A mechanistic study of Lewis acid-catalyzed covalent organic framework formation

Abstract

Three boronate ester-linked covalent organic frameworks (COFs) were synthesized using a new approach that employs polyfunctional boronic acid and acetonide-protected catechol reactants in the presence of the Lewis acid catalyst BF₃·OEt₂. This transformation avoids the use of unstable and insoluble polyfunctional catechols. The COF-5 and COF-10 hexagonal lattices were obtained from a triphenylene tris(acetonide) and the appropriate diboronic acid linker, whereas a square Ni phthalocyanine COF was prepared from the appropriate Ni phthalocyanine tetra(acetonide). The powder X-ray diffraction, infrared spectra, and measured surface areas of these materials matched or exceeded previously reported values. A mechanistic study of this transformation revealed that the dehydrative trimerization of boronic acids to boroxines and the formation of a nonproductive aryl boronic acid–BF₃ complex strongly affect the rate of boronate ester formation. Crossover experiments employing substituted boronate ester derivatives suggest that esterhydrolysis is the most likely exchange mechanism during COF formation under BF₃·OEt₂-catalyzed conditions.

---

Introduction

Covalent organic frameworks (COFs) are an emerging class of crystalline, porous materials composed of organic subunits linked by covalent bonds.^1–5 The high surface areas (over 4000 m² g⁻¹) and record low densities⁶ of three-dimensional (3D) COFs are attractive for gas storage applications.^7–14 Two-dimensional (2D) COFs^15–17 are layered materials whose aromatic subunits are often stacked into columns ideally suited for efficient energy and charge transport.^2,4,18–20 In spite of their significant promise, relatively few COFs have been crystallized and materials incorporating functional building blocks are even more rare. New synthetic methods are needed for COFs to realize their full potential.

The largest class of COFs features boronate ester linkages formed by condensing polyfunctional boronic acids and catechols under solvothermal conditions, but few such catechol motifs have produced crystalline materials. These building blocks are prone to oxidation^1,21 and are poorly soluble in most organic solvents, drawbacks that make them difficult to prepare and incorporate into COFs. In contrast, acetonide-protected catechols offer improved solubility and stability, and we have found that they can undergo in situ deprotection to form boronate esters directly from boronic acids in the presence of the Lewis acid catalyst BF₃·OEt₂ (Scheme 1). We used this strategy to synthesize a 2D phthalocyanine-containing COF (Pc-PBBA COF)⁴ of interest for organic photovoltaic devices, an advance in complexity over frameworks that had been reported previously.

Scheme 1 Lewis acid-catalyzed formation of catechol boronate esters.

Here we generalize this approach to three additional COFs: a square Ni phthalocyanine framework (NiPc-PBBACOF)¹⁹ as well as the hexagonal lattices COF-5 and COF-10 (Scheme 2).^1,22 While optimizing their syntheses, several questions arose about the mechanisms and rates of boronate ester formation and exchange. We studied these processes using the model reaction shown in Scheme 1 and have developed a mechanistic picture of boronate ester-linked COF formation under BF₃·OEt₂-catalyzed conditions. These studies will establish paradigms for predicting effective crystallization conditions, which currently must be optimized empirically for each new material, and guide catalyst improvements.

Scheme 2 The synthesis of COF-5, COF-10, and NiPc-PBBACOF under BF₃·OEt₂-catalyzed conditions.

Results and discussion

COF synthesis and characterization

We synthesized two known frameworks, COF-5 and COF-10 (Scheme 2), to compare their crystallinity and porosity to reported values. It is necessary to benchmark new COF syntheses in this way because efficient bond formation is not sufficient to obtain ordered materials. Amorphous networks result if the rates of bond formation, exchange, and crystal nucleation are not appropriately balanced.²³ We obtained COF-5 by condensing a 2 : 3 molar ratio of triphenylene tris(acetonide) 4 and phenylenebis(boronic acid) (PBBA, 5) in a 1 : 1 v/v mixture of mesitylene : dioxane in the presence of 1 equiv of BF₃·OEt₂ per boronic acid functional group at 90 °C for 72 h. COF-10 was prepared under similar conditions by using 4,4′-biphenylbis(boronic acid) 3 in place of 2. We also synthesized the Ni derivative of our previous free-base phthalocyanine material⁴ (NiPc-PBBACOF) by condensing NiPc tetrakis(acetonide) 7 and 5 in a 3.3 : 1 mixture of 1,2-dichloroethane : mesitylene at 120 °C for 144 h.

