Mechanically-Driven Vase-Kite Conformational Switch in Cavitand Cross-Linked Polyurethanes.

The eligibility of tetraquinoxaline cavitands (QxCav) as molecular grippers relies on their unique conformational mobility between a closed (vase) and an open (kite) form, triggered in solution by conventional stimuli like pH, temperature and ion concentration. In the present paper, the mechanochemical conformational switching of ad hoc functionalized QxCav covalently embedded in an elastomeric polydimethylsiloxane and in a more rigid polyurethane matrix is investigated. The rigid polymer matrix is more effective in converting mechanical force into a conformational switch at the molecular level, provided that all four quinoxaline wings are covalently connected to the polymer.

Introduction

The past few decades have seen extraordinary progress in the construction of molecular machines and switches.1 Much effort has been devoted to the development of molecular architectures capable to perform mechanical‐like movements (output) as a consequence of external stimuli (input).2 Among the variety of switches that mimic different macroscopic objects, molecular grippers stand out for the uptake/release of molecular objects.3 To act as a gripper, a molecule should close and open upon external stimulation to grab and release molecular objects.

Scaffolds that have emerged as particularly promising for the design of these molecular tools are tetraquinoxaline cavitands (QxCav).4 These resorcinarene‐based macrocycles are obtained through the four‐fold bridging of the resorcinol scaffold with 2,3‐dichloroquinoxaline moieties, which delimit a deep cavity. QxCav shows the ability to reversibly switch between two spatially well‐defined conformations: an expanded kite (C2v symmetry) and a contracted vase (C4v symmetry), according to the position of 1,4‐diazanaphtalene ‘flaps’, that can occupy either the equatorial or the axial positions, respectively (Figure 1).5 Interestingly, since all intermediate conformers are energetically disfavoured, only these two discrete forms exist.6 The energy barrier for a single flap interconversion has been estimated in 7.6 kcal mol−1 by DFT calculations.7 The vase‐kite interconversion has been extensively investigated in solution via different stimuli, such as pH,8, 9 temperature,10 Zn2+ coordination11, 12 and redox.3, 13

Figure 1

Vase‐kite conformational interconversion for a generic cavitand in solution (above) and in polymeric matrix (below).

The effective integration of these molecules into functional nanodevices, however, requires a precise control of their conformation on surfaces or in the solid‐state. Recently, we proved the retention of pH‐driven conformational switch of QxCav covalently included in poly(butyl methacrylate) (PBMA) and polydimethylsiloxane (PDMS) matrices.14, 15

The incorporation of these gripper‐like macrocycles in polymers paves the way for the trigger of vase‐kite switching with unprecedented stimuli. In fact, when embedded in polymer matrices, QxCav can be considered as a “conformational” mechanophore, a term that indicates a molecule that undergoes a predictable conformational modification triggered by a mechanical force. Most of the mechanophores reported so far are characterized by the presence of strategically weakened covalent bonds, in the form of strained rings or isomerizable bonds, prone to react once mechanical forces are transferred to them from the polymer chains.16, 17, 18, 19, 20 Spiropyran‐based (SP) mechanophores are the most employed stress‐sensitive units reported, thanks to their highly effective colour‐generating nature.21

Over the past few years, thanks to the progress in mechanophore design, a broad range of mechanically‐triggered transformations was attained.22, 23, 24, 25, 26, 27, 28 Among these, particularly attractive are mechanophores that do not require scission of covalent bonds for the activation and change of their properties.29, 30, 31, 32, 33 In the case of rotaxanes, the activation process does not involve any bond cleavage, leading to a low activation energy of the process.30

In the study reported herein, we extensively investigated the vase‐kite switching of tetraquinoxaline cavitands in polymers through the application of tensile strain. To achieve this goal, we applied the tool kit of polymer mechanochemistry to induce the conformational switching.

Results and Discussion

Considering the essential role played by the nature of the polymer in terms of the efficiency of the mechanochemical transduction, we investigated two matrices (Scheme 1), an elastomeric polydimethylsiloxane (PDMS) and a more rigid polyurethane (PU).

