Deciphering the Self-Catalytic Mechanisms of Polymerization and Transesterification in Polybenzoxazine Vitrimers

The use of internal catalysts has emerged as a pivotal design principle to facilitate dynamic exchanges within covalent adaptable networks (CANs). Polybenzoxazines, specifically, have shown considerable potential in generating vitrimers through thermally induced transesterification reactions catalyzed internally by tertiary amines. This study aims to investigate the chemical complexities of transesterification reactions within benzoxazine vitrimers. To achieve this, model molecules using various phenolic acids and amino-alcohol derivatives were synthesized as precursors. The structure of these model molecules was fully elucidated by using nuclear magnetic resonance (NMR). Differential scanning calorimetry (DSC) and rheology experiments evidenced the accelerated network formation of the precursors due to the presence of aliphatic −OH groups. Thermogravimetric analysis coupled with microcomputed gas chromatography (TGA-μGC) was used to provide evidence of transesterification reactions. The results showed that the spatial proximity between tertiary amine and hydroxyl groups significantly enhances the rate exchange, attributed to a neighboring group participation (NGP) effect. Interestingly, kinetic experiments using complementary NMR techniques revealed the thermal latency of the tertiary amine of benzoxazine toward transesterification reactions as its opening is needed to trigger the dynamic exchange. The study highlights the crucial role of steric hindrance and tertiary amine basicity in promoting the dynamic exchange in an internally catalyzed system.

where Ea corresponds to the average activation energy, Tp is the peak of the first exothermic transition, A is the frequency factor, β the heating rate of the curing reaction, R is the gas constant, and c a constant.The Ea can be obtained from the slope of the linear regression of ln( β Tp 2 ) versus 1 Tp (eq.1) or ln(β) versus 1 Tp (eq.2).
The apparent Ea of benzoxazine ROP (EαROP) was determined using a numerical optimization of the isoconversional Friedman analysis 4 .As this method is closely related to the conversion rate, a deconvolution of the first exothermic peak was performed using a Gaussian fitting function on OriginPro version 2019b (eq.3): (3) y= y0 + ^( -4ln(2)(x-xc) 2 w 2 ) √  4 (2)   The isoconversional Freidman method calculates the dependence of activation energy on the degree of conversion (α, eq.4).The Eα can be obtained from the slope of the linear regression of ln( d∝ dT ) versus 1 Tp (eq.4).The model-free numerical optimization implemented by NETZSCH Kinetics Neo (Professional Edition, product version 1.2.6.2.) was used for the accurate calculations of EαROP from the deconvoluted peaks.The method is based on the optimization of the results of the Friedman analysis (sum of squares of deviations between measured and simulated value) and intends to minimize the optimization function while the iterations are repeated until no any numerical improvements happens.
The EαROP was calculated as the average mean between the conversion rate of 0.2 and 0.8.

TGA-µGC:
A thermogravimetric analysis coupled with micro-gas chromatography apparatus (TGA-µGC) was used to identify and quantify the gaseous methanol stream adduct of transesterification reactions [3] .The crucible containing 10 mg of sample is directly inserted at the isotherm temperature and 50 micro GC scans of 120 s were acquired spread over the entire TGA test (4 h).For the quantitative analysis, the analyser was calibrated using a standard methanol sample.The integration of the area of the alcohol peak is normalized to the methanol standard and is used to monitor the progress of transesterification reactions.Details on the retention time of the carrier gas and standard gas sample used for the calibration are provided in Table S2.550 mL were transferred into an NMR tube.The kinetic of thermally-induced reactions was monitored at 140ºC overnight using an AVANCE III HD Bruker spectrometer operating at a proton frequency of 600 MHz.Conditions for 1 H NMR experiment: spectrum recorded every 5 minutes, 16 scans, 2.72 s acquisition time, 1 s relaxation delay.
Conditions for 1 H-15 N HMBC NMR experiment: spectrum recorded every 39 minutes, 8 scans, 0.13 s acquisition time in 1 H dimension, 5.27 µs acquisition time in 15 N dimension, 1 s relaxation delay).
Computational calculations: Model structures were drawn in ChemDraw ® software (PerkinElmer Informatics, version 19.0.0.22).The molecular modelling was visualized in the Chem3D plugin and energy was minimized using the implementation of MM2.The calculation of the partial charge density was performed using the Extended Hückel method and the algorithm provided therein.