COF-5, COF-10, and NiPc-PBBA COF were each obtained under the above conditions as insoluble microcrystalline powders. Powder X-ray diffraction (PXRD) of COF-5 showed (100), (110), (200), (210), (310) and broad (001) diffraction peaks at 3.44°, 5.96°, 6.88°, 9.12°, 12.48°, and 26.6°, respectively (Fig. 1(a)). Pawley refinement of the diffraction pattern suggested a P6/mmm-symmetric hexagonal framework (a = b = 3.00 nm) containing cofacially stacked aromatic units (c = 0.34 nm), which match the reported (a = b = 2.97 nm; c = 0.35 nm) and predicted (a = b = 3.00 nm; c = 0.34 nm) lattice constants closely.^1,22COF-10 exhibited a similar diffraction pattern whose peak locations (2.80°, 4.72°, 5.56°, 7.36°, 9.80°, 25.67°) correspond to the (100), (110), (200), (210), (310), and (001) diffractions of its larger unit cell (a = b = 3.78 nm; c = 0.35 nm). Fourier transform infrared (FTIR) spectroscopy of the COF powders produced under Lewis acid-catalyzed conditions also indicated boronate ester formation and displayed attenuated hydroxyl stretches relative to the reactants (see ESI†). The COF-5 and COF-10 samples were also characterized by gas sorption analysis (Fig. 2) to evaluate their accessible surface area. COF-5 exhibited a reversible Type IV isotherm with a Brunauer–Emmett–Teller (BET) surface area of 1670 m² g⁻¹.^24,25 This value is quite close to that first reported for COF-5 (1590 m² g⁻¹).²⁶COF-10 exhibited a Type IV isotherm with a BET surface area of 1320 m² g⁻¹. COF-10's reported Langmuir surface area of 2080 m² g⁻¹ is quite close to that derived from our isotherm (1970 m² g⁻¹). Thus, samples of COF-5 and COF-10 obtained from BF₃·OEt₂-catalyzed esterifications exhibit similar crystallinity, spectroscopy, and porosity compared to those obtained from the direct condensation. It should also be noted that triphenylene tris(acetonide) 4 is obtained in one step from catechol acetonide²⁷ and is stable under ambient conditions, making it an attractive COF building block.

Fig. 1 PXRD patterns of (a) COF-5 (blue trace) and COF-10 (red trace) and (b) NiPc-PBBACOF (blue trace) and simulated PXRD (red trace). The major diffraction peaks are labeled.

Fig. 2 N₂adsorption isotherms of COF-5, COF-10, and NiPc-PBBA COF synthesized under BF₃·OEt₂-catalyzed conditions.

The PXRD pattern of the NiPc-PBBACOF (Fig. 1(b)) is similar to that of the free-base Pc-PBBA COF but exhibits true P4/mmm symmetry. The peaks observed at 3.88°, 7.76°, 11.72°, and 15.64° correspond to the (100), (200), (300), and (400) diffractions. The weak diffraction at 26.9° indicates an interlayer stacking distance of 3.35 Å along the (001) vector,^28,29 indicating that the sheets are in van der Waals contact. Refinement of these data provided unit cell parameters a = b = 22.84 Å, a value similar to both the free base structure Pc-PBBA COF (a = b = 22.85 Å) and the NiPc-PBBACOF (a = b = 23.12 Å) prepared via direct condensation.¹⁹NiPc-PBBACOF absorbs light over a broad range of the solar spectrum (250–1100 nm) as a consequence of its H-aggregated structure^30,31 and displays excellent thermal stability to 500 °C as determined by thermogravimetric analysis. NiPc-PBBACOF (Fig. 2) exhibited a Type IV N₂adsorption isotherm with a higher BET surface area (776 m² g⁻¹) than was reported for the material prepared by direct condensation¹⁹ (624 m² g⁻¹) or for Pc-PBBA COF prepared via BF₃·OEt₂ catalysis (506 m² g⁻¹).⁴ We attribute the increase in surface area of the NiPc-PBBA COF relative to Pc-PBBA COF to its improved crystallinity, as evidenced by its sharper PXRD peaks. The lower surface areas of phthalocyanine-containing COFs compared to COF-5 and COF-10 are a consequence of the larger size of the Pc macrocycle in the unit cell compared to triphenylene and the higher density of a square relative to a hexagonal lattice.

