Tracking Local Mechanical Impact in Heterogeneous Polymers with Direct Optical Imaging

Abstract Structural heterogeneity defines the properties of many functional polymers and it is often crucial for their performance and ability to withstand mechanical impact. Such heterogeneity, however, poses a tremendous challenge for characterization of these materials and limits our ability to design them rationally. Herein we present a practical methodology capable of resolving the complex mechanical behavior and tracking mechanical impact in discrete phases of segmented polyurethane—a typical example of a structurally complex polymer. Using direct optical imaging of photoluminescence produced by a small‐molecule organometallic mechano‐responsive sensor we observe in real time how polymer phases dissipate energy, restructure, and breakdown upon mechanical impact. Owing to its simplicity and robustness, this method has potential in describing the evolution of complex soft‐matter systems for which global characterization techniques fall short of providing molecular‐level insight.

Abstract: Structural heterogeneity defines the properties of many functional polymers and it is often crucial for their performance and ability to withstand mechanical impact. Such heterogeneity,h owever,p oses at remendous challenge for characterization of these materials and limits our ability to design them rationally.H erein we present ap ractical methodology capable of resolving the complex mechanical behavior and trackingm echanical impact in discrete phases of segmented polyurethane-a typical example of as tructurally complex polymer.U sing direct optical imaging of photoluminescence produced by as mall-molecule organometallic mechano-responsive sensor we observe in real time how polymer phases dissipate energy,r estructure,a nd breakdown upon mechanical impact. Owing to its simplicity and robustness,t his method has potential in describing the evolution of complex soft-matter systems for whichglobal characterization techniques fall short of providing molecular-level insight.
Understanding the mechanisms that define mechanical stability of polymers is crucial for designing robust materials resistant to wear and failure.A lthough conventional macroscopic techniques such as rheology or dynamic mechanical thermal analysis (DMTA) may be sufficient to characterize most polymers,t he macroscopic picture provided by these techniques often lacks molecular-level resolution. This limitation is particularly valid for functional polymers that are typically non-uniform and heterogeneous on multiple length scales.T ou nderstand properties of materials composed of several distinct or mixed phases ac onceptually different characterization approach would be required.
Acomplex multi-phase material at the core of this work, is ac hain-extended polyurethane (PU)-a copolymer composed of flexible polyether or polyester "soft" blocks and rigid "hard" blocks that form ah ydrogen-bonded crystalline phase ( Figure 1). [1] Theu biquity of PU in practical applications stems from the great variability of their properties, which can be achieved by tuning the chemical nature and relative content of different phases.S uch diversity,h owever, comes at the expense of the drastically increased structural complexity of PUs.A sac onsequence,alocal, phase-and component-specific picture of their behavior under mechanical stress remains at opic of intense debate eight decades after the first PU was prepared.
Seeking to overcome the limitations of global characterization techniques lacking molecular-level resolution, several force-responsive molecular probes have been developed over the last decade.W ith few exceptions utilizing nanoparticle moieties, [2] them ajority of such probes are small organic fragments built into the polymer main chain. Upon exposure to external force,t hese organic moieties,o ften termed as mechanophores,c an undergo mechanochemical activation [3] producing aspecific response,for example,luminescence [4] or ac hange in absorbance and fluorescence characteristics. [5,6] Alternatively,am echanical stimulus can alter the arrangement of fluorescent mechanophore within the polymer matrix, which results in ad ramatic change of their emission properties. [7] Thevast majority of these mechanophores relies on transformations with no or limited reversibility.F or this reason, common mechanophores can only probe the state of the polymer system before and after the mechanical impact while the dynamic evolution of the polymer system overtime remains beyond their reach.
Herein we address this challenge and develop as imple and straightforward methodology that allows the direct monitoring of the real-lime evolution of discrete phases in ac omplex polymer system under mechanical stress.I nstrumental to this work was the development of new robust organometallic mechanophores that combine high aerial stability and facile response with straightforward chemistry for their incorporation into distinct phases of polyurethanes. Relying on direct optical imaging as the primary analysis technique,the methodology holds great potential for general use in soft matter characterization.
