Roles of chemistry modification for surface wettability of Polyether-Ether-Ketone (PEEK) by ultraviolet laser ablation

Tailoring the functionality of Polyether-ether-ketone (PEEK) is critical for enhancing its application, which can be accomplished by the modification of surface morphology and chemistry. In the present work, we experimentally demonstrate the correlation of modified chemical composition of textured PEEK surface by 355 nm UV nanosecond pulsed laser ablation with enhanced surface wettability. Specifically, the impact of UV laser processing parameters on microgroove morphology and ablated surface quality of PEEK surface is evaluated, with which high precision grid surface textures with uniform ablation quality are fabricated. The modification of chemical elements and functional groups of textured PEEK surface by the laser ablation is further analyzed by XPS spectra characterization, which demonstrates the substantial change of C=O and O–C=O bonds, as well as freshly generated polar carboxylic acid groups. Experimental results indicate that the surface composition modification greatly increases surface polarity and surface free energy of textured PEEK surface accompanied by enhanced surface wettability.


Introduction
Polyether-ether-ketone (PEEK) is one of the most popular high-performance engineering polymers currently available, due to its low cost and outstanding properties of high mechanical strength, moderate thermal stability, strong chemical resistance, excellent biological properties, high wear resistance and superior anticorrosive nature [1][2][3]. For instance, PEEK has been increasingly used as a base biomaterial for parts of orthopedic and spinal implants [4][5][6], for which the biocompatibility is one important indicator for evaluating their performance. It has been reported that the biocompatibility of parts is closely associated with the surface wettability characterized by contact angle. Specifically, a smaller contact angle, which indicates better hydrophilicity, is conducive to enhancing the osteogenic differentiation ability of cells and improving the osseointegration ability of PEEK material, i.e., better biocompatibility [7]. Therefore, enhancing the surface wettability of PEEK is crucial for promoting the biocompatibility and performance of PEEK parts for biomedical application.
The surface wettability is not only determined by the intrinsic surface free energy of materials [8], but also tailored by changing surface morphology such as texturing [9]. For instance, laser surface texturing (LST) has been widely used to modulating surface wettability of a variety of materials by introducing surface microstructures, for its advantages of low cost, high machining accuracy and wide range of processed materials [10][11][12][13]. In particular, recently LST has also been used to modify the surface wettability of PEEK. Li et al [14] used a femtosecond laser with a wavelength of 1026 nm to fabricate grooves with a spacing of 0.8 mm on PEEK surface, which change the surface wettability from hydrophilic to hydrophobic. Corderoet et al [15] fabricated parallel microgrooves with different intervals on PEEK using an ArF excimer laser, and reported that the texture geometry strongly influences the biocompatibility of laser ablated areas for MC3T3-E1 pre-osteoblastic cells, in Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. terms of both the contact angle and biological response. Michaljanicova et al [16] found that the incidence angle of the KrF excimer laser has a strong impact on the texture geometry on PEEK, which in turn affects both the contact angle and biological activity of the ablated surface.
Although previous work demonstrated the LST-induced modification of surface wettability of PEEK, most of previous laser ablation mainly utilized 1064 nm wavelength. Since laser ablation process is strongly affected by the absorption coefficient of ablated material at specific laser wavelength, the laser ablation of polymers that absorb the laser energy in a non-linear manner is strongly related with the utilized laser wavelength. In particular, PEEK has a higher absorption coefficient at ultraviolet 355 nm wavelength than that at infrared 1064 nm wavelength [17]. Riveiro et al [18] investigated the impact of laser wavelength on surface morphology of PEEK after laser ablation, and found that laser ablation with the wavelengths of 1064 nm and 532 nm leads to the formation of adhesion debris around microgrooves, which is not beneficial for cell adhesion. In contrast, the utilization of 355 nm wavelength results in high precision of microgrooves accompanied with a slight melting of the material. Ledesma et al [19] used a picosecond laser with two different wavelengths (355 and 1064 nm) to ablate PEEK surface, and reported that the laser ablation with 1064 nm wavelength is dominated by a thermal process, while the laser ablation with ultra violet (UV) 355 nm wavelength is governed by direct photochemical ablation. While the laser ablation process with low absorption coefficient is effective in high accuracy of texturing, it is important to investigate the UV laser ablation on PEEK surface, in particular the relationship between textured surface morphology and laser parameters.
