Deciphering How the Viscoelastic Properties of Mussel-InspiredMetal-Coordinate Transiently Cross-Linked Gels Dictate Their TackBehaviorErica Lai,† Bavand Keshavarz,‡ and Niels Holten-Andersen*,†

†Department of Materials Science and Engineering, ‡Department of Mechanical Engineering, Massachusetts Institute of TechnologyCambridge, Massachusetts 02139, United States

*S Supporting Information

ABSTRACT: In recent years, researchers have incorporatedmussel-inspired metal-coordinate cross-links into various typesof gels to improve their mechanical properties, particularlytoughness and self-healing. However, not much is understoodabout how the linear mechanical properties of these gelsdictate their tack properties. In this study, we use shearrheology and tack tests to explore correlations between linearviscoelastic properties (i.e., plateau modulus, Gp, andcharacteristic relaxation time, τc) and tack behavior (i.e.,peak stress, σmax, and energy dissipation per volume, EDV) oftransiently cross-linked hydrogels comprised of histidine-functionalized 4-arm PEG coordinated with Ni2+. By usingthe Ni2+-histidine ratio and polymer wt % of the transientnetworks to control their viscoelastic properties, we demonstrate a strong dependence of σmax on Gp and τc. The observedcorrelation between network dynamics and mechanics under tensile load is in good quantitative agreement with a theoreticalframework for σmax, which includes the linear viscoelastic properties as parameters. EDV is also influenced by Gp and τc, and theEDV after reaching σmax is additionally dependent on the polymer wt %. These findings are consistent with previously proposedmolecular mechanics of reversible HisxNi

2+ cross-links and provide us with new insights into the correlations between bulkmechanics and adhesive dynamics of gels with transient metal-coordinate cross-links.

■ INTRODUCTIONResearchers seeking to design bioinspired adhesives forunderwater or medical device applications have increasinglydrawn inspiration from proteins within the mussel byssalthreads.1−3 In particular, focus has been on using modifiedamino acid ligands like 3,4-dihydroxyphenylalanine (Dopa),which is known to play an important role in surface adhesionand bulk cohesion within the mussel adhesive plaque (pleasesee Figure 1a for a general schematic). Dopa contributesstrongly to interfacial adhesion via bidentate coordinationbonds, hydrogen bonds, and other noncovalent interactions.Within the mussel adhesive plaque, diDopa covalent cross-linksand dynamic metal−ligand coordination cross-links jointlycontribute to the formation of a strong and highly dissipativematerial.4−6 It has been well established that the incorporationof mussel-inspired metal coordination cross-links into cova-lently cross-linked hydrogels can improve their bulk mechan-ical properties, particularly toughness and self-healing.7,8

However, it remains unclear how metal-coordinate-basedcross-link mechanisms affect the adhesive properties of ahydrogel.Significant research efforts have been dedicated to under-

standing how the viscoelastic properties of reversibly cross-linked hydrogels can be controlled with mussel-inspired

metal−ligand coordination chemistry. Specifically, the role ofmetal-coordinate cross-link dynamics in controlling gelmechanics in the linear mechanical regime has been extensivelystudied through the tuning of multiple physical-chemicalparameters,9−15 particularly in gels composed of four-armpolyethylene glycol (4PEG) end-functionalized with histidine(His).7,12,14,15 Histidine is an amino acid also found within themussel byssal thread, primarily in the thread core, which cancoordinate with divalent transition metals like Ni2+ to formreversible HisxNi

2+ complex cross-links (please see Figure 1b).Using the metal−ligand ratio to control the concentration of

HisxNi2+ complexes with different functionalities present

within a gel (i.e., His3Ni2+ and His2Ni

2+, or tris- and bis-coordination complexes, respectively), the impact on plateaumodulus (Gp), and characteristic relaxation time (τc) wasdetermined.14

Through spatiotemporal hierarchical coordination, Dopaand histidine coordination complexes together are hypothe-

