Massively parallel cantilever-free atomic force microscopy

Atomic force microscopy (AFM), since its inception in 1986, has become a leading technique for surface topography and functional property analysis at micro- and nanoscales. Conventional AFM uses a micro-cantilever to sense the minute forces between a sharp tip and the substrate, with deflection detected by an optical lever. However, this serial nature leads to a trade-off: higher resolution comes at the cost of a smaller field of view. Efforts to improve this include higher-bandwidth probes and probe arrays (e.g., IBM Millipede), but current arrays are limited to around 30 probes, highlighting the difficulty in parallelizing cantilever-based sensing. While limited in AFM, probe arrays are common in scanning probe lithography (SPL). To increase SPL throughput, cantilever-free architectures have been explored, where probe arrays rest on a compliant film. This offers scalability to millions of probes but loses the force-sensing capability of the cantilever. This research addresses the limited throughput of serial AFM and SPL by demonstrating massively parallel AFM using a cantilever-free architecture with scalable optical detection termed the distributed optical lever. The study aims to establish the feasibility and precision of this technique for large-area high-resolution imaging applications.

The existing literature extensively discusses the limitations of conventional AFM regarding the trade-off between resolution and field of view. Many studies have focused on improving the bandwidth of individual AFM probes to increase scan speeds and resolution. However, the inherent serial nature of single-probe AFM restricts its application to large-area high-resolution imaging. The use of probe arrays, such as the IBM Millipede, has been explored as a solution for parallelization, but the complexity and limitations in the number of probes remain significant challenges. In the field of scanning probe lithography (SPL), cantilever-free approaches have been developed to achieve high throughput patterning, yet these methods lack the force-sensing capabilities of traditional AFM. This study builds on previous research in cantilever-free architectures and aims to bridge the gap between high throughput and high-resolution imaging by introducing a novel optical detection mechanism.

The researchers developed a cantilever-free AFM system based on a distributed optical lever. The system uses an array of optically reflective conical probes fabricated on a compliant polydimethylsiloxane (PDMS) film on a sapphire wafer. The fabrication process involved spin-coating the PDMS onto a sapphire wafer, followed by two-photon polymerization direct laser writing (2PP-DLW) to create the conical probes. An aluminum coating was then applied to make the probes reflective. The compliant PDMS film acts as a distributed optical lever; vertical probe motion causes deflection of the film, changing the reflected light intensity, allowing for the detection of probe-sample contact. A model was developed to relate the optical contrast to probe motion and force. The model predicts a linear relationship between reflected light intensity and probe motion, allowing for sub-10 nm vertical precision. An array of 1088 probes was fabricated and mounted in a scanning probe instrument. Bright-field optical images were acquired through the sapphire wafer, and the probes were brought into contact with a sample using force feedback. The change in optical contrast upon contact was used to calibrate the distributed optical lever for each probe. The intensity of the reflected light was linearly related to the sample height, and this relationship was used for topographical image reconstruction. A raster scan of the sample was performed, and optical images were captured at each position. The images were processed using a Hough transform to locate the probes, and the reflected light intensity was used to calculate the sample height at each probe location. The overlapping regions of the image from neighboring probes allowed for correcting probe height variations and frame-to-frame variations in probe array extension (Z). A point spread function (PSF) was used to remove artifacts caused by the physical coupling between neighboring probes. The results were validated by comparison to AFM measurements on a calibration sample.

The study successfully demonstrated massively parallel AFM using a cantilever-free architecture and a distributed optical lever. A linear relationship between optical contrast and probe motion was observed, enabling sub-10 nm vertical precision. Simultaneous imaging with 1088 probes was achieved, mapping sample height with 100 nm lateral resolution and 9 nm vertical precision across a 0.5 mm area. The developed image reconstruction algorithm effectively corrected for various imaging artifacts, including probe height variations, frame-to-frame variations in Z, and crosstalk between neighboring probes. The final image exhibited a high level of fidelity and accuracy, as demonstrated by its agreement with measurements from traditional AFM. Analysis of an AFM calibration sample with a fiducial arrow feature revealed a measured step height of 126 nm which was within 15% of the AFM measurement. Furthermore, imaging of an intricate region of the calibration sample showcased the approach’s capability to generate images of multiscale surfaces, with an observed variation in step height among four probes measuring comparable regions of only ~3%. This precision is remarkable given the aspect ratio of over 10,000 and the large horizontal span of over 0.4 mm. The study also explored the potential for measuring lateral forces and gradients of sample topography using the asymmetric deformation profiles of the probes. The findings strongly suggest that this methodology could significantly enhance the efficiency of AFM in applications such as integrated circuit metrology, optical metasurface characterization, and multi-scale biological tissue studies.

The results demonstrate the potential of massively parallel cantilever-free AFM to overcome the limitations of conventional AFM in terms of field of view and throughput. The high resolution and large area capabilities of this technique make it suitable for various applications where both high-resolution and wide-field imaging are necessary. The ability to obtain high-quality images with minimal artifacts and high precision opens possibilities for imaging intricate multiscale samples, enhancing the utility of AFM across many research fields. The developed methodology shows significant promise for enhancing the speed and scalability of AFM, providing a more efficient alternative for various applications compared to conventional approaches. The success of this study provides a compelling rationale for further investigation into this approach.

This research successfully demonstrated massively parallel atomic force microscopy (AFM) using a novel cantilever-free approach. The use of a distributed optical lever for probe-sample contact detection enabled high-resolution imaging over a large area using a 1088 probe array. The developed image reconstruction algorithms effectively corrected for imaging artifacts. The high precision and scalability demonstrated in this proof-of-concept study opens exciting avenues for numerous applications in materials science, biology, and nanofabrication. Future studies could explore further improvements in probe design, higher probe densities, and integration with other imaging modalities.

While the study achieved high-resolution imaging over large areas, some limitations exist. The current system uses a hexagonal probe array, and future work could investigate different probe arrangements. The current image reconstruction algorithm assumes a linear relationship between probe deformation and crosstalk. The effect of probe wear on the image quality over extended imaging sessions requires further investigation. Also, the study primarily focused on topographical imaging; future work should investigate the feasibility of applying this approach to other types of AFM measurements, such as force spectroscopy.