Surface Wettability Control Without Knowledge of Surface Topography and Chemistry—A Versatile Approach

Abstract

This paper introduces a versatile approach to achieve specific surface functionality with-out the need for detailed knowledge of surface topography. This is accomplished for applications targeting wettability properties by integrating contact angle measurement into a micromachining setup with a nanosecond pulsed UV laser, allowing for fully automated programs to find optimal functionalization without requiring knowledge on the topography or on possible laser-induced chemical changes itself. This study investigates the impact of various processing parameters, including laser pulse energy, scanning speed, hatching distance, jump speed, and laser repetition rate, on the wetting properties of two widely used polymers: polyethylene (PE) and ethylene propylene diene monomer (EPDM). A design of experiment (DOE) approach is used for experimental design and subsequent modeling. Finally, the effectiveness of this new approach is evaluated and compared with conventional methods.

1. Introduction

A targeted modification of workpiece surface properties is a crucial and widely utilized method for improved functionality and tailored material performance. Surface functionalization has thus emerged as an enabling approach across a broad spectrum of applications, ranging from enhanced biocompatibility [1] and interfacial adhesion [2], reduced friction [3], to the realization of self-cleaning [4,5] and anti-fouling surfaces [6]. Traditional surface functionalization methods include grinding, brushing, coating and deposition processes, plasma treatment, etching, photo-transplantation, and embedding techniques [7,8,9,10]. The growing interest in laser-based methods can be attributed to the versatility and numerous advantages of this technology. These include geometric freedom, minimal or no post-processing, avoidance of toxic or environmentally harmful chemicals, high repeatability, short process times, and the ability to be material-independent, as long as enough absorption is available [11,12]. However, the process of surface functionalization using lasers has also its disadvantages. It is inherently complex, as the laser process itself is subject to a multitude of parameters such as repetition rate, pulse energy, and wavelength, as well as system parameters such as scanning speed, apertures, and scanning delay times, which also influence each other, in both linear and nonlinear ways. Additionally, during the laser–material interaction, additional material-specific parameters, such as absorption behavior, number of free electrons, molecular weights, and specific heat capacity, play a significant role and exert a substantial influence on processing result [13]. Given that the material is exposed to temperatures of a few thousand degrees Celsius during laser processing, it is inevitable that chemical alterations will occur in addition to the ablation and topographical changes [14,15]. All this means that in numerous instances, neither the required surface topography nor the necessary process parameters for a particular function are known. In this case, an empirical approach may be selected, which typically entails a significant investment of time and effort with no assurance of success. A more systematic approach is to use a design of experiments (DOE), a statistical methodology for planning, conducting, and analyzing experiments. Through extrapolation beyond the tested conditions, it enables the derivation of statistical models from experimental data, thereby providing predictive insight into the entire design space [16,17,18,19]. Therefore, it is not necessary to test all parameter combinations. Instead, the number of experiments can be substantially reduced while still enabling the construction of a statistical model that predicts the outcomes of untested conditions with a defined level of confidence. The greatest advantage is that such a model enables application optimization solely on the basis of the laser parameters, without requiring prior knowledge of material or chemical properties.

In many cases, the devices for analyzing functional properties of surfaces are detached from the laser processing system, which precludes the possibility of automated data acquisition. This requires many human inputs, which is the most common source of potential errors [20]. Particularly in manual contact angle measurements, inaccuracies frequently occur during droplet deposition, sample positioning, or camera focusing. Furthermore, the water contact angle may change over time after surface treatment. To ensure that the measurement results are comparable, it may therefore be important to carry out all measurements at equal intervals after treatment. This publication presents a fully automated system that can assess the wetting characteristics of a laser-treated surface through contact angle analysis. This has several advantages: On the one hand, a substantial amount of data can be generated in a reliable and reproducible manner within a relatively short time frame without the need for human intervention. On the other hand, the surface energy can be determined through the dispensing of water and diiodomethane, which serves as the basis for numerous functionalization applications [21,22].

