Self‐Healable Fluorinated Copolymers Governed by Dipolar Interactions

Abstract Although dipolar forces between copolymer chains are relatively weak, they result in ubiquitous inter‐ and/or intramolecular interactions which are particularly critical in achieving the mechanical integrity of polymeric materials. In this study, a route is developed to obtain self‐healable properties in thermoplastic copolymers that rely on noncovalent dipolar interactions present in essentially all macromolecules and particularly fluorine‐containing copolymers. The combination of dipolar interactions between C─F and C═O bonds as well as CH2/CH3 entities facilitates self‐healing without external intervention. The presence of dipole‐dipole, dipole‐induced dipole, and induced‐dipole induced dipole interactions leads to a viscoelastic response that controls macroscopic autonomous multicycle self‐healing of fluorinated copolymers under ambient conditions. Energetically favorable dipolar forces attributed to monomer sequence and monomer molar ratios induces desirable copolymer tacticities, enabling entropic energy recovery stored during mechanical damage. The use of dipolar forces instead of chemical or physical modifications not only eliminates additional alternations enabling multiple damage‐repair cycles but also provides further opportunity for designing self‐healable commodity thermoplastics. These materials may offer numerous applications, ranging from the use in electronics, ion batteries, H2 fuel dispense hoses to self‐healable pet toys, packaging, paints and coatings, and many others.


Methods
Statistical p(TFEMA/nBA) copolymers were synthesized using free radical polymerization. In a typical experiment, each TFEMA/nBA monomer molar ratios were dissolved in THF at the concentration of 1.5g/mL, and 0.05 wt% of AIBN was added to the round bottom flask. N 2 gas was purged for 30min prior placing the flask into a 75 ºC oil bath. Polymerization was conducted at 75ºC for 8 hrs while stirring at 400rpm. The resulting reaction p(TFEMA/nBA) copolymer was diluted with 10mL THF and precipitated out in MeOH. The last step involved centrifuging at 9000 rpm for 5min and removal of upper layer solvent, followed by drying overnight at 75 ºC in vacuum oven. p(TFEMA/nBA) copolymer films were prepared by dissolving copolymers in THF at the concentration of 0.8g/mL, and dried in Teflon TM mold placed in sealed container under ambient conditions for 2 days, followed by drying at 75 ºC in vacuum oven for 24 hrs.

Analytical methods
Gel permeation chromatography (GPC) measurements were conducted on Tosoh EcoSEC GPC calibrated with HPLC grade poly(methyl methacrylate) (PMMA) standards using refractive index (RI) detector. Prior each measurement, copolymers were dissolved overnight in HPLC grade THF (purchased from Fisher Scientific) at the concentration of 1.5mg/ml followed by passing through 0.2µm PTFE filter. Optical images of self-healing of p(TFEMA/nBA) copolymer films were recorded using RENISHAW inVia TM Raman microscope (20×).

Mechanical properties
Self-healing efficiencies and mechanical properties of copolymers before damage and after self-healing were measured using Instron Model 5500R1125 (stress-strain). In a typical experiment, the gauge length and the strain rate were set to 1.5cm and 80mm/min respectively. To determine the self-healing capability of p(TFEMA/nBA) copolymers, 0.3×1.5×0.2 cm (W×L×T) films were damaged using stainless-steel razorblades forming cuts 50µm in width and ~60 µm in depth. These damaged films were allowed to heal under ambient conditions at room temperature (25°C) for 48hrs.
The same stress-strain measurements were performed for undamaged and damaged samples. Stress-strain measurements were repeated 10 times for each copolymer composition. 1 H and 19 F NMR spectroscopic measurements were conducted on 300MHz JEOL Model ECX-300 spectrometer with 1.5s relaxation delay and each spectrum represents 64 added scans. 2D 19 F NOESY experiments were performed on the same JEOL spectrometer with 1.5s relaxation delay, 0.5s mixing time and each spectrum represents 8 coadded scans. 2D 1 H NOESY NMR experiments were conducted on 500MHz AVANCE NEO spectrometer with 1.5s relaxation delay, 0.5s mixing time and 8 coadded scans collected for all undamaged, damaged, and self-healed samples. 2D 1 H COSY spectra represents 32 coadded scans.
NMR samples were prepared at the concentration of 1mg/ml in CDCl 3 for 1D 1 H and 19 F NMR, and 15mg/ml for 2D 1 H COSY, NOESY, and 19 F NOESY NMR. 2D 1 H COSY, 1 H NOESY, and 19 F NOESY NMR spectra were processed using MestReNova software.
Undamaged, damaged, and healed 2D 1 H NOESY NMR spectra were normalized to (4.37, 0.97) assigning to -OCH 2 and -CH 3 of pTFEMA. 19 F NMR spectra were processed using TopSpin software. 2D NMR experiments, p(TFEMA/nBA) film (5×5×0.5 mm) were damaged with 20×20 cuts on both top and bottom sides prior dissolving in CDCl 3 at the concentration of 15mg/ml for 10min without agitation, resulting in ~ 42% ratio of damaged area. 1 Healed copolymers were prepared using identical damaged sample which allowed to heal at 37 ºC for 24hrs before dissolving for 10min in CDCl 3 at the concentration of 15mg/ml without agitation.
Molecular dynamic (MD) simulations were conducted using Materials Studio software (v 5.5.0.0) (distributed by BIOVIA). Amorphous cell module and DRIEDING force field under isothermal (NVT) conditions were utilized to determine cohesive energy density (CED) as a function of copolymer compositions. For each p(TFEMA/nBA) composition, seven p(TFEMA/nBA) copolymer chains containing 60 monomer units each were placed into an amorphous unit cell at the density of 1.125 g/cm 3 . The pTFEMA and pnBA homopolymer densities were 1.181 g/cm 3 and 1.087 g/cm 3 . According to the literature, 2 during free radical propagation, stereochemistries of -COOCH 3 in MMA and -COOCH 2 CF 3 in TFEMA side chains are indistinguishable. We conducted parallel studies with the same reactivity ratio as MMA and nBA system, which are 2.60 for r TFEMA and 0.39 for r nBA . Prior to equilibration, copolymers in each unit cell were geometrically optimized using 1000. Subsequently, each cell was allowed to equilibrate for 100 psec (NVT, isothermal at 298K, at a time-step of 0.33 fsec, Berendsen thermostat) to obtain primary values of minimized energies. The cohesive energy densities for equilibrated compositions (CED eq ), vdW forces densities at equilibrium (vdW eq ) were calculated using Forcite cohesive energy density module.

