Miniaturized all-optical photoacoustic
microscopy based on
microelectromechanical systems mirror scanning
Sung-Liang Chen,1,†Zhixing Xie,1,†Tao Ling,2L. Jay Guo,2Xunbin Wei,3and Xueding Wang1,*
1Department of Radiology, University of Michigan, Ann Arbor, Michigan 48109, USA
2Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, Michigan 48109, USA
3Med-X Research Institute, Shanghai Jiao Tong University, Shanghai 200030, China
*Corresponding author: firstname.lastname@example.org
Received August 6, 2012; accepted August 31, 2012;
posted September 12, 2012 (Doc. ID 173862); published October 11, 2012
Achieving photoacoustic microscopic imaging through a miniaturized scanning head is a crucial step toward
high-resolution photoacoustic endoscopy. In this work, we have developed a miniaturized probe head using a
microelectromechanical systems (MEMS) based mirror for raster scan of the laser beam and our newly developed
super broad bandwidth microring resonator based ultrasound detector for photoacoustic signal detection. Through
this all-optical design, which offers unique advantages for endoscopic applications, this system is capable of three-
dimensional (3D) imaging with high resolution of 17.5 μm in lateral direction and 20 μm in axial direction at a
distance of 3.7 mm. After the performance of this system was validated through the experiments on printed grids
and a resolution test target, microscopic imaging of the 3D microvasculatures in canine bladderswas also conducted
successfully, demonstrating the potential of novel photoacoustic endoscopic in future clinical management of
bladder cancer.© 2012 Optical Society of America
OCIS codes:170.3880, 170.5120, 180.5810.
A majority of cancers originate in the epithelial tissues of
the human body, including those in the oral cavity and
pharynx, digestive system, respiratory system, genital
system, and urinary system. For example, bladder cancer
is the fifth leading new cancer diagnosis in the United
States and is fourth among men . The long-term survi-
val and the need for lifelong routine monitoring and treat-
ment lead to a cost per patient for bladder cancer that is
the highest among all cancers [2,3]. Despite the advances
in diagnosis and therapy in the past decades, bladder can-
cer remains an important public health problem. One of
the major reasons is the lack of powerful screening and
Photoacoustic imaging (PAI) is an emerging modality
and has great potential for in vivo biomedical applica-
tions . It provides image information of optical-
absorption contrast or functional properties by detecting
ultrasonic waves, generated based on thermoacoustic
effect. It has been demonstrated that a potential marker
to obtain prognostic information about bladder cancer is
tumor neoangiogenesis , which can be quantified by
morphometric characteristics, such as microvascular
density. Photoacoustic microscopy (PAM), by rendering
three-dimensional (3D) map of microvasculature with
unbeatable sensitivity and high resolution, holds promise
for evaluation of neoangiogenesis associated with blad-
der cancer generation and progression . To achieve
in vivo noninvasive evaluation of a bladder tumor with
microscopic resolution, the best method is to perform
PAI in an endoscopic manner, similar to conventional
white light cystoscopy. However, despite the rapid
advancement of PAI technology in the past decade,
achieving high-resolution endoscopic imaging with PAI
is still very challenging and in need of development.
One reason is that most current PAI systems are usually
bulky and do not have sufficient detection bandwidth
required for microscopic resolution.
Miniaturized probes for PAI have been studied recently
[7–10]. Mechanical scanning based on micromotor and a
single transducer can acquire only two-dimensional (2D)
and side-view images . Another design using a taper
reflector and a ring transducer array suffers from poor
transverse resolution . We expect that applying micro-
electromechanical systems (MEMS) mirrors capable of
2D scanning is the best way to enable forward-view scan-
ning  and high-resolution imaging . In this work, we
demonstrate feasibility to miniaturize the probe head of a
PAM system without sacrificing the image quality, by
using a MEMS mirror for fast scanning. In this system,
an optical microring resonator with a noise equivalent
detectable pressure (NEDP) of 29 Pa is employed for sen-
sitive ultrasound detection . The detection mechan-
ism has been introduced before in the literature .