Scheme 1

PDMS (a), (L)PDMS (b), and PU (c) polymers containing covalently embedded cavitands.

Tetraquinoxaline cavitands were obtained via a convergent approach where either functionalized or non‐functionalized quinoxalines were introduced on resorcinarenes or partially bridged cavitands. Since DFT calculations on cavitands suggested that all intermediate states between the two discrete conformers are energetically disfavoured, we initially assumed that the application of a tensile force to two quinoxaline wings in distal position would elicit the opening of the other two, thus promoting the vase‐kite interconversion. Based on this hypothesis, we performed the synthesis of distal‐difunctionalized tetraquinoxaline cavitands bearing suitable moieties for the covalent incorporation into the polymeric network. In particular, terminal vinyl groups were introduced to ensure the embedding into the PDMS matrix, while hydroxy groups were exploited for the PU one. The anchoring site for polymer grafting was connected to the cavitand core via easily accessible ester linkages and through aliphatic spacers, in order to prevent a possible decrease in reactivity for steric effects. Moreover, since any modification at the upper rim walls might destabilize the vase conformation, we exclusively functionalized the quinoxaline walls in position 6 or 7, thus minimizing steric hindrance.

For both matrices, specimens with the corresponding tetrafunctionalized cavitand were prepared and mechanically tested, in the hypothesis that the use of a QxCav molecule acting as cross‐linker into the polymer could result in a more efficient channelling of the applied force to the mechanophore unit. Also in this case, cavitand walls decoration was performed via a convergent approach through a bridging reaction with pre‐functionalized 2,3‐dichloroquinoxalines.

The above‐described cavitands were employed as co‐monomers in the preparation of polymer specimens. Furthermore, in order to irrefutably correlate the applied mechanical stimulus to the desired response, excluding any other possible activation mechanism (e. g. thermal activation), the corresponding control specimens, with physically dispersed, mono‐ or non‐quinoxaline anchored QxCav, were prepared and mechanically tested.

Alkene‐Functionalized Cavitands for PDMS

As demonstrated by Craig and co‐workers,34 polysiloxanes are suitable matrices for mechanochemical applications since, besides their well‐known high stretchability and mechanical strength, they present an efficient transduction of applied forces to embedded mechanophore units. The conventional protocol for the covalent incorporation of stress‐responsive elements into PDMS relies on hydrosilylation reaction, requiring the presence of terminal double bonds.

Accordingly, cavitand C2A (Scheme 2), bearing two distal alkenyl groups suitable for mechanochemical transduction, was prepared following a previously described procedure.14 Monofunctionalized cavitand C1A was synthesized by bridging ω‐decene‐2,3‐dichloroquinoxaline 1 14 on the corresponding triquinoxaline scaffold, obtained from the unsubstituted QxCav C by selective monoquinoxaline excision reaction with catechol under basic conditions.35 The four‐fold bridging of dichloroquinoxaline on a resorcinarene scaffold afforded the tetrafunctionalized QxCav C4A as racemic mixture of two enantiomers with C4 symmetry. This regioselectivity can be explained with steric effects; in fact, after the first statistical bridging event, hindrance of the introduced aliphatic chain influences the following insertions, thus leading to the functionalization in alternate positions at expense of the overall yield. The high symmetry, revealed by very sharp signals in the 1H NMR spectrum, was unambiguously confirmed by single crystal X‐ray diffraction (Figure S1). To provide an analogous cross‐linker in control films, QxCav C4A, bearing four double bonds at the lower rim, was synthetized adapting previously reported procedures.36, 37

Scheme 2

Overview of alkene functionalized employed in PDMS polymers preparation. Unsubstituted cavitands, reported in the inset, have been prepared as reference compounds.

