Ultrathin Film Hydrogels with Controlled Swelling and Viscoelastic Properties Deposited by Nanosecond Pulsed Plasma Induced‐Polymerization

Development of ultrathin film (utf) hydrogels for cutting‐edge biomedical applications (i.e. artificial skins) is receiving increasing attention. Nonetheless, achieving accurate control on the structure and thickness of utf‐hydrogels becomes extremely complex when assessed through conventional techniques. In this work, an atmospheric‐pressure plasma‐assisted deposition technique is reported, showing great thickness accuracy and versatility, to design utf‐hydrogels with customized properties. For the first time, specific and independent control on the generation and nature of cross‐links by only changing the plasma exposure frequency (fPE) during the synthesis process are reported. Thus, utf‐hydrogels are successfully prepared with tuned swelling ratios and viscoelastic properties (ranging from 150 to 20 kPa). Moreover, a thickness accuracy of 9 nm is reported, permitting the accurate synthesis of utf‐hydrogels below 150 nm. Exhaustive structural and topographical analyses allow elucidating the effects of the fPE on the cross‐link generation mechanism, discarding any undesired effect on the thickness accuracy. To support the structural results obtained, quartz‐crystal microbalance with dissipation (QCM‐D) coupled with spectroscopic ellipsometry are put in the spotlight as an efficient and viable alternative for the characterization of the resulting properties of ultrathin film soft materials, including the presence of a hydrated layer at the interface.


Introduction
The great potential of hydrogel materials have been reported for a wide range of disciplines, among which biomedicine stands DOI: 10.1002/admi.202300644out particularly. [1]1e,2] Certainly, the most cutting-edge technologies reported rely on the design of hydrogels with controlled architecture down to the nanometric scale, showing tremendous potential for intelligent drug delivery, [3] enhanced sensitivity, [4] or enhanced bioadhesion. [5]Besides, ultrathin film hydrogels (i.e., <1 μm) can show increased swelling ratios, overcoming their theoretical confinement limits. [6]For this reason, developing useful tools to synthesize and characterize ultrathin film hydrogels with controlled physical and appropriate morphological and mechanical properties is crucial. [7]urrently, the most conventional approach to obtain hydrogel thin films is based on the spin coating technique, which provides good homogeneity and thickness control at the micrometric scale. [8]However, some important drawbacks appear when further reducing the thickness of the films, as the required accurate control on the viscosity of the fluids might interfere with the polymerization and crosslink generation processes. [9]This fact results in a loss of the control on the resulting properties of the hydrogel, while clearly increasing the complexity and limitations of the syntheses routes.9a] Atmospheric pressure (AP) plasma-deposited polymers are reported as an attractive alternative to the spin coating technique due to their excellent thickness control at the nanometric scale. [10]uch control is achieved because the plasma-initiated polymerization is solvent-free.That is, polymerization is triggered as a result of the interaction of a precursor with the energetic species of the plasma gas.Therefore, liquid precursors can be accurately deposited on the desired surface (or directly injected in the plasma gas flow) [11] in the form of aerosols for its subsequent polymerization in a single-step approach.Among other advantages, [12] AP plasma-assisted polymerization routes are far less time-consuming and foremost versatile; supporting a wide range of substrates without any specific geometrical restrictions and favoring an easy industrial scalability by its simple implementation of the plasma applicator in roll-to-roll processes or robotic devices.
AP plasma-initiated polymerization kinetics are characterized by a radical polymerization mechanism competing with a fragmentation/recombination process. [13]While generation of radical species is required to trigger the polymerization, unbalanced number of radical species versus monomer content can also favor chain termination, spontaneous crosslinking, and fragmentation/recombination reactions losing the chemical structure of the initial monomer.Hence, plasma polymers typically present lower molecular weights and a higher variety of crosslinks, side chains, or end groups as compared with polymers prepared by conventional wet polymerization routes.Consequently, the hydrogel-like plasma deposited films reported in the literature also present a high number of crosslinks, resulting in a very poor control of the resulting viscoelastic and swelling properties. [14]ots of efforts have been put to elucidate the optimum experimental conditions for controlled polymerization, as the kinetics are very sensible to a wide range of parameters, such as the atmosphere composition (presence of oxygen and reactive oxygen species), [13a] the initial set-up (precursor is injected in plasma indischarge or post-discharge, distance of the plasma discharge to the substrate…) [11] or plasma power, [15] among much others. [16]n this context, our group has recently published a work [17] reporting the successful plasma polymerization copolymerization of different acrylic monomers in the liquid state.Interestingly, we also demonstrate that plasma exposure time can be effectively used to control the kinetics of polymerization.However, the aforementioned optimization has been scarcely explored for hydrogel materials.In this sense, achieving an accurate control on the number of cross-links (comparable to that obtained with wet chemical routes) [18] specifically for ultrathin film hydrogels, is extremely attractive but also challenging.Indeed, enhancing the control on the final hydrogel properties can be considered as the principal challenge to overcome to properly establish plasma technologies as a viable route for developing biomedical applications.
In this work, we have continued our previous works with nanosecond pulsed plasma discharges to prepare, for the first time, ultrathin film hydrogels with controlled thickness, swelling, and mechanical properties through AP dielectricbarrier-discharge (DBD).Advantageously, this control has been achieved by optimizing the plasma conditions, rather than changing the chemical nature of the initial precursors.Methacrylic acid (MAA) has been chosen as the monomer and ethylene glycol dimethacrylate (EGDMA) is used as the cross-linking agent.Due to the size and wet nature of ultrathin film hydrogels, reliable characterization is difficult to assess through conventional techniques. [19]Herein, we establish the use of quartz-crystal microbalance with dissipation (QCM-D) coupled with spectroscopic ellipsometry (SE) operating under static and dynamic conditions as a powerful tool to characterize the mechanical properties of wet ultrathin films.Detailed discussion on its procedures, advantages, and data interpretation is also provided.Complementarily, exhaustive structural and morphological measurements have been carried out to support the QCM-D/SE results obtained.

