As discussed in the appendices A and B of my book, reference and calibration samples are extremely important for AFM. Calibration samples allow the AFM to make accurate, quantitative measurements. Reference samples can be used for training, and certification, and just to check if the instrument (or the probe) are performing as they should. There's a list of commercially available calibration and reference samples here.

AFM piezoelectric scanners tend to change their responses and age over time, so it's necessary they they be recalibrated periodically (typically annually).  All these repeated uses can lead to reference samples becoming dirty and contaminated. This leads to serious problems, because a contaminated samples cannot be properly used for calibration of an instrument, and may also be difficult to use as a reference sample. The article below explains some techniques that can be used to clean AFM calibration and reference samples. Note that this method o sample cleaning first appeared at I learned the "NewSkin" method while visiting their headquarters.



 Although they can be expensive, reference samples are often improperly stored and/or generally mishandled in lab environments, leading to an accumulation of contamination and particulates on the sample surface. Once they're contaminated, it becomes difficult to use the reference sample as a calibration for an AFM. Contamination makes probe approach difficult and causes false feedback that makes scanning a challenge. Take a moment before running your reference sample to view the sample under an optical microscope, where the larger surface contamination can be readily observed. The photo on the below illustrates heavy contamination present on a standard silicon test grid reference sample.

Cleaning Heavy Contamination from a Silicon Test Grid

There are many ways to clean reference samples, but one novel approach for a heavily contaminated silicon test grid utilizes an inexpensive product readily purchased from a pharmacy, called NewSkin.This is a dilute polymer layer, which can be used to remove gross contamination from surfaces. Here's the procedure. First, coat the contaminated reference with the NewSkin. There's a brush built right into the bottle that makes this an easy task. This is shown below. It's best to dab the liquid on, not brush the surface as you might damage the structures.


Next, allow the New Skin to dry on the sample. This can take several hours, so it's best to apply in the late afternoon and allow the New Skin® to harden overnight. The image below shows a dried layer of NewSkin on our calibration sample.


Once the NewSkin is dry, it pops up a bit, making a convex surface. This makes it easier to remove with a pair of tweezers, as shown below. Be careful not to damage the surface of the silicon sample.

By pulling the hardened NewSkin off of the reference sample with tweezers, you're also removing larger surface contaminants from the reference.

Finally, take a look at your silicon test grid reference sample under the optical microscope to see if the larger contaminants are now removed.

Cleaning Microscopic Contamination for AFM Calibration

Although the New Skin® method is successful in removing large contaminants such as dust and fibers, you will need a technique that can remove microscopic contaminants and organic layers, such as airborne hydrocarbons, if the sample is to be used for calibrating the AFM.

 Short wavelength UV light, in air, can produce ozone, which is a strong oxidising agent. The combination of UV irradiation (<300nm), and ozone can rapidly decompose a wide variety of organic contaminants, leading ultimately to the production of carbon dioxide and water vapour. Thus, this is a rapid and effective procedure for removal of organic matter from sensitive surfaces. Such devices can be constructed fairly simply*, and commercial devices are also available (e.g the surfinator).

Samples can be cleaned by UV irradiation consistently only if gross contamination is removed first. Typically this pre-cleaning step would need to involve a sequence of cleaning steps involving multiple solvents. These include swabbing, rinsing, and ultrasonic agitation. For suitable pre-cleaning, the cleanliness of the solvents themselves is also important. Depending on the efficiency of the UV and ozone generation, and the distance from the UV source, this UV/ozone cleaning procedure can require between 20 seconds to several hours to completely clean a surface. In addition, some samples are not resistant to the pre-cleaning procedures, for example water-sensitive materials. In the case of cleaning either AFM probes, or calibration samples, solvent resistance is not necessarily an issue, but the surfaces are very sensitive to mechanical damage. For example, swabbing with even the softest materials is not appropriate for calibration specimens, since tiny changes to the shape of the surface features caused by mechanical scraping, even at the nanometer level, will be a significant problem in terms of AFM calibration. Removal of large contaminants from AFM cantilevers / probe tips is highly problematic, since they are so mechanically fragile.

 An alternative to the UV-ozone cleaning procedure is carbon-dioxide snow cleaning (2,3). In this procedure, a compressed CO2 gas source, and a specialised nozzle are used to produce a flow of carbon dioxide “dry ice” particles. These can mechanically remove particles from the surface, without causing significant physical abrasion. The carbon dioxide also typically forms liquid CO2 at the surface, acting as a solvent, allowing removal of many organic contaminants. This procedure is claimed to be highly effective, however, it requires a compressed CO2 source, as well as the specialised nozzle which costs several thousand dollars.

Overall, there is a need for a low-cost mechanism to remove gross contamination from surfaces such as AFM calibration specimens simply, and without risk of mechanically damaging such surfaces. The method above, using NewSkin, can fulfil this need. overall, the two methods discussed here, can remove large and small contaminations, leaving clean calibration samples, suitable for use with the AFM. 

(1) J. Vig, J. Vac. Sci. Technol. A3 (1985)

(2) R. Sherman and P. Adams, Proceedings, Precision Cleaning, p.290 (1996).

(3) D. A. Chernoff and R. Sherman, J. Vac. Sci. Technol. B28, 643 (2010).