We report on a sub-cellular resolution photoacoustic microscopy (PAM) system that employs microcavity synchronous parallel acquisition technique for detecting the weak photoacoustic (PA) signal excited by a modulated continuous wave (CW) laser source. The gas microcavity transducer is developed based on the fact that the bulk modulus of the gas is far less than the solid and the change of the air-gas pressure is inversely proportional to the gas volume, making it extremely sensitive to the tiny PA pressure wave. Besides, considering PA wave expends in various directions, detecting PA signals from different position and adding them together can increase the detecting sensitivity and the signal to noise ratio(SNR), then we employs two microphone to acquire PA wave synchronously and parallelly. We show that the developed PAM system is capable of label-free imaging and differentiating of the hemoglobin distribution within single red blood cells under normal and anemia conditions.
©2012 Optical Society of America
Photoacoustic microscopy (PAM) is a recently developed photoacoustic (PA) imaging technique capable of label-free and functional imaging of biological tissues with a cellular and subcellular spatial resolution [1–5]. In the current PAM, it is necessary to use a short pulsed laser source to generate the photoacoustic signal from the tissue that is strong enough so that the traditional PA transducers, e.g., piezoelectric ceramic transducer  and polyvinylidene fluoride needle hydrophone [7, 8], can be used for detection. However, probably because the PA signal generated by the continuous wave (CW) laser source is too weak to detect by the currently commercial available ultrasonic PA transducers, we need a special designed transducer to achieve PAM in cellular or even subcelluar scale with CW laser. It’s an attractive aspect of CW-mode PA transducer that PAM can be compatible to the standard laser scanning confocal microscopy (LSCM) system, facilitating the system integration for multiple functional imaging of the sample with both the optical absorption and scattering contrasts . While the pulse-mode PAM is hard to be integrated to standard LSCM, because the common laser source of LSCM is CW laser.
The most important thing for CW-mode PAM is to increase the detection sensitivity so as to effectively improve the contract of PAM image, because the PA signal induced by CW laser is too weak. There are some special designed CW-mode transducers have been reported [9–11]. In our previously research, we had reported a microcavity PA transducer which has the enough sensitivity to achieve CW-mode PAM. In order to have subcellular structural and functional PAM images, it’s necessary to further improve the sensitivity of CW-mode PA transducer. As we know, the sensitivity of transducer is proportional to its capture area, the larger capture area the higher sensitivity. Employing multiple sensors is equivalent to using a transducer with large capture area. In this paper, we report on a microcavity synchronous parallel acquisition technique which employed multiple sensors to detect weak CW-mode PA signal and achieved subcellular resolution imaging with higher contract and signal-to-noise ratio (SNR).
2. Microcavity synchronous parallel acquisition technique
2.1 Principle of multi-sensors microcavity PA transducer
The mechanism of our microcavity PA microscopy is based on Rosencwaig-Gersho theory, when chopped light impinges on a solid sample in an enclosed cell, an acoustic signal is produced within the cell, which is then transported to the microphone. According to the Rosencwaig-Gersho theory [12, 13], any light absorbed by the solid is converted, in part or in whole, into heat by nonradiative deexcitation processes within the solid. Consider a semi-spherical cell with a radius R, as shown in Fig. 1 . The solid sample is mounted inside the semi-spherical cell so that its front surface is exposed to the gas within the cell and its back surface is against an optical window which has poor thermal conductor. When a modulated illuminated light focuses on a point of the solid sample, an acoustic signal arises from the periodic heat flow from the solid to the surrounding gas. The periodic diffusion process produces a periodic temperature variation in the gas. According to the Rosencwaig-Gersho theory , the temperature in the gas attenuates rapidly to zero with increasing distance from the surface of the solid. At a distance of only, whereis the thermal diffusion length, the periodic temperature variation in the gas is effectively fully damped out. Thus, we can define a boundary layer, as shown in Fig. 1, whose radius is(0.2mm at = 2500Hz), and maintain to a good approximation that only this radius of gas is capable of responding thermally to the periodic temperature at the surface of the sample . Because of the periodic heating of the boundary layer, this layer of gas expands and contracts periodically and thus can be thought of as acting as an acoustic piston on the rest of the gas in the semi-spherical cell, producing an acoustic pressure signal that travels through the entire semi-spherical cell.
If we assume that the rest of the gas in the semi-spherical cell responds to the action of this piston adiabatically, then the acoustic pressure in the cell due to the displacement of this gas piston is derived from the adiabatic gas law.
Then, the Eq. (2) can be expressed as:
Obviously, under the same condition of the temperature change (rise or drop), the incremental pressureis inversely proportional to the volume of the cell, in other words, the pressure response of the transducer is significantly greater for the small gas-volume than that of the big gas-volume. Thus, reducing the volume of cell can effectively increase the detection sensitivity of PA transducer.
