Open Access
Add to BibSonomyAbstract
We describe a three-dimensional microscopy technique based on spectral and frequency encoding. The method employs a wavelength-swept laser to illuminate a specimen with a spectrally-dispersed line focus that sweeps over the specimen in time. The spatial information along each spectral line is further mapped into different modulation frequencies. Spectrally-resolved detection and subsequent Fourier analysis of the back-scattered light from the specimen therefore enable high-speed, scanner-free imaging of the specimen with a single-element photodetector. High-contrast, three-dimensional imaging capability of this method is demonstrated by presenting images of various materials and biological specimens.
© 2015 Optical Society of America
1. Introduction
Two-dimensional (2D) image sensors such as charge-coupled device (CCD) and complementary-oxide-semiconductor (CMOS) detectors are commonly employed to construct optical imaging systems. Advances in the image sensor technologies have enabled imaging with improved acquisition speed and broader spectral responses. However, the use of single-element detectors is beneficial or inevitable in some applications, especially for the cases where high-sensitivity segmented detectors are unavailable or prohibitively expensive (e.g., imaging in UV, mid-infrared and terahertz spectral regimes).
Laser-scanning microscopy (LSM) such as confocal and nonlinear microscopes is a representative single-pixel imaging technique [1–4]. In LSM, a tightly focused probe beam is scanned through a specimen in three-dimensions, and a signal at each position is de-scanned and measured with a single-element detector. The LSM provides high-resolution, high-contrast images of the specimens, but its operation typically compromises the acquisition time due to the necessity of mechanical scanning of probe beam or specimen, along with serial data acquisition. LSM schemes based on multiple foci generated with a microlens array and high-sensitivity 2D image sensors have been developed [5, 6]. These methods, however, are costly to implement, and require dedicated beam scanning devices and high-sensitivity 2D image sensors.
Spectrally encoded confocal microscopy (SECM) represents one of the high-speed line-scanning microscopy techniques [7, 8]. SECM utilizes a combination of diffraction grating and objective to illuminate a specimen with spectrally dispersed foci, so that each wavelength maps onto a position along one axis in the specimen. By measuring the reflected light via spectrally-resolved detection, the sample information along the axis is captured. However, it requires beam scanning in the other direction for 2D image construction. Another strategy for single-pixel imaging techniques exploits patterned illumination of the probe beam, while the signal is collected by a single-pixel photodetector [9–11]. For instance, in compressed imaging, a series of predetermined illumination patterns is projected onto a specimen, collecting the transmitted light with a single-element detector. The sparsity of the specimen information in certain domains, in combination with a system model, enables this method to obtain the 2D images [12].
A recently introduced single-element imaging method employed a spinning modulator to map spatial coordinates into different modulation frequencies [13]. In this scheme, an illumination beam is focused in a line on a specimen, along which the spatial coordinates are modulated with linearly varying frequencies. The spatial information along the line can then be obtained by the Fourier transform of the time-domain signals. Yet, scanning either the line focus or the sample is required to obtain a 2D image.
Here, we describe a novel microscopy technique, termed frequency- and spectrally-encoded confocal microscopy (F-SECM), which enables scanner-free, high-speed imaging of materials and biological specimens with a single-element detector. This method combines the features of spatially chirped modulation with spectral encoding. A rapid wavelength-swept light source illuminates a specimen with a line focus through a diffraction grating. As the diffraction angle varies with the wavelength, the focused line scans over the specimen in time as the wavelength sweeps. The spatial information along each spectral line is further encoded into different modulation frequencies. Hence, Fourier analysis of the spectrally resolved time-domain signals enables scanner-free imaging of the specimen with a single-pixel photodetector. We also demonstrate depth-resolved imaging capability of F-SECM by implementing F-SECM in reflection mode and placing a slit aperture in the detection path. Due to its simplicity and cost effectiveness, the technique could potentially be applied to various applications that call for high-speed imaging in various spectral regimes.