Synthesis of MAA/EGDMA Ultrathin Film Hydrogels
Methacrylic acid (MAA) and ethylene glycol dimethacrylate (EGDMA) were purchased from Sigma-Aldrich (99% purity).MAA/EGDMA ultrathin film hydrogels were prepared by means of a customized AP-DBD setup exhaustively described elsewhere. [10]Schematically represented in Figure 1, MAA/EGDMA (90/10%mol) mixture was sprayed on single-side polished silicon wafers (dopped, 10-20 Ohm cm, diameter = 50.8± 0.3 mm, <100>, Siegert Wafer GmbH) using a venturi nebulizer (VITO).The precursor was injected in the venturi system at a molar flow rate of 0.93 mmol min −1 .Nitrogen gas flow (1.8 SLM (standard liter per minute), 99.999%) was used to generate the droplets and as carrier gas carrier gas (0.5 SLM).Then, the liquid thin layer was exposed to the AP-DBD plasma to trigger the polymerization.Argon (20 SLM, 99.999%) was used as the plasma gas, while the plasma discharge was shaped into a sequence of square discharges generating nanosecond pulses (200 ns, time on) that were applied spaced in time (time off).Hence, the plasma exposure frequency (f PE ), which accounts for the number of nanopulses per second irradiated on the samples (see inset in Figure 1), is defined as an important experimental parameter under consideration.On the other hand, thickness of the samples was easily controlled by repeating the whole procedure for a desired number of runs.This process has been automatized through the implementation of a moving table acting as the grounded electrode.
MAA/EGDMA ultrathin film hydrogels were deposited onto silicon wafers which were previously activated with 5% oxygen in the plasma gas at 100 Hz for 51 s.Samples were prepared considering both f PE (1 and 3 kHz) and number of runs (5 to 100).Besides, MAA and EGDMA polymers (PMAA and P(EGDMA), respectively) were also deposited separately by injecting 100% mol of MAA and EGDMA as precursors.Note that EGDMA polymers have been explicitly abbreviated as P(EGDMA) to avoid confusions with PEG-DMA (where only the ethylene glycol polymerizes).Plasma conditions were set at 100 runs and f PE = 1 and 3 kHz to ensure proper structural characterization.

Characterization Methods
Fourier transform infrared (FTIR, Bruker Vertex 70) and Xray photoelectron spectroscopy (XPS, ThermoFisher Scientific Nexsa-G2) analyses were performed to elucidate the effects of f PE on the samples.Water contact angle (WCA, Krüss DSA100 Drop Analyzer) experiments were carried out to further validate the structural information derived from the spectroscopic techniques.Additionally, gel permeation chromatography (GPC, ThermoFisher Scientific Dionex UltiMate 3000 LC) measurements were conducted only to PMAA due to solubility requirements.Second, topological characterization of the samples was mainly performed by means of atomic force microscopy (AFM, Oxford Instrument MFP3D Infinity microscope).Thickness and roughness parameters (average roughness, S a ; average height, S z ; and mean square roughness, S q ) were obtained for the ultrathin films in the dry state, while phase imaging was collected to study the homogeneity of polymerization and crosslinking on the samples.To provide further discussion on the most appropriate technique for characterizing soft materials, AFM results were alternatively corroborated with optical microscope (Keyence VK-X Series) pictograms and spectroscopic ellipsometry (SE, J.A. Woollam Alpha-SE) measurements.Detailed information on the equipment and measurement procedures can be found in the Supporting Information.

