Change of Adhesion Properties of Bioinspired Laser-Induced Periodic Nanostructures towards Cribellate Spider Nanofiber Threads by Means of Thin Coatings

Abstract

We investigated the effect of additional continuous functional coatings, which changed the hydrophilic–hydrophobic properties of the surface without heavily influencing the surface topography at the nanoscale, on the antiadhesive properties of bioinspired laser-induced periodic nanostructures. These nanostructures mimic the antiadhesive structures on the silk-combing area on the legs of cribellate spiders, the calamistrum. The thin films were deposited by matrix-assisted laser deposition and characterized by infrared spectroscopy, X-ray photoelectron spectroscopy, water contact angle measurements, and adhesion tests using capture threads from the cribellate spider webs. In all cases, the nanoripples were preserved and these structured surfaces showed lower adhesion forces compared to flat controls, although not significant. However, this effect was totally overwhelmed by the difference between the adhesion forces on surfaces with different chemical compositions. The largest adhesion forces were observed on hydrophilic surfaces and the lowest ones on hydrophobic surfaces. The fact that the antiadhesion between nanofibers and the nano-structured areas depends strongly on the chemical composition of the surface can be explained by the specific adhesion between individual chemical groups due to frequency dependencies in the theory of van der Waals forces. However, explaining these adhesion properties just by the categories “hydrophilic” or “hydrophobic” is oversimplified.

1. Introduction

Material scientists and engineers have long been thrilled by the application potential of nanofibers as their surface-to-used material ratio is beneficial for, e.g., biomedical applications [1]. As outlined in [2], nanofibers are widely used in biomedical applications for (bone, skin, cardio vascular, and neural) tissue engineering [3], in drug delivery applications [4], and wound dressing [5]. Furthermore, nanofibers are used in composite applications [6] as supports for catalysts [7], air filtration [8], wastewater treatment [9,10], food packaging [11], and other applications [12]. Another interesting field of applications is nanofiber-based flexible transparent electrodes, which were reviewed in [13]. However, technical nanofiber processing, transportation, or even simple things such as spooling are often hindered by their strong attraction to any surface by van der Waals forces, originating from static or induced electrical dipole–dipole interactions [14,15]. Recently, a biological example has been explored, aiming to tackle these problems biomimetically: cribellate spiders, that is, spiders that use dry capture threads with mechanical stickiness instead of capture threads coated with viscid glue [16]. Their capture threads contain thousands of 15–30 nm-thick nanofibers, organized into structural complex threads with wooly puffs surrounding thicker axial fibers [17]. To process the nanofibers, cribellate spiders bear at their hind legs a specialized comb, the calamistrum [18,19]. The <40 nm-thick fibers in the puffs do not stick to the calamistrum due to a special ripple-like nanostructure. This structure causes the nanofibers to not smoothly adapt to the surface of the calamistrum, but rather minimizes contact and thus reduces the adhesive forces between the nanofibers and the calamistrum [2,20]. Figure 1a shows a magnified photograph of a cribellate spider of the species Uloborus plumipes, Figure 1b shows a scanning electron microscopy (SEM) image of a thread with its nanofiber capture wool (organized in characteristic puffs), Figure 1c shows a SEM image of a thread on the surface of a fruit fly (i.e., a typical prey of this spider), and Figure 1e,f shows SEM images of the surface of the calamistrum and of some of the bristles with the antiadhesive ripple-shaped nanostructure and some nanofibers of the capture wool artificially placed onto the calamistrum for size demonstration.

As we have explained in more detail in [20], the calamistrum, with its bristles covered by nanoripples (Figure 1d–f), functions as an antiadhesive surface and prevents the spider from adhering to its own nanofibers (Figure 1b,e,f) during their extraction. These results were confirmed by comparing the antiadhesive properties of legs with native calamistra to those with shaved-off calamistra. The nanofibers adhered significantly better to the legs without a calamistrum and in all cases where the calamistrum was removed, residues of fibers clotted the original location of the calamistrum. Therefore, a special feature of the calamistrum is to provide antiadhesive properties towards nanofibers, i.e., by a reduction of van der Waals forces.

We have demonstrated that these ripple features can be mimicked by laser-induced periodic surface structures (LIPSS) on polyethylene terephthalate (PET) foils, rendering these foils also antiadhesive [20]. LIPSS, i.e., those of type LSFL-2 [21], are formed by the interference of the linearly polarized pulsed laser light with diffracted or scattered radiation remnants in the surface after multi-pulse irradiation at fluences well below the ablation threshold. Here, the orientation of the structures is parallel to the laser polarization. We fabricated LIPSS on PET foils and other polymeric films using a 248 nm KrF* laser with variable ripple spacing and height in the ranges of 203–613 nm and 63–161 nm, respectively [22]. All features in this parameter range are antiadhesive as well.

