The Nanoscale World

200N/m Deflec

rated by 0 users
Not Answered This post has 0 verified answers | 3 Replies | 2 Followers

Top 100 Contributor
10 Posts
Points 116
meaton212 posted on Wed, Sep 21 2016 9:13 AM

Hello,

 

I do a lot of QNM experiments of my research. Typically, I use the RTESPA series. Out of the 3 I use (5N/m, 40N/m, 200N/m), I find the calibration of the 200N/m the hardest to get consistent numbers for. Which is kind of a pain, since the research I do requires very accurate values of Modulus measurements. Below are Deflection Sensitivity checks on the same 200N/m tip with the Bruker Sapphire sample. Both yield widely different deflection sensitivity values 

Deflection Sensitivity 1 (129.37 N/m) 

Deflection Sensitivity 2 (106.89 N/m)

 

In terms of Spring Constant, I am still having more issues. I usually do the Sader method, since the thermal tune doesn't work for stiff tips, but again I get a lot of variability in the values of Q that are given. For 200N/m tips, these range from like Q = 800-1100, leading to Spring Constants of 140-210N/m, all for the same tip.

 

I've tried using the relative method, using the Bruker Polystyrene sample, rated 2.7GPa, but it always seems to over estimate by a lot. with starting at 200N/m, the modulus is given as something like 40-50GPa, and I usually have to decrease the spring constant to well below 100N/m until a value of 2.7GPa emerges.

 

I guess my question revolves around... how to get consistent numbers for deflection sensitivity and spring constant for the stiff 200N/m that are not all over the place?

 

Thanks

  • | Post Points: 12

All Replies

Top 50 Contributor
20 Posts
Points 220

For your two examples of different deflection sensitivity with the same probe, did you change the laser alignment between the measurements? It is normal for the deflection sensitivity to change if the alignment is changed, even with the same probe. If the alignment did not change, there is a problem somewhere.

In your screen grabs, you are plotting deflection against Z - you might want to try changing the X data type to Z sensor to avoid any issues with nonlinearlty in the Z piezo.

I have heard that it is better to measure the deflection sensitivity offline using nanoscope analysis (as opposed to using the nanoscope realtime software) as the NSA performs a linear fit to the curve between the markers while NSRT just subtracts the z position and deflection at one marker from the other and divides the difference in Z by the difference in deflection.

It is possible that the tip is both elastically deforming and blunting with each approach curve, due to the high force from the stiff cantilever and the hard sapphire surface. The elastic deformation of the tip will depend on its size, so if it is blunting you will have different deformation with each curve and so a different deflection sensitivity. You could test if this is happening by taking several curves one after the other and seeing if the tip eventually becomes blunt enough the the deflection sensitivity approaches a constant value. However, if you are applying enough force to blunt the tip, there will certainly be elastic deformation happening too. This will mean that your deflection sensitivity in not accurate, even if it approaches a constant value, so you will have a systematic error in your measurement. Using a tip made from a harder material (e.g. diamond) or switching to a different measurement (e.g. nanoindentation) are possible ways around this.

It is also possible that you are picking up contamination from the sapphire - it would be a good idea to image the sapphire sample to check for this.

For k calibration, have you tried the thermal method, or just been advised that it will not give good results? Often the thermal method is not recommended for stiff probes because the thermal peak is small and can be difficult to fit. If you are able to obtain a good thermal spectrum with a clear peak, it is likely that it will work fine, as long as you can solve the deflection sensitivity problem above. 

For the Sader method, how are you measuring the Q? As for the thermal method, if you are using the thermal noise spectrum to get the Q, the peak must be clear enough to get a good fit - if the spectrum is not good enough for the thermal method, it is unlikely to give you an accurate enough Q for the Sader method.

Using the Q calculated from a driven cantilever tune usually gives inaccurate results, as the peak often has contributions from the cantilever holder, tapping piezo, spring and support chip. These can distort the amplitude and phase response so giving inaccurate Q values.

One way to get an accurate Q if you do not have a clear thermal noise spectrum is to drive the cantilever at resonance, then set the drive amplitude to zero and use the HSDC function to record the cantilever waveform as it rings down. You can fit the exponential decay of the amplitude as a function of time and get an accurate Q from that.  

Hope that helps.  

  • | Post Points: 12
Top 100 Contributor
10 Posts
Points 116

Thank you so much for your response

 

How do you run the HSDC at the resonance frequency? According to the manual, it says Amplitude monitoring occurs at 500kHz. How do you adjust this? Also, how do you switch off the drive frequency in the middle of a capture?

  • | Post Points: 12
Top 50 Contributor
20 Posts
Points 220

meaton212:

How do you run the HSDC at the resonance frequency? According to the manual, it says Amplitude monitoring occurs at 500kHz. How do you adjust this? Also, how do you switch off the drive frequency in the middle of a capture?

The amplitude channel responds too slowly for this measurement - it is best to do it using the vertical deflection channel, which can be sampled much faster.

To measure the rigndown using HSDC, set up the AFM as normal, but keep the probe well away from the surface. Tune the cantilever on resonance, then open the HSDC panel and set it as follows: rate: 50 MHz, data type: "vertical deflection", trigger type: "edge", slope: "negative", trigger channel: "amplitude", set the level to a value just below your free amplitude, delay: -100 ms (the minus sign is important!), duration: 330 ms.

Now false engage the microscope (Ctrl, Alt & F), the microscope should begin scanning but well above the surface. 

Press "arm trigger" on the HSDC panel, then set the drive amplitude to zero. The HSDC will trigger on the reducing amplitude, but the -100 ms delay means it will keep the data from 100 ms before you turned off the drive.

If you open the resulting HSDC file, you'll see a wide "stripe" for the first 100 ms, which is the trace of the oscillation cycles of the cantilever while the drive signal is on (zoom in and you'll see the sinusoidal profile of the oscillation), after which the stripe gets much thinner as the amplitude drops down to zero. If you zoom in around the 100 ms point, you'll see the gradual decay of the amplitude after the drive is switched off. Fit this decay to a single exponential and the number of cycles it takes for the cantilever amplitude to reach 1/e of its initial value is the Q. 

"Vibrations and waves" by French has a good chapter on driven damped oscillators if you want to read up on theory.

Cheers,

Nic

 

  • | Post Points: 10
Page 1 of 1 (4 items) | RSS
Copyright (c) 2011 Bruker Instruments