1. Introduction

The traditional single-photon confocal microscope (SPM) with a Gaussian beam as illumination has been widely used in various research fields due to its advantages, such as high resolution, low excitation power, and tomographic imaging capabilities [1–3]. However, SPM imaging depth is very limited compared to two-photon and multi-photon imaging due to the strong scattering of short-wavelength excitation light by biological tissue [4–7]. Although the two-photon and multiphoton excitation microscope can achieve large imaging depths, they require higher excitation light power, which will induce some phototoxicity for biological samples and photobleaching to probes [6–13]. Moreover, to obtain volume images of three-dimensional (3D) structures, confocal microscopy requires 3D tomographic scanning, which inevitably increases the acquisition time and thus significantly reduces the volume frame rate. Slow frame rates also have many side effects, such as motion fuzzy in volume imaging that monitors blood flow in transparent zebrafish [7, 14–16]. Therefore, the SPM is not ideal for real-time volumetric imaging. To solve these problems, anti-diffracting Bessel and Airy beams have been explored to increase acquisition speed by converting the axial large depth of focus information into a two-dimensional (2D) image [17–23]. The extended focal length substantially extends the depth of field, providing fast volumetric acquisition speed. Therefore, anti-diffracting beam volumetric imaging can avoid a lot of time wasted by axial scanning. Although the advantages of the anti-diffracting Airy beam are obvious, the focal length of the anti-diffracting Airy beam is still limited. To overcome this limitation, Weng et al. developed a versatile optical pen to manipulate the number, amplitude, position, and phase of energy oscillations, so that the ultralong anti-diffracting beam can be realized in free space [22]. Due to the ultralong anti-diffracting distances, the UAD beam offers promising applications in many research fields, such as optical imaging, optical trapping, optical communication, and laser-assisted guiding. So far, studies exploring the performance of UAD beams for optical imaging under SPM have not been reported.

In this letter, we describe the first use of the UAD beam to perform on SPM. The UAD beam SPM with super-long focal length projects the axial images stack onto a single frame 2D image, which can fast obtain volumetric images by minimizing axial scanning. The ability of volumetric SPM with the UAD beam is evaluated by imaging through a strongly scattering environment mixed with fluorescent microspheres and agarose gel, as well as blood vessels in the transgenic zebrafish.

2. Experiment setup

Figure 1 is the schematic diagram of our home-built single-photon excitation microscope system. In the SPM system, a 473 nm CW laser (Becker & Hickl GmbH, BDL-473-SMN) is used as the excitation light source with a maximum output power of 10 mW, and the output light is expanded by a 4f lens system (L1: 100 mm, L2: 250 mm). A collimated beam passes through a half-wave plate to adjust the polarization state of linearly polarized light to match the spatial light modulator (SLM). The linearly polarized light is modulated by the SLM (Holoeye Pluto-2) with the UAD mask and the Airy mask that incorporate a fixed period of blazed grating, which converts the Gaussian light into the UAD light and the Airy light. The SLM is conjugated to a pair of X-Y galvanometric scanning mirrors (Thorlabs, GVS202TSH515) through a second 4f lens system (L3: 200 mm, L4: 100 mm). In the second 4f system, a pinhole is used to filter out other diffraction orders of diffracted light, reserving only the light of negative first-order diffraction. The galvanometric scanning mirrors through a third 4f lens system (L5: 100 mm, L6: 200 mm) are further conjugated to the rear aperture of an objective (Nikon, 40X, NA= 0.75). The fluorescence signal is collected by the same objective and separated from the major beam path by a dichroic mirror (Chroma, T505LPXR), and then is focused on the hybrid photodetector (HPD, Hamamatsu, H13223-40) by a lens (L7: 50 mm). Between the HPD and the lens (L7) is a bandpass filter (Chroma, ET525/40) which can remove any residual excitation light. The Airy and UAD modes can be switched by time-sharing loading the Airy or UAD mask to the SLM. The Gaussian mode is achieved by only loading a fixed period of the blazed grating. The Airy and UAD beams can be reconstructed by recording the light intensity in different z planes and the stage can be moved along the optical axis by a motorized precision translation stage.

 

Fig. 1. Schematic of the experimental setup. L1–L7: lenses; HWP: half-wave plate; SLM: spatial light modulator; P1: pinhole; DM: dichroic mirror; Scanner: X, Y linear scan; HPD: hybrid photodetector; Obj: objective; Z stage: motorized precision translation stage.