QCM-D Measurements
Quartz crystal microbalance with dissipation (QCM-D) measurements were conducted using a Q-Sense Explorer QCM-D from Biolin Scientific with a quartz sensor reaching up to seven odd harmonics.Before each measurement, sensors (QSX 304, Q-Sense) were placed as a substrate in the AP-DBD setup for its subsequent coating under the same identical conditions reported for silicon wafer.Static measurements in air were performed to obtain the dry thickness of the hydrogels.To do so, the fundamental and overtones resonant frequencies were collected from the sensors before and after plasma deposition and compared by means of the Sauerbrey equation (see Supporting Information for more details).
The Q-Sense Explorer was equipped with an ellipsometer flow module (coupled QCM-D/SE cell, QELM 401, Q-Sense) with temperature control allowing static and dynamic measurements on the samples submerged in liquid media.Dynamic experiments were designed as a four-block environment protocol to study the thickness evolution, effect of ionic strength, and viscoelastic properties of the samples as follows: 1) de-ionized water, 2) Dulbecco's phosphate buffered saline (D-PBS) solution, 3) de-ionized water; and 4) air (drying).Solutions were injected at 20 μL•s −1 (IPC, ISMATEC) and the working temperature of the analysis chamber was set to 24 °C.Measurements were allowed to equilibrate for > 2 h for each block.Coupled QCM-D/SE measurements were considered to further confirm and refine the results obtained.
Data analyses to obtain the film thickness d and the shear modulus μ was performed using the software Q-Tools provided by the manufacturer.More specifically, the fittings from the experimental data were obtained considering an extended viscoelastic model modeling the hydrogel layer as a frequency-dependent Voigt model adapted to mathematical power relationship. [6]