The change of surface wettability by LST is also closely associated with the modification of surface chemistry for metallic materials, such as stainless steel [20], aluminum [21] and titanium alloy [22], etc. Specifically, the interaction of laser energy with metal surface leads to generation of large numbers of hole electron pairs that results into the instability of surface state, which in turn reduces its surface energy by the adsorption of foreign chemicals at the surface defect sites [20]. However, the laser-polymer interaction is significantly differently from the laser-metal interaction. For metallic materials, the laser ablation leads to high-frequency vibrations of free electrons, which trigger weak transmission waves and strong reflection waves, and finally are absorbed in the form of electron bremsstrahlung accompanied by conversion into heat energy. For non-metallic materials, however, most of the irradiation laser is absorbed and only a small proportion is reflected. Recently, the surface chemistry modification of PEEK by laser ablation has been reported. Henriques et al [23] used a two beam interference ultrashort pulse laser with a wavelength of 1064nm to realize the preparation of multi-scale surface structures on PEEK surfaces. Meiling [24] used a picosecond laser to process circular dimples on PEEK surface to obtain a surface with stable friction coefficient and less wear debris. In wettability, Wilson et al [8] used a nanosecond pulse with a wavelength of 1064 nm to process discrete points on the PEEK surface, which increases the proportion of oxygen elements on the surface, thus improving the surface wettability. Lu et al [25] used Nd: YAG laser with a wavelength of 1064 nm to fabricate microgrooves with a spacing of 1 mm on PEEK surface, and found that the content of oxygen after laser ablation is increased, accompanied with lowered hydrophilicity due to the non-polarity of ablated surface. However, the change of surface chemical composition of PEEK by UV laser ablation is largely unknown. In particular, the electronic excitation of polymer molecules triggered by the absorption of ultraviolet photons results into polymer chain breakage, which is significantly different from the vibration and rotation excitation of polymer molecules by the absorption of infrared photons. Thus, revealing the underlying correlation between chemical composition modification and surface wettability of PEEK surface by UV laser ablation is essentially needed.
Thus, we carry out the LST on PEEK surfaces using UV nanosecond laser, with an emphasis on the surface wettability and chemical composition modification. Firstly, optimal UV laser ablation parameters for achieving high precision of ablated microgroove are derived. Secondly, PEEK surface is ablated by UV laser to produce high-precision grid textures, and the optimal line pitch of the texture for the lowest contact angle and surface roughness is discovered. Finally, the modification of chemical compositions of ablated PEEK, as well as its correlation with surface wettability, are discovered.

Preparation of PEEK material
The utilized PEEK 5600G workpiece has a dimension of 22 mm in diameter and 10 mm in thickness, as shown in figure 1(a). Table 1 lists the material properties of PEEK 5600G. Prior to LST, the workpiece is mechanically turned to obtain a homogeneous surface with surface roughness (Ra) less than 1 μm, as shown in figure 1(b). Then the workpiece is subjected to ultrasonic cleaning in ethanol for 15 min at a temperature of 25°C to remove surface contaminants, followed by drying in a closed cabinet. Figure 2 shows the experimental configuration of LST experiments, which consists of a nanosecond pulsed laser with the wavelength of 355 nm, a vertical linear slide, a galvo scanner and a sample holder. Table 2 lists the specifications of the UV nanosecond pulsed laser. During the LST processing, the PEEK surface is exposed to a horizontally polarized Gaussian laser beam for ablation, in which the focusing length is guaranteed by the precisely adjustment of vertical Z-axis. The diameter of focused laser beam is 20 μm for a focusing length of 186.88 mm. Meanwhile, the design of textures to be ablated is realized in the Ezcad2 laser marking software, which is also capable of adjusting laser ablation parameters. Optical microscopy and scanning electron

Characterization of surface wettability and chemical composition
To investigate the surface wettability of textured PEEK, the analysis of contact angle is evaluated following the method of axisymmetric drop shape analysis-profile with two kinds of test liquids of modified-simulated body fluid (M-SBF) and distilled-deionized (DD). In particular, the M-SBF is an apatite supersaturated solution containing calcium ions and phosphate ions that is close to that of human blood plasma, which is widely used in the evaluation of biological activity in vitro. Prior to the measurement of contact angle, the sample surface is cleaned within ultrasonic cleaner with ethanol for 5 min, followed by naturally dried. During contact angle measurement, a test droplet of 5 μl is continuously dropped on the sample surface with a microsyringe and allowed to stand for 16 s. The elliptical fitting method is applied to estimate the static contact angle. Five repeated measurements at random position are carried out to verify the accuracy of the data collection. The chemical compositions of textured and non-textured PEEK surfaces are measured by XPS (EscaLab 250Xi) using a monochromatic Al Kα x-ray source. Survey scans under binding energy ranging from 0 to 1200 ev and 100 ev constant pass energy are performed, after that the C1s XPS measurement under a 20 eV constant pass energy is performed. The C-C peak of 284.58 eV is used as reference to calibrate the electron binding energy of peaks. Data processing is done using XPS PEAK41.

