Direct Observation of the Interplay of Catechol Binding and Polymer Hydrophobicity in a Mussel-Inspired Elastomeric Adhesive.

Marine organisms such as mussels have mastered the challenges in underwater adhesion by incorporating post-translationally modified amino acids like l-3,4-dihydroxyphenylalanine (DOPA) in adhesive proteins. Here we designed a catechol containing elastomer adhesive to identify the role of catechol in interfacial adhesion in both dry and wet conditions. To decouple the adhesive contribution of catechol to the overall adhesion, the elastomer was designed to be cross-linked through [2 + 2] photo-cycloaddition of coumarin. The elastomer with catechol moieties displayed a higher adhesion strength than the catechol-protected elastomer. The contact interface was probed using interface-sensitive sum frequency generation spectroscopy to explore the question of whether catechol can displace water and bond with hydrophilic surfaces. The spectroscopy measurements reveal that the maximum binding energy of the catechol and protected-catechol elastomers to sapphire substrate is 7.0 ± 0.1 kJ/(mole of surface O-H), which is equivalent to 0.10 J/m2. The higher dry and wet adhesion observed in the macroscopic adhesion measurements for the catechol containing elastomer originates from multiple hydrogen bonds of the catechol dihydroxy groups to the surface. In addition, our results show that catechol by itself does not remove the confined interstitial water. In these elastomers, it is the hydrophobic groups that help in partially removing interstitial water. The observation of the synergy between catechol binding and hydrophobicity in enabling the mussel-inspired soft adhesive elastomer to stick underwater provides a framework for designing materials for applications in tissue adhesion and moist-skin wearable electronics.

Introduction

Mussels, caddisflies, sandcastle worms, barnacles, and many other species use biological adhesives to form interfacial bonds to various substrates in the presence of water,[1−4] and among these examples, the mussel holdfast is one of the most well-studied biological adhesive system.[5−7] Mussels secrete a series of adhesive proteins which solidify to form the byssus thread—the holdfast of mussels.[8] The byssus thread comprises of more than 15 different mussel foot proteins (Mfp),[9−12] and among them, a set of low molar mass (<11 kDa) foot proteins (Mfp -3F, -3S, and -5) are found at the adhesive interface.[4,13] These low molecular weight proteins are rich in l-3,4-dihydroxyphenylalanine (DOPA, more than 20 mol %).[7,13−16] The DOPA-rich proteins adhere strongly to various surfaces (adhesion energy ≈ 3–15 mJ/m2), and it has been suggested that the strong adhesion of DOPA is due to its Janus-like nature.[17−22] Depending on the surface, the catechol binding group of DOPA is proposed to either interact through hydrogen bonding (to polar and metal oxide surfaces), coordination bonds (to metal oxides and metallic surfaces), or hydrophobic interactions (to nonpolar surfaces).[17,19,20,23−28] Single molecule atomic force microscopy (AFM) experiments[23,29−32] have shown that catechol interacts with TiO2 surfaces with a binding energy of 22 kcal/mol, which is similar to the bond energy of a covalent single bond.[23] When the surfaces of main group metal oxides such as SiO2 and Al2O3 were studied, lower adhesion strengths were found, and it has been suggested that they interact exclusively through hydrogen bonding in ambient conditions.[30] Although AFM studies provide detailed information on binding energies, there are unresolved inconsistencies in the values of DOPA-TiO2 pull-off forces in the literature.[23,29,32]

Inspired by the strong interactions of catechol to various surfaces, a substantial amount of adhesive polymers containing catechol have been synthesized and examined for their ability to improve adhesion to surfaces.[33] These mussel-inspired adhesives have been shown to adhere to various surfaces such as metal oxides and biological tissues.[34−39] But only in a few of these studies have adhesive joints been constructed in the presence of water.[34,40−42] Moreover, most of these studies in the DOPA-polymer literature utilize lapshear strength measurements to demonstrate the importance of catechol in increasing adhesion strength.[41,43,44] The lapshear test cannot accurately distinguish the contribution of interfacial interactions and cohesive strength to adhesion. Catechol can also play an important role in increasing the cohesive strength of mussel-inspired polymers.[45] For example, catechol can be oxidized and cross-linked to increase the cohesive strength of the material, thereby making it difficult to distinguish the role of catechol in interfacial adhesion.[39,46,47] Also, these adhesion studies lack control samples of similar bulk properties[48] and without catechol groups, which is necessary to understand the role of catechol in increasing underwater interfacial adhesion. Herein, taking these factors into consideration and inspired by our recent design of a photocurable, mussel-inspired adhesive with remarkable underwater lapshear adhesion strength (0.65 ± 0.09 MPa),[49] we have designed an experimental approach to test the role of catechol in underwater adhesion.