Special Issue: Intermolecular Forces and Interfacial Science

Received: September 3, 2019Revised: October 21, 2019Published: October 21, 2019

Article

pubs.acs.org/LangmuirCite This: Langmuir 2019, 35, 15979−15984

© 2019 American Chemical Society 15979 DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

Dow

nloa

ded

via

MA

SSA

CH

USE

TT

S IN

ST O

F T

EC

HN

OL

OG

Y o

n D

ecem

ber

23, 2

019

at 2

1:15

:51

(UT

C).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

sized to contribute to the energy dissipative and self-healingbehavior of mussel adhesive threads.7,16−19 However, in theseinitial investigations of the adhesive properties of mussel-inspired metal-coordinate cross-linked hydrogels, we chose tofocus exclusively on histidine as the coordinating ligandbecause of its insensitivity to auto-oxidation. Accordingly, theadhesive tack properties of fully transient (i.e., dynamicallycross-linked) hydrogels consisting of 10 kDa 4PEG-Hiscoordinated with Ni2+ were assessed for the first time. It isimportant to note that this fully transient model gel system istechnically a viscoelastic fluid even if it has gel-like behavior onthe time scales studied. By controlling the bulk gel networkdensity and dynamics via the Ni2+-His ratio (0.25−0.5) andthe polymer weight percent (5 and 10 wt %), we were able toquantify the dependence of nonlinear properties, such as peakstress (σmax) and energy dissipation per volume (EDV)extracted from tack tests performed at different constant strainrates (ε), on the linear viscoelastic properties Gp and τc. We areable to account for the experimental measurements of σmaxwith a theoretical framework which incorporates Gp and τc.The amount of energy dissipated by the gel networks beforeand after peak stress is as expected also influenced by Gp and τc,even though it is not captured by our theoretical model.Furthermore, the EDV after reaching σmax appears to display anadditional dependence on network density, plausibly due to aneffect of network remodeling dynamics during deformation.Thus, our experimental findings are consistent with previouslyproposed molecular mechanics of the dynamic HisxNi

2+ cross-links. By deciphering the contributions of metal-coordinate

cross-linking to the tack properties of physical hydrogels, thisstudy provides insights into how metal coordinate dynamiccross-linking can be used to potentially tune the adhesiveproperties of future hydrogel materials.

■ EXPERIMENTAL SECTIONSynthesis of 4PEG-His. The full synthesis procedure for 10 kDa

four-arm histidine-modified polyethylene glycol (4PEG-His) can befound in Grindy et al. but is summarized here.14 First, 1−3 g of four-arm PEG-NH2·HCl (0.25 equiv PEG, 1.0 equiv -NH2 groups) (10kDa, from JenKem U.S.A.) is mixed with Boc-His(Trt)-OH (1.5equiv) and BOP reagent (1.5 equiv) and is dissolved in 15 mL ofdichloromethane (DCM). N,N-Diisopropylethylamine (DIPEA) (535equiv) is added, and the reaction proceeds for 2 h under N2 at roomtemperature. Then, the product is purified by precipitation in diethylether one time, methanol at −20 °C one time, and diethyl ether onceagain, and then the product is vacuum-dried. To remove theprotecting groups from the histidines, a cleavage solution of 95 mL oftrifluoroacetic acid (TFA), 2.5 mL of triisopropylsilane, and 2.5 mL ofwater is added and the reaction proceeds for 2 h at room temperature.The solvent is removed under reduced pressure, and the product isthen purified by redissolution in methanol, precipitation in ether threetimes, and vacuum-drying.

Synthesis of 4PEG-His + Ni2+ Hydrogel Samples. Theappropriate volume of 0.1 M phosphate buffer solution, pH 7.4, isadded to the polymer in a centrifuge tube, and the mixture is vortexeduntil the solution is well-mixed. This 4PEG-His solution is thenpipetted onto a piece of Parafilm and combined with a small volumeof 0.2 M solution of nickel(II) chloride hexahydrate (NiCl2·6H2O),resulting in a 10 wt % gel with the desired stoichiometric ratio of Ni2+

to histidine; in this study, the ratio range was 0.25−0.5. The gel iskneaded using the Parafilm, until it is homogeneous, and stored in a1.7 mL centrifuge tube until testing to prevent dehydration. For a 5 wt% gel, more phosphate buffer is added to dilute the 10 wt % gel.

