Acta Mechanica Solida Sinica, Vol. 28, No.

4, August, 2015 ISSN 0894-9166

Published by AMSS Press, Wuhan, China

ANALYTICAL ANALYSIS OF MEMS CAPACITIVE

PRESSURE SENSOR WITH CIRCULAR DIAPHRAGM
UNDER DYNAMIC LOAD USING DIFFERENTIAL
TRANSFORMATION METHOD (DTM)

Marzieh Molla-Alipour1⋆ Bahram Azizollah Ganji2

1
( Department of Engineering, South Tehran Branch, Islamic Azad University, Tehran, Iran)
(2 Department of Electrical Engineering, Babol University of Technology, Babol, Iran)

Received 6 January 2014, revision received 25 June 2014

ABSTRACT In this paper, first a circular diaphragm is modeled using the classical plate theory.
An analytical solution based on differential transformation method (DTM) and Runge-Kutta
method is employed for solving the governing differential equation for the first time. Then the
influences of various parameters on central deflection of the diaphragm, stress distribution and
capacitance of pressure sensor with a time-dependent pressure are examined. Several case studies
are compared with simulations to confirm the proposed method. The analytical results compared
with ABAQUS simulation show excellent agreement with the simulation results. This method is
very promising for time saving in calculating micro-device characteristics.

KEY WORDS MEMS, capacitive pressure sensor, circular diaphragm, deflection, analytical model,
theoretical analysis

I. INTRODUCTION
Pressure sensing is a fundamental measurement in many industries. In the automotive industry,
for example, a typical passenger car has pressure sensors as part of the power-train, brake and airbag
systems, in tire pressure monitoring, and increasingly in safety and ergonomic systems[1, 2] . Capacitive
sensors are interesting products for their ability to sense motion, chemical composition, electric field,
pressure, acceleration, fluid level and fluid composition[3] . Piezoresistive pressure sensors have been
widely studied during the past 20 years as one of the most popular MEMS sensors. But capacitive
pressure sensors provide very high-pressure sensitivity, low power, low noise, large dynamic range and
low temperature sensitivity compared with the piezoresistive pressure sensors, and are preferred in
many emerging high-performance applications[4–9] .
In most cases, these microelectromechanical devices are manufactured from rectangular or circular
diaphragms and typically consist of a plate and a substrate, both conductive, separated by insulating
media (e.g. silicon dioxide and air as shown in Fig.1)[10] . The capacitance of the device increases when
pressure is applied on the plate causing its deflection toward the substrate. These sensors can be designed
to work in normal mode, where the deflection of the plate is smaller than the cavity height or in touch
mode when the plate touches the insulator on the bottom electrode. As the deflection properties of
the moving electrode have the most influence on defining the capacitive behavior of the sensor, it is
⋆ Corresponding author. E-mail: m.mollaalipour@gmail.com
 Vol. 28, No. 4 Marzieh Molla-Alipour et al.: Analysis of MEMS Capacitive Pressure Sensor · 401 ·

important to have a good understanding of the deflection, stress and strain in order to design capacitive
pressure sensors properly. An analysis of the static deflection of plates is documented in the classical
plate bending literature[11] .

Fig. 1. The schematic diagram of the MEMS capacitive pressure sensor with clamped circular plate.

The finite element analysis provides most accurate results despite the long time taken by the data
computing and posting process in particular. Previous work on mathematical models for capacitive
pressure sensors has been presented[12–15] . The problem has been approached with different techniques,
such as assuming a power series solution[12–14] and using the pseudo spectral method[14] , where in both
cases the principle of virtual work is used to find the displacement for large deflection. In this work, we
shall show an extended analytical solution for normal mode capacitive pressure sensors (NMCPS) under
dynamic load in linear elastic regime that is for small deflections. AMEMS pressure sensors are modeled
using the classical plate theory and a. semi-analytical solution based on the differential transformation
method (DTM) and Runge-Kutta method is employed for solution of the governing differential equation
for the first time.
In the following discussion three main topics will be covered. In §II, the governing differential equation
and boundary conditions of axisymmetric circular plate are presented based on the classical plate theory.
In §III the differential transformation method (DTM) and the fourth-order Runge-Kutta method are
employed for solving the governing differential equation in order to obtain the capacitive pressure sensor.
In §IV, the numerical solution and a comparison between the model and the simulation results will be
presented. Finally the paper is ended in §V with conclusions.