Thep reparation of the new mechanophore and its selective incorporation in PUs was the first objective of our study.W eb ased our mechanophore design on an ew class of recently developed phosphorescent complexes that were shown to reversibly respond to mechanical stress (CuN 4 , Figure 1b). [8,9] Themechanism governing their force-induced response,depicted in Figure 1b,isconceptually different from those found in common organic mechanophores. [9] Instead of bond scission and related irreversible transformations,t hese copper halide complexes vary their photoluminescence (PL) via aforce-induced restriction of their intramolecular isomerization that represents an on-radiative relaxation process. Slower isomerization, induced by the restriction on the polymer chain mobility under stress,r esults in brighter phosphorescence with higher quantum yields and longer PL lifetimes.F ast isomerization, that is typical for highly mobile polymer chains results in lower PL intensities.T his response mechanism would be suitable to track force-induced phenomena in PUs in real time but falls short of being practically useful due the poor air stability of the Cu I halide complexes.
Forp ractical reasons,the target stress-responsive mechanophores based on the CuN 4 core should be robust enough to withstand continuous exposure to air and UV light. A common strategy for stabilizing Cu I towards oxidation is the use of strong donor auxiliary ligands,for example,phosphines or N-heterocyclic carbenes (NHC). To identify potential mechanophores we evaluated as eries of Cu I complexes with different NHC ligands using DFT calculations (S6 of Supporting Information) targeting complexes with the isomerization mechanism identical of that of previously described Cu halides.S pecifically,w ea nalyzed how mechanophore isomerization is influenced by substitution of the halide ligand in the parent halide CuN 4 by NHC ligands with different steric bulk. We identified one of the smallest known NHC ligands derived from N,N'-dimethylimidazolium carboxylate ( Figure 1) to be the suitable auxiliary ligand that would not alter mechanophore isomerization mechanism and therefore its force response.I nl ine with predicted behavior,r eference complex 1,prepared to validate our analysis,indeed exhibits isomerization dynamics identical to that of CuN 4 and remains structurally similar to the parent complex according to the single crystal X-ray diffraction data (see Figure 1C and Section S5 of Supporting Information). This behavior,particularly important for mechanophore selection, suggests an isomerization mechanism of 1 is identical to that of parent halide mechanophores [9] involving intermediate formation of 4-coordinate Cu centers,t hus,b eing associative.F inally, complex 1 is remarkably stable and withstands nearly aw eek long exposure to air in dichloromethane solution, making it suitable for further practical use as am echanophore.
We further incorporated this phosphorescent motif into hard and soft phases of polyurethane samples.Inspired by the strategy reported by Braun and co-workers [5] we developed ao ne-pot protocol targeting the macrocyclic ligand L ( Figure 1) to prescribed locations.N amely,w ep re-formed two building blocks depicted in Figure 1a"soft" one terminated with polyether groups (pTHF) and a" hard" one terminated with hexamethylene diisocyanate (HDI) extended with 1,4-butanediol (BDO) to direct the mechanophore precursor to soft or hard PU domains,r espectively.R espective PU samples were denoted as SP/HP for the type of PU phase hosting mechanophore and given the index "40" indicating the molar fraction of the BDO chain extender. NMR spectroscopy confirmed the incorporation of preloaded amounts of ligand L in all PUs and their identical composition regardless of the phase labelled (See Supporting Informa- tion). FTIR spectroscopy confirmed no impact of mechanophore incorporation on the hard phase of SP40 samples and indicated am arginal increase in the relative content of disordered carbonyl groups in HP40 further confirming that the labelling impacted specific synthetically targeted phases (See Figure S29 in the Supporting Information). Size exclusion chromatography indicated similar molecular weights of all samples ranging from 70 to 90 kDa with nearly identical polydispersity of 1.9-2.1. Finally,the polymer-bound ligand L was complexed with Cu I following the procedure developed for the synthesis of complex 1 to produce emissive polymers with 1 HNMR spectra nearly identical to those of the reference complex 1 (See Section S3 of Supporting Information). Thel atter,i mportantly,c onfirms the identity of mechanophore embedded in the PU material and suggests no ligand exchange between mechanophore and polymer matrix taking effect.
