Forums
Bruker Media
Community Media
Bruker AFM Probes
SPM Digest
Application Notes
NanoTheater
Website
中文
Brochures & Datasheets
Publications
Probes Catalog
Events
Manuals & Documentation
Presentations
Guide to AFM Modes
News
Journal Club
Webinars & Video
Nanovations
Other
Hello all,
we have a problem regarding the force curves recorded by the Catalyst: sometimes (without any obvious reproducibility) the baseline is tilted. Further the trace/retrace correspondence seems to make no sense then. To illustrate the problem, I add an extreme example:
The height sensor data does not differ between constant and tilted baselines. However the effect seems to get stronger with increased ramp sizes, which leaded us to the opinion it could be a piezo-issue. First we thought the problem is solved with closed-z-loop but this observation turned out to be wrong. Also I wanted to ask if it will make more sense to rotate the whole curve back for analysis or if the baseline tilt should not influences the indentation part. In our experiments on hard substrate we had to tilt the indentation part too in order to get reasonable calibration curves (right angle in force separation).
I am glad for every hint,
with best regards,
Leon
Hi Leon,
There can be a lot of different reasons for sloped or curved force curve baselines. Some are artifacts and others are real contributions to deflection. Drift, long range forces (e.g. electrostatic), optical interference, and poorly shaped reflected laser spots are common reasons. Tip scanning AFMs often have some inherent baseline tilt because the laser spot moves on the detector as the probe moves up and down. But this tilt should be quite small, linear and very consistent. You note that you are only observing excessive tilt intermittently, so I expect that you are seeing something more than just this expected tilt. Next time you see the issue I would suggest adjusting the laser alignment slightly to see if that improves it. Sometimes the reflected laser spot will be misshapen if the alignment is off and that can lead to deflection artifacts.
There's not always an obvious right answer about how to correct force curve baselines offline during analysis. Some of the causes of tilted or sloped baselines go away or change once the tip touches the surface, so applying a uniform correction to the whole curve is not always the right answer. But applying different corrections to the non-contact and contact portions of the curve isn't always easy either because the contact point in the curve is not always well defined. I've seen both approaches taken in publications and commercial force curve analysis packages.
One thing that might help would be to use as soft of a cantilever as possible to maximize the deflection at the desired force. Effectively you want to make the measured deflection large such that any uncertainty due to baseline errors is small in comparison.
Dear Leon,
From your message is probably missing the most important information, namely what kind of samples you are working on.
Since you are using a Catalyst facility I guess you may rather be working on cell samples.
It comes out that for cells, there can be some proteins floating around on top of the cells that will be acting as brushes and may be generating the kind of deflections you are seeing.
Because of this effect, I don’t think that I would advise you to make a baseline correction, since the needed correction will not be the same for the non-contact and contact part of the curve.
Nevertheless, the presence of such baseline slope may make it difficult to figure out the position of the curve origin. And given that the Young’s modulus calculation (with the Hertz model for example) result will be different for different positions of the curve origin, this may be quite problematic to analyze such curves.
I have published a paper (P. Carl, S. Hermann, Elasticity measurement of living cells with an atomic force microscope: data acquisition and processing, Pflugers Arch - Eur J Physiol. 457, 551–559 (2008)) where I describe how to calculate the Young’s modulus value (using the Hertz model) by just performing a simple linear fit and without having to determine the origin position of the curve.
Thus with this method, the analysis of indentation data of sloped baseline curves becomes really trivial and there is anymore no need for you to try to fix this problem.
Feel free to contact me for any additional question or purpose.
Let the force be with you,
Philippe