Zuo-Guang. Ye Advanced Dielectric Piezoelectric and Ferroelectric Materials: Synthesis, Characterisation and Applications

Подождите немного. Документ загружается.

Handbook of dielectric, piezoelectric and ferroelectric materials740

Usually, the PZT film, especially the thin sol–gel film, tends to crack at

the edge of the bottom electrode. Electric short-circuit occurs when the

patterned top electrode interconnection passes across the edge of the bottom

electrode although this part is often designed to be covered by PZT film.

Therefore, making a good interconnection between the top electrode and its

electric contact pad is one of the challenges in pMUT fabrication. In our

design, a polyimide layer is utilized to cover the top of the PZT film at the

edge part to overcome this problem. Figure 24.14 shows that the smooth

surface and well-curved slope covered at the edge of PZT layer facilitate the

interconnection of the top electrode. In addition, the very dense polyimide

layer prevents the electric short-circuit at the edge part of the bottom electrode.

24.5.3 Etching of the PZT thick film

In general, there are two approaches to micromachine the PZT film: dry

etching and wet etching. Reactive ion etching (RIE), inductively coupled

plasma (ICP) and electron cyclotron resonance (ECR) excitation, have been

reported to etch PZT films. These studies are driven primarily by ferroelectric

memory applications, where the film thickness generally does not exceed

250nm. However, the dry PZT etching technique is not well developed

because there is no common halogenous gas that forms a volatile compound

with all three elements (Pb, Zr, Ti) to guarantee the residue-free removal of

the film. The volatility of the reactive products of individual Pb, Zr and Ti is

different. Heating the substrate or increasing the ion bombardment is often

Polyimide

Active area

Diaphragm

PZT

24.13

Top profile image of a fabricated pMUT diaphragm. Backside

silicon is wet etched by the KOH. The PZT film is 5µm in thickness

and the total thickness of the support layer is 13µm. The effective

top electrode area of the element is defined by an opened window

on polyimide layer.

WPNL2204

Piezoelectric thick films for MEMS application 741

a way to obtain a uniform etch of PZT layer. It is very difficult to achieve a

high etch rate of the PZT with adequate etch selectivity to electrode materials

and photoresist mask. As a result, the dry etching technique is too slow to

etch thick composite PZT films with thickness up to 20µm. The wet etching

of PZT films remains the only good choice for thick PZT etching in our

pMUT fabrication.

Since the wet etching process is isotropic, lateral etching cannot be avoided.

We define the active pMUT element with patterned bottom electrode together

with an open window through polyimide layer to avoid the etching of small

lateral PZT patterns. Large ‘V’s are etched on PZT films for the access to

bottom electrode bonding pads. Since this bonding pads area is far away

Polyimide

PZT

Ti/Pt

Polyimide

PZT

(a)

(b)

24.14

Polyimide layers on the fabricated pMUT facilitate the electric

interconnection of the top electrode and reduce the parasitic

capacitance. (a) Isolation layer for defining the effective top electrode

area. (b) Polyimide protection layer at the device edge.

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials742

from the active part of the pMUT element, undercutting does not affect the

operation of the pMUT.

In our work, 50ml concentrated HCl (37%), 50 ml water and 0.1ml HF

(48%) were mixed as the etch solution. A 5 µm-thick AZ9260 photoresist

was patterned as an etching mask. The etching process was conducted before

the final annealing of the composite PZT film because the etching is relatively

faster if the PZT film is not fully crystallized. The etch rate at room temperature

was about 0.8µm/min. The whitish metal-fluoride residues formed on the

etch area during PZT film etching were rinsed off by de-ionized (DI) water

after every minute of etching. The residue-free etching of PZT film can be

achieved in few etch–rinse cycles. Figure 24.15 shows an optical image of a

wet etched composite film covered by polyimide layer. It can be seen that the

etching edge is neat and straight at this magnification. However, under etching

as large as 20µm and some profile irregularity can be observed, which

means small etching patterns less than 40 µm are impossible by using wet

etching. Fortunately, for our pMUT design, the achieved wet etching results

are good enough at the device edge which is far away from the pMUT

diaphragm.

