Articleshttps://doi.org/10.1038/s41565-018-0227-7

Tunable thermal transport and reversible thermal conductivity switching in topologically networked bio-inspired materialsJohn A. Tomko1,10, Abdon Pena-Francesch2,3,10, Huihun Jung2,3, Madhusudan Tyagi4,5, Benjamin D. Allen6,7, Melik C. Demirel   2,3,7* and Patrick E. Hopkins1,8,9*

1Department of Materials Science and Engineering, University of Virginia, Charlottesville, VA, USA. 2Center for Research on Advanced Fiber Technologies (CRAFT), Materials Research Institute, Pennsylvania State University, University Park , PA, USA. 3Department of Engineering Science and Mechanics, Pennsylvania State University, State College, PA, USA. 4NIST Center for Neutron Research, Gaithersburg, MD, USA. 5Department of Materials Science and Engineering, University of Maryland, College Park, MD, USA. 6Department of Biochemistry and Molecular Biology, Pennsylvania State University, University Park, PA, USA. 7Huck Institutes of the Life Sciences, Pennsylvania State University, University Park, PA, USA. 8Department of Mechanical and Aerospace Engineering, University of Virginia, Charlottesville, VA, USA. 9Department of Physics, University of Virginia, Charlottesville, VA, USA. 10These authors contributed equally: John A. Tomko, Abdon Pena-Francesch. *e-mail: MDemirel@engr.psu.edu; phopkins@virginia.edu

SUPPLEMENTARY INFORMATION

In the format provided by the authors and unedited.

NATurE NANoTECHNology | www.nature.com/naturenanotechnology

2

TableI.DataofthermalconductivityswitchesfromFigure4cinManuscriptMaterial Switching

Temperature(K)

Thermal ConductivitySwitchRatio(κmax/κmin)

SwitchingMechanism

TR-n4 295 1.65±0.25 Hydration

TR-n7 295 2.74±0.30 Hydration

TR-n11 295 4.10±0.46 Hydration

TR-n25 295 3.92±0.31 Hydration

TR-n∞ 295 4.54 Hydration

H2O1 273 2.66 Phasechange(S/L)

LixCoO22 295 1.45 Electrochemicallithiation

C703 298 1.45 Appliedpressure

PZT4 298 1.11 Ferroelectric domainscattering(electricfield)

Bnanoribbon5,6 300 1.61 Wetting

SWCNT/n-C18H387 303 2.5 Phasechange(S/L)

Ga8–10 303 1.57 Phasechange(S/L)

PEG400011 313.15 1.42 Phasechange(S/L)

K9,10 337 1.80 Phasechange(S/L)

VO212 340 1.41 Metal-insulator

transition(S/S)

RM25713 353 1.55 Molecular reorientation(magneticfield)

3

Time-domainthermoreflectance(TDTR)Inthiswork,weuseatwo-colorsetup,whereatrainof80MHz,90fs,800nmlaserpulses

emanatesfromaTi:Sapphireoscillator(SpectraPhysicsTsunami)andissplit intoapumpandprobepath.Thepumppulsesareelectro-opticallymodulatedatafrequencyof8.8MHzandthenfrequency-doubledto400nm,allowingustoperformourTDTRmeasurementswithbothAlandAutransducers(duetoincreaseabsorptioninAuat400nm).Theprobepulseremainsat800nmand ismechanically delayed to create a temporal delay between the pump and probe. Thethermoreflectanceof theprobe ismeasuredvia lock-indetectionasa functionof thepump-probedelaytimetoobtainthethermaldecayfromthemetaltransducerintothesurroundingfilms. Note, we havemeasured the thermal conductivity of the protein films under varyingexperimentalconditions,namelywithdifferentobjectivesandpowerdensitiesduringtheTDTRmeasurements.Themost repeatedexperimental conditionswere theuseofa10Xobjective,leadingtoapumpdiameterof~19µmandaprobediameterof∼9µmwithtypicallylessthana15mWpumppowerwhilemodulatingthepumppulsesat8.8MHz.Theprobepowerwas∼5mW.TheselaserpowerswerespecificallyaimedtowardsavoidingdenaturationoftheproteinfilmduetolargelocaltemperaturerisesduringtheTDTRscanwhilestillobtainingalargeenoughout-of-phasesignalforreliabledata.

