2023 Volume 64 Issue 9 Pages 2206-2213
The parameters affecting the hydrophilicity and its aging effect on the polymer surface treated with atmospheric pressure plasma were investigated. A series of experimental procedures were performed according to various parameter combinations using the DoE method. The main factors having the most impact were found by quantifying the effect of each process parameter in plasma treatment. In addition, based on the results, multiple regression analysis was conducted to optimize the parameters of the experiment under conditions that maximize the hydrophilicity of the polymer surface. Finally, process capability analysis showed the superiority of optimized conditions statistically, physically, and chemically. A series of experiments and analytical procedures were characterized with the Minitab 20 program, X-ray photoelectron spectroscopy (XPS), and atomic force microscope (AFM).
Airborne molecular contaminations (AMCs) refer to contaminants in the form of gases generated in or flowing into a cleanroom. As the semiconductor process develops, the process becomes increasingly miniaturized and sophisticated.1) Defects caused by AMCs in the semiconductor process include corrosion defects,2) CU plate/anneal blisters,3) and T-topping.4) As a result, the influence of AMCs on the semiconductor process is getting bigger, and the ability to control it has become very important. Gaseous contaminants are controlled in the cleanroom, and among them, the role of the filter is the most important. However, these filters are either very expensive or perform poorly.5) A water showering system (WSS) is attached to the front of the filter to assist the filter and increase the lifespan of the filter. WSS uses a gas-liquid contact material to absorb water-soluble gases, so the hydrophilicity of the product is essential.
Hydrophilic polymers are a promising technology in various application fields, such as filters and oil-water separation membranes. However, it is difficult to directly utilize them due to problems such as a tiny number and low mechanical properties. Accordingly, various attempts have been made to modify the surface of non-hydrophilic materials to be hydrophilic, such as dip-coating/spray,6) chemical treatment,7) and surface graft treatment.8) However, these methods have disadvantages such as unsatisfactory hydrophilicity of the object to be treated, additional contamination, and expensive. The plasma treatment method, in which activated ions collide with the surface to be modified by supplying energy to gas,9) can easily modify the surface without causing significant damage to the polymer surface.10)
Atmospheric pressure plasma attracts attention because it can modify the surface with only plasma source and substrate without a particular device, such as a vacuum system or solvent treatment.11,12) However, the hydrophilic properties obtained on polymer surfaces are generally not permanent. Surface modification using plasma is applied only to the polymer’s outermost surface layer, and the surface’s composition can be changed when surface tension acts due to the segmental mobility of the polymer.13,14) It is well-known that when a polymer surface is subjected to plasma treatment using a specific gas, the surface is modified, and its properties disappear or decrease over time. However, the effects of each process parameter and optimization condition have yet to be discussed in detail.
Design of experiments (DoE) is a statistical technique for rapidly optimizing the performance of experimental results using input process parameters15,16) by analyzing the results of experiments to determine which factors have a statistically significant effect on the results. It is possible to know which and how factors are interdependent, and finally, the experimental conditions can be optimized for the experiment. A process capability index is a tool for determining whether a set of process results meets a certain quality. Experimental results with optimized process parameters can statistically verify process consistency and yield through process capability analysis.
In this study, experimental factors that affect the hydrophilic properties of polymer surfaces treated with the atmospheric pressure plasma method under various conditions designed by the design of experiments are statistically analyzed. Based on the statistical analysis results, parameters affecting the hydrophilicity of the plasma-treated gas-liquid contact material were identified through physicochemical analysis, and optimized conditions for improving the hydrophilicity were derived.
The experiment used plasma-generating equipment (Enerconstech, ECT-2K-KO0408, Korea). The flow rate of the supplied gas was controlled using a mass flow controller (Line-tech, M3030A, Korea), and plasma treatment was performed under various conditions by placing an LTPE non-woven fabric (Kored, Korea) on a ceramic board (Provis, Korea).
2.2 Experimental design 2.2.1 Experimental setupParameters known to affect plasma treatment are as follows: plasma power,17) gas type/composition,18) treatment time,19) and excitation frequency.20) In the case of plasma power, it affects the concentration of reactive species. Depending on the type of gas supplied, the characteristics of the plasma-treated object are different. In addition, if several types of gas are mixed, the ratio of gases also affects the result. The plasma treatment effect increases as the plasma treatment time increases, but when the treatment time is excessively long, the performance deteriorates as the temperature increases. The excitation frequency of the equipment controls the pore size of the processing object.
2.2.2 Input parameterIn this paper, since the purpose is to make the surface of the polymer hydrophilic, oxygen and nitrogen gas were used according to the characteristics of the object to be treated during plasma treatment. In addition, experiments were conducted under various conditions using it as a factor to confirm each parameter’s effect. In Table 1, the experimental range was set considering the possible conditions of the equipment. It is essential to set this range before experimenting because the design of the experiment proceeds based on this table.
