In a previous experiment, we characterized the drift along the XY axis by comparing a sequence of AFM images. To acquire a complete understanding of drift, we must also characterize drift along the Z axis. We define Z drift as an undesirable change in the distance between the probe and the sample surface.
The Z position of the probe is recorded as the Z piezo control voltage (Z Voltage) and/or the Z piezo position sensor voltage (Z Sensor). These two signals was evaluated by imaging a 1 [nm] field using 512x512 points at a frequency of 1 [Hz]. The raw trace and retrace signal for Z voltage and Z sensor is recorded and then processed and analyzed in Matlab. The 2D image data set is collapsed into a 1D time varying signal as shown in Figure 1. The blue stars Figure 1 represents the beginning of a scan. In the first 4 scans, the tip was disengaged before the scan. During the engage, the Z voltage changes from 0 [V] to ~70 [V] in an instant. Piezo actuators are not fast enough to respond to the demand to extend 70 [V] or 7.5 [μm] instantly, so as the piezo catches up to the control voltage, the control voltage must withdraw to keep the probe at the desired setpoint.
In the next 6 scans, the tip remains engaged. Since the piezo actuator is not moved very much, the Z Voltage is a lot more stable. Z Sensor records the actual displacement of the piezo actuator. From the Z Sensor reading, we observe a gradual drift for approximately 1 hour before it stabilizes. Once it stabilizes, the Z Sensor position varies by less than 10 [nm]!
Figure 1: The Z position of the AFM tip can be recorded as the Z piezo control voltage or the Z piezo position sensor voltage
Figure 2 compares the Z Voltage and Z Sensor signals. The signals have been offset for clarity. During this 1 seconds scan, the Z Voltage changes by 1.6 [nm] but the signal fluctuates by ~100 [pm]. On the other hand, the Z Sensor signal remains steady at 0 [nm], but fluctuates by ~ 1000 [pm]. In other words, the Z Sensor signal is accurate but noisy and the Z Voltage signal is clean but not accurate. For a clean and accurate signal, one must record the Z Voltage in a stable environment.
Figure 2: The Z Voltage signal is cleaner but less accurate than the Z Sensor signal
The Z Voltage and Z Sensor data teaches us about the characteristics of the piezo actuator and reveals the differences between monitoring/recording the Z Voltage or Z Sensor, but does not tell us whether or not the probe is drifting away from the sample surface. To figure this out, we perform many single force measurements over time.
In a single force experiment, the probe is extended towards the sample until it reaches a trigger point. Then it retracts a preset distance and idles there. The Z Sensor and Amplitude reading is recorded for each experiment. This single force experiment is ran 20 times at 7 minute intervals. Figure 3 shows the start and end position as a function of time. Observe that after 21 minutes or 3 experiment, the start and end position remain the same. When the system is stable, the gap between the probe and sample surface vary by less than 20 nm!
Figure 3: Z drift is less than 20 nm when the system is stable
From this investigation we learned that the position of the Z piezo is tracked via the piezo control voltage (clean but not accurate) and the piezo position sensor (noisy but accurate). After approximately 1 hour, the system reaches a stable state where the piezo drift is less than 10 [nm] and the gap drift is negligible. Under the right conditions, it is possible to idle an AFM probe a fixed distance above a sample surface.