Toensurethemetalfilmisopticallyspecular,weelectron-beamevaporateeithertheAlorAufilmonamorphousSiO2substratesandthensolvent-casttheTRfilmontheopposingsideofthedeposited80nmmetaltransducer;theproteinfilmsaredriedovernightandthenwashedwithDIwater.

We perform the TDTR measurement through the transparent substrate in a “probe-up”geometry;aphotographofthisuprightschematicisshowninFig.S1aswellasdepictedinthemanuscriptinFig.1a.Theexperimentallymeasureddecaycurveisfittoabidirectionalthermalmodel. An example of the fit is shown in Fig. S2 for TR-n25 in both ambient and hydratedconditions.Forthehydratedmeasurements,wesimplyapply500µLofDIwaterontopoftheTRfilmintheprobedregion;toreturnto‘ambient’conditions,thisdropispipettedoffthesurface.Wefindthatthethermalconductivityswitchesontheorderoftypicalliquiddiffusiontimes(i.e.,ontheorderofseconds);weareabletorepeatedlymeasurethischangeinthermalconductivityviaTDTRwithnoexperimentalwaitingtimebetweenscans.Note,wefindanasymmetryinthe“switch on” time upon hydration) and “switch off” time when the water is removed; thisphenomenaisdiscussedindetailinthefollowingsectionoftheSupplementaryInformation.Asthebest-fitvalueinthismodelreliesonthedependenceofknowingthethermalpropertiesof the SiO2 substrate and 80 nmmetal film,wemeasure the film/substrate system prior todepositionoftheproteinaceousfilms.Ourthermalmodelaccountsforthefollowingvalues:weexperimentallymeasureκSiO2=1.35Wm−1K−1,insurethefilmthicknessis80nmviaprofilometryandpicosecondacousticmeasurements,andκAl=130Wm−1K−1basedonelectricalresistivityandTDTRmeasurements.Theheatcapacityofthesubstrateandfilmarebasedon literaturevaluesof1.66and2.42MJm−3K−1respectively.Furthermore,duetothethree-layersystemandrelatively lowthermalconductivityof theTRfilms,TDTR isequallysensitivetoboththeheatcapacityandthermalconductivityofthematerial(i.e.,thethermalmodelcanonlyfitforthermaldiffusivity, "#$,inthisregion.Aswemeasurethevolumetricheatcapacityofourproteinfilmsin

bothambientandhydratedconditionsviaDSC,asdiscussedinalatersection,weareabletofitfor solely the thermal conductivityof theTR film. Thevaluesused forourTDTRanalysis aretabulatedinTableII.

4

Table II. Measured material parameters for TR films and comparison to R.T. thermalswitchesSample TR-n4 TR-n7 TR-n11 TR-n25

n(repeatunits) 4 7 11 25

M.W.(kDa) 15 25 42 86

Cv,ambient(MJm-3K-1) 1.76 1.71 1.73 1.72

κambient(Wm-1K-1) 0.35±0.09 0.34±0.11 0.27±0.10 0.33±0.10

Cv,hydrated(MJm-3K-1) 2.47 2.60 2.77 2.82

κhydrated(Wm-1K-1) 0.58±0.09 0.93±0.10 1.12±0.13 1.30±0.10

Ratio(κmax/κmin)1.65±0.26 2.74±0.30 4.10±0.46 3.92±0.31

Material SwitchingTemperature(K)

ThermalConductivitySwitchRatio(κmax/κmin)

SwitchingMechanism

H2O1 273 2.66 Phasechange(S/L)

LixCoO22 295 1.45 Electrochemicallithiation

C703 298 1.45 Appliedpressure

PZT4 298 1.11 Ferroelectric domain scattering(electricfield)

Bnanoribbon5,6 300 1.61 Wetting

SWCNT/n-C18H387 303 2.5 Phasechange(S/L)

Ga8–10 303 1.57 Phasechange(S/L)

PEG400011 313.15 1.42 Phasechange(S/L)

K9,10 337 1.80 Phasechange(S/L)

5

VO212 340 1.41 Metal-insulatortransition

(S/S)

RM25713 353 1.55 Molecular reorientation(magneticfield)