As is discussed in detail, for instance, in [23], fouling, i.e., unwanted adhesion, is a complex and undesirable process where materials from the environment, such as macromolecules, microorganisms, or suspended particles, adhere reversibly or irreversibly to a surface. This process is a widespread obstacle, causing problems in medical, marine, and industrial applications. Several means have been developed to realize antifouling strategies, including the modification of the surface chemistry, surface topography, and architecture. The first two approaches emphasize the change in surface characteristics often by applying a coating, while the role of the coating interior is included in the latter.

For adhesion phenomena, in general, there is often a discussion about whether adhesion on the micro- and nanometer scale depends rather on the “topography” or “chemistry” of the surface (see for instance [23]). As we know that such topographic ripples, as described above, can have antiadhesive effects for nanofibers, we tested in the current study whether additional functional coatings influence this property too. As a deposition method, we chose matrix-assisted pulsed laser evaporation (MAPLE). This method is excellently suited to deposit thin organic films [24]. We chose two different chemical compositions resulting in hydrophobic and hydrophilic surface properties.

2. Materials and Methods

2.1. MAPLE

A sketch of the experimental setup for thin-coating deposition by MAPLE is shown in Figure 2. More details about the setup and the method can be found in [25,26,27]. The materials utilized in this study for the fabrication of thin coatings were polyethylene oxide (PEO) (as hydrophilic coating) with the formula (–CH₂–CH₂–O–)_n and Nafion (hydrophobic coating), i.e., polytetrafluoroethylene, also known as Teflon, with the formula (–CF₂–CF₂–)_n, where some of the F-atoms are substituted by a hydrophilic sulfonate side group to make it soluble in aqueous solvents. Both materials were purchased from Sigma Aldrich. To prepare the PEO-containing solutions for use as targets in the MAPLE experiments, PEO (with an average molecular weight of 100 kDa) was dispersed in double-distilled water (serving as the matrix) to create 4 wt% solutions through sonication for 30 min. Similarly, 5% Nafion was suspended in a mixture of lower aliphatic alcohols and water. Subsequently, the polymer solutions (for different experiments) are flash-frozen in liquid nitrogen, resulting in solid targets used in the MAPLE experiments.

The frozen targets were then transferred in a vacuum chamber and the polymers were deposited onto substrates by means of a 266 nm pulsed Nd:YAG UV laser with ns laser pulses (Surelite II, Continuum, Milpitas, CA, USA) at a repetition rate of 10 Hz. The target-to-substrate distance was, in all cases, 3.7 cm. The experimental conditions are summarized in

Table 1. Experimental conditions of MAPLE films on PET samples with LIPSS.

Sample ID	Target (Starting Solution)	Laser Fluence (J/cm2)	Laser Pulse No.
PEO	PEO in water	0.4	90,000
Nafion1	Nafion in ethanol/water	0.6	54,000
Nafion2	Nafion in ethanol/water	0.8	108,000
Nafion3	Nafion in ethanol/water	1	90,000

.

2.2. Atomic Force Microscopy (AFM) Measurements

The AFM images were acquired using a XE100 microscope (Park System, Suwon, Republic of Korea). Commercial silicon cantilevers were used (OMCL-AC240TS, Olympus, Tokyo, Japan) with a 70 kHz nominal frequency and 2 N/m nominal force constant. The scanning of the surfaces was carried out in non-contact mode in ambient conditions. Multiple regions and dimensions (40 × 40 µm² and 5 × 5 µm²) were investigated to obtain data on the surface topography and roughness (R_q, which is the root-mean-square roughness), as well as the skewness (S_k) and the excess kurtosis (K).

2.3. Infrared (IR) Spectra

Fourier-transform infrared (FTIR) spectra were recorded with a Nicolet™ FT-IR iS™ 50 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an Attenuated Total Reflection (ATR) module. A ZnSe crystal was used for ATR, with the following characteristics: 1.5 mm diameter, one internal reflection at an incidence angle of 42°, depth of penetration 2.03 μm at 1000 cm⁻¹, and 2.4 refractive index.

All spectra were acquired by absorption measurements in the 700–4000 cm⁻¹ range, at room temperature, as a mean of 16 scans, at 4 cm⁻¹ resolution, and CO₂/H₂O correction.

