Plasma depositions were carried out on mirror polished stainless steel (SS) disks (304-8ND, APERAM, 2 cm diameter, 0.8 mm thickness) and two face-polished silicon wafers (Sil'tronix), notably for IR and AFM analyses, respectively. A home synthesized N-(3,4-dihydroxyphenethyl)acrylamide, called dopamine acrylamide (DOA) was produced according the procedure reported in Patil et al. (2015). The 2-hydroxyethyl methacrylate monomer (HEMA; 97%) was purchased from Sigma–Aldrich and used as received or after a purification step using a silica column. The Micro-BCA protein assay kit was purchased from Thermo Scientific. The liquid precursor mixture was obtained by dissolving solid DOA in HEMA during 30 min at a final 3.3 mg mL⁻¹ concentration. SS substrates were cleaned by successive ultrasonic washings in acetone (5 min), and absolute ethanol (5 min) and further dried under a nitrogen flux. Prior deposition, SS substrates were activated through an Ar: O₂ (19 standard liter per minute (slm):1slm) plasma treatment in continuous mode at 1.6 W cm⁻² during 2 min. This plasma treatment ensures an ultimate cleaning and promotes the anchoring of the deposited functional layer.

Coatings were deposited using an atmospheric pressure dielectric barrier discharge (AP-DBD) process. The discharge was produced between two plane-parallel high voltage electrodes covered by alumina and a moving table as the ground electrode, thus generating a ~19 cm² plasma discharge area. Samples were placed on the moving table and the electrode gap adjusted at 1 mm. The precursor solution was nebulized using a OneNeb injector system (Agilent) and a syringe pump to control its flow delivery (5 μL min⁻¹). The droplet size distribution of the precursor spray, examined using the Spraytech laser diffraction system from Malvern, was centered around 1 μm. During the deposition process, the plasma was generated using a Corona generator (SOFTAL electronic GmbH) with a 10 kHz sinusoidal signal. Coatings were deposited using a modulated electrical excitation, i.e., pulsed wave with different t_ON and t_OFF pulses. The total process gas flow was fixed at 20 slm of argon (99.999 %, Air Liquide). The applied voltage was measured with an HV probe (Lecroy, PPE, 20 kV, 100 MHz). The current was estimated using a Lecroy (CP030, 50 MHz) Hall Effect current probe. The charge was determined by measuring the voltage drop across a 9.4 nF capacitor connected in series with the ground electrode. Data were recorded with a digital oscilloscope (Lecroy, WS 42XS, 300 MHz bandwidth). The average power (W_a) values were calculated according to Equation (1).

Optical Emission Spectroscopy (OES) measurements have been carried out using an ARC SpectraPro-2500i spectrometer equipped with a PIXIS 400 CCD detector. Process conditions were a peak power of 30 W and a pulsed discharge at 1:400 ms in pure argon.

The mass of the films was estimated by weighing the substrate before and after deposition using a Sartorius ME-36S microbalance. Each deposition condition was carried out, at least, on 4 substrates. As a result, the reported value corresponds to the average values. The thickness of the layer was estimated using a contact profilometer (P-17, KLA Tencor). The growth rates in weight and thickness were determined considering the measured masses, thicknesses and the effective deposition duration.

The coating morphologies were observed by Scanning Electronics Microscopy (SEM) using a Hitachi SU70 HRSEM instrument. Prior observations, in order to avoid charging effects, the samples were coated with a 7 nm platinum layer using a BAL-TEC MED020 metallizer equipment.

Atomic force microscopy (AFM) analyses were carried out in constant amplitude tapping mode (2.8V) with scanning speeds ranging from 1 Hz to 2. The tip (HQ15NSC15) came from Micromasch (325 KHz, 40 N/m). Images were reprocessed with the SPIP software. Roughness parameters were calculated by subtracting the mean plane.

UV-visible absorption spectroscopy analyses were carried out using a PerkinElmer Lambda 950 UV–vis-NIR (InGaAs) spectrophotometer equipped with an integrating sphere. Samples were dissolved in a 2 mL absolute ethanol solution.