Compounds C1A, C4A and C4A were characterized by 1H NMR and MALDI‐TOF spectrometry. NMR spectroscopy is diagnostic of cavitand conformation; indeed, it is widely reported in the literature4 that the resonance of methine protons in solution is highly conformation‐dependent, with a difference in chemical shift of more than 2 ppm between vase (≈5.6 ppm) and kite (≈3.9 ppm). Thus, the presence of a signal at around 5.6 ppm in all cases confirmed the retention of the vase conformation. Finally, cavitand C , bearing four methyl groups in apical position that force it in kite conformation by steric effects, was synthesized according to previously reported procedures.38

PDMS Functionalization and Mechanical Tests

RTV 615, the selected matrix for PDMS specimens, is a two component (base and curing agent, typically mixed in a 10 : 1 ratio) silicone elastomer comprising of vinyl terminated poly(dimethylsiloxane), poly(methylhydrosiloxane‐codimethyl siloxane) copolymer and a platinum catalyst. QxCav C1A, C2A, C4A and C4A were covalently incorporated into the matrix via hydrosilylation protocol to give specimens PDMS‐C1A, PDMS‐C2A, PDMS‐C4A and PDMS‐C4A, respectively (Table S2, concentrations from 0.05 to 0.7 % w/w). Effective covalent embedding was confirmed by Soxhlet extraction with chloroform, performed on PDMS‐C2A as representative (Figure S2a). The extracted polymer and the organic phase were analysed by UV‐Vis spectroscopy: the absorption pattern of the cavitand was observed only in the polymeric specimen, while it was not present in the organic phase. This experiment, although being crucial for proving the covalent incorporation of cavitands in the polysiloxane network, does not demonstrate the effective reaction of all their functionalities. Unfunctionalized QxCav C and C, instead, were physically dispersed in PDMS; the resulting PDMS‐C and PDMS‐C specimens provided a reference of the two conformers in the matrix (Table S3). In this case Soxhlet extraction with chloroform caused the almost complete removal of dispersed cavitands from the specimens, as evidenced by the UV‐Vis spectra of the two phases (Figure S2b).

Comparison between the UV‐Vis spectra of C and C in solution and in PDMS revealed the retaining of the optical properties at the solid state, with two maxima of absorption at 318 and 330 nm for vase QxCav, and at 335 and 348 nm for the kite form (Figure S3). Moreover, UV‐Vis spectroscopy was already used to confirm the protonation‐driven interconversion for PDMS‐C2A, suggesting that, even when covalently embedded into the polymer matrix, cavitands possess enough free volume to open up into the kite conformation.14 In this work, cavitands interconversion upon tensile stress was tested. Specimens containing mono‐ and di‐functionalized QxCav C1A and C2A at different concentrations were uniaxial stretched and UV‐Vis spectra were recorded at progressively increasing applied force (Figure 2).

Figure 2

Digital photographs of the device employed for mechanical tests: relaxed sample (top) and loaded sample at 300 % stretching (bottom).

In particular, PDMS‐C1A samples were tested to isolate the effects of bidirectional strain. In all cases, no maxima variation was observed. Figure 3a, showing the experiment for PDMS‐C2A, is representative of the general behavior observed for all tested specimens. Since the only visible effect is the λ‐independent decrease in absorbance due to the thinning of polymer films upon elongation, we can conclude that the visual monitoring of vase‐kite force‐induced switching via UV‐Vis spectroscopy was not possible.

Figure 3

UV‐Vis spectra of (a) PDMS‐C2A (0.5 % w/w, Table S2 entry V) and (b) PDMS‐C4A (0.3 % w/w, Table S2 entry X) upon progressive elongation. Vertical grey lines correspond to maxima in UV‐Vis spectra for CHCl3 solution of vase and kite conformers.

To test the hypothesis that all four quinoxaline wings need to be stretched in order to induce interconversion, active and control specimens with tetrafunctionalized QxCav C4A and C4A, respectively, were prepared. Unlike QxCav C4A, cavitand C4A, bearing the anchoring sites at the lower rim, is not expected to be affected by any conformational change upon stretching (Figure S4). The UV‐Vis spectrum of cavitand C4A in solution shows a slightly shifted maximum (323 nm instead of 318 nm for QxCav C), while the diagnostic shoulder attributed to the protonated kite form remained unaltered after functionalization (Figure S5). PDMS‐C4A specimens were uniaxially stretched; once again, UV‐Vis spectra recorded at progressively increasing applied force did not show any maxima variation upon elongation (Figure 3b). Considering the high number of cross‐linking sites in pristine elastomeric PDMS, we supposed that the unresponsiveness of our system could be justified by a complete or partial prevention in the transmission of mechanical energy to the mechanophore.