2. F-SECM implementation
2. 1 System configuration
Figure 1 illustrates a schematic of F-SECM. A 1.31-μm high-speed wavelength-swept laser with a spectral bandwidth of 110 nm (Axsun Technologies Inc.) was employed as the light source. The time-averaged optical power of the laser was measured to be 20 mW. The wavelength tuning operated at a repetition rate of 50 kHz. The output of the laser was connected to a 10/90 fiber coupler, and 10% of the light was directed to a fiber Bragg grating (FBG). An optical pulse was produced when the wavelength of the light source matches the FBG reflection wavelength. This pulse was detected and converted into a transistor-transistor logic (TTL) signal that is coincident with the optical pulse. The TTL signal (TRIG) was fed to a high-speed digitizer (AlazarTech Inc., ATS9350) to trigger signal acquisition for each wavelength sweep.
Fig. 1 A schematic of F-SECM. Two-dimensional spatial information of a specimen is encoded by modulation frequencies and wavelengths of probe light. The back-scattered light from the specimen is measured by a single-element detector, and its spectrally-resolved detection and subsequent Fourier analysis enables two-dimensional image re-construction. The slit (S) in the detection path also allows depth-resolved imaging. SS: swept source, SS-OUTPUT: swept source output, FBG: fiber Bragg grating, SMF: single-mode fiber, C: collimator, L1-L5: achromatic spherical lenses, CL: cylindrical lens, FM: frequency modulator, BS: beam splitter, DG: diffraction grating, OBJ: objective lens, S: slit, PD: photodetector, DAQ: data acquisition system, TRIG: DAQ triggering signal, SIG: measurement signal
In the F-SECM probe arm, the light from the fiber was collimated with an achromatic doublet lens (C; Thorlabs Inc., F810APC-1310), and line-focused with a length of 14 mm on a patterned disc through an optical system composed of a cylindrical lens (CL; Thorlabs Inc., ACY254-150-B, f = 150 mm) and a lens L1 (Thorlabs Inc, AC254-300-C, f = 300 mm). The patterned disc was set to rotate at 3000 rpm, serving as a frequency modulator (FM) that altered the light intensity with different frequencies along the radius. Detailed description on the frequency modulator is given in Sec. 2.2. The frequency-modulated beam was then directed on a reflection diffraction grating (Edmund Optics, 600 lines/mm, 1600-nm NIR ruled grating) via the lens L2 (Thorlabs Inc, AC254-400-C, f = 400 mm), and diffracted as a function of wavelength. The diffracted beam then illuminated a specimen with a focused line via the objective (Nikon Corp., 20 × /0.5).
In the detection path, the light back-scattered from the specimen was collected by the same objective, and imaged onto a slit aperture (S; Thorlabs Inc., 15 μm × 3 mm) via the lens L3 (Thorlabs Inc., AC254-050-C, f = 50 mm). The magnification ratio from the sample plane to the slit was set to 6. The light was then de-magnified by a telescope composed of the lenses L4 (Thorlabs Inc., AC254-050-C, f = 50 mm) and L5 (Thorlabs Inc., A240TM-C, f = 8 mm) to be detected by a single-element photodetector (Thorlabs Inc., PDA10CF). The de-magnification ratio of the telescope was set to 1/6.25, so that the entire signal from the sample could be captured by the detector.
2.2 Frequency modulator
Various strategies can be employed to modulate the light intensity along the extent of the line focus [14]. In our case, we utilized a patterned rotating disc similar to that used by Futia et al. [13], as it provided a simple and cost-effective means to modulate the light intensity. However, we noted that for a disc pattern with continuous frequency distribution, imperfect centering of the printed pattern and motion jitter of the motor spindle could cause signal mixing between adjacent pixels, thereby significantly distorting the images. Hence, we designed the pattern to have discrete frequencies, rather than continuous frequency distribution, to minimize the effect of the motion jitter of the rotating disc. The expression for the pattern (m) is given by
where and denote the radial and angular coordinates in the disc, respectively, k0 is the center modulation frequency in cycles/rev, Δk is the modulation frequency ranges over the radius of 0 to R in the disc, and is the number of frequency steps.For this proof-of-concept experiment, we designed the pattern to have the frequencies from 11 to 460 cycles/rev over the radius of r = 13 to r = 45 mm, and each adjacent pixel was separated by 1 cycle/rev. As the FM rotated at a speed of 3000 rpm, the maximum modulation frequency range that could be achieved by the rotating disc was 550 to 23,000 Hz. In our case, the length of the line focus on the disc was ~14 mm, and the modulation frequency range of 7 - 17 kHz was used. This operating condition resulted in ~200 pixels along the frequency (y) axis.
