Controlling Reactivity by Geometry in Retro-Diels-Alder Reactions

Article pubs.acs.org/JACS

Cite This: J. Am. Chem. Soc. 2017, 139, 16768-16771

Controlling Reactivity by Geometry in Retro-Diels−Alder Reactions under Tension Richard Stevenson and Guillaume De Bo* School of Chemistry, University of Manchester, Oxford Road, Manchester M13 9PL, United Kingdom S Supporting Information *

ABSTRACT: Mechanical force, with