Small Angle Oscillatory Shear (SAOS). Shear rheology isperformed on an Anton Paar MCR 302 rheometer with a 10 mm-diameter parallel plate (PP10), i.e., R0 = 5 mm, a glass substrate atroom temperature (∼22 °C), and a gap of 0.300 mm, scanning afrequency range of 100 to 0.1 rad/s with a strain amplitude of 0.1%. Asolvent trap is used during the test to minimize dehydration.

Tack Tests. All tack tests are performed on an Anton Paar MCR302 rheometer at room temperature using PP10 as a probe. Thebottom substrate is a glass slide that sits atop a support containing a45° mirror allowing bottom-up video recording of the gel sampleduring experiments. A fresh gel sample is used for each individual run.During a test, the probe is brought into contact with the gel sample at50 μm/s. The gel is trimmed to the dimensions of the probe when thegap size is 0.325 mm. Once the gel is compressed to an initialthickness, h0, of 0.300 mm, it is allowed to relax to zero normal force.While the exact contact time was not precisely controlled, it wasensured that the contact time was at least ten times longer than thecharacteristic relaxation time scale (i.e., at least 30 s), and theassumption is that the gel has reached chemical equilibrium with theprobe surface. At that point, the probe is retracted at a constant strainrate (ε), which is obtained by creating an exponential moving profileon the rheometer (please see Figure 2a). This generates a force−displacement curve, which is then converted to an engineering stress−strain curve up to 178% engineering strain (to 0.834 mm) todetermine peak stress, σmax, and the energy dissipation per volume,EDV (in units of J/m3). A distinction is made between the EDVbefore σmax is reached, “pre-peak EDV”, and the EDV after σmax isreached, “post-peak EDV”. A schematic representation of theseparameters is presented in Figure 2b.

Theoretical Prediction for the Temporal Evolution of theTack Force. We use the following theoretical framework to accountfor the evolution of tack force over time. During a tack test, the gapbetween the probe and substrate remains much smaller than the plateradius, i.e., h(t) ≪ R0. Thus, for noncavitation cases, we can assumethat the flow in the thin gap is dominated by viscous effects and the

Figure 1. (a) Schematic of LxMn+ complexes in a 4PEG-ligand (L) +

metal ion (Mn+) transient hydrogel and their proposed contributionsto gel cohesion and interfacial adhesion. (b) Chemical structure of aHis2Ni

2+ complex.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

15980

pressure gradient under the plate is mainly balanced by the gradient ofviscous stresses (similar to squeeze flow or Stefan’s flow).20 We canalso assume that the apparent adhesion force is due to the suctionpressure that is generated through the action of viscous stresses in theflow. Spiegelberg et al. studied this flow for viscous Newtonian fluidsand found an approximate solution for the flow kinematics in thisproblem.21 It has been shown that the described Newtoniankinematics is fairly accurate even for viscoelastic liquids.22 Theapproximate kinematics can be coupled with linear constitutiveequations, such as the Maxwell model (one spring and one dashpot inseries), which leads to a theoretical prediction for the temporalevolution of the adhesion force, F(t), in eq 1

F tG R

tG

Rh

WiWi

Wit t

Wit

Wit

( ) ( )

31 2

exp 2 exp

exp 23

12

exp 3

p p

c c

c c

02

0

0

2

πσ

τ τ

τ τ

=

=−

− − −

− − −

ikjjjjj

y{zzzzz

Ä

Ç

ÅÅÅÅÅÅÅÅÅÅÅikjjjjj

y{zzzzz

ikjjjjj

y{zzzzz

É

Ö

ÑÑÑÑÑÑÑÑÑÑÑÄ

Ç

ÅÅÅÅÅÅÅÅÅÅÅikjjjjj

y{zzzzz

ikjjjjj

y{zzzzz

É

Ö

ÑÑÑÑÑÑÑÑÑÑÑ (1)

where the geometrical parameters R0 and h0 represent the plate radiusand the initial gap height, and Wi is the Weissenberg number (Wi ≡τcε). This relationship shows that the maximum adhesion force is setby a combination of the linear rheological properties, geometricalfactors, and Wi. The maximum of this expression can be easilycalculated at different Wi, and the suggested model can be compared

to experimental measurements. A more detailed derivation is providedin the Supporting Information (S1).