Structural Characterization
The effect of f PE on the resulting structure of the hydrogels is elucidated in this section.Other experimental parameters, such as the monomer flow rate or the EGDMA cross-linker concentration, were previously optimized, establishing as a criterion the appearance of a solid and adherent coating (by qualitative naked eye and microscope inspection and finger rubbing test) showing suitable water stability (not shown).Once fixed, FTIR analyses were initially conducted on the hydrogels synthesized at f PE = 1 and 3 kHz (hereafter named as MAA:EGDMA-1 and MAA:EGDMA-3, respectively), using PMAA (3 kHz), MAA and EGDMA monomers as references.As can be seen in Figure 2, both hydrogels show the distinctive vibrations at 1703 and 1723 cm −1 attributed to the ester moiety (C═O stretching) contained in the MAA and EGDMA monomers, respectively. [20]he clear reduction of the vinyl vibrations (C═C stretching) at 1637 cm −1 demonstrates the proper polymerization of MAA and EGDMA (generating cross-link) triggered by the double bond consumption.Moreover, a shift to higher wavelength numbers of the C═O band can be observed for polymerized samples, supporting the free-radical polymerization mechanism proposed. [17]oth samples exhibit most of the characteristic MAA and EGDMA absorption bands (as highlighted in Figure 2) proving that the structural integrity of the monomers is mainly preserved.However, fragmentation/recombination phenomena cannot be neglected in AP plasma-induced polymerization, as observed by the widening, splitting and shifting of some MAA peaks, like the C─O stretching in ─COOH at 1186 cm −1 , the triplet centered at 1389 cm −1 (attributed to CH 3 symmetric/antisymmetric deformations and skeletal vibrations) [20] or the new doublet at 1246 and 1271 cm −1 attributed to EGDMA (see FTIR spectra obtained for P(EGDMA) in Figure S1, Supporting Information).Finally, it is worth noting that the increment observed on the CH 2 -related vibrations at 2983 cm −1 (CH 2 stretching, see inset in Figure 2b) and 1481 cm −1 (CH 2 deformation vibrations) also evidences the proper polymerization of MAA. [21]ertainly, to develop a robust method to control the resulting properties of hydrogels it would be extremely interesting to control separately the evolution of the cross-linker (i.e., EGDMA) from the polymerization kinetics of the backbone chains (i.e., MAA).Therefore, the effect of f PE on the MAA polymerization has been firstly examined by means of FTIR and GPC techniques (Figures S2 and S3, Suporting Information).Accordingly, PMAA samples present an apparent molecular weight of 1700 g mol −1 with a relatively high polydispersity index of 1.7.Interestingly, none of the spectra obtained show any representative differences when increasing the f PE from 1 to 3 kHz, settling down a solid base to characterize the effect of f PE on the generation and nature of cross-links exclusively.
Exhaustive analyses of the spectra depicted in Figure 2 shed some light on the possible cross-link mechanism and effect of f PE .On one side, the presence of the O─H stretching vibrations at the 3100-3600 cm −1 region attributed to the carboxylic group of MAA for both hydrogels allows to discard the generation of crosslinks through the nucleophilic activation of the carboxylic acid by means of EGDMA (,-unsaturated carbonyl compound), resembling an oxa-Michael addition reaction. [22]Although it was a quite expected result, it has important implications as maintaining the hydroxyl group in MAA (together with the ester groups present in EGDMA) strongly contributes to the swelling behavior of the resulting hydrogel.Consequently, the most plausible cross-linking mechanism relies on the alkene reaction of EGDMA with the methylene group of MAA generated as a result of the C═C cleavage during polymerization (Scheme 1).Therefore, comparing the intensity of the CH 2 stretching band (2983 cm −1 ) in conjunction with the skeletal fingerprint vibrations of the hydrogel (i.e: 816, 1296, 1321, and 1389 cm −1 ) could provide only a tentative characterization (due to FTR sensitivity and peak overlapping between PMMA and P(EGDMA), see Figure 2a; Figure S1, Suporting Information) of the cross-link generation.Indeed, MAA:EGDMA-3 shows a slightly higher amount of CH 2 groups (compared to the CH 3 symmetric/antisymmetric stretching vibrations at 2939 and 2956 cm −1 ) and a significantly lower intensity for the skeletal vibrations, suggesting a higher number of cross-links (see inset in Figure 2d).
Consequently, the most remarkable difference between hydrogels can be found in the ratio between MAA and EGDMA, defined by the intensity of their characteristic C═O stretching vibrations (inset in Figure 2c), being a 9% higher for MAA:EGDMA-1, and thus, demonstrating the presence of less cross-linking agent.Moreover, apart from the aforementioned doublet (1246 and 1271 cm −1 ), MAA:EGDMA-3 presents a shoulder of the peak at 1163 cm −1 attributed to the C-O stretching vibration of EGDMA (see inset in Figure 2d).Such increase in the C-O species compared to the ketone group might indicate the homopolymerization of small units of EGDMA by attacking the ketone group through the plasma activation of the vinyl group of another EGDMA unit (Michael addition reaction).
That is, during the chain-growth polymerization higher f PE leads to higher EGDMA initiators, allowing their propagation with MAA (generation of cross-link) but also with other EGDMA monomer.More specifically, while initially the two vinyl of EGDMA have equivalent reactivity, when the first vinyl group is incorporated to the growing MAA chain the reactivity of the Scheme 1. Cross-link mechanism proposed.
remaining vinyl group considerably decreases due to conformational and steric restrictions. [23]Hence, the reactivity of the hanging vinyl has to be triggered by an external source (plasma), another initiator (EGDMA, homopolymerization), or another pendant radical group; all of them resembling chain termination reactions.Note the evident role of increasing the plasma exposure in all the three scenarios.Accordingly, MAA:EGDMA-1 has a higher amount of unreacted C═C groups, as evidenced in the inset of Figure 2 by its characteristic vibration at 1637 cm −1 .
To further confirm the results obtained by FTIR, in-depth XPS studies were considered.Figure 3a,b compare the XPS C 1s of PMAA (1 and 3 kHz, Figure 3a), MAA:EGDMA-1 and MAA:EGDMA-3 samples (Figure 3b).As it can be observed, all samples present the expected peaks located at 285.0, 285.4,286.7, and 289.2 eV attributed to C─(C─H), C─COO, C─O and O═C─O species, respectively. [24]While the presence of EGDMA can be clearly confirmed by the increase of the C─O peak, no evident differences between hydrogel samples can be detected (Figure 3c).Although the results might seem contradictory, they were supported by water contact angle (WCA) measurements showing very similar results for both samples (59.8°± 1.3°and 59.3°± 2.3°for MAA:EGDMA-1 and MAA:EGDMA-3, respectively).Therefore, plasma effects on the upmost top surface properties of the thin films by varying the f PE have been discarded.Nonetheless, XPS measurements at the bulk (see the sputtering experimental details in the Supporting Information) reveal an increase of the C─O species in front of O═C─O for MAA:EGDMA-3 (Figure 3d), which are in complete agreement with the FTIR results reported.Similarly, same quantification analysis has been conducted for PMAA samples in Figure S4 (Supporting Information) showing no differences between f PE conditions.Finally, the carbon-oxygen ratio (at.%) was calculated for all the samples showing a very similar result for all of them (71.2±1at.% for carbon) corroborating that any sample undergoes a dramatic fragmentation/recombination process due to f PE .The values determined for each sample are listed in Table S1 (Supporting Information).
Overall, the structural characterization of the hydrogels reveals the presence of a higher amount of EGDMA when raising the f PE , potentially increasing the number of cross-links.Surprisingly, choosing the appropriate f PE range and fixing the rest of experimental conditions shows an unprecedented control on the generation and nature of cross-links without affecting the polymerization kinetics of the main polymer chains neither modifying the chemistry at the surface of the hydrogels.Despite the fact that a higher number of cross-links normally results in a higher cross-link density associated with lower swelling ratios and increased rigidity, results also suggest that EGDMA is homopolymerizing at 3 kHz resulting in longer cross-link segments (improved flexibility and swelling capacity), clearly competing with the aforementioned phenomena.Such result is assessed in the following sections of this work.Because of its intrinsic nature, detailed and unequivocal structural characterization of hydrogels becomes very difficult.18c,d] Although the structural results reported must be taken cautiously and have to be supported by further characterization of the hydrogel properties, the approach presented here might be useful for further structural determination of hydrogels prepared by means of AP plasma-assisted techniques.