Copper-containing PU films fabricated by mold casting were photoluminescent and emitted at around 580 nm upon excitation with UV light. Importantly,t he emission of PU films was identical to that of reference complex 1 in solution but not in the solid state.This confirms that mechanophore in the PU is well dispersed and not aggregated as the aggregates would produce emission similar to that of crystalline 1, which is significantly blue shifted to approximately 555 nm. Films of Cu-HP samples incorporating the mechanophore in the hard domain exhibit nearly three-fold longer PL lifetimes compared to their soft phase labelled counterparts Cu-SP suggesting that Cu-SP indeed contains mechanophore in the phase with significantly higher chain mobility as expected when the mechanophore is embedded in the soft phase of PU.
Once discrete phases of PU samples were labelled with mechanophore,w ef ocused on tracking the differences in force-induced behavior of those phases.T odothat, PU films were subjected to uniaxial extension with simultaneous detection of the photoluminescence intensity produced by these films.F or tensile tests the isotropic PU films were prestretched to 250 %o ft heir initial length to prevent necking and associated changes in sample geometry during the imaging experiment.
To observe both deformation and stress relaxation kinetics we adopted the testing protocol ( Figure 2) similar to that used by Braun [5] and Boyce. [10] Namely,s amples were extended with ac onstant rate of 5o r5 0% L 0 s À1 to produce af ixed (7.5 MPa) increase in stress at each loading step followed by a5second hold allowing for stress relaxation to be monitored. Afterwards,t he loading/relaxation was repeated until the sample failed. Ty pical curves produced in these tests are depicted in Figure 2.
As expected, external stress applied to PUs reduces mechanophore isomerization rate and leads to an increase of the PL intensity during all tests.The evolution of PL response during stress relaxation step,however, depended strongly on the location of the phosphorescent probe and strain rates applied during the elongation. Slow deformation of the Cu-SP40 containing mechanophore in the soft phase produces an immediate increase of PL intensity that further decays when the sample is allowed to relax (Figure 2A). Simultaneously with the decrease of the PL intensity,t he global mechanical stress decreases.T he drop on the PL intensity of the soft phase embedded mechanophore suggests that gradual redistribution of local stresses within the PU sample from the soft phase to the hard phase takes place during the relaxation stage.W et herefore assume that even at low loading rates deformation of the soft phase is impacted by the deformation of the hard phase that accounts for decrease in the detected PL response.
Interestingly,mechanophore localized in the hard domain as in Cu-HP40 produces ad ifferent PL profile that contains as hort plateau of approximately 1.2-1.8 sd uration that precedes the PL decrease associated with the same stress redistribution phenomena that we observed in Cu-SP40. Because the appearance of plateau was specific to the hard phase we assigned it to the reorientation of the hard phase along the elongation direction that was observed experimentally (See Figure 2and Figure 3and Supporting Information) and reported for other common PUs. [1,11] Ac ommon feature of slow deformation in either HP or SP samples is the immediate increase of PL intensity in response to deformation that indicates an immediate restriction of the chain mobility upon loading regardless of the mechanophore location.
Fast deformation of the hard phase labelled Cu-HP40 produces ad ramatically different PL response.U pon deformation the PL intensity instantaneously drops while subsequent stress relaxation and associated decay of global mechanical stress is accompanied by am ono-exponential increase of PL intensity ( Figure 2D). This sharply contrasts the behavior of the PL probe in Cu-SP40 at the same deformation rate ( Figure 2C)w here no pronounced droprecovery behavior was observed.