This study examines the modification of two selected polymers (Polyethylene (PE) and ethylene-propylene-diene monomer (EPDM)), which are commonly utilized in industrial applications. EPDM is widely used in the automotive industry as a hose or sealing material, and in the printing industry as a material for printing plates, where surface functionalization can either prevent deposits or enable specific printing performance [23]. PE, being the most extensively produced polymer worldwide [24], is utilized as a base material for a number of applications, for instance, the development of antibacterial surfaces [25], biomedical applications aimed at reducing friction [26], and the enhancement of shear strength in adhesive joints [27].

The focus of this study is on examining the automation approach, including the use of the DOE method. Both the materials used and the process technology employed are to be understood as examples. No claim is made to have developed an overall optimized machining process for the selected materials. Nor is the aim to expand the understanding of the laser functionalization process itself or to examine the applicability of the functionalized surfaces in more detail.

2. Materials and Methods

2.1. Materials

Polyethylene and ethylene-propylene-diene monomer were selected as the materials for investigation in this study. The PE is a sintered, low-molecular-weight material produced by the company Wintersteiger (Ried im Innkreis, Austria), while the EPDM is manufactured by ContiTech Elastomer-Beschichtungen GmbH (Hannover, Germany). Ultrapure water, HPLC grade from Alfa Aesar (Ward Hill, MA, USA) was used for water contact angle measurements.

2.2. Experimental Setup

Figure 1a shows a schematic representation of the fully automated experimental setup. Laser processing is performed using a Coherent Avia LX 355-20-40 SB 1.5 HE (Coherent, Saxonburg, PA, USA) (355 nm, 20 W) diode-pumped, solid-state, Q-switched nano-second laser (τ_p < 25 ns) with an adjustable pulse repetition rate settable from single pulse to 100 kHz. A SCANLAB SK1010 (Scanlab, Puchheim, Germany) is used as the galvanometer scanner combined with a telecentric F-Theta objective from Sill Optics (Wendelstein, Germany) (f = 110 mm) as the focusing optic. To ensure almost seamless power control, a λ/2 waveplate was installed in a piezoelectric rotary axis followed by a polarization splitter in beam direction. For the calculation of the water contact angle using sessile drop method, the image is captured with a UK1158-M SXGA+/1.4 megapixel/monochrome CCD camera from ABS Optronics GmbH (Unterföhring, Germany) along with the aid of Quartz Tungsten-Halogen Lamp from Thorlabs (Newton, NJ, USA) as illuminator. The operating principle of droplet generation is shown in Figure 1b. Two compressed air levels are applied. First, a low material pressure of 0.6 bar ensures that a constant amount of material is pressed through the diaphragm valve (DV-5625SS, Vieweg GmbH Dosier- und Mischtechnik, Kranzberg, Germany—hereinafter, “Vieweg GmbH”), regardless of the cartridge fill level. Second, an additional pressure (2.0 bar) is applied that controls the opening of the diaphragm valve. Since the valve should always remain open for the same duration, this opening pressure is initially regulated by a dosing device (DC1160, Vieweg GmbH). By means of an analog signal from the control PC, the valve opening time can thus be controlled reproducibly and in an automated manner. Each droplet had a volume of approximately 14 µL, as smaller droplet volumes of 2–6 µL, recommended in DIN EN ISO 19403, tended to remain attached to the needle upon contact with the surface (cf. Section 3).

A xyz-axis system (M-521.PD1 from Physik Instrumente GmbH & Co. KG, Karlsruhe, Germany) which enables the sample material to be moved between the scanner and the dosing unit, allows for the direct measurement of contact angles following laser processing. A central computer controlled all subcomponents and analyzed the recorded drop image to determine the resulting contact angle. To determine the contact angle, the captured image is first imported, and brightness values are assigned to the pixels using the pyDSA_core (V1.4.1) library [28]. Subsequently, the contour and baseline are detected using the canny edge detection algorithm. An elliptical fit is then superimposed over the detected contour. Finally, the angles between the baseline and the tangents to the drop on the left and right sides are calculated, and their mean value is reported as the contact angle. Despite the automated evaluation, all automatically analyzed images were reviewed for accuracy and documented for further analysis. The entire procedure is illustrated in Figure 2.