Topological Studies
1e,2,3b] For this reason, in this section we study the effect of f PE and number of runs on the topography of the prepared samples.More specifically, AFM and SE have been considered as complementary techniques to determine the thickness of the hydrogels.
The thickness dependence on the number of runs is depicted in Figure 4a.As it can be seen, a clear linear relationship (R 2 = 0.9) is obtained for small number of runs showing ultrahigh  thickness accuracy of  d = 9 nm/#run.Interestingly, f PE does not considerably affect the resulting thickness of the hydrogels in the linear regime.However, a clear divergence between f PE at 100 runs (already noticeable for 55 runs) is observed, where the energetic/mass balances of f PE cannot be neglected.Anyhow, it is very interesting noticing how f PE = 1 kHz still shows a fair linear dependence at 100 runs, which supports the fact that apart from increasing the number of cross-links, f PE = 3 kHz favors the homopolymerization of EGDMA.
Comparison between AFM and SE results deserve special attention as both techniques arise from completely two different approaches.AFM measurements are obtained by the step-height procedure, i.e., by scratching the coating surface with the tip of a needle, which allows distinguishing within the surface of hydrogel and substrate (silicon wafer), and thus, direct and unequivocal measurements can be obtained.Apart from the fact that it can be considered a partially destructive technique (due to scratching), the principal challenge relies on obtaining representative data.To do so, different regions comprising 20 × 20 μm 2 are averaged (representative measurements are shown in Figure S5, Supporting Information), leading to a time-consuming procedure.On the other hand, SE approach is much faster and offers more versatility.However, the thickness determination is obtained indirectly through modeling and mathematical fitting of the exper-imental data.Therefore, SE approach strongly depends on the proper modeling which is not trivial for soft materials due to the presence of water, roughness, or porosity, among others. [25]For the case under study, a simple model considering a Cauchy layer for the hydrogel layer and a standard Si substrate (parameterized semiconductor layer) were assumed.Representative spectra and fittings obtained are presented in Figure S6 (Supporting Information).The parity plot presented in Figure 4b evidences the excellent correlation between both techniques (points found in the ± [5,10] % interval range), confirming the validity of the SE model used; including the values obtained for the refractive index (detailed discussion is provided in Figure S7, Supporting Information).
Finally, topographic images of some the most representative hydrogel samples are presented in Figure 5a,b and Figure S8 (Supporting Information).Notice how XPS measurements report homogenous systems (check the conservation of the carbonoxygen ratio) discarding any kind of layer stratification.Further inspection shows very smooth and homogenous surfaces (S q = 0.8 ± 0.4 nm, in average) independently of the experimental conditions used.The most relevant topographical results have been summarized in Table 1.It is worth remarking that from a macroscopic point of view, optical microscope (Figure S9, Supporting Information) displays the presence of humps that can be roughly 1 to 2 μm higher than the flat surfaces for samples treated at 100 runs.On the other hand, SEM analyses of a utf-hydrogel synthesized at f PE = 1 kHz and 15 runs were considered as a complementary characterization approach (Figure S9, Supporting Information).Despite of being in agreement with the thickness measurements obtained by means of SE and AFM, SEM inspection has not been further considered as it presents some important drawbacks related with the procedure analyses (i.e., drying method or manually scratching the sample prior analysis).Nonetheless, a close examination of the phase images depicted in Figure 5c reveals a consistent small phase shift for samples treated at 3 kHz already anticipating mechanical differences between hydrogels. [26]Accordingly, samples presented in Figure S8 (Supporting Information) present the same trend.Despite of that, phase images show very homogenous cross-link distributions across the ultrathin films areas analyzed (see the welldefined Gaussian distributions), independent of the experimental conditions used (f PE and number of runs).