Mechanistic studies of boronate ester formation

The interplay between the bond formation and exchange processes that produce crystalline COFs is poorly understood, and the BF₃·OEt₂-catalyzed boronate esterification is itself a new functional group transformation. Although BF₃·OEt₂ is an inexpensive and readily available catalyst, we wondered why high loadings were necessary to obtain crystalline COFs. These boronic acids trimerize reversibly to form boroxines and H₂O,^32–34 an equilibrium that provides boroxine-linked COFs^1,3,18 when established in the absence of catechols. The role of the trimerization during boronate ester-linked COF formation is unknown, but boroxine formation decreases the concentration of free boronic acid and increases the concentration of H₂O in the reaction mixture. Thus it should affect the rates of boronate ester formation and exchange associated with COF crystallization. 1 rapidly dehydrates to the corresponding boroxine 8 when dissolved in anhydrous CDCl₃ at 28 mM.^35,36 The equilibrium favors the boroxine under these conditions, as 89% of the available aryl boronate species exist in the boroxine form (a 2.8 : 1 mole ratio of 8 : 1), as determined by integrating the ¹H NMR resonances of each species. We also isolated 8 by refluxing 1 in PhMe under Dean–Stark conditions, allowing us to study its reactivity independently.

We measured the rate of formation of 3 under pseudo-first order conditions (20 equiv 2; 0.5 equiv BF₃·OEt₂), using either boronic acid1 or boroxine 8 reactants (Fig. 3). When the boronate source is 1, the reaction mixtures initially contain mostly 8 along with residual 1 and the H₂O derived from the dehydration (Fig. 3(a), top spectrum). When 8 serves as the boronate source, only trace amounts of free 1 are observed (Fig. 3(b), top spectrum). These differences affect the rate of formation of 3 significantly. The reaction progress is characterized by two distinct kinetic domains when 1 is used as the boronate source. 3 forms rapidly during the first few minutes of the reaction, while the resonances associated with free 1 disappear (Fig. 3(a), middle spectrum). Once 1 is consumed, the rate of formation of 3 slows. During this time, the resonances of 8 disappear as 3 is formed, but no resonances that correspond to free 1 appear. These observations suggest that free 1 is rapidly converted to the boronate ester in the presence of 2 and BF₃·OEt₂. Once the boronic acid is consumed, boroxinehydrolysis becomes rate limiting. Experiments employing 8 as the boronate source are consistent with this interpretation. The rapid initial formation of 3 occurs in the first 2 min, originating from trace amounts of free 1 present in the solution. Thereafter, the reaction appears hydrolysis limited. We also found that adding H₂O to partially complete reactions that employed 8 as the boronate source caused the rate of formation of 7 to immediately increase (Fig. S16, ESI†).

Fig. 3 (a, b) Partial ¹H NMR spectra (CDCl₃, 298 K) of BF₃·OEt₂-catalyzed boronate ester formation reactions employing different boronic acid sources. (a) utilizes 4-t-butylphenylboronic acid 1 (28 mM) and (b) utilizes the corresponding boroxine 8 (9.3 mM). Both samples contain 560 mM catechol acetonide 2 and 14 mM BF₃·OEt₂. (c) Plots of the formation of 3 as a function of time for both experiments.

In situ infrared spectroscopy provided improved temporal resolution of the fast kinetic regime associated with the conversion of free 1 to 3. The formation of 3 was monitored through the appearance of a well-resolved B–O stretch characteristic of the boronate ester at 1335 cm⁻¹. Plots of the intensity of this peak as a function of time (Fig. 4) reproduced the two-stage kinetic behavior observed in the NMR experiments. Doubling the excess [2]₀ from 28 mM to 56 mM (10 equiv to 20 equiv relative to 1) caused the initial rate to approximately double, suggesting that the fast process is first order in the acetonide component. In contrast, the slower kinetic regime does not depend on [2]₀ strongly, which is consistent with rate-limiting boroxinehydrolysis. The hydrolysis-limited reaction is slightly slower with increased [2]₀, which we attribute to the competing deprotection of 2 to catechol that reduces the [H₂O] available for hydrolysis. Indeed, at the higher of the two [2]₀, we observed the buildup of free catechol before 8 had been consumed.