24.6 Performances of thick film pMUT

24.6.1 Piezoelectric response and simulated properties of

the fabricated thick film

As a transducer, the pMUT is supposed to work at around its resonance

frequency. Thus the dynamic piezoelectric response of the fabricated pMUT

was investigated. The resonance frequency of a multilayer PZT/silicon

diaphragm is basically determined by the dimension and the mechanical

Polyimide/LTO

Polyimide/Pt/LTO

LTO

Pt/LTO

Pt/PZT

Pt/Polyimide/PZT

Polyimide/PZT

24.15

Optical image of the edge part of a fabricated pMUT element.

The wet etched composite film is covered by a transparent polyimide

layer. The etching edge of PZT is clear and straight. The area marked

with Pt/PZT is the active part of a 250 × 250µm

pMUT element.

WPNL2204

Piezoelectric thick films for MEMS application 743

properties of the PZT and silicon. As long as the fundamental resonance

frequency of the PZT/silicon diaphragm is measured, the mechanical properties

of the thick PZT film can be extracted based on the ANSYS modal analysis

method.

A fabricated 2 × 2mm

pMUT diaphragm was measured and used for

FEA. The measured thicknesses of the PZT film and support layer are 5 µm

and 13µm, respectively, as shown in Fig. 24.13. An external ac electric field

with sweeping frequency was used to excite the vibration of the pMUT and

the vibration velocity at the central point was recorded as the function of the

frequency. Figure 24.16 shows the measured velocity response by using a

Polytec PSV300 scanning laser vibrometer. Only two vibration modes at

41.6 kHz (fundamental mode) and 146.8kHz (sixth mode) were detected.

Assuming the density and Poisson’s ratio of the PZT film are 7500kg/m

and

0.25, respectively, the Young’s modulus of the film was determined as 75.2 GPa

when the simulated fundamental resonance frequency equals 41.6kHz. The

bulk silicon material properties used for the simulation and the estimated

mechanical properties of the thick composite film are listed in Table 24.3. In

the FEA simulation, the influence of the thin electrode and SiO

layers was

ignored.

After determining the mechanical properties of the composite PZT film,

we further estimated the piezoelectric stress coefficient of the film based on

its quasi-static piezoelectric response. Figure 24.17 shows the quasi-static

piezoelectric displacement at the center of the same pMUT element driven

by a sine wave voltage of 500 Hz, which is much lower than its resonance

frequency. The bias dc voltage is 20 V and the peak-to-peak driving voltage

Velocity (a.u.)

41.6kHz

(1,1)

146.8 kHz

(1,3)

25 50 75 100 125 150 175

Frequency (kHz)

24.16

Frequency response of vibration velocity at central point of a

square 2mm × 2 mm diaphragm driven by 5 µm-thick PZT film. About

13µm-thick support layer was left after DRIE.

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials744

is 30V, resulting in a peak-to-peak center displacement of the PZT/silicon

diaphragm at about 200nm, corresponding to a quasi-static piezoelectric

center displacement/voltage sensitivity of 6.67 nm/V. This deflection/voltage

sensitivity was then used to estimate the piezoelectric stress coefficients of

the composite PZT films. The other materials properties determined by model

analysis shown in Table 24.3 remained unchanged during simulation. The

obtained piezoelectric coefficients e

and e

of the composite PZT film

were –14.8 and 18.8 C/m

, respectively, quite close to those of the bulk PZT

ceramics. The good piezoelectric properties are attributable to the applied dc

bias voltage.

Figure 24.18 shows the displacement magnitude/phase-frequency properties

of the same pMUT around the fundamental resonance frequency. During the

measurement, a dc bias of 20 V was applied to ensure the film has sufficient

polarization. A 30V peak-to-peak ac voltage was superimposed on the dc

bias as a driving voltage. It is seen that in the vicinity of the resonance

Table 24.3

Estimated mechanical properties of PZT film from mode

analysis and material properties of silicon used in simulation

Materials Young’s modulus Poisson’s ratio Density

(GPa) (kg/m

)

PZT 75.2 0.25 7500

Silicon 150 0.17 2330

2×2mm

, 20V

Bias

, 30V

Driving volttage (V)

Displacement (nm)

150

100

–50

–100

–150

01 23 456 7

Time (ms)

24.17

Quasi-static piezoelectric center displacement of the 2 × 2mm

diaphragm measured at 500Hz for estimation of piezoelectric

properties of the thick PZT film.

