             in   bond regions on whicha spreading layerofmetal               was
operation, it is used to immobilize the       proof   mass to      deposited andpatterned in the channels.Finallythebond
protect the flexures and particularly the tunneling tip.           metal was deposited and patterned on top of the spreading
     The active element (proof mass) of theaccelerometer           layers in such a way that it protruded       above the     wak
can be electrostatically suspended at the null position in a       surface andwas narrower than the spreading layer such that
    w i c Bcnd
    Aman Pm

  Fig. 1 AccelerometerCross Section            Fig. 2 TopView of Proof Mass Fig.             3 TopView of Force     Plate
                                                                   If bonded to a silicon cube the accelerometers are
                                                              thermally inert and their proof masses can electro-
                                                              statically levitated to null static fields (i.e. earth or planet
                                                              gravity). A 'smart' three-axis accelerometer/seismometer is
                                                              realizedbythe       integration of an ultra precision voltage
                                                              reference (temperature-controlled   zener diode), precision
                                                              DAG, and a three-decade overlap in accelerometer scalings.
                                                                   Deployed as a terrestrial seismometer this single unit
                                                              could replacethe dual Streckeisen and low-g seismometers
                                                              used for earthquakemapping.Theextremelysmall                sue,
                                                              thermal insensitivity, and robustness (electrostatic 'caging'
                                                              protecting @omshockloadingandquiescenthandlingl
                                                              makes these systems ideal sensors use down bore holes
                 Si,N, Cr
                                                              andfor p l a n e t q seismometFy.
         \ /                                                       TOLERANT GRADIOMETER
                   Figure 4 Tip-Plate Die                               NASA has an interest in an accurate (< 1 milligal)
it's volume was less than the volume of the channel. When          gravity field measuring instrument for deployment on the
two dice prepared in this wayare brought in contact aid            Shuttle, or a ' h e flyer', to map the global earth gravity
heated, the bond metal melts and spreads by wicking and            field. Such an instrument would enable geophysicists to
capillary forces reducing the spacing between the wa6er            understand plate dynamics, plumes and mantle structure
surfaces to zero.                                                  and provide oceanographers precise
                                                                                                      a     geoid            &x
     The sensitivity of this sophisticated accelerometer was       determining oceancurrents and other Ocean phenomena.
determined to be sub micro-G and superior to the best CE                The component micro-Gaccelerometersrequired lbr
the commercial micro-G                  (QA3000).
                           accelerometers                          such a gradiometer are single axis devices that exhibit very
Thermal dependence was not discernible and stability and           low cross-coupled interference. The alignment. of a cubic
accuracy were belowthe noise floor of the test system and          array of 81 such accelerometersto the gradiometer structure
believed to meet it's design criteria of 10-*-10"G over a          (arranged in 27 vector triads) will be a major activity, as
frequency rangeof 0.0001 - 25 Hertz.                                                                        each
                                                                   will theempirical determination of accelerometer's
     Aside interest
          @om                in the
                                               by                  vector relative to the orthogonal axis of the gradiometer.
international space faring community the largest market is         While in principle only four such triads are sufficient, the
in sensitive seismometery and geological density gradient          extrameasurement permit the removalof all the most
survey applications  particulmly down deep bore holes.             serious errors - self gravity, gain, alignment mismatch, and
                                                                   vibration rectification. The instrument would consist of a
MULTI-AXIS SEISMOMETER                                             solid cube of less than 10 cm dimension.