Fig. 4 Plots of the intensity of a B–O stretching frequency (1335 cm⁻¹) characteristic of boronate ester3 obtained from in situIR spectroscopy of reaction mixtures containing either 10 equiv (blue squares) or 20 equiv (red circles) of 2. [1]₀ = 2.8 mM and [BF₃·OEt₂] = 2.8 mM for both experiments. The blue line is a linear fit to y = 8.12 × 10⁻⁶x; the red line is a linear fit to y = 1.54 × 10⁻⁵x, suggesting that the initial process is first order with respect to 2.

High catalyst loadings (0.5–1 equiv BF₃·OEt₂ per ArB(OH)₂ group) necessary to obtain crystalline COFs motivated us to study the catalyst loading needed to effect boronate ester formation in the model system. We measured the rate of formation of 3 under pseudo-first order conditions at various [BF₃·OEt₂] (Fig. 5(a)). The reaction is sluggish when fewer than 0.3 equiv of BF₃·OEt₂ relative to 1 are employed but proceeds rapidly above this threshold. The formation of 3 also slows when [1]₀ is increased at constant [BF₃·OEt₂] (Fig. 5(b)). These observations suggest the formation of a nonproductive BF₃·1 complex that does not deprotect 2. We are unaware of previous reports of boronic acid-BF₃ complexes but have obtained spectroscopic evidence for their formation. CDCl₃ solutions of 1 and BF₃·OEt₂ show downfield shifts of the 2,5-¹H resonances of free 1 from 7.70 to 7.82 ppm, and the methylene groups of the Et₂O resonate at 3.48 ppm, the chemical shift of the uncomplexed species. ¹⁹F NMR spectra of solutions of 1 containing fewer than 0.3 equiv of BF₃·OEt₂ exhibit three resonances shifted downfield from those of BF₃·OEt₂. ¹⁹F resonances of BF₃·OEt₂ are not observed until at least 0.3 equiv are present, a transition corresponding to the loadings needed to catalyze boronate ester formation. The separation of the ¹⁹F resonances of the BF₃·1 complex into three well-resolved signals suggests that the BF₃ is rigidly bound by other noncovalent interactions. BF₃ does not bind catechol acetonide 2, boroxine8, or the boronate ester product 3 (see ESI†). ¹⁹F NMR spectra of reaction mixtures taken to high conversions of the boronate ester show restoration of the BF₃·OEt₂ resonance after the boronic acid1 is consumed (Fig. S22, ESI†).

Fig. 5 (a) Plots of the conversion of 1 to 3 as a function of time of reactions containing various [BF₃·OEt₂]. Each reaction mixture contained [1]₀ = 28 mM and [2]₀ = 280 mM. (b) Plots of the formation of 3 as a function of time of reactions containing varying amounts of [1]₀. Each reaction mixture contained [2]₀ = 560 mM 2 and [BF₃·OEt₂] = 19 mM.

The above results suggest a mechanistic picture (Fig. 6) that is relevant to boronate ester formation during COF synthesis. The boronic acid reactant 1 partially trimerizes to the boroxine 8, liberating much of the H₂O of condensation at the outset of the reaction. The remaining 1 binds BF₃ nonproductively. Free BF₃·OEt₂ catalyzes acetonide hydrolysis, and the resulting catechol rapidly condenses with 1 to form the boronate ester product 3. Acetonide hydrolysis is rate-limiting before free 1 is consumed, after which the rate-limiting step changes to boroxine hydrolysis (formation of 1). Much of the boronate ester formation in the model system occurs under the latter kinetic regime, suggesting that the rates of COF formation might be quite sensitive to adventitious or added H₂O.

Fig. 6 Proposed mechanism of BF₃·OEt₂-catalyzed boronate ester formation.

After establishing this model of boronate ester formation under Lewis acid-catalyzed conditions, we identified the likely mechanisms of boronate ester exchange. These processes must occur at appreciable rates during COF formation to obtain crystalline products. We evaluated boronate ester exchange by ¹H NMR through crossover experiments between the simple boronate esters9 and 10 as well as their boronic acid and catechol constituents (Fig. 7). 9 and 10 do not exchange aryl boronate groups in anhydrous CDCl₃ over the course of 24 h either in the presence or absence of BF₃·OEt₂. However, these species form a statistical mixture of the four possible boronate esters within minutes in CDCl₃ saturated with H₂O (46 mM). We also found that equimolar mixtures of 10 and catechol exchange to form a statistical mixture of boronate esters within minutes in anhydrous CDCl₃. In contrast, mixtures of 9 and phenylboronic acid require ∼1 h to reach equilibrium (Fig. S27, ESI†). Exchange is slowed further in the presence of 1.5 equiv BF₃·OEt₂, requiring more than 3 h to reach a statistical equilibrium. This is most likely because BF₃ binds to the free boronic acid. These observations suggest that boronate esterhydrolysis is the most important exchange mechanism that occurs during BF₃·OEt₂-catalyzed COF formation, because free catechols are not present in large amounts and boronic acid-based exchange occurs an order of magnitude more slowly. Boronate esterhydrolysis and exchange also provide an unexplored opportunity to reactivate or repair the COF after its formation.