2.4. X-ray Photoelectron Spectroscopy (XPS)

The XPS survey spectra and high-resolution XPS spectra were acquired for the PET, PET/LIPSS, PEO, and Nafion coatings deposited by MAPLE using an ESCALAB Xi + system, Thermo Scientific. The survey scans were acquired using an Al Kα gun with a spot size of 900 μm, pass energy of 50.0 eV, and an energy step size of 1.00 eV, while the high-resolution XPS spectra were acquired with 20 eV pass energy and a 0.10 eV energy step size. In total, 5 scans were accumulated for the XPS survey spectra and 10 scans for the high-resolution spectra. The high-resolution spectra were fitted using Voigt functions (Gaussian–Lorentzian convolutions) and the Shirley line-shapes background, using a procedure described previously [26]. The component positions were fixed in accordance with the binding energies (BE) of each sub-peak (component) and its width [full width at half maximum (FWHM)] was kept the same for similar sub-peaks.

2.5. Water Contact Angle Measurements

Contact-angle measurements using the sessile drop method were conducted using a KSV CAM101 microscope (KSV Instruments Ltd., Espoo, Finland) equipped with a video camera. All contact-angle measurements were obtained by dispersing water droplets with a volume of 2.5 ± 0.5 µL. Five different points were measured for every sample and the contact angle reported is the average of these measurements.

2.6. Production of LIPSS on PET Foils

The formation mechanism and the experimental setup for LIPSS formation on PET were described in our previous work, for instance in [22,28]. In brief, flat, biaxially stretched PET foils with a thickness of 50 µm (Goodfellow Ltd., Huntingdon, UK) served as a substrate for LIPSS mimicking the ripple structures on the calamistrum of the cribellate spiders. A KrF* excimer laser (LPX 300, Lambda Physik, Göttingen, Germany) was used to generate UV light pulses with a 20 ns duration and 248 nm wavelength. For sample fabrication, a pulse repetition frequency of 10 Hz and N = 6000 pulses were applied. The samples were typically fabricated using fluences between 3.9 and 4.3 mJ/cm². These parameters were chosen because they resulted in regular ripples on the whole processed area. A beta-bariumborate (BBO) polarizer (Thorlabs GmbH, Bergkirchen, Germany) generated linearly polarized light. This output was then imaged onto the samples by two fused silica lenses in a telescope configuration. A high-power variable attenuator with a dielectric coating (magnetron sputtered and with anti-reflection layer, Laseroptik GmbH, Garbsen, Germany) enabled beam energy adjustment by being mounted on a rotatable stepper motor. We measured beam energy with a pyroelectric joulemeter (Gentec from Soliton Laser-und Messtechnik GmbH, Gilching, Germany, model “ED-500”); this was placed after the last lens. With the help of a rotatable sample holder, the laser hit the surface at an angle of incidence of 30°. The resulting processed areas were approximately rectangular and about 2 cm² large.

2.7. Spiders and Adhesion Measurement with Spider Threads

Ethics: The species used in the experiments are not endangered or protected. Special permits were not required. All applicable international, national, and institutional guidelines for the care and use of animals were followed.

Study animals: Adult spiders of the species Uloborus plumipes of similar sizes were caught in garden centers across Aachen (Germany). They were held separately in 11 cm × 11 cm × 6.5 cm boxes with roughened surfaces, where they could build their webs. The spiders were fed weekly with fruit flies Drosophila melanogaster and water was provided via soaked cotton balls biweekly. Capture threads were collected from the webs before each experiment without stretching them. To ensure equal thread lengths, threads were placed on a 7.7 mm-wide sample holder.

Antiadhesion experiments: We tested structured (LIPSS) and unstructured (control) foils. To ensure edgeless contact and thus prevent the tangling of the fibers, the foils were cut into 1.5 cm-long strips and clamped in a loop form. We took caution not to kink the foils. A motorized micromanipulator (MT30-50; Standa Ltd., Vilnius, Lithuania) was used to bring thread samples into contact with foil samples. A small deformation of the thread confirmed contact, whereupon the thread was slowly and constantly (v = 2 mm/s) withdrawn from the foil until it detached. Via a video recording microscope (VW-9000C; Keyence Corporation, Osaka, Japan), the deflection was recorded at a 10- to 20-fold magnification and 60 fps. Adhesion was determined by measuring the maximum deflection perpendicular to the initial thread position using the Keyence VW-9000 motion analyzer software (version 1.4.0.0, Keyence Corporation, Osaka, Japan). We present the mean of twelve measurements, using a new thread for each experiment. The results are compared with adhesion measurements on gold-coated foils at room temperature and humidity from a prior work [28], where also the measurement method is described in further detail. The measurement procedure for polymer foils is also schematically described in Figure S1 in the Supplementary Material S1. This figure includes a photograph of the setup as well.

2.8. Statistics

For adhesion measurements, statistical analyses and data processing were performed in RStudio (Posit, Boston, MA, USA). The normal distribution of all data sets was confirmed using the Shapiro test. Data sets were then compared between coatings using ANOVA and Tukey post hoc test or between LIPSS and control foils using an unpaired t-test. Statistical significance was assumed for p < 0.05. The data are presented as violin plots with integrated boxplots.