■ RESULTS AND DISCUSSIONLinear Viscoelastic Properties (Gp and τc) of Tran-

siently Cross-linked 4PEG-His + Ni2+ Hydrogels. Asintroduced above, the model system consists of 10 kDa 4PEG-His polymers forming strong yet reversible bonds with Ni2+ tocreate a transiently cross-linked viscoelastic hydrogel. Todetermine how their linear viscoelastic properties affect tackbehavior, we first performed oscillatory frequency sweeps inshear on all hydrogel samples. Because the hydrogels exhibitedsingle-mode Maxwellian behavior, a single Gp and τc wasobtained from each frequency sweep when fitted, as shown inFigure S1. The trends in the average Gp and τc with respect tothe Ni2+-His ratio match those previously reported in Grindyet al. (see Figure S2),14 wherein the value of Gp wasdemonstrated to scale with the predicted amount of elasticallyactive chains in the gel network, as described by eq 2

GHis Ni 3 His Ni

Hisppredicted 2

23

2

total∝

[ ] + [ ][ ]

+ +

(2)

The τc trend also matches that found in Grindy et al., even ifthe polymer molecular weight and gel wt % studied weredifferent.14,23

Qualitative Categorization of the Tack Tests viaVisual Documentation. On the basis of video recordedduring the tack test, each tested sample was qualitatively placedinto one of three categories: (a) no cavitation, no rupture, (b)no cavitation, rupture, and (c) cavitation. Representative cross-sectional sample snapshots for each category can be found inFigure S3; corresponding stress−strain curve shapes areincluded as well. In the cases where no cavitation or ruptureis detected, the sample stays well-adhered at the probe andsubstrate interfaces and is drawn into a single central fibril.“Rupture” is defined as a rip that forms when the pressureinside the elongating fibril suddenly equalizes with itsenvironment, but the sample does not debond from thesubstrate or the probe. The transition from “no rupture” to“rupture” occurs when the strain rate is increased and/or thegel composition results in a stiffer, slower-relaxing sample andis represented in the shape of the stress−strain curve by a sharpdecrease in stress. “Cavitation” is defined as the formation andgrowth of air bubbles from the interface of the fixed glasssubstrate and potentially into the bulk gel due to negativehydrostatic stress created in this sample geometry (i.e., a radiusmuch larger than its thickness).24 The presence of cavitation isindependent of the presence of rupture, but cavitation tends tooccur at higher strain rates and in stiffer, slower-relaxing gelsamples; if rupture does occur, it usually happens at lowerstrains when cavitation also exists.Unlike in traditional pressure sensitive adhesives, which

typically contain much longer-lived cross-links, no stressplateau corresponding to fibrillation is observed after thepeak stress.25 Instead, we observe an exponential decay afterthe peak stress (or after a rupture event occurs). We attributethis to the fully transient nature and narrow distribution ofdynamic cross-link time scales of the metal-coordinate modelsystem; however, more investigation is required to confirm thishypothesis.

Relationship between Peak Stress (σmax) and LinearViscoelastic Properties. To determine how the tack

Figure 2. (a) Schematic of a constant strain rate tack test (dimensionsare not to scale). (b) Representative stress−strain tack test curvedefining peak stress (σmax) and pre-peak and post-peak energydissipation per volume (EDV).

Langmuir Article

DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

15981

properties of a hydrogel may be influenced by its linearrheological behavior, i.e., Gp and τc, we first investigated howσmax changes with ε (0.1−1.6 1/s) and gel composition (Figure3). The Y-axis error bars indicate the lowest and highest valuesrecorded for a given data point after averaging the results.