     The architecture of the accelerometer is such that large           In space applications themainadvantage        of such a
static or slowly varying accelerations can be electrostatically    gradiometer, over existing designs, is its much     greater
compensated allowing nano-g resolution in one-g fields (Le.               spacecraft
                                                                   tolerance                     shortcomings, including lice
earth) with seven decades    of       dynamic range. The           propellants and other moving masses, and vibrations h m
acceleration sensitivities of these micro-machined                 articulated components. Thus, thespacecraftneednot        be
accelerometers are proof mass selectable. If accelerometers        designed around the gradiometer, and other payloads may
are designed to measure full-scale accelerations of g and
                                                    one            easily
                                                                  more                                         Also, the extm
one Kilo-g, with accuracies    of
                               micro-g           and mili-g        accelerometers that
                                                                                mean performancedegrades                   quite
respectively, their co-location provides      an acceleration      gracefully should individual accelerometers fail.
measuring instrument that spans nano-gKilo-g.              An                 long
                                                                        It has been          known that gradiometers actually
orthogonal triad oftheyethreeaccelerometers ( 109-10"g,            measure components of an "intrinsic tensor" - a
10d-lOog and 10"-IO g) would three-axis a                          combination of the gradient and various angular velocity
seismometer with 12 decades of dynamic range.                      and acceleration terms. The direct removal of these terms
                                                                   by practical gyros, or other attitude measurements, is not
done tosufficientaccuracy.However,         dynamic estimation     set of locations and driving momentumactuators         (linear
based on the rotational translational
                          and                    equations d       motors) to null thesensed response.
motion ofthe free floatinggradiometerandthe spacerraft                   The advantages of this approach over active platform
attitude measurements greatly improve gradient estimation.         isolation are that the    entire environment is improved
     Allaccelerometersand       gradiometers are subject to        throughoutthe structure, not just that of theplatform.
attraction of local masses and masses fixedin instrument           While the platform disturbances are attenuated the rest cf
coordinates generally an
                      cause             output bias that is          structure
                                                                   the       too,
                                                                             is             through improved damping and
indistinguishable from bias fiom other sources, and                reduction of settling time following transient disturbances
irrelevantfor geophysical purposes. However, propellant            and excitation of structural modes. VSS essentially mimics
motions, articulated objects, outer drag fiee motion, and          freestanding shockabsorbers. Small VSSs could be placed
thermal distortions may cause signals with power spectra           at the extremities of flimsy panels (i.e.. solar panels) to
similar to the earth or planetary     gravity tensor signal.       attenuate transient disturbances and prevent the excitation
This is self-gravity which, except for propellant motions,         of their structural modes. The mass of VSS could
are all measurable, and can be removed by calculation. In          potentially be less than the mass savings representedby
differencing gmbometers, accelerometer gainmismatch and            more flimsy    panels. For microgravity experiments all
input axis misalignment both yields errors proportional to         disturbances above the sub-Hertz hquencies will be
the common acceleration. In the proposed instrument emx            eliminated but not     at the expenseaof       neighbor or
calibration and the ensemble of      accelerometers should         contributing to the excitation of the hundreds of structural
greatly improve the detection and removal of these d e             resonances of space vehicle.
factor errors.                                                           Whilethe microgravi@ experimenterswouldbethe
     Vibration rectifcation results h m compliance in the          immediatebeneficiaries of VSS it would be of global
accelerometer mounting structure in a differencing                 benefitwhen qplied to largestructures(terrestrial         as
gradiometer leading to a DC signal even if the vibration          well as in space) and potentially be a saving technologv
frequency lies outside the pass band of the accelerometers.       for Space Station Alpha.
Howeverby including athree-dimensionalmodelof             the
accelerometer mounting structure and redundant                    DERIVATIVE M W       E
accelerometers (81 verses 27) a     clear signature of the              These two derivative devices utilize the zero thickness
gradients is expected in spite of rectification.                  bonding, encapsulation, wiring, micro flexures and
                                                                  electrostatic caging technologies spawnedby the nano-g
VIBRATION SUPPRESSION                                             accelerometer development. The electrostatic ‘caging’
     There are two dynamic means of mitigating                    innovation also spawned the concept for an electrostatic
disturbances that may be applied to microgravity payloads         peristaltic pump, and inter wafer eutectic wiring spawned
on board the Shuttle or Space Station. One means, an              techniques for implementing ultra dense electronic circuit
active isolation system, takes the      approach       that the   fabrication.
