June 15, 2009 / Vol. 34, No. 12 / OPTICS LETTERS 1771

Laser-scanning optical-resolutionphotoacoustic microscopy

Zhixing Xie,1 Shuliang Jiao,2 Hao F. Zhang,1,* and Carmen A. Puliafito3

1Department of Electrical Engineering and Computer Science, University of Wisconsin–Milwaukee,Milwaukee, Wisconsin 53211, USA

2Department of Ophthalmology, University of Southern California, Los Angeles, California 90033, USA3Keck School of Medicine, University of Southern California, Los Angeles, California 90033, USA

*Corresponding author: zhang25@uwm.edu

Received January 27, 2009; revised March 31, 2009; accepted April 23, 2009;posted May 13, 2009 (Doc. ID 106613); published June 2, 2009

We have developed a laser-scanning optical-resolution photoacoustic microscopy method that can potentiallyfuse with existing optical microscopic imaging modalities. To acquire an image, the ultrasonic transducer iskept stationary during data acquisition, and only the laser light is raster scanned by an x–y galvanometerscanner. A lateral resolution of 7.8 �m and a circular field of view with a diameter of 6 mm were achieved inan optically clear medium. Using a laser system working at a pulse repetition rate of 1024 Hz, the dataacquisition time for an image consisting of 256�256 pixels was less than 2 min. © 2009 Optical Society ofAmerica

OCIS codes: 110.5120, 110.0180.

Developing multimodal microscopic imaging technol-ogy covering different contrast mechanisms toachieve comprehensive physiological information ofbiological tissues is becoming increasingly importantin biomedical research [1–3]. However, as the onlyavailable high-resolution modality to image opticalabsorption contrast noninvasively, photoacoustic mi-croscopy (PAM) has not been successfully integratedwith other optical microscopic imaging technologiesthat provide, for example, optical scattering and fluo-rescence contrasts.

Photoacoustic (PA) imaging is a scalable imagingtechnology that detects laser-induced ultrasonicwaves to image the three-dimensional distribution ofthe optical energy deposition, which is related to theoptical absorption. PA imaging has been investigatedintensively during the past decade owing to its abilityto achieve both physiologically specific optical ab-sorption contrast and good spatial resolution in deeptissue. Several different designs were reported, andpotential applications in a wide spectrum of both bio-medical research and clinical diagnostics also weredemonstrated [4,5].

Among all the PA imaging technologies, only PAM[6–8] has demonstrated capabilities to image biologi-cal tissue at a microscopic scale. PAM usually em-ploys a focused single-element high-frequency ultra-sonic transducer and a spatially overlapped opticalillumination. Depending on the relative sizes of theoptical and ultrasonic focuses in PAM, the spatial(lateral) resolution can be provided either ultrasoni-cally or optically.

To image biological tissue beyond one optical trans-port mean free path, where photons are randomlyscattered, ultrasonic resolution is required owing tothe much weaker ultrasonic scattering in tissue. Us-ing a 50 MHz focused ultrasonic transducer, a lateralresolution of 45 �m was achieved at a maximumdepth of 3 mm [7]. When the object is optically thin(within one transport mean free path) or optically

0146-9592/09/121771-3/$15.00 ©

transparent, optical resolution can be employed tofurther improve the spatial resolution of PAM. Anoptical-resolution PAM (OR-PAM) was recently re-ported, where a lateral resolution of 5 �m wasachieved [8].

The integration of PAM with existing optical micro-scopic imaging modalities such as confocal micros-copy, two-photon microscopy, and optical coherencetomography has been hampered, because existingPAMs require mechanical scanning of the ultrasonic-optical assembly, which is not compatible with otheroptical microscopic imaging modalities. Since PAMuses ultrasonic detection whereas other methods useoptical detection, the only possible solution for com-patibility is to use the same optical scanning mecha-nism as other optical microscopic imaging modalities.

In this Letter, we report on the laser-scanning OR-PAM (LSOR-PAM) to demonstrate the feasibility ofemploying optical scanning in PAM. The field of view(FOV) and spatial resolutions were quantified by im-aging various phantoms, and in vivo images ofmicrovessels in mouse ears were acquired.

In LSOR-PAM (Fig. 1), the ultrasonic detector iskept stationary, and only the laser light is rasterscanned within the FOV during data acquisition. Atunable dye laser (Cobra, Sirah Laser and Plas-

Fig. 1. (Color online) Schematic of the LSOR-PAM experi-

mental setup.

2009 Optical Society of America

1772 OPTICS LETTERS / Vol. 34, No. 12 / June 15, 2009

matchnik GmbH) pumped by an Nd:YLF laser sys-tem (IS8II-E, EdgeWave GmbH; pulse duration: 6 ns)was used as the irradiation source. The laser systemcan operate with a pulse repetition rate (PRR) of upto 5 kHz in an optical wavelength tuning range be-tween 540 nm to 900 nm. In this experiment, the la-ser system worked in an external trigger mode, and asingle optical wavelength of 580 nm was used. Theoutput laser light from the dye laser was spatially fil-tered by an iris and expanded to reach a beam diam-eter of 8 mm. After attenuated by a neutral densityfilter, the expanded laser beam then passed throughan x–y galvanometer scanner (6230H, CambridgeTechnology) and a 60 mm focal length objective lens(AC254-060-A1, Thorlabs).