To conclusively identify the role of catechol in dry and wet interfacial adhesion, we have designed two polymers containing either catechol (deprotected) or protected-catechol (protected) pendant groups. To decouple the interfacial and cohesive contributions to the overall adhesion, the polymers were designed with pendant coumarin units, which undergo a [2 + 2] photocycloaddition to provide catechol independent cross-linking.[49−52] Both these polymers have similar glass transition temperatures and viscoelastic properties, thereby minimizing the influence of bulk physical property variations on adhesion measurements. We have chosen protected elastomer as our control instead of varying the molar ratio of the catechol groups because the later would vary the bulk physical properties and make the interpretation of adhesion results difficult.

Johnson Kendall Roberts (JKR) adhesion geometry[53] is used to measure force (pull-off force) and work required to separate two substrates by bringing a glass lens in contact with the elastomeric adhesives in dry and wet conditions. The JKR adhesion test in conjunction with low pull-off velocities helps in reducing the contribution from the energy spent in deforming the bulk material from the total work done in separating the two substrates.[54] Because the adhesive is cross-linked and soft (G′ ≈ 7 kPa), we obtain good molecular contact between the adhesive and the glass probe, thus reducing the effect of roughness. In addition to adhesion strength measurements, we have used interface-sensitive sum frequency generation (SFG) spectroscopy to directly probe the contact interface in dry and wet conditions. SFG is a second-order nonlinear optical spectroscopic technique that provides direct information on the concentration and orientation of the interfacial molecules.[55,56] SFG probes a depth of a few nanometers compared to hundreds of nanometers probed by the infrared or Raman spectroscopic technique. We have combined the interface-sensitivity of SFG with an experimental design of protected or deprotected elastomers placed in contact with a sapphire prism in either dry conditions or in the presence of water. By observing the shift of the sapphire O–H peak after bringing in contact with the protected or deprotected elastomers, we directly measured the interactions of catechol with the surface O–H groups and the role it plays in hydrogen bonding with a hydrophilic surface.[57,58] The current work provides the framework for the design of effective underwater adhesive elastomers based on the information provided from SFG and adhesion measurements. Also, this study gives critical insight into the interplay of catechol in hydrogen bonding and polymer hydrophobic groups in removing interstitial water as a means of providing effective underwater adhesion.

Results and Discussion

Elastomer Design

To create an elastomer, a statistical copolyester was synthesized using a N,N′-diisopropylcarbodiimide (DIC) assisted polyesterification reaction of three N-functionalized diols and sebacic acid (Scheme S1 and Figure ).[59] The first diol with pendant aliphatic hydrocarbons (S) provides hydrophobicity and lowers the glass transition temperature (Tg ≈ – 45 °C), which makes it easier to spread at room temperature.[52] The second coumarin diol (C) undergoes [2 + 2] cycloaddition when exposed to UV light of wavelength (λ) ∼350 nm, which converts the viscous polymer to an elastomer.[50,51] The third component is an acetonide protected-catechol diol (Dpr), which is deprotected under acidic conditions to provide catechol and is expected to increase underwater adhesion upon deprotection.[60] The feed ratio of monomers to form the polymer was chosen to be S:C:Dpr 65:5:30 (mol %), and from 1H NMR (Figure S1A) the actual composition was calculated to be 63:5:32 and the molar mass as detected from GPC (Mn,GPC) of the protected polymer was 11.3 kDa (dispersity = 1.6). The polymer with catechol groups (deprotected) was obtained by the reaction of protected polymer (500 mg) with trifluoroacetic acid (5.0 mL) in methylene chloride (10 mL) for 2 h at room temperature under N2 (details provided in the Supporting Information). The disappearance of protons from the 1,2-acetonide group (δ = 1.63 ppm, -C(CH3)2) in the 1H NMR spectrum (Figure S1B) and the disappearance of acetonide C–H bend signal at 1498 cm–1 along with the appearance of a broad absorption band at 3100–3650 cm–1 assigned to O–H groups of catechol in the FT-IR spectrum (Figure S2) indicate the successful deprotection of acetonide groups. Choosing the protected elastomer as the control avoids the problem of comparing polymers with a different molar mass and/or Tg (Table S1). From the rheological measurements of cross-linked protected and deprotected polymers, the frequency responses of storage () and loss () moduli were quantified (Figure S3). The rheological response of both the elastomers after cross-linking were similar, indicating the appreciable similarity of bulk behavior of the protected and deprotected elastomers and minimal interaction between catechol units themselves. The usage of coumarin for cross-linking also gives similar modulus for the elastomers (discussed in detail later), which is a prerequisite for comparing the adhesive properties since the adhesion values could be influenced by both interfacial and bulk mechanical properties.