Up to this point, our observations suggest that at high deformation rates the hard phase restructures differently with mechanophore becoming highly mobile upon mechanical impact and being slowly immobilized over time.Importantly, the extent of plastic deformation in Cu-HP40-5/50 is negligible and the distinctly different hard phase behavior takes place within the same range of applied stresses and strains to those registered throughout all tests (see expansions in Figure 2). Taken together,t hese data suggest that observed PL response differences in Cu-HP40-5/50 are specific to the hard phase of PU and stem solely from the elongation rate dependence.H aving observed such ap rofound difference in the PL response for Cu-HP40 we further aimed at assigning these differences to physical events taking place in PUs under stress.T his would ultimately allow linking the data provided by our methodology to the knowledge framework developed for PUs using conventional characterization techniques.
Them ajor structural change occurring in PUs upon deformation is associated with the reorientation and segmentation of the crystalline hard phase. [1,11] Using linear dichroism IR spectroscopy we found that slow deformation of Cu-HP40 leads to al arge increase of hard phase orientation along the stretching direction. Theo rder parameter in that case increases from the initial value of 0.37 for pre-stretched samples before tensile tests to approximately 0.7 (Figure 2, top). Thei ncrease of the order parameter is in line with the expected behavior of PUs,w hich are known to produce oriented morphologies upon deformation though the slow rotation of hard domains that orients them along the principal deformation axis. [11] Theobserved mechanophore behavior in the slow deformation experiment is indeed consistent with this conventional deformation mechanism. Theincrease of PL intensity and the following slow decay can, therefore,b ea ssigned to two distinct events in PUs upon deformation:1)restriction of mobility of extended polymer chains (increase of PL) and 2) rotation of hard domains that allows for higher chain mobility as the stress relaxation takes effect (slow decrease of PL).
Aiming to identify the origin of adifferent PL response profile observed in the case of fast deformation we extended our structural studies to Cu-HP40-50 sample. Interestingly,wef ound no apparent difference in the close range packing in Cu-HP40-5/50 examined by wide angle X-ray scattering (WAXS,s ee S7 in Supporting Information). IR spectroscopy,s imilarly indicated no detectible change in the overall hard phase composition in all Cu-HP40 samples as evidenced by identical absorbance in the characteristic regions for urethane NH and C=Ov ibrations ( Figure 3B and section S7 in Supporting Information). Theh ard phase alignment in Cu-HP40-50, however was found to be strikingly different from that of Cu-HP40-5.N amely,t he fast deformation did not affect the hard phase order parameter that remained at the value of 0.37 after the tensile tests.This suggests that hard domains in PU were unable to rotate and re-orient upon the fast deformation that resulted into aP Uw ith an entirely different morphology.T his notion is in part supported by atomic force microscopy data ( Figure S28 in the Supporting Information) indicating am ajor restructuring of the PU surface as the PU samples are subjected to mechanical impact.
Further structural differences in Cu-HP40-5/50 were revealed by the 2D-SAXS analysis depicted in Figure 3A.All samples after the pre-stretch and tensile tests exhibited at ypical 4-spot pattern often referred to as "chevron pattern" common among HDI-based PUs. [12] This scattering pattern indicates the presence of two groups of hard domains tilted at identical angle with respect to the deformation direction as depicted in Figure 3A.I n contrast to the linear dichroism IR data suggesting the reorientation of the urethane functional groups in slow deformation experiments and the lack of thereof in fast deformation tests,SAXS data indicates that hard domains in Cu-HP40-50 produced by the fast loading are actually more ordered macroscopically (See Figure 3B). Namely,amore defined "chevron" pattern is revealed by the azimuthal scattering intensity distribution data and adetectible shift was observed for the broad peak in SAXS profile at ca. 0.08 À1 corresponding to the hard domain repeating unit. Altogether,these findings confirm that deformation rate dependence of the mechanophore response in Cu-HP40 indeed originates from differences in hard phase deformation mechanism.
