Mario Viani

We have used a simple process to fabricate small rectangular cantilevers out of silicon nitride. They have lengths of 9–50 μm, widths of 3–5 μm, and thicknesses of 86 and 102 nm. We have added metallic reflector pads to some of the cantilever ends to maximize reflectivity while minimizing sensitivity to temperature changes. We have characterized small cantilevers through their thermal spectra and show that they can measure smaller forces than larger cantilevers with the same spring constant because they have lower coefficients of viscous damping. Finally, we show that small cantilevers can be used for experiments requiring large measurement bandwidths, and have used them to unfold single titin molecules over an order of magnitude faster than previously reported with conventional cantilevers.

Small cantilevers allow for faster imaging and faster force spectroscopy of single biopolymers than previously possible because they have higher resonant frequencies and lower coefficients of viscous damping. We have used a new prototype atomic force microscope with small cantilevers to produce stable tapping-mode images (1 μm×1 μm) in liquid of DNA adsorbed onto mica in as little as 1.7 s per image. We have also used these cantilevers to observe the forced unfolding of individual titin molecules on a time scale an order of magnitude faster than previously reported. These experiments demonstrate that a new generation of atomic force microscopes using small cantilevers will enable us to study biological processes with greater time resolution. Furthermore, these instruments allow us to narrow the gap in time between results from force spectroscopy experiments and molecular dynamics calculations.

One of the most popular methods for calibrating the spring constant of an atomic force microscope cantilever is the thermal noise method. The usual implementation of this method has been to position the focused optical spot on or near the end of the cantilever, acquire a force curve on a hard surface to characterize the optical lever sensitivity and to then measure the thermal motion of the cantilever. The equipartition theorem then allows the spring constant to be calculated. In this work, we measured the spring constant as a function of the spot along the length of the cantilever. The observed systematic variation in the spring constant as a function of this position ranged from for a short 60 µm cantilever up to for a 225 µm cantilever we examined. In addition, the thermally calibrated spring constants systematically disagreed with spring constants calibrated using the Sader and Cleveland methods: by for the short 60 µm cantilever and by for the longest, 225 µm cantilever. By using a model that accounts for the spot diameter and position on the cantilever, the thermally measured spring constants were brought into better than 10% agreement with the other methods.