platform to be isolated be allowed to “float” or “sway” to        Active Optical Filter
a certain extent within some sway space. Active Isolation               The fust derivative device under development is an
uses dynamic control to improve its performance            by     active optical filtermanufactured       as an assembly d
deploying microgravity measuring accelerometers mounted           micromachined silica wafers. The wafers cany stationary
on the isolated platform to calculate the necessary forces to     and movable  mirrors               a
                                                                                           that formreflectiveFabry-Perot
be applied through a set of celocated actuators to cancel         cavity. To make the device feasible the mirrors must be
the forces that are transmitted to the isolated platform          aligned and  keep their alignment. The position of the
through the stifhess, damping, and fiictional efkcts              mirrors has to be controlled and measured while the
inherent the
           in        coupling and umbilical cables. An            geometry of the filters must not change with time or the
adversarial effect of active isolation is that the motion the
                                                          of      environment.
isolated platformhas a reciprocaland amplification effect on            In the reflectiveFabry-Perotinterferometerthe       p f
the motion of host
                the            structure (i.e. will aggravate     mass and spring flexures of the accelerometer are used as a
environment of neighbors). Also if multiple experiments           mirrorplatformto      maintain surhx parallelismaccurately
are installed on a single isolated platform, servicing one        and constrain the cross-axial location of the movable mirror
experiment will disturb all the others.                           to the sub-pm level. The accelerometer tunneling tip is
     Conversely a Vibration Suppression System (VSS)              removed from the    center quad plate (Figure 4) and the metal
consists of a set of linear proof mass actuators mounted to a     platens are covered with a thin insulation layer to facilitate
payload or structure that apply the forces required to cancel     ‘caging’ of the mirror. A thin metal layer is deposited on
the effectsof disturbances. Becausethe primary cause d            the lower s&e of an optically flat transparent substrate
large accelerations is the amplification of     disturbances by   followed by a dielectric mirror coating. The quad plate die,
the lightly damped structural modes of a Shuttle or Space         two proofmass      dice and thin spacing d c e are bonded
Station, a system that    produces     forces that cancel the     together to formthe adjustable cavity. Figure 5 illustrates
excitations due to disturbances effectively increases the         how the dielectric coated mirror is placed over this cavity to
damping of structural modes and significantly reduce the          produce reflection
                                                                          a        Fabry-Perot          cavity with the    facing
structure’s response. A Vibration Suppression System does         surface ofthe movable mirror (proof mass). The electrostatic
this by sensing the (acceleration) response at an optimized       platens on either side of the ‘proof mass’ fice electrostatic
                                                                  platens onthequaddie        and behind the dielectricmirror
  enabling monitoring and    control of the spacing of     the     are then made to both M e plates and each set of quad
 optical elements via   capacitancemeasurementand           time   plates from bothsides of the assembled unit.
  feedback through electrostatic actuators.                             During handling,   launch     etc., the baffle plate is
       Such devices promise significant advantages in              clamped to either of the quad plates by means of a small
  instrumentation for space     astrophysics and quantitative      battery. The separationcapacitance’s ofeach quad plate,
  imaging science in general    and      may
                                                 a                 withrespectto      its’ hcing M e plate, is measuredand
 building block of the future optical imaging systems cf           applied, viacontrol
                                                                               a                algorithm, to holdthe      hailk
 microspacecraft probes, and rovers.                               stationary in the center of the cavity. When the ba& is
       One of the many potential applications of such a filter     mounted perpendicularto the gas flow vector of interest the
 is to augment the Hubble Space Telescope (HST) advanced           pitot static force(pressure x aperture area) applied firr>m
 radialcamera(WFPC-111)spectralfilter          set. The use cf     either side of the   anemometer is determined from     the
 ramp filters on WFPC-I1 has demonstrated the advantages           aggregate voltages on each set of quad plates.