Fourier Transform Infrared Spectroscopy (FT-IR) analyses were performed in the Attenuated Total Reflectance (ATR) mode on a Hyperion 2000 spectrometer (Bruker). The spectra, acquired at 32 scans and a 4 cm⁻¹ spectral resolution, were processed using the Opus 7.5 software.

X-ray Photoelectron Spectroscopy (XPS) analysis was carried out using the Kratos AXIS Ultra DLD instrument, equipped with a hemispherical energy analyzer and a monochromatic X-ray source (Al Ka, hν = 1486.6eV) operating at 150 W. Global and high-resolution XPS spectra were recorded with a pass energy of 160 and 40 eV, respectively. Spectra were reprocessed with the Casa XPS software. All spectra were calibrated using the aliphatic C1s carbon peak at 285 eV. The C 1s core level was fitted with 5 main components: (i) C-C/C-H (285 eV), (ii) C-CO-O (285.7 eV), (iii) C-O / catechol (286.7 eV), (iv) C = O/O-C-O (288 eV), and (v) O-C = O (289 ± 0.1 eV) bonds (Beamson and Briggs, 1992). The full width at half maximum (FWHM) and the Gaussian/Lorentzian ratio were fixed at 1.0 ± 0.1eV and 30%, respectively. The experimental uncertainty related to the XPS surface elemental composition is around 2 at %.

Mass spectrometry analysis was carried in order to evaluate the structure of the plasma deposited layers. Electrospray ionization mass spectrum was recorded using Waters Synapt G2-Si mass spectrometer. Firstly, the coating was dissolved in an absolute ethanol (2 mg mL⁻¹). The solution was then diluted 10 times in methanol and 4 μL of a NaI solution (4 mg mL⁻¹ in acetonitrile) was added. This solution was then directly infused in the ESI source with a typical flow rate of 5 μL min⁻¹ using a capillary voltage of 3.1 kV, a source temperature of 100°C and a desolvation temperature of 200°C. The quadrupole was set to pass ions from m/z 100 to m/z 2,000. All ions were transmitted into the pusher region of the time-of-flight analyzer (Resolution~20,000) for mass-analysis with 1 s integration time. Data were acquired in continuum mode until acceptable average data were obtained (<5 min).

To estimate the catechol content in the deposited thin films, two different methods were exploited, both relying on the catechol oxydo-reduction property. The first technique is based on a commercial Micro BCA kit, originally exploited for protein quantification and later extended to catechol (Liu et al., 2015). Briefly, the method relies on the reduction of Cu²⁺ to Cu⁺ by catechol. The Cu⁺ then complexes with bicinchoninic acid (BCA) which absorbs light at 562 nm (Slocum and Deupree, 1991). The second technique, issued from own developments, is based on the combination of AgNO₃ immersion tests with potentiometric measurements. Indeed, immersed in the silver salt solution (Ag⁺), catechol-bearing thin films operate as a reducing agent, thus converting silver ion into silver nanoparticle (Ag⁰) (Lee et al., 2007, 2008). The silver salt reacting with catechol group in a 1:1 molar ratio (Equation 2), the amount of Ag⁺ consumed during the reaction should equal the initial amount of catechol groups. As an indirect measurement, the estimation of the Ag⁺ concentration in the AgNO₃ solution, before and after the layer immersion step, indicates the catechol amount in the deposited films.

In the dark, SS coated with plasma layers of known thicknesses were immersed in a 6 mL of an AgNO₃ solution (1mg mL⁻¹) during 24 h and under stirring. Afterwards, the solutions were recovered and titrated by potentiometric measurements. A known volume of a NaCl solution (5.339 10⁻⁴ M) was added while the evolution of the potential difference, between a silver electrode and an E_Ag/_AgCl reference electrode, was followed. Hence, by graphical reading, it was possible to deduce the equivalent volume, V_eq, allowing to determine the final Ag⁺ concentration as reported in Equation (3). Preliminary measurements carried out on selected plasma polymerized (pp) layers confirm the reliability and reproducibility of the method. Indeed, Ag⁺ consumptions of 1 and 26 % (compared to the initial Ag⁺ concentration) are measured for pure pp(HEMA) (i.e., layer of reference) and pp (DOA-HEMA), respectively.