The modulator was patterned by a UV printer with a resolution of 1280 × 1280 dpi (Mimaki Engineering Co., Ltd., UJF-6042). The pattern was printed on a 101.6-mm diameter glass substrate (Corning, Inc., Eagle 2000). An exemplary disc pattern is shown in Fig. 2. Note that the disc in Fig. 2 is composed of 40 frequency steps to provide a better visualization of the discrete frequency steps. In the experiment, a disc with 450 frequency steps was used.
Fig. 2 F-SECM frequency modulator. Light from the light source is focused into a line along the radius of the modulator. Rotation of the disc produced the light modulated with different temporal frequencies along the radius.
2.3 F-SECM signal acquisition and image reconstruction
Figure 3 shows a timing diagram for F-SECM signal acquisition during a single revolution of the frequency modulator (FM). Since the FM rotational speed was set to 3000 rpm and the wavelength tuning rate of the laser was 50 kHz, the corresponding FM rotational period (TFM) and the period of wavelength sweep (Tλ) were 20 msec and 20 μsec, respectively. Upon receiving the trigger pulse (TRIG) from the FBG filter for each wavelength sweep, the acquisition system recorded Nx = 1536 samples with an internal on-board clock (DAQ-CLK) at 100 MS/s. Since the number of wavelength sweeps during the single revolution of the disc was 1000, it can be seen that a line focus corresponding to λi illuminated the spatial positions at xi 1000 times during a single revolution of the modulation disc. Here, λi and xi are the i-th wavelength and the corresponding spatial coordinate in the x axis, respectively, with i < Nx = 1536. One can also note that for each wavelength sweep, the positions at xi were illuminated with a line focus modulated with different spatial frequency along the y axis, which is determined by the rotating frequency modulator. Therefore, for a given instant, the acquired sample corresponds to the signal at xi with a spatial frequency of uj, where uj is the j-th spatial frequency in the y direction with j < 1000. The spatial information in the y-axis can thus be obtained by taking the inverse Fourier transform of the signal acquired at xi during a single revolution of the disc.
Fig. 3 Timing diagram for F-SECM signal acquisition. TFM: FM rotational period, TRIG: 50-kHz data acquisition (DAQ) trigger signal, SS-OUTPUT: swept source output, DAQ-CLK: 100-MHz DAQ sampling clock, SIG: measured signal at a photodetector. Upon receiving the TRIG pulse for each wavelength sweep, the DAQ system acquired Nx = 1536 samples with the 100-MHz DAQ sampling clock (DAQ-CLK).
Figure 4 illustrates the F-SECM image reconstruction procedure. The acquired data during a single revolution of the disc was first arranged into two-dimensional matrix in the (λ, t) or (x, u) domains (Fig. 4(a)). Shown in Fig. 4(b) is the signal measured at a particular wavelength or x-coordinate indicated by the dashed lines. Note that this signal corresponds to the spatial frequency spectrum in the y axis, as indicated by the axis on the top of Fig. 4(b). The spatial information along the y direction could then be obtained by taking the inverse Fourier transform of the signal (Fig. 4(c)). By executing this process for all the wavelengths or x coordinates, a 2D image of the sample could be re-constructed (Fig. 4(d)). The dashed line in Fig. 4(d) corresponds to the spatial information found in Fig. 4(c).
Fig. 4 F-SECM image reconstruction procedure. (a) The F-SECM temporal signal acquired during a single revolution of the disc was sampled at each wavelength, so that it can be arranged into two-dimensional matrix in the (λ, t) and (x, u) domains. Here, u denotes the spatial frequency in the y axis. (b, c) The spatial information in the y axis or as a function of modulation frequency can then be obtained by taking the inverse Fourier transform of the signal at a particular wavelength. Presented in (d) is an example of the reconstructed F-SECM image.