II. MATHEMATICAL MODELLING

The capacitive pressure sensor of this study is composed of two circular frames with a cavity. These
sensors typically have one fixed and one flexible conductive plate. The capacitance between the plates
changes as a function of their proximity, which in turn is a function of the pressure[16, 17]. These sensors
can be fabricated using MEMS technology. Figure 1 shows a schematic diagram of the MEMS capacitive
pressure sensor with clamped edges, radius R, thickness h, distance between the two frames d, Young’s
modulus E, and the Poisson’s ratio ν.
Under an applied pressure, the upper deformable frame undergoes a deflection w(r, t) at any point
and produces a change in the capacity of the sensor(see Fig.2).
A rather lengthy calculation, reported in Appendix, leads to the following equilibrium differential
equation of transverse motions for a circular micro-plate plate in the polar coordinate.
 4
2 ∂3 1 ∂2
 2
∂2
 
∂ 1 ∂ ∂ 1 ∂
D 4
+ 3
− 2 2
+ 3
w(r, t) + T 2
+ w(r, t) + ρh 2 w(r, t) = P (ωs , t) (1)
∂r r ∂r r ∂r r ∂r ∂r r ∂r ∂t

where D = Eh3 / 12 1 − ν 2 is the flexural rigidity of the plate, w(r, t) is the deflection of the upper



frame under the applied pressure P , T is the initial tension of the diaphragm per unit length, ρ is
a density of the diaphragm, and h is diaphragm thickness. In this paper, we consider a periodically
 · 402 · ACTA MECHANICA SOLIDA SINICA 2015

Fig. 2. Cross section view of the typical pressure sensor under applied pressure.

applied pressure on the upper plate as follows:

P (ωs , t) = P0 cos(ωs t) (2)
where P0 and ωs are the amplitude and frequency of the pressure wave, respectively.
The boundary conditions of clamped plate are defined as follows:
dw
w(b) = 0, |r=b = 0 (3)
dr
Equation (1) is a fourth-order differential equation, so four boundary conditions are required to obtain
the complete solution. Two of these boundary conditions can be found from the conditions of the outer
edge of the circular plate which are mentioned in Eq.(2). The remaining two conditions can be defined
based on the regularity conditions at the center of the plate. In the case of an axisymmetric deflection
of plate, the regularity conditions at the center (r = 0) of the solid circular plate is expressed as follows:
 3
d w 1 d2 w

dw 1 dw
|r=0 = 0, Vr |r=0 = + − 2 |r=0 = 0 (4)
dr dr3 r dr2 r dr
where Vr is the effective radial shear force per unit length. Also, the corresponding initial conditions
of the plate are:
w(r, 0) = 0, ẇ(r, 0) = 0 (5)
where ẇ is dw/dt. For convenience the following dimensionless parameters are defined to transform
Eq.(1) into non-dimensional form.
r h d
r̄ = , h̄ = , d¯ = (6)
R R R
For simplicity, in the remaining part of the paper the bar (-) symbol will not be shown.

III. PROBLEM SOLVING

In this section, the differential transformation method (DTM) and the fourth-order Runge-Kutta
method are employed for solving the governing differential equation. In the DTM, the governing dif-
ferential equations and the relevant boundary conditions of the system are transformed into a set of
algebraic equations. For solution of the governing differential equations, the unknown functions w(r, t),
ẇ(r, t) and ẅ(r, t) are defined as follows:
∞
X ∞
X ∞
X
w(r, t) = w̄i,j (r − r0 )i , ẇ(r, t) = w̃i,j (r − r0 )i , ẅ(r, t) = ŵi,j (r − r0 )i (7)
i=0 i=0 i=0

where j is the time step counter. Also w̄i,j , w̃i,j and ŵi,j are coefficients of Taylor’s series at each time
step.
 Vol. 28, No. 4 Marzieh Molla-Alipour et al.: Analysis of MEMS Capacitive Pressure Sensor · 403 ·

In practical applications, the functions w̄i,j , w̃i,j and ŵi,j are usually expressed by a finite series.
Therefore, Eq.(7) can be rewritten as:
K
X K
X K
X
w(r, t) = w̄i,j (r − r0 )i , ẇ(r, t) = w̃i,j (r − r0 )i , ẅ(r, t) = ŵi,j (r − r0 )i (8)
i=0 i=0 i=0
∞ ∞ ∞
w̄i,j (r − r0 )i , w̃i,j (r − r0 )i and ŵi,j (r − r0 )i are small and neglected. In the
P P P
whose
i=K+1 i=K+1 i=K+1
present research, the K value is so chosen that the calculated results converge with high accuracy .
First, by substituting Eq.(8) into Eq.(4), the regularity conditions at the center of the circular plate
(r = 0) can be transformed as follows:
w̄1,j = 0, w̄3,j = 0 (9)
By substituting Eq.(8) into the governing equation (1) and using Eq.(9) the transformed form of
Eq.(1) can be obtained as
K
X K
X K
X
D (i + 4)2 (i + 2)2 w̄i+4,j ri + T (i + 2)2 w̄i+2,j ri + ρh ŵi,j ri = P (ωs , t) (10)
i=0 i=0 i=0