In the dark, SS coated with plasma layers were immersed in a 6 mL of an AgNO₃ solution (1 mg mL⁻¹) during 24 h and under stirring. Afterward, the solution was collected and a drop placed on microscope glass slide. Visualization was performed using Cytoviva® hyperspectral imaging system (Cytoviva Inc., Auburn, Alabama, USA) mounted on Olympus BX-43 optical microscope. Images were captured at 60-fold magnification with oil immersion using hyperspectral camera controlled by environment for visualization ENVI software (version 4.8 from Harris Corporation, Melbourne, FL, USA and modified by CytoViva®, Inc.).

Figure 1 illustrates the open-air atmospheric aerosol-assisted Dielectric Barrier Discharge process operating with argon. A mixture of DOA and HEMA (DOA:HEMA = 0.3:99.7 wt/wt) was nebulized and injected in the plasma zone. OES of the argon dielectric barrier discharge was used to identify various chemical species in the plasma gas phase. In the 300–400 nm region, an intense spectral line from •OH at 309 nm and small emission lines due to NH and N₂ at 338 and 359 nm, respectively were identified (Fang et al., 2016). The •OH emission peak was probably caused by fragmentation of H₂O molecules, which diffused from ambient air. Similarly, the detected nitrogen species might originate from the diffusion of ambient air into the discharge. Finally, in the 700–800 nm range, atomic oxygen emission at 777 nm and different emission lines attributed to argon were also observed. For a fixed 1.6 W cm⁻² power density (Figures S1, S2), the influence of the pulsing of the plasma discharge (and thus on different switching ON and OFF periods) on the deposition rate, chemistry and morphology of the films is investigated.

In a first set of experiment, t_ON is varied from 1 to 15 ms while t_OFF is fixed at 15 ms to ensure the renewing of the gas mixture between the electrodes during each ON-time. Figure 2A reports an example of pulse (i.e., t_ON + t_OFF) composed of a 5 t_ON and 15 ms t_OFF. Figure 2B displays the evolution of the layer thickness and mass deposited per pulse as a function of different t_ON. For a 1 ms t_ON, the mass deposited is 1.13 ng cm⁻² with a layer thickness of 7.4 10⁻³ nm per pulse. Increasing t_ON induces a depletion in the mass deposition per pulse with the value that decreases to 0.2 ng cm⁻² pulse⁻¹ with a t_ON of 15 ms. This result suggests that part of the monomer dissociates into volatile fragments for higher t_ON, consequently it does not contribute to the film growth. Interestingly, considering the dynamic character of the deposition technique, the thickness of the coating can be easily adjusted by programing the number of runs (Figure S3).

According to SEM and AFM pictures (Figure 3) of 100 nm thick coatings, it appears that pulsing of the discharge is likely to affect both deposition rate and film morphology. Indeed, a short t_ON duration (i.e., 1 ms, Figure 3A) provides a smooth and pinhole-free coating that is covering homogeneously the entire substrate surface. By increasing t_ON from 5 to 10 ms, rough coatings are formed, with the roughness that is increasing with t_ON (0.9 nm vs. 10.3 nm). Numerous particles are also observed that present an average diameter size up to 267 nm for a 10 ms t_ON (Figure 3E).