Optical detection of photoacoustic signals leads to a very
unique all-optical design. The advantages of the micror-
bandwidth, size-independent sensitivity, and free from
electromagnetic interference, particularly benefit the
design of an endoscopic probe. Moreover, to gain enough
signal-to-noise ratio, the piezoelectric transducer used in
conventional PAM systems should connect directly to an
amplifier circuit/system before the signal is routed to the
data acquisition board. This creates significant challenge
for endoscopic application imposed by the limited space
constraints. In contrast, microring sensors with optical
fiber connections eliminate the signal routing and deliv-
ery problem easily by using external high-speed and high-
The schematic of the system is shown in Fig. 1(a). A
diode-pumped solid-state Nd:YAG laser (SPOT-10-200-
532, Elforlight Ltd., UK) is employed to provide pulses of
2 ns duration at wavelength of 532 nm. The laser beam is
coupled into a fiber (SMF-28e+, Thorlabs, Newton, New
Jersey) by a focusing lens, and is delivered to a MEMS tilt
October 15, 2012 / Vol. 37, No. 20 / OPTICS LETTERS4263
0146-9592/12/204263-03$15.00/0© 2012 Optical Society of America
mirror (TM-2520, Sercalo Microtechnology Ltd., Switzer-
land) with a laser spot size of1 mmafter a collimator. The
MEMS mirror is used to perform 2D scanning of laser
beam. Figure 1(b) shows a photograph of the MEMS mir-
ror. An aspheric lens (A220-A, Thorlabs) is used as an ob-
jective lens. The diameter of the commercial MEMS
mirror and the objective lens is 9.2 and 7.2 mm, respec-
tively. A MEMS driver board provided by Sercalo Ltd. is
used to control the mirror tilt angle. A digital-to-analog
converter in the driver board delivers voltages to set
lized by setting different voltages of two pairs of electro-
des, which is used to control two-axis tilt. Figure 1(c)
shows calibrated response oftiltangle versusapplied vol-
tages. Compared with our previous all-optical PAM sys-
tem , a fiber-optic pulsed laser delivery, a MEMS
mirror, and a small objective lens are adopted in the cur-
rent system to meet the needs of miniaturization.
In this system, the laser works in an external trigger
mode. An output trigger signal is used to synchronize
the laser triggering, the MEMS tilt, and the data acquisi-
tion. Laser-generated photoacoustic signal is received by
a microring detector, which is placed under the sample
and works on a transmission mode. The microring reso-
nator with a ring diameter of 60 μm is a wideband single-
element ultrasound detector with a measured receiving
bandwidth up to ∼100 MHz . The photoacoustic sig-
nal modulated optical signal from the microring is de-
tected by a photoreceiver module (1801-FC-AC, New
Focus). The detected signal is then digitized at a sampling
rate of 200 MHz for 2.5 μs and stored by a data acquisition
board in a PC. The microring resonator is kept stationary
during laser scanning and no signal averaging is needed.
We have acquired images from both phantom and ex
ging system. Given the 11 mm focal length of the
objective lens and the maximum tilt angles of the MEMS
mirror (?6° and ?9° for the two axes, respectively), the
system is able to cover an area of up to 2.3 mm×
voltages are set according to the calibrated nonlinear re-
sponse in Fig. 1(c). Due to the 50 Hz limit of the MEMS
driver board, a 2D raster scan of 256 × 256 steps takes al-
most half an hour, which also includes the data transfer
time. However, as we know, a MEMS mirror with step re-
sponse setting times <100 μs is commercially available,
which can drastically improve the imaging speed.
Figure 2 shows the imaging results of phantoms. We
placed the microring detector at ∼3.7 mm under the
phantom surface. Both phantom and the detector were
immersed in the deionized water for acoustic coupling.