By performing some manipulations, Eq.(10) can be rewritten as

P (ωs , t) δ(i) − T (i + 2)2 w̄i+2,j − ρhŵi,j
w̄i+4,j = (11)
D(i + 4)2 (i + 2)2
The boundary conditions of clamped plate for (r = b) are defined as follows:
K K
X dw X
w(b) = w̄i,j bi = 0, |r=b = (i + 1)w̄i+1,j bi = 0 (12)
i=0
dr i=0

Also by using Eq.(8), the initial conditions can be transformed as follows:

w̄i,0 = 0, w̃i,0 = 0 (13)
In the fourth-order Runge-Kutta numerical time integration method, the relations between the accel-
eration and velocity parameters at the end of each time step and parameters of the beginning of the
mentioned time step are:
 
4 1 2
ŵi,j+1 = wi,j+1 − wi,j − ∆tw̃i,j − ∆t ŵi,j (14)
∆t2 4
∆t ∆t
w̃i,j+1 = w̃i,j + ŵi,j + ŵi,j+1 (15)
2 2
Simultaneous solution of the governing equations of motion in terms of the displacement parameters,
Eq.(11), using Eq.(13) gives values of the displacement parameters at the end of the time step considered.
However, in solving the resulting system of equations, the transformed forms of the edge, regularity,
and initial condition Eqs.(16), (19) and (20), have to be satisfied. The complete solving procedure for
the governing equations of motion includes the following steps:
(1) Initial values (j = 0) of the acceleration parameters (ŵi,0 ) are determined based on the Eq.(13)
and the governing equation (11).
(2) Determination of the acceleration parameters at the end of the first (j = 1) step is based on
Eq.(10).
(3) Using Eqs.(11) and (7) for computing the velocity and displacement parameters corresponding
to the end of the time step.
(4) Repeat items (2) and (3) till the final time step is reached.
The capacitance between two parallel conductive plates with air gap shown in Fig.1 can be calculated
as[12]
A
C =ε (16)
d
 · 404 · ACTA MECHANICA SOLIDA SINICA 2015

where, A is the area of the plates, d is the air gap and ε is the dielectric constant. It is assumed that
the circumference of the plates is much bigger than the air gap. Thus, the fringe capacitance at the
edges of the plates can be neglected. But in the clamped circular flat diaphragm the deflection of the
center of plate extends further than the other side (see Fig.2), so the capacitance of pressure sensor
with clamped circular diaphragm is given by[18]
Z Z 
1 1
ZZ
Cf = ε rdrdθ = ε rdrdθ (17)
d − w(r, θ) N
P k
d− Wk,i r
k=0

IV. RESULTS AND DISCUSSION

In order to study the dynamic behavior of capacitive pressure sensors, a clamped circular micro-plate
is considered with the geometrical and material properties as listed in Table 1.

Table 1. Material and geometrical parameters of the micro-plate

Young’s modulus, E 160 GPa

Poisson’s ratio, ν 0.22
Dielectric constant, ε0 8.85×10−12 F/m
Pressure range 10-20 (Pa)

Deflection of a plate with boundary conditions stated in Fig.1 can be analytically calculated according
to Equations in §II. Figures 3-5 show a comparison of the device characteristics from the analytical
solution and ABAQUS simulation. The values of non-dimensional parameters h̄ and d¯ are 0.05 and
0.02.

Fig. 3. Central deflection of diaphragm under two dynamic loads versus time.

Figure 3 shows a comparison of analytical and simulation value of diaphragm center deflection under
two dynamic pressure loads with ωs = 500 and ωs = 1000 versus time. The simulation and calculation
results show that as the pressure decreases, the diaphragm center deflection decreases and as the pressure
increases, the diaphragm center deflection also increases. The theoretical results are well matched by
the simulation results.
Figure 4 shows a comparison between analytical and simulation results for central deflection of
diaphragm under two dynamic pressure load versus radius at four different times. The simulation
results show close agreement with the analytical values.
Figure 5 shows a comparison of analytical and simulation results for stress distribution of diaphragm
under two dynamic pressure loads versus radius at four different times. The results show that as the
diaphragm radius, r, increases, the stress distribution of diaphragm increases under the same load. The
simulation shows good agreement with the theoretical prediction.
Figures 6 and 7 show the effect of air gap change on capacitance of sensor under two dynamic
pressure loads versus time. It can be seen that the capacitance will be changed if the pressure has been
 Vol. 28, No. 4 Marzieh Molla-Alipour et al.: Analysis of MEMS Capacitive Pressure Sensor · 405 ·

Fig. 4. Top plate deflection versus radius for different times.

