Dominik Kohl

Auto-tuning PI controller for surface tracking in atomic force microscopy - a practical approach

Correct tuning of feedback gains is important for AFMs to cope with uncertainties of the system dynamics coming from a large range of different samples, cantilevers and scan parameters. For state of the art AFMs gains have to be adjusted manually by the operator. The typical operator such as biologist, physicist or material scientist may not have detailed knowledge about control engineering. In order to increase usability and acceptance an easy and intuitive approach is needed. The method presented in this paper is based on the commonly applied manual tuning strategy for AFM-imaging and copies the behavior of an experienced user. By spectral analysis ringing of the feedback loop is detected when feedback gains are increased beyond the stability margins. Increasing gains is stopped when an 1/f stop criteria is reached. The algorithm is successfully tested in simulation and practical AFM topography measurements with different cantilever - sample combinations, demonstrating that auto-tuning can be applied to achieve imaging performance close to ideal settings.

Evaluation of surface charge shift of collagen fibrils exposed to glutaraldehyde

Collagen fibrils are a major component of the extracellular matrix. They form nanometer-scale “cables” acting as a scaffold for cells in animal tissues and are widely used in tissue-engineering. Besides controlling their structure and mechanical properties, it is crucial to have information of their surface charge, as this affects how cells attach to the scaffold. Here, we employed Kelvin-probe Force Microscopy to determine the electrostatic surface potential at the single-fibril level and investigated how glutaraldehyde, a well-established protein cross-linking agent, shifts the surface charge to more negative values without disrupting the fibrils themselves. This shift can be interpreted as the result of the reaction between the carbonyl groups of glutaraldehyde and the amine groups of collagen. It reduces the overall density of positively charged amine groups on the collagen fibril surface and, ultimately, results in the observed negative shift of the surface potential measured. Reactions between carbonyl-containing compounds and proteins are considered the first step in glycation, the non-enzymatic reaction between sugars and proteins. It is conceivable that similar charge shifts happen in vivo caused by sugars, which could have serious implications on age-related diseases such as diabetes and which has been hypothesised for many years.

Water desorption in Kelvin-probe force microscopy: a generic model

Nanoparticles or similar, nanoscale objects such as proteins or biological fibrils usually have to be deposited from aqueous suspension onto a solid support surface for further characterization by atomic force microscopy (AFM) and related methods such as Kelvin-probe force microscopy (KFM). Here we show, on the examples of functionalized nanoparticles and collagen fibrils, that water desorption after sample preparation affects their electrostatic potential determined by KFM in a predictable manner. We explain this effect with a simple, analytical model based on the capacitance of the partially dielectric-filled tip-sample system. We also propose practical measures to avoid false interpretation of electrical AFM-based experiments. As the phenomenon is very generic it may have significant implications in the application of AFM to nanoparticles and other nanostructures including biological ones.

Automatic fourier synthesis based input-shaping for scanning piezo-electric actuators

This paper presents an automatic Fourier synthesis-based input-shaping algorithm for repetitive force-curve trajectories, applied to a piezo-electric actuator. A periodic scan-signal is generated by Fourier synthesis, consisting of the fundamental frequency and its even and odd harmonics. In order to compensate for the system dynamics of the used piezo-electric actuator, amplitude and phase response at these frequencies are identified. By adjusting amplitude and phase of each individual frequency component, the tracking error is reduced significantly. The presented method also demonstrates its capability of improving the non-linear behavior of piezoelectric actuators. The technique is experimentally evaluated on a piezo-electric stack-actuator, used as fast Z-actuator within an atomic force microscope (AFM). Comparing the presented method to an uncompensated triangular signal as well as input shaping methods, presented in literature, shows a significant reduction of tracking error of 84% and 65%, respectively.

Active damping by Q-control for fast force-distance curve measurements in atomic force microscopy

This paper investigates the benefit of active damping by an analog Q-control circuit for measuring fast force-distance curves in atomic force microscopy. By active damping of the cantilever oscillation after snap-off, the down-ring time-constant is reduced significantly from 385 μs to 23 μs. Experimental results demonstrate that the number of force-distance curves per second can be increased by a factor of more than 30.

Non-parametric robustness analysis for feedback motion control for a high precision stage with large mass uncertainty

Robust feedback control with large parameter uncertainties in a high precision stage with large sample mass variation is presented. Instead of expressing the nominal plant and its uncertainty by parametric models, they are expressed using the identified frequency response data. The open loop transfer function is mapped onto a Nyquist plot, to assess robust stability. Next, the complementary sensitivity function is derived from this data to describe tracking performance, which is verified by experiments. The advantage is that the parameter estimation step is omitted, without compromising robust stability and performance.