Equipment and characterizations
Nuclear Magnetic Resonance (NMR) spectroscopy was performed on an AVANCE III HD Bruker spectrometer equipped with a 5 mm BBO-probe operating at a proton frequency of 600 MHz.All chemical shifts are given as δ value (ppm) referenced to tetramethyl silane (TMS) as an internal standard.Assignments were performed using a combination of COSY, HSQC, and HMBC spectra.Peak multiplicity was indicated as follows: singlet (s); doublet (d); triplet (t) or multiplet (m).The coupling constants (J) were reported in Hertz (Hz).
Differential scanning calorimetry (DSC) thermograms were recorded on a Netzsch DSC 204 F1 Phoenix device in standard pierced aluminum crucibles (40 μL) and a sample mass of 5 mg.A linear heating ramp at a constant heating rate was applied from 25 to 300°C under a nitrogen flow rate (N2, 40 mL.min -1 ).
Thermogravimetric analysis coupled with micro-gas chromatography (TGA-micro GC) was completed on the Mettler Toledo TGA 2 device coupled with SOLIA 490 micro-GC.TGA experiment was conducted in a ceramic alumina pan in isothermal conditions under an inert atmosphere (N2, 20 mL.min -1 ).The micro GC apparatus includes three analytical modules equipped with a micro thermal conductivity detector (10 µL injection time).The relative integration of the methanol peak' area was performed on module C using Soprane II software.

Each data point corresponds to an individual micro-GC injection.
Table S1 Retention time of the carrier gas and methanol standard on the micro GC column.

Module
Retention time carrier gas (s) Rheological measurements were recorded using an Anton Paar Physica MCR 302 rheometer equipped with a CTD 450 temperature control device.The isothermal rheo-kinetic measurements were performed using small quantities of the samples loaded in a parallel plate-plate geometry (Ø= 25 mm, gap 0.5 mm).The polymerization measurements were recorded in the oscillation mode at a controlled strain of 0.1% (1 Hz).Heating ramps of 20°C•min -1 were applied to reach the targeted temperature.The sample deformation was ramped linearly from 1% to 0.2% to remain within the instrument's limitation and to maintain a linear viscoelastic behavior as the moduli (G' storage modulus, G'' loss modulus) increase by several orders of magnitude upon curing.

Figure
Figure S1 a) 1 H and b) 13 C NMR spectra of Me-PA .

Figure
Figure S2 a) 1 H and b) 13 C NMR spectra of Me-DPA.

FigureFigureFigureFigure
Figure S3 a) 1 H and b) 13 C NMR spectra of Me-pHBA.

Figure
Figure S7 a) 1 H and b) 13 C NMR spectra of Me-PA -fa.

Figure
Figure S8 a) 1 H and b) 13 C NMR spectra of Me-DPA-mea.

FigureFigureFigureFigure
Figure S9 a) 1 H and b) 13 C NMR spectra of Me-DPA-fa.

Figure
Figure S13 a) 1 H and b) 13 C NMR spectra of pPP-fa.

Figure S15
Figure S15 HSQC spectrum of Me-PA -fa.

Figure S22 Figure S25
Figure S22 DSC thermograms of pPP-fa with the heating rate ranging from 2 to 20°C.min −1 .

Figure S37
Figure S37Area methanol peak determined on the µGC spectrum of Me-PA -fa.

Figure S36 Figure S38
Figure S36Isothermal TGA curves of Me-PA -fa.

Figure S39 Figure S40
Figure S39Area methanol peak determined on the µGC spectrum of Me-DPA-mea.

Figure S41 Figure S42
Figure S41Area methanol peak determined on the µGC spectrum of Me-pHBA-mea.

Figure S43
Figure S43 Area methanol peak at 190°C determined on the µGC spectrum of Me-PA -mea,Me-PA -dga, and Me-PA -tga.

Figure S45
Figure S45 Particular cross peak involving tertiary amine and phenol labile protons in 15 N-1 H HMBC kinetic of Me-DPA-mea.

Figure S46
Figure S46 Isothermal evolution of the complex viscosity as function of time of Me-PA -fa.

Figure S47
Figure S47Isothermal evolution of the complex viscosity as function of time of pPP-fa.

Table S2
Ratio of closed oxazine ring determined by 1 H NMR experiment.
a determined by experimental integration of the corresponding peaks in 1 H NMRELECTRONIC SUPPORTING INFORMATION

Table S3
Onset temperature of ring-opening reaction determined by DSC experiment and activation energy of benzoxazine ring-opening polymerization determined by Kissinger model, FWO equation, and numericaloptimization of isoconversional Friedman method (performed on deconvoluted peaks).

Table S4
Extended Hückel calculation on opened and closed benzoxazine.