FEM Simulation of CMUT Cell for NDT Application

Aditi, R. Mukhiya, Ram Gopal, V. K. Khanna

MEMS and Microsensors, CSIR- Central Electronics Engineering Research Institute (CEERI),
Pilani-333031, Rajasthan, India.

ABSTRACT

This paper presents the simulation of various electromechanical performance parameters for an efficient and broadband
air coupled Capacitive Micromachined Ultrasonic Transducer (CMUT) for Non Destructive Testing and Evaluation
(NDT/NDE) using FEM simulation tool, Coventor Ware. The non contact ultrasonic inspection performed in ambient air
reduces the complexity and cost. Various critical parameters like collapse voltage, resonant frequency, coupling
coefficient, squeeze film damping, bandwidth, quality factor and transient response of the single cell of CMUT are
discussed. We employ a hexagonal CMUT cell for the modeling to improve the CMUT’s transmission/reception
efficiency as the average membrane displacement is high.
Keywords: CMUT, collapse voltage, NDT/NDE, FEM

1. INTRODUCTION

Recently CMUT based on electrostatic actuation has taken over traditional piezoelectric transducers for NDT/NDE
application for various aspects like enhanced sensitivity, broad bandwidth, ease of batch fabrication and have better
acoustic impedance match with air [1]. Ultrasonic transducers are used to generate short pulses which propagate in a
solid or liquid medium. Flaws or discontinuities in the medium cause reflected pulses which are then detected. In
general, a wide bandwidth around the pulse center frequency is desirable in order to distinguish closely spaced
reflections [2]. Figure 1 shows the schematic of a single cell CMUT.

CMUT’s have been fabricated using Surface micromachining and wafer bonding techniques. The surface
micromachining process introduces limitations on the cavity and membrane size of a CMUT. In the wafer-bonding
technique, the membrane and the cavity are defined on separate wafers that are bonded. Active research on CMUTs has
been reported [3, 4, and 5]. The results demonstrate that CMUTs optimized with respect to such design parameters as
device size, membrane radius, thickness, shape, gap height, and region of operation can perform comparably to
piezoelectric transducers in terms of bandwidth, frequency range and dynamic range [6]. Simulations for the different
shapes have been done and hexagonal membrane is taken in this paper, then the dimensions were optimized for the NDT
application.

Pol silicon Membrane

Silicon Dioxide

Silicon Nitride

Silicon

Figure 1. Schematic of a single cell of CMUT

16th International Workshop on Physics of Semiconductor Devices, edited by Y. N. Mohapatra, B. Mazhari,

M. Katiyar, Proc. of SPIE Vol. 8549, 85491H · © 2012 SPIE · CCC code: 0277-786/12/$18 · doi: 10.1117/12.926046

Proc. of SPIE Vol. 8549 85491H-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 In order to generate acoustic signal a DC bias is applied between the membrane and an AC signal is superimposed. In the
receive mode, only DC bias is applied and incident acoustic pressure generates an AC detection current which is the
output signal. A single cell hexagonal CMUT is designed using CoventorWare. Various electromechanical parameters
which influence the operation of CMUTs are calculated. The cell acts as a parallel plate capacitor with doped polysilicon
as top membrane and doped silicon as bottom electrode.

2. SIMULATION DETAILS

Various modules of CoventorWare are used for the FEM simulation. MemMech module is used for the calculation of
resonant frequency and transient analysis. A modal analysis calculation computes the natural resonant frequency of a
mechanical structure at equilibrium. CoSolve module is used for the collapse voltage and coupling coefficient
calculation. A pull-in analysis is an electromechanical analysis that is conducted in CoSolveEM. A voltage trajectory is
specified, and the displacement is calculated for each voltage up to the pull-in voltage, which is the greatest voltage for
which an equilibrium position can be found. At each point of this trajectory the mechanical solver MemMech and the
electrostatic solver MemElectro are employed. HarmonicEM module is used for the frequency response curve to
calculate the bandwidth and quality factor. DampingMM module is used for the squeeze film damping coefficient
calculation. The top membrane is clamped at the edges. The large gap size is decided to allow high transmission and
reception efficiencies at higher operating voltages and also on the basis of the fabrication limitations. The 3D solid mesh
modal can be seen in (fig. 2). We employ a hexagonal CMUT cell for the modeling to improve the CMUT’s
transmission/reception efficiency as the average membrane displacement is high compared to other membrane shapes
like, square and rectangular.

Figure 2. 3D Solid mesh model

3. RESULTS & DISCUSSIONS

For NDT application one of the important factors to decide the resolution of the device is the resonant frequency. Based
on the dimensional parameters listed in the Table.1 the resonant frequency of the cell comes to be 2.47 MHz. The
CMUTs were designed to operate at 1MHz center frequency with 2.5µm gap between the membrane and substrate
[7].The structure vibrates in mode shapes when excited at its resonant frequencies. Figure 3 shows the modal frequencies
of the first three modes of vibration. The first mode is along z-axis direction. Second mode is across x-axis and third
mode is across y-axis. Higher resonant frequency of second and third modes signifies that the cross-axis sensitivities are
lower than the prime axis sensitivity for the structure and make it out of plane single-degree-of-freedom (DOF).

