High-speed 3D microscopic imaging methods have led to numerous biological discoveries. In this Letter, we present a sync-free light sheet microscope (LSM) based on a static phase mask that eliminates the need for coupling the detection plane, thereby enabling high-speed volumetric imaging that is only limited by the speed of cameras, e.g., cross sections/s. In the sync-free design, the emission signals are first guided to the back of a galvanometric mirror, which laterally scans the emissions across a 2D phase mask, converting it to axial scanning that automatically compensates the focal shifts in the detection optics. Parametric models are developed to guide the phase mask design as well as to relate the axial scanning depth and magnification to design parameters. To quickly evaluate the different mask designs, a liquid-crystal-based spatial light modulator (LC-SLM) is used in the system. In the experiments, we scanned pollen and tissue samples via both the 2D phase mask and a piezoelectric objective scanner. The results show that the new method can generate clear images with comparable quality throughout the scanning range. The overall efficiency of the LSM is . It is worthwhile to note that the efficiency will be significantly improved by replacing the LC-SLM with a custom-made lens. The new method realizes a compact LSM for high-speed 3D imaging that may find important applications in in vivo biological imaging.
© 2017 Optical Society of America
The capability to visualize a volume instead of a cross section is vital to study and understand biological events occurring at different depths simultaneously and interactively . Examples include tracking particles/cells  and monitoring neuronal activities [3,4] or embryo development [5,6] in vivo. In 2004, the selective plane illumination-based techniques, i.e., light sheet microscope (LSM), were conceived and introduced by Stelzer’s group  to overcome the speed limitation of point-scanning systems. In this system, the excitation laser is focused by a cylindrical lens into a thin light sheet and directly illuminates a cross section of the sample; the emissions are collected by a separate detection objective positioned orthogonally to the excitation laser. Due to the parallel nature of the imaging process, real-time volumetric imaging may be realized. In addition, photobleaching is reduced as only the plane of observation is illuminated, enabling long-term volumetric imaging . To perform volumetric imaging, a direct method is to scan the samples axially with an XYZ stage; however, the speed is limited to . Besides, moving soft biological samples at high frequency often causes sample deformation or other negative effects. To achieve higher speeds without disturbing the samples, a galvanometric scanner (GS) is integrated in LSMs to scan the light sheet axially . In order to obtain clear images throughout the scanning process, the detection plane must be synchronized and scanned simultaneously with the light sheet, i.e., the excitation plane; the separation of the excitation and detection planes will result in blurry and defocused images on the charge coupled device (CCD). Much work has been done to improve the scanning speed of the detection plane, which has been the bottleneck of LSM performance. For example, to mechanically scan the detection objective lens using a piezoelectric actuator , the speed of this method is limited to 10 s Hz due to the weight of the objective lens. Following the same concept, one may pair an electrically tunable lens (ETL) with the detection objective to modulate the focal length ; while this solution is viable, it requires complex calibration processes, and the speed of the ETL is limited to . The defocusing issue may also be relieved by axially elongating the point spread function (PSF) of the detection objective lens; this can be achieved by introducing (1) a high refractive index medium  or (2) cubic phases to the emission signals in the detection path , where the elongated PSF reaches a length of when paired with a 10× objective lens. More recently, a single-objective LSM was developed based on an angled and confocally aligned planar light sheet, where the oblique light sheet automatically descans the detection plane, realizing in vivo animal studies at a speed of 20 volumes per second ; however, the resolution is slightly compromised due to the underfilled back aperture of the objective lens. A light sheet technique that does not require active synchronization components and compromise resolution has yet to be developed.