WPNL2204

Piezoelectric thick films for MEMS application 745

frequency, the deflection amplitude of the central point of the diaphragm

reaches about 2.8 µm, and the phase shifts about 180°. The fundamental

vibration mode is the interested mode which can be utilized as both ultrasound

transmitters and receivers.

24.6.2 Output sound pressure level

23,24

The performance of the pMUT as an ultrasound transmitter can be evaluated

by output sound pressure level (SPL), which is expressed by

SPL = 20 log

rms

ref









24.1

where P

rms

is the output sound pressure at the measuring point which should

be far away from the transmitter, P

ref

is reference sound pressure and equals

2 × 10

–5

Pa.

The continuous ultrasound wave emitted by the transmitter is measured in

air with Brüel and Kjær 2825 test system, including a 7521 signal analyzer

interface module, 3016 input module, and 2670 microphone. The highest

frequency of the system can reach 100kHz. The output SPL of the same

transmitter as shown in Fig. 24.18 is measured at the same driving conditions.

Figure 24.19(a) shows the test result. The distance between the transmitter

and the microphone is 12 mm. In comparison with Fig. 24.18, Fig. 24.19(a)

indicates that the output SPL of a pMUT also reaches its maximum at the

resonance frequency. The output SPL is 107.8dB at 41.3 kHz. Among the

20 25 30 35 40 45 50

Frequency (kHz)

2 × 2mm

Bias 20V, 30V

8 nm/mV 41.53 kHz

42.15 kHz

Magnitude (mV)

400

300

200

100

Phase (degree)

180

120

–60

–120

24.18

Frequency spectra of magnitude/phase of the 2 × 2mm

diaphragm at around fundamental resonance frequency.

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials746

fabricated pMUTs, the highest SPL obtained is around 120dB at a distance

of 12mm. Figure 24.19(b) shows another example. The SPL is 116.8 dB at

76.3kHz for a square pMUT element with side length of 1.5mm. As a

ultrasonic transmitter, this output SPL is quite good.

We also investigated the effect of the dc bias, the driving voltage and the

measuring distance on the output SPL and resonance frequency of a pMUT

with 7µm-thick composite PZT layer. Figure 24.20 shows the results. The

output SPL increases with the increase of the applied dc bias and saturated

at certain dc bias level (30 V as shown in Fig. 24.20a) at which the polarization

of the film approaches saturation. Figure 24.20(b) shows that the output SPL

increases with the driving voltage and decreases with the increase of the

measuring distance.

41.3kHz, 107.8 dB

20V dc bias, 30 V

P-P

36 38 40 42 44 46 48

Frequency (kHz)

(a)

Sound pressure level (dB)

120

110

100

76.3 kHz, 116.8 dB

20 V dc bias, 20 V

P-P

70 72 74 76 78 80

Frequency (kHz)

(b)

Sound pressure level (dB)

120

100

24.19

Frequency dependence of output sound pressure level of

pMUTs with different diaphragm sizes: (a) 2 mm × 2mm, (b)1.5 mm

× 1.5 mm.

WPNL2204

Piezoelectric thick films for MEMS application 747

24.6.3 Directivity of 2D pMUT array

The pMUT array has potential applications in object detection and

environmental recognition, either at low ultrasonic frequency in air or at

high frequency in liquid medium or human body. In the design of pMUT

array for these applications, the number of elements in the array has to be

increased to meet the requirement of the resolution and signal to noise ratio.