Results and discussion
3.1. Laser parameters rationalization for LST of PEEK While the functionality of surface textures is closely related with the texture morphology, the effect of laser processing parameters on the ablated microgrooves morphology is assessed. While the material removal in nanosecond pulsed laser is dominated by thermal effect-induced evaporation, there is inevitably heat affected zone (HAZ) observed in the ablated surface. Previous work demonstrated that the HAZ in nanosecond laser ablation can be minimized by parametric optimization of laser processing parameters [26][27][28]. Therefore, in the present work three laser processing parameters, as average power, scanning speed and repetition frequency, respectively, are considered to achieve microstructures fabrication with high surface quality and minimized thermal damage.
The effect of average power ranging from 0.5 W to 3.0 W on the laser processing of single microgroove on PEEK surface is firstly studied. For each average power, the scanning speed is 100 mm s −1 and repetition frequency is 100 kHz. Figure 3(a) shows microgrooves on PEEK surface by UV laser ablation with different average powers. As seen from figure 3(a) that with increasing average power, the groove ablation intensifies accompanied with obvious HAZ. Figure 3(b) presents the ablated microgroove morphology measured by white light interferometer at the average power of 1.25 W, indicating the high straightness of microgroove. Furthermore, the accumulation of resolidified particles is observed mainly on one side of the microgroove, which is correlated with the leading side of laser beam. The surface roughness of ablated surface is 2.64 μm, which is higher than the 1 μm for the un-textured one, due to the presence of surface pile up around ablated microgroove.
Feature sizes of laser ablated microgrooves on PEEK surface are accurately derived from the characterization by white light interferometer. Figure 4 plots variations of the average depth and width of microgrooves with average power. It is shown either depth or width of microgrooves is proportional to the average power, starting from 0.5 W to 3.0 W. The resulting violent melting and evaporation of ablated material lead to the increase of either depth or width of microgrooves due to the enhanced laser intensity in the ablation zone.  Figure 5 demonstrates that the microscopic morphology of ablated microgrooves is significantly affected by the average laser power. Specifically, the laser ablation effect of PEEK materials increases with increasing average power, and the depth and width of microgrooves also increase accordingly. However, strong ablation and irregular material removal occur on the ablated surface of PEEK with  the further increase of laser average power, resulting in increased groove surface roughness caused by the accumulation of resolidified particles. Therefore, an optimal average power of 1.25 W is discovered and used in subsequent experiments to ensure the high quality of ablated microgrooves.