A black mesh grid printed on a transparency with a de-
signed gap size of 381 μm was imaged. The scanning area
was 1 mm × 1 mm. The maximum amplitude projection
(MAP) image was presented in Fig. 2(a). The sizes of
the grids measured in the image were in the range of
360–390 μm, showing a good agreement with the de-
signed grid size. Figure 2(b) presents a typical A-line sig-
nal and its amplitude after the Hilbert transform
(envelope). We determined the axial resolution using
the criteria of the FWHM. The quantified axial resolution
is better than 20 μm at a detection distance of 3.7 mm.
The good axial resolution is benefited from the wide
bandwidth of the microring resonator. Higher axial reso-
lution can also be achieved by reducing the sound pro-
pagation distance .
To quantify the system lateral resolution, we imaged a
USAF 1951 resolution test target (T-20-P-TM, Applied Im-
age, Rochester, New York). The MAP image of the reso-
lution target was shown in Fig. 2(c). Figure 2(d) showed
the one-dimensional (1D) profile measured along the line
as marked in Fig. 2(c). Spatial averaging over ∼12 pixels,
as indicated by the bar in Fig. 2(c), was applied to im-
prove the contrast-to-noise ratio (CNR) in Fig. 2(d).
The smallest resolvable bar spacing is 17.5 μm (group
5, element 6), which indicates the lateral resolution is
better than 17.5 μm. Since the same detector is used in
our current and previous system , the NEDP and axial
resolution are expected to be similar. The lateral resolu-
tion and imaging speed are worse in the current system.
However, the former can be improved by increasing the
NA of the objective lens and the latter can be increased
system. (b) A photograph of the MEMS mirror. (c) The cali-
brated response of mirror tilt angle versus applied voltages.
(Color online) (a) Schematic of MEMS-based PAM
(b) A-line photoacoustic signal and its amplitude after the
Hilbert transform. (c) The MAP image of a USAF resolution test
target. (d) 1D profile plotted along the lines marked in (c).
(Color online) (a) MAP image of a printed mesh grid.
4264OPTICS LETTERS / Vol. 37, No. 20 / October 15, 2012
by using a MEMS mirror with fast response time, as men- Download full-text
To evaluate the performance of this system in imaging
biological tissues, fresh canine bladders were employed.
This study on canine bladder model also helps to explore
the potential of adapting the novel PAM technique to
bladder cancer diagnosis. Before the imaging experi-
ment, the harvested whole bladder was kept fresh in a
refrigerator. The laboratory animal protocol for this work
was approved by the UCUCA of the University of Michi-
gan. Figure 3(a) shows an anatomical photograph of the
excised inner wall of the bladder.
Other than imaging using the microring resonator, we
also acquired images from the same sample using a nee-
dle hydrophone for comparison with the results from the
microring resonator. The highly sensitive custom-built
needle hydrophone has a center frequency of 35 MHz
and a −6 dB bandwidth of 100%. The needle hydrophone
was operated on the reflection mode, while the microring
detector worked in the transmission mode. In this way,
the same area in the same sample can be scanned with
the microring resonator and the hydrophone simulta-
neously, facilitating a better comparison. The MAP
images showing the microvasculature in the canine blad-
der obtained by the microring resonator and the hydro-
phone are shown in Figs. 3(b) and 3(c), respectively. The
maximum CNRs are similar: 32 dB for Fig. 3(b) and 33 dB
for Fig. 3(c). Larger field of view was obtained by the mi-
croring resonator because of its small active element size
(60 μm) compared with that of the needle hydrophone
(∼1 mm). The MAP image on the XZ plane acquired
by the hydrophone shows obscure features along the
axial direction, which is due to relatively narrower band-
width and, as a result, worse axial resolution associated
with the needle hydrophone. In contrast, at least two
layers of tissue features can be recognized along the Z
direction in the MAP image on the XZ plane obtained
by the microring resonator.