2.4 F-SECM performance parameters
2.4.1. Field of view (FoV)
The field of view in the spectral axis () can be estimated as
where ƒOBJ is the focal length of the objective and is the total range of angles dispersed along the spectral axis at the back aperture of the objective [8]. For a diffraction grating with a period of Λ and a light source with a tuning range of , can be approximated to befor small . Here, the first-order diffraction only was considered. Using the parameters in our setup (nm and mm/lps), is estimated to be ~0.066 rad and the corresponding was found to be ~660.2 μm.The FoV in the frequency axis () can be obtained simply by projecting the line focus on the frequency modulator onto the sample plane through the imaging system composed of lens L2 and the objective. Since the length of the line focus at the modulation disc was 14 mm and the de-magnification from the disc to the sample plane was 1/40, was found to be 350 μm.
2.4.2. Lateral resolutions and the numbers of resolving points
In F-SECM, the lateral resolution in each axis is determined differently, due to its different optical arrangements and encoding schemes. In this Section, we examine the lateral resolutions and the number of resolving points in each axis.
The actual spatial resolution in the spectral axis (i.e., x axis) is determined by the poorest element among the quantities set by the effective spectral resolution and the slit confocal imaging setup in F-SECM [7]. The effective spectral resolution is, in turn, determined by the larger of the instantaneous linewidth of the light source or the spectral resolution of the grating. The spectral resolution of the diffraction grating () can be obtained as [8]
where is the center wavelength of the light source and D is the incident beam diameter along the grating. Evaluation of Eq. (4) with our F-SECM configuration yields 0.23 nm. Comparing this value with the instantaneous linewidth of the laser (~0.06 nm, according to the spec. from the laser manufacturer), it is clear that the effective spectral resolution in F-SECM is governed by the diffraction grating. The lateral resolution limited by the F-SECM spectral resolution can then be obtained as μm. The resultant number of resolving points is 440.We now examine the lateral resolution determined by the slit confocal imaging system in F-SECM. It has been well understood that the lateral response in a slit confocal microscope is almost as sharp as a confocal microscope in the slit axis, but the same as a conventional microscope in the other direction [15]. Therefore, the spatial resolution in the slit axis can be approximated as that for a confocal microscope with a circular pinhole, of which size is the same as the slit width. The effect of pinhole size on the spatial resolutions in confocal microscopes has been extensively studied [3, 16]. We employ the analytical results in [3], to estimate the FWHM lateral resolution of F-SECM in the slit axis. The analysis allows us to estimate the lateral and axial resolutions of confocal imaging systems based on the size of pinhole or slit on the sample plane in optical unit (OU). We first recall definitions of lateral and axial OU [17]:
where λ is the wavelength, n is the refractive index, α is the half of the angular aperture of an optical system, and r and z are the radial and longitudinal coordinates, respectively. Noting , Eqs. (5) and (6) can be re-written as:We now evaluate the effective slit size on the specimen plane in F-SECM. The effective slit size on the sample plane can be estimated by projecting the 15-μm slit onto the sample plane through the imaging optics and diffraction grating in the detection path. Since the light source in our setup is characterized by an instantaneous linewidth of 0.06 nm, we assumed a Gaussian light source centered at 1.31 μm with the FWHM spectral bandwidth of 0.06 nm for imaging the slit onto the specimen plane. A simple calculation found the slit width on the sample plane to be ~3.6 μm. The obtained value is larger than a simple geometrical estimation (i.e., 15 μm/6 = 2.5 μm) because the diffraction grating at the back focal plane of the objective shifts the slit image at each wavelength laterally on the sample plane, broadening the effective width of the slit image. This half width of the slit is then found to be ~4.3 in the lateral OU using Eq. (7). Using this value and referring to Fig. 3.1 in [3], the half of FWHM lateral response (, using the same notation in [3]) was found to be 1.6 OU. The corresponding FWHM lateral resolution in the slit direction can thus be obtained as 1.33 μm.