In this Letter, we present a new sync-free LSM, where the detection plane is automatically synchronized via a 2D phase mask, enabling high-speed 3D imaging that is only limited by the speed of cameras. Figure 1 presents the optical configuration of the sync-free LSM. The design of the illumination system resembles a generic selective plane illumination microscope (SPIM), where the light source is a 488 nm continuous-wave laser (MLD 488 nm, Cobolt Inc.); the laser beam is first expanded to form a light sheet via a cylindrical lens (CL); the focal length of the CL (50 mm) is selected to achieve a reasonable field of view (FOV), i.e., . Next, the laser stripe is focused onto a GS to perform volumetric scanning. After that, a 4-f system, i.e., L1 and L2 (; ), adjusts the beam size to fill the back aperture of the illumination objective (IO; Nikon Plan 10x) as well as to control the light sheet thickness. Note that the GS is conjugated with the IO back aperture. Lastly, a thin light sheet, i.e., illumination plane, is generated in the focal volume and scanned axially by the GS.
For detection, the emission signals are first collected by the detection objective (DO; Nikon APO LWD 40x/1.15, water immersed), which is paired with L3 () to control the beam size. Next, the emissions are guided to the back of the GS, which routes the emissions from different depths to different lateral regions on a liquid-crystal-based spatial light modulator (LC-SLM; Leto, Holoeye)—achieving automatic synchronization. The encoded phase patterns then fine-adjust the focal lengths of the emission signals at different regions on the LC-SLM so that their focal planes always coincide with the CCD camera (ProEM; Princeton Instruments) after the tube lens L4 (). In other words, the phase mask converts the lateral scanning from the GS to axial scanning, compensating the focal shifts in the detection optics. To achieve better efficiency, both sides of the galvanometric mirror have been polished and coated. It is worthwhile to note that the LC-SLM is used in the system to quickly evaluate the different phase mask designs, and during volume scanning processes, the phase patterns on the LC-SLM remain unchanged. A polarizer is included in the light path, in front of the LC-SLM, which is sensitive to polarization.
In the following, we present the design and optimization of the 2D phase mask. The concept of exploiting transverse displacements for compensating defocus errors is inspired from the spherical aberration phenomenon. As illustrated in Fig. 2, when a lens profile deviates from its ideal aspherical profile, the incident beam at different radial positions will be focused to different axial positions. In other words, with properly designed phase patterns, one can guide the incident beam from different radial or transverse positions to the desired axial positions, and vice versa. To determine the required compensation during axial scanning processes, we first relate the axial scanning distance () to the GS rotation angle () as expressed in Eq. (1),2): 2) is derived by considering the detection optics as a lens combination imaged at the detection plane of the EMCCD. Next, we present a general approach to determine the lens profile based on Snell’s law; combining with Eq. (2), the solution is the 2D phase mask we seek for the sync-free LSM. The lens profile along and traverse to the scanning direction is defined as the and directions respectively, as shown in Figs. 1 and 2, where the phase pattern in the direction functions as a normal lens of a focal length . To obtain the lens profile in the direction, let the focal points along the optical axis (i.e., , ) be a function of the radial position of the incident beam (i.e., , ), i.e., . Considering a local lens region at (), an incident beam will be refracted and bent towards , and the angles of incidence and refraction, i.e., and , are related by Snell’s law. Next, the lens is divided into regions, and the slope for each region can be solved simultaneously with the angle of incidence and refraction, as illustrated in Fig. 2(b). Accordingly, the lens profile in the axis can be found as a continuous piecewise linear function. When is large, the function approaches the ideal lens profile. In practice, is usually selected to match the resolution of the LC-SLM, e.g., 3 to 5 divisions per LC-SLM pixel. Table 1 presents the pseudo-codes of the steps for obtaining the lens profile. Note that the required focal length () for the point (, ) is obtained from Eq. (2). The final cross-sectional profile of the lens is . Lastly, the lens profile is converted to an equivalent 2D phase map to be programmed to the LC-SLM. The phase along the direction, i.e., the GS scanning direction, can be obtained from Eq. (3), where is the wave number: 4), which is the phase of a spherical lens under the paraxial condition, 5): 5) to an LC-SLM, the modulus of the phase is converted to a [0–255] grayscale digital image. Figure 3 presents the calculated 2D phase map, where the focal length varies from , , , to 250 mm as the light scans from the left end to the right end. Aside from compensating the axial shifts in an LSM, this new method can further be generalized to arbitrarily control the focal point in the axial direction.