In this work, we have demonstrated a miniaturized
PAM probe head capable of optical-resolution PAI. As
demonstrated by the results from phantoms and ex vivo
tissues, this system based on the MEMS mirror scanning
and the microring detector shows satisfactory image
quality. High resolution of 17.5 μm in lateral direction
and 20 μm in axial direction have been achieved. Micro-
vasculatures in canine bladders have also been mapped
with good spatial resolution, high CNR, and satisfactory
continuity, suggesting the potential of PAM in diagnostic
imaging and guided treatment of bladder cancer and
other epithelial cancers.
However, before a clinically usable endoscopic PAM
system could be fabricated, several technical problems
still need to be resolved. First, a MEMS mirror with smal-
ler element size in the range of 3–5 mm is needed if the
mercial cystoscopic catheters). We also expect that the
MEMS mirror could allow wider scanning angle for larger
imaging area and faster scanning speed, both major con-
ond, we are also trying to build a microring resonator that
is optically transparent at the wavelength of the pulsed
laser for photoacoustic signal generation. With such a de-
sign, the scanning laser beam can go through the micror-
ing resonator without being interfered, facilitating a
reflection mode PAM. As demonstrated in our previous
study, building an optically transparent microring resona-
tor is feasible, by using a transparent silica substrate .
This work was supported by the National Institutes of
Health under grants R01AR060350 and R01AR055179,
University of Michigan-Shanghai Jiao Tong University
(UM-SJTU) Joint Grant, and Basic Radiological Sciences
(BRS) Grant Award G012284.
†These authors contributed equally to this work
1. A. Jemal, R. Siegel, J. Xu, and E. Ward, CA-Cancer J. Clin.
60, 277 (2010).
2. E. B. C. Avritscher, C. D. Cooksley, H. B. Grossman, A. L.
Sabichi, L. Hamblin, C. P. Dinney, and L. S. Elting, Urology
68, 549 (2006).
3. M. F. Botteman, C. L. Pashos, A. Redaelli, B. Laskin, and R.
Hauser, Pharmacoeconomics 21, 1315 (2003).
4. L. V. Wang, ed., in Photoacoustic Imaging and Spectro-
scopy (Taylor & Francis/CRC Press, 2009).
5. Y.-K. Zee, J. P. B. O’Connor, G. J. M. Parker, A. Jackson,
A. R. Clamp, M. B. Taylor, N. W. Clarke, and G. C. Jayson,
Nat. Rev. Urol. 7, 69 (2010).
6. B. H. Bochner, R. J. Cote, N. Weidner, S. Groshen, S.-C.
Chen, D. G. Skinner, and P. W. Nichols, J. Natl. Cancer Inst.
87, 1603 (1995).
7. J.-M. Yang, K. Maslov, H.-C. Yang, Q. Zhou, K. K. Shung, and
L. V. Wang, Opt. Lett. 34, 1591 (2009).
8. Y. Yuan, S. Yang, and D. Xing, Opt. Lett. 35, 2266 (2010).
9. L. Xi, J. Sun, Y. Zhu, L. Wu, H. Xie, and H. Jiang, Biomed.
Opt. Express 1, 1278 (2010).
10. S.-L. Chen, T. Ling, H. W. Baac, and L. J. Guo, Proc. SPIE
7899, 78992T (2011).
J. Zou, and L. V. Wang, J. Biomed. Opt. 17, 080505, (2012).
12. Z. Xie, S.-L. Chen, T. Ling, L. J. Guo, P. L. Carson, and X.
Wang, Opt. Express 19, 9027 (2011).
13. C.-Y. Chao, S. Ashkenazi, S.-W. Huang, M. O’Donnell, and
L. J. Guo, IEEE Trans. Ultrason. Ferroelectr. Freq. Control
54, 957 (2007).
wall of a canine bladder. (b) MAP images on the XY and XZ
planes of the microvasculature in the bladder wall obtained
by microring detectors. (c) MAP images of the same area ac-
quired by the needle hydrophone. All images share the same
scale bar as shown in the left-upper image.
(Color online) (a) Anatomical photograph of the inner
October 15, 2012 / Vol. 37, No. 20 / OPTICS LETTERS4265