A PC-controlled (scan-controlling PC) analog-output board (PD2-AO-16, United Electronic Indus-tries) synchronized the laser triggering and the opti-cal scanner. The induced PA waves were detected bya commercial, unfocused ultrasonic transducer(V312, Olympus NDT; center frequency: 10 MHz;bandwidth: 80%; active element diameter: 6 mm).The ultrasonic transducer was placed 30 mm awayfrom the sample and tilted about 15°. The detectedPA signals were first amplified by a wideband ampli-fier (ZFL 500LN, Mini-Circuits) and then digitizedand stored by a data acquisition board (CS12400,Gage Applied) controlled by the data acquisition PCfor further processing. A fast photodiode (DET10A,Thorlabs) detected the laser pulses and triggered thedata acquisition to avoid the impact of laser jittering.The energy of each laser pulse was also recordedfrom the photodiode to compensate for the pulse en-ergy instability. The sampling rate of the data acqui-sition board was 200 MS/s, and the PA signals wererecorded for 3 �s. No signal averaging was employedduring data acquisition.

We quantified the FOV and spatial resolution ofthe LSOR-PAM through phantom imaging. The FOVof LSOR-PAM is determined by the ultrasonic beamsize of the unfocused ultrasonic detector, since onlythe optical focus was raster scanned. According to oursimulation on the ultrasonic beam profile the beamdiameter �−6 dB� of the unfocused ultrasonic detectoris comparable to its active element size (6 mm in di-ameter) within at least 30 mm away from its trans-ducer surface.

Figure 2(a) gives the experimental measurement ofthe FOV in LSOR-PAM. The sample was a blackmesh grid printed on a transparency using a com-mercial laser printer. The measured grid size was865 �m using an optical microscope (Axiovert 200,Zeiss). During the experiments, the grid and theultrasonic detector were immersed in deionized wa-ter inside a water tank. The scanning area was8 mm�8 mm. The laser pulse energy was 0.04 �J.After the volumetric data were acquired, themaximum-amplitude-projection (MAP) image wascalculated [6]. As Fig. 2(a) shows, a maximumcontrast-to-noise ratio (CNR) of 26 dB was achievedwithout any image processing except that high-frequency noise above 100 MHz was digitally re-

moved. Based on the size of the grid, a circular FOV

with a diameter of �6 mm was achieved (at leastseven grids can be clearly observed with a localCNR�6 dB).

The lateral resolution of the LSOR-PAM was quan-tified by imaging an USAF 1951 resolution test tar-get (IT-20-P-TM, Applied Image). Figure 2(b) showsthe MAP image of the resolution target. Figure 2(c)shows the magnitude of the PA signals measured atthe two locations as marked in Fig. 2(a), where thesmallest resolvable bar spacing is 7.82 �m (group 7,element 1). Hence, the lateral resolution is betterthan 7.8 �m. Considering the influence of the refrac-tion at the water surface, we estimated the theoreti-cal optical focal diameter to be 7 �m. This experi-mental result shows 11% degradation. In tissueimaging, optical scattering reduces the spatial reso-lution; however, when the sample is optically thin,the deterioration of the lateral resolution is insignifi-cant [8]. The lateral resolutions along the x and yaxes shown here are not identical, which is possiblydue to optical misalignment. It can be solved by abetter alignment of the objective lens and the twogalvanometer mirrors.

To quantify the axial resolution, a typical A-line[Fig. 2(d)] was used as an approximation of the 1Daxial point-spread-function (PSF). We first shiftedthe 1D PSF by a distance and added it to the originalPSF. Then the Rayleigh criterion was applied to thesum. The smallest shifted distance that still allowsresolving the two peaks (by 10% amplitude differ-ence) was considered as the resolution along that di-rection. Hence, the estimated axial resolution isaround 50 �m, which is primarily determined by theultrasonic center frequency and bandwidth [6].

In vivo imaging of microvasculature in the ear of aSwiss Webster mouse (Charles River Labs; body

Fig. 2. Quantification of the FOV and the spatial resolu-tions. (a) LSOR-PAM image of the printed mesh grid. (b)LSOR-PAM image of the USAF resolution target. (c) PAvalues at the cross sections highlighted by a-a and b-b in(b). (d) Estimate of axial resolution from a typical A-line ac-quired from blood vessel. sd, shift distance.

weight: 20 g) was acquired. Before experiments, the

June 15, 2009 / Vol. 34, No. 12 / OPTICS LETTERS 1773

hairs on the right ear were gently removed usingcommercial human hair-removing lotion [Fig. 3(a)].During experiments, the animal was placed on ahomemade animal holder, and a mixture of 1% isof-lurane with medical grade oxygen was ventilated tothe animal through a commercial nonrebreathing an-esthesia system (M3000, LEI Medical) at a flow rateof 1 l /min to keep the animal motionless. Unlikeother work [9] no optical clearing agent was appliedto the skin here. All experimental animal proceduresconformed to the laboratory animal protocol ap-proved by the Animal Care and Use Committee of theUniversity of Wisconsin–Milwaukee.

The LSOR-PAM image of the microvasculaturewithin the marked region in Fig. 3(a) is shown in Fig.3(b). As previously demonstrated [8], OR-PAM is ableto resolve capillary network with high quality and of-fers the potential to study associated physiologicalchanges in the capillaries. Although LSOR-PAM haslower spatial resolution comparable with OR-PAM, itshows comparable image quality in imaging mi-crovascular structures. Up to seven orders of vesselbranching can be observed, and the smallest vessel isimaged by a single pixel in the image. Parallel veinsand arteries, as pointed out in Fig. 3(b), are clearlyobserved.

The laser exposure was 84 mJ/cm2 at the opticalfocus. Although this value is higher than the ANSI la-ser safety limit �20 mJ/cm2� at 580 nm [10], it is stillwell below the reported skin damage threshold [9].

The average CNR in Fig. 3 is 30 dB, which is lowerthan that from the OR-PAM image. Two reasons con-tribute to the reduced CNR: (1) A flat ultrasonictransducer was used in LSOR-PAM, while the previ-ous OR-PAM used a strongly focused ultrasonictransducer. Since geometrical focusing increasesultrasonic detecting sensitivity and rejects unwantedsignals from the out-of-focus regions, the confocal de-sign in OR-PAM greatly improves CNR. A focusedultrasonic transducer can be employed in LSOR-PAM

Fig. 3. (Color online) LSOR-PAM image of vasculature ina Swiss Webster mouse ear in vivo. (a) Photograph of theear. (b) LSOR-PAM image of vasculature within themarked region in (a). Two pairs of parallel veins and arter-

ies are highlighted by 1 and 2.

to increase the CNR, as well; however, a trade-off be-tween the CNR and the FOV needs to be reached, be-cause the focused transducer has a reduced beam di-ameter. (2) The distance between the ultrasonicdetector and the sample in LSOR-PAM �30 mm� ismuch longer than that in OR-PAM �5 mm�, whichleads to lower detected PA signal amplitude owing tothe decay of a propagating spherical PA wave.

The advantages of LSOR-PAM stem from severalaspects. First, removing the bulky load (all the opti-cal and ultrasonic components) that used to be trans-lated together in OR-PAM can potentially achievehigher scanning speed. For imaging systems withlow-laser PRR where the data acquisition speed isprimarily limited by the PRR, the use of optical scan-ning is not critical; however, when the PRR is higherthan 10 kHz, the bottleneck for increasing the dataacquisition speed is the scanning system. Hence, op-tical scanning presents an opportunity to furtherspeed up the data acquisition. Second, optical scan-ning allows sophisticated scanning patterns to beperformed, such as a complete raster scanning fol-lowed by localized measurements on a single vessel.Lastly, the optical delivery and scanning mechanismin LSOR-PAM are similar to most existing optical mi-croscopic modalities, which lay the foundation for fu-ture integration of LSOR-PAM to other modalities.

This work was supported in part by the Universityof Wisconsin–Milwaukee startup fund and by a grantfrom The Lynde and Harry Bradley Foundation. Theauthors would like to thank Michelle Schoenecker forproofreading the manuscript.References

1. D. A. Sipkins, X. Wei, J. W. Wu, J. M. Runnels, D. Côté,T. K. Means, A. D. Luster, D. T. Scadden, and C. P. Lin,Nature435, 969 (2005).

2. C. A. Patil, N. Bosschaart, M. D. Keller, T. G. vanLeeuwen, and A. Mahadevan-Jansen, Opt. Lett. 33,1135 (2008).

3. S. Tang, T. B. Krasieva, Z. Chen, and B. J. Tromberg, J.Biomed. Opt. 11, 020502 (2006).

4. A. A. Oraevsky and A. A. Karabutov, in BiomedicalPhotonics Handbook, T. Vo-Dinh, ed. (CRC Press,2003), Vol. PM125, Chap. 34.

5. X. Wang, Y. Pang, G. Ku, X. Xie, G. Stoica, and L. V.Wang, Nat. Biotechnol. 25, 114 (2003).

6. H. F. Zhang, K. Maslov, and L. V. Wang, Nat. Protocols4, 797 (2007).

7. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, Nat.Biotechnol. 24, 848 (2006).

8. K. Maslov, H. F. Zhang, S. Hu, and L. V. Wang, Opt.Lett. 33, 929 (2008).

9. V. P. Zharov, E. I. Galanzha, E. V. Shashkov, N. G.Khlebtsov, and V. V. Tuchin, Opt. Lett. 31, 3623 (2008).

10. Laser Institute of America, American NationalStandard for Safe Use of Lasers ANSI Z136.1-2007

(American National Standards Institute, Inc., 2007).