Combination of Different Methods to Characterize Micromechanical Sensors
- PDF / 1,158,277 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 2 Downloads / 189 Views
t. Res. Soc. Symp. Proc. Vol. 459 01997 Materials Research Society
piezoresistors
a) Setup
mechanical stress
pressuresfulloou
u
passivation layers
b) Transfer function ful,
j/ su
scale (FS)
pre
silicon diaph
ideal ra
__ real
silicon bulk
0
max. pressure
>-P
c) Operation principle MehaiclChange
of el.
Stress
cnductivity
Mechanical
Mechano-Electrical
Electrical
Pressure
Subsystem
Subsystem
Subsystem
Output
p
Diaphragm
Piezoresistors
Wheatstone-bridge
Voltage Vout
Fig. 1: Setup, transfer function and operation principle of piezoresistive pressure sensors APPROACH The performance of piezoresistive pressure sensors is limited by several physical effects [1], e. g.: mechanical effects: - geometric nonlinearities of the mechanical subsystem, - thermo-mechanical stress, - relaxation and recrystallization of passivation or contact layers electricaleffects. - space charge induced resistance modulation (SCIRM), - formation of channels by positive (negative) surface charges due to depletion (enhancement), - parasitic currents due to adsorbed and absorbed charges or due to insufficient isolation, - "aging" due to carrier-induced electrochemical reactions. Due to the complexity of the sensor setup and manufacturing it is hardly possible to assign instabilities to the related physical effects as described above. Finite-Element-(FE-)
Capacitance-Voltage-(CV-)
Simulation
Measurements
4 A.k
Determination of Metrological Behavior Sensor Output otg
,,
PressureP
Mechanical
SSubsystem
Piezoresistors Optical and
of Determination Thermo-mechanical Material Properties
Michelson Interferometry
Geometrical Inspection
Fig. 2: Survey on investigation methods
244
Electrical Subsystemo
Vut o
Measurement of metrological behavior (output voltage) phenomenologically records the totality of relevant instability mechanisms. Considering the actual boundary conditions (e. g. stimulation regime, material properties or manufacturing data) this totality can be separated by signal analysis. This separation allows the determination of dependences of sensor performance from design and technological parameters. The combination of this knowledge with results from other methods (Fig. 2) represents the basis for the correct interpretation of ongoing physical mechanisms. As mentioned above the major part of the investigation is the determination and analysis of metrological sensor behavior consisting of the following steps: - stimulation with multidimensional regimes (pressure, temperature, relative humidity), - regression of the multidimensional transfer function, - heuristic modeling of separated transfer channels, - evaluation of determinate dynamic (L) and static (VJ)transfer coefficients, - evaluation of channel noise V0t (stimulation-induced) and Vi,, (intrinsic noise) as deviation of measured data from transfer function. a) Multidimensional stimulation
b) Heuristic signal model
Dynamic Sbytm
VstT
out=f(p,T)
Static Subsystem stochastic
systeimatic
t p.
.. ...... . ........
P
PO
.. in
LH~)V
.a
Data Loading...