To gain insight into the coating chemistry, FT-IR measurements are carried out. Figure 4A presents the IR spectra of pure HEMA (as reference) and coatings deposited at different t_ON. Irrespective of the deposition conditions, one can notice that the film spectra present some HEMA units fingerprint peaks, such as C-O bending in CH₂-O at 1,454 cm⁻¹, C-C-O stretching at 1,166 cm⁻¹ and C-O in primary alcohol at 1,079 and 1,031 cm⁻¹ stretching. The disappearance of the HEMA C=CH₂ peak at 3,107 cm⁻¹ along with the shift toward higher wavenumber of the C=O peak (from 1,718 cm⁻¹ to 1,726 cm⁻¹) suggests an efficient polymerization reaction through the C=C bonds, associated to the loss of conjugation for the C=O bond. In contrast, any peak relative to the DOA contribution was noticeable. One explanation may be the initial low DOA content in the precursor mixture. Interestingly, films deposited at high t_ON present: (i) a band broadening around 1,273 cm⁻¹, usually related to strong monomer fragmentation processes and resulting in high crosslinked films (Veuillet et al., 2017) and (ii) a new peak contribution at 1,623 cm⁻¹, probably related to the formation of C=C bonds. Complementary bulk UV-Vis analyses are carried out in order to detect the presence of catechol group in the films deposited at different deposition conditions. Preliminary assays reveal an absorbance at 300 nm only detected in a DOA-HEMA solution and related to catechol (Figure S4). Such result is consistent with the absence of absorbing group in plasma polymerized (pp) layer from (HEMA), noted pp(HEMA), thus confirming the relevance of this analysis. Figure 4B reports the UV spectra of different pp(DOA-HEMA) layers (presenting an averaged 75 nm thickness) in the 200–300 nm region. Interestingly, a maximal of absorbance at 260 nm, attributed to catechols, is observed for films deposited at 1 and 5 ms t_ON. The variation in the maximal wavenumber position might be attributed to different catechol environments, the film being dissolved in ethanol prior UV measurements (Figure S5A). Indeed, plasma deposition carried on polished mirror stainless-stell disks and analyzed in UV reflectance mode indicate a catechol peak at 300 nm (Figure S5B). Considering the layer deposited at 10 ms and 15 ms t_ON, a small shoulder is only noticeable for 10 ms, suggesting a lower catechol content. To support the presence of catechol in the film, complementary Cytoviva (i.e., Enhanced darkfield Hyperspectral microscope) analyses have been carried out. This technique enables rapid characterization of nanoparticles and is mainly employed by biologists to investigate histological samples (Vallotton et al., 2015; Mehennaoui et al., 2018). In this work, the technique was exploited to track silver nanoparticles issued from the reduction of silver salt by the catechol. Indeed, it is now well-established that silver nanoparticles (AgNPs) are efficient at absorbing and scattering light and have a color that depends upon the size of the particle (Théoret and Wilkinsona, 2017). Here, a pp(layer) deposited at 1:15 ms was immersed in AgNO₃ solution overnight. The solution was collected and analyzed via Cytoviva. The dark field microscopy image, reported in Figure 5A, highlights the presence of numerous bright blue and green points. As shown in Figure S6A, the bright blue color, due to a surface plasmon resonance (SPR), is peaked at a 450 nm wavenumber and might be correlated to the presence of AgNPs with a 60 nm size. In Figure S6B, the maximal intensity of the spectral profile for the green points is around 540 nm. Such profile shift to higher wavelengths is related to an increased diameter of the AgNPs, estimated around 100 nm. Similarly, a pp(DOA-HEMA) film deposited at 10:15 ms was reacted with AgNO₃. In Figure 5B reporting the dark field microscopy picture, no bright point is visible, indicating the absence of AgNPs. The presence of rods might be related to some film fragments issued from the film. Hence, in accordance with the reported UV results (i.e., Figure 4B), it can be concluded that using a 10 ms t_ON induces the degradation of the catechol-based monomer, explaining its absence in the deposited film. These results are in agreement with the IR ones, suggesting that depositions carried out at high t_ON values lead to coatings with poor monomer structure retention.

Figure 6 summarizes the results of the XPS analyses of pplayers from pure HEMA and DOA-HEMA deposited at 1 and 10 ms t_ON. Irrespective of the precursor nature, coatings deposited at 1 ms t_ON present a similar C:O content around 71:29 at.% (Figure 5A). Nitrogen is not detected, suggesting neither air contamination in the deposited film considering pp(HEMA) nor DOA detection in pp(DOA-HEMA). This latter observation may be explained by the extremely low DOA content in the monomer mixture. Indeed, nitrogen is present in the precursor mixture at 0.01 at. %. Pp(HEMA) and pp(DOA-HEMA) deposited at 10 t_ON present similar C and O amounts as well as a N content around 1–2 at.%. Here, the presence of nitrogen in pp(HEMA) coating might be related the influence of air contamination due to predominant gas phase reactions in such operating condition. From the XPS C1s overlappings of pp(DOA-HEMA) layers deposited at 1:15 ms and 10:15 ms (Figure 5B), it appears that the t_ON value strongly influences the carbon environments. Indeed, the C1s deconvolution of layers deposited at short t_ON (Figure 5A) present higher ester contribution (16 vs. 11 %), which is consistent with IR results. It is worth noting that both deposition conditions contribute also to the HEMA fragmentation as both layers contain a novel contribution attributed to C=O bond.