In order to maximize the channelling of tensile force to the cavitand, we modified the polymer backbone by replacing the conventional RTV 615 curing agent, consisting of the poly(methylhydrosiloxane‐codimethylsiloxane) copolymer acting as cross‐linker, with a ditopic H‐terminated‐poly (dimethylsiloxane). Thus, in the resulting polymer (L)PDMS‐C4A, the unique cross‐linking agent is represented by tetrafunctionalized cavitand C4A. Different polymeric samples were prepared, both varying the ratio between vinyl‐terminated and H‐terminated pre‐polymers and the amount of cross‐linker (Table S4), to find the optimal conditions in terms of mechanical properties of the resulting specimens. Tetra‐ω‐alkene functionalized QxCav C4A, bearing double bonds at the lower rim, was used as cross‐linker in the preparation of control specimens (L)PDMS‐C4A, in order to provide a reticulation similar to the active ones. Mechanical testing on (L)PDMS‐C4A failed to evidence a clear maxima variation in the UV‐Vis spectra, although an increase in the absorbance at 335 nm was observed against the trend of thinning of the specimen upon elongation (Figure 4a). Interestingly, the apparent increase in the adsorbance, occurring at the wavelength of the first maxima for the kite conformer, was not observed in control specimens (L)PDMS (Figure 4b).

Figure 4

Normalized UV‐Vis spectra of (a) (L)PDMS‐C4A (1 % w/w, Table S4 entry VI) and (b) (L)PDMS‐C4A (1 % w/w, Table S4 entry IX) upon progressive elongation. Vertical grey lines correspond to maxima in UV‐Vis spectra for CHCl3 solution of vase and kite conformers.

A mechanically‐driven vase‐kite interconversion for PDMS‐embedded QxCav was not clearly detected. A possible reason could be the low degree of force‐elicited conversion, already reported for other mechanophores,39, 40, 41 which makes the detection of cavitands conformational switching via UV‐Vis spectroscopy very difficult with respect to a visible colour change. Thus, a more efficient system ensuring a higher degree of conversion for the mechanophore unit is needed. Therefore, we decided to replace the elastomeric PDMS with a more rigid matrix, such as PU.

Hydroxyl Functionalized Cavitands for PU

Following a protocol reported in the literature,41 the covalent inclusion of cavitands in PU matrix was obtained via step growth polymerization. The mechanophore units used for this study are summarised in Scheme 3. In particular, mono‐ω‐hydroxy QxCav C1H and bis‐ω‐hydroxy QxCav C2H were employed as co‐monomers in the step growth polymerization, while tetra‐ω‐hydroxy QxCav C4H acts as cross‐linker in active specimens. Finally, tetra‐ω‐hydroxy velcrand C was designed as a reference for the covalently embedded kite form into PU. The convergent synthetic approach, analogously to the previous one, was based on the bridging reaction of ad hoc functionalized dichloroquinoxalines onto variously substituted cavitand scaffolds. In detail, 2,3‐dichloroquinoxaline 2, bearing a hydroxy group protected as tert‐butyl dimethyl silyl ether (TBDMS), was prepared from the easily accessible 2,3‐dichloroquinoxaline‐6‐carbonyl chloride through a convenient acyl substitution with mono‐protected hexanediol under mild conditions. Nucleophilic substitution of 2 by triquinoxaline and AC‐diquinoxaline QxCav hydroxyl groups, followed by deprotection of the TBDMS group, afforded respectively cavitands C1H and C2H, suitable for the covalent incorporation in PU (see SI for the detailed synthetic procedures). Tetrafunctionalized cavitand C4H, designed for acting as unique cross‐linker in the linear polymer network, was obtained by the four‐fold bridging of the corresponding resorcinarene with 2, under microwave irradiation, and subsequent deprotection. The high resolution of 1H NMR signals for C4H confirmed that, also in this case, only the two enantiomers bearing alkyl functionalities in alternate positions were obtained. This regioselectivity, due to steric effects, justifies the low yield of the bridging reaction. Finally, QxCav C, forced in the kite conformation, was obtained from the corresponding resorcinarene scaffold via four‐fold bridging of 2 and subsequent deprotection. Unlike QxCav C4H, the broad signals in the 1H NMR spectra suggest that both cavitand C and its TBDMS‐protected precursor 6 were obtained as an inseparable mixture of constitutional isomers. The lack of regioselectivity for the bridging reaction is probably due to the expanded conformation of the resulting QxCav, where the absence of steric effects does not impose any geometrical constraint to the subsequent insertions of dichloroquinoxaline 2.