QCM-D Technique: Determination of the Swelling and the Viscoelastic Properties
Characterizing the resultant properties of ultrathin film hydrogels becomes a particularly complex task when assessed through conventional techniques.First, obtaining the swelling ratio (i.e., weight difference, Equation S2, Supporting Information) can lead to enormous inaccuracies as the typical measurement errors (specifically when determining the weight of the hydrogel in the dry state as reference) might fall in the measurement range itself.Second, determination of the corresponding mechanical properties can be only obtained by means of AFM or nanoindentation techniques because of the fact that most ultrathin films are deposited directly onto desired substrates (i.e., they are not selfstanding).Nevertheless, reporting reliable data is demanding due to the soft nature of the samples and the presence of water during the measurements.Generally, these measurements are timeconsuming (requiring specific cantilever tips or sample preparation processes) and become complex when reliable quantitative results are needed. [19,26]lternatively, QCM-D represents a suitable technique for characterizing ultrathin film soft materials under both dry and wet states. [6,28]The fundamental principle, valid only under the rough assumption that the total thickness of the sample is maximum ≈500 nm, [29] relies on measuring differences in the resonant frequencies (directly related to mass increment) and vibration dissipation (viscoelastic behavior).Certainly, QCM-D is a non-destructive, versatile, and fast technique with extraordinarily high sensitivity and accuracy; able to perform measurements under different media and working temperatures.28b] Like the SE technique, the principal challenge of QCM-D resides in using a mathematical model that is properly converging into a solution with physical sense.Besides, most of the models used are over-determined leading to multiple solutions.This problem can be partially solved by collecting multiple resonant frequencies (odd harmonics).However, QCM-D measures differences in weight, and thus, is impossible to independently determine the thickness and density of the samples.For this reason, coupling both SE and QCM-D during all the measurements is extremely attracting.Briefly, SE can be applied to determine the thickness which is going to be used to determine the density in the dry state by QCM-D.Exhaustive details on all the experimental procedures are reported in the methods section and in the Supporting Information.Consequently, the density in the dry state is reported for ultrathin film hydrogels prepared using f PE = 1 and 3 kHz, and 15 and 10 number runs (thickness of 141±6 and 67±6 nm, respectively); respectively (hereafter named as ut-MAA:EGDMA-1 and ut-MAA:EGDMA-3, respectively).ut-MAA:EGDMA-1 presents a dry density of 1.030 g cm −3 while for ut-MAA:EGDMA-3 is of 1.054 g cm −3 , aligned with a higher cross-link density for the latter as suggested by the structural characterization.
Although measurements in the wet state become easier because water density is generally assumed, [6,28a] coupling SE and QCM-D is still interesting because SE cannot measure the hydration layer at the interface of the hydrogel while QCM-D considers the complete system. [6]Thus, characterization of both hydrogel thickness and hydration layer can be obtained.For illustrative purposes Figure 6a depicts a general scheme of the coupled SE/QCM-D system in conjunction of the raw data of the frequency and dissipation variations obtained from ut-MAA:EGDMA-3 sample under wet dynamical measurements (Figure 6b).Complementarily, the fittings and frequency normalization graphs are presented in Figure S11 (Supporting Information).As it can be observed, qualitative information can be easily obtained considering that higher frequency variations lead to higher mass changes, and thus, higher increased thickness and swelling ratio.Moreover, using different media is interesting as the gradient in concentration exerts pressure on the hydrogel that can be understood in terms of a mechanical response.On the other hand, stability tests can be easily conducted considering different swelling-drying cycles.All the samples show excellent stability (no mass loss detected), suggesting the absence of completely unreacted monomer in the network.
Quantitative analyses are reported for both hydrogels in Figure 7.The comparison between thickness variations depicted in Figure 7a shows a higher increase for ut-MAA:EGDMA-3 sample.Moreover, it can be clearly observed a drastically different response when changing the media.That is, ut-MAA:EGDMA-1 almost does not deform under the pressure exerted by the gradient in concentration, while ut-MAA:EGDMA-3 undoubtedly shrinks, anticipating different viscoelastic properties.In Figure 7b, the total thicknesses obtained by means of SE and QCM-D in aqueous media are presented.As it has been explained, the hydrated layer can be obtained by comparing both thicknesses.Surprisingly, even though the superficial chemistry is the same for both samples, the hydrated layer is significantly higher for the ut-MAA:EGDMA-3.We attribute these differences to the thermodynamic equilibrium between absorbed water and the hydrated layer which (apart from the interface) also depends on the swelling properties of each sample.
However, the thickness obtained for the hydrated layer (i.e., subtraction between QCM-D and SE results) is higher than expected. [6,30]These results have motivated a complete revision of the mathematical models used for the fittings for both techniques including the modeling of water environment for SE (from preestablished theoretic water models to experimental modeling through Cauchy layers in blank sensors) or the exploration of other local minima for QCM-D models (results not shown).As far as we know, the mathematical models used seem to provide robust and reproducible fittings.Hence, the hydration layers derived from measurements can be attributed to real physical phenomena.To obtain more insights into the hydrated layer, a new two-step protocol was designed consisting of the measurement under heavy water (deuterium oxide, D 2 O) and de-ionized water.To avoid concentration gradients, samples were exposed to air between each step.Figure 7c shows schematically the protocol followed and the QCM-D results obtained for ut-MAA:EGDMA-3 sample.As expected, several differences can be detected in the SE/QCM-D measurements when changing the media (Figure S12, Supporting Information).Accordingly, D 2 O favors both water swelling and an increase of the hydration layer (proportional to the hydrogel swelling).
Despite of that, the most interesting result arise when analyzing the behavior of the samples when the chamber was exposed in air.As can be observed in Figure 7c, the hydrogel maintains the water absorbed and the viscoelastic response (i.e., Δf and ΔD are not zero).Moreover, the resultant thickness analyses presented in Figure 7d show a slight decrease in the thickness determined by SE (as expected) but a significant reduction of the hydrated layer.Indeed, under this specific case, both SE and QCM-D results are in complete agreement with the literature showing much smaller differences, which results in a hydrated layer of ≈7 nm. [6,30]The fact that the measured hydration layer is much higher for immersed samples could be attributed to the existence of a water layer superstructure that is not completely bounded to the surface (i.e., hydrated layer found in proteins or polymer brushes) [30] but shows some specific order and structure (less bounded water) which also vibrates according to the QCM-D sensor.All these assumptions are currently being further explored by studying the swelling kinetics of ultrathin film hydrogels.Nonetheless, the possibility of characterizing the evolution of the hydrated layer is of special relevance for the design of novel materials with specific interface-related phenomena.Specially, the approach presented here is very attractive to understand and optimize the anti-biofouling and/or anti-microbial response of hydrogels under different wet conditions. [31]igure 7e displays the swelling ratios obtained by SE and QCM-D in aqueous media.Note that due to conclusions derived from Figure 7c,d, the swelling ratio determined by SE can be used as reference for wet samples exposed to air.Even though a higher density of cross-link has been attributed to ut-MAA:EGDMA-3, the clear greater swelling capacity can be explained by the presence of longer cross-link segments which support the reactivity mechanisms of EGDMA proposed.Nonetheless, both samples show considerably high swelling ratios, which can even reach six times their weight when considering the hydrated layer (see ut-MAA:EGDMA-3 sample measured using QCM-D technique).Moreover, the presence of small homopolymerized EGDMA segments also introduces more flexibility to the system, being in complete agreement with the extraordinary difference reported for the viscoelastic modulus, as seen in Figure 7f.As can be observed, the viscoelastic modulus has been also obtained from wet samples exposed in air, showing similar results (differences have been attributed to the reduction of the hydration layer).Hence, both the QCM-D models and all the conclusions derived from the synthesis of the hydrogels can be validated.
Summing up, all the hypotheses proposed from structural mechanism have been demonstrated by means of the characterization of the resulting properties of the hydrogels.Although increasing f PE to 3 kHz increases the cross-linking density, it also creates longer cross-linking chains responsible for slightly increasing the swelling capacity as far as drastically reducing the viscoelastic of the hydrogels.It is worth remarking that the results presented here do not imply that there exists an optimum f PE .On the contrary, they demonstrate that hydrogels with customized swelling capacity and more notably, specific viscoelastic properties can be obtained by only tuning the f PE parameter.As an example, while increased rigidity (ut-MAA:EGDMA-1) could be very useful for developing artificial skins (exposed sweat and/or external humidity), the viscoelastic modulus obtained for ut-MAA:EGDMA-3 sample (μ = 20±1 kPa) falls in the ideal range for developing advanced biomedical applications such as enhanced antimicrobial behaviors. [32]AA:EGDMA-3 sample following the D 2 O-H 2 O-Air protocol, (d) Total thickness analyses obtained for wet samples exposed to air atmosphere.Similarly, the hydrated layer has been also included, (e) Swelling ratio (in DI water medium) determined by SE and QCM-D; and (f) Viscoelastic modulus.for the first case, the viscoelastic is maintained for both DI water and DPBS.