We then compare the spatial resolutions determined by the F-SECM spectral resolution and the slit optical setup. Since the spatial resolution is determined by the larger value between the estimations from the two contributors, it can be concluded that the theoretical FWHM lateral resolution in the spectral axis is 1.40 μm.
The lateral resolution in the frequency axis is, on the other hand, determined by the larger between the resolutions set by the numerical aperture (NA) of the imaging optics and by the frequency resolution of the modulator. The diffraction-limited FWHM spatial resolution is obtained as [18], which gives ~1.34 μm. The number of resolving points in the frequency axis can be obtained by considering a line focus with a length of ∆x, along which the coordinates are mapped by different modulation frequencies. In our F-SECM prototype, we employed a frequency modulator with discrete frequencies. In this case, the number of frequency steps along the radius of the disc, in effect, determines the number of resolving points in the frequency axis, which is 200 in our setup. Hence, the pixel resolution set by the frequency resolution is obtained as 1.75 μm. Comparing this value against the diffraction-limited resolution, it is clear that the theoretical lateral resolution along the frequency axis is determined by the pixel resolution and our F-SECM can resolve the features larger than 3.5 μm by Nyquist sampling criterion.
2.4.3. Axial resolution
Depth responses of slit confocal microscopes have been extensively investigated elsewhere [3, 17, 19]. As in the analysis of the F-SECM lateral resolutions (Sec. 2.4.2), we refer to [3] to obtain theoretical estimation of the axial resolution of our slit imaging system. For the half width of the slit of 4.3 in lateral OU, the half width of the FWHM axial response ( using the same notation in [3]) was found to be 6.3 in the axial OU. Using this value, an approximate expression for the FWHM axial resolution can be derived as:
From Eq. (9), the theoretical FWHM axial resolution can be estimated to be 9.80 μm.3. Results
3.1 Measured lateral and axial resolutions
In order to characterize the imaging performance of F-SECM, we first imaged a 1951 USAF resolution target (Thorlabs, Inc., R1DS1N,) as shown in Fig. 5(a). The small features in group 6 and 7 could be well discerned (Fig. 5(b)). Figures 5(c) and 5(d) show the intensity profiles along the lines indicated in Fig. 5(b). The spatial frequency of the elements 1 in group 7 is 128 lps/mm (i.e., 3.9 μm per line). We examined the FWHMs of their corresponding first-order derivatives. The measured FWHM resolutions were found to be 3.82 μm and 1.89 μm in the frequency and spectral axes, respectively. These results correspond to 1.09 × and 1.35 × larger values than the twice the pixel resolution in the frequency axis and the theoretical estimation in the spectral axis, respectively.
Fig. 5 (a) F-SECM image of a 1951 USAF resolution target and (b) a magnified view of the region indicated in (a). (c, d) show the intensity profiles along the solid red lines in (b). The spatial frequency of the elements 1 in group 7 is 128 lps/mm, which corresponds to 3.9 μm per line. The spectral line focus scanned the target from left to right directions, meaning that the left side of the image corresponds to the shorter wavelength.
The axial resolution of the F-SECM was then measured by acquiring the images as a planar mirror scanned through the focus in the sample plane. Figure 6 shows the measured axial response obtained by averaging the pixel values over the entire FoVy at the center wavelength. The measured FWHM resolution was found to be ~13.5 μm, which is ~1.38 × larger than the theoretical estimation.
We attribute the discrepancies between the theoretical and measured resolutions to the mismatch between the design wavelength of the objective (400 ~800 nm) and our operating wavelength (1.3 μm). In addition, the differences may also be partly accounted for by the tilt of the line focus produced by the cylindrical lens CL relative to the grating plane. The light was incident on the diffraction grating with an angle of ~51.5 degrees, and thus the resultant focal shifts on the grating plane along the extent of the line focus may lead to a wavefront distortion at the back focal plane of the objective. Based on our analysis with ZEMAX software, we noted that chromatic aberration of CL was found to produce the focal shifts of upto ~430 μm on the grating. This may also contribute to the degradation of spatial resolutions.
