HAL Id: hal-01302015
https://hal.sorbonne-universite.fr/hal-01302015
Submitted on 13 Apr 2016
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
A capacitive micromachined ultrasonic transducer probe
for assessment of cortical bone
Audren Boulmé, Sophie Ngo, Jean-Gabriel Minonzio, Mathieu Legros,
Maryline Talmant, Pascal Laugier, Dominique Certon
To cite this version:
Audren Boulmé, Sophie Ngo, Jean-Gabriel Minonzio, Mathieu Legros, Maryline Talmant, et al.. A
capacitive micromachined ultrasonic transducer probe for assessment of cortical bone. IEEE Trans-
actions on Ultrasonics, Ferroelectrics and Frequency Control, Institute of Electrical and Electronics
Engineers, 2014, 61 (4), pp.710-723. �10.1109/TUFFC.2014.2959�. �hal-01302015�
For Review Only
1
A capacitive Micromachined Ultrasonic Transducer
probe for assessment of cortical bone
Audren Boulmé
(1)
, Sophie Ngo
(2)
, Jean-Gabriel Minonzio
(3)
, Mathieu Legros
(4)
, Maryline Talmant
(3)
, Pascal
Laugier
(3)
, Dominique Certon
(1)
(1) François Rabelais University, GREMAN UMR-CNRS 7347, 10 Bd Tonnelle, Tours, France.
(2) STMicroelectronics, 10 rue Thalès de Milet, 37071 TOURS Cedex 2
(3) UPMC Univ Paris 06, LIP UMR-CNRS 7623, 15 rue de l’école de médecine, Paris, France.
(4) Vermon SA, 180 Av. Général Renault, Tours, France.
Abstract
A wide range of ultrasound methods has been proposed to assess the mechanical strength of bone. Axial
transmission technique, which consists of measuring guided elastic modes through the cortical shell of long
bones such as the radius and the tibia, has recently emerged as one of the most promising approaches of all bone
exploration methods. Determination of dispersion curves of guided waves is therefore of prime interest as they
provide a large set of input data required to perform inverse process, and hence to evaluate bone properties
(elastic and geometric). The cortical thickness of long bones ranges from approximately 1 to 7 mm, resulting in
wide inter-individual variability in the guided wave response. This variability can be overcome by using a single
probe able to operate with a tunable central frequency typically, within the 100 kHz – 2 MHz frequency range.
However, there are certain limitations in the design of low frequency arrays using traditional PZT technology,
and these limitations have triggered active research to find alternative solutions. Capacitive Micromachined
Ultrasonic Transducers (cMUTs) present the potential to overcome these limitations and to improve axial
transmission measurement significantly. The aim of the study presented here was to design and construct a new
cMUT-based axial transmission probe and to validate the approach. We report all the steps followed to construct
such a prototype, from the description of the fabrication of the cMUT (based on a surface micromachining
process) through to probe packaging. The fabricated device was carefully characterized using both electrical and
optical measurements in order to check the homogeneity of the device first from cMUT to cMUT and then from
element to element. Finally, axial transmission measurements carried out with the prototype cMUT probe are
shown and compared to results obtained with a counterpart PZT-based array.
Index Terms: cMUTs – Acoustic linear arrays – Cortical bone – Axial transmission – Lamb waves – Guided modes –
Osteoporosis – Surface micromachining
Page 1 of 26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Review Only
2
I. INTRODUCTION
The most common methods used for assessment of skeletal status use X-ray-based techniques to determine
bone mineral density (BMD) [1]. Although BMD is an important predictor of bone strength, its measurement
alone fails to predict fracture risk accurately. Additional bone characteristics, including geometric
microstructural and material properties, are required in order to assess the biomechanical competence of the bone
more accurately. Ultrasound techniques are among the most promising modalities to fulfill this requirement [2].
Recent ex vivo experiments [3-7] carried out on long bones have shown that the cortical envelope of long bones
behaves like an acoustic waveguide for frequencies between 100 kHz and 2 MHz. Guided modes are sensitive to
elastic and geometric properties of the waveguide. These results have encouraged new research to extend the
concept of the axial transmission technique in order to exploit the waveguide character of cortical bone. Several
axial transmission techniques have been described [8-14]. They are all based on the same measurement principle,
consisting of using a set of emitters and receivers placed on the same side of the long bone, aligned with the
bone axis. However, according to the operating conditions, different working frequency bandwidths and
different signal processing techniques have been implemented, resulting in different waveforms being detected
and subsequently processed: these include the first arriving signal (FAS) [15], an energetic late arrival signal [4]
or a set of multiple guided waves [3, 7, 14]. Only FAS-based techniques have been clinically validated and they
showed they were able to predict fracture risk [16]. However, this capacity is at best equivalent to that of X-ray
densitometry, but not better. One major limitation of the FAS-based approach is that it measures only one
parameter, i.e. FAS velocity, and it fails to yield a complete assessment of the various determinants of bone
strength, such as cortical thickness, stiffness and porosity. Other approaches dedicated to a more comprehensive
analysis of multiple recorded waveforms have recently been proposed [6-7, 14, 17] and tested successfully on
bone mimicking phantoms [4, 14] and on ex vivo bone specimens [7, 18]. Briefly, cortical bone is considered as
a waveguide and the guided mode spectrum is estimated [6-7, 14, 19] and a fit of the experimental spectrum to a
model of the waveguide eventually yields estimates of the waveguide characteristics such as its thickness and
stiffness coefficients [4, 17, 20].
Our group has developed specific multi-emitter and multi-receiver ultrasonic transducers dedicated to axial
transmission measurement [13] consisting of a 1-D linear array. In contrast to arrays used in pulse-echo imaging,
the emitters and receivers are separated, so that each element of the array works either as an emitter or a receiver.