Fig. 5. Stress distribution of diaphragm versus radius for different times.

Fig. 6. Theoretical value of capacitance with two dynamic loads versus time for ωs = 500.

changed versus time. It is seen that the capacitance of the device increases when the applied pressure
is increased. Also it can be seen that the capacitance will be decreased with an increase of the air gap.
The sensitivity of sensor can be calculated using Eq.(18):

∂C
S= (18)
∂P

As shown in Figs.6 and 7, the sensitivity of the sensor when capacitance has maximum value is as
presented in Table 2.
It can be concluded that the sensitivity is increased with an increase of ωs and decrease of d.
 · 406 · ACTA MECHANICA SOLIDA SINICA 2015

Fig. 7. Theoretical value of capacitance with two dynamic loads versus time for ωs = 1000.

Table 2. The sensor sensitivity

ωs = 500 ωs = 1000
d = 0.01 S = 1.2f F (Pa) S = 3.9f F (Pa)
d = 0.02 S = 0.3f F (Pa) S = 0.9f F (Pa)

V. CONCLUSION
In this paper, a MEMS capacitive pressure sensor with clamped circular diaphragm was studied.
An analytical solution based on differential transformation method (DTM) and Runge-Kutta method
are employed for solution of the governing differential equation for the first time. This solution is a
promising method for quick estimation of diaphragm deflection, stress distribution and capacitance of
diaphragm with a time-dependent pressure. It can give us accurate results of device characteristics in
good agreement with simulation results. Also, this method, which implements the power series solution,
is very convenient being time-saving compared with others.

References
[1] Schmidt,R., Elektronic Praxis 20, 2000, 62-64.
[2] Schuster,J.P., Automotive silicon based pressure transducers for fluid power applications. In: Proceeding
of the National Conference on Fluid Power, Annual Meeting, 1986, 1921-1929.
[3] Marco,S., Samitier,J., Ruiz,O., Morante,J.R. and Esteve,J., High-performance piezoresistive pressure sen-
sors for biomedical applications using very thin structures membranes. Measurement Science and Technol-
ogy, 1996, 7: 1195-1203.
[4] Lee,Y.S. and Wise,K.D., A batch-fabricated silicon capacitive pressure transducer with low temperature
sensitivity. IEEE Trans. Electron Dev., 1982, 29: 42-48.
[5] Ko,W.H. and Wang,Q., Touch mode capacitive pressure sensors, Sens. Actuators, 1999, 75: 242-251.
[6] Suzuki,K., Suwazono,S. and Ishihara,T., CMOS integrated silicon pressure sensor. IEEE J. Solid-State
Circuits SSC, 1987, 22: 151-156.
[7] Clark,S.K. and Wise,K.D., Pressure sensitivity in anisotropically etched thin-diaphragm pressure sensors.
IEEE Trans. Electron Devices 1979, 26: 1887-1896.
[8] Sander,C.S., Knutti,J.W. and Meindl,J.D., A monolithic capacitive pressure sensor with pulse-period out-
put. IEEE Trans. Electron Devices, 1980, 17: 927-930.
[9] Ko,W.H., Solid-state capacitive pressure transducers. Sens. Actuators, 1986, 10: 303-320.
[10] Chavan,A.V. and Wise,K.D., A batch-processed vacuum-sealed capacitive pressure sensor. IEEE Trans-
ducers, (1997), 1449-1452.
[11] Timoshenko,S. and Woinowsky-Krieger,S., Theory of Plates and Shells. McGraw-Hill Book Company Inc.,
1959.
[12] Ko,W.H. and Wang,Q., Touch mode capacitive pressure sensor. Sensors and Actuators, 1999, 75: 242-251.
[13] Meng,G. and Ko,W.H., Modeling of circular diaphragm and spreadsheet solution programming for touch
mode capacitive sensors. Sensors and Actuators, 1999, 75: 45-52.
[14] Daigle,M., Corcos,J. and Wu,K., An analytical solution to circular touch mode capacitor. IEEE Sensors
Journal, 2007, 7: 502-505.
[15] Wang,X., Mingxuan,L. and Wang,C., Numerical analysis of capacitive pressure micro-sensors. Science in
China Series E: Engineering & Materials Science, 2005, 48: 202-213.
 Vol. 28, No. 4 Marzieh Molla-Alipour et al.: Analysis of MEMS Capacitive Pressure Sensor · 407 ·