2.3. Experimental Method

Prior to the establishment of the test plan through the DOE, it is essential to define the parameters to be investigated and the boundary conditions. The initial boundary condition (R1) is derived from the objective of achieving scalability of the functionalization process to large areas. Consequently, the number of passes per parameter was set to n = 1. Furthermore, direction-dependent effects must be avoided. This leads to the second boundary condition (R2). In the present study, for each parameter combination a grid consisting of horizontal and vertical lines was ablated. The following parameters were varied as part of the study: scan speed, pulse energy, repetition rate, hatch distance, and jump speed. The latter refers to the velocity of the laser beam during non-exposure, when the galvanometric scanner repositions itself between consecutive marking vectors. The values of the individual parameters are listed in

Table 1. Parameters used to create the DOE test plan, which were varied accordingly.

Parameter	Symbol	Unit	Minimum Value	Maximum Value	Intermediate Steps
Scan Speed	vsca	mm/s	50	1000	100; 250; 500
Pulse Energy	Ep	µJ	20	100	50
Repetition Rate	fRep	Hz	20,000	40,000	-
Hatch Distance	dh	µm	20	200	50; 100
Jump Speed	vjump	mm/s	2000	50,000	10,000; 20,000

. The selection of parameters was based on the specifications of the laser, the material limits, and a reasonable process time, which were estimated from preliminary tests. Furthermore, the objective was to test a wide range of parameters.

The objective of this study is to examine the impact of the laser modified surface on the water contact angle. Many well-functioning types of surface structures have already been identified for this application [29], allowing for a direct comparison of the structures and potentially validating this approach. The water contact angle was chosen as the response parameter.

The worksheets were generated using the DOE software MODDE 13 (Sartorius Stedim Data Analytics AB, Umeå, Sweden). First, an initial parameter screening is conducted to identify the fundamental relationships between the experimental parameters and the response.

Based on these findings, a subsequent worksheet, referred to as system characterization, is generated. This step serves to refine the experimental design and to provide a more comprehensive understanding of the system’s behavior. Given the proof-of-concept character of this work, two distinct design strategies were deliberately employed to evaluate whether both approaches yield consistent and reliable insights. In this study, a screening with the proven D-optimal design was selected for the PE material, while a reduced combinatorial design was employed for the EPDM material, each with an interaction model. A reduced combinatorial design simplifies the full factorial set by omitting runs for practical reasons, whereas a D-optimal design selects runs statistically to maximize information and minimize estimation errors. An interaction model refers to a statistical model that, in addition to main effects, explicitly includes interaction terms to capture how the effect of one factor depends on the level of another. The methodological foundations of DOE cannot be discussed in detail within the scope of this work. Comprehensive explanations are provided in the relevant technical literature [30].

The experimental worksheet for PE contained 22 runs for parameter screening and 15 runs for system characterization. Due to the alternative model for EPDM, the test plan was designed to include 35 tests for screening and 20 tests for system characterization. A comprehensive overview of all parameter combinations and the resulting contact angles can be found in the Supplementary Materials. To verify repeatability, the last three runs of the parameter screening were conducted under identical parameter settings.

With the common COST (Consider One Separate factor at a Time) method, 480 experimental runs are required in order to draw conclusions. As explained in the Introduction, a major advantage of the DOE method is that the experimental worksheet is built with fewer experimental runs based on a chosen model that examines the influence of each parameter with a high likelihood.

In this study, an array of separate grids, one for each parameter combination, gapped by 2 mm to the next one, covering an area of 7 mm × 7 mm, and consisting of crossed horizontal and vertical lines, was generated by laser ablation with a single laser pass per line. When screening PE, two of the test runs in this study had to be excluded because the combination of high pulse energy paired with low scanning speed and low hatching distances resulted in significant material damage, whereupon contact angle measurement was no longer possible. Consequently, a total of 20 tests were therefore carried out on the material samples for the screening evaluation. The system characterization then proceeded to examine the most relevant factors in greater detail.

3. Results and Discussion

Figure 3a illustrates the effect plot from the screening experiments with the PE material, which serves as a critical analysis tool for determining the significance of each of the parameters and their interaction effects. The effects of the parameters can be observed in order of increasing influence. The hatch distance was identified as the most significant factor, with a change of 11.86° observed when it varied from its low level to its high level, with other factors held at their respective averages. The repetition rate and scan speed exhibited individual effects, with changes of 5.41° and 4.88°, respectively. Based on the measured contact angles, the influence of pulse energy and jump speed was found to be less significant. However, pulse energy demonstrated to be a more significant interaction effect in combination with hatch distance. This interaction effect resulted in a change of 6.94° in the contact angle response.