Proc. of SPIE Vol. 8549 85491H-2

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 Table 1. Device specification

Parameters Value

Shape of the cell Hexagonal

Membrane thickness 1 µm
Gap height (Silicon dioxide) 1 µm
Silicon nitride (Passivation) 1500 Å
Edge length of the membrane 46 µm

kx

Modal Displacement mag.:

um
0.0E+00 2.5E-01 5.0E-01

247391E+006Hz
it
7.5E- 01 1.0E+00

COVENTOR
Modal rnsaiecemem may.: 0.0E +00 2.5E-01

Figure 3. Modal Frequencies

50E-01

5.11486E+006 Hz
1t
7.5E-01 1.0E+00

COVENTOR
IhAe
maI DISPICMIeM Ma, .: 0.0E +00
111!
2.5E -01 50E -01

5.13473E+006 Hz
t 7.5E -01 1.0E+00

COVENTOR

When a DC bias is applied on the membrane, the electrostatic force pulls the membrane down and the restoring force
balances it. The voltage at which the electrostatic force exceeds the restoring force, the membrane collapses; that is
known as the collapse voltage of the membrane. The collapse voltage calculated is 133.7 V as shown in (fig.4).

1.0

08 ---0 --Logic
g0
aaá

a
00

09
_
i
t4IIU
0011
. :
....- :

00
MEL1211
o 20 40 00 m ' ,00 t20 11p tOp

VoRege(V)

Figure 4. Displacement Vs voltage

Proc. of SPIE Vol. 8549 85491H-3

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 s

Figure 5. Frequency Vs voltage

S 2 Q 2 R
For operation at the first harmonic, Nthe DC rvbias04 voltage
04 04 is rv
made larger than the time varying voltage [6]. Usually the DC
bias is 80% of the collapse voltage and the time varying voltage is 10% of the DC bias. So the applied DC bias on the
(ZHwo Á'J1p116Q1:

membrane for device operation is 107 V and the AC is 10 V. When a DC bias is applied across the membrane, the
resonant frequency shifts as compared to that of without DC bias. The shift in the frequency with the applied bias can be
seen in (fig.5). The softened frequency at 107 V comes to be 2.15 MHz.

When the CMUT operates with the electrical input, the coupling coefficient relates amount of mechanical energy
delivered to the load to the total energy stored in the device. In this paper, we have used Berlincourt's [7] approach to
calculate the coupling coefficient that relies on the use of the fixed (CS) and free (CT) capacitance of the transducer. The
fixed capacitance is defined as the total capacitance of the transducer at a given DC bias:

Q(x)
CS = x DC , VDC
V
The free capacitance is defined as the slope of the charge voltage curve:

dQ(x)
CT = x DC , VDC
dV
CS
and the coupling coefficient is given by: k T2 = 1 −
CT

The coupling coefficient calculated from (fig.6a, b) comes out to be 0.13.

Proc. of SPIE Vol. 8549 85491H-4

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 0 05

0.052

0.062
Mel"
ummoraim
6D

VaMape
10

N)
100 120 110

Figure 6. (a) Capacitance Vs voltage (b) Charge Vs voltage

As the medium is air, the air film acts as a damper and this damping is called squeeze-film damping. The damping force
is dependent on gap height; smaller the gap results into larger damping force. When the gap is large, the pressure build
up is negligible and at higher frequencies spring force is dominant over damping force as shown in (fig.7a). The air
captured in the cavity is squeezed. At low frequencies, air can escape with little resistance so force is low. At high
frequencies, the air is held due to its inertia. So, the time is not enough for air to move out. Therefore the air compresses,
resulting in spring force. Since the damping force is caused by the viscous stresses, if the gas gets compressed and does
not move much, the damping force will be lower. This accounts for the reason why damping force is smaller when
frequency increases. The damping coefficient (c) calculated comes out to be 1.96 N/ (m/s), as shown in (fig.7b).

200 --
-- -.- 0.0001

_ . Damp S Fora !NzlÍ

o _ r Spot Fa (ti YmÙ 6E-06

160
á E

ÌI
LL
OI z
t 6E-06

100

ó
LL
ME O 4E-06

äE
d 2EO6
i i

i.--4
i

:....:....:...;

o lt
`
0 1E+06 2E+06 3E+06 1E+06 0 1E+06 2E+06 3E+06 4E+06

Frequency (Hz) Frequency(Hz)

Figure 7. (a) Damping and spring forces Vs frequency (b) Damping coefficient Vs frequency

The damping ratio is then calculated using formula:

c c ,
ξ= =
c 0 2mω 0

where, ξ is the damping ratio, c0 is the critical damping and ω0 is the free vibration frequency of the system. The
damping ratio comes as 0.18. By using the damping ratio harmonic analysis is done to calculate the bandwidth and the
quality factor of the device. (fig. 8a) shows the frequency response when squeeze film damping is not considered. The

Proc. of SPIE Vol. 8549 85491H-5

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 applied bias is 107 V DC and 10 V AC. (fig. 8b) gives the frequency response when the damping ratio is 0.18, due to the
squeeze film effect. The bandwidth of the transducer comes to be 1MHz and the quality factor 2.15. The maximum
displacement of the membrane is 0.126 µm.