Download Full Size | PDF

3. Results and discussion

In the SPM imaging system, we performed imaging in three modes: Gaussian, Airy, and UAD beams. In the experiment, we first investigated the spot imaging properties under the three light field modes using fluorescent beads (Excitation: 505 nm; Emission: 515 nm) with a diameter of 0.2 µm. Figure 2(a), 2(c), and 2(e) show the lateral images of the fluorescence beads in Gaussian, Airy, and UAD beams SPM, respectively. The intensity profiles marked by the white dotted lines in Fig. 2(a), 2(c), and 2(e) are shown in Fig. 2(g). From the Fig. 2(g), we could see that the full width at half-maximum (FWHM) of spot images in Gaussian, Airy, and UAD beams were 0.48 µm, 0.55 µm, and 0.60 µm, respectively. Although the image spots diameters of Airy and UAD beams were larger than that of Gaussian spot, the difference was not large. Both the Airy and UAD image spots had side lobes, and the side lobe of the UAD spot was two times that of the Airy spot. Furthermore, to get more information on spot images in three modes, we obtained the axial light field distribution information using Z-axis scanning. The axial light field distribution of Gaussian, Airy, and UAD spot images are shown in Fig. 2(b), 2(d), and 2(f), respectively, and the maximum signal intensity along the axial direction, which take the selected longitudinal average strength value of the white dot box, is shown in Fig. 2(h). The axial FWHMs were 6.5 µm, 46 µm and 90 µm for Gaussian, Airy, and UAD modes, respectively. The axial FWHM of the UAD beam was approximately fourteen times larger than that of the Gaussian beam and approximately two times larger than that of the Airy beam, which indicates that the axial imaging range of UAD was significantly expanded. Therefore, the volume imaging with UAD beams is worth exploring. It should be noted that due to the symmetrically modulated Airy beam, the focal length of the UAD beam had a symmetrically curved trajectory, so the projected volume image and UAD side lobes were not perpendicular, which may slightly blur the image.

 

Fig. 2. Images of the fluorescent bead in Gaussian (top), Airy (middle), and UAD (bottom) mode with a NA= 0.75 objective. (a), (c), and (e) are lateral images in the three modes, respectively; (b), (d), and (f) are axial images in the three modes, respectively; (g) Intensity profiles in lateral images marked with white dotted lines in (a), (c), and (e); (h) Maximum signal intensity along the axial direction for (b), (d), and (f). Red, blue, and green lines represent the UAD beam, the Airy beam, and the Gaussian beam, respectively. Both signal intensities of the three modes beam are normalized.

Download Full Size | PDF

Then, the system performance was evaluated by volumetric imaging of spherical fluorescent beads. We prepared a thick sample by mixing 2 µm diameter fluorescent beads (Excitation: 505 nm; Emission: 515 nm) with a low melting point agarose gel (2% AGAR mixed with water). The sample of 90 µm thickness was scanned under Gaussian mode with 0.5 µm axial step size. Figure 3(a) shows a 3D projection of the sample under Gaussian mode, where the depth is coded by different colors. It took about three minutes to record a 3D projection image. There were about 90 beads randomly distributed on different planes of the sample. A single frame 3D projection image under Gaussian mode recorded these beads. When switched to the Airy mode, it only took 2 frames to cover most of the fluorescent beads under Gaussian mode [Fig. 3(b)]. When we switched to the UAD mode, only single frame image could cover most of the fluorescent beads under Gaussian mode or Airy mode [Fig. 3(c)]. Due to the ultra-long focal length of the UAD beam, a single frame of 2D volumetric was able to image the sample of all the beads, which greatly improved the acquisition speed. It is important to emphasize that each bead in the Airy mode was decorated with a comet tail due to the side lobes of the Airy beam. While some of the beads were decorated with a pair of symmetrical comet tails and some with a single comet tail in the UAD mode, meaning these beads were not in the same layer. The unique symmetrical curved trajectory related axial depth to penetration position [Fig. 3(d)]. We selected three points A, B, and C in the axial direction of the spot, representing different penetration positions (+40 um, 0 um, -40 um), and their corresponding lateral light field distribution were shown in the white box, as shown in Fig. 3(d). By studying the lateral and axial positional relationship of fluorescent beads, we ascertained that a single comet tail to the left indicated that the beads were on the surface, a symmetrical comet tail indicated that the beads were in the middle, and a single comet tail to the right indicated that the beads were on the bottom. In addition, there is a slight shift of the bead position in UAD mode compared to images in standard Gaussian and Airy modes. Compared to that of the Gaussian beams, the axial imaging range of Airy beams was substantially elongated with little difference in lateral resolution. Compared with those of the Airy beams, the images in the UAD mode had almost the same lateral resolution, while the axial imaging range produced by UAD beams was greatly expanded. Thus, the speed of volume imaging can be meaningfully improved. The UAD beam also had obviously larger penetration depth and better anti-scattering ability, as shown in the Fig.S1.

 

Fig. 3. Fluorescent beads volumetric imaging with an objective (40×, NA: 0.75). (a) Projection of Gaussian image stack of fluorescent beads color coded by depth. Step size: 0.5 µm; Frame rate: 1 Hz; (b) Projection of 2 frames Airy image stack of fluorescent beads color coded by depth; (c) Single frame UAD image of fluorescent beads; (d) Lateral light field distribution at different axial positions of the UAD beam. The excitation laser powers were 0.20 mW for Gaussian beams, 1.25 mW for Airy beams, and 2.50 mW for UAD beams.