Figure 1

Chemical structures of the polymers designed for the adhesion and spectroscopic measurements.

Adhesion Measurement

The adhesion of a hemispherical glass lens to protected and deprotected elastomers coated on oxidized polydimethylsiloxane (PDMS) elastomer was tested (PDMS provides an elastic backing; see Supporting Information for details). The polymers coated over PDMS elastomers were then cross-linked using UV-A irradiation (λ = 350–420 nm, intensity on the substrate = 50 mW/cm2) for 10 min to form an elastomer. To ensure a uniform coating of elastomer on PDMS sheets, the elastomer-coated PDMS sheets were analyzed by fluorescence microscopy. The elastomer-coated films fluoresced under a DAPI filter in contrast to bare PDMS, confirming the uniform coverage of the elastomer on the PDMS sheets (Figure A).

Figure 2

Characterization of the substrates used for adhesion force measurements. (A) (L–R) Fluorescence microscopy images of PDMS sheet, protected and deprotected elastomers coated PDMS sheets before and after periodate treatment. The images shown are under 5× magnification and the scale bars (white box at bottom right) in the images correspond to 400 μm. (B) The UV–vis absorption spectra of the polymer films on quartz substrates before and after cross-linking. The presence of UV absorption peak λmax ≈ 280 nm indicates the presence of unoxidized catechol after cross-linking. (C) Single bounce ATR/FT-IR spectra of PDMS sheet and both elastomers coated on PDMS sheets recorded by exposing the coated side to the IR beam. The spectra of the elastomer-coated sheets match those of the respective polymers and are different from the FT-IR spectra of the PDMS sheet.

Since catechol is prone to oxidation reactions, it is important to show that the catechol moieties are intact after exposure to the cross-linking conditions.[61] The chemical stabilities of the protected and deprotected elastomers on UV-A exposure were analyzed by UV–vis and ATR/FT-IR absorption spectroscopies after the polymer films were exposed to UV-A irradiation. The UV–vis spectra of elastomers show that the catechol absorption peak (π → π*, λmax = 280 nm) remains intact after cross-linking (Figure B), indicating the stability of catechol upon UV-A exposure (upon oxidation, red shift is expected).[62] The absorption bands (λmax = 310, 320, and 340 nm) corresponding to coumarinyl groups disappear with UV-A exposure confirming the completion of cross-linking reactions.[51] Single bounce ATR/FT-IR absorption spectra of elastomers were also collected by placing the elastomer-coated side of the substrates toward the IR beam. The characteristic absorption peaks in Figure C match the IR absorption signature of corresponding polymers (Figure S2). In the protected elastomer spectrum, it is seen that the acetonide protection group (C–H bend, 1498 cm–1) remains intact after cross-linking reactions. The results from characterization of the elastomers provided confidence for the following experiments which examine the role of catechol in dry and wet adhesion.

Figure A shows the schematic diagram of the in-house experimental set up used for the adhesion force measurements using a JKR geometry. During the adhesion strength measurements, a hemispherical glass lens is brought in contact with the elastomers in the absence (dry) and presence (wet) of water to a preload of −1 mN. The lens is then retracted back after letting it equilibrate to measure the maximum force for separating the contact, which is recorded as the pull-off force (Figure B).[53]

Figure 3

Adhesion strength measurements. (A) Schematic diagram of the home-built set up showing the JKR geometry used for adhesion force measurements. (B) Representative force runs of the protected and deprotected elastomers in dry and wet environments showing the force and work done to separate the two surfaces. (C) Pull-off forces (left axis), work of adhesion (right axis), and (D) work done to separate the protected and deprotected elastomers in dry and wet environments from the glass surface when loaded to −1 mN force followed by steady hold for 3 min and unloaded at a rate of 0.4 μm/s. The data represented here are presented as mean ± standard deviation (SD), and “*” represents the statistical significance among the samples using a Tukey mean comparison test (p < 0.05). Error bars (SD) are evaluated using at least three measurements for each condition. Higher forces are required to separate the contact of the glass lens with the deprotected elastomer than the contact of the protected elastomer in dry and wet environments. There is a reduction in work of adhesion for both the elastomers in the presence of water.