3. Results

3.1. Results of AFM Measurements of PET Samples with LIPSS with and without MAPLE Coatings

This study explores the interplay between the surface chemistry and topography in influencing the antiadhesive properties of nanostructured surfaces, drawing inspiration from natural systems, i.e., cribellate spiders. Thus, we applied PEO and Nafion by MAPLE directly onto the PET with LIPSS, and we characterized the samples by AFM (Figure 3).

We have shown previously [26] that by tuning the experimental conditions, it is possible to obtain continuous and conformal PEO coatings by MAPLE. Moreover, the PEO coatings present uniform surfaces with few randomly distributed spherical-shaped droplets, which are expected for semicrystalline polymers such as PEO. The PEO coating shown in Figure 3b follows the morphology of the LIPSS structures (the thickness is below 50 nm). Like what has been observed previously, random “droplets” (of a micrometric size) can be seen also on the PEO coatings deposited on the LIPSS structures.

The thickness of the coatings as well as the statistical parameters such as the root-mean-square roughness (R_q), the skewness (S_k), and the excess kurtosis (K) are summarized in

Table 2. The statistical parameters, i.e., the root-mean-square roughness (R_q), the skewness (S_k), and the excess kurtosis (K) of the PET with LIPSS before and after coating by MAPLE with either PEO or Nafion extracted from the 5 × 5 µm² AFM images shown in Figure 3.

Sample ID	Thickness [nm]	Ra [nm]	Sk	K
PET w LIPSS before PEO	-	54	0.54	2.09
PET w LIPSS after PEO	40	48	0.15	2.2
PET w LIPSS before Nafion1	-	52	0.48	1.89
PET w LIPSS after Nafion1	17	50	0.08	1.68
PET w LIPSS before Nafion2	-	55	0.56	2.09
PET w LIPSS after Nafion2	41	62	−0.73	4.51
PET w LIPSS before Nafion3	-	55	0.30	1.67
PET w LIPSS after Nafion3	48	52	−0.27	2.76

. S_k assesses the extent to which a variable’s distribution is symmetrical. K is a measure of whether the distribution is too peaked (a very narrow distribution with most of the responses in the center). A positive value for the kurtosis indicates a distribution more peaked than a statistically normal distribution. In contrast, a negative kurtosis indicates a shape flatter than normal [29,30].

As a general observation, from the AFM investigations of the Nafion coatings, more pronounced features like ridges or peaks are noticed by applying higher fluences, while lower fluences result in smoother surfaces. This can also be seen in SEM images of the prepared coatings, which are presented as Figure S2 in the Supplementary Material S2. The appearance of the nano and micro features, i.e., droplets, etc., has been explained by molecular dynamic simulation studies, as the result of an explosive process in which the polymer molecules are ejected as part of matrix–polymer droplets [31]. The S_k coefficient of the Nafion1 coating obtained at 0.6 J/cm² (see Table 1) is close to zero, suggesting a symmetric distribution of the height values. In contrast, by increasing the laser fluence to 0.8 J/cm² and 1 J/cm², the S_k coefficients of the Nafion2 and Nafion3 coatings are negative, which indicates the presence of height values below the average, i.e., pits or depressions in the surface. In accordance, the results presented in Figure 3 prove that the Nafion coatings deposited at 0.6 J/cm² are continuous and conformal, i.e., they follow the LIPSS structures fabricated onto the PET substrates.

3.2. Results of IR Spectroscopy of PEO- and Nafion-Derived Coating Samples

The IR spectra of both PEO- and Nafion-derived coating samples, i.e., obtained by drop-cast and MAPLE, are shown in Figure 4.

The drop-casted PEO thin film and MAPLE-processed PEO thin coating are shown in Figure 4a. The assignment of the vibrational bands of PEO was based on already published data [26,32,33]. The characteristic absorption peaks have been identified as follows: O–H stretching vibration of hydroxyl groups at 3480 cm⁻¹, C–H₂ symmetric stretching of alkane groups at 2885 cm⁻¹, C–H₂ scissoring vibration in alkane groups at 1466 cm⁻¹, C–H₂ wagging vibration in alkane groups at 1343 cm⁻¹, C–H₂ twisting vibration in alkane groups at 1280 and 1241 cm⁻¹, a combination between C–H₂ scissoring and rocking vibration in alkane groups at 964 cm⁻¹, C–O symmetric stretching vibration in ether groups at 876, 843, and 816 cm⁻¹, and C–C skeleton vibration (rocking) at 739 cm⁻¹. In addition, the presence of the triplet peak of the C–O–C stretching vibration in the region 1000–1200 cm⁻¹ is evidence of the PEO crystalline phase.