[16] Bakhoum,E.G. and Cheng,M.H.M., Capacitive pressure sensor with very large dynamic range. IEEE Trans-
actions on Components and Packaging Technologies, 2010, 33: 79-83.
[17] Pedersen,T., Fragiacomo,G., Hansen,O. and E.V., Thomsen: Highly sensitive micromachined capacitive
pressure sensor with reduced hysteresis and low parasitic capacitance. Sensors and Actuators A: Physical,
2009, 154(1): 35-41.
[18] Blasquez,G., Naciri,Y., Blondel,P., Ben Moussa,N. and Pons,P., Static response of miniature capacitive
pressure sensors with square or rectangular silicon diaphragm. Rev. Phys. Appl., 1987, 22: 505-510.

APPENDIX: THE GOVERNING EQUATION OF

THE AXISYMMETRIC CIRCULAR PLATE
Usually, employing the classical plate theory leads to sufficiently accurate results for the thin plates.
According to the classical plate theory, the displacement field in the radial direction of the axisymmetric
circular plate can be expressed as follows:
∂w
u = −z (19)
∂r
where w is the transverse deflection. For small deflections, the strain-displacement relations can be
written as
∂u u
εr = , εθ = (20)
∂r r
The order of the transverse normal strain is usually ignorable in comparison with the order of the
in-plane strains. So Hooke’s stress-strain law can be expressed as follows:
E E
σr = (εr + νεθ ) , σθ = (εθ + νεr ) (21)
1 − ν2 1 − ν2
where E and ν are the Young’s modulus and Poisson’s ratio, respectively. Based on Eqs.(19)-(21), the
stress-deflection relations will be:
 2
∂ 2w
  
Ez ∂ w 1 ∂w E 1 ∂w
σr = − + ν , σθ = − + ν (22)
1 − ν 2 ∂r2 r ∂r 1 − ν 2 r ∂r ∂r2
The governing equations of the circular plate can be derived by either using the minimum total potential
energy principle or Hamilton’s principle. Based on the first principle, we can write:
δΠ = δU + δK − δW = 0 (23)
where δU , δK and δW are increments of the strain energy, kinetic energy (energy of the inertial loads),
and work of the externally applied loads, respectively:
Z
δK = ρ (ẅδw) dV
ZV
δU = (σr δεr + σθ δεθ )dV (24)
ZV  
∂w ∂w
δW = pδw + T δ dA
A ∂r ∂r
where ρ and T are the mass density and tensile force per unit length, respectively. P (ωs , t) is the
harmonic load and is defined as
P (ωs , t) = P0 sin(ωs t) (25)
By substituting Eqs.(24) into Eq.(23) and using Eqs.(19), (20) and (22), Eq.(23) can be rewritten as
follows:
Z   2     Z   
∂ w zσθ ∂w ∂w ∂w
δΠ = −zσr δ − δ + ρ ẅδw dV − pδw + T δ dA = 0
V ∂r2 r ∂r A ∂r ∂r
Or Z   2      (26)
∂ w Mθ ∂w ∂w ∂w
δΠ = −Mr δ − δ + I0 ẅδwqδw + T δ dV = 0
A ∂r2 r ∂r ∂r ∂r
 · 408 · ACTA MECHANICA SOLIDA SINICA 2015

where the stress resultants (Mr and Mθ ) and inertia ( I0 ) are defined as
Z h/2 Z h/2  2 
E ∂ w 1 ∂w
Mr = zσr dz = 2
(εr + vεθ )dz = −D +ν
−h/2 1 − ν ∂r2 r ∂r
Z−h/2
h/2 h/2
∂2w
 
E 1 ∂w
Z
Mθ = zσθ dz = 2
(εθ + νεr ) dz = −D +ν 2
−h/2 −h/2 1 − ν r ∂r ∂r
Z h/2 (27)
I0 = ρdz = ρh
−h/2
Z h/2
E Eh3
D= 2
z 2 dz =
−h/2 1 − ν 12(1 − ν 2 )

where D is the flexural rigidity of the plate. By substituting Eqs.(27) into Eq.(26) and performing some
manipulations, the motion equation of the plate in polar coordinate will be written as
 4
2 ∂3 1 ∂2
 2
∂2
 
∂ 1 ∂ ∂ 1 ∂
D + − + w(r, t) + T + w(r, t) + I0 w(r, t) = P (ωs , t) (28)
∂r4 r ∂r3 r2 ∂r2 r3 ∂r ∂r2 r ∂r ∂t2