We noticed three trends that correspond well to those foundin traditional pressure sensitive adhesives (PSAs),25 eventhough our model system is fully transient. First, withincreasing ε in the range studied, σmax increases toward aplateau. Second, for each Ni2+-His ratio, 10 wt % gels reachhigher σmax values than their 5 wt % counterparts. And third,within a given wt %, a gel with a higher Gp reaches higher σmaxvalues. Like in PSAs, σmax increases with strain rate due toviscoelastic losses and correlates well with the storage modulus(G′), which at the strain rates tested is effectively Gp.

25

To determine the effect of the linear mechanical propertiesof Gp and τc on the value of σmax, dimensionless variables arecreated; ε is multiplied by τc to generate the Weissenbergnumber, Wi, while σmax is divided by Gp (please see Figure S4for a full data plot of each individual gel sample accordingly).Generally speaking, σmax is dictated by the competitionbetween elastic chain deformation and plasticity introducedby dissociating cross-links. However, the onset of cavitationreduces σmax and the propagation of the cavities tends toinduce inconsistent catastrophic rupture and/or failure. Goingforward, we therefore only focus on gel samples that did notundergo cavitation, which could be experimentally determinedvia an empirical cavitation threshold of approximately 40 kPa(i.e., the dotted line in Figure 3). In Figure 4, we plot thesamples that did not undergo cavitation alongside the linearmechanical theoretical model (as derived from eq 1). Theexperimental data points are found to match the modelprediction well, which strongly indicates that the gel networkdensity, as quantified by Gp, and the gel network dynamics, asquantified by τc, indeed determine σmax.The gel samples that exhibited rupture are observed to fall

slightly below the theoretical prediction, which may be caused

by the expansion of initially optically invisible air bubbledefects allowing for premature rupture to occur thereby slightlyreducing σmax.

Relationship between Energy Dissipation and LinearViscoelastic Properties. Because our model material tendsto draw into a single fibril and undergo large extension ratiosrather than debond at the interface, adhesion energy (J/m2) isnot a relevant measure for characterizing its tack properties.26

Instead, we defined EDV with units of J/m3 by accounting forthe initial thickness of the sample as well. Like the σmaxanalyses, EDV comparisons were only done with gels thatdid not undergo cavitation. Like σmax vs ε, EDV, pre-peakEDV, and post-peak EDV have similar trends regarding ε, wt%, and Gp; EDV increases with ε, 10 wt % gels have higherEDVs than 5 wt % gels, and a gel with a higher Gp has a higherEDV. These trends, which can be observed in Figure S5a−c,are due to an increase in viscoelastic losses as ε increases andare consistent with previous studies.25 When the axes are madedimensionless, i.e., EDV/Gp vs Wi (Figure S 5d), moreinsightful trends can be observed, especially when split into thepre-peak and post-peak EDV/Gp vs Wi graphs shown in Figure5.The results suggest that in addition to being influenced by

Gp and τc, post-peak EDV is also dependent on wt %. Thesefindings are consistent with the proposed molecular mechanicsof reversible cross-links, in the following way. Prior to reachingσmax, the gel predominantly undergoes elastic deformationunder the applied load; the ligands in the coordination cross-links have not yet detached and relaxed, and so the elasticallyactive chains (whose fraction is dictated by the metal−ligandratio) determine both the pre-peak EDV and Gp. Therefore,pre-peak EDV/Gp should not have an additional dependenceon gel composition. However, after reaching σmax, remodeling,i.e., the ligand detachment and/or exchange betweencomplexes, results in a reduction of elastically active chains.Since the probability for successful ligand exchange to occurdecreases with increased polymer wt %, possibly due tohindered molecular motion as a ligand tries to find a differentcomplex to join, the observed increase post-peak EDV/Gp

Figure 3. Peak stresses (σmax) achieved by 4PEG-His + Ni2+ gelsduring tack tests at a given strain rate (ε). A composition-independentthreshold was empirically found at around 40 kDa (indicated by adashed line), below which no gel sample cavitation occurred.