Figure 4b shows the IR spectra of a Nafion thin coating obtained by MAPLE at 0.8 J/cm² laser fluence compared to the IR spectra of a Nafion film deposited by drop-casting. The peak assignment was performed with [34]. The spectra of both Nafion films show the asymmetric and symmetric CF₂ stretch modes at 1150 and 1207 cm⁻¹, while a satellite band at 1057 cm⁻¹ can be assigned to the symmetric SO₃ stretching mode of the sulfonic acid groups, which are typical for Nafion. The O–H stretching vibration at 3448 cm⁻¹ corresponds probably to physically adsorbed water and the peak observed at 981 cm⁻¹ could be attributed to the C–O–C stretching. During laser ablation, the sulfonic group containing side chains extending outward from the free surface may fold back toward the bulk of the ionomer film, exposing the nonpolar part, which results in a hydrophobic surface.

In the PEO coatings obtained by MAPLE, changes in hydrogen bonding have caused a shift from 1092 cm⁻¹ to 1100 cm⁻¹. This can be attributed to a more crystalline structure of the PEO processed by MAPLE compared to the drop-casted PEO. Furthermore, two peaks, i.e., at 1720 cm⁻¹ and 1748 cm⁻¹, can be noticed in the MAPLE-processed PEO spectra, which could originate from the substrate.

3.3. XPS Results

We acquired XPS survey spectra, ranging from 0 to 1350 eV, to define the elements in the PET, PET/LIPSS, PEO, and Nafion samples (see Figure 5). The PET, PET/LIPSS, and PEO XPS survey spectra shown in Figure 5 indicate that carbon and oxygen are the main elements on the surface of the samples. In the case of Nafion, the main constituents (carbon, fluorine, and oxygen) appear as well-defined peaks, whereas the sulfur yielded a small peak with a poor signal-to-noise ratio.

For a detailed analysis of the surface chemistry of the samples, high-resolution XPS spectra are recorded in the region of C1s and O1s for PET, PET/LIPSS, and PEO, while for the Nafion sample, the high-resolution spectra of fluorine are also recorded. The high-resolution XPS spectra are shown in Figure S3 in the Supplementary Material S3.

The deconvoluted C1s and O1s XPS spectra of PET and PET with LIPSS are shown in Figure S3a–d. The C1s spectrum of the pristine PET is deconvoluted into four components, i.e., C1 at 284.8 eV due to the C–C and C–H bonds, the C2 component at 286.3 eV due to C–O bonds, the C3 component at 288.8 eV due to the O–C=O bonds, and the C4 component at 291.5 eV due to π–π* the shake-up transition associated with the aromatic ring at 291.5 eV [35]. Following the best fitting procedure, 54.2% C1, 24.1% C2, 16.8% C3, and 4.9% shake-up percentage areas were obtained. As a result of the laser treatment, some changes in C1’s peak can be noticed. Similarly, it can be decomposed into four components, i.e., C1 at 284.8 eV, C2 at 286.3 eV, C3 at 288.8 eV, and C4 at 291.5 eV. However, the relative area of the 284.8 eV decreases (i.e., by 43.1%) compared to the unprocessed PET. In addition, the concentration of the oxidized carbon species on the polymer surface increases after laser processing the PET surface, i.e., C2 is 23.3% and C3 is 29.8% for the PET/LIPSS sample. The oxygen is incorporated on the polymer surface introducing the C–O and O–C=O polar groups, which leads to an increase in the PET’s surface energy and reduction in the water contact angle (see Figure 6, where the contact angle of the PET/LIPSS is around 42° compared to the contact angle of the pristine PET which is around 80°).

The O1s spectrum of the untreated PET sample shows the expected O1s doublet peak (BE at 531.6 eV) due to carbonyl O atoms and the O1s corresponding to ester O atoms (BE at 533.2 eV). In addition, two more peaks at 535 eV assigned to surface contamination (adsorbed water, hydroxyl groups (OH), or other oxygen-containing contaminants on the polymer surface) and 537.5 eV appear in both PET and PET/LIPSS samples with carboxylate (COO-) groups, from the hydrolysis or degradation of ester linkages in the PET polymer.