To confirm the sensitivity and uncertainty associatedwith our bidirectionalmodel to the

thermalconductivitiesofmaterialsontheorderofourTRfilms,weuseidentical80nmAlfilmsonSiO2substratesaswereusedforthebiopolymerfilms,andinsteadspin-coat200nmPMMA,Poly(methyl-methacrylate),filmsontheAlfilm(PMMA/Al/SiO2).Additionally,weuseaseparatefilmandaddwater to the topof theAl film (water/Al/SiO2). Inbothcases,weuse thesameparametersandmodelasused for thebiopolymer filmmeasurements. Inaccounting for thepropagationoferrorinvaluesusedforourthermalmodel,asdiscussedbelow,wemeasurethethermalconductivityofwateratroomtemperaturetobe0.59±0.1Wm-1K-1andthethermalconductivityofPMMAtobe0.18±0.07Wm-1K-1;thesevaluesareinexcellentagreementwithliteraturevaluesforthetwomedia.Althoughwe‘reference‘thesupportingmetaltransducerandunderlyingSiO2substrate,asdis-

cussedabove,thepropagationoferrorwithinthesevaluesleadstoincreaseduncertaintyinthemeasuredthermalconductivityofourTRfilms.ToaccuratelydeterminethetrueuncertaintyinourTDTRmeasurements,weconsiderthemeansquaredeviationofourbi-directional,three-layer thermal model to the experimental TDTR data obtained for the various TR films; thiscontour analysis uncertainty calculation is discussed in detail in previous references14,15. Tosummarize,thecontourdisplaysthenormalizedresidual,orthemodel’smeansquaredeviation,relative to the experimentallymeasured thermal decay curve; an increase in thenormalizedresidual isequivalenttoapoorfittothedata.Forthebi-directionalthermalmodel,withthethermaleffusivitiesinvestigatedinthiswork,themostsensitiveparameteristhethicknessofthemetaltransducer.Thus, inconsideringthemaximumpossibleuncertaintyinthevaluesofbestfit forourthermalconductivitymeasurementsoftheTRfilms,weextendthissensitivityanalysisto5%errorinfilmthickness,asshowninFig.S3;wefindthatourfitstotheexperimentaldata lie within the inner most contour, in dark blue. Given such, this contour denotes theuncertaintyinthermalconductivityfora5%uncertaintyinfilmthickness(thethermalmodel’smost sensitiveparameter).Additionally, repeating these contours for various thermophysicalparameters inourmodel, suchas thermal conductivityof the SiO2 substrate leads tonearlyidenticalcontourplots (i.e., themodel isnothighlysensitive to thisparameter).Weplot theinner-mostcontourforTR-n25asanexamplewiththereferencevalueoftheSiO2substrate(1.35Wm-1K-1)and±10%ofthisvalueinFig.S4.

Toruleoutpotentialmetal/proteininteractionsthatmaybeleadingtothemeasuredchangeinthermalconductivityuponhydration,werepeatthemeasurementsusingAutransducersofthesamethickness.TheresultsareshowninFig.S5.Asseen,thetwoareingoodagreementandprovidesimilarswitchingvaluesbetweenthetwostates.

Tofurtherreinforcetheaccuracyofthethermalconductivityvaluesreportedinthiswork,itshouldbenotedthataminimumoffoursamplesperTRfilmweremeasured;weconductedno

6

lessthan40TDTRscansoneachofthesesamples(e.g.,thethermalswitchmeasurementforTR-n25wasperformedover200timesalone).Additionally,themajorityofthesemeasurementsareconductedonvariouspointsthroughoutthesampletoaccountforanyinhomogeneityineitherthemetaltransducerorTRfilm.Thestandarddeviationofthesemeasurementsisaccountedforinourreportederror.

SwitchtimesandwatercontentInthedatareportedinthiswork,weconductTDTRscanoncethesampleshaveachievedtheir