Figure 4. σmax/Gp vs Wi for gel samples with no cavitation (symbols)compared to the theoretical prediction derived from eq 1 (line). Thecolor of the symbol also indicates whether the corresponding gelsample experienced rupture (blue) or not (pink).

Langmuir Article

DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

15982

makes sense.27−30 The effect of wt % on cross-link density isencapsulated in Gp, but its effect on network remodeling in thepost-peak region of the stress−strain curve is not, hence theadditional dependence.

■ CONCLUSIONS

In this study, we demonstrated how strongly the tackparameters of peak stress (σmax) and energy dissipation pervolume (EDV) correlated with the linear rheological propertiesof plateau modulus (Gp) and characteristic relaxation time (τc)for 4PEG-His + Ni2+ gels during tack tests with no cavitation.For σmax, a theoretical model was developed that had thetemporal evolution of the tack force dependent on Gp, τc, theWeissenberg number (Wi), and geometrical factors; this modelwas then compared to experimental data to furtherdemonstrate this relationship. For energy dissipation pervolume, in addition to the dependency of EDV on linearrheological properties, the polymer wt % affects the EDV afterthe peak stress is reached because network density affects theability of the gel to remodel at higher strains. This studydemonstrates that when tensile tests cannot be performed, tacktests can provide comparable information. With mussel-inspired metal coordinate cross-links providing methods totailor the relaxation spectrum of a viscoelastic hydrogel, thesefindings can inform the design of better adaptable adhesives.While this study already provides pertinent information

regarding the possible use of metal coordination in futurehydrogel adhesives, further investigation is needed in twoareas. First, since adhesion is a balance of cohesive andinterfacial strength, attention must be paid to behavior at thegel-substrate interface; this can be done by performing tests onmultiple substrates with different surface energies or roughnessor by including a primer layer. Interfacial physics andmechanics on the mesoscopic length scales should also bestudied. Second, knowledge obtained from the Ni2+-Histransient gel model system can be applied to a Fe3+-Dopacross-linked gel in an effort to decouple the combined effectsof coordinate and covalent bonds within this hybrid system;the mix of covalent and metal-coordinate bonds better matcheswhat is likely present in a mussel’s adhesive plaque. The

performance of tack tests on Fe3+-Dopa gels and theconnection of tack parameters to linear rheological propertieswould also provide us with an understanding of the adhesivedynamics of these gels and thus inform the design ofunderwater or medical adhesives.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.lang-muir.9b02772.

Full derivation of the theoretical prediction for thetemporal evolution of the tack force, a representativefrequency sweep for one of the 4PEG-His + Ni2+

hydrogels studied, average Gp and τc for the gels, aschematic depicting the three qualitative categories ofsample behavior, σmax/Gp vs Wi with the cavitation casesincluded, EDV vs strain rate and total EDV/Gp vs Wigraphs for noncavitation cases of tack tests on the gels(PDF)

■ AUTHOR INFORMATIONCorresponding Author*E-mail: holten@mit.edu.ORCIDBavand Keshavarz: 0000-0002-1988-8500Niels Holten-Andersen: 0000-0002-5318-9674NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work is financially supported in part by the Office ofNaval Research (ONR) under the Young InvestigatorsProgram Grant ONR.N00014-15-1-2763 and the MRSECProgram of the National Science Foundation under AwardDMR-1419807. In addition, E.L.’s work is supported in part bythe Anne M. Mayes Fellowship through the Department ofMaterials Science and Engineering at MIT as well as by theNational Science Foundation via the Graduate Research

Figure 5. EDV/Gp, (a) pre-peak and (b) post-peak, vsWi for noncavitation cases of tack tests on 4PEG-His gels. The results for the Ni2+-His ratiosof 0.33 and 0.4 are not plotted because there were not enough noncavitation cases to analyze. The alignment of the pre-peak EDV data indicatesthat like for σmax, pre-peak EDV is predominantly influenced by Gp and τc. Post-peak EDV, on the other hand, has an additional dependence onpolymer wt %.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

15983

Fellowship Program under Grant No. 2388357. The authorswould like to thank their colleagues in the Laboratory for Bio-Inspired Interfaces, their collaborators Daniel Darby and Dr.Jonathan Pham (University of Kentucky), and members of theadhesion community for their insights and support.