By following the strategy for evaluating the influence of average power on the laser ablation process of PEEK surface, the effect of scanning speed in the range of 40-200 mm s −1 is evaluated. The scanning speed dominates the spot overlap ratio, which is critical for the microgroove formation. For each scanning speed, the average power is 1.25 W and the repetition frequency is 100 kHz. Figure 6 plots changes of groove feature sizes as a function of scanning speed. It is shown that the microgrooves depth and width are inversely proportional to the scanning speed. The increasing scanning speed leads to the decrease of spot overlap ratio, accompanied with the decrease of energy accumulation on PEEK surface, which results into the decrease of the microgrooves depth and width . Figure 7 further shows SEM images of microgrooves under increasing scanning speeds in the range of 50-200 mm s −1 . When the scanning speed is small, the laser ablated microgroove with large depth has poor surface quality, due to significant HAZ around microgrooves and strong thermal damage in the microgroove bottom accompanied by the high laser overlap ratio. With increasing scanning speed, the microgroove bottom becomes relatively flat and smooth. As the scanning speed is further increased, the decreased laser overlap ratio leads to the decrease of material removal and formation of irregularity surface of microgroove. Therefore, an optimal scanning speed is chosen as 100 mm s −1 in subsequent experiments.
Finally, the effect of repetition frequency in the range of 40-145 kHz on the UV laser ablation of PEEK surface is studied. For each repetition frequency, the average power is 1.25 W and the scanning speed is 100 mm s −1 . Figure 8 plots variations of feature sizes of laser ablated microgrooves with repetition frequency. It is shown that both the groove depth and width are inversely proportional to the repetition frequency, while the width and the depth fluctuate slightly when the repetition frequency is higher than 100 kHz. It indicates that with increasing repetition frequency, the pulse energy has a strong influence in determining the width and the depth of laser ablated microgrooves. Figures 9(a)-(d) further shows SEM images of individual microgrooves at different repetition frequencies of 40 kHz, 70 kHz, 100 kHz and 130 kHz, respectively. When the repetition frequency is small, the ablated microgroove has a poor quality with significant material melting and resolidification occurring at microgroove edge. With increasing repetition frequency, the laser ablated microgroove surface becomes flat and smooth. When the repetition frequency is further increased, the ablated area is rough and has irregularity surface with debris accumulated around microgrooves. Therefore, an optimal repetition frequency of 100 kHz is discovered and selected in subsequent experiments. Above results indicate that the thermal damage of PEEK under UV nanosecond laser ablation can be well controlled by optimizing laser processing parameters. We also note that the laser ablation-induced thermal damage can be also controlled by different methods, for instance utilizing ultra-short femtosecond [29] and picosecond [23] pulsed lasers, or employing additional fields of inert gasassisted [30,31], water mist-assisted [32], water-conductive laser [33], etc. Those methods could also be applied in future work of UV nanosecond laser ablation of PEEK for achieving low thermal damage surface.

Surface wettability of grid texture on PEEK surface
With the obtained combination of optimal laser processing parameters, (an average power is 1.25 W, a scanning speed is 100 mm s −1 and a repetition frequency is 100 kHz), grid texture with a line pitch of 100 μm is fabricated on PEEK surface by UV laser ablation. In the laser processing process of the grid texture, laser beam firstly ablates the horizontal microgrooves on the PEEK surface followed by the vertical microgrooves to form mesh-type textures. Figure 10(a) shows the SEM image of microgrooves with a line pitch of 100 μm, which demonstrates that each unit mesh divided by horizontal and vertical microgrooves is regularly distributed with uniform size. Figure 10(b) further shows SEM image of enlarged view of unit mesh on grid textured surface, which indicates that the bottom of the laser ablated microgroove is uneven with bulges and rims. Figures 11(a) and (b) shows the evolution of contact angle as well as droplets profile of DD water over time on non-textured and grid textured PEEK surface with a line pitch of 100 μm, respectively. It can be seen that the contact angle of DD water of non-textured surfaces of PEEK is less than 90°, which indicates that the PEEK surface is hydrophilic. However, the contact angle of textured surfaces is larger than 90°, indicating the surface wettability changes from hydrophilic to hydrophobic, which is not desired for the design of textured PEEK  surface with enhanced surface wettability. Above results indicate that the line pitch of 100 μm possessing low ablation accuracy of microgroove and deteriorated surface wettability which is not satisfied, and it is necessary to discover the optimal line pitch for enhanced surface wettability.