“fully on” (hydrated) or “fully off” (ambient) state. Note, throughout this work, we defineambient as being the TR protein in an unsaturated, ambient environment; for our TDTRmeasurementsattheUniversityofVirginia,wemeasuretherelativehumidityofourlaboratorytobebetween28-35%.Conversely,thehydratedmeasurementsareperformedonapositionofthe TR protein film that is fully submerged in an aqueous environment, as described in theprevioussection.AsasingleTDTRscantakesapproximatelytwominutes,wecannotconductfullTDTRscansduringthetransitionfromambienttohydratedtogaininsight intothetimescaleassociatedwiththeswitchingofthermalconductivityintheTRfilms.However,toquantifytheorderofmagnitudeofthisthermalconductivityswitchingtime,weinsteadusetheTDTRsignalasarelativemonitorofthetemperaturechangebymonitoringtheTDTRsignalatasinglepump-probedelaytimeasafunctionoflabtimewhilemanuallypipettingofwateronandoffofthesampletorepeatedlyhydrateanddrythesample.Wehaveusedthisthisapproachof“staring”attheTDTRsignal intoobserverealtimechanges inthermalconductivity invarioussamplespreviously4,16.Wenotethatinourcurrentcase,thisproceduredoesindeedproducevariabilitiesin the absolute switching times due to inadvertent changes in the manual application andremoval of the water, along with sample to sample variabilities in film thickness which willchangeliquiddiffusiontimesthroughthematerials.Thus,whilethesemeasurementspreventusfromabsolutelyquantifyingtheintrinsicswitchingtimesofthesematerialswithhighprecision,thisdoesallowfororderofmagnitudequantificationoftheswitchingtimes.Whilewediscussthisprocedureinmoredetailbelow,wenotethatallreportedthermalconductivityvaluesinthiswork are determined when the system has equilibrated in either the full hydrated or fullyambientstates,aspreviouslymentioned.Inotherwords,thefullTDTRscansthatresultinthethermalconductivitydatapresentedinthemainmanuscriptareonlyperformedwhenthereisnotemporalfluctuationinthemeasuredlock-insignal.

Togiveinsightintotheswitchingtime,wefixourpump-probedelaytimeto500picosecondsandmeasurethelock-insignalasafunctionoflabtime.Fromthere,werepeatedlyhydrateanddry the sample, via manual pipetting of water, over the course of minutes. To convert themeasuredsignaltothermalconductivityasafunctionoftime,wenormalizethesignaltothemeasuredratiofromTDTRscansfortheambientandhydratedstatesat500picoseconds.Theresults for the TR films are shown in Fig. 4a of the primary manuscript for TR-n25. Uponhydration,thechangeinthermalconductivityisontheorderofseconds.

Note: the reported thermal conductivity values in this work are when the system hasequilibratedineithertheambientorhydratedstate. Inotherwords,theTDTRscansareonlyperformedwhenthereisnofluctuationinthemagnitudeofthelock-insignal.Inanattempttoquantifytheroleofvaryingwaterconcentrationsassociatedwiththisswitch,

weperformedourTDTRmeasurementsonTR-n11and-n25upto92%relativehumidityusingacustom-built chamber (see Fig.S1). Although the chamber and sample have excessive

7

condensation,as shown inFig.S6b,wherewaterdroplets similar to thoseused forour initialswitching experiments are visible on every surface in the chamber including the insetwhichshowsourTR-n11sample,ourprobedregionremainsdrydueto localized laserheating.Thisfindingisinconceptualagreementwiththerapidswitchtimefoundwhenoursampleisdriedthroughmanualpipettingofawaterdroplet;ourTDTRmeasurement leads toasteady-statetemperature rise from laser heating, which causes rapid local evaporation from the probedregion. In the caseof humidity/concentrationmeasurements, this local heatingdisallows forcondensationinthemeasuredarea.

Time-domainBrillouinscatteringWemeasurethegroupvelocityofcoherentacousticphononsthroughpump-probeBrillouin

scattering measurements using a two-tint version of our TDTR apparatus. In pump-probereflectionmeasurements,suchasTDTR,the initialpumppulsecreatesaheatingevent inthetransducer;theenergyimpartedfromthelaserpulsepairedwiththistemperatureriseleadstotheformationofcoherentacousticwavewithinthesurroundingmedia.Asacoherentacousticwavetraversesthismedia,inourcasetheTRfilms,theprobebeamispartiallyreflectedduetothechangeinrefractiveindexassociatedwiththepressuregradient;thetransmittedportionofthebeamcontinuestowardsthemetaltransducerandundergoesthermo-reflectionaswithatraditionalTDTRmeasurement.Asdescribedinpreviousworks17–19,thesetworeflectedbeamswill interfere and lead to periodic oscillations in the time-dependent reflectance signal. Theintensityofthistime-independentsignalcanbedescribedvia

% &, ( = * ∗ exp −Γ& ∗ cos 456 & − 7 − 8 ∗ exp − 9

: (S1)

whereA,B,andδareessentiallyscalefactors,Γisthedampingoftheacousticwave-frontdue

to energy dissipation, T is the period of the pressure front, and the second term, exp(−t/τ)accounts for the thermal decay associatedwith traditional thermo=reflectance at themetaltransducer.Whentheprobebeamisatnormalincidence,thefrequency,f,(i.e.,theinverseoftheperiod,T)canberelatedtothesoundspeedofthepartiallytransparentmaterialvia

< = 4=>

? (S2)whereristheindexofrefraction,λisthewavelengthoftheprobebeam,andvisthevelocityofthis coherentwave-front. In thecaseofourTR films, rhasbeendetermined inourpreviouswork20andourprobebeamis800nm.