0.25 0.14

0.12
02

E
-s_0.15 _.... ....

E E 0.08
N N
U U
ao
W
0.1
l0
0_
N
Ö Ö 0.06

0.05
0.04

.
pp

1E+06 2E+06 3E+06 4E+06 5E+06 1E +06 2E +06 3E +06 4E +06

Frequency (Hz) Frequency (Hz)

Figure 8. (a) Displacement Vs frequency when ξ = 0.1 (b) Displacement Vs frequency when ξ = 0.18

VI

Figure 9. Transient response of the CMUT cell

O
N
Transient analysis is used to predict the settling time of aO CMUT. The membrane is loaded with two components, one is
static DC bias and other is short duration transient pulse. The value of damping is required for the transducer, which is
(Lull) luaweaeldsb
taken from the squeeze-film damping. (fig.9) shows the transient response of the transducer when a DC bias of 80 V and
a pulse of amplitude 5 V for duration of 50 ns is applied. The pulse length generated by the device is small which gives
higher resolution. Lack of transducer ringing allows high-frequency imaging at distances close to the transducer.

4. CONCLUSIONS

An air coupled CMUT cell design was accomplished which has significant implications in non-destructive evaluation. A
broadband transducer is achieved which is suitable for detecting cracks in the near field region. Hexagonal membrane
was used as the average membrane displacement is high compared to other membrane shapes like, square, rectangular
and fabrication is easy than the circular one, which has slightly better response. Our approach intents at increasing the

Proc. of SPIE Vol. 8549 85491H-6

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms

 transduction efficiency of the transducer, with 1MHz bandwidth and 1µm gap, are comparable with the other CMUT’s
for non-destructive testing. Further array design optimization of the CMUT is to be done to get high power.

5. ACKNOWLEDGEMENT

The authors would like to thank the director, CEERI, for the financial and motivational support and the IC Design group,
CEERI, for the software access.

6. REFERENCES

[1] Ahmad, B.; Pratap, R , "Elasto-Electrostatic Analysis of Circular Microplates Used in Capacitive Micromachined
Ultrasonic Transducers," Sensors Journal, IEEE , vol.10, no.11, pp.1767-1773, Nov.(2010).
[2] Greve, D.W.; Oppenheim, I.J, "Coupling of MEMS ultrasonic transducers," Sensors, 2003. Proceedings of IEEE ,
vol.2, no., pp. 814- 819 Vol.2, 22-24 Oct. (2003).
[3] Oralkan, O.; Ergun, A.S.; Johnson, J.A.; Karaman, M.; Demirci, U.; Kaviani, K.; Lee, T.H.; Khuri-Yakub, B.T.; ,
"Capacitive micromachined ultrasonic transducers: next-generation arrays for acoustic imaging?,"Ultrasonics,
Ferroelectrics and Frequency Control, IEEE Transactions on , vol.49, no.11, pp.1596-1610, Nov. (2002)
[4] Y. Huang; “Fabricating capacitive micromachined ultrasonic transducers with wafer bonding technology,” J. Micro-
electromech. Syst.,12, , p. 128- 137,( 2003)
[5] Baris Bayram; “Diamond-based capacitive micromachined ultrasonic transducers” Elsevier (2011)
[6] Haller, M.I.; Khuri-Yakub, B.T., "A surface micromachined electrostatic ultrasonic air transducer," Ultrasonics,
Ferroelectrics and Frequency Control, IEEE Transactions on , vol.43, no.1, pp.1-6, Jan (1996).
[7] Xuefeng Wang; Ying Fan; Wei-Cheng Tian; Hyon-Jin Kwon; Kennerly, S.; Claydon, G.; May, A.; , "An Air-
IEEE ,Coupled Capacitive Micromachined Ultrasound Transducer for Noncontact Nondestructive
Evaluation," Sensors, vol., no., pp.1464-1467, 28-31 Oct. 2007
[8] Yaralioglu, G.G.; Ergun, A.S.; Bayram, B.; Haeggstrom, E.; Khuri-Yakub, B.T., "Calculation and measurement of
electromechanical coupling coefficient of capacitive micromachined ultrasonic transducers," Ultrasonics,
Ferroelectrics and Frequency Control, IEEE Transactions on , vol.50, no.4, pp.449-456, April (2003).

Proc. of SPIE Vol. 8549 85491H-7

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 05/18/2015 Terms of Use: http://spiedl.org/terms