Supplementary Discussion
In an effort to eliminate the possibility that sample preparation for 2D 19 F and 1 H NMR experiments may have impacted p(TFEMA/nBA) copolymer conditions, a series of control experiments were conducted on filtered undamaged and damaged copolymer specimens (filter pore size: 0.2µm). As shown in Figure S3, A and B, 2D 19 F NOESY NMR spectra show identical sequence of resonances for both undamaged and damaged samples. DLS measurements showed the same particle size distribution for both filtered samples. This control experiment shows that the inversible resonances result from interactions within the insolubilized parts of copolymers.
A typical 2D NOESY or COSY NMR experiment takes 100 min. To eliminate the possibility that inter-chain interactions may be altered during solvation process when dispersing p(TFEMA/nBA) copolymers in CDCl 3 , a series of controlled experiments were conducted using DLS capable of measuring 10-1000 nm particle size. The DLS sample preparation paralleled 2D NOESY or COSY NMR experimental conditions and showed that after 100 min without agitation the particle size exceeds the instrument upper detection limits.
The phase sequences resume identical for 2D 19 F NOESY NMR spectra ( Figure S5 C-C'') across the undamaged, damaged and post-damaged process for 40/60 copolymers, whereas for 60/40 copolymers, the phase sequences ( Figure S5 F-F'') are reversed when comparing undamaged and damaged samples, the phases of post-damaged samples after two days are the same as the damaged one due to the lack of selfhealing capability.
The following normalization resonance (4.37, 0.97) in 2D 1 H NOESY NMR were used, instead of previously mentioned assigning 0.97ppm assigned to -CH 3 in pTFEMA, terminal -CH 3 on the side chain of nBA. In this case, the requirement of unchanged interactions through damage-repair cycles may not be satisfied. However, by comparing the pre-normalized spectra of undamaged, damaged, and healed specimens, the integration of (4.37, 0.97) resonance are consistent, which eliminated this possibility. Table S2 summarizes monomer feed (f) and actual (F) ratios in p(TFEMA/nBA) copolymers which were determined using 1 H NMR spectroscopy. 3 Junction densities (v j ), stored entropy (ΔS s ), and mol. wt. between junction points (M j ) were determined using dynamic mechanical analysis (DMA). In a typical DMA experiment, storage modulus (E'), loss modulus (E''), tan δ (log(E'')/log(E')), and viscoelastic length transitions (VLT) as a function of temperature were obtained. Using experimental VLT values from a single DMA measurement and applying rubber elasticity theory allows us to calculate the ν j using the following relationship ν j = ζ R / [RT (α -1 /α 2 )]; where: ζ R is the retractive stress and α is the elongation ratio (L/L 0 ) obtained from the DMA analysis. Using ∆S = -(Rν j /2) × [α 2 +2/α-3] 4 , this approach also allows determination of stored entropy ∆S s = -T εmax S εmax + T i S i ; where: T εmax and S εmax are temperature and entropy at max elongation (ε max ), and T i and S i represent values before elongation. 5 These ∆Ss and v j are shown as a function of copolymer composition (Table   S3).
To determine the impact of monomer distribution on CED and self-healing properties, NPT and NVT simulations were performed on a series of TFEMA/nBA hextads. In a typical experiment, 14 identical hextads were placed into a unit cell at 0 GPa with the density of 1.125 g/cm 3 . Afterwards, all the hextads were allowed to reach optimum packing density under NPT equilibration for 100ps. Subsequently, the hextads were loaded into a unit cell with the equilibrium densities calculated via NPT ensemble, followed by 100ps isothermal equilibration via NVT. The average cohesive energy (CE p ) values were calculated. Figure S7 illustrates multiple damage-repair cycles conducted on p(TFEMA/nBA) copolymer films with TFEMA/nBA = 45/55 monomer molar ratio films under ambient conditions. Each damage-repair cycle represents approximately 1.5~2 hrs between subsequent cuts. Figure S8 illustrates the GPC traces and Table S4 summarizes molecular weights and dispersity of selected p(TFEMA/nBA) copolymers.     (I) self-healed after fourth cycle of multiple cuts. This process can be repeated multiple times.

Video S1
Self-healing of p(TFEMA/nBA) copolymer films recorded under optical microscope