To quantitatively estimate the catechol content in the deposited film, a micro-BCA assay kit, already applied to pDA coatings (Shin et al., 2011), was firstly exploited. However, a lack of specificity toward catechol compound was observed, with catechol detection for native SS substrates and pp(HEMA). Therefore, a method relying on the oxidation of catechols with an excess of silver nitrate, followed by potentiometric titration of excess silver cation is carried out before and after being in contact with plasma deposited layers. This technique allows quantifying the catechol content in pp(layers) (see experimental section for details) with an accuracy of 1%. This analytical method is applied to two pp(DOA-HEMA) deposited at different t_ON. The catechol densities are estimated to be 34 and 18 μg cm⁻² for the pplayer deposited at 1 and 5 ms t_ON, respectively. This result is fully consistent with the previous bulk and surface layer characterizations, highlighting a lower monomer structure retention when working at higher t_ON duration. Indeed, using pulsed discharges with long t_ON duration ensure plasma depositions according to a so-called monomer deficient regime. During t_ON, high energy is transferred to monomer molecules. The collision rate between monomers and electrons/ions is increased, leading to an increased monomer dissociation rate and a loss of the precursor structure in the deposited film (Kakaroglou et al., 2015). Gas phase reactions dominate the film growth, thus favoring particles/agglomerates formation and rough-to-powder-like coatings.

In the next section, the lowest t_ON value (i.e., 1 ms) is selected for its low monomer fragmentation impact. The study is extended to explore the effects of varying the t_OFF in an attempt to determine the optimal t_ON/t_OFF ratio to maximize the functional group retention for catechol-bearing plasma polymers.

Figure 7 displays the layer thickness and mass deposited per pulse according to different t_OFF duration. Interestingly, the thickness and mass rates exhibit a similar trend, characterized by a logarithmic layer growth for cycle values ranging from 15 to 400 ms and followed by a plateau. Therefore, it appears that the layer growth is the highest for a t_OFF of 400 ms.

SEM and AFM analyses demonstrate that, irrespective of the depositions conditions, smooth and pinhole-free coatings covering homogenously the substrate are achieved with a maximal average roughness (Sa) of 1.5 nm. As an example, Figure 8 presents the SEM and AFM pictures of a 150 nm layer deposited at 1:400 ms.

Also, IR spectra of the deposited films are similar in all cases (Figure 9A). The absence of the monomer vinyl bands at 3,107 cm⁻¹ and 1,638 cm⁻¹ confirms an efficient polymerization via carbon double bonds C=C bonds. The series of peaks located at 1,454 cm⁻¹ (C-O), 1,166 cm⁻¹ (C-C-O) and 1,079 and 1,031 cm⁻¹ (C-O in primary alcohol) are in line with a well-retained poly(HEMA) structure in the deposited layers. Complementary UV analyses (Figure 9B) definitely confirm the DOA incorporation in all deposited layers as the result of the catechol peak absorbance at 264 nm.

According to Figure 10 reporting the XPS analysis of pp(DOA-HEMA) deposited at 15 and 400 ms t_OFF, it appears that the coatings present a similar chemical composition but different carbon environments. Indeed, increasing t_OFF from 15 to 400 ms considerably improves the monomers structure retention in the layers, namely ester (COOR) and catechol (C-O/catechol) from HEMA and DOA, respectively, and limits the formation of secondary groups, such as C=O groups resulting from HEMA fragmentation.

The catechol content in films deposited at 1:50 ms and 1:400 ms is then estimated by the potentiometric titration as described above. Extending t_OFF from 50 to 400 ms significantly increases the catechol content from 51 to 138 μg cm⁻². These results are fully consistent with the previous measurements, mentioning a 33 μg cm⁻² catechol density for a 1:50 ms deposited layer. Optimal depositions are therefore achieved with a 1:400 ms pulsed discharge leading to smooth and catechol rich layers.

