Scanning superlens microscopy for non-invasive large field-of-view visible light nanoscale imaging.

Microsphere-based SSUM system and imaging equipment The BaTiO 3 microspheres (supplied by Cospheric) were attached to the cantilever using an ultraviolet curable glue (NOA63, Edmund Optics) as shown in Fig. 1c,d . Different illumination conditions were achieved by adjusting the two stops in the Köhler illumination system (Thorlabs). A high-speed scientific complementary metal oxide semiconductor camera (PCO. Edge 5.5) was used to record the images. Illumination was provided by an intensity-controllable light source (C-HGFI, Nikon, Japan), and the peak illumination wavelength of the system was set to ∼550 nm by the optical components for white-light imaging. A × 50 objective (Nikon TU Plan EPI ELWD) and a × 100 objective (Nikon LU Plan EPI ELWD) were used in these experiments. A split photodiode (QP50-6-18U-TO8, First Sensor) and a 635 nm circular beam laser diode module (#83-838, Edmund Optics) were applied in the custom AFM system ( Supplementary Fig. 1 ). Polystyrene ( M w =123 kg mol −1 , M w / M n =1.08, from Alfa Aesar) was dissolved in toluene and spin-coated onto the AgNWs that were dispersed onto silicone substrates. The optical beam profiler used to measure the illumination conditions was made by Thorlabs (BP209-VIS). The SEM images were taken using a Zeiss EVO MA10.

Cell culture and imaging Mouse myoblast cells

(C2C12) and human breast cancer cells (MCF-7) were cultured on common Petri dishes in DMEM or RPMI-1640 containing 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C (5% CO 2 ) in an incubator (Model 371, Thermo Scientific). The solvent used to fix the cells imaged in this study was 4% paraformaldehyde. For fluorescence microscopy of actin cytoskeleton, the cells were treated with Alexa Fluor 488-phalloidin (A12379, Thermo Fisher Scientific) at a concentration of 0.165 μM according to the process recommended by the supplier. A mercury lamp filtered by a fluorescein isothiocyanate (FITC) emission filter (MDF-FITC, Thorlabs) was used to excite fluorescence. The fluorescent imaging by SSUM ( Fig. 6d and Supplementary Figs 7d and 8d ) was conducted before the white-light imaging to prevent fluorescence quenching.

Imaging with SSUM

The microsphere-attached AFM probe was mounted on a custom holder. After the common adjustment procedure used in AFM, the sample was moved towards the microsphere using a three-dimensional translation stage during which the AFM feedback was acquired to monitor whether the microsphere touched the sample surface and to adjust the distance between the microsphere and sample. As the distance or force reached a set-point value, the optical microscope was driven by a vertical translation stage with nano-scale resolution (IMS100V, Newport) to acquire the microsphere-generated virtual images. The AFM feedback was open or closed in contact or non-invasive mode. In non-invasive mode, Δ z was precisely adjusted by the PZT scanner (P-733.3CL, Physik Instrumente, Germany), and the sample stage was adjusted to be horizontal to decrease the difference in the distance between the microsphere apex and sample stage. The images were recorded by an external signal that triggered a high-speed camera from a controller ( Supplementary Figs 1a and 16a ) during scanning. The recorded area of the camera could be adjusted before scanning to the proper size to satisfy the overlap (15–20% in our experiments) required for image stitching and to be within the area of the FOV of the microsphere superlenses without apparent aberration, which effectively reduces the data processing time and enables rapid image processing during or after scanning. These recorded images could be directly used in the image stitching procedure without preprocessing. A commercial piece of software (Topostitch, Image Metrology) was used to stitch the acquired image tiles of a CPU, a Blu-ray disc, and the cell shown in Fig. 5b,d , whereas the free ImageJ software with the Stitching plugin 49 based on a phase correlation algorithm 50 was used to stitch the cellular images shown in Figs 5f,h and 6b,d according to the procedures shown in Supplementary Fig. 16c–i . These two software platforms took the recorded image tiles as input and the set overlap percentage. A few layout arrangements were provided for selection to arrange the image tiles in trigger-signal order ( Supplementary Fig. 16a ). Several fusion methods are also provided for selection by ImageJ. We mention both of them here to provide more options for different purposes or applications. There were 1,056 ( Figs 2c , 4b and 5d,h ), 561 ( Fig. 6b,d ), 272 ( Fig. 3d ) and 320 ( Supplementary Fig. 5d ) image tiles used in the image stitching processes. The performances of and time consumed by these two software platforms are compared in Supplementary Fig. 16 and Supplementary Table 1 . The quality of the stitched images can be further improved by selective processing using a band-pass filter algorithm or a recursive bilateral filtering algorithm 51 . Analyses of microsphere imaging properties In this study, the imaging conditions were similar to the constant-height scanning mode (that is, the microsphere was attached to a TESP cantilever, which was fixed to maintain the distance between the microsphere and objective, as shown in Fig. 7a ). Initially, the microsphere superlens was in contact with the Blu-ray disc’s surface (that is, Δ z =0 μm) without pre-stress, which can be achieved by monitoring the position-sensitive device signal of the AFM system. Then, the microscope, including the objective driven by a motorized vertical stage, was adjusted to a position where the virtual images generated by the microsphere superlens could be clearly observed. Before scanning, a pre-stress was applied by moving the PZT towards to the objective; this produced an ∼1 μm-deep pre-deformation at the position, where the microsphere was attached. Then, the Blu-ray disc was carried by the PZT scanner away from the microsphere at different illumination conditions ( Fig. 7b,c ). The virtual images generated by microsphere superlens in the vertical scanning processes were recorded using a high-speed camera (inset of Fig. 7a ).

Data availability

The data sets generated during and/or analysed during the current study are available from the corresponding authors on reasonable request.

Supplementary Material Supplementary Information Supplementary Figures, Supplementary Table, Supplementary Notes and Supplementary References. Supplementary Movie 1 Dynamic scanning of a CPU surface by SSUM. Supplementary Movie 2 Dynamic scanning of a skeletal muscle cell (C2C12) by SSUM.