Fig. 3: Effects of time-varying cantilever resonance in AFM-IR. | Nature Communications

Fig. 3: Effects of time-varying cantilever resonance in AFM-IR.

From: Closed-loop atomic force microscopy-infrared spectroscopic imaging for nanoscale molecular characterization

Fig. 3

a Amplitude of the cantilever deflection voltage demodulated at fixed 1007 kHz (6th resonance mode) resulting from driving the piezo amplitude from 0 to 1 V within 0.5 s (single scan). The recorded deflection signal was scaled by an averaged measurement of responsivity to match the closed-loop signal |u| while retaining the noise behavior of the raw deflection signal. Data points represent samples taken from a single representative scan plotted with mean and 95% confidence intervals (CI) of 50 repeat scans. Signal variance (noise) was observed to be constant for low signal (below. 1V) and then increases linearly with the cantilever deflection. b Amplitude of the closed loop (CL) responsivity corrected signal under the same conditions as (a). The CL signal exhibits constant noise equal the minimum noise achieve in the OL signal (i.e. at 0V from (a)). (c) Signal to noise ratio (SNR) for the OL and CL ramped signals from (a) and (b) respectively. SNR was also measured using the signal produced by a 400 nm SU8 film heated by a Quantum Cascade Laser (QCL) at discrete laser power levels for both OL and CL modes. The laser power was correlated to the ramp voltage (x-axis) using the relation between the CL signal and laser power. The SNR of the QCL driven signal exhibits the same behavior as the piezo ramp measurements suggesting the OL noise scales with cantilever deflection (not laser power) most likely due to time-dependent variations in the cantilever’s resonance behavior. d, e OL amplitude and phase responsivity corrected AFM-IR images of chromosomes from a metastatic breast cancer cell line (MCF10A series) collected at fixed 360 kHz (3rd resonance). f, g Same sample and conditions as (d) and (e) except operated in CL. Both OL and CL were collected with resonance-enhanced sensitivity; however, the OL signal is degraded by rapid time-variations in resonance resulting in high noise and inaccurate phase signal due to processing the ratio of noisy measurements.

Back to article page