3.2 Imaging result with biological specimens
We then assessed the imaging capability of F-SECM by imaging an excised mouse ear, a pine cone axis, an onion, and eggshells (Fig. 7). Distribution of corneocytes could be clearly visualized in the mouse ear tissue (Fig. 7(a)), and tracheid structures in pine cones were observed with high contrast (Fig. 7(b)). Distinct arrangements of onion cells and porous structures in both the interior and exterior surfaces of the eggshells were also clearly visible. We attribute the artifacts observed in Fig. 7(c) (i.e., sharp horizontal discontinuity in the image) to the excessive motion jitter of the frequency modulator during the image acquisition.
Fig. 7 Representative F-SECM images of (a) a mouse ear (b), pine cone axis (c), onion cells and (d, e) interior and exterior structures of an eggshell, respectively. The spectral line focus scanned the specimens from left to right directions. The left side of the images corresponds to short wavelength. The scale bar represents 50 μm.
We further evaluated depth-resolved imaging capability of F-SECM by imaging an eye of a dragonfly. Shown in Figs. 8(a)-8(c) are the representative F-SECM images at the surface of the eye, and 10 and 20 μm deep in the sample. An array of ommatidia could be clearly discerned. Figure 8(d) is a three-dimensional representation of the eye, which was re-constructed from 80 images separated by 2 μm in depth. Each frame was acquired at 50 fps. A movie (Media 1) presents three-dimensional views of the dragon fly eye in different angles.
Fig. 8 (a)-(c) F-SECM images of a dragonfly eye acquired at its surface, and 10 and 20 μm below the surface, respectively. The spectral line focus scanned the specimens from left to right directions. The left side of the images corresponds to short wavelength. (d) shows a three–dimensional representation of the dragon fly eye reconstructed from 80 images acquired with a step size of 2 μm in depth (see Media 1). The scale bar denotes 50 μm.
4. Discussion
We presented a novel microscopy technique capable of scanner-free, depth-resolved imaging of biological and material specimens with a single-pixel detector. Two-dimensional spatial information of the specimen was encoded into different modulation frequencies and wavelengths of the probe beam. Therefore spectrally-resolved measurement of the back-scattered light from the specimen and subsequent Fourier analysis allowed for high-contrast, high-resolution imaging of the specimens. The slit aperture in the detection path also offered depth-sectioning capability. This scanner-free, single-pixel operation of F-SECM could greatly reduce the cost of imaging systems and may be applicable to high-speed, high-resolution imaging in the spectral regimes (e.g., IR, THz), in which the segmented image sensors are either unavailable or expensive.
In our F-SECM prototype, the number of pixels in the y axis was determined by the number of patterned frequency steps in the modulation disc. Yet, it would be worth to examine the limiting factor that governs the number of resolving points for the case of continuously chirped modulation frequencies. Let us consider a line focus with a length of ∆y, along which the coordinates are mapped by different modulation frequencies over the range of ∆ƒ. F-SECM resolves the spatial information in the y axis by taking the inverse Fourier transform of the acquired signal at a particular wavelength during a rotational period of the FM, i.e., TFM. Then, the frequency resolution is approximated as 1/TFM, and thus the number of pixels in the frequency axis can be obtained as Ny = TFM Δf, which is equal to the time-frequency bandwidth product. In our case, the maximum modulation frequency range that can be achieved is limited to 25 kHz by the Nyquist sampling criterion. With this limit, the number of resolving points can be increased by acquiring more wavelength sweeps; however, the overall image acquisition rate will decrease.
For the spectral axis, the number of resolving points is determined by a spectral bandwidth of the light source and the spectral resolution of the imaging system. The spectral resolution in our setup was limited by the diffraction grating in the imaging optics. Therefore, it could be improved by enlarging the beam incident on a grating and using a diffraction grating with a higher dispersive power.
