Acoustofluidic scanning fluorescence nanoscopy with large field of view

Nanoscale fluorescence imaging with a large-field view is invaluable for many applications such as imaging of subcellular structures, visualizing protein interaction, and high-resolution tissue imaging. Unfortunately, conventional fluorescence microscopy has to make a trade-off between resolution and field of view due to the nature of the optics used to form an image. To overcome this barrier, we have developed an acoustofluidic scanning fluorescence nanoscope that can simultaneously achieve superior resolution, a large field of view, and enhanced fluorescent signal. The acoustofluidic scanning fluorescence nanoscope utilizes the super-resolution capability of microspheres that are controlled by a programable acoustofluidic device for rapid fluorescent enhancement and imaging. The acoustofluidic scanning fluorescence nanoscope can resolve structures that cannot be achieved with a conventional fluorescent microscope with the same objective lens and enhances the fluorescent signal by a factor of ~5 without altering the field of view of the image. The improved resolution with enhanced fluorescent signal and large field of view via the acoustofluidic scanning fluorescence nanoscope provides a powerful tool for versatile nanoscale fluorescence imaging for researchers in the fields of medicine, biology, biophysics, and biomedical engineering.


INTRODUCTION
Fluorescence microscopy has become an indispensable technique in the fields of biology and medicine 1 with applications ranging from microscale imaging of live cells to nanoscale imaging of DNA sequencing protocols. 2,3 However, due to the structure of the objective lens in conventional fluorescence microscopy, a trade-off is made in terms of the resolution and field of view. A higher-resolution image from a conventional fluorescence microscope can be achieved by using a higher magnification (typically also with a higher numerical aperture) objective lens, but it is typically at the cost of a reduced field of view. One effective approach to solve the issue of increasing resolution while maintaining a large field of view is through utilizing scanning dielectric microspheres. [4][5][6][7][8][9][10][11][12][13] When the dielectric microsphere has a higher refractive index than its outer medium, the propagated light is focused from the inside of the microsphere, and a highly localized electromagnetic beam is generated near its surface, a phenomenon known as the photonic nanojet, that allows for super-resolution imaging below the diffraction limit.
In this article, we introduce an enhanced acoustofluidic scanning nanoscope for fluorescence imaging and amplification. Under the same imaging conditions, the acoustofluidic fluorescent scanning nanoscope can distinguish structures that cannot be resolved from a conventional fluorescent microscope under the same objective lens, and enhance the fluorescent signal by a factor of ~5 without altering the field of view of the image. With these features, the acoustofluidic scanning fluorescence nanoscope achieves rapid and superior fluorescence imaging with a large field of view that could be utilized in the fields of biology, chemistry, materials science, engineering, and medicine. nanoscope. Super-resolution imaging is achieved when a microsphere was placed on the target sample as shown in the yellow dotted box in Figure 1(a). The sample consists of fluorescent nanoparticles that are drop-cast on a cover glass. The fluorescent particles are then covered by a thin layer of PDMS film to lock their positions on the cover glass and avoid drifting in the imaging process. A large field-of-view image is achieved by stitching the super-resolution image from each scanning microsphere. The scanning of microspheres is achieved by activating a propagating acoustic wave following the same method as we used before 34,35 or counter-propagating acoustic waves that is demonstrated in this work. The advantage of using counter-propagating acoustic waves is that we can easily control the direction of the scan, which could not be achieved with a propagating acoustic wave. source combined with a blue bandpass filter was used to illuminate the sample, which was passed through a dichroic mirror and focused through a 60x objective lens. The green fluorescent light from the green fluorescent nanoparticle was collected on a CMOS camera (shown as a red camera #1 in Figure 1(b)) through the same objective lens combined with a green bandpass filter (denoted as emission filter in Figure 1(b)).

Configuration of the acoustofluidic scanning fluorescence nanoscope
A 50:50 beam splitter was placed in the light pathway to add a second camera (shown as a blue camera #2 in Figure 1(b)) in the system without the emission filter in order to track the position of each microsphere. The position of the two cameras (#1 and #2) are adjusted in a way so that both the sample and the microsphere can be imaged simultaneously on camera #1 and camera #2, respectively. The red box in Figure 1  (a) 3D schematic of the system. A hard PDMS membrane on the target sample helps achieve the desired focal distance and demonstrate high-resolution images, as shown by the yellow box in the 2D schematic on the right. (b) Schematic of the optical setup. A 50:50 beam splitter delivered images into two different cameras, for both fluorescent detection (Camera #1, red box) and microsphere tracking (Camera #2, blue box). (c) Experimental results of enhanced fluorescent amplification of 500 nm fluorescent nanoparticle images (Camera #1, red box) through microspheres and microsphere particle tracking (Camera #2, blue box). Camera #2 focused on the center of the microspheres, as shown in the blue box. Only camera #1 was connected to an emission filter. Scale bars are 20 µm. The simulation confirmed that the light is well focused by the microsphere to a spot with a full width at half maximum (FWHM) of 720 nm and a distance of 17 µm away from its surface (defined as its focal length) as shown on the vertical graph in Figure   2    which is defined as the ratio of the averaged intensity of the fluorescence with a microsphere to that without a microsphere, is found to be ~5 (Figure 3(c)).