Figure C shows the pull-off force (left axis) and work of adhesion (right axis) calculated using the JKR model for the protected and deprotected elastomers in dry and wet conditions. Under dry conditions, the work of adhesion of the deprotected elastomer (1.80 ± 0.21 N/m) is significantly higher than the protected elastomer (0.51 ± 0.01 N/m). During the experiments, the loading and unloading cycles were monitored using an optical microscope from which it was observed that the mode of failure for protected and deprotected elastomers was adhesive and cohesive for the dry measurements, respectively (Figure S5). Hence it is possible that the interfacial adhesion strength of the deprotected elastomer could be even higher than what is described here. Since the protected and deprotected elastomers have a similar mechanical response, it is safe to conclude that the differences are due to interfacial interaction of the elastomers.

In wet conditions, both protected (0.15 ± 0.03 N/m) and deprotected (0.38 ± 0.05 N/m) elastomers showed appreciable adhesion in contrast to the PDMS sheet (∼0 N/m) (Figure S4C). This indicates that in these experiments PDMS is not contacting the glass substrate. However, in wet conditions both the elastomers have a significant decrease in adhesion strength as compared to their corresponding dry values. Catechol containing polymers have been shown to have enhanced adhesion to mica surfaces under acidic conditions since they are known to oxidize readily at higher pH.[19] To investigate the pH dependence of adhesion, we measured the adhesion strength of our elastomers at pH 3, 6.5, and 9 (Figure S4A). For deprotected elastomer, the adhesion strength is insensitive to pH changes (based on statistical comparison), whereas the protected elastomer showed a reduction in adhesion (0.06 ± 0.04 N/m) at pH 9. At pH 9, the glass surface becomes more negative and can lead to electrostatic repulsion between hydrophobic polymers and the glass surface. This repulsion of the similar charge densities might be the reason for reduced underwater adhesion of protected elastomer at higher pH.[63] However, we did not observe any statistical difference in the pull-off forces for the deprotected elastomer with the increase in pH. We speculate that at a higher pH, the hydroxyl group of catechol groups deprotonates to form either coordination bonds with silicon or quinone, which then act as an efficient hydrogen bond acceptor to the surface hydroxyl groups and maintain the adhesion strength.[20,64] After wet measurements at pH 9 were performed, both the elastomers were submerged in an aqueous solution of 10 mM sodium periodate (NaIO4) for 2 h. Periodate (IO4–) treatment can cause oxidation of catechol to form quinone and its tautomer and decrease adhesion of the deprotected elastomer.[19,43] We observed that the deprotected elastomer does not stick after periodate treatment, but the protected elastomer still retains its adhesion compared to pH 9 (Figure S4A). It is possible that the deprotected elastomer undergoes extensive oxidation during periodate treatment, resulting in the loss of adhesion, while the protection prevents this oxidation reaction. But, it was intriguing that upon oxidation, the deprotected elastomer did not retain adhesion comparable to the protected elastomer. We found that the periodate treatment not only caused the catechol oxidization reaction but also ruptured the elastomer film and exposed the oxidized PDMS sheet. This introduces roughness which was evident in fluorescence microscopy images (Figure A). Both roughness and exposure of PDMS lead to loss of underwater adhesion.

A similar trend for the work of adhesion was observed when the work done was calculated by integrating the area under the unloading curve of force as a function of displacement (Figure D). The time axis of the graph in Figure B can be converted into displacement by multiplying time with the unloading rate of 0.4 μm/s. This work comprises the energy required to break the interfacial bonds and also the elastic/viscous work done in stretching the elastomer.[54] From Figure D, it can be observed that more work is required to separate the contact for the deprotected elastomer than the protected elastomer. To further investigate the detailed adhesion mechanisms of these elastomers, we performed SFG spectroscopy experiments of the contact interface, which are described in the following section.

SFG Spectra of the Elastomer-Substrate Contact Interface

For SFG experiments, we have used total internal reflection geometry to probe the contact interface in both dry and wet conditions. The experimental details are provided in the methods section. The generation of SFG signals requires a breakdown in the symmetry of dipole orientation, and this happens only for the interfacial molecules at the contact interface. This interface selectivity allows us to interpret the presence or absence of molecules at the contact interface. For example, if the contact is dry, we should not observe water bands between the 3100–3600 cm–1 (or 2300–2700 cm–1 for D2O) region. If water is present at the contact interface, then the location of the water peak indicates the nature of hydrogen bonding of the confined water. Since the bulk is centrosymmetric, the SFG signals are interface-specific and are not swamped by the signals from the bulk elastomers.