Similarly, we carried out peak assignment for the Nafion films deposited by MAPLE onto PET/LIPSS (Figure S3e–g). The fitting of the C1s region revealed several lines, each corresponding to the Nafion chains in the polymer structure. The lines at 284.8 eV, 286.6 eV, 288.5 eV, and 291.2 eV depict C-C, C-F bonds, or carbon atoms adjacent to the CF₂ groups, reflecting the partial electron-withdrawing effects of fluorine atoms in the fluoropolymer matrix, C-O bonds found in the ether links or C-SO₃H groups, where the carbon is bonded to the sulfonic acid group, C-SO₃, and CF₂ bonds, respectively [36]. Furthermore, oxygen yields signals at 533 eV and 535.4 eV and a less intense peak at 538.8 eV. The O1 and O2 peak components in the O1s spectra of Nafion are consistent with previous studies, and their chemical assignment is correlated to the oxygen present in the ether functional group (535.4 eV) and oxygen in the sulfonate group (533 eV). In addition, the low-intensity peak at 537.9 eV binding energy could be assigned to water molecules or hydroxyl groups adsorbed on the surface of the Nafion coating.

The F1’s spectra show a high peak at 689.5 eV, which is attributed to the C-F bonds within the fluoropolymer structure, a peak at 691.8 eV, and another smaller peak at 695.1 eV binding energy. The peak at 691.8 eV could be assigned to C-F bonds adjacent to electron-withdrawing groups or end-group or defect-related fluorine, while the peak at 695.1 eV could be assigned to shake-up satellites.

In the case of PEO coating by MAPLE, the high-resolution carbon (C1s) and oxygen (O1s) XPS spectra together with their deconvolutions are shown in Figure S3h and S3i, respectively. The C1 region is fitted considering four components, which are assigned according to the position of the carbon binding energy in PEO [26,37,38]. The first C1 component at 284.8 eV is assigned to the C-C found in the backbone of polyethylene oxide. The peak at 286.2 eV can be attributed to C-O single bonds, i.e., to the ether linkage (–O–) present in the repeating unit. Moreover, the peak at 288.9 eV is associated with carbon atoms in more oxidized states, i.e., C=O (carbonyl groups) or O-C-O (carbonate groups), in particular related to minor oxidized impurities or defects. Finally, the peak at 291.2 eV binding energy is attributed to shake-up satellites associated with oxidized species.

The O1s region was fitted considering the binding energy at 531.0 eV (O1, double bond O in O=C), at 532.9 eV (O2 single-bonded O in –O*–C, C–O, HO*–C), and at 536.3 eV (O3 assigned to O in water or O–C–OR).

3.4. Results of Water Contact Angle Measurements of PET Samples with and without LIPSS and MAPLE Coatings

The results are summarized in Figure 6. The raw data are shown in Table S1 in the Supplementary Material S4. The as-deposited Nafion films (both on PET flat substrates and on PET with LIPSS) are hydrophobic, with static contact angles in the range of 85–115°. In contrast, the contact angle of the PEO thin films is below 60°. For the PEO films, a minimum contact angle of approx. 36° was observed. The reason is that the large number of hydrophilic groups on the thin-film surface plays a favorable role in increasing the hydrophilicity of the thin film.

3.5. Results of Adhesion Measurements of Spider Threads on PET Samples with and without LIPSS and MAPLE Coatings

We tested the influence of previously characterized coatings on the adhesion of the natural nanofibers of U. plumipes cribellate threads to biomimetic PET foils, as well as to non-rippled control foils. Figure 7 shows the results of the antiadhesion measurements. The legends in the figure refer again to the sample names in Table 1. The same samples were characterized before by water contact angle measurements (see Figure 6). Gold-coated samples produced by sputter coating are shown for comparison. The raw data of Figure 7 are listed in Table S2 in the Supplementary Material S5. The measured quantity is the deflection. To convert the measured deflection in parameters like force and stress, one would need the mechanical properties of the individual threads from the spider webs, which are not available.

In all cases, the PET samples with LIPSS showed slightly lower adhesion values than flat controls; however, these were not significant except for the gold coating (n = 12 per coating, n = 25 for gold). The lower adhesion of the PET samples with LIPSS is in full accordance with our earlier results from the references [20,22,28], where the differences to the flat control samples were significant in most cases. Even a higher reduction of the adhesion forces could be expected if the LIPSS structures could be combined with structures in the µm range, mimicking the bristles of the calamistrum of cribellate spiders like in [2], which is, however, not feasible with the materials used in this study. Although no (significant) antiadhesive effect of the LIPSS could be shown after coating the samples, the hydrophobic Nafion coating nevertheless showed a reduction in adhesion by about 50% (compared to the uncoated PET control, Figure 7). The hydrophilic PEO coating, on the other hand, strongly increased adhesion, especially for some of the data points for PEO without LIPSS.