Conclusion
In this work we demonstrate that the plasma exposure frequency (f PE ) applied in atmospheric-pressure (AP) nanosecond pulsed plasma polymerization processes can be effectively used to independently control the generation and nature of cross-links; without affecting the chemistry of backbone chains, the chemical nature of the interface nor the topography of the resultant hydrogels.Hence we report, for the first time, the synthesis of AP plasma-assisted ultrathin film MAA:EGDMA hydrogels with customized swelling ratios and viscoelastic properties, resembling the control achieved using conventional chemical routes.Moreover, such control is achieved without hindering the exceptional thickness accuracy (9 nm) reported in plasma processes permitting an accurate control for utf-hydrogels below 150 nm.
From a structural point of view, increasing the f PE leads to the higher density of cross-links but also to longer cross-link segments due to favored homopolymerization.Therefore, swelling ratio is increased while viscoelastic modulus obtained is reduced.On the other hand, lower f PE diminishes the number of crosslinks and restricts the cross-linker growth.Such results have been confirmed by quartz-crystal microbalance with dissipation (QCM-D) coupled with ellipsometric spectroscopy (SE) measurements.Both coupled techniques have been put in the spotlight as promising alternatives to the conventional characterization methods of soft ultrathin film materials.More specifically, they provide insights into the presence of a hydrated layer at the interface, which can play a crucial role in interface-related technologies such as the design of ultrathin film hydrogels with antibiofouling or anti-microbial properties.We do believe that such unprecedented results will encourage other researchers to reconsider plasma-assisted techniques for the design of ultrathin film hydrogels with customized properties.