In the case of a 2-level factorial design, the experimental design’s statistical power changed according to the center point of each block and the number of iterations. Therefore, as shown in Fig. 1, a power curve was prepared to confirm the power according to each condition. In this study, the experiment was conducted based on the design of 38 experiments with three central points and two repetitions with the best power in Table 2. In the plasma experiment, the LTPE non-woven fabric was treated by setting the voltage, the flow rate of O2 and N2, the plasma treatment time, and the excitation frequency based on each condition. After that, the non-woven fabric was cut into 1 * 10 (cm) and measured three or more times according to the prescribed water absorption evaluation method (KS K 0642: time taken to rise 6 cm from a point immersed 3 cm in water), and the average was recorded. Each condition and its results are presented in Supplementary 1. The order of experiments was randomized to prevent the order of experiments from affecting the results.
Power curve for 2-level factorial design.
Minitab 20 software was used as a tool to analyze the basic experimental design and its effects. Various physicochemical analyzes were conducted to analyze the effects. Differences in surface damage under specific conditions were observed through an Atomic Force Microscope (AFM, Park Systems, XE-150, Korea). In addition, WCA was measured with a contact angle analyzer (DSA60, Krüss, Germany) to compare hydrophilicity under various conditions. Surface analysis was performed with an X-ray Photoelectron Spectrometer (XPS, Thermo Fisher Scientific Brno s.r.o, Nexsa, Czech Republic) to confirm the change of functional groups on the surface of the substrate.
Factorial design analysis was conducted based on the experimental results. The normal plot in Fig. 2(A) shows the standardized effects relative to the distribution fitted line for the case where all effects are zero. Effects that are far from the normal probability plot are statistically significant. Figure 2(B) is a Pareto chart with a baseline indicating which effect is statistically significant. Table 3 is an analysis of the variance table for each factor. The smaller the p-value, the more statistically significant the factor. If the P-value was less than 0.05, it was judged to have a significant effect. Based on this, the results were statistically significant in the order of treatment time, voltage, O2 * treatment time, N2 * frequency, O2, voltage * O2, and voltage * treatment time.
(A) Normal plot and (B) pareto chart of the standardized effects in the initial condition.
Figure 3(A) shows the tendency according to the level of each parameter. The slope of the graph increases as the factor has a more significant influence, and the tendency of a specific factor can be known according to the direction of the slope. Figure 3(B) is an interaction plot on whether the two factors affect each other. The closer the two graphs are orthogonal, the larger the result of the interaction of the two factors and the smaller the effect when they are parallel. It can be seen that the two graphs are not parallel in the case of essential data.
(A) Main effects plot and (B) interaction plot for initial condition experiment results.
The voltage, O2, and treatment time are confirmed to change non-woven fabrics’ physical properties significantly. In order to compare the hydrophilicity of the non-woven fabric surface according to the intensity of voltage, the water contact angle was measured, as shown in Fig. 4. After plasma treatment for 10 seconds from 0.02 kW to 0.05 kW in increments of 0.01 kW, the hydrophilicity of the surface was confirmed. As a result, all of them were hydrophilic, but the time required for complete absorption differed for each condition. Videos of the absorption of water on the surface of the non-woven fabric can be seen in supplement 2. In order to find out the difference between the reacting gas, the surface of the plasma-treated sample was observed for each reactant gas, as shown in Fig. 5, through AFM analysis. Figure 5(a) shows the root mean height (Sq) of the surface height measured by AFM, an index to evaluate the non-uniformity of the surface height. When plasma treatment is performed with O2 gas, it collides with the surface of the non-woven fabric with strong energy. The Sq value is significant, creating a strong hydrophilic functional group on the surface.
Water contact angle images of various voltage parameters.
AFM 3D images of (a) non-treatment, (b) O2 plasma, (c) N2 plasma treatment non-woven.
On the contrary, the N2 can be confirmed that the surface unevenness is reduced to the exposed sample without causing a strong reaction. It was found that the plasma treatment time also affected the result value. As shown in Fig. 6, it can be seen that the hydrophilic performance improves as time increases for the first 60 seconds. However, the performance rapidly decreases after a certain period (120 seconds). These results explain that when plasma treatment is performed over time, the equipment temperature is excessively high, and the reaction does not occur properly.
Absorption time with various treatment time parameters.
In the case of an interaction in which two or more factors interact with each other and affect the result, there are a total of four cases: O2 * treatment time, N2 * frequency, voltage * O2, and voltage * treatment time. Lei et al.22) reported that the longer the treatment time with O2, the more the polymer surface is oxidized, and the surface properties may change. According to Kyu,21) the growth rate decreases under high-power plasma conditions because the object to be processed is etched by excessively energetic particles, or the excessively energetic particles cause a vapor phase reaction. In addition, since the current density does not continue to be maintained/increased as the voltage increases in Yigang, et al.,23) but rapidly decreases when it exceeds a certain level, care must be taken not to proceed at an excessively high voltage. Unlike voltage, O2, and treatment time, which are known to have a significant effect on each major factor, N2, and frequency did not significantly affect the result, as a single factor appeared to have an effect. Referring to Eugene G,24) N2 ionized with high energy has a high deposition or etch rate at lower frequencies because it forms a long tail.