A sapphire (Al2O3) prism-like glass (Figure S6) with surface hydroxyl functional groups (O–H) was brought in contact with the protected and deprotected elastomer-coated PDMS lenses in dry and wet environments (Figure A). We expect SFG signals to be generated from only a few nanometer-thick interfacial layer between the elastomer and the sapphire substrate. In addition, the shift in the surface sapphire O–H peak can be used to calculate the strength of the acid–base interactions (hydrogen bonding is a subset of acid–base interactions).[57] For example, in the previously reported study, the ester groups in poly(methyl methacrylate) showed stronger interaction (O–H peak is shifted to 3580 cm–1 compared to free O–H peak at 3720 cm–1) than functional groups in polystyrene (3645 cm–1) when in contact with a sapphire substrate.[58] Similarly, since the interfacial adhesion of the deprotected elastomer is higher, we speculated that there might be a larger shift for the deprotected elastomer as compared to protected elastomer, which would indicate a stronger bond between the hydroxyl groups of catechol and the sapphire.

Figure 4

SFG spectra of elastomer–substrate contact interface. (A) Schematic of total internal reflection geometry used to probe the mechanical contact of elastomer-coated PDMS lens with a sapphire prism in dry and wet (D2O) environments. (B) Top and bottom panels show SFG spectra (SSP polarization) of protected and deprotected elastomers in contact with a sapphire substrate in dry and wet (D2O) conditions, respectively. Spectra were collected in two regions 2700–3200 cm–1 and 3100–3800 cm–1 separately and plotted together to show the differences. These spectra show the hydrocarbon signature from the elastomer and sapphire hydroxyl groups. (C) SFG spectra (SSP polarization) in the O–D stretching region of the protected and deprotected elastomers in contact with a sapphire substrate under wet (D2O) conditions. Features in the region from 2200–2800 cm–1 indicate the presence of D2O in the contact region.

Figures B, C, and S7 show the SFG spectra in both SSP (Figure B and C) and PPP (Figure S7) polarizations (the polarization combination is for three beams: SFG, visible, and IR, respectively, where S and P are components of electric field perpendicular and parallel to the plane of incidence). Depending upon the polarization combination, information about the molecular group orientation can be inferred.[65]Figure B shows the SSP spectra comparison of both elastomers in dry (top panel) and wet (bottom panel) environments in the hydrocarbon region (C–H stretch, 2800–3100 cm–1) and the sapphire region (O–H stretch, 3200–3800 cm–1). The wet measurements were done in D2O to avoid the overlap of signal from O–H stretches of water (H2O) and sapphire. In both elastomers, the peak at ∼ 2960 cm–1 represents the vibration of the aliphatic side chain and polymer backbone C–H groups. The contact of PDMS with sapphire substrates results in a very different SFG spectrum, and this again confirms that the protected or deprotected elastomer layers are intact upon contact with the sapphire substrate.[66,67] The sapphire O–H peak positions for both dry contacts of the protected and deprotected elastomers are very similar. On the basis of three independent measurements, using three different lenses coated with elastomers from two different molecular polymers, the averaged peak position of sapphire O–H region was 3552 ± 6 cm–1 for protected elastomer and 3557 ± 22 cm–1 for the deprotected elastomer. These peak positions were obtained by fitting the data using a Lorentzian equation (details in Supporting Information). Surprisingly, the similar shifts in sapphire O–H peak indicate a similar strength of acid–base interactions for both elastomers.

The shift in sapphire O–H was further scrutinized with a first moment analysis of the peak distribution (details in Supporting Information). The average sapphire O–H shift was found to be 3557 ± 10 cm–1 and 3557 ± 12 cm–1 for protected and deprotected elastomers in dry contact, respectively using the first moment analysis. This confirms the identical acid–base interaction strength of both protected and deprotected elastomers. On the basis of this shift of sapphire free O–H (3707 cm–1, obtained experimentally) and using the Badger–Bauer equation (energy of interaction, ΔH = m × Δν + C, where m = 1.09 × 10–2 kcal/mol cm, C = 0.03 ± 0.01 kcal/mol for sapphire, and Δν is the shift of the O–H peak), we estimate that the interaction corresponds to an adhesion energy of 7.0 ± 0.1 kJ/(mole of O–H groups).[57] A sapphire surface typically has around nine O–H groups per nm2, and hence macroscopically, this interaction can contribute to a maximum interfacial threshold energy (G0) of 0.10 J/m2 for both elastomers assuming all the surface O–H groups are participating equally in this interaction.[54,58]