3.6. Results of Control Experiments Using Different Materials

To see if there is a clear correspondence of hydrophilicity/hydrophobicity and the adhesion of a material to cribellate spider silk, we tested the interaction of capture threads from the cribellate spider Uloborus plumipes to the hydrophilic materials quartz glass and a mixture of cellulose and cellulose acetate, as well as the hydrophobic materials Teflon and bee’s wax. We took Teflon purchased from RS Components (Gmünd, Austria), quartz glass from GVB GmbH (Herzogenrath, Germany), bee’s wax from Carl-Roth (via Lactan, Graz, Austria), and cellulose acetate from Sigma (St. Lewis, MO, USA).

Freshly harvested threads were brought into contact with PET and the deflection at the breaking of the interaction was taken as a reference. Then, the same thread was brought into contact with the different materials and the deflection was normalized to the deflection observed with PET, i.e., the deflection for PET was set to 100%. Therefore, values lower than 100% mean smaller deflections (and values higher than 100%, larger deflections) when compared to the deflection observed with PET. The results are depicted in Figure 8. Clearly, there is no direct correspondence of adhesion and hydrophobicity/hydrophilicity. While the hydrophilic quartz glass and the hydrophobic Teflon hardly interact with spider silk, there is a strong interaction with the hydrophilic cellulose and the hydrophobic bee’s wax.

4. Discussion

We could demonstrate that it is possible to deposit high-quality functional coatings onto PET foils with LIPSS, mainly preserving the surface topography (see Figure 3). As demonstrated by the water contact angle measurements, it is possible to obtain highly hydrophobic, as well as hydrophilic properties (see Figure 6). We found that surface coatings have a strong influence on the adhesion of the cribellate threads (see Figure 7). Although hydrophobic Nafion did not yet reach similar values to samples coated with gold, it was able to reduce the adhesion significantly, especially compared to the samples coated with PEO. Gold could have the advantage that its conductivity eliminates electrostatic effects that would otherwise overshadow the reduction of van der Waals forces.

At first glance, it is counter-intuitive that cribellate spider silk adheres strongly to the hydrophilic PEO, while it hardly adheres to the hydrophobic Nafion-derived coating, as in nature, the cribellate catching threads adhere to the extremely hydrophobic epicuticular waxes of prey insects [39,40]. However, cribellate silk also consists of a surprisingly high amount of hydrophilic amino acids, which might explain why we could detect a stronger interaction between spider silk and hydrophilic surfaces [17]. A closer look at the van der Waals (vdW) interaction, mainly responsible for solid body adhesion, might give another explanation for this surprisingly increased adhesion (for details on vdW theory, see, e.g., [41]).

The van der Waals potential, describing the interactions of dipole fluctuations and their induced dipoles, depends on the materials involved and the geometry of the problem under investigation. In Hamaker’s theory, one can look over all microscopic van der Waals potentials and use the assumption that the resulting potential is additive (which is not the case in reality). To avoid the problem of non-additivity, the potential can alternatively be calculated using the Lifshitz’s theory.

Generally, the van der Waals potential between bodies of materials A and B, separated by a material m (e.g., air) of thickness l, is given as G_vdW = −A_Ham·K_geo(l) where K_geo(l) is a geometry factor describing the shapes of the bodies (which depends on the thickness of the separating layer m) and A_Ham is the Hamaker constant. A_Ham depends on the polarizability of the materials. Often one can read

$$A_{Ham}~\frac{{\epsilon }_A-{\epsilon }_m}{{\epsilon }_A{+\epsilon }_m}·\frac{{\epsilon }_B-{\epsilon }_m}{{\epsilon }_B{+\epsilon }_m},$$

(1)

with ε_i as the dielectric functions (corresponding to the polarizability) of the different materials. However, these dielectric functions of the materials are functions of the frequency f or circular frequency ω = 2πf, respectively. To calculate the Hamaker constant, one would need the complex dielectric function ε(ω) = ε′(ω) + iε″(ω). Here the imaginary part corresponds to energy dissipation in the material. Via the Kramers–Kronig transformation, one can obtain ε(ω). In the Lifshitz theory, the vdW interaction can be described as a function of ε(iξ_n), which is ε(ω) evaluated at the discrete thermal Matsubara frequencies ξ_n = 2πnk_BT/ħ where n = 0, 1, …, ∞, k_B is the Boltzmann constant, ħ is the reduced Planck constant, and T is the absolute temperature. Thus, if the distance l of the interacting bodies is small, so that the finite speed of light can be neglected, one can obtain:

$$A_{Ham}\approx \frac{3k_{B}T}{2}{\sum }_{n=0}^{\infty }\frac{{\epsilon }_A\left({i\xi }_n\right)-{\epsilon }_m\left({i\xi }_n\right)}{{\epsilon }_A{\left({i\xi }_n\right)+\epsilon }_m\left({i\xi }_n\right)}·\frac{{\epsilon }_B\left({i\xi }_n\right)-{\epsilon }_m\left({i\xi }_n\right)}{{\epsilon }_B\left({i\xi }_n\right){+\epsilon }_m\left({i\xi }_n\right)}.$$

(2)

To have strong interactions, the Hamaker constant needs to be large. If now the separating material is air or vacuum, the ε_i of the materials A and B at the Matsubara frequencies should match. This yields what Parsegian [41] called the “dance of charges” or resonance of charge fluctuations and thus corresponds to the lowest free energy.