3.1.4 OptimizationSo far, each major factor in the atmospheric pressure plasma reaction and its interaction has investigated how and why it affects the hydrophilicity of the treated substrate. Therefore, based on this, we optimize the parameters under which conditions show the most excellent hydrophilicity. In the case of the plasma treatment reaction, each factor tends in a specific direction, but rather than one side unconditionally prevailing, it is essential to find an appropriate middle point to find the optimal condition in a parabolic form. For this purpose, multiple regression analysis was performed. As shown in Fig. 7, all parabolic shapes were shown through multiple regression analysis, except for N2 and frequency, which are insignificant factors. Each minimum point among them is shown in Table 4 as the optimal condition. In Table 4, it was possible to confirm the optimal conditions in which the plasma-treated substrate, confirmed through Fig. 7 above, showed the most outstanding absorption. The expected optimal conditions are voltage 0.227, O2 26.162, N2 600, frequency 5.0, treatment time 77.374, and absorption time 26.8780 seconds. However, according to the 95% prediction interval, the range was too extensive, so we tried to confirm this through an additional process to verify the results. One hundred experiments were conducted based on the optimization conditions, shown in Fig. 8. As a result, the Cpk value was 1.55, indicating a significantly high yield corresponding to 4.5 sigma.
Multiple regression analysis for absorption results in the initial condition.
Process capability report of absorption time of optimized condition.
Evolution of adsorption time for plasma-treated non-woven as a function of the storage in air.
Plasma treatment has the advantages of being simple and inexpensive but has the disadvantage of aging effects, as shown in Fig. 9 over time. Therefore, to confirm the effect of each factor, samples experimented under the same conditions as supplement 1 were stored in the air for one month. Then the degree of effect over time was analyzed, and the results are shown in supplement 3.
3.2.1 Factorial design analysisWhen interpreting Fig. 10 and Table 5 regarding the interpretation method in 3.1, only N2 was identified as a significant factor, unlike Fig. 2, immediately after plasma treatment.
(A) Normal plot and (B) pareto chart of the standardized effects in the aging effect experiments.
In Fig. 11(A), in the case of factors affecting the aging effect, only N2 was significant and showed the largest slope. Compared to Fig. 3(A), the O2, frequency, and treatment time trends are similar, but in the case of voltage and N2, it can be confirmed that they are reversed. In the case of the interaction plot, it can be partially confirmed that the interaction is estimated because the two graphs are overlapped rather than parallel. However, it is difficult to see that it has a significant effect when referring to Fig. 11.
(A) Main effects plot and (B) interaction plot for aging effect experiment results.
Only N2 is considered to be a significant factor in the aging effect because, referring to Fig. 5, since N2 is not as aggressive as O2,25) it does not have much effect immediately after treatment. However, it forms amino functional groups, which are long-lasting functional groups. Also, as shown in Fig. 12, the functional group is maintained for a long time compared to O2 plasma, where the hydrophilic functional group disappears quickly. It is known that carbonyl and carbonate groups, which are functional groups generated by reacting with oxygen gas, disappear first among functional groups generated by plasma treatment.26)
XPS (a), (b) area and (c), (d) spectra of (a) O 1s, (b) N 1s scan area after 0 day and 30 days, and C 1s after (c) 0 day and (d) 30 days for plasma treatment samples.
As described above, it can be seen that only nitrogen has a significant effect on the hydrophilicity of the plasma-treated non-woven fabric stored for one month. Therefore, to improve the hydrophilicity, the amount of nitrogen should be increased during the initial plasma treatment. However, more information is needed to optimize the parameters, and in order to improve it, follow-up research on innovative ways to reduce the aging effect is needed.
This study investigated the effect of atmospheric pressure plasma treatment and optimization on the hydrophilicity and maintenance of the polymer surface according to parameters. Statistical and physicochemical analysis revealed the effect and extent of each parameter on hydrophilicity. Among the atmospheric pressure plasma process factors, the processing time, voltage, and O2 showed a nonlinear trend, such as parabolic curves, and significantly affected the hydrophilicity of non-woven fabrics.
This research was supported by Basic Science Research Program through the National Research Foundation (NRF) of Korea funded by the Ministry of Science, ICT (MSIT) (NRF-2022R1A2C1011616) and the Technology Innovation Program (20017434, Development of Maximum Diameter 400A UHP grade Pipe and Particle Free High Precision Module) funded By the Ministry of Trade, Industry & Energy (MOTIE, Korea).