Note,weutilizeatwo-tintpump-probeapparatus(i.e.,thepumpbeamisalso800nm)forthesemeasurementstoavoiddamageoftheTRfilm;asdescribedinourpreviouswork21,directinteractionbetweentheprotein filmandthe400nmpumpused inourTDTRmeasurementsleads to gradual degradation of the film. In the hydratedmeasurements, the sample is fullysubmergedbeneath3mmofDIwater;thesignal-to-noiseratiointhehydratedcaseisgreatlyreducedcomparedtothatofthefilmsinambientconditions.Nonetheless,well-definedpeaksremainvisible,showingclearreductioninsoundvelocitiescomparedtotheambientsamples.Theresultsfromourmeasurements/calculationsareprovidedinTableIIIoftheSupplementaryInformation.

8

TableIII.TDBSmeasurements

Sample TR15(-n4) T42(n11)

AmbientVelocity(m/s) 2374±45 1768±200

HydratedVelocity(m/s) 2088±291 1805±291

AmbientCv,(MJm-3K-1) 1.76 1.73

HydratedCv,(MJm-3K-1) 2.47 2.77

AmbientMSD(Å2) 0.55 0.55

HydratedMSD(Å2) 1.5 1.5

CalculatedSwitchingRatio ∼3.70 ∼4.90ProteinstructureAs shown in Figure 1b of the primarymanuscript, the proteins are composed ofβ-sheet

crystallites,whicharerichinalanine,andare2x2.5nminsize,andareconnectedby∼3nmamorphousstrands,whicharerichinglycine.

AsX-raydiffractiontechniquesarenotoptimaltoinvestigatethestructureandconformationofproteins,sincetheyonlyprovideinformationabouttheordereddomains(β-sheets)butnotabout the disordered strands, we have performed structural analysis of the ambient andhydratedproteinsbyFouriertransforminfrared(FTIR)spectroscopy(FigureS8).AmideIband(1600-1700cm-1)correspondstocarbonylstretchingvibration,andthuscontains informationabouttheproteinbackboneconfirmationandoverallsecondarystructure.TheamideIbandofambient and hydrated proteins is similar, showing that there are not significant structuralchangesuponhydration.Forfurtheranalysisanddeterminationofthecrystallinity(i.e.,β-sheetcontent),weanalyzedthesecondarystructureviaFourierself-deconvolution(FSD)oftheamideIbandandcorrespondingfittingof individualGaussianfunctions (FigureS9)22.Thesecondarystructurecontentofambientandhydratedfilmsaresimilar(withinerrorofeachother),withaβ-sheetcontentof54%,andamorphousstrandsadopting20%randomcoil,7%α-helix,and18%turnconformations.

RheologySRTproteinswerecompressedinPDMSmolds(hydrated,70◦C)intodiskshapedsamplesof

2mmindiameterand1mminheight.MeasurementswereperformedinaRheometricScientificARESrheometerwith3mmdiameterparallelplategeometrywitha10mLliquidreservoirinthebottomplate. Sampleswere adhered to theplatesusingClickBondCB200adhesive (2hourcuring time at room temperature, withmodulus in the GPa range). Liquid was fed into thereservoirwithaperistalticpump(UltraLowFlowMini-pumpmodel3384,ControlCompany)with

9

aflowrateof0.4mL/mintocompensateforevaporation,andthesystemwasequilibratedat70◦Cfor90minutes.

Dynamicmechanicalanalysis(DMA)DMAwasperformedinaTA800QDMAinstrument,whichhasaforcerangeof10-4to18N,

forceresolutionof10-5N,strainresolutionof1nm,andafrequencyrangeof0.01to200Hz.DMAwasused for smalldeformationanalysis (oscillatory strainand strain ramp) inorder tomeasuretheglassymechanicsandglasstransitionofSRTproteins.Film-tensionandsubmersion-tensionclampswereusedforambientandhydrated (saturation)measurementsrespectively.Sampledimensionswere15mm×2.5mm×0.2mmfortheambienttestand25mm×2.5mm×0.2mmforthehydratedtest.Oscillatorytemperaturesweepexperimentswereperformedat1Hz,with amplitudeof 2µmand a rateof 2 ◦C perminute. Stress-strain experimentswereperformedwithastrainrateof1%perminuteandapreloadof0.01N.

