Polymeric materials that convert local fleeting signals into global macroscopic responses

Polymers that support self-propagating reactions are used to create materials that change global wetting properties in response to specific fleeting, local stimuli.


Introduction
The leaves of Venus ytraps and Mimosa pudica (touch-me-nots) close when mechanically stimulated. This macroscopic change in the structure of the leaves is controlled by biological materials within the leaves that are capable of translating a local, oen eeting signal (e.g., brief touch) in one section of a leaf into a response in the entire leaf. [1][2][3][4] Here we describe a general design strategy for creating polymeric materials that also display the ability to translate local, eeting stimuli into global macroscopic changes in the properties of the entire material. In a proof-ofconcept demonstration, we illustrate this capability using a material that switches from hydrophobic to hydrophilic (Fig. 1a), but the design strategy should be compatible with a variety of other macroscopic responses by simply mixing and matching functionality on the polymer that makes up the material.
Global, autonomous changes in the properties of these materials occur via self-propagating reactions within the materials that are mediated by specic functionality on each repeating unit of the polymers that make up the materials. These selfpropagating reactions simultaneously communicate a detection event to distant portions of a material while also changing the properties of the material, all without requiring electronics, reagents from the surroundings, input from a user, or tethering to a specic location. In comparison, traditional synthetic stimuliresponsive materials have no mechanism for altering a distant portion of a material as a result of a local detection event. [5][6][7][8] Thus, this work provides the foundation for a new class of bio-inspired stimuli-responsive materials. 9,10 This unique capability, once developed further, should enable a new generation of plastics, coatings, adhesives, and other materials that autonomously recongure themselves in response to low intensity, limited duration exposure of a specic applied stimulus, even when the stimulus interacts with only a section of the material.

Design strategy
Our design for these bio-inspired materials requires only a single, functionalized copolymer (gray spheres in Fig. 1b), where the AB repeating units in the polymer are distributed randomly in a 1 : 2 ratio, respectively, throughout the polymer. The functionality on the polymer includes: (i) sensing groups (pink rectangles on repeating unit A in Fig. 1b), (ii) functionality that mediates a self-propagating signal amplication reaction (pink squares with a wedge removed on repeating unit B), and (iii) polymer-bound chemical reporters (green wedges). The less abundant sensing functionality on repeating unit A responds selectively to an applied signal by undergoing a chemical reaction that changes its properties (in this case, becoming hydrophilic, Fig. 1a), while also communicating the occurrence of the detection event to other portions of the material by releasing the chemical reporters (green wedges). The functionality on repeating unit B responds to the chemical reporter by simultaneously releasing more copies of the reporter while also transforming the molecular structure attached to repeating unit B (again by becoming hydrophilic, Fig. 1a). Thus, in theory, a single detection event can trigger a sequence of reactions that transform the physical properties of the material as a whole, in this case, rendering the material hydrophilic.
An example of a polymer that supports self-propagating reactions in response to eeting stimuli Fig. 2 depicts a specic poly(norbornene) (1) that we designed to demonstrate the overall concept of materials that change globally in response to local and eeting stimuli. In this example, the detection functionality is an o-nitrobenzyl carbamate, which reacts when exposed to 300 nm light (the applied signal) to release a pendant carbamate and ultimately four molecules of uoride (the chemical reporters). The released uoride is free to diffuse across the lm to propagate the change in wetting properties by reacting with pendant tertbutyldimethyl silyl groups (TBS groups) on repeating units B, breaking the Si-O bond, which ultimately leads to release of four additional molecules of uoride. 11,12 Upon completion of these reactions, the formerly hydrophobic polymer has lost signicant contributors to its hydrophobicity (e.g., uorine atoms, aromatic rings, and TBS groups) and now possesses alcohols. Consequently, the resulting polymer is signicantly more hydrophilic than the starting polymer.
We used uoride as the chemical reporter in this example because we need a signal (i) that accurately transfers the message from the detection event to initiation of the signal amplication reaction; (ii) readily diffuses across macroscopic distances along (and perhaps within) the material; (iii) is sufficiently stable and non-volatile to prevent its loss; and (iv) is sufficiently reactive with the TBS group on repeating unit B to initiate the self-propagating reaction on a reasonable timescale. We chose the TBS group on repeating unit B to receive the signal from the chemical reporter because of the high selectivity of the TBS group for uorideinduced heterolytic cleavage of Si-O bonds, thus minimizing the possibility for spurious initiation of the self-propagating reaction in the absence of the applied signal. Finally, we used light as the applied signal for this demonstration because it is a convenient signal for providing spatiotemporal control for proof of concept illustration of local and eeting signals.