F-SECM acquires spatial information of a specimen in the x axis with a spectrally dispersed light. Therefore, the signal along the axis may be influenced by the wavelength-dependent scattering and absorption properties of the specimen. While this feature may limit the types of specimens for observation, it could be exploited for spectral imaging of biological and material samples with chemical specificity [20]. For example, biological cells and tissues are known to exhibit distinctive wavelength-dependent absorption and scattering signatures depending on their physiological conditions at the UV-visible wavelengths [21, 22]. Thus, F-SECM in this spectral range can be potentially utilized to probe the physiological states of biological tissue with some modifications in the setup. Our F-SECM prototype, on the other hand, operates in the NIR spectral range (i.e., 1250 - 1360 nm). This NIR spectral range is typically used for deep tissue imaging, as the biological tissues exhibit reduced absorption and scattering in this spectral range. Therefore, our setup would be particularly useful for three-dimensional imaging of biological tissue samples.
It should be noted that the imaging speed of F-SECM can be further improved by employing faster wavelength-swept lasers and frequency modulators. In our F-SECM setup, the patterned disc mounted on a brushless DC motor was utilized as the frequency modulator. One can thus note that motors with a higher load capacity are expected to enhance the speed by an order of magnitude, with a spin velocity limited by material fracture limits. For example, commercially available high-speed DC motors are capable of providing rotational speeds of >100,000 rpm, which would result in increasing the modulation frequency range. Further increased speeds can be achieved by use of inertia-free systems, such as acousto-optic deflectors (AODs). For the light sources, polygon-based wavelength-swept lasers have demonstrated a sweep rate of 400 kHz [23]. Multi-MHz sweep rates can also be achieved using Fourier-domain mode-locked lasers [24, 25]. A combination of these advanced wavelength-swept lasers and novel strategies for high-speed frequency modulators would enable F-SECM to be a viable high-speed imaging platform, with potential applications in real-time imaging and unveiling fast responses in many areas.
In summary, we demonstrated a novel scanner-free, single-pixel imaging modality. In contrast to conventional single-pixel imaging techniques, our method exploits frequency and spectral encoding to map the spatial information of specimens, and thus does not require a dedicated beam or sample scanning device. Future development may include improvement of imaging speed and integration with fluorescence imaging modalities for structural and functional studies.
Acknowledgments
This research was supported by the research programs of National Research Foundation of Korea (NRF) (NRF-2012R1A1A1003867) and by the Korean Health Technology Research and Development Project of the Ministry for Health, Welfare and Family Affairs (HI10C19110300).
References and links
1. D. Shotton and N. White, “Confocal scanning microscopy: three-dimensional biological imaging,” Trends Biochem. Sci. 14(11), 435–439 (1989). [CrossRef] [PubMed]
2. W. Denk, J. H. Strickler, and W. W. Webb, “Two-photon laser scanning fluorescence microscopy,” Science 248(4951), 73–76 (1990). [CrossRef] [PubMed]