■ REFERENCES(1) Lee, B. P.; Messersmith, P. B.; Israelachvili, J. N.; Waite, J. H.Mussel-Inspired Adhesives and Coatings. Annu. Rev. Mater. Res. 2011,41, 99−132.(2) Brubaker, C. E.; Messersmith, P. B. The Present and Future ofBiologically-Inspired Adhesive Interfaces and Materials. Langmuir2012, 28, 2200−2205.(3) Cui, M.; Ren, S.; Wei, S.; Sun, C.; Zhong, C. Natural and Bio-Inspired Underwater Adhesives: Current Progress and NewPerspectives. APL Mater. 2017, 5, 116102.(4) Lee, H.; Scherer, N. F.; Messersmith, P. B. Single-MoleculeMechanics of Mussel Adhesion. Proc. Natl. Acad. Sci. U. S. A. 2006,103 (35), 12999−13003.(5) Waite, J. H. Mussel Adhesion − Essential Footwork. J. Exp. Biol.2017, 220 (4), 517−530.(6) Ahn, B. K. Perspectives on Mussel-Inspired Wet Adhesion. J. Am.Chem. Soc. 2017, 139 (30), 10166−10171.(7) Fullenkamp, D. E.; He, L.; Barrett, D. G.; Burghardt, W. R.;Messersmith, P. B. Mussel-Inspired Histidine-Based TransientNetwork Metal Coordination Hydrogels. Macromolecules 2013, 46(3), 1167−1174.(8) Fullenkamp, D. E.; Barrett, D. G.; Miller, D. R.; Kurutz, J. W.;Messersmith, P. B. PH-Dependent Cross-Linking of Catecholsthrough Oxidation via Fe 3+ and Potential Implications for MusselAdhesion. RSC Adv. 2014, 4 (48), 25127−25134.(9) Holten-Andersen, N.; Harrington, M. J.; Birkedal, H.; Lee, B. P.;Messersmith, P. B.; Lee, K. Y. C.; Waite, J. H. PH-Induced Metal-Ligand Cross-Links Inspired by Mussel Yield Self-Healing PolymerNetworks with near-Covalent Elastic Moduli. Proc. Natl. Acad. Sci. U.S. A. 2011, 108 (7), 2651−2655.(10) Barrett, D. G.; Fullenkamp, D. E.; He, L.; Holten-Andersen, N.;Lee, K. Y. C.; Messersmith, P. B. PH-Based Regulation of HydrogelMechanical Properties through Mussel-Inspired Chemistry andProcessing. Adv. Funct. Mater. 2013, 23 (9), 1111−1119.(11) Kawamoto, K.; Grindy, S. C.; Liu, J.; Holten-Andersen, N.;Johnson, J. A. Dual Role for 1,2,4,5-Tetrazines in Polymer Networks:Combining Diels-Alder Reactions and Metal Coordination toGenerate Functional Supramolecular Gels. ACS Macro Lett. 2015, 4(4), 458−461.(12) Grindy, S. C.; Learsch, R.; Mozhdehi, D.; Cheng, J.; Barrett, D.G.; Guan, Z. B.; Messersmith, P. B.; Holten-Andersen, N. Control ofHierarchical Polymer Mechanics with Bioinspired Metal-Coordina-tion Dynamics. Nat. Mater. 2015, 14 (12), 1210−1216.(13) Li, Q.; Barrett, D. G.; Messersmith, P. B.; Holten-Andersen, N.Controlling Hydrogel Mechanics via Bio-Inspired Polymer−Nano-particle Bond Dynamics. ACS Nano 2016, 10 (1), 1317−1324.(14) Grindy, S. C.; Lenz, M.; Holten-Andersen, N. EngineeringElasticity and Relaxation Time in Metal-Coordinate Cross-LinkedHydrogels. Macromolecules 2016, 49 (21), 8306−8312.(15) Grindy, S. C.; Holten-Andersen, N. Bio-Inspired Metal-Coordinate Hydrogels with Programmable Viscoelastic MaterialFunctions Controlled by Longwave UV Light. Soft Matter 2017, 13(22), 4057−4065.(16) Vaccaro, E.; Waite, J. H. Yield and Post-Yield Behavior ofMussel Byssal Thread: A Self-Healing Biomolecular Material.Biomacromolecules 2001, 2 (3), 906−911.(17) Zhao, H.; Waite, J. H. Proteins in Load-Bearing Junctions: TheHistidine-Rich Metal-Binding Protein of Mussel Byssus. Biochemistry2006, 45 (47), 14223−14231.(18) Harrington, M. J.; Waite, J. H. Holdfast Heroics: Comparingthe Molecular and Mechanical Properties of Mytilus CalifornianusByssal Threads. J. Exp. Biol. 2007, 210 (24), 4307−4318.