Scheme 3

Overview of hydroxy functionalized cavitands employed in PU polymers preparation.

PU Functionalization and Mechanical Tests

The selected matrix is a relatively rigid polymer obtained by step growth polymerization between poly(tetramethylene glycol) (PTMG, Mn=1000) and isophorone diisocyanate, followed by the addition of hexamethylene diisocyanate (HDI) as chain extender. The use of isophorone instead of the more common methylene diisocyanate meets the need of optical transparency for the resulting films, thus facilitating the visualization of vase‐kite switching via optical spectroscopy.

Firstly, unfunctionalized QxCav C and C were physically dispersed in the polymer and the resulting specimens PU‐C and PU‐C provided a reference of the two conformers in the solid state. UV‐Vis spectroscopy confirmed that the optical properties are essentially retained when moving from solution to the solid state (Figure S7). QxCav C1H, C2H, C4H and C bearing hydroxy groups, were employed as co‐monomers in the step growth polymerization. Thus, tetra‐ω‐hydroxy functionalized QxCav C4H was employed in the preparation of force‐responsive films PU‐C4H, while functionalized QxCav C1H, C2H and C, together with QxCav C, were used for control films PU‐C1H, PU‐C2H, PU‐C and PU‐C (Table S5‐S6, concentrations from 0.01 to 0.5 % w/w). Differential scanning calorimetry (DSC) analysis performed on PU‐C4H (Figure S6) revealed a glass transition temperature (Tg) of −36 °C.

Cavitands interconversion upon mechanical stress was tested via UV‐Vis absorption spectroscopy. The stretching of PU‐C4H induces the rise of a shoulder at 348 nm, corresponding to the second maximum of the kite conformer. Figure 5a, illustrating the experiment for PU‐C4H (0.3 % w/w, Table S5 entry VII), is representative of the general behaviour observed for all tested specimens, where changes in the UV‐Vis spectra become sizable for a degree of stretching of about 200 %. The same mechanical tests performed on control specimens failed to evidence comparable changes in the UV‐Vis spectra (Figure S8); Figure 5b, showing the spectra recorded for PU‐C upon elongation, resumes the general behaviour observed for all control samples (Figure S9 and S10).

Figure 5

Normalized UV‐Vis spectra of (a) PU‐C4H (0.3 % w/w, Table S5 entry VII) and (b) PU‐C (0.3 % w/w, Table S6 entry II) upon progressive elongation. Vertical grey lines correspond to maxima in UV‐Vis spectra for CHCl3 solution of vase and kite conformers.