Fig. 7 Summary of the results of exchange experiments between boronic esters, H₂O, catechols, or boronic acids.

Conclusions

We have developed a general strategy to form boronate ester-linked COFs from operationally convenient, soluble, and stable acetonide-protected polyfunctional catechols. This approach broadens the scope and complexity of available COF materials. Mechanistic studies on soluble model compounds reveal a complex interplay of equilibria that accompany boronate ester formation. Understanding these processes will allow rational optimization of COF crystallization conditions and impact other emerging classes of boronate-linked materials.^37–39 We continue to develop new strategies for COF synthesis and will apply these strategies to prepare materials with advanced functionality.

Acknowledgements

This research was supported by startup funds provided by Cornell University and the NSF CAREER award (CHE-1056657). E.L.S. gratefully acknowledges the award of the American Competitiveness in Chemistry postdoctoral fellowship (ACC-F) from the NSF (CHE-0936988). M.R.G. was supported by the Rawlings Cornell Presidential Research Scholar (RCPRS) undergraduate research program. S.L.W. was supported by the Research Experience for Undergraduates (REU) program within the NSF Materials Research Science and Engineering Centers (MRSEC) program (DMR-0520404). We thank David B. Collum for helpful discussions.

Notes and references

1. A. P. Côté, A. I. Benin, N. W. Ockwig, M. O'Keeffe, A. J. Matzger and O. M. Yaghi, Science, 2005, 310, 1166–1170.
2. S. Wan, J. Guo, J. Kim, H. Ihee and D. L. Jiang, Angew. Chem., Int. Ed., 2008, 47, 8826–8830.
3. R. W. Tilford, W. R. Gemmill, H. C. zur Loye and J. J. Lavigne, Chem. Mater., 2006, 18, 5296–5301.
4. E. L. Spitler and W. R. Dichtel, Nat. Chem., 2010, 2, 672–677.
5. R. W. Tilford, S. J. Mugavero, P. J. Pellechia and J. J. Lavigne, Adv. Mater., 2008, 20, 2741–2746.
6. H. M. El-Kaderi, J. R. Hunt, J. L. Mendoza-Cortés, A. P. Côté, R. E. Taylor, M. O'Keeffe and O. M. Yaghi, Science, 2007, 316, 268–272.
7. J. L. Mendoza-Cortés, S. S. Han, H. Furukawa, O. M. Yaghi and W. A. Goddard, J. Phys. Chem. A, 2010, 114, 10824–10833.
8. B. Assfour and G. Seifert, Microporous Mesoporous Mater., 2010, 133, 59–65.
9. H. Furukawa and O. M. Yaghi, J. Am. Chem. Soc., 2009, 131, 8875–8883.
10. S. S. Han, H. Furukawa, O. M. Yaghi and W. A. Goddard, J. Am. Chem. Soc., 2008, 130, 11580–11581.
11. R. Babarao and J. W. Jiang, Energy Environ. Sci., 2008, 1, 139–143.
12. G. Garberoglio, Langmuir, 2007, 23, 12154–12158.
13. S. S. Han, J. L. Mendoza-Cortés and W. A. Goddard III, Chem. Soc. Rev., 2009, 38, 1460–1476.
14. Y. Li and R. T. Yang, AIChE J., 2008, 54, 269–279.
15. X. Feng, L. Chen, Y. Dong and D. Jiang, Chem. Commun., 2011, 47, 1979–1981.
16. M. Dogru, A. Sonnauer, A. Gavryushin, P. Knochel and T. Bein, Chem. Commun., 2011, 47, 1707–1709.
17. N. A. A. Zwaneveld, R. Pawlak, M. Abel, D. Catalin, D. Gigmes, D. Bertin and L. Porte, J. Am. Chem. Soc., 2008, 130, 6678–6679.