Characterization of a self-propagating response in a polymeric material
Prior to studying the polymer, we tested whether monomer 4 ( Fig. 2b) was capable of supporting a self-propagating autoinductive reaction when exposed to substoichiometric quantities of uoride. Thus, we treated 4 (1 mM in 10 : 4 : 1 MeCN-H 2 O-pyridine) with substoichiometric uoride and quantied the time-dependent disappearance of 4 using liquid chromatography coupled to a mass spectrometer (LCMS). This experiment (ESI Fig. 1 †) revealed that (i) all of 4 is consumed when exposed even to 0.05 equiv. uoride; (ii) the rate of consumption of 4 is faster when exposed to quantities of uoride that are higher than 0.05 equiv.; and (iii) the kinetics are sigmoidal. These three observations are consistent with selfpropagating autoinductive reactions. 11 Moreover, the expected amino dialdehyde product of the reaction (3, Fig. 2a) was observed in the LCMS chromatograms (ESI Fig. 2 †), thus supporting the mechanism for conversion of 4 to the expected products outlined in Fig. 2.
Ring-opening metathesis polymerization of 4 provided homopolymer 6 (Fig. 2b), which was obtained in 92% yield with a number average molecular weight (M n ) of 290 kDa and a polydispersity index (PDI) value of 1.2. Spin casting of a 5 mg mL À1 chloroform solution of 6 onto polypropylene provided a 4.3 nm AE 0.1 nm thick lm with x,z-dimensions of 1 cm Â 0.5 cm. Immersion (ESI Fig. 3a †) of 42 replicas of this lm in sealed vials, each containing 0.3 mL of 100 mM uoride in 10 : 4 : 1 i-PrOH-H 2 O-pyridine was expected to induce the autoinductive selfpropagating reaction in the solid lms. Several lms were removed from the solution and dried at various intervals aer initial exposure to uoride (six lms per interval) and their propensity to wet with water was measured. As expected, the contact angle measurements decreased sigmoidally for lms with increasing durations of exposure to uoride: global contact angles changed from 90.8 AE 0.4 to 74.9 AE 2.0 (ESI Fig. 3b †) (the error is provided at 90% condence intervals). 13 Atomic force microscopy (AFM) of these thin lms, measured before and aer exposure to different initial concentrations of uoride, reveals Fig. 1 Schematic illustration of a synthetic polymeric material that is capable of changing physical properties globally in response to specific applied signals that are fleeting and/or interact with only one portion of the material. (a) A hydrophobic film that responds to a fleeting local signal (generating the blue hydrophilic region) and then converts entirely from hydrophobic (pink) to hydrophilic (blue) via self-propagating reactions that continue even in the absence of the signal. (b) The material is made from a random poly(norbornene) AB copolymer that contains repeating units (spheres) that are functionalized either with detection functionality (rectangles attached to repeating unit A) or functionality that mediates the self-propagating reaction (squares with missing wedges attached to repeating unit B). The green wedges represent the chemical reporters, which initially are covalently attached to the polymer, but subsequently are freed to diffuse and react with the functionality that mediates the selfpropagating reaction.
statistically insignicant changes in surface roughness as a result of the self-propagating reaction: e.g., 0.6 nm AE 0.2 nm surface roughness before exposure to 100 mM uoride changing to 1.3 nm AE 0.7 nm features aer 48 h of exposure (ESI Fig. 4 †). Moreover, exposure of a lm of 6 for 48 h to 10 : 4 : 1 i-PrOH-H 2 O-pyridine in the absence of uoride resulted in no change in contact angle (i.e., 90.8 AE 0.4 before exposure to the solvent and 89.9 AE 0.6 48 h aer exposure; the error is provided at 90% condence intervals). Thus, the 16 decrease in contact angle aer 48 h of exposure of 6 to uoride is the consequence of changing molecular composition of the surface of the lm (as designed), due to loss of uorine atoms, aromatic rings, as well as other functionality during the self-propagating reaction (Fig. 2).
Characterization of the response of a polymeric lm to a eeting stimulus Taken together, these solution and solid phase experiments demonstrate the success of the self-propagating reaction as well as the ability of uoride to act as a chemical reporter. We next evaluated the feasibility of the proposed photochemical reaction in lms of homopolymer 7 (Fig. 2b), which contain only the functionality that responds to 300 nm light (the applied signal).
Attenuated total reection infrared spectroscopy (ATR-FTIR) measurements (ESI Fig. 7 †) of spin-cast lms of 7, aer various durations of exposure to 300 nm light, revealed complete loss of the asymmetric stretch of the carbamate from each repeating unit aer 40 min of exposure. Loss of this diagnostic carbamate stretch is consistent with a successful photochemical reaction, as veried by an LCMS experiment using monomer 5. Specically, the photochemical reaction of 5 causes release of the pendant aniline, which ultimately releases four molecules of uoride in subsequent azaquinone methide-mediated reactions. LCMS analysis of the products of the reaction of 5 with 300 nm light revealed the expected nitrosobenzene photochemical product as well as aminodialdehyde 3 (ESI Fig. 5 and 6 †).
Demonstration of a material that alters its macroscopic properties in response to eeting stimuli Since homopolymers 6 and 7 individually performed successfully in their designed roles, we prepared copolymer 1, which contains a random distribution of repeating units A and B in a 1 : 2 ratio (Fig. 2a), as conrmed by analysis of peak areas in the 1 H NMR spectrum of the polymer (ESI Fig. 25 †). This copolymer (M n value of 160 kDa; PDI value of 1.5) readily forms lms when pyridine in a sealed container and storing them in the dark at 23 C provided an appropriate environment for the selfpropagating reaction. At various intervals, six of these lms were dried and tested for their propensity to wet with water. The contact angles of the lms continued to decrease over a period of $24 h, long aer the signal (300 nm light) had been removed, for a total change of 16 . A plot of water contact angle versus exposure time to i-PrOH-H 2 O-pyridine reveals a temporal change in contact angle (Fig. 3a) that is commensurate with the kinetics of the self-propagating autoinductive reaction for homopolymer 6 (ESI Fig. S3 †). The graph in Fig. 3a is not unambiguously sigmoidal as was observed when homopolymer 6 was treated with 100 mM uoride (ESI Fig. S3 †). However, for copolymer 1, all of the photoresponsive groups were consumed during the 40 min irradiation with UV light (ESI Fig. S7 †), therefore it is likely that considerable quantities of uoride were generated during the photochemical reactions. High concentrations of uoride would reduce the induction period for the autoinductive reaction, thus yielding kinetics that are comparable to the response when homopolymer 6 was treated with 1 mM or 10 mM uoride (ESI Fig. S3 †).
Control experiments support the conclusion of a signalinduced self-propagating reaction, as illustrated in ESI Fig. 8. † For example, exposure of copolymer 1 to i-PrOH-H 2 O-pyridine but not 300 nm light resulted in lms with unchanged contact angles aer 24 h. Likewise, exposure of homopolymer 6 to 40 min of 300 nm light followed by 24 h of wetting with i-PrOH-H 2 O-pyridine again provided statistically insignicant changes in contact angle. However, exposure of homopolymer 7 to UV light and solvent did cause a 4 initial decrease in contact angle due to the photochemical reaction, as expected, but did not continue to change further over time. Thus, the only lm that provides a continuous bio-inspired response to the eeting signal is copolymer 1.
Global changes in the properties of a material in response to a local stimulus In analogy to Fig. 1a, copolymer 1 also is capable of communicating a local detection event to the entire lm, thus inducing global changes in wetting properties of the lm. Fig. 3b illustrates this capability, where half of the lm (4.0 nm AE 0.1 nm thick with x,z-dimensions of 2 cm Â 0.5 cm) was covered in foil to block the 300 nm light, while the other half was exposed to the signal. Changes in wetting properties of the unexposed half of the lm tracked with the exposed half, 14 albeit with a slight delay as expected from the bio-inspired design since the selfpropagating reaction had to communicate a macroscopic change from the exposed region to the unexposed portion of the material (Fig. 3b). Hence, the lms behave similarly to Venus ytraps and touch-me-nots in their ability to impart a macroscopic response to a eeting signal.
The physical (rather than chemical) mechanism of signal propagation has yet to be established, although on going studies are probing this question. It may be possible for the uoride to diffuse through the polymer matrix during the selfpropagating reaction, particularly since the solvent may plasticize and/or swell the lm. We suspect, however, that the observed changes in contact angles in this proof-of-concept study arise predominantly from diffusion of uoride through the surrounding uid, rather than through the polymer matrix.