The similarity in the shift of the O–H peak for protected and deprotected elastomers was unexpected. We were anticipating higher G0 for the deprotected elastomer than the protected elastomer from the Badger–Bauer equation-based calculations in accordance with the observations from the adhesion measurements (Figure C). To investigate more, we measured the interaction of catechol-d2 in CHCl3-d (0.07 M) adsorbed on sapphire substrate using SFG. Interestingly, the peak of sapphire O–H interacting with catechol was observed at ∼3593 ± 2 cm–1 (Figure S8). The bimodal peak observed in Figure S8 is for catechol-d2 and CHCl3-d interacting with the sapphire O–H groups. The position of CHCl3-d peak is at the similar location with or without adding catechol-d2. This indicates that the presence of catechol-d does not alter the interaction strength of CHCl3-d groups. The shift in the position of the O–H peak as result of the interaction of catechol to sapphire O–H was intermediate to acetone (∼3610 cm–1) and pyridine (∼3575 cm–1) interactions.[57] Also, the lower O–H shift in the case of catechol than the elastomers suggest that the other polar groups present in the elastomer may interact more strongly than the catechol.

For the wet contact (Figure B, bottom panel), there is a notable decrease in the intensity of sapphire O–H region as compared to hydrocarbon signature. To understand this change in intensity, we need to also compare the changes in the peak in the O–D stretching region (Figure C). Both protected and deprotected elastomers showed two distributions of peaks: one corresponding to O–D stretching in “liquid-like” water at ∼2500 cm–1 and another corresponding to O–D stretching next to hydrophobic interface at ∼2650 cm–1.[63] The higher the number of hydrogen bonds per water molecule, the O–D or the O–H stretch mode moves to lower wavenumbers. So, the “liquid-like” water peak is ∼2500 cm–1 (∼3400 cm–1 for H2O) compared to more strongly hydrogen bonded peak for ice, which is ∼2400 cm–1 (∼3200 cm–1 for H2O). If there are fewer numbers of hydrogen bonds compared to liquid-water, the peaks are shifted to ∼2600 cm–1 (∼3500 cm–1 for H2O), whereas the non-hydrogen bonded O–D peak appears at ∼2700 cm–1 (∼3700 cm–1 for H2O). The decrease in the intensity of sapphire O–H region (∼ 3550 cm–1) as compared to hydrocarbon signature along with the distribution of O–D stretch next to the hydrophobic interface (∼2650 cm–1) could be due to exchange of the sapphire proton to deuterium (O–H → O–D) after exposing the sapphire to deuterated water. The exchange can cause decrease in sapphire O–H intensity and is expected to show peaks in the 2600–2750 cm–1 region. The second reason could be the presence of weakly hydrogen bonded water between the elastomer and sapphire interface.[68] This weakly hydrogen bonded water layer may not decrease adhesion. Zhou et al. also observed similar water structure (∼2630–2700 cm–1) at the polyurethane-sapphire interface after exposing the sample to low humidity and the presence of this water layer did not completely disrupt the polyurethane-sapphire interactions.[69] In both these scenarios, the exchange of surface O–H to O–D or presence of weakly hydrogen bonded water (∼2650 cm–1) may not disrupt adhesion of the protected or deprotected elastomers when contacted with the sapphire substrate in the presence of water. The important observation here is the presence of the O–D peak near 2500 cm–1 corresponding to “liquid-like” confined water, which can disrupt underwater adhesion.

Interestingly, in the hydrocarbon region (Figure B, bottom panel) an aromatic =C–H signature at ∼3030 cm–1 was observed only for the deprotected elastomer in wet contact. This signature observed in the SFG spectra of the deprotected elastomer exclusively in wet conditions indicates the presence of catechol at the underwater contact interface.[70,71] This peak also overlaps with the features from other unsaturated hydrocarbons in the polymer side chains. The lack of such a signature for the dry and underwater contact of protected elastomer and dry contact of deprotected elastomer confirms the presence of catechol groups at the underwater contact interface.

Now relating these spectroscopic observations to adhesion measurements, we can understand the precise role of catechol in increasing adhesion. The similar acid–base interactions for both elastomers negate the possibility of hydroxyl groups of catechol forming stronger acid–base bonds with hydrophilic substrates compared to other polar groups present in the polymer. This leaves the following three possible explanations for the increased adhesion of the deprotected elastomer. First, the differences observed in adhesion strength measurements could originate from the differences in the deformation during loading. To scrutinize such possibility, we measured the effective modulus of the contacting surfaces. The load dependent contact deformation of the elastomer-lens during the approach and retraction was captured by observing the changes in contact area as a function of force using an optical microscope over the set up shown in Figure A. This data along with a JKR model was used to measure the effective modulus (K) of contacting surfaces in dry condition,[53] which were calculated to be ∼1.9 and ∼1.5 MPa for the protected and deprotected elastomers coated PDMS sheets, respectively (Figure S9). The similar values of K (Figure S9) and rheological properties (Figure S3) imply that the differences in adhesion measurements arise from the interfacial adhesion and not because of differences in the bulk properties. Since the elastomers display adhesion hysteresis (Figure S9), the adhesion strength reported here may also contain the energy spent in stretching interfacial chains and potentially a contribution from the energy dissipated in stretching the bulk polymeric chains during the unloading cycle. Therefore, the work of adhesion measured using the JKR geometry during pull-off is not equal to but is proportional to the interfacial threshold strength (G0) of the protected or deprotected elastomers. Since the protected and deprotected elastomers have similar moduli and bulk properties, we expect that the ratio of G0 for these two elastomers would be similar to the corresponding ratio of their work of adhesion.[54]

