In the last three decades, atomic force microscopy (AFM) has evolved to be one of the most powerful and versatile tools enabling nanoscale analysis and exploration in many areas such as material science, physics, chemistry, tribology and nanomanufacturing. Especially, there has been intensively increasing interest in applying AFM in biological researches due to its unmatched capability of studying biological samples like protein, DNA molecule and live cell in their native physiological environment with high resolution. Dynamic AFM is a widely used AFM mode because it greatly reduces the lateral force between tip and sample through intermittent tip-sample contact during scanning and it is insensitive to thermal drift of cantilever. However, there is a major drawback in conventional dynamic AFM which inhibits further innovation and full development of its potentials: its capability of high speed scanning with high spatial resolution retained is limited by several fundamental issues involving the tapping dynamics of cantilever, the instrument hardware and the amplitude modulation principle of dynamic AFM operation.
In this dissertation, the fundamental aspects in modeling, actuation, sensing and control of AFM are investigated. Accurate multi-mode tip position reconstruction, high speed and high precision active tip motion control, precise sensing and direct control of dynamic interaction force are realized to overcome the limitations in conventional dynamic AFM and to enable the potential of high speed and high resolution dynamic AFM imaging.
As scanning speed increases, cantilever’s responses of high dynamic modes can be excited noticeably by tip-sample interaction force. These responses are distorted in the optical lever measurement system due to measurement sensitivity difference among distinct dynamic modes, which leads to tip position measurement error and therefore compromises image resolution. In this research work, cantilever dynamics of multiple dynamic modes is modeled and multi-mode measurement sensitivity calibration is realized. Based on the multi-mode cantilever model and the calibrated measurement sensitivities, two approaches, i.e., the modal projection filtering method and the multi-mode state estimation method, are developed to reconstruct actual tip position from the distorted optical lever measurement signal with sub-Angstrom level accuracy to help retain image resolution during high speed scanning.
AFM system relies on adjustment of cantilever’s z position/tip position to track sample topography. Therefore, speed and accuracy of tip position control are of paramount importance to the scanning speed and image resolution. In conventional dynamic AFM, tip positioning speed is severely limited by the low-bandwidth piezo actuator used as z-positioner and tip positioning accuracy is compromised by the thermal fluctuation of cantilever. In this research work, a novel active tip motion control system is developed to realize high speed and high precision tip positioning. In this system, a collocated electromagnetic actuation mechanism, used as a high-bandwidth z-positioner, and a model-based controller, designed according to the multi-mode cantilever model, are employed to extend the bandwidth of tip motion control over cantilever’s fundamental dynamic mode to enable rapid tip stepping in each tapping cycle. Cantilever’s thermal fluctuation is also actively suppressed through high-bandwidth feedback control so that tip positioning accuracy of sub-Angstrom level can be achieved even in liquid.
Amplitude modulation is commonly used for tip-sample interaction regulation in conventional dynamic AFM, whereas its regulation bandwidth is limited by the transient response of cantilever as well as the oscillation amplitude measurement delay. In this research work, dynamic sensing and direct regulation of the tip-sample interaction force are realized, which avoid the bandwidth limitation of conventional amplitude modulation method and enable the potential of high speed dynamic AFM imaging. By introducing the physical interaction process between tip and sample into a Kalman-filter-based state estimator and by estimating and compensating cantilever dynamic parameter variation, dynamic interaction force is precisely estimated with estimation resolution reaching the physical thermal force limit. Peak of the estimated interaction force in each tapping cycle is thus detected and directly regulated with a feedback controller through adjustment of tip-sample distance with the high speed and high precision active tip motion control system.
The enabling technologies developed in this research work are integrated to deliver a one-of-a-kind AFM probing system. A high-performance digital controller based on field programmable gate array (FPGA) is built, wherein various sensing, estimation and control algorithms are implemented for real-time computation and control with 1 MHz update rate. A standard calibration grid composed of circular hole array with 20 nm depth and 5 um pitch is imaged to prove the scanning speed advantage of developed AFM probing system over conventional dynamic AFM. Its scan rate is shown to be only limited by cantilever bandwidth. The image resolution is illustrated by simulated scanning of a sample of repeated spherical structure with 0.5 nm height and 6.2 nm pitch. Topography imaging and mechanical property mapping of live MCF7 human breast cancer cell in its native physiological environment are realized to demonstrate this system’s potential of biological applications.