To further improve the detection sensitivity, a synchronous parallel acquisition technique is introduced to our microcavity PA transducer. As descripted above, the detection sensitivity is proportional to the capture area of transducer. Considering PA wave expends spherically, setting several sensors at different direction but at the same wave front, as shown in Fig. 1, is equivalent to using a spherical transducer with large capture area. Because the sensors response the pressure of PA wave synchronously, the total output of microcavity PA transducer with N sensors becomes N times as the output of one sensor.
2.2 Realization of microcavity synchronous parallel acquisition technique
Theoretically, the more sensors can achieve higher sensitivity, but the structure of transducer is more complicated and hard to install. Here, we employed two sensors to illustrate the feasibility of microcavity synchronous parallel acquisition technique. But it can be observed that more sensors can be equipped on our transducer in the future.
Figure 2(a) illustrates the realization of the microcavity synchronous parallel acquisition technique. It consisted of a microcavity, a resonant cavity, a microchannel and two microphones. The microcavity was with a diameter of 0.5 mm and a height of 1mm where the sample cell was situated and attached to its opening. A resonant cavity with the same dimension as the microcavity was introduced into the design, with its opening attached with two microphones for detecting the pressure waves generated at the sample. The microcavity and resonant cavity were bridged by a microchannel, with a diameter of 0.25mm and a length of 200 mm. The microchannel was used to minimize the gas-volume contained within the microcavity PA transducer, thus further enhancing the sensitivity of the transducer. The microphone was specially designed that has a bandwidth of 2.5 kHz, a sensitivity of 10mv/Pa, a capture area of about 3mm2 and a bias voltage of 2V. The whole transducer was a microphone-base device and was air-tight.
The schematic of the experimental PAM setup that uses a continuous wave laser as the irradiation source is illustrated in Fig. 2(b), where the CW laser is an Argon ion laser with the output wavelength of 514.5 nm. We modulated the output beam by an optical chopper. The modulated beam was then delivered to a field flattening objective with a magnification power of x60 (NA = 0.75) that focused the beam to a tight spot on the sample. The sample (RBCs) was placed at the surface of a microscope cover-slide, facing the microcavity of the PA transducer. The PA transducer was fixed on a motorized translation stage, which was used to fine tune the position of the transducer, so that the focused spot of the probe beam is situated at the sample. The amplitude of PA signal at each focused spot is determined by the absorption coefficient at that spot. Then, to achieve imaging, a 2D galvanometer scanner (6231C, Cambridge Technology) was used before the objective lens to scan the probe beam in the x-y direction.
With the beam diameter of 8 mm before the objective lens, the lateral resolution of the system i.e., the diffraction-limited optical focal diameter was calculated to be 0.42 μm at 514 nm. The system lateral resolution was experimentally checked by the Guangzhou Municipal Standard Bureau, reporting a measured value of 1.25μm, which is worse than the diffraction-limited estimation, likely due to optical aberration in the system. The resulted PA signals were then sensed by the two microphones. Because the amplitude of PA signal was too low to be detected directly, the detected PA signals were first amplified by a preamplifier (Stanford Research Systems SR550), and then sent to a lock-in amplifier (Stanford Research Systems SR830) for demodulation and further amplification. Benefiting from frequency-selecting characteristics of lock-in amplifier, we can pick up the amplitude of PA signal from noise. Consequently, the SNR and the sensitivity can be increased. Finally spatial distribution of the amplitude of PA signal was acquired and stored by a data acquisition board (PCI6115, National Instrument) in the computer.
3. Results and discussions
Considering hemoglobin within RBCs is the major source of endogenous optical absorption in biological tissue, we elected to use RBCs to demonstrate the imaging capability of the developed PAM at sub-cellular level resolution. To show whether PAM is sensitive enough to differentiate the hemoglobin distribution in the RBC, we conducted experiments on the anemia RBC cells. The blood/RBCs were drawn from the anemic patients attending the clinic at Guangzhou Zhujiang Hospital and then prepared by the hematologists at the hospital for PAM imaging. For this study, the RBCs were smeared directly onto the cover glass, which was then placed against transducer. Note that all the RBCs used in this study were already accurately diagnosed using standard clinical diagnostic procedures, with their degree of anemia known before the PAM imaging. Three kinds of RBCs, i.e. normal, mild and serious anemia RBCs, were prepared for PA imaging test.
3.1 Improvement from microcavity synchronous parallel acquisition technique
Figure 3(a) and Fig. 3(b) illustrates the PAM image from the normal RBCs, acquired by raster scanning of the focused beam spot on the sample with and without the microcavity synchronous parallel acquisition technique. The image had a field of view of 40x40 μm2. The focus depth of PAM using CW laser is not short enough to distinguish the pit of normal RBCs which is donut shape. But for the normal RBCs, the hemoglobin is distributed uniformly, thus the PAM image obtained from each single RBC should appear as solid circular, which is confirmed from our experiments. The SNR of PAM using microcavity synchronous parallel acquisition technique is 34dB, compared to the one without such technique is 28dB, when the input power of laser is 3mw. Then, the contrast of image is enhanced accordingly due to the addition of the PA signal.