Different kinds of fluctuations correspond to different frequencies or characteristic times τ, respectively:

• Bound electrons corresponding to ultra-violet (UV) and higher frequencies with characteristic times τ <≈ 10⁻¹⁷ s;
• Vibrating molecules corresponding to IR with τ ≈ 10⁻¹⁶ to 10⁻¹² s;
• Rotations at microwave frequencies with τ ≈ 10⁻¹¹ to 10⁻⁶ s;
• Mobile charges with τ down to 0.

For non-polar organic materials, we expect the IR to UV frequencies to dominate for the Hamaker constant. Looking at the spectra of PEO [42], Nafion [43], and spider silk [44], respectively, one can see that in the UV and in the visible (VIS) range, Nafion is almost transparent and lacks IR-peaks that correspond to amides, in contrast to PEO (for comparison see also Figure 4). However, spider silk shows absorption at different UV-VIS wavelengths and has some dominant IR peaks due to amide oscillations.

In contrast, for highly polar materials with a large ε_i at very low frequencies, like water (ε(0) ≈ 80), interacting with non-polar materials, such as hydrocarbons (ε(0) ≈ 2), the magnitude for zero frequency is (80 − 2)/(80 + 2), which is almost equal to 1. Clearly this first (static) term in the summation over n in formula (2) dominates its (oscillating) successors.

Thus, although not quantitative, this analysis indicates that from the fact that material A is hydrophobic, one cannot conclude that a material interacting with another hydrophobic material, B, will indeed strongly interact with material A, too. Hence, whether our observation of the increased adhesion of nanofibers to hydrophilic surfaces can be transferred to artificially spun nanofibers remains questionable too. A detailed analysis of the vdW interaction over the whole frequency range would be necessary.

As insect surfaces are normally hydrophobic, one would presume that evolution would have achieved this type of surface. But certainly, the explanation of the adhesion forces between a nanofiber and a (corrugated) surface just through the hydrophilic or hydrophobic categories is oversimplified (see Figure 8). One could also discuss this topic in terms of oleophobic and oleophilic, which are not just equivalent to hydrophilic and hydrophobic, especially not for fluorinated polymers. Further significant influential factors might be the electrical conductivity or specific adhesion between individual chemical groups, for instance. Meanwhile, other interaction mechanisms of functional surfaces, such as the exchange of ions [45,46], are probably of minor relevance in the case of our polymer materials. We think that we have, with the new findings of this work, laid the starting point for further interesting investigations in the field—both for theorists and experimentalists.

Our new finding is that antiadhesive properties towards nanofibers of the nano-structured areas like those at the calamistra of cribellate spiders can depend strongly on the chemical composition of the surface. This raises the question of why do these spiders have nano-structured surfaces on their calamistrum at all. One point here is that their surface may have further functions, for instance, water or dirt repellence or robustness against wear, which reduces the possible choice of surface chemistries. Another point, which is exemplarily addressed in Figure S4 in the Supplementary Material S6, is that we showed in our previous work that our bioinspired laser-induced periodic nanostructures are always considerably shallower than the original ones in the calamistrum [20], which makes them certainly less antiadhesive; however, there is no large influence of the nanostructure periodicity on the antiadhesive properties [22]. Also, the hierarchic combination of microstructures, i.e., the bristles of the calamistrum, and the nanoripples increases the antiadhesive properties [2]. This altogether may provide the cribellate spiders with sufficient nanofiber repellence based on a topographic effect, mainly or at least to a considerable extent.

5. Conclusions

Our approach to functionalize bioinspired laser-induced periodic nanostructures by means of MAPLE was successful and we could prepare functional surfaces which follow the LIPSS structures fabricated onto the PET substrates. We found that both hydrophilic and hydrophobic coatings deposited by MAPLE affected the antiadhesion of biomimetic polymer foils significantly. In contrast to our expectations, cribellate fibers tend to stick better to the tested hydrophilic than to hydrophobic surfaces. We think that the explanation lies in the missing overlap of the resonance frequencies between the hydrophobic Nafion-derived coating and the spider silk, in contrast to the hydrophilic PEO, resulting in small contributions to the dynamic Hamaker constant at the corresponding discrete Matsubara frequencies. Our results may help to develop antiadhesive tools for improved nanofiber handling which can open new application fields.