Temperaturemodulateddifferentialscanningcalorimetry(TMDSC)TAInstrumentsQ200DSCwasusedfortheglasstransitionmeasurementsofSRTproteins.

TheinstrumentwascalibratedforspecificheatCpmeasurementsusingTzeroaluminumpansandsapphire standard materials, including cell resistance and capacitance, cell constant andtemperaturecalibration.SRTsampleswereannealedonahotplateat100◦Cfor30minutesinordertoremovewaterbeforetheexperiment.HydratedproteinswereimmersedinDIwaterfor1hour,excesswaterwasremoved,andsamplesweresealedinhermeticpans.Sampleswereheatedat2◦C/minfrom25◦Cto210◦C(beginningofthermaldegradation)withamodulationperiodof60secondsandatemperatureamplitudeof0.318◦C.Hydratedsampleswererunfrom5 ◦C to 60 ◦C with the same rate, modulation period, and amplitude parameters. The glasstransitionwasanalyzedfromthereversingspecificheatcapacityCpdata.Theincreaseinheatcapacity ∆Cp is calculated as the difference between the glassy and rubbery tangent lines:Cp(T)rubbery=Cp(T)glassy+∆Cp.Theglasstransitiontemperature iscalculatedasthestepchangemidpoint:Cp(Tg)=Cp(Tg)glassy+∆Cp/2.Cp(T)glassyandCp(T)rubberyaretheextrapolationofthetangenttothespecificreversingCpbelowandabovetheglasstransitionrespectively.ThemeasuredheatcapacityvaluesinboththeambientandhydratedstatesarelistedinTableII.

It shouldbenoted that theaforementionedaluminumhermitic pans allow for theuseofliquidsandpreventevaporationasthesampleisheated.ThisisnotaproblemwhenmeasuringTRproteins inambientconditions, since theproteinsarepreviouslyannealed to removeanymoisture.Althoughtheycanbeheatedupto200◦C,theythermallydegradeasotherstructuralproteins23,24.However,hightemperaturecharacterizationandthermaldegradationarenotthefocusofthisworkandarenotdiscussedinthismanuscript.Ontheotherhand,absorbedwaterinhydratedproteinsdoessupposeaproblemforhightemperaturecalorimetry.Althoughthehermetic pans prevent the evaporation of water, water vapor pressure buildup at highertemperatures can breach the sealed pan. Besides changing the hydration conditions of thesampleof interest,awatervaporbreachwill generatea falseendothermartifactdue to theevaporation, and can potentially damage the calorimeter cell. For these reasons, TMDSCmeasurementsofhydratedproteinsampleswereperformedfrom5to50◦Conly,andtheywerenot cycled to prevent any water vapor pressure buildup inside the hermetic cells. Themeasurementswererepeatedforeachsampletoensurereproducibilityandsignaloverlap(twomeasurementsofeachareplottedinFigureS10).

10

NeutronScatteringQuasielastic neutron scattering experiments were performed at NIST Center for Neutron

Research(NCNR),GaithersburgMD.ExperimentsonambientSRTproteinswereperformedonthehigh-fluxbackscatteringspectrometer (HFBS)25,withenergyresolutionof1µeV,dynamicrangeof±15µeVandQ-rangeof0.25A˚ -1 - 1.8A˚ -1, at295K. The resolution functionwasmeasuredat4K(signaliscompletelyelastic).QuasielasticexperimentsonD2O-hydratedSRT-TRpolypeptides were performed on the disk chopper time-of-flight spectrometer (DCS)26, withenergyresolutionof64µeV(wavelengthλ=6Å),dynamicrangeof±0.5meVandQ-rangeof0.1Å-1-2Å-1,at295K.VanadiumwasusedastheresolutionfunctioninDCSmeasurements.

†MDemirel@engr.psu.edu‡phopkins@virginia.edu1. Slack,G.A.Thermalconductivityofice.PhysicalReviewB22,3065-3071(1980).2. Cho,J.,Losego,M.D.,Zhang,H.G.,Kim,H.,Zuo,J.,Petrov, I.,Cahill,D.G.,&Braun,P.V.

Electrochemically tunable thermal conductivity of lithium cobalt oxide. NatureCommunications5,4035(2014).

