Dear Xianglei,

Actually, It is possible to use AFM force spectroscopy technique for rough surface. But there are at least two problems:

1. Approaching. When you shift a point for force curves, you have new height of relief. It is possible to break a cantilever.

2. Contact area. If your substrate is very rough a probe can contact with surface non only by a tip. See an image (sorry, I am not good painter :)) 

What kind of sample do you have?

If your sample consists from particles and they are visible by light microscopy, you can stick these particles to glass or mica surface and adjust by light microscope a cantilever to single particle before force measuring. This technique is used for AFM force spectroscopy of living cells.    

Good luck!

Try to use etched tips: DOI: 10.1016/j.micron.2015.01.005

Remember, that all AFM's samples should be flat. That is why AFM may measure only distance in X, Y, Z axis. QNM/PeekFofce is only model - not straight measurement and results depends on many aspects.

Hi Janusz,

What did you mean by "model"? If it is QNM-visualization (each pixel is a force curve) - yes, it is modelling of mechanical properties distribution by Bruker software. If you have a single force curve (XY scan size = 0, ramp) - it is not a model, it is the real physical data. But in all cases of force spectroscopy we have the main question - what forces we have measured?! :) 

Xianglei, what model of AFM do you have?

If it is Veeco/Bruker be careful by approaching and put 0 to "XY scan size" field. 

Dear Denis,

Thank you very much for your answer. It's helpful!

Xianglei

@Denis Korneev: where is strain gauge? Nownere. How is the value of spring constant of cantilever taken? - from model. How spring constant of different tips change during test? No one knows. If it is measurement, could I take a Young Modulus value from any place on diamond/polymer/wood composite? No. How the inserting the tip impact on material? And how much of atoms of materials stays on tip, changing its radius? No one knows... All QNM is based on models, asumptions, and has no relation with the standard measurement of Young's modulus for example, in the static tensile test or hysitron indenter (in which I believe more).

@Denis, I have an AFM agilent 5500. @Janusz, Thank you Janusz for your great idea. Could you please send me a copy of the referenced paper?

@Janusz, I got the full text from your site. Thank you for sharing.

Dear Janusz, in general you are right, but it is necessary to be careful in terminology.

What is a modeling? We take some parameters and make a simple construction - a model, examine it and translate some obtained results to a real object. We say "we have examined a model, but it is equal to a real object, our results is representative!". It is modeling.

What is a measuring? We have a device (balance, voltmeter, etc.) and say: it can convert some real physical value to a number. It can be precision or very rough - in this case it does not matter - we have an act of measuring, it is non modeling.     

In our case: when we have a map of Young Modulus, which was created by AFM software - it is modeling. When we have just force-distance curve - it is result of measuring, not modeling.

It is not precision measuring - the force constant is not exactly measured (ThermalTune is just evaluation), there are some technical artifacts. Yes. But, it is a measured datum, not result of modeling.

When we start to analysis these curves, we can not escape from modeling - we use Hertz model, JKR, DMT, COS, etc. 

The vicious circle :)       

@Xianglei, I have no experience with Agilent AFM, I have used only Veeco/Bruker and NT-MDT. But, I have an universal advice - be careful! :) I have an experience of rough carbon particles AFM examination - a lot of cantilevers were broken...

Good luck!

@Denis: "The vicious circle :)" - terrific and accurate comment - honestly. But I am materials scientist and - according to ISO/ASTM standards when I talking about Young Modulus I know - with no doubt - what it is and how was obtained. Other "young modulus' is for biologists, other physicists, other for Bruker (DMT). What you know or what you believe - it is your choice.

@Xianglei, maybe piuma optics (http://optics11.com/products/piuma) product - dedicated to 'nanomeasurement' will be much moe accurate? Unfortunately working in same mode/manner as AFM in CM.

@Denis, Thank you very much for your suggestions. The Agilent AFM is not ready yet. I have to use the Dimension AFM. What is the size of your carbon particles?

@Janusz. Is the "Young's Modulus" for the cantilever? Could you please explain in details your "Young Modulus"? Thank you!

Young's modulus E is the ratio of the increase in tension in a corresponding increase of elastic extension. Force is measured via strain gauge  - but only in the region of full elastic deformation (of sample), according to Hooke's law. The main reason to see how E is important is fact, that it is a material's constant, which means that the material will be characterized by the same modulus always, if measured under similar conditions. That is why there is no way to put - even polymer sample to do some measurement DMTMODULUS with AFM and compare it to Young Modulus measured according to ISO/EN/ASTM/PL/Other standard. There is a lot of ideas to measure modulus of elasticity even using ultrasound, but it is always model.

Young modulus of cantilever should be differ and depend on production process, covering, crystalographic orientation, etc. Usually You will no need to know E of silica cantilever because spring constant is more important for assumptions to estimation of material's elastic/plastic/deformation/adhesion/other properties. Usually spring constant is obtain from thermal oscillation (another model).

@Xianglei, I had some large (100-500  µm) carbon particles, stuck them to a glass slide by cyanoacrylate and positioned the cantilever by light microscope to single particles before the scanning.  

@Janusz, you are right. But it is not all the difficulties. The geometry of probe's tip is very important. What is the contact area? Which objects are on the tip? Etc. Especially it is important for the force spectroscopy. We use TEM for AFM probes visualization.         

@Denis Korneev, yes, know it but I though we talking about modelling above - tip estimation is implemented in all AFM/SPM software (yes, another model). If not, SPIP or Gwyddion got it. Taking into consideration shape, I modified shapes of it with FIB... as mentioned. And, depends on model (yes, another one) the contact area may be cone or ball/sphere what has 25 - 45% impact on DMTmodulus value :)

@Denis, how did You measure tip with TEM? What kind of holder You got? It is Your own construction? I did observation but first have to prepare special holder for  SEM and FIB.

Best Regards and good luck!

@Janusz, it is possible to use a standard high tilt holder. We use Jeol EM-21010 for TEM Jeol JEM 1400 with max (60-65 deg) tilt of goniometer.

Good luck!