1.1 Scope

Cleanliness verification is growing in its importance in many industries, e.g. aerospace, biomedical engineering, and semiconductor fabrication [1], and the sessile drop technique stands out as an inexpensive, versatile, and portable way to probe the wettability of a surface [2], which is correlated with the cleanliness of hydrophilic (i.e. metallic and metallic oxide) surfaces. This chapter addresses the underserved community of users who clean surfaces that are too large to be placed in a small instrument or for whatever reason are not amenable to offline study. Hydrophobic oils, greases, and soils are their most common contaminants. This chapter may be less relevant to the user who is able to analyze small parts, wafers, and coupons using a commercial goniometer with subdegree accuracy and precision.

The ultimate goal is to facilitate the confident use of the sessile drop technique on large surfaces for contamination control and cleanliness verification. By “confident use,” we mean techniques and procedures that possess verifiable accuracy and precision suitable for incorporation into a company’s quality management infrastructure. By “cleanliness verification,” we mean verification that the surface is suitably prepared for the next production step or end use. This is not necessarily the same as a contamination-free or pristine surface. Rather, knowledge of the surface energy required for optimum performance will be useful in writing specifications for cleanliness, and this chapter should assist in the verification that the desired surface energy specification has been achieved.

This chapter examines the utility of sessile drop contact angle measurement for surface energy determination and cleanliness verification. A review is given on the available methods, commercial instruments, patents, and literature describing the state of the art in contact angle measurement. Then, a description is given on contact angle measurement techniques that have been modified for use on large surfaces. The negative effects of these changes on accuracy and precision are discussed, and remedies are given including the use of standard reference objects (SROs) [3,4] that mimic the size and shape of sessile drops.

1.2 Surface Cleanliness and Surface Energy

Surface energy (free energy per unit area) and surface tension (force per unit length) are essential concepts for describing the characteristics of solid–liquid interactions [5]. A clean metal or metal oxide surface will typically have a high surface energy. Liquids, adhesives, and polymer melts will spontaneously coat a high-energy surface as long as the surface tension of the liquid is lower than the surface energy of the solid. Contamination, especially hydrocarbon soils, will lower the surface energy of the substrate leading to incomplete coating, adhesion failure, delamination, etc. Therefore, knowing the surface energy, types of soils, and surface tension characteristics of coatings and adhesives is essential for confident and functional cleanliness verification.

The surface tension of a liquid is typically measured directly using tensiometry, while the surface energy of a solid is determined indirectly using the wetting behavior of test liquids with known surface tensions [6]. The wetting behavior is easily measured using the contact angle θ of a sessile drop. The interplay of surface energy, surface tension, and contact angle has been described in great detail since the 1960s by Zisman [7] and many others.

Zisman developed a standard technique to determine the surface energy of a smooth, planar surface by studying the contact angle behavior of probe liquids of varying surface tension [7]. Although examples are given in Zisman’s work of polar liquids and polar surfaces, the Zisman Plot uses a one-parameter approach to the surface tension of the liquid phase and surface energy of the solid substrate. Owens and Wendt [8], Rabel [9], and Kaelble [10] added a polar parameter, and eventually a hydrogen-bonding parameter to form a system similar to that of Hansen [11]. Alternatively, a description can be made in terms of dispersion and Lewis acid–base interactions as developed by van Oss, Chaudhury, and Good [12]. To save time in reviewing the above developments, the interested reader will find the review of these theories concisely delivered on the Krüss web site [13].

But for those new to surface energy determinations, an example of a simple Zisman analysis is given. In a Zisman plot, the cosine of the contact angle, cos θ, of each liquid is plotted against the surface tension of each liquid (γ_lv—the liquid–vapor interfacial tension). A line is fitted to the contact angle measurements and extrapolated to find the critical surface tension (γ_c) where spontaneous wetting occurs (i.e. cos θ = 1). Any liquid with a surface tension less than γ_c will completely wet the surface. Figure 5.1 is a Zisman plot generated using aqueous solutions of sodium dodecyl sulfate on an aluminum surface.

Although the Zisman plot is simple enough to analyze, we have found that a slight modification of the plot speeds the analysis considerably. In the modified Zisman plot (Fig. 5.2), the surface tension of each probe liquid is plotted versus “1 − cos θ”. Almost all plotting packages are capable of displaying a trend line with a linear or polynomial fitting equation. The critical surface tension is the y-intercept in this modified plot. An analysis of variance routine on γ_lv vs 1 − cos θ yields the critical surface tension (65.5 mN/m) with an estimate of the standard error (0.4 mN/m). The modified plotting technique’s ability to estimate the uncertainty in the critical surface tension is a distinct advantage over the traditional plotting technique.

If one is using very pure liquids, then the literature values of surface tension may be used. But if solutions are used as probe liquids, then one must measure the surface tension. The DuNouy tensiometer [14,15] uses a platinum–iridium ring that is pulled out of the vapor–liquid interface, and the force pulling on the ring is used to calculate γ_lv.

The authors have shown how a digital version of this tensiometer may be constructed from an analytical balance and a hydraulic press [16]. The balance must have a hook for weighing objects below the balance and must be able to communicate with a computer. The top platen of a Carver-type press typically has a hole in it. A ball chain may be allowed to hang through this hole from an analytical balance that is resting on the top platen. The other end of the ball chain holds the platinum–iridium ring. The lower platen holding the test solution is raised until the ring is submerged. Then, the hydraulic fluid in the press is released to slowly lower the liquid while the balance is recording the force on the ring as it passes through the liquid–air interface.

The ring tensiometry technique is easily checked against pure liquids to ensure that the instrumentation and the technician are producing accurate results. The weakness of all the surface energy analyses of Zisman et al. for cleanliness specification and verification lies in the uncertainty of the contact angle measurement. The contact angle is affected by surface contamination, roughness changes, surface tilt, liquid purity, liquid viscosity, surface reactivity, etc. Kumar and Prabhu review many of these factors in detail [17]. This strong dependence on the state of the surface illustrates the excellent sensitivity displayed by the sessile drop contact angle.

2.1 Traditional and Newly Digitized Contact Angle Methods

Before discussing contact angle measurement validation, it is appropriate to describe some of the contact angle measurement techniques and analysis methods. By far the most common commercially available instruments view the drop profile with back illumination (Fig. 5.3). The analysis has been automated using computerized image analysis algorithms. These instrument vendors provide their own method validation procedures, and some supply validation slides (SROs) [3,4]. One drawback, however, is the inability of most of these commercial instruments to travel outside of the laboratory to the production floor, paint bay, or field.

Since most of the contact angle analysis methods are based on the geometry of a perfect sphere, one must use small drops on a level planar surface, although nonlevel and curved surfaces [18,19] have been addressed. According to Extrand and Moon [20] based on Eqn (5.1), a 10 µL water droplet will be spherical if it adopts a shape with a contact angle between 10° and 140°. Equation (5.1) describes the maximum spherical volume (V_max in µL) in general terms suitable for any liquid where g is the acceleration due to gravity (9.81 m/s²), γ is the liquid surface tension in mN/m, and ρ is the liquid density in g/cm³.

(5.1)

For models that depend on drop volume (e.g. Bikerman [21]), the uncertainty in drop volume is a large source of uncertainty in the resulting contact angle measurement. Some models (half-angle and Brugnara [22]) are insensitive to volume as long as the drop is spherical. Some drop shape analysis (DSA) routines (LB-ADSA [23]) model the gravity-induced shape, and still others (DropSnake [24]) do not depend on the shape of the drop at all. Each of these methods and their freely available software packages are described below.

2.2 Suggested SRO in the Literature or in Industry

The concept for a >90° 3D contact angle standard has been mentioned, for instance in ASTM D 5725-99 [37]. But this and other mentions in the literature did not contain enough detail to instill confidence in the production and use of such an SRO. Some instrument manufacturers provide [3,4] high-quality contact angle images imprinted on glass slides which are placed on the goniometer stage in place of an actual sessile drop (Fig. 5.11). These slides are examples of a 2D SRO, and they work well in situations where the illumination, sample pedestal, and camera have a stable, side-on geometry. But these slides are less useful if one is adapting the illumination and camera positions to enhance the contrast and focus of an actual sessile drop on a large surface that cannot be brought to the lab. These slides cannot be used for top–down or reflected-angle methods, either. In general, it is the best when the calibration object closely mimics the size, shape, and reflective characteristics of an actual sessile drop.

2.3 SRO Materials

A >90° contact angle standard can be constructed using a 3.18 mm chromium steel ball (MSC Industrial Supply Co. #72660) mounted in various drill gage card (Grainger #5C732) holes to mimic a series of contact angles (Fig. 5.12). Less than 90° contact angle standards can be constructed using 6.35 mm and 9.53 mm diameter balls (MSC Industrial Supply Co. #72702 and #72744) mounted under small sheets of punched aluminum (Fig. 5.13). The ball diameters and holes were chosen to produce faux sessile drops with volumes near 15 µL.

These types of standards have proven to be useful in comparing four different side-on methods and multiple student technicians [25]. The main advantages of these standards are the fact that they are rigid, nonevaporating, very spherical in shape, portable, and resistant to damage.

The <90° SROs exhibit one drawback. Since the ball is protruding through a cylindrical hole, there is a gap near the three-phase point. A machined hole with a matching spherical contour along the walls of the hole would eliminate this gap almost completely.

The possibility of printing a 3D SRO was explored using a uPrintSE (Stratasys, Inc) [38]. It was found that the extrusion nozzle of our 3D printer was too large for a high-fidelity reproduction of a 10 µL sessile drop. It was apparent to the naked eye, and under microscopic inspection, the drops were very rough and resembled bee hives.

3.1 Personnel Training

A select set of sessile drop images can be used to test the operator’s competence with the software. This is a technique used at our university to ensure that new research students are able to precisely and accurately analyze the sessile drop image data. But there is more to contact angle analysis than image analysis.

A realistic 3D SRO allows the evaluation of an operator’s ability to align, illuminate, and capture high-quality images of sessile drops. We have found that drop illumination and camera elevation angle are the two skills with the steepest learning curve. The SRO provides an objective target to this seemingly subjective category of image quality. A quality image produces an accurate contact angle, and accurate contact angles instill confidence in surface cleanliness decisions.

3.2 Method Comparisons

Strength of the 3D SRO is the ability to compare methods. The ball in a hole allows the comparison of all the side-on methods. But this SRO is not appropriate for the top–down and reflected-angle methods. The ball protruding through a hole allows the comparison of all the methods described in this chapter.

4.1 Examples of Imaging Choices

The choice between a glass prism (Edmund Optics) and a bent electropolished metal mirror (Rimex Inc, Edison, NJ) was evaluated against the Ramé-Hart 2D SRO. The metal mirror is preferred for cost and durability factors. As an example, the results are shown for the 119.5° drop image in Table 5.1. It is important to note that the uncertainty with the mirror is only slightly more than the prism. However, the accuracy is compromised by the slightly distorted image of the bent metal mirror (Fig. 5.14). This judgment is impossible unless one can use an SRO to calculate the accuracy values, and therein lies the whole motivation for using an SRO in a cleanliness verification quality management plan.

4.2 Example of Performance Comparison

The prism was used in the arrangement shown in Fig. 5.6 where the sessile drop was replaced by the 2D SRO (Fig. 5.11). Three images of each drop profile were analyzed in ImageJ using the half-angle method. The absolute error was calculated by subtracting the accepted value from the experimental value. Table 5.2 shows the accuracy of this imaging method in terms of the mean absolute error and the precision of this imaging method in terms of the standard deviation of each set of three observations. The pooled standard deviation for this imaging method is 0.3°.

4.3 Example of Personnel Training

The 3D SROs were used to evaluate the performance of three operators (DP, DW, and EN). The operators varied in their experience from over 1 year (DW) to 1 week (EN) to 1 day (DP). The contact angle measurement methods were the Circle and Ellipse methods of Brugnara and the DSA and Snake methods of Sage [25]. The accepted value for the SRO was determined using the calibrated half-angle method. Figure 5.15 shows the accuracy and the precision of the operators and methods.

Clearly, DP had accuracy and precision problems with the Ellipse method and should be retrained. The operator NE initially had difficulty with the Snake method, but self-corrected when the analyses were repeated. The ability to evaluate accuracy objectively using an SRO instills confidence that new operators are performing within acceptable limits.

4.4 Examples of Method Comparisons

The Langmuir method described herein has been evaluated against two 3D SROs constructed to exhibit contact angles of 48.1° and 61.0°. The half-angle analysis of these SROs gives the accepted values of 45.8 ± 0.2° and 61.6 ± 0.3° (N = 6 each). Figure 5.16 shows the comparison of the accuracy and precision of the Langmuir (L) method to the half-angle (H) method against these standards.

The Langmuir (L) method matched the half-angle (H) method to within a degree (Fig. 5.16). The mean error of this method was slightly below zero at −0.6 ± 0.5° (N = 12). The standard deviation of this method (0.5°) is acceptable for many operations, and is comparable to the standard deviation of the half-angle method (s = 0.2°).