Surface modification of AFM silicon probes for adhesion and wear reduction

Abstract

Tip wear of silicon probes used for an atomic force microscope (AFM) is a critical issue. Wear can result in an increase of tip radius and adhesion between tip and sample, thus reducing the image resolution and introducing artifacts. In order to reduce adhesion, friction, and wear so as to reduce tip related artifacts, liquid lubricant (Z-TETRAOL), self-assembled monolayers (pentafluorophenyltriethoxysilane (PFPTES)), and fluorocarbon polymer (Fluorinert™) were applied on the silicon probe. A comprehensive investigation of adhesion, friction, and wear of the uncoated/coated tips in both ambient air and various humidity levels as well as the influence of the coatings on the image resolution was performed. Experiments showed that the coatings reduced the adhesion, friction, and wear of the silicon tip, improved the initial image resolution, and exhibited less deterioration as compared to that of uncoated tip in the long-term test.

Appendix: Tip characterization principles

The tip geometry can be measured independently using scanning electron microscopy (SEM) or transmission electron microscopy (TEM) and can also be measured by AFM imaging an ultrasharp geometry (sharper than AFM tip), known as tip characterizer, and analyzing the data for tip reconstruction using a tip deconvolution algorithm. An SEM or TEM measurement takes a long time and requires additional handling of tip. Therefore AFM technique is widely used for tip geometry measurement. In order to use this procedure, the actual shape of the tip characterizer must be known with some certainty. If it is not practical to obtain the actual characterizer geometry, blind estimation of the tip shape is an alternative.

The tip reconstruction methods are discussed as follows. Based on [34], an AFM scanned image is a dilation of the real sample surface and the reflection of the probe tip surface. That is

$$i(x,y)=\mathop{\max}\limits_{(x^{\prime},y^{\prime})}[s(x-x^{\prime},y-y^{\prime})+p(x^{\prime},y^{\prime})]$$

(A.1)

where (x, y) and (x′, y′) are in plane coordinates, and i, s, p are the functions of the top surfaces of the scanned image, the real sample (tip characterizer), and reflected tip through the origin, respectively. Since p is the reflection of the tip, it relates to the function t, describing the tip surface, by

$$p(x,y)=-t(-x,-y)$$

(A.2)

In Eq. (A.1), if the sample surface (s) is known, then the tip shape can be obtained from the scanned image and the sample surface:

$$p_{r}(x,y)=\mathop{\min}\limits_{(x^{\prime},y^{\prime})}[i(x+x^{\prime},y+y^{\prime})-s(x^{\prime},y^{\prime})]$$

(A.3)

where p _r is an upper bound on the true tip surface (P _r ⫆ ⊇p).

Different tip characterizers with known geometry have been developed. Examples of the characterizers used are colloidal gold [35], polystyrene and glass sphere [36], and biomolecules [37].

In the case the tip characterizer surface (s) is unknown, the upper bound on the true tip surface can be determined by giving an initial value (p ₀ ) and iteratively deriving the upper bound by [34]

$$p_{i+1}=\mathop{\min}\limits_{x^{\prime}\in D_I}\left\{{\mathop{\max}\limits_{d\in D_{p^{\prime}}}\left\{{\min\left[{i(x+x^{\prime}-d)+p_i(d)-i(x^{\prime}),p_i(x)}\right]}\right\}}\right\}$$

(A.4)

where D _I and D _p′ designate the domain of I and p′, respectively, and d is forbidden value.

In equation (A.4), the tip shape is expressed as the function of the measured image (I) and the tip surface itself (p). The sample surface is not involved in determining the tip shape. This is called “blind tip reconstruction” [34]. In order to get a desirable estimation of the tip shape using blind tip reconstruction method, the sample should contain sharp features and high relief so that all the points on the tip surface are in contact with the sample surface.

Tip characterizers based on the “blind tip reconstruction” principle have been developed. Spikes and columns [38] are examples using this principle.