Bidirectional acoustofluidic scanning of microspheres
To perform the efficient 2D scanning process in a large field of view, we designed and fabricated a bidirectional acoustofluidic scanning device. As shown in Figure 4

Image distortion correction and large-field-of-view imaging
The off-axis fluorescent image from a microsphere suffers large image aberrations as manifested in the image (the comet-like tails in the images located at the edge of the microsphere) shown in Figure 3(b). Since each of the distorted images comes from a single nanoparticle, its distortion is corrected by a Matlab algorithm that allows us to adjust different types of lens distortion by changing the value of an input parameter.

CONCLUSION
In this work, we developed an acoustofluidic scanning fluorescence nanoscope that can achieve superior resolution without sacrificing the field of view of the image. In contrast, a trade-off has to be made between resolution and field of view for most conventional fluorescence microscopes. The presence of a microsphere in the scanning fluorescence nanoscope lead to enhanced fluorescence compared to that without a microsphere. The bidirectional acoustofluidic scanning design allowed excellent freedom to control the scan of the microsphere. The dual-camera configuration enabled us to collect the fluorescent signal as well as the position information of each microsphere to form the image with a large field of view. Finally, the image-correction algorithm significantly reduced image distortion, resulting in a clearer and more accurate representation of the sample. Based on these features, the acoustofluidic scanning fluorescence nanoscope can be valuable for biomedical imaging and lab-ona-chip systems.

Optical characterization
As shown in Figure 1
The distance between the two transducers was 6 mm.

Microsphere preparation and experimental setup
To perform the microsphere imaging, we chose 20 µm polystyrene microspheres (refractive index: 1.6, Sigma-Aldrich, USA). The microspheres were diluted with deionized water before being placed on the sample surface. To maintain a consistent water channel height between the device and sample, a square cover glass (#1.5, 10 × 10 mm, Ted Pella, USA) was placed at both ends of the device. The MATLAB (version: R2021) script was designed and executed to control the function generator (FY6600, FeelTech, China) and CMOS cameras simultaneously. These cameras were used to collect the image data. An acoustic burst mode with 0.2 second intervals was applied, and the image acquisition for the two cameras was executed every 0.2 seconds.

Simulation of the acoustic field
To understand the acoustic energy distribution within the device, a model of an acoustic device was designed in COMSOL Multiphysics®. The model included two piezoelectric transducers, a cover glass, and water under the cover glass. A time domain study was used to visualize the transducer excitation. A 2.1 kHz and 4 VPP signal was applied to the transducer using the electrostatics module. A low reflection boundary water layer with open channel conditions was applied to the cover glass layer. We observed the vibration profile and acoustic streaming to determine the proper microsphere manipulation area, which was the space between the two transducers.

Imaging sample preparation
To experimentally demonstrate the scanning performance of the system, we fabricated a fluorescence nanoparticle sample with a hard PDMS (PP2-RG07, Gelest, Inc., USA) membrane. The green fluorescence nanoparticle sample (200 nm: FSDG002, 500 nm: FSDG003, Bangs Laboratories, Inc., USA) was diluted with deionized water and loaded on the cover glass (24x50 mm C8181-1PAK, Sigma-Aldrich, USA). Then the sample was dried at room temperature for 3-6 hours. After drying, we applied a hard PDMS mixture to the sample and ran the spin-coat (WS-650-23, Laurell Technologies, USA) process. Then, the sample was baked at 60°C for 30 minutes in the oven.

Image processing method
To generate the final scanned image, the collected images were processed in the following order. First, a circle-finding algorithm was executed on the image from Camera #2, as seen in the bottom panel of Figure 1(c), which stored information on the microsphere coordinates and radius. The magnification factor was calculated using the ratio of the sample grating line pitch length between Camera #1 and Camera #2. The calculated magnification factor (0.984) was then multiplied with the coordinates and radius information and applied to the images from Camera #1, as shown in the top panel of Figure 1(c). Next, the microsphere magnified circle images were cropped from the images of Camera #1. Finally, the cropped images were pasted onto the final image with a lens distortion restoration technique to ensure clear matching between the images.
Each image was processed recursively in the same manner. The final scanned image was generated by the repetitive image processing algorithm.

SUPPORTING INFORMATION
Microspheres floating when a voltage higher than 4 VPP is applied from a function