In order to confirm unambiguously the occurrence of the mechanically‐triggered conformational switching, fluorescence measurements were performed. Preliminary studies in solution on QxCav C and C suggested a solvatochromic nature of the emission band, which redshifts with increasing solvent polarity (Figure S11). When dissolved in solvents of medium‐high polarity, the fluorescence band of the vase conformer is located at roughly the same position as the band of the kite conformer, in low‐polarity solvents the band of the vase is located at shorter wavelengths. This suggests that, in low‐polarity solvents, the fluorophoric units in the vase conformer are less solvated than in the kite conformer. This can be rationalized assuming the formation of an ‘intramolecular quadrimer’ in the contracted conformer. Analogous behaviour was observed comparing PU‐C with PU‐C and PU‐C4H with PU‐C (Figure S12): in both cases, the emission band of the vase conformer is blue shifted with respect to the kite one (15 nm and 30 nm, respectively), suggesting that the vase conformer is less solvated than the kite one also in the solid state. Remarkably, while the contracted form shows a redshift of 10 nm when passing from being physically dispersed (PU‐C) to being covalently linked to the polymeric network (PU‐C4H), the expanded one moves of 30 nm (PU‐C vs PU‐C). This observation, supported by concentration dependence studies in solution (Figure S13), strongly suggests that the kite compound, when dispersed, tends to form intermolecular aggregates, resulting in an emission behaviour closer to the ‘intramolecular quadrimer’ situation showed by the vase compound in low‐polarity media. On the other hand, covalent embedding into the polymeric structure prevents the formation of intermolecular aggregates, leading to an emission behaviour more similar to the one observed in medium‐high polarity solvents. By contrast, the aggregation effect for the vase conformer is essentially non‐influenced by the nature of cavitand incorporation into PU, having mainly an intramolecular origin.

Once verified that the four‐fold opening of the quinoxaline wings leads to a bathochromically‐shifted emission in PU, mechanical tests on PU‐C4H were performed. Because of the thinning of the stretched specimens, a quantitative evaluation of the absolute emission intensity upon interconversion was considered unreliable. Thus, only the shape of the emission spectra is discussed. As expected, emission spectra of control films PU‐C did not reveal any significant change upon elongation (Figure S14); by contrast, in agreement with UV‐Vis measurements, a change in the emission profile for PU‐C4H was observed (Figure 6a).

Figure 6

a) Emission spectra representative of the most common behaviour observed for PU‐C4H (0.01 % w/w, Table S5 entry X) upon tensile elongation (thick black and red lines). Fitting of the curves (results as thin black and green lines) as sums of two Gaussian bands (filled areas, dashed for the vase and solid for the kite conformers). b) Same as panel a, but reporting the behaviour of a few selected specimens. c) Emission spectra of PU‐C (0.01 % w/w, Table S6 entry III, blue line) and PU‐C (0.01 % w/w, Table S5 entry XII, red line), reported as references.

Interestingly, mechanical tests performed on several specimens pointed out a different behaviour. Figure 6a resumes the most common behaviour observed in active films: the emission spectrum of the pristine film (black thick line) shows a single maximum at 390 nm and, upon stretching, the emission band broadens in the long‐wavelength side, with a 10‐nm redshift of the maximum (green thick line). A few specific specimens, instead, have an emission spectrum showing two maxima already for the pristine film: a main band at 390 nm and a strong shoulder at about 440 nm (Figure 6b, black thick line). This shoulder becomes the most intense band under tensile stress (green thick line in Figure 6b).

Deconvolution of the emission spectra (on the frequency scale) with Gaussian lines was performed, starting from the spectra in Figure 6b. The emission band of the pristine film (black thick line) is nicely fitted as the sum of two Gaussian functions (thin black line). The two deconvoluted components, the green dashed area for the vase and the grey area for the kite conformer, have a clear resemblance with the emission spectra of PU‐C blue line in Figure 6c) and PU‐C (red line in Figure 6c), respectively. In order to fit the emission spectrum of the stretched film (green thick line), the position and width of the two Gaussians were kept fixed and only their relative intensity was varied, obtaining the green thin line as the best fit. In this way, we estimated a relative increase of the long‐wavelength band (green area) of a factor amounting to 1.53. The same analysis was performed on experimental spectra in Figure 6a; since in this case the two components of the emission band are more difficult to recognize, we chose to fix the position of the two Gaussian lines as extracted from the previous fitting, while allowing the optimization of their width and intensity. In this case, analogously to the previous one, a relative increase of the long‐wavelength band of about a 1.44 factor was estimated upon tensile elongation. Therefore, for all the specimens, two components in the fluorescence spectrum can be recognized, one corresponding to the emission of the contracted form PU‐C and the other, strongly redshifted, similar to the emission of the expanded form cross‐linked in the polymeric network, PU‐C. The relative contribution of this latter band clearly increases under tensile stress, proving the occurrence of the mechanochemical interconversion. The deconvolution analysis suggests that a certain amount of kite structure is already present in the pristine films; a likely explanation can rely on the nature of the chosen matrix, whose rigidity exerts, during the polymerization process, enough tension on cavitand molecules to induce a partial vase‐kite interconversion. For most of the samples (case exemplified in Figure 6a), an about 30 % contribution of the open conformation is already recognized in the pristine films, while in a few specific samples (Figure 6b) this contribution amounts to about 50 %. In all cases, however, upon a mechanical solicitation, the open component increases of a relative factor of about 1.5.