18. S. Wan, J. Guo, J. Kim, H. Ihee and D. L. Jiang, Angew. Chem., Int. Ed., 2009, 48, 5439–5442.
19. X. Ding, J. Guo, X. Feng, Y. Honsho, J. Guo, S. Seki, P. Maitarad, A. Saeki, S. Nagase and D. Jiang, Angew. Chem., Int. Ed., 2011, 50, 1289–1293.
20. J. W. Colson, A. R. Woll, A. Mukherjee, M. P. Levendorf, E. L. Spitler, V. B. Shields, M. G. Spencer, J. Park and W. R. Dichtel, Science, 2011, 332, 228–231.
21. W. Niu, M. D. Smith and J. J. Lavigne, Cryst. Growth Des., 2006, 6, 1274–1277.
22. A. P. Côté, H. M. El-Kaderi, H. Furukawa, J. R. Hunt and O. M. Yaghi, J. Am. Chem. Soc., 2007, 129, 12914–12915.
23. For examples of other dynamic systems linked by boronate esters, see: (a) B. M. Rambo and J. J. Lavigne, Chem. Mater., 2007, 19, 3732–3739; (b) N. Christinat, R. Scopelliti and K. Severin, Angew. Chem., 2008, 120, 1874–1878; (c) N. Nishimura, K. Yoza and K. Kobayashi, J. Am. Chem. Soc., 2009, 132, 777–790.
24. O. Talu, C. J. Guo and D. T. Hayhurst, J. Phys. Chem., 1989, 93, 7294–7298.
25. N. Floquet, J. P. Coulomb, G. Weber, O. Bertrand and J. P. Bellatt, J. Phys. Chem. B, 2003, 107, 685–693.
26. Cooper et al. have since prepared COF-5 samples with BET surface areas exceeding 2200 m² g⁻¹, though much of this increase was attributed to changes in external surface area associated with differences in particulate size. See: N. L. Campbell, R. Clowes, L. K. Ritchie and A. I. Cooper, Chem. Mater., 2009, 21, 204–206.
27. S. R. Waldvogel, A. R. Wartini, P. H. Rasmussen and J. Rebek, Tetrahedron Lett., 1999, 40, 3515–3518.
28. P. Ballirano, R. Caminiti, C. Ercolani, A. Maras and M. A. Orru, J. Am. Chem. Soc., 1998, 120, 12798–12807.
29. S. Senthilarasu, Y. B. Hahn and S. H. Lee, J. Appl. Phys., 2007, 102 , 043512/043511–043512/043516.
30. A. Gouloumis, D. Gonzalez-Rodriguez, P. Vazquez, T. Torres, S. G. Liu, L. Echegoyen, J. Ramey, G. L. Hug and D. M. Guldi, J. Am. Chem. Soc., 2006, 128, 12674–12684.
31. G. F. Lu, Y. L. Chen, Y. X. Zhang, M. Bao, Y. Z. Bian, X. Y. Li and J. Z. Jiang, J. Am. Chem. Soc., 2008, 130, 11623–11630.
32. A. L. Korich and P. M. Iovine, Dalton Trans., 2010, 39, 1423–1431.
33. B. M. Rambo and J. J. Lavigne, Chem. Mater., 2007, 19, 3732–3739.
34. T. M. B. Islam, K. Yoshino, H. Nomura, T. Mizuno and A. Sasane, Anal. Sci., 2002, 18, 363–366.
35. J. Kua and C. R. Gyselbrecht, J. Phys. Chem. A, 2008, 112, 9128–9133.
36. Y. Tokunaga, H. Ueno, Y. Shimomura and T. Seo, Heterocycles, 2002, 57, 787–790.
37. P. De, S. R. Gondi, D. Roy and B. S. Sumerlin, Macromolecules, 2009, 42, 5614–5621.
38. A. L. Korich, A. R. Walker, C. Hincke, C. Stevens and P. M. Iovine, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 5767–5774.
39. C. Z. Cui, E. M. Bonder, Y. Qin and F. Jäkle, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 2438–2445.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details of the synthesis and characterization of additional compounds and materials; ¹H and ¹⁹F NMR and FTIR spectra associated with the mechanistic study and boronate ester exchange experiments. See DOI: 10.1039/c1sc00260k

---