Conclusions
Overall, these results demonstrate the success of self-propagating reactions to impart synthetic materials with the ability to respond globally to signals that are both eeting and localized. The modular design of this system should be compatible with responses to a variety of applied signals other than light by The open circles depict the average contact angles obtained from the unexposed half of the film, while the closed circles represent the values obtained from the exposed half (six films were measured to provide the average values for the data points). The error bars in both graphs reflect the uncertainty in the measurements at 90% confidence intervals, and the lines provide the fit of an autocatalytic equation to the data (see the ESI † for additional details).
simply exchanging the detection functionality on repeating unit A (Fig. 2) with a functionality that responds to another signal, such as Pd(0), 11,15 H 2 O 2 , 16 enzymes, 17 thiols, 18 and others. 19,20 Likewise, the chemical reporters need not be limited to uoride: alternative privileged reporters (such as H 2 O 2 , 16 piperidine, 20 or thiols 18 ) that have been demonstrated in small molecule autocatalytic, autoinductive, and networked reactions should be compatible with this design strategy as well. The modular design of the system also should be compatible with various structural changes to the functionality that supports the self-propagating reaction and simultaneously provides the macroscopic readout. Such changes, in theory, could be used to impart a range of macroscopic responses in materials beyond altering surface wetting properties. Finally, we anticipate that the approach will be compatible with a wide range of polymerization methods. Thus, this bioinspired strategy may provide the foundation for realizing a diverse set of synthetic materials that display response properties that extend beyond traditional stimuliresponsive materials [21][22][23][24][25][26][27][28] and begin to resemble the behaviour of biomaterials, at least at a rudimentary level.