The second possibility is that the catechol moieties can form multiple hydrogen bonds (multimodal acid–base interactions),[72] and breaking multiple bonds simultaneously would require higher energy as proposed by the single molecule AFM and surface force apparatus measurements of catechol containing molecules.[20,31] Our experimental evidence substantiates this model. By comparing the sapphire O–H peak shifts in the adsorption of catechol (Figure S8) and the elastomer-sapphire contacts (Figure B), the interaction of catechol hydroxyl groups (3593 ± 2 cm–1) was observed to be weaker than the interaction of both protected (3552 ± 6 cm–1) and deprotected (3557 ± 12 cm–1) elastomers. In the SFG measurements, the maximum sapphire O–H peak shift corresponds to the strongest monomodal interaction of functional groups with sapphire. Therefore, we can conclude that the interaction strength of individual hydroxyl groups of catechol is lower than some of the functional groups in the elastomers. However, with two adjacent hydroxyl groups, catechol can form multiple weaker monomodal interactions in a localized area to constitute multimodal interactions, and as a result, the elastomer with catechol groups show a higher adhesion strength than the protected elastomer.

The third possibility is that the presence of catechol in the deprotected elastomer increases the number of polar groups (∼60 mol % increase). This increases the number of potential hydrogen bonds that can be formed by the deprotected elastomer compared to the protected elastomer. Since the shift in the sapphire O–H peak of both the elastomers is very similar, the level of polar interactions with the substrate is identical. Hence, we conclude that the localized multiple hydrogen bonding by dihydroxy groups is primarily responsible for the increase in the interfacial adhesion in dry conditions. The probability of other interactions such as metal-coordination is unlikely, as it has been shown that catechol interacts exclusively through hydrogen bonding with SiO2 and Al2O3 surfaces.[31]

In the case of wet contact of both the elastomers, “liquid-like” water is present (Figure C, bottom panel, peak ∼2500 cm–1). Subtle changes in the water structure have been shown to influence the interfacial phenomena such as adhesion and friction.[73] The statistically similar O−D spectral signatures in the wet contact of both the elastomers eliminate the possibility of water structure causing the differences in adhesion strength. The presence of “liquid-like” water and the significant underwater adhesion indicate that the contact is patchy. There are certain regions where the elastomers are in direct contact with the sapphire substrate and other regions where there is “liquid-like” water trapped between the elastomers and the sapphire substrate.[57,74] This patchy contact interface explains the drop in underwater adhesion compared to the dry adhesion. Besides, catechol itself does not play an important role in removing interfacial water next to hydrophilic interface, which is consistent with the observations by Kirpat et al.[75] The observed catechol signature in underwater contact of the deprotected elastomer (Figure B, bottom panel, peak ∼3030 cm–1) is from the population of catechol that is interacting with water.

The nature of the underwater contact depends on the surface energy of the materials.[74] For example, two hydrophobic surfaces make true molecular contact after removing interfacial water, and the contact between two hydrophilic surfaces retains a thin film of water at the interface,[73,76] whereas a hydrophobic material makes patchy contact with hydrophilic surfaces.[66] In our case, both elastomers are hydrophobic (polymer water contact angles >95°, Table S1) and are in contact with a hydrophilic surface, and hence the contact area is expected to be patchy. Perhaps a patchy contact could also form as the draining of interstitial water requires more time. If this was the case, increasing the contact time should increase the dry molecular contact and thus lead to higher adhesion.[66]Figure S4B shows the work of adhesion at contact equilibration times of 0.5, 3, and 15 min. The underwater adhesion values remain lower than those measured in dry contact, indicating that the kinetics of drainage is not playing an important role in underwater adhesion. The deprotected elastomer showed higher underwater adhesion as compared to the protected elastomer. Since the percentage reduction in the adhesion between dry and wet contacts for both protected (∼72%) and deprotected (∼79%) elastomers are similar, we infer that the fraction area of patchy water contact is similar for both elastomers.