(19) Schmidt, S.; Reinecke, A.; Wojcik, F.; Pussak, D.; Hartmann, L.;Harrington, M. J. Metal-Mediated Molecular Self-Healing inHistidine-Rich Mussel Peptides. Biomacromolecules 2014, 15 (5),1644−1652.(20) Stefan, J. Versuche Uber Die Scheinbare Adhasion. Ann. Phys.1875, 230 (2), 316−318.(21) Spiegelberg, S. H.; Ables, D. C.; McKinley, G. H. The Role ofEnd-Effects on Measurements of Extensional Viscosity in FilamentStretching Rheometers. J. Non-Newtonian Fluid Mech. 1996, 64 (2−3), 229−267.(22) Jaishankar, A. The Linear and Nonlinear Rheology of MultiscaleComplex Fluids; Massachusetts Institute of Technology: 2014.(23) Parada, G. A.; Zhao, X. Ideal Reversible Polymer Networks. SoftMatter 2018, 14 (25), 5186−5196.(24) Crosby, A. J.; Shull, K. R.; Lakrout, H.; Creton, C. Deformationand Failure Modes of Adhesively Bonded Elastic Layers. J. Appl. Phys.2000, 88 (5), 2956−2966.(25) Lakrout, H.; Sergot, P.; Creton, C. Direct Observation ofCavitation and Fibrillation in a Probe Tack Experiment on ModelAcrylic Pressure-Sensitive-Adhesives. J. Adhes. 1999, 69 (3−4), 307−359.(26) Zosel, A. Molecular Structure, Mechanical Behaviour andAdhesion Performance of Pressure Sensitive Adhesives. Adhes SealantsInd. Mag 2001, 193−203.(27) Yount, W. C.; Loveless, D. M.; Craig, S. L. Small-MoleculeDynamics and Mechanisms Underlying the Macroscopic MechanicalProperties of Coordinatively Cross-Linked Polymer Networks. J. Am.Chem. Soc. 2005, 127 (41), 14488−14496.(28) Yount, W. C.; Loveless, D. M.; Craig, S. L. Strong Means Slow:Dynamic Contributions to the Bulk Mechanical Properties ofSupramolecular Networks. Angew. Chem., Int. Ed. 2005, 44 (18),2746−2748.(29) Mozhdehi, D.; Neal, J. A.; Grindy, S. C.; Cordeau, Y.; Ayala, S.;Holten-Andersen, N.; Guan, Z. Tuning Dynamic MechanicalResponse in Metallopolymer Networks through SimultaneousControl of Structural and Temporal Properties of the Networks.Macromolecules 2016, 49 (17), 6310−6321.(30) Tang, S.; Olsen, B. D. Relaxation Processes in SupramolecularMetallogels Based on Histidine-Nickel Coordination Bonds. Macro-molecules 2016, 49 (23), 9163−9175.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b02772Langmuir 2019, 35, 15979−15984

15984