3. Soldatov, A. & Sundqvist, B. Molecular rotation in C70 at high pressures: A thermalconductivitystudy.JournalofPhysicsandChemistryofSolids57,1371–1375(1996).

4. Ihlefeld,J.F.,Foley,B.M.,Scrymgeour,D.A.,Michael,J.R.,Mckenzie,B.B.,Medlin,D.L., Wallace, M., Trolier-Mckinstry, S., & Hopkins, P. E. Room-Temperature VoltageTunablePhononThermalConductivityviaReconfigurableInterfacesinFerroelectricThinFilms.NanoLetters15,1791-1795(2015).

5. Kim, K.&Kaviany,M. Thermal conductivity switch:Optimal semiconductor/metalmeltingtransition.PhysicalReviewB94,155203(2016).

6. Yang,J.,Yang,Y.,Waltermire,S.W.,Wu,X.,Zhang,H.,Gutu,T.,Jiang,Y.,Chen,Y.,Zinn,A.A.,Prasher,R.,Xu,T.T.,&Li,D.EnhancedandswitchablenanoscalethermalconductionduetovanderWaalsinterfaces.NatureNanotechnology7,91-95(2012).

7. Harish,S.,Ishikawa,K.,Chiashi,S.,Shiomi,J.,&Maruyama,S.Anomalousthermalconductioncharacteristicsofphasechangecompositeswithsingle-walledcarbonnanotubeinclusions.JournalofPhysicalChemistryC117,15409–15413(2013).

8. Samsonov, G. V. Handbook of the Physicochemical Properties of the Elements. SpringerScience&BusinessMedia(2012).

9. Ho,C.Y.,Powell,R.W.,&Liley,P.E.ThermalConductivityoftheElements.JournalofPhysicalandChemicalReferenceData1,279-421(1972).

10. Dyos,G.T.,&Farrell,T.ElectricalResistivityHandbook.IET(1992).11. Chen,R.,Cui,Y.,Tian,H.,Yao,R.,Liu,Z.,Shu,Y.,Li,C.,Yang,Y.,Ren,T.,Zhang,G.,&Zou,

R.ControllableThermalRectificationRealizedinBinaryPhaseChangeComposites.ScientificReports5,8884(2015).

12. Oh,D.W.,Ko,C.,Ramanathan,S.,&Cahill,D.G.Thermalconductivityanddynamicheatcapacity across themetal-insulator transition in thin filmVO2.Applied Physics Letters96,151906(2010).

11

13. Shin,J.,Kang,M.,Tsai,T.,Leal,C.,Braun,P.V.,&Cahill,D.G.ThermallyFunctionalLiquidCrystalNetworksbyMagneticFieldDrivenMolecularOrientation.ACSMacroLetters5,955-960(2016).

14. Feser, J. P., & Cahill, D. G. Probing anisotropic heat transport using time-domainthermoreflectancewithoffsetlaserspots.ReviewofScientificInstruments83,104901(2014).

15. Giri,A.,Niemela,J.P.,Tynell,T.,Gaskins,J.T.,Donovan,B.F.,Karppinen,M.,&Hopkins,P. E. Heat-transportmechanisms inmolecular building blocks of inorganic/organic hybridsuperlattices.PhysicalReviewB93,115310(2016).

16. Donovan,B.F.,Tomko,J.A.,Giri,A.,Olson,D.H.,Braun,J.L.,Gaskins,J.T.,&Hopkins,P.E. Localized thin film damage sourced and monitored via pump-probe modulatedthermoreflectance.ReviewofScientificInstruments88,054903(2017).

17. Ma,W.,Miao,T.,Zhang,X.,Kohno,M.,&Takata,Y.ComprehensiveStudyofThermalTransportandCoherentAcoustic-PhononWavePropagationinThinMetalFilm-SubstratebyApplyingPicosecondLaserPump–ProbeMethod.The JournalofPhysicalChemistryC119,5152-5159(2015).

18. Tas,G.,&Maris,H.J.Picosecondultrasonicstudyofphononreflectionfromsolid-liquidinterfaces.PhysicalReviewB55,1852-1857(1997).

19. Morath, C. J., & Maris, H. J. Phonon attenuation in amorphous solids studied bypicosecondultrasonics.PhysicalReviewB54,203(1996).

20. Yilmaz,H.,Pena-Francesch,A.,Shreiner,R.,Jung,H.,Belay,Z.,Demirel,M.C.,Ozdemir,S.K.,&Yang,L.StructuralProtein-BasedWhisperingGalleryModeResonators.ACSPhotonics4,2179-2186(2017).