 of ”tunable” filers. Fast moving objects can create very large         The micromachinedanemometer dimensions are less
 red shifts. It is nearly impossible to provide a set of fixed     than 2 cm square and 2 mm thick, it does not need to be
       that would cover     all eventualities. The proposed        mounted in a tube and two (or three)ofthemmounted
 tunable filter can provide complete spectralcoverageand           orthogonaly (60”) will   provide    wind velocity. A unit
 simultaneous imaging that is not possible with ramp filters.      mounted horizontally would provide a       measurement c f
 This capability will enable WFPC-111 to operate as an             vertical drafts. The bi-polar forcebalancedarrangement c f
 imaging spectrophotometer. Another major advantage of a           this anemometer provides for a wide dynamic range (lo>
 micro-machined tunable filter is the lack      of blue shift      and a bandwidth from sub-Hertzto hundreds of Hertz. The
 degradation that  affects fmed filters.                           zero flexure deflection design and electrostatic actuator
       The Fabry-Perot  based        filter permitsvariable        control ensure the anemometer is athermal and the relatively
 bandwidth,continuoustuninganda             possibility ofseIf     large M e plates (0.5 cm’) and compliant suspension
 calibration. It ofers low weight, low energyconsumption           ensure high sensitivity.
and an extreme thermal stability. The device has numerous               Themarket for Martiananemometerswouldnot               be
 commercial applications     particularly as a multicolor $her     commerciaI(y compelling, however, the marketfor sensitive
for camcorders and digital cameras.                                anemometers in industrial    and
                                                                                                domestic                  WAC
       FPbry-Rrot I n t a f e r y c e Cavity   mdeeQie Mirror
                                                                   applications b large.

                                                                        The work         was
                                                                                 described performed                    Jet
                                                                                                                    at the
                                                                   Propulsion Laboratory, California Institute of Technology
                                                                   undercontract          National
                                                                                     to the                  and
                                                                                                   Aeronautics Space
                                                                   Administration. The author would like to thank Prof. Paul
                                                                   Zavracky of Northeastern Universityfor his collaboration on
                                                                   the nano-G accelerometer and DavidSonnabend, formally cf
                                                                   JPL, for his collaboration in the design of a gradiometer.

       Figure 5 Fabry-Perot Filter Cross Section                   REFERENCES:
                                                                     P.M. Z a m k y , F. Hartley, N. Sherman, T. Hansen, and
Anemometer                                                         K. Warner, “A NewForceBalancedAccelerometer using
The second derivative device under                                 Tunneling Tip Position Sensing,” 7th Int.    Conf.     on
development is           a   sensitive                             Sensors and Actuators, Yokahama,Japan, June 7-10, 1993.
anemometer use           on Mars to                                2
measure  wind speeds of the low
                                                                   Frank T. Hartley, “Development Of A Small, Stable,
pressure CO, atmosphere. Herethe                                   Rugged Microgravity Accelerometer”, August 1996, JPLD.
proof mass of the accelerometer and
its spring flexures are deployed not
as a proof mass but as a large bawe
                                                                   *ak              and
                                                                            T. Hartley        James H. Wise, “Caging,
                                                                   Calibration, Characterization
                                                                                            and         Compensation d
plate that will be presented to the                                          Transducers”,
                                                                   Microstructural                          No.
                                                                                                    Patent Serial
‘wind’ and on which the pitot                                      08/106,448, August 16, 1993.
static force will apply.
     The tipless quad platen is                                    4Ben P. Dolgin,                   and Zavracky,
                                                                                  Frank T. Hartley Paul
again used but with the center 50%                                 “Micromachined Tunable Filters for Optical Applications”,
of each of the quad plates removed.                                May 2 1, 1994, NPO 19456, C IT No. 9060
The anemometer is then fabricated
out of two       M e plates (proof                                 5
                                                                   D.Sonnabend,             “A Tolerant Gradiometer”
mass) and two perforated quad dice                                 OSSI proposal NRA 92-OSSA-6 July, 1992.
eutectically bonded together as Fig Cross Section Of
illustrated in Figure6. Connections Anemometer {NTs)

To top