Like many other wavefront modulation-based axial scanning methods, the system magnification slightly varies as the focal shifts axially. The relationship between the magnification () and axial scanning distance () is mathematically described in Eq. (6):1)–(5), Fig. 4(a) presents the relationship between the equivalent LC-SLM focal length () and the axial scanning distance (), and Fig. 4(b) presents the corresponding magnification variation as the illumination plane scans axially. In this analysis, the target scan range () is 100 μm, and thus a better range for axial scanning lies between and 80 μm, where and 0.44, respectively. Note that the scanning performance, i.e., range and resolution, may be further increased if the LC-SLM has a larger aperture and smaller pixel sizes.
In the following, we characterize the axial scanning performance of the sync-free LSM using a stained pollen sample (, mixed pollen grains, Carolina Biological Supply, USA). Figures 5(a) and 5(b) present the volumetric imaging results of a sample region that contains multiple pollens laying at different depths, scanned by the 2D phase mask and piezoelectric objective scanner (PIFOC, Physik Instruments), respectively. Images of three selected depths, i.e., , 0, and 10 μm, are presented, where the images of the same depths are grouped in the same column, showing comparable quality and resolution throughout the scanning range; the slight blurriness in Fig. 5(a) is mainly attributed to the unmodulated light reflected from the SLM, i.e., the DC component. From the results, we can conclude that the 2D phase mask can effectively perform high-speed axial scanning with comparable resolution versus mechanical scanning. In Fig. 5, one may also observe the field magnification effect; for example, the left and right columns are slightly magnified and demagnified, respectively, matching well with the predicted field magnification in Fig. 4(b). Note that in all experiments the phase patterns on the LC-SLM remain unchanged, and the volumetric imaging speed is only limited by the GS or camera, e.g., 1000 frames per second for a typical EMCCD camera. Additional tissue imaging results of fluorescently labeled bovine pulmonary artery endothelial cells (BPAEC) as well as discussions on aberrations are provided in Supplement 1. Video demonstrations of the pollen sample and BPAEC are presented Visualization 1 and Visualization 2, respectively.
Figure 6 presents the PSF of the LSM along the axis at four different depths, i.e., , 0, 30, and 60 μm, where the 0 position refers to the inherent focal plane of the DO. In the experiments, the PSF is measured via 1-μm-diameter fluorescent beads (F8819, Thermo Fisher Scientific) with a 40× DO (Nikon CFI S Plan Fluor ELWD; ). From the results, one may observe that the resolution of the LSM along the axis is 1.6/1.0/8.0 μm in the inherent focal plane (). As expected, the lateral resolution slightly worsens as it scans away from the original focal plane; the resolution at , 30, and 60 μm along the axis becomes 2.4/1.4/8.0 μm, 1.8/1.0/6.0 μm, and 1.8/1.0/6.0 μm, respectively. Importantly, the measured resolutions are comparable with most remote scanning methods used in LSMs. It is worthwhile to note that the lateral resolution along the and axes is different, where the resolution along the axis is compromised due to the aberration introduced by the galvanometric scanning process.
The current sync-free system has an overall efficiency of , where most energy is lost from the linear polarizer (50%) and the LC-SLM (50%); the DC component in the LC-SLM introduces both energy loss and background noises that degrade the image quality. In practice, the system will have much higher efficiency when the LC-SLM is replaced by a custom-designed lens.