21. Jung, H., Szwejkowski, C., Pena-Franchesh, A., Tomko, J. A., Allen, B., Ozdemir, S. K.,Hopkins,P.E.,&Demirel,M.C.UltrafastLaser-ProbingSpectroscopyforStudyingMolecularStructureofProteinAggregates.Analyst142,1434-1441(2017).

22. Sariola, V., Pena-Francesch, A., Jung, H., Cetinkaya, M., Pacheco, C., Sitti, M., &Demirel, M. C. Segmented molecular design of self-healing proteinaceous materials.Scientificreports5,13482(2015).

23. Pena-Francesch,A.,Florez,S.,Jung,H.,Sebastian,A.,Albert,I.,Curtis,W.,&Demirel,M.C.MaterialsFabricationfromNativeandRecombinantThermoplasticSquidProteins.AdvancedFunctionalMaterials24,7401-7409(2014).

24. Hu,X.,Kaplan,D.,&Cebe,P.Effectofwateronthethermalpropertiesofsilk fibroin.ThermochimicaActa461,137-144(2007).

25. Meyer, A., Dimeo, R. M., Gehring, P. M., & Neumann, D. A. The high-fluxbackscattering spectrometer at the NIST Center for Neutron Research. Review ofScientificInstruments74,2759-2777(2003).

26. Copley, J. R. D. & Cook, J. C. The Disk Chopper Spectrometer at NIST: a newinstrument for quasielastic neutron scattering studies.Chemical Physics292,477-485(2003).

12

SupplementaryFigure1.a)Photographofourbeam/samplealignment forprobingupwards,throughthetransparentsubstrate,ofoursamples;acalibrationsample(SiO2)isbeingabovetheobjective. b) Photograph of our optical access window to the humidity chamber used. Thechamberrestsonastagewith3-axismovementforback-alignmentofthesample.

SupplementaryFigure2. Theacquired ratioofVin/Voutsignals forTR-n25 in theambientandhydratedcase.

13

SupplementaryFigure3.Thenormalizedresidual(i.e.,theaccuracyofthemodelsfitrelativetotheobtainedthermaldecaycurve)forvaluesofbest-fitforourbi-directionalthermalmodelforvariousmediaon80nmAl/SiO2supports;themediaarea)ourTRfilmsinanambientstate,b)hydrated TR-n25 (the largest thermal conductivity measured in this work), c) PMMA as areference for the sensitivity of TDTR to low thermal conductivity films in a bidirectionalexperimentalsetup,andd)adropofwater.

14

SupplementaryFigure4.Theinner-mostcontour,orthemodel’sbestfit,asshowninFig.S3,forthermalconductivityvaluesofSiO2within10%ofthemeasuredvalue(1.35Wm-1K-1).Asseen,thethermalmodelismostsensitivetothethicknessofthemetaltransducer.

SupplementaryFigure5.Measuredthermalconductivityasafunctionofinverserepeatunits,1/n,forboth80nmAlandAumetaltransducers.

15

SupplementaryFigure6.a)Azoomed-inviewofthemeasuredthermalconductivityasafunctionof real time for TRn25, providing insight to the time-scale associatedwith switching from ahydratedtoambientstate.Note,theambient-to-hydratedswitchisconsistentlywithinthetimeconstant of our lock-in amplifier (300 ms), thus appearing instantaneous. b) Image of ourhumiditychamberatarelativehumidityofapproximately85%.TheinsetshowsaTR-n11sample(supported by an 80 nmAl transducer and 1mm SiO2 substrate underneath) on the opticalwindowinlettothechamber.Asseen,condensationoccursheavilythroughoutthechamber.

16

Supplementary Figure 7. Examples of TDBSmeasurements on n-4 in the ambient (top) andhydratedstates(bottom),withvaluesofbestfittabulatedforeach.

17

SupplementaryFigure8.FTIRamideIanalysisofambientandhydratedTR-n11protein.

SupplementaryFigure9.FTIRamideIFSDanalysisofambientandhydratedTR-n11protein.

18

Supplementary Figure 10. Specific reversing Cpmeasurements on ambient and hydrated TRproteins by TMDSC. Two measurements (solid line for first measurement, dashed line forrepetitions)persampleareplotted.Verticallinedenotedthetemperatureofinterest(20◦C)forthecalculationofCp.