Download Full Size | PDF

To demonstrate the advantages of the UAD volumetric imaging, we performed the dynamic single-photon fluorescence volumetric imaging of the cardiovascular system of zebrafish labeled with specific fluorescent proteins using a 473 nm laser. Figure 4(a) shows the projection of the Gaussian images stack (120 frames with 1 µm step size) of the zebrafish blood vessels labeled with an enhanced green fluorescent protein (EGFP), the effective imaging depth is about 120 µm and coded in different colors to show the 3D projection of the cardiovascular system of zebrafish. Figure 4(b) is the 2D Airy volumetric images stack (2 frames) of the zebrafish blood vessels. From the Fig. 4(b), we can see that the superposed Airy image information basically reproduced the Gaussian projection image. The information from 2 frames of the Airy images stack basically reproduced that from 120 frames of the Gaussian images stack. Figure 4(c) is a frame 2D UAD volumetric image of the zebrafish blood vessels. From Fig. 4(c) we can see that the information of the volumetric image almost covered Fig. 4(a) or Fig. 4(b). Therefore, a single 2D volumetric image in the UAD mode could be used to replace the projection image of the 3D stack image with 2 frames in the Airy images or 120 frames in the Gaussian images, which greatly improved the volumetric imaging acquisition speed. Furthermore, we use the single-photon UAD beam to monitor the dynamic changes of the cardiovascular system of zebrafish (See Visualization 1). Blood vessels of zebrafish approximately 30 µ m thick were dynamically monitored (Fig. 5). Figure 5(a) and Fig. 5(b) show three positions of substance transport in zebrafish blood vessels (red arrow) in the Gaussian mode and UAD mode respectively. Compared with the Gaussian mode (Gaussian mode was tomographic imaging), more detailed material transport in blood vessels was captured in the UAD mode because of the significantly stretched axial focal length (UAD mode was volume imaging). Based on single-photon imaging and dynamic monitoring experiments in the UAD mode, it was clear that the single-photon imaging of UAD beams substantially improved the capture speed and avoid the side effects of motion blur in vivo imaging, which has the potential for broad application in dynamic imaging of living organisms.

 

Fig. 4. Transgenic zebrafish blood vessels volumetric imaging. (a) Projection of Gaussian image stack of transgenic zebrafish blood vessels color coded by depth. Step size: 1 µm; Frame rate: 1 Hz; (b) Projection of 2 frames Airy image stack of transgenic zebrafish blood vessels color coded by depth; (c) Single frame UAD image of transgenic zebrafish blood vessels. The excitation laser powers were 0.20 mW for Gaussian beams, 1.25 mW for Airy beams, and 2.50 mW for UAD beams.

Download Full Size | PDF

 

Fig. 5. Dynamic volume monitoring. (a) Three processes for dynamic volume monitoring under Gaussian mode; (b) Three processes for dynamic volume monitoring in the UAD mode (See Visualization 1). Since there were some autofluorescent substances in the blood vessels, we could see the dynamic flow of substances in the blood vessels.

Download Full Size | PDF

4. Conclusions

In order to ensure the fairness of the comparison experiments, we used the same set of optical systems in our experiments, and the only difference was that different mask was loaded on the SLM to generate the required light field. Although the axial length of Airy beam can be adjusted by controlling the size of the cubic phase mask on the SLIM, the axial length of UAD beam can also be controlled using the same operation. In order to ensure the comparability of the experimental conditions, in our experiments the masks that generate the Airy and UAD beams are loaded on the SLM according to the size of 1:1. In conclusion, we first obtained light spot images in the three modes with 0.2 µm diameter fluorescent beads and a NA= 0.75 objective. Compared with that of the Gaussian beam and the Airy beam, the axial length of the UAD beam was substantially longer, approximately 14 times that of the Gaussian beam and approximately 2 times that of the Airy beam, meaning the imaging depth was deeper. Secondly, we verified the performance of the system through high scattering media samples that mixed fluorescent beads with AGAR. The information in a single frame 2D UAD volumetric image covered that of Gaussian images stack (90 frames) for a 90 µm thick sample. Thirdly, we found that the lateral light field distribution properties of the UAD beams were related to the axial depth position, which may be used to locate biological structures. But it only works for mini objects in the side-lobes directions, and can only estimate rough axial positions. Finally, we compared the volumetric images of the transgenic zebrafish blood vessels in the three modes and compared the dynamic transport process of substances in zebrafish vessels with Gaussian and UAD modal volume imaging. Due to the long focal length of the UAD beam, the axial volume imaging range was deeper, and more structural information was covered within a frame volumetric image, while the volumetric imaging needed to be mapped layer by layer in Gaussian mode. Thus, single-photon laser scanning microscopy with the UAD beam greatly improved the volume imaging frame rate. Although the sidelobe of the UAD beam caused some image blurriness, the significantly stretched axial imaging distance captured all axial images within the effective focal length, which is more robust for live sample imaging and thus avoids the axial motion of the sample. Therefore, it is worth sacrificing some image quality. In addition, the position of the sample in UAD mode was shifted slightly as compared with the image in Gaussian mode, which is induced by the symmetrical curved axial trajectory of the UAD beam. In summary, these advantages make the single-photon laser scanning fluorescence microscope with ultralong anti-diffracting beams as a potential tool for real-time monitoring of deep internal tissue activity of living organisms in a large volume. It is expected that the present study will promote the application of ultralong anti-diffracting beams in optical imaging and biological research.