In conclusion, we have developed and characterized a compact sync-free LSM based on a 2D phase mask, which effectively converts the lateral scanning from the GS to axial scanning, compensating the focal shifts in the detection optics. Parametric models have been developed to guide the phase mask design and deterministically link system performance, i.e., scan range and magnification, to the design parameters. Imaging experiments have been devised and performed to demonstrate the speed and effectiveness of the phase mask scanner; the results show clear optical cross sections of pollen samples with comparable resolution to mechanical scanners. Without the need for scanning the detection plane, the new sync-free LSM realizes high-speed 3D imaging and may find important applications in in vivo volumetric imaging. In addition, the new phase-mask-based scanning unit can be used in other laser systems that require high-speed axial scanning—for example, laser scanning microscopy or 3D laser fabrication.
HKSAR Research Grants Council, General Research Fund (CUHK 14206517); HKSAR Innovation and Technology Commission, Innovation and Technology Fund (ITS/179/16FP).
The authors thank Mr. Dong Keun Lee and Prof. Mi-Ryoung Song, School of Life Sciences, Gwangju Institute of Science and Technology (GIST), Gwangju, Republic of Korea, for providing tissue samples used in the imaging experiments; and Prof. Jing Yuan, Britton Chance Center for Biomedical Photonics, Wuhan National Lab for Optoelectronics, Huazhong University of Science and Technology, Wuhan, China, for loaning us the LC-SLM unit used to perform part of the imaging experiments.
See Supplement 1 for supporting content.
1. P. J. Keller, F. Pampaloni, and E. H. Stelzer, Curr. Opin. Cell Biol. 18, 117 (2006). [CrossRef]
2. K. Jaqaman, D. Loerke, M. Mettlen, H. Kuwata, S. Grinstein, S. L. Schmid, and G. Danuser, Nat. Methods 5, 695 (2008). [CrossRef]
3. W. C. Lemon, S. R. Pulver, B. Höckendorf, K. McDole, K. Branson, J. Freeman, and P. J. Keller, Nat. Commun. 6, 7924 (2015). [CrossRef]
4. T. F. Holekamp, D. Turaga, and T. E. Holy, Neuron 57, 661 (2008). [CrossRef]
5. P. J. Keller, A. D. Schmidt, A. Santella, K. Khairy, Z. Bao, J. Wittbrodt, and E. H. Stelzer, Nat. Methods 7, 637 (2010). [CrossRef]
6. R. Tomer, K. Khairy, and P. J. Keller, Curr. Opin. Genet. Dev. 21, 558 (2011). [CrossRef]
7. J. Huisken, J. Swoger, F. Del Bene, J. Wittbrodt, and E. H. K. Stelzer, Science 305, 1007 (2004). [CrossRef]
8. E. G. Reynaud, U. Krzic, K. Greger, and E. H. K. Stelzer, HFSP J. 2, 266 (2008). [CrossRef]
9. P. J. Keller, A. D. Schmidt, J. Wittbrodt, and E. H. Stelzer, Science 322, 1065 (2008). [CrossRef]
10. M. B. Ahrens, M. B. Orger, D. N. Robson, J. M. Li, and P. J. Keller, Nat. Methods 10, 413 (2013). [CrossRef]
11. F. O. Fahrbach, F. F. Voigt, B. Schmid, F. Helmchen, and J. Huisken, Opt. Express 21, 21010 (2013). [CrossRef]
12. R. Tomer, M. Lovett-Barron, I. Kauvar, A. Andalman, V. M. Burns, S. Sankaran, L. Grosenick, M. Broxton, S. Yang, and K. Deisseroth, Cell 163, 1796 (2015). [CrossRef]
13. O. E. Olarte, J. Andilla, D. Artigas, and P. Loza-Alvarez, Optica 2, 702 (2015). [CrossRef]
14. M. B. Bouchard, V. Voleti, C. S. Mendes, C. Lacefield, W. B. Grueber, R. S. Mann, R. M. Bruno, and E. M. C. Hillman, Nat. Photonics 9, 113 (2015). [CrossRef]