Using a line scan camera and an acousto-optic deflector (AOD), we constructed a high-speed confocal laser line-scanning microscope that can generate confocal images (512×512 pixels) with up to 191 frames/s without any mechanically moving parts. The line scanner consists of an AOD and a cylindrical lens, which creates a line focus sweeping over the sample. The measured resolutions in z (depth), x (perpendicular to line focus), and y (direction of line focus) directions are 3.3 µm, 0.7 µm and 0.9 µm, respectively, with a 50× objective lens. This confocal microscope may be useful for analyzing fast phenomena during biological and chemical interactions and for fast 3D image reconstruction.
© 2005 Optical Society of America
Since the first development by Minsky in 1961 , various types of confocal microscope [2–10] have been produced and used in many fields, including biology and medicine. The confocal microscope enables us to noninvasively obtain ultra-high-resolution 2D or 3D images, which is important for quantitative studies of biological samples.
When rapidly changing phenomena need to be monitored throughout 2D planes or 3D volumes, the scan speed or frame rate becomes critical . Also, the advantages of high-speed imaging are evident when in-vivo images are to be obtained, such as with the imaging of human cornea [3,4]. The frame rates of the current commercially available confocal microscopes are the video rates (~24 frames/s). Actually, much higher scan speed is attainable with the Nipkow disk  (~360 frames/s). However, the image-acquisition speed of the Nipkow disk-based system is limited to the video rate mainly because of the low optical throughput. Several confocal microscope designs (e.g., a polygon mirror  or a resonant galvanometer ) have been introduced with frame rates slightly faster than the video rate. However, these confocal microscopes are expensive and employ complicated optical and electronic designs for synchronization between the horizontal resonant scanner and the vertical galvanometer. One easy way to increase the frame rate is to reduce the number of scan lines proportionally. One group reported that they obtained the frame rate of 120 frames/s for an image of 256×240 pixel size. For a 512×512 image, the same system produced 30 frames/s .
To improve the frame rate even further, developers created line-scanning systems, which use a line focus instead of a single point focus. In this way, the scanning over the sample is performed only in one direction, and the frame rates can be improved significantly. An oscillating mirror  or a double-sided mirror  was employed for line scanning over live tissues. Recently, a commercial line-scanning confocal microscope (LSM 5, Carl-Zeiss International), which can operate at 120 frames/s, was introduced. One drawback of this type of line-scanning system is that it is confocal only in one dimension, which is perpendicular to the line focus. As a result, the resolution is slightly worse than that of a regular confocal microscope. However, the line-scanning system still has the depth-resolving capabilities, and the resolution is superior to conventional microscopes. Because these line-scanning confocal microscopes use the schematic in which only one half of the objective aperture is illuminate, the lateral resolution becomes deteriorated [3,4,11]. To overcome this problem, we designed a system in which the objective is filled completely.
In this paper, we present a high-speed confocal line-scanning microscope (HSCLM) that is faster than the video rate and has a relatively simple optical design and no mechanically moving parts in its scanning system. Our system can produce confocal images with up to 191 frames/s for 512×512 pixels.
2. Optical system for line scanning
The schematic of the HSCLM is shown as Fig. 1(a). While most line-scanning confocal microscopes use halogen or Xe lamps as a light source [5,6], we use a HeNe laser to obtain sufficient power in a diffraction-limited line. The laser beam is expanded by two convex lenses to fill the aperture of an objective lens, while most line-scanning confocal microscopes use half-filled objective aperture.
We use a polarizing beam splitter (PBS) and a quarter-wave plate (WP) to maximize the laser power collected by the detector. A cylindrical lens (CL1) and an acousto-optic deflector (AOD, Isomet, Inc., LS55-V) in the dashed rectangle in Fig. 1 are used as a line scanner, which ultimately converts the point source into a line source parallel to the y-axis over the sample. The detailed ray-traces of the line scanner are shown in Figs. 1(b) and 1(c). The circular beam is initially changed into an elliptical beam parallel to the x-axis by CL1. The AOD performs the sweeping in the direction represented by the angular arrow. Use of the AOD has several advantages over other scanners: stability, easy access, quietness, and variable frequency. The first-order diffracted beam is used as a light source. The first-order diffraction efficiency of the AOD is approximately 80% and the maximum diffraction angle is 2.75°. The elliptical beam diffracted by the AOD is transferred into the objective lens (50×, NA=0.8) by two lenses. The objective lens then forms two perpendicular line foci in two axial locations separated by a small distance, since our optical system is anamorphic. As the major axis of the incident elliptical beam (before the objective lens) is parallel to the x-axis and its length is close to the lens aperture, the focused beam in the sagittal plane (SP) is parallel to the y-axis. In this way, the width of the focal line (in x-axis) is close to the size of a diffraction-limited spot. Provided that the focused line beam is collimated after the objective lens, the length of the focused beam can be given as l=a·f 3·f ob (f CL1·f 4), where a is the radius of the circular beam measured after the lens L2 and fs are the focal lengths of the lenses. Since the beam is not collimated, this formula gives only an approximate value. Two convex lenses (L3 and L4) are used to guide the beam into the objective lens and keep the beam stationary at the objective lens [see Fig. 1(b) and (c)].
The beam that is reflected from the sagittal plane (SP) becomes circular again after being reflected from the PBS, and then it enters the detection arm. A cylindrical lens (CL2) in the detection arm changes the circular beam to the elliptical beam parallel to the z-axis. Another cylindrical lens (CL3) makes the focused beam conjugate to that in the SP. The light reflected from the SP is selectively detected with adjustment of a cylindrical lens (CL3) and a slit (S).
A line-CCD (ATMEL, Inc., AVIVA SM2) camera is employed to obtain the confocal images. As the line acquisition rate of the L-CCD camera is 98 kHz/line, the image acquisition speed can reach up to 191 frames/s for 512×512 pixels. This inexpensive and fast line-CCD camera is available from several manufacturers. The readout signal is sent to the frame grabber in a personal computer to form 2D images. The computer-controlled micro-actuator (resolution =1 µm) is used to move the sample on a stage along the z-axis. All the confocal images in this paper were taken with the frame rate of 191 frames/s. We can monitor the confocal images with a computer monitor of the same frame rate.
Since the scan speed of the AOD is externally controlled by analog signal, the only limit for higher frame rate is the CCD-camera read-out time. Therefore, the frame rate can be further improved when faster CCD is available.
3. Results and discussion
Figure 2 shows images of an air-force target (Edmund optics, 36408) taken by our HSCLM. All the confocal images in this paper are raw images which were taken with 191 frames/sec and they were not treated by digital image processing techniques just yet. Actually, the resolution in x-axis is slightly better than that in y-axis. This is because the direction of the focused line on the sample is parallel to y-axis, therefore, the resolution in y-direction is approximately the same as that of conventional optical microscope . This is the limitation of all the line-scanning confocal microscopes as explained before.
The lateral resolutions in x- and y-directions were also measured by obtaining the line spread function of the edge response  in Fig. 2, and the results are shown in Figs. 3(a) and 3(b), respectively. The full width at half maximum (FWHM) of the line spread function in x-direction is approximately 0.7 µm while that in y-direction is approximately 0.9 µm. As mentioned above, the resolution in y-direction is slightly worse than in x-direction owing to the intrinsic limitation of the line-scanning method.
We tested the axial response of our system by measuring the power of light passing through the slit as a mirror was moved across the focal sagittal plane (Fig. 4), which followed the procedure described in the work by Corle et al. . A usual confocal signal was observed as shown in Fig. 4. Positive values of z represent the increased distance between the objective lens and the mirror. The FWHM is approximately 3.3 µm.
Figure 5(a) shows images, taken by a scanning electron microscope, of a micrometer-size object fabricated by optical lithography. Using an atomic force microscope, we separately found that this object has a step structure with a height of 3.33 µm. The same object was imaged using a conventional optical microscope, and the results are shown in Figs. 5(b) and 5(c). As expected, the image becomes blurred when the sample stage is moved across the focal plane.
To demonstrate the depth resolving capabilities of our HSCLM system, we obtained images of the same object by using the HSCLM, and the results are shown in Figs. 6(a)–6(c). Each picture was taken as the stage was moved by 1 µm in depth direction. It is clear that the 3D feature of the object can be resolved with our system. As shown in Figs. 6(a)–6(c), the confocal images contain noise similar to interference artifacts, which might be caused by inappropriate slit alignment. These noises may be eliminated by optimizing the slit location or by digital image processing. Figure 6(d) shows the reconstructed 3D image captured with commercial software.
Since our system is constructed mainly to measure the specularly reflected light inside transparent media such as cornea, several modifications have to be made so that it can be used for general applications. First of all, since AOD is used for sweeping the laser beam, our system does not work well for collecting fluorescence light because of the dispersion in the AOD. To correct this problem, a Galvano mirror may be used in place of the AOD. Second, since the input beam polarization has to be maintained to be collected by the detector, the system has to be modified when the scattering sample is to be imaged. To solve this problem, the quarter-wave plate may be removed and the nonpolarizing beam splitter may replace the PBS.
We present what we believe is a novel high-speed confocal line-scanning microscope, which can be devised with rather simple and low-cost configuration. Using a line CCD camera and an AOD, frame rate can be significantly improved. The confocal images of 512×512 pixels can be obtained at a rate of 191 frames/s and with no mechanically moving parts. This high-speed confocal line-scanning microscope may be useful for analyzing fast reactions and for fast 3D image reconstruction in the clinical and biological applications.
This study was supported by a grant of the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (02-PJ3-PG6-EV07-0002).
References and links
1. M. Minsky, “Microscopy apparatus,” U.S. patent 3,013,467 (December 1961).
2. R. Y. Tsien and B. J. Bacskai“Video-rate confocal microscopy,” in Handbook of Biological Confocal Microscopy, J. B. Pawley (Plenum, New York, 1995), pp. 459–478.
5. G. Q. Xiao, T. R. Corle, and G. S. Kino, “Real-time confocal scanning optical microscope,” Appl. Phys. Lett. 53, 716–718 (1988). [CrossRef]
6. M. Rahadhyaksha, R. R. Anderson, and R. H. Webb, “Video-rate confocal scanning laser microscope for imaging human tissues in vivo,” Appl. Opt. 38, 2105–2115 (1999). [CrossRef]
7. J. Cushion, F. N. Reinholz, and B. A. Patterson, “General purpose control system for scanning laser ophthalmoscopes,” Clin. Exp. Ophthalmol. 31, 241–245 (2003). [CrossRef]
8. W. B. Amos and J. G. White“Direct view confocal imaging systems using a slit aperture,” in Handbook of Biological Confocal Microscopy, J. B. Pawley (Plenum, New York, 1995), pp. 403–415.
9. Y. S. Sabharwal, A. R. Rouse, L. Donaldson, M. F. Hopkins, and A. F. Gmitro, “Slit-scanning confocal microendoscope for high-resolution in-vivo imaging,” Appl. Opt. 38, 7133–7144 (1999). [CrossRef]
11. R. H. Webb, “Confocal optical microscopy,” Rep. Prog. Phys. 59, 427–471 (1996). [CrossRef]
12. C. J. R. Sheppard and X. Q. Mao, “Confocal microscopes with slit apertures,” J. Mod. Opt. 25, 1169 (1998).