A summary of fit plot provides the statistical overview of all findings obtained from the model fitted using multiple linear regression (MLR). The interaction model presented here can be seen in Figure 3b. The four columns represent R2, Q2, model validity, and reproducibility. If all four values were to equal 1, it would indicate that the model is in complete conformity with the trails. R2 indicates the model fit, where Q2 shows the prediction precision. Model validity tests for various model problems such as presence of outliers, inaccuracies, or a transformation issue. Reproducibility is the replicate variability to overall variability indicator. RSD shows the variation in the response not explained by the model. DF represents the Degrees of Freedom. A highly robust model has been designed here, as evidenced by the high R2, Q2, model validity, and reproducibility.

As the model was assessed to be good, this study was further proceeded with the system characterization to understand the influence of the most influential factors. From the analysis of the effects of individual parameters and their interaction effect, it was concluded that scan speed, hatch distance and repetition rate can be considered as the significant parameters among the five parameters investigated. However from the understandings of the interaction effects along with their individual significance in the contact angle, pulse energy tends to give the maximum contact angle at its highest setting and hence to further proceed with the system characterization, it was set to this constant value of 100 μJ. All other parameters are varied in accordance with their previously defined values (cf. Table 1) with the objective of further enhancing the model and achieving higher contact angles. From this interaction model, the software MODDE creates a new test plan with 15 experimental runs for system characterization.

The effect plot for the EPDM material is presented analogously in Figure 4a, with the corresponding summary of fit plot shown in Figure 4b. Notably, for both materials, the individual factors hatch distance (8.06°), repetition rate (−5.66°), and scan speed (5.20°) exert the strongest influence on the resulting contact angle, although the effect is somewhat less pronounced for the PE material. Pulse energy and jump speed appear to have only a minor or negligible impact when considered as single parameters. Nevertheless, these parameters exhibit comparatively strong effects when acting in combination (cf. pulse energy with scan speed and hatch distance). Furthermore, it is notable that, despite the higher number of experiments, the repeatability values remain high, while the model validity has decreased to below 0.6. However, a substantial lack of fit is only considered present when the model validity falls below 0.25, so the current model can still be regarded as providing a good fit. The effects of jump speed on the wettability were decided to be insignificant and hence in this study to further proceed with system characterization, this factor was set to a constant of 2000 mm/s. As with the previous example, all other parameters are varied according to their previously defined values (see Table 1).

After the system characterization experiments have been performed, further information on the interactions of the significant parameters can be obtained. The analysis of the effect and interaction model plot in Figure 5 demonstrates that the applied DOE essentially represents a statistical evaluation. Although the interaction model shows good agreement with the experimental results, it becomes evident that the repetition rate, neither as an individual factor nor in interaction with hatch distance or scan speed, exerts a significant influence. Within the framework of the present system characterization, this parameter can therefore be considered negligible, despite indications to the contrary from the initial parameter screening. By contrast, the effect plot clearly indicates that hatch distance and scan speed are the dominant factors. To validate the model and to determine an optimal set point, the repetition rate was kept constant at 40 kHz, as higher repetition rates tend to be advantageous, particularly with respect to process scaling.

Response contour plots offer effective visualization when two factors dominate the system behavior, as shown in Figure 6a. The response contour plot depicts a design space with scan speed and hatch distance as the outer axes, which plots the response contours for different values of these factors. With constant values of pulse energy, jump speed, repetition rate and number of passes in vertical and horizontal scan direction, a minimum of 80.54° and a maximum of 144.56° are predicted for these parameters from the design analysis. The high response values in the range between 140° and 144.56° are obtained within the scan speed range of 333 mm/s–953 mm/s and hatch distance in the range of 100 µm–200 μm. The optimal condition to achieve maximum response was as shown in the figure with a double circle and is supposed to give a response of 144.56°. The experimental testing of this optimal set point resulted in a measurement contact angle of 143.74°, which reciprocated the validity and prediction precision of the studied model. The drop volume of 2–6 µL, as recommended in DIN EN ISO 19403, resulted in the drop remaining attached to the needle upon contact with the surface. This suggests that the adhesive forces between the needle and the drop are stronger than those between the water and the laser-modified surface. The drop volume had to be increased to 14 µL, until the drop detached from the needle. Therefore, the drop volume was set to 14 µL for all experiments, as noted in Section 2.2.