Figure 1 .
Figure 1.Experimental setup used for the preparation of the samples.

Figure 3 .
Figure 3. High-resolution XPS results.a) High-resolution spectra of the C 1s region of the PMAA samples synthesised at 1 and 3 kHz and b) highresolution spectra of the C 1s region of the MAA:EGDMA-1 and MAA:EGDMA-3 samples.Atomic distribution of the C 1s carbon of the MAA:EGDMA-1 and MAA:EGDMA-3 obtained at c) surface and d) bulk.

Figure 4 .
Figure 4. a) Thickness values for different number of runs and f PE determinated by means of AFM and SE techniques.Linear regime is highlighted in the inset.b) Parity plot comparing the correlation between AFM and SE techniques.

Figure 5 .
Figure 5. AFM topographical studies of the most representative samples.a) 3D height topographic images, b) 2D height image, and c) Phase images.

Figure 6 .
Figure 6.a) Scheme of the cell used for coupling QCM-D and SE techniques.b) QCM-D raw data obtained for ut-MAA:EGDMA-3 dynamic experiments.Black line shows the fitting obtained.The advantages of using different media (following the protocol described in the methods sections) are highlighted.

Figure 7 .
Figure 7. Quantitative results obtained for ut-MAA:EGDMA-1 and ut-MAA:EGDMA-3 by means of coupled QCM-D/SE techniques.a) Dynamic thickness increment obtained under different media, and b) total thickness analyses obtained for de-ionised (DI) water.The hydrated layer has been also included.c) Raw data obtained for ut-MAA:EGDMA-3 sample following the D 2 O-H 2 O-Air protocol.d) Total thickness analyses obtained for wet samples exposed to air atmosphere.Similarly, the hydrated layer has been also included.e) Swelling ratio (in DI water medium) determined by SE and QCM-D, and f) viscoelastic modulus.For the first case, the viscoelastic modulus is maintained for both DI water and DPBS.

Table 1 .
Most relevant topographical parameters obtained from AFM 3D reconstruction measurements.