From the analysis of adhesion measurements and SFG spectra, we propose an overall model for the underwater adhesion for these elastomers (Figure ). Regardless of the presence of catechol, both elastomers form patchy contact with hydrophilic substrates underwater. The wet patches which contain water (blue) between the polymer and substrate constrain the adhesion. The dry patches (yellow) are true contacts between the polymer and substrate which provide underwater adhesion.[66,74] We anticipate that the patches are smaller than the resolution of the optical probes (less than microns). Both the protected and deprotected elastomers succeed in contacting the substrate underwater. However, the interfacial strength of the dry patch of deprotected elastomer is higher due to the localized multimodal interactions of O–H groups in catechol with the sapphire substrate, resulting in higher underwater interfacial adhesion than the protected elastomer. We have reported recently that a hydrophilic adhesive with catechol does not adhere underwater compared to a relatively hydrophobic adhesive without catechol groups.[49] Therefore, for effective underwater adhesion, initially the polymer should make interfacial contact with the surface and remove bound water, which in this case is achieved by a conformable hydrophobic adhesive. Also, polar functional groups with strong interfacial interactions are essential to display significant adhesion underwater. If we use hydrophobic PDMS in contact with hydrophilic substrates, we observe patchy contact. However, the wet adhesion is very weak for PDMS (Figure S4C) due to the absence of strong interfacial polar interactions present in mussel-inspired catechol polymers. Additionally, the hydrophobic functional groups in Mfp[19] and mussel-inspired polymers[77] have also been shown to improve adhesion by shielding catechol groups from oxidation reactions.

Figure 5

Proposed model for the adhesion mechanism in wet condition (pH = 6.5) for the deprotected elastomer. The diagram on the left side is a representative underwater contact of the deprotected elastomer with hydrophilic substrate. In the dry patch (top right), the elastomer is in contact with sapphire and can interact with substrate through acid–base interactions. Catechol makes multimodal hydrogen bonds with the substrate. In the wet patch (bottom right), water interferes the interaction of the elastomer with the substrate.

Conclusion

We have used polymers with well-defined chemistry, JKR geometry for adhesion measurements, and interface-sensitive SFG spectroscopy to understand the role of catechol in underwater adhesion of mussel-inspired polymers. Having both the elastomers of similar bulk mechanical properties, but a difference in only the surface-active catechol groups being protected and deprotected, helped to identify the importance of the dihydroxyl groups and the hydrophobic groups in adhesion. The dry adhesion of deprotected elastomer is higher than the protected elastomer. Direct probing of the contact interface using SFG spectroscopy reveals that the strength of the acid–base interactions of polar groups with surface O–H groups on the sapphire substrate is very similar for both protected and deprotected elastomers. The stronger adhesion strength obtained from macroscopic measurements can be explained by the dihydroxy chemistry of catechol despite similar interaction strength as observed in SFG spectra. The energy required to break multiple bonds concurrently is higher than the sequential breaking of monodentate bonds of similar strength, thus increasing the adhesion strength of deprotected elastomer compared to protected elastomer.

In wet environment, the deprotected elastomer has a higher adhesion compared to protected elastomer. The SFG experiments reveal that the presence of “liquid-like” water in the contact interface for both the polymers. The SFG and adhesion data combined show the interplay of hydrophobicity and catechol binding to enhance adhesion. Both elastomers have a patchy contact underwater. The protected and deprotected elastomers have a similar water contact angle (similar hydrophobicity) and similar decrease in wet adhesion compared to dry adhesion indicating the assistance of hydrophobicity in achieving similar patchy contact underwater. The formation of patchy contact between a hydrophobic and hydrophilic interface underwater is consistent with the previous studies.[66,78] The higher adhesion strength of the deprotected elastomer is due to multimodal bonding of catechol with the hydroxylated surface. Hence, the interplay of hydrophobicity to remove water from the interface and the role of catechol in forming multiple hydrogen bonds are proposed from the combination of SFG spectroscopy and adhesion measurements in this study. Although the synergistic effect of lysine to displace cations and water from the surface for enhanced interfacial binding of catechol is also essential to understand the adhesion of Mfp, our current focus was to study the importance of hydrophobicity in improving underwater adhesion of mussel-inspired polymers.[79,80] In the future, the critical understanding of the interplay of hydrophobicity and mussel-inspired dihydroxy chemistry in synthetic adhesives can be put to use in designing adhesives for applications in tissue adhesion where the presence of moisture or physiological fluids cannot be avoided.