Our data consistently links the differences in PL response to the development of distinct PU morphologies.F igure 3C summarizes our proposed mechanism for their formation. In line with previous observations and our structural characterization data, slow deformation allows for the rotation of hard domains in PU and their alignment with the stretching direction. This deformation has no major effect on the hard phase morphology and preserves its macroscopic structure as confirmed by 2D-SAXS.Asthe PL response in hard and soft phase of Cu-SP/HP40 was similar at low elongation rates,we assume that deformation is largely dominated by the soft phase uncoiling and subsequent relaxation due to hard phase rotation ( Figure 3C).
In contrast to this,t he fast deformation regime is apparently destructive for the hard phase of the PU at least during the initial deformation step.T his interpretation is supported by the sharp drop in the PL intensity during the fast loading in Cu-HP40-50 that suggests that the mechanophore is mobilized upon the impact. This can proceed either via ap ull-out of individual hard segments containing the mechanophore or ap artial breakdown of the hard phase crystallites.I mportantly,r egardless of which of these phenomena takes place,itdoes not induce hard domain rotation, suggesting that mechanical impact restructures the hard phase without allowing it to diffuse and reorient macroscopically. Based on these considerations,wecome to conclusion that the hard phase deformation should involve the scission of weak hydrogen bonds between individual hard segments.
As the hydrogen bonds are non-covalent, they can form and break reversibly, [13] allowing fragmented hard domains to re-associate when deformation stops.T his behavior could indeed be responsible for the recovery of PL response observed during fast deformation experiments.C onsidering that the recovery of PL signal in Cu-HP40-50 is strictly monoexponential we conclude that the PL response profile in fact describes as ingle process that is likely crystallization of the hard phase with an observed rate constant of 0.93 AE 0.1 s À1 calculated from exponential PL recovery traces-a value corresponding to aprocess lifetime in the order of one second.
Although highly unusual to occur at room temperature, the hard phase crystallization is consistent with our experimental data-in PUs crystallization is af ast process with at ime scale of multiple seconds at room temperature. [14] Direct analysis of non-isothermal crystallization kinetics of Cu-HP40 performed using DSC,places an estimated crystallization rate constant at 4.7 s À1 at 25 8 8C( S4 in Supporting Information)-a value similar to that observed in imaging experiments.A lthough the H-bond scission in PUs likely impacts aminor fraction of the hard phase,our methodology appears to have sufficient sensitivity to produce ar eadily detectible response to these transient physical events that otherwise remain undetected by techniques lacking temporal or spatial resolution.
In conclusion, our work provides ad irect observation of diverse and distinctly different phase behavior patterns within as ingle polymer family and offers an unprecedented insight into the mechanically induced evolution of complex polymer system. Theresponse of robust mechanophores has been used as ad iagnostic tool to identify the deformation mechanisms leading to the development of distinct polymer morphologies  Figure 2. B) Total scattering intensity and azimuthal intensity distribution plots for scatteringdata depicted in panel (A) and IR spectroscopy data for these samples. C) Proposed deformationm echanism of Cu-HP40 at slow and fast elongation regimes assigned to observed PL profiles. that can hardly be predicted using global characterization techniques.S ince the response of new mechanophores originates at the molecular level, it allows us to describe force-induced phenomena locally.Atthe same time,the nondestructive nature of mechanophore action allows these dynamic phenomena to be tracked in real time.This methodology is efficient in dealing with the case of structural complexity in polyurethanes.W e, therefore,see no limitation for it to become acommon tool assisting materials scientists in resolving soft matter complexity in the near future.

Experimental Section
Full synthesis and characterization data and testing methodology description can be found in Supporting. CCDC 1849321 contains the supplementary crystallographic data for this paper.These data can be obtained free of charge from TheC ambridge Crystallographic Data Centre.