The surface wettability represented by contact angle is related to not only internal surface energy but also external surface roughness. There are two types of wetting models for rough surface, as the Wenzel Model [34] for hydrophilic surface and the Cassie Model for hydrophobic surface [35]. In particular for the contact angle less than 90°, the droplets on the hydrophilic surface completely infiltrate the microgrooves to form the Wenzel model, for which the contact angle is calculated as follows: q q = ⋅ r cos cos , w o in which q w represents the contact angle on rough surface under the Wenzel model, q 0 represents the contact angle on smooth surface, and r represents the ratio of the actual area of the rough surface to the geometric area. Thus, the Wenzel model indicates the strong correlation of surface wettability with surface roughness, which is also closely related with line pitch of textures.  Therefore, the influence of line pitch ranging from 0 μm to 700 μm with an interval of 100 μm on the PEEK surface roughness and the contact angle of grid textured surface is studied. Figure 12 plots the variation of average surface roughness with line pitch. It can be seen from figure 12 that with increasing line pitch, the textured surface roughness increases first and reaches the maximum of 12.0 μm at the line pitch of 100 μm, followed by a rapid decrease to a constant value of 4.5 μm. The processing area with smaller line pitch has higher surface roughness, due to the overlap between microgrooves and the pronounced re-solidification of ablated material. With increasing line pitch, the number of microgrooves decreases within a fixed area, which leads to the flatten of ablated area under lowered laser energy accumulation. Figure 12 also plots the variation of contact angle with line pitch for DD water droplets on grid textured PEEK surfaces, which demonstrates the diffusion of DD water droplets is heavily dependent on line pitch. When the line pitch is small, the accumulation of re-solidificated material prevents the diffusion of droplets, so that the textured PEEK surface results into larger contact angle than the non-textured PEEK surface. When the line pitch is large, the droplets spread out under the prevent by only partial resolidified particles, leading to smaller contact angle for grid textured PEEK surface than that for non-textured PEEK surface. Interestingly, the variation of contact angle with line pitch has similar characteristics to that of surface roughness. It can be seen from figure 12 that on grid textured PEEK surface with different line pitches, the contact angle on the grid textured surface with line pitch of 400 μm for DD water droplet is the smallest, which also has a low surface roughness of 4.8 μm.
In the LST process of grid texture with a line pitch of 400 μm, the molten material produced in the laser ablation process has sufficient cooling time to prevent multiple melting and re-forging at a single location, thus producing a high surface quality. Figure 13(a) specifically shows the SEM image of microgrooves with the line pitch of 400 μm, which indicates that the surface of ablated microgrooves is clear and smooth, with no obvious damage and cracks. Figure 13(b) further presents optical image of the grid texture with measurements of  geometries. It can be seen that the average pitch between the two horizontal microgrooves and vertical microgrooves is 464 μm and 455 μm, respectively. The deviation in line pitch with the desired value of 400 μm is 0.500% in the horizontal direction and 0.625% in the vertical direction, indicating the high accuracy of laser processing of textured PEEK surface. Figures 14(a) and (b) show the evolution of contact angle and droplet profile over time on non-textured PEEK and grid textured PEEK with a line pitch of 400 μm for DD water, respectively. It is shown in figure 14 that for each surface, the contact angle decreases with time due to the reduction of surface tension. Compared with the non-textured PEEK surface, the droplet profile on textured PEEK surface spreads more and the profile of droplet is flattened. Therefore, the contact angle on non-textured surface is higher than that on textured PEEK surface, indicating the enhanced wettability of textured PEEK surfaces. The actual solid-liquid contact area on the rough surface with microgrooves is larger than that on the smooth surface, that means, the ratio of geometric area to the actual contact area is smaller than 1. According to the Wenzel model for hydrophilic surface, the contact angle between droplets and material surface decreases with increasing surface roughness. Figure 15 further plots the changes of contact angles of two kinds of test liquids on non-textured and textured PEEK surfaces with time. It can be seen from figure 15 that the contact angle of M-SBF on PEEK surface is larger than that of the DD water, due to the higher density of the M-SBF than the DD water that leads to higher surface tension. The droplet diffusion kinetic function y = k·x n is fitted using the data presented in figure 15, which is employed to describe the relationship between contact angle and time [36], where y is the contact angle while x is the time, k is the empirical coefficient of initial contact angle, and n is the diffusion coefficient. Table 3   lists the droplet diffusion kinetic function values of n and R 2 by matching the wettability results shown in figure 15. The correlation coefficient R 2 represents the degree of fit of the regression equation, and its value is in the range of 0.69-0.98. The diffusion coefficient n represents the degree of droplet diffusion, and the higher value of n, the faster the diffusion [37,38]. Table 4 indicates that the surface texture has a strong influence on the diffusion behavior of droplets, and the textured surface has a lower surface diffusion coefficient than that of the non-textured surface. In addition, the diffusion coefficient of DD water droplet is 0.00553 for the textured surface, which is about 10% higher than that of 0.00498 for M-SBF. Therefore, the grid texture generated by laser ablation has a strong influence on the diffusion of droplets on textured PEEK surfaces [39].