As a result, the whole array will occupy a large area on the silicon wafer. In

this case, it would be increasingly difficult to maintain uniform thickness of

the diaphragm throughout a whole array by wet silicon etching. This problem

can be solved by making use of SOI wafers and DRIE etching since the

Sound pressure level (dB)

114

112

110

108

106

104

102

100

10 20 30 40 50 60

bias (V)

(a)

30 V

p–p

, 12mm

Thick R-A

PZT 7 µm

2 mm × 2mm

Driving voltage (V)

0 10 2030405060

Bias 30 V, 12 mm

Bias 30 V, 30 V

p-p

Thick R-A

0 50 100 150 200 250

Distance (mm)

(b)

Sound pressure level (dB)

120

116

112

108

104

100

24.20

Output sound pressure level at resonance frequency versus (a)

dc bias (at driving voltage of 30 V

p-p

), (b) peak to peak driving

voltage (at dc bias of 30V) and distance between measuring point

and transmitter (bias 30V, 30 Vp-p).

WPNL2204

Handbook of dielectric, piezoelectric and ferroelectric materials748

etching will stop automatically at the SiO

layer and the diaphragm thickness

will be determined by the device layer of the SOI wafer. This can ensure a

very uniform diaphragm thickness over the whole wafer, resulting in a uniform

resonance frequency of the individual elements of pMUT array. Figure 24.21

shows fabricated pMUT arrays on a 4-inch (100 mm) SOI wafer for ultrasound

beam forming and beam steering applications. The individual devices were

separated by wafer dicing and electrically connected by wire bonding.

We investigated the directivity of the ultrasound beam emitted by the

pMUT array at resonance frequency. The tested array consists of 37 elements

with a size of 1.5 × 1.5 mm

. An ac driving voltage of 10 V

p-p

superimposed

on a dc bias of 30V is applied across the 6µm-thick piezoelectric layer. The

support silicon layer is 40 µm thick. The array surface is rotated from –90°

to +90° with respect to the normal direction of an acousto-optic sensor. The

displacement output of the sensor in response to the radiated ultrasound

from the array was recorded at every rotation step of 1° and was utilized as

the measure of sound pressure. The ultrasound radiation pattern of the array

measured at its resonant frequency of 160 kHz is shown in Fig. 24.22. It can

be seen that the main lobe is very sharp and tall, which reveals that the

fabricated transmitter array can be employed in beam forming and beam

steering applications.

24.21

Piezoelectric ultrasonic transmitter arrays fabricated on a 4-inch

(100 mm) silicon wafer and a separate array device after wafer dicing

and wire bonding.

WPNL2204

Piezoelectric thick films for MEMS application 749

24.6.4 Electromechanical coupling coefficient

The effective electromechanical coupling coefficient k

of pMUT is defined as

= 1 – (f

)

24.2

where f

means the resonance frequency and f

the anti-resonance frequency

in an impedance frequency spectrum, respectively.

Figure 24.23 shows the measurement results of the impedance and phase

frequency spectrum for a 0.5 × 0.5mm

pMUT element with a 7µm-thick

PZT film and 10µm-thick support layer. The impedance frequency spectrum

of the pMUT was characterized using an Agilent 4294A Precision Impedance

Analyzer. The measured f

and f

are 259.87 and 261.73 kHz, respectively,

resulting in a coupling coefficient of 1.42% as calculated according to equation

(24.2). The measured value is similar to that of a diaphragm prepared with

sol–gel PZT film.

With optimum thicknesses of PZT and silicon in the

diaphragm, a better coupling coefficient can be expected by using our thick

composite PZT film.

24.6.5 Equivalent circuit parameters of the pMUT

The Q factor of the pMUT can be determined by the real part of the impedance

frequency spectrum, which is defined as Q = f

/∆f, where the resonance

frequency f

is the frequency at which the real part of the impedance reaches

its maximum, ∆f is the width of the peak at its half height, so-called 3dB

Displacement (nm)

2.5

2.0

1.5

1.0

0.5

0.0

0.5

1.0

1.5

2.0

2.5

180

150

120

330

300

270

240

210

24.22

The directivity pattern of a pMUT array fabricated with SOI

wafer and measured by using a diaphragm type acousto-optic

sensor.

WPNL2204