The electronic driving circuits of the emitters and receivers are therefore simplified due to elimination of
electronic protection circuits used in classical pulse-echo imaging. However, acoustic specifications and using
Page 2 of 26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Review Only
3
conditions associated with axial transmission measurements require a probe with a wide bandwidth, able to
cover the frequency bandwidth used for cortical bone testing, typically from 100 kHz up to 2 MHz. In order to
compensate for strong attenuation of the bone tissue, signals with high excitation amplitude are used to drive the
probe and this generally leads to an increase in the temperature of the probe. The probe must thus withstand high
temperature increases. Secondly, a wide angle directivity pattern is suitable in order to generate a large set of
guided modes in the cortical bone. Finally, since all measurements are carried out in the near field of the probe,
the electrical and acoustic cross-coupling phenomena must be treated with great care to avoid the presence of
artifacts among the guided modes measured. In the study reported here we assumed that the characteristics
offered by cMUT-based transducer technology (capacitive Micromachined Ultrasonic Transducers) should
overcome some of the limitations encountered with PZT probes for our purposes. We present the fabrication and
testing of an axial transmission probe based on cMUT technology and compare its performance with the results
obtained with a PZT probe.
cMUTs are micro-electromechanical systems [21] introduced by P. Khuri-Yakub [22] and P. Eccardt [23] in
1996. They work like in-air microphones used in audio applications where mechanical vibration is produced by
electrostatic forces instead of the piezoelectric effect used for standard technology. The topology of an acoustic
cMUT-based antenna is defined by the pitch of the array, the kerf, the elevation and the number of
emitters/receivers. Each acoustic radiator of an array is created by few thousand micro-membranes driven
individually by electrostatic forces. The shape of each cMUT depends on the acoustic design: square, circular,
polygonal or rectangular. The most advanced technological demonstrations with cMUT probes have been
published in the field of classical pulse-echo imaging by O. Oralkan [24], R. O. Guldiken [25], G. Gurun [26], A.
Savoïa [27], M. Legros [28] and A. Novell [29]. Why is cMUT technology suitable for axial transmission
measurements? First, weak internal cross-talk exists in a cMUT probe, their directivity pattern is dependent only
on the width of the emitter. In contrast, a standard PZT-based probe is a multi-layered structure in which guided
modes can exist and create cross-talk which leads to narrowing of the directivity pattern. Secondly, the
temperature increase is very low, as demonstrated by S. H. Wong et al. [30]. In a PZT probe the temperature
increase is mainly caused by internal dielectric and mechanical losses while in the case of a cMUT probe there
are no internal losses, since the main source of losses in a cMUT probe is the radiation of the acoustic pressure
field. Finally, preliminary studies undertaken by Sénégond et al. [31] have demonstrated that a cMUT probe can
operate at a central frequency fixed by its coupling with the fluid, but they can also generate efficient acoustic
fields at very low frequencies, i.e. in a quasistatic regime. The reason is that, when a cMUT works at very low
Page 3 of 26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Review Only
4
frequencies, it produces a very considerable displacement amplitude (close to the gap height value) which can
somewhat compensate for the acoustic mismatching. However, despite strong non-linearities which characterize
the low frequency behavior of cMUTs, we demonstrate here that such an operating regime can be used to enlarge
the probe’s bandwidth without any impact on the detection of guided modes in the plate under test.
This article comprises three parts. The first focuses on description of the probe design and its fabrication:
from micromachining the cMUT up to the final packaging. A dedicated pre-amplification circuit was placed in
the head probe to increase the signal to noise ratio. To help explain the design concept, we have added a
reminder of the principle of axial transmission measurement. Simulations performed to establish sizes and
topologies of the cMUT cells are also reported in the first section. The second section focuses on the
characterization of the cMUT probe. It reports basic post-process measurements such as static deflection of the
diaphragm performed on a non-packaged cMUT and acoustic pressure field measurements. The aim is to provide
complete data for the probe from the unit cell up to the overall acoustic performance of the probe. Among all the
output data provided, the homogeneity of the prototype from cell to cell and then from element to element is
discussed. Finally, axial transmission tests carried out with the cMUT probe are reported in the last section. The
acoustic phantom was a bone mimicking phantom. The results are discussed and compared with a PZT probe
with the same topology.
II. DESIGN AND FABRICATION OF THE CMUT PROBE FOR AXIAL TRANSMISSION MEASUREMENT
A. Principle of Axial Transmission Measurement: Topology of the Array
Using a transducer in contact with the skin (Fig. 1), one acoustic pulse is emitted towards the cortical bone
and then collected throughout its propagation using a set of transducers regularly spaced along the bone. If the
emitter width is small enough at the skin/bone interface, according to classical refraction laws, the acoustic rays
emitted in the skin are converted into guided modes in the cortical bone. The central frequency and bandwidth of
the acoustic pulse select the type of guided modes propagated. In practical terms, the bandwidth to inspect bone
tissue should be 100 kHz - 2 MHz. A set of RF signals is collected by the receivers to construct a data matrix
that is dependent on time and distance, from which the dispersion curves of the guided modes can be determined
[14]. Bossy et al. developed a specific biaxial transmission probe [13] to improve the signal to noise ratio and to
compensate for the thickness of the soft tissue (mainly skin and fat). The designed probe of this study has similar
topology (Fig. 2). It is divided into five zones: two emission arrays placed at the probe extremities, one reception
array placed at the center of the probe and two dead zones for the propagation of waves. The specifications of
Page 4 of 26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