Chemical composition modification of PEEK by LST
The laser ablation applied on PEEK surface can not only form surface textures, but also change the surface chemical composition. Therefore, the content of chemical elements and functional groups on the laser ablated PEEK surface is further analyzed by XPS spectra characterization. Figure 16 shows the XPS survey spectrum of non-textured and textured PEEK surfaces, which indicate that the two surfaces are both composed of carbon and oxygen elements from the C1s (284.6 eV) and O1s (532.0 eV) peaks, while no new elements is introduced into PEEK surface after laser ablation. Table 4 lists the results of carbon and oxygen content and their ratio on non-textured and textured PEEK surfaces with different line pitches. It can be seen from the table 4 that after laser ablation, the surface carbon content decreases and the oxygen content increases, resulting into a decrease of the C/O ratio. At the same time, the overall change of carbon and oxygen content is small, since only part of the textured surface is treated and the rest remains unchanged.
Although figure 16 demonstrates that no new elements is introduced into PEEK surface after laser treatment for each surface, the relative intensity of carbon peak and oxygen peak of textured surface are different from that of non-textured one. Specifically, the carbon content of textured PEEK surface of decreases accompanied with Figure 15. Changes of contact angle over time for two kinds of test liquids on non-textured and textured PEEK surfaces. the increase of oxygen content, indicating that oxygen content dominates the wettability of PEEK surface. To further study the influence of carbon functional group content on the surface wettability of textured PEEK, high resolution peak analysis of C1s at textured PEEK surface is performed. By fitting the C1s spectra, the content of carbon functional groups in C1s peak on textured PEEK surface with different line pitches are shown in figure 17.  Figure 18 further presents variations of carbon content and contact angle with line pitch, which shows a similar trend with each other. In particular when the line pitch is 400 μm for the minimum contact angle, the carbon content is also the minimum, while the oxygen content is the maximum. In the laser ablation process, some carbon elements on the PEEK surface are ablated and combine with oxygen form oxygenated compounds, i.e., newly formed O-C=O bond and C-O bond by C-C bond breaking, thus increasing the content of oxygen elements on the surface. The change in the oxygen functional group content suggests that the original C=O groups of the PEEK surface is replaced by the newly formed polar carboxylic acid groups. While the laser ablates the material surface, the laser energy is absorbed to cause chain scission, resulting into the formation of new polymer chain terminated by carbonyl groups on the surface.
The increase of oxygen element content by laser ablation leads to enhanced surface polarity and surface free energy of textured PEEK, thus promoting its surface wettability.

Conclusions
In conclusion, grid textures on PEEK materials have been precisely fabricated, and the wettability of the textured surfaces is also evaluated. We find the most appropriate values of laser ablation parameters, as a repetition frequency of 100 kHz, a scanning speed of 100 mm s −1 and an average power at 1.25 W, which are chosen to achieve high quality microgrooves with high precision surface morphology on PEEK surface. Subsequently, a precise grid texture on PEEK surface is fabricated, which has a line pitch of 400 μm, a depth of 26 μm and a width of 55 μm. Subsequent surface wettability evaluation shows that the as-prepared microgrooves effectively  facilitate the droplet spreading, while the contact angle decreases for either DD water or M-SBF. The XPS test results show that the surface polarity of textured PEEK is significantly increased due to the formation of C-O bond and O-C=O bond on the ablated surface, which in turn leads to the enhancement of the surface polarity of PEEK accompanied with enhanced surface wettability.