Figure 6b shows the surface, which was created with the parameters of the optimal setpoint. It is notable that the topography exhibited only minimal change during laser processing. Nevertheless, the laser processing is clearly visible, as the color of the material changes from black to white. This observation suggests that the hydrophobic behavior may have been achieved through laser-induced chemical changes. However, the hydrophobization arises most likely from both topographical and chemical changes, though this remains speculative. Yet, the precise origin of hydrophobization is less of an issue for this study, as the aim is to directly relate laser parameters to the resulting contact angle. On the one hand, this allows the use of high scanning speeds (>600 mm/s) and correspondingly high area rates, which is particularly important for the functionalization of large areas. On the other hand, numerous challenges are associated with this approach, as chemical functionalization, in contrast to functionalization on a topographical basis, can undergo changes over time, so that the functional effect can weaken or even disappear completely. Moreover, the fundamental principles underlying chemical functionalization are considerably more challenging to ascertain. Nevertheless, hydrophobic structures have also been identified on PE, which are based on the topographical change [29], similar to the structures on the EPDM material, which will be discussed in greater detail below. Ultimately, the determination of the approach will depend on the specific application.

The results of the system characterization of EPDM revealed a behavior similar to that observed for PE. Here as well, the repetition rate as an isolated parameter appears to exert only a negligible effect. Nevertheless, in interaction with the hatch distance, the repetition rate exhibits a pronounced influence. As illustrated in the interaction plot between hatch distance and repetition rate (c.f. Figure 7), it becomes evident that increasing the repetition rate combined with decreasing hatch distance leads to an increase in the measured contact angle. While the contact angles obtained at a repetition rate of 20 kHz cluster closely around approx. 132°, substantial differences emerge at 40 kHz. At this higher repetition rate, the contact angle decreases to 126.1° at a hatch distance of 200 µm, whereas it increases considerably to 139.2° and 145.8° at hatch distances of 100 µm and 50 µm, respectively. Based on this relationship, the hatch distance was kept constant at 50 µm and the repetition rate at 40 kHz for the subsequent investigations aimed at determining the optimal set point.

In contrast to PE, the applied pulse energy represents an additional significant factor. Since the repetition rate and hatch distance had already been kept constant, the contour plot was generated to illustrate the contact angle as a function of scan speed and pulse energy, as shown in Figure 8a. A design space for achieving high contact angles is observed at lower scan speeds combined with low to moderate pulse energies. Several clear trends can be identified. For instance, increasing the scan speed leads to a decrease in the expected contact angle. It should be noted, however, that the scan speed cannot be reduced arbitrarily. Excessively low scan speeds not only result in prohibitively long processing times but may also cause damage to the material, as shown in the parameter screening. Nevertheless, the parameters of the optimal set point—50 µJ at 50 mm/s and a hatch distance of 50 µm —appeared to remain suitable for the material. Furthermore, it is noteworthy that the optimal pulse energy, according to the model, is consistently around 50 µJ, independent of scan speed. Deviations from this value, either toward higher or lower pulse energies, result also in a reduction in the contact angle.