Conclusions

In conclusion, two different polymeric matrices were investigated for promoting the vase‐kite interconversion of tetraquinoxaline cavitands via mechanochemical stimulus. Elastomeric PDMS, despite being known for the successful activation of spiropyran mechanophores,34, 42, 43 failed to evidence any maxima variation via UV‐Vis spectroscopy for both di‐ and tetrafunctionalized cavitands. The use of a more rigid PU matrix, together with the incorporation of tetrafunctionalized QxCav units as unique cross‐linkers into the polymeric network, resulted in the expected responsiveness upon tensile elongation, detected both via UV‐Vis and fluorescence spectroscopy. Considering the lack of any significant spectroscopic variation for control specimens, we conclude that the observed response can be traced back to a mechanically induced conformational change of QxCav units embedded into the PU matrix, even if not all the cavitands reacted to the mechanical stimuli. Moreover, experimental data suggest that, contrary to the initial suppositions, all four quinoxaline wings need to be covalently anchored to the polymer backbone in order to experience an efficient transduction of tensile stress. Furthermore, fluorescence spectra allowed a preliminary quantification of the degree of activation in terms of relative increase in the intensity of the long‐wavelength band.

Experimental Section

Detailed experimental procedures, characterization of all new compounds and spectroscopic investigation of polymeric specimens are available in the Supporting Information (SI).

PDMS Samples Preparation

RTV 615 base (3.0 g) and a THF cavitand solution were homogenized with a Vortex in a 15 mL Falcon tube. RTV 615 curing agent (0.3 g) was added, and the tube was extensively shaken with a Vortex. The homogenous mixture was degassed under vacuum and poured onto a PTFE plate. After a second degassing, the sample was cured in an oven at 60 °C for 16 h. Once cured the film was peeled away and cut into stripes for testing.

(L)PDMS Samples Preparation

RTV 615 base and a THF cavitand solution were homogenized with a Vortex in a 15 mL Falcon tube. Ditopic H‐terminated PDMS (RTV 615 Base/H term. PDMS ratios reported in Table S4) and Karstedt's catalyst were added, and the tube was extensively shaken with a Vortex. The homogenous mixture was degassed under vacuum and poured onto a PTFE plate. After a second degassing, the sample was cured in an oven at 60 °C for 16 h. Once cured the film was peeled away and cut into stripes for testing.

PU Samples Preparation

A solution of isophorone diisocyanate (59 eq.) in dry THF was slowly added to a solution of cavitand and DABCO (8.42 eq.) in dry THF. The mixture was stirred at 60 °C for 2 h. The solution was cooled at room temperature and added to a solution of poly(tetramethylene glycol) (Mw 1000; 1216 eq.) and DABCO (1.05 eq.) in dry THF. The mixture was stirred under vacuum and heated at 50 °C to remove volatiles. After the addition of hexamethylene diisocyanate (1180 eq.), the viscous pre‐polymer was degassed under vacuum and poured onto a PTFE plate. After a second degassing, the sample was cured in an oven for 48 h at 60 °C and for 48 h at room temperature before testing. Once cured, the film was peeled away and cut into stripes

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.