To gain insight into the plasma deposited layer (pplayer) structure and composition, complementary mass spectrometry analyses are carried out (Jackson and Simonsick, 1997). In this purpose, Electrospray Ionization (ESI) Mass spectrometry analyses are performed by using a direct injection of the diluted pplayer solution. The as-obtained mass spectrum is reported in Figure 11. Interestingly, various distributions of oligomer ions separated by 130 mass units are unambiguously identified. This mass difference is perfectly consistent with HEMA monomer mass unit, thus confirming a layer formation from radical polymerization reactions occurring during t_OFF. In addition, MS analyses clearly highlighted many different end-groups (Figure S7). This result might be explained by the numerous and various chemical species issued from the monomer fragmentation generated during t_ON. Hence, in Figure 11, red dots distribution corresponds to poly(HEMA) with, formally CH₄ as end-groups, and characterized by DP superior or equaled to 18. Considering the green dot distribution, this latter may be attributed to poly(HEMA) with one formaldehyde molecule loss. Many signals separated by 2u lower are also detected and might be attributed to H₂ losses. Finally, by comparing pp(HEMA) and pp(DOA-HEMA) MS spectra, no significant modification of the mass spectrum is observed. This result might be explained by the fact that (i) only the soluble part of the polymer is analyzed by ESI-MS, (ii) the poor DOA concentration in the precursor mixture along with the huge number of different oligomer species present in the 250–1,500 mass range. Hence, using pulsed discharges favoring long t_OFF vs. short t_ON durations ensure plasma depositions according to a so-called monomer sufficient regime. This latter is defined by a concentration of active species involved in plasma polymerization lower than the monomer molecules density in the plasma zone, the layer growth being directly governed by the modulation of the electrical discharge. Hence, during each short t_ON (i.e., 1 ms), there is monomer activation and fragmentation as well as the generation of reactive sites at the substrate surface via UV irradiation, ion and electron bombardments. During the subsequent t_OFF, so in the absence of any UV, ion or electron induced damage, conventional and heterogeneous free-radical polymerization reactions occur between radicals at the surface and fresh monomer molecules continuously injected and being absorbed at the surface (Ward et al., 2003). Such mild deposition conditions allow the formation of smooth layers with extremely high levels of structural retention of the precursors, here, catechol as well as ester functional groups.

Finally, as an attempt to correlate deposition kinetics and physico-chemical properties of deposited films, the evolution of the catechol content and deposition rates are plotted as a function of a W_a/F factor (W_a and F corresponding to the average power in W estimated according to Equation 1 and the monomer flow rate fixed here at 8.9 10⁻⁸ Kg s⁻¹ (i.e., 5 μL min-1), respectively) (Figure 12). Indeed, originally introduced for low-pressure plasma processes, this factor has been transposed to AP-DBDs (Hegemann et al., 2007, 2009; Nisol et al., 2016). According to Figure 12, two distinct deposition regimes are identified with a competition zone identified around 9.3 MJ/Kg corresponding to an optimal deposition rate of 0.62 nm/s. In zone I (i.e., monomer sufficient regime), defined by low values of W_a/F (in the 0.4–9.3 MJ/Kg range), the deposition rate is increasing from 0.14 to 0.62 nm/s. In zone II (i.e., monomer deficient regime), defined by W_a/F value exceeding 9.3 MJ/kg, the deposition rate decreases down to 0.05 nm/s. Figure 12 is therefore a useful and simple tool to select the appropriate conditions to prepare specific layers with targeted thickness, catechol content and morphology. Importantly, the experimenter should respect the t_ONand t_OFF values fixed in each zone as well as the precursor flow rate. Indeed, using similar W_a/F values do not imply the formation of similar coatings (Kakaroglou et al., 2015). Hence, as an example, in <2.5 min, 50 nm thick and smooth coatings containing 100 ug cm⁻² catechol can be synthesized by depositing at a W_a/F of 1.7 MJ/kg (i.e., 1 ms t_ON and 200 ms t_OFF). Similarly, 50 nm thick and rough coating containing 25 ug cm⁻² catechol can be deposited in 8 min. using a W_a/F of 50 MJ/kg (i.e., 3 ms t_ON and 15 ms t_OFF).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.