3.2 Quantitative measurement for the degree of anemia
Next we tested whether the developed PAM system is capable of differentiating the hemoglobin distributions within anemia RBCs between the mild and serious anemia. Pathologically, the anemia RBCs have the hemoglobin distributed towards the edge of the cell, leaving the central region appearing more or less transparent in the OM images, depending on the seriousness of the anemia. The more severe the anemia, the more the hemoglobin is distributed to the edge.
Figure 4(a) gives the PAM image obtained from the seriously anemia cells. Because the hemoglobin is distributed around the edge of the cell, the optical absorption due to the hemoglobin is strongest in the cell edge. On the contrary, the central region of anemia cell has almost no hemoglobin, thus, the optical absorption there is very weak, leaving an appearance of donut shape for the RBCs in the PAM image. This result matched very well with the image photographed by the optical microscope (OM) (Fig. 4(c)). To show the degree of the hemoglobin towards the edge, we plotted in Fig. 4(b) the amplitude of the PA signal along the central line of one serious anemia cell (marked in Fig. 4(a)). The diameter of the cell is measured at d1 = ~8.5μm and the diameter of the central part of cell is d2 = ~6μm, giving an area ratio of 49.8% between the central region and the whole cell. Note that d1 was the FWHM value of the normalized PA signal, while d2 was provided by least-square fitting of the normalized PA signals at the central region to a reversed Gaussian function, and then taken the FWHM value of it. We evaluated n = 15 cells from this cohort of anemic cells. The average area ratio was 50 ± 10%. Such area ratio calculated from the single cells in the PAM image may be used to evaluate the degree of anemia in the red blood cells. Figure 4(d) shows the amplitude of optical opacity from the same cell observed from the optical microscope. Here, the contrast of the signal is low, hardly useful to determine hemoglobin distribution of the anemia cells.
Figure 4(e) shows the PA image of the mildly anemia cells. Similarly, Fig. 4(f) gives the amplitude of PA signal plotted along the central line of one anemic cell (marked in Fig. 4(e)). d1 is measured at ~8.5μm and d2 is ~2.5μm, giving the area ratio of 8.7% between the central region and the whole cell. Again, we evaluated n = 15 cells from the mild anemic cells, that gave the averaged area ratio of 9 ± 8%. This value is about lower than that of the seriously anemic cells, indicating less seriousness of anemia for the cells examined. In the case of the mild anemia, the OM image (Fig. 4(g)) is similar to that of the seriously anemia cells, where it is difficult to distinguish the difference between them. Thus clinically, using optical microscope alone would not be able to determine the degree of the anemia in the RBCs. These results demonstrate the usefulness of PAM in the label-free imaging of anemic RBCs.
We have developed a PAM imaging system with a resolution at sub-cellular level by the use of a CW laser as the excitation light source. A microcavity synchronous parallel acquisition technique has been applied in detection of the weak thermoelastic pressure waves generated by the CW laser. This technique can effectively enhance the detection sensitivity and improve the contrast of PAM images. We have demonstrated the feasibility of the developed PAM imaging system in label-free imaging and differentiation of the hemoglobin distributions within the RBC cells under normal, mild and severe conditions of anemia.
This work is supported by the National Natural Science Foundation of China (Grant No. 60877068), the Research Fund for the Docoral Program of Higher Education of China (Grant No.20104407110008).
References and links
2. D. K. Yao, K. Maslov, K. K. Shung, Q. F. Zhou, and L. V. Wang, “In vivo label-free photoacoustic microscopy of cell nuclei by excitation of DNA and RNA,” Opt. Lett. 35(24), 4139–4141 (2010). [CrossRef] [PubMed]
8. Y. Q. Lao, D. Xing, S. H. Yang, and L. Z. Xiang, “Noninvasive photoacoustic imaging of the developing vasculature during early tumor growth,” Phys. Med. Biol. 53(15), 4203–4212 (2008). [CrossRef] [PubMed]
9. Z. L. Tan, Z. L. Tang, Y. B. Wu, Y. F. Liao, W. Dong, and L. N. Guo, “Multimodal subcellular imaging with microcavity photoacoustic transducer,” Opt. Express 19(3), 2426–2431 (2011). [CrossRef] [PubMed]
10. R. S. Quimby, “Real‐time photoacoustic microscopy,” Appl. Phys. Lett. 45(10), 1037 (1984). [CrossRef]
12. A. Rosencwaig and A. J. Gersho, “Photoacoustic effect with solids: a theoretical treatment,” Science 190, 556–557 (1975).
13. A. Rosencwaig and A. J. Gersho, “Theory of the photoacoustic effect with solids,” J. Appl. Phys. 47(1), 64–69 (1976). [CrossRef]
14. A. Rosencwaig, Photoacoustics and Photoacoustic Spectroscopy (Wiley, 1980), Chap.9.