3. T. Wilson, Confocal Microscopy (Academic, 1990).
4. C. Xu, W. Zipfel, J. B. Shear, R. M. Williams, and W. W. Webb, “Multiphoton fluorescence excitation: new spectral windows for biological nonlinear microscopy,” Proc. Natl. Acad. Sci. U.S.A. 93(20), 10763–10768 (1996). [CrossRef] [PubMed]
5. J. Bewersdorf, R. Pick, and S. W. Hell, “Multifocal multiphoton microscopy,” Opt. Lett. 23(9), 655–657 (1998). [CrossRef] [PubMed]
6. T. Tanaami, S. Otsuki, N. Tomosada, Y. Kosugi, M. Shimizu, and H. Ishida, “High-speed 1-frame/ms scanning confocal microscope with a microlens and Nipkow disks,” Appl. Opt. 41(22), 4704–4708 (2002). [CrossRef] [PubMed]
7. G. J. Tearney, R. H. Webb, and B. E. Bouma, “Spectrally encoded confocal microscopy,” Opt. Lett. 23(15), 1152–1154 (1998). [CrossRef] [PubMed]
8. C. Boudoux, S. Yun, W. Oh, W. White, N. Iftimia, M. Shishkov, B. Bouma, and G. Tearney, “Rapid wavelength-swept spectrally encoded confocal microscopy,” Opt. Express 13(20), 8214–8221 (2005). [CrossRef] [PubMed]
9. R. G. Baraniuk, “Single-pixel imaging via compressive sampling,” IEEE Signal Process. Mag. 25(2), 83–91 (2008). [CrossRef]
10. M. Liang, R. L. Stehr, and A. W. Krause, “Confocal pattern period in multiple-aperture confocal imaging systems with coherent illumination,” Opt. Lett. 22(11), 751–753 (1997). [CrossRef] [PubMed]
11. R. G. Baraniuk, “Compressive sensing,” IEEE Signal Process. Mag. 24(4), 118–120 (2007). [CrossRef]
12. W. L. Chan, K. Charan, D. Takhar, K. F. Kelly, R. G. Baraniuk, and D. M. Mittleman, “A single-pixel terahertz imaging system based on compressed sensing,” Appl. Phys. Lett. 93(12), 121105 (2008). [CrossRef]
13. G. Futia, P. Schlup, D. G. Winters, and R. A. Bartels, “Spatially-chirped modulation imaging of absorbtion and fluorescent objects on single-element optical detector,” Opt. Express 19(2), 1626–1640 (2011). [CrossRef] [PubMed]
14. J. S. Sanders, R. G. Driggers, C. E. Halford, and S. T. Griffin, “Imaging with frequency-modulated reticles,” Opt. Eng. 30(11), 1720–1724 (1991). [CrossRef]
15. C. Sheppard and X. Mao, “Confocal microscopes with slit apertures,” J. Mod. Opt. 35(7), 1169–1185 (1988). [CrossRef]
16. T. Wilson and A. R. Carlini, “Size of the detector in confocal imaging systems,” Opt. Lett. 12(4), 227–229 (1987). [CrossRef] [PubMed]
17. J. Pawley, Handbook of Biological Confocal Microscopy (Springer, 2010).
18. G. S. Kino and T. R. Corle, Confocal Scanning Optical Microscopy and Related Imaging Systems (Academic, 1996).
19. T. R. Corle, C.-H. Chou, and G. S. Kino, “Depth response of confocal optical microscopes,” Opt. Lett. 11(12), 770–772 (1986). [CrossRef] [PubMed]
20. Y. Bar-Ilan and D. Yelin, “Spectral imaging using single-axis spectrally dispersed illumination,” Opt. Lett. 39(17), 5177–5179 (2014). [CrossRef] [PubMed]
21. J. Q. Brown, K. Vishwanath, G. M. Palmer, and N. Ramanujam, “Advances in quantitative UV-visible spectroscopy for clinical and pre-clinical application in cancer,” Curr. Opin. Biotechnol. 20(1), 119–131 (2009). [CrossRef] [PubMed]
22. H. Akbari, L. V. Halig, D. M. Schuster, A. Osunkoya, V. Master, P. T. Nieh, G. Z. Chen, and B. Fei, “Hyperspectral imaging and quantitative analysis for prostate cancer detection,” J. Biomed. Opt. 17(7), 076005 (2012). [CrossRef] [PubMed]
23. W.-Y. Oh, B. J. Vakoc, M. Shishkov, G. J. Tearney, and B. E. Bouma, “>400 kHz repetition rate wavelength-swept laser and application to high-speed optical frequency domain imaging,” Opt. Lett. 35(17), 2919–2921 (2010). [CrossRef] [PubMed]
24. R. Huber, M. Wojtkowski, and J. G. Fujimoto, “Fourier Domain Mode Locking (FDML): A new laser operating regime and applications for optical coherence tomography,” Opt. Express 14(8), 3225–3237 (2006). [CrossRef] [PubMed]
25. T. Klein, W. Wieser, C. M. Eigenwillig, B. R. Biedermann, and R. Huber, “Megahertz OCT for ultrawide-field retinal imaging with a 1050 nm Fourier domain mode-locked laser,” Opt. Express 19(4), 3044–3062 (2011). [PubMed]