The laser scanning microscope image of the topography created using the optimal set point was characterized to be a structure that appeared similar to micro-columns or spike patterns, as shown in Figure 8b. By studying the topography under varying process parameters, it was found that the depth and height of the micro-columns are governed by both laser power and scan speed, while their density is primarily affected by laser power and their lateral arrangement is dictated by the hatch distance. These observations are well in line with the underlying physical mechanisms. The pulse energy predominantly determines the ablation depth as well as the ablated volume per pulse. At higher pulse energies, deeper and wider grooves tend to be generated. The scan speed likewise has a direct impact on the line energy input. Lower scan speeds generally result in deeper ablation trenches, whereas higher scan speeds lead to shallower trenches. Finally, since the hatch distance dictates the path planning of the scanning system, it is the governing parameter for the spatial arrangement of the micro-columns. Since many laser and scanner parameters influence each other, this emphasizes the importance of concurrently considering multiple parameters and their interactions when optimizing surface topography. The surface functionalization method described here, based on a change in topography, has already been investigated on numerous occasions, resulting in a comprehensive understanding of the underlying causes. In the Cassie–Baxter model, the droplet is not in complete contact with the surface, but rather only at the tips of the microstructures of the surface and the entrapped air [31]. The presence of air entrapped between the solid structures creates a barrier that inhibits the wetting of the surface by the water droplet over a large area. The confinement of the droplet by air from multiple directions results in uniform interfacial interactions, culminating in a high contact angle and the formation of a nearly spherical droplet.

Although the presented approach relies solely on varying the laser parameters, a more detailed analysis of the topography of the two highly hydrophobic surfaces obtained under the model-predicted optimal conditions is conducted. Therefore, in Figure 9 SEM images of the two surfaces are shown. The tip of the spike-like structure in EPDM is displayed on the left side of Figure 9, while the laser-irradiated surface of the PE material (cf. the white area in Figure 6b) can be seen on the right side. The middle images show enlarged versions of the corresponding sections. It is evident that the microstructures formed on the surface of the EPDM material are significantly deeper than those generated on PE. Nevertheless, both surfaces exhibited pronounced hydrophobic behavior after processing. Whether this effect can be attributed to a chemical functionalization or to a change in topography cannot be conclusively determined within the scope of this study.

4. Conclusions and Outlook

This study has provided valuable insights into the potential for developing a model for the functionalization of surfaces with varying wettability through the use of DOE and laser parameter variation, without the necessity of prior knowledge regarding the surface topography and possible laser-induced chemical material changes. The objective was achieved through the complete automation of the process, from the laser process to contact angle measurement and evaluation. By employing DOE, a systematic investigation of the parameters was enabled, while simultaneously reducing the number of experimental runs by more than 90% compared to the COST method. Nevertheless, the interaction model generated by the experiments demonstrated a high level of agreement with the experimental results.

The novel approach was validated positively in order to comprehend the complex interactions between wettability and the effects of important laser processing parameters, including pulse energy, scan speed, hatch, jump speed, and repetition rate on topographic characteristics. The design of experiments strategy was demonstrated to be a dependable method that permitted the investigation of the parameter space to be organized and effective. This statistical technique enabled the quantification of the individual and combined effects of the parameters on surface changes and the determination of the optimal settings for laser processing. The systematic evaluation of the experimental results has contributed to the creation of a predictive model that improves the ability to customize surface properties for specific wetting functions. The model optimizes for the maximum contact angle, without consideration of the cause of the functional effect. It should be noted, however, that hydrophobicity is not associated with a single surface architecture, as various structures or chemical compositions can exhibit hydrophobic behavior. This became evident here as well, since the structures produced at the optimal setpoint according to the DoE model differed significantly between PE and EPDM. In the case of EPDM, the model revealed a well-known hydrophobic structure that can be attributed to the physical principle of Cassie–Baxter wetting, whereas the structures generated in PE exhibited no significant topographical changes. Instead, the material exhibited a pronounced color change from black to white.

It should also be mentioned that only a limited number of cases were examined in this study. However, this field offers considerable scope for further investigation. For example, additional studies could address other surface properties such as oleophobicity and omniphobicity, facilitated by the flexibility of the self-built contact angle measurement setup, which allows the liquid of interest to be selected. Another important aspect that could not be addressed within the scope of this work is long-term stability. Since surface properties may change over time, future research should focus on whether this system can also provide insights into the durability of the induced structures. Moreover, future studies may extend the functionalization approach to other materials and further elucidate the physical origins of hydrophobization as well as phenomena such as the discoloration observed in PE.

This approach has the potential to have a significant impact on a variety of industries and challenges, such as biomedicine, materials research, or large-area functionalization using laser direct structuring due to limiting the number of passes to one. Improving wettability and surface topography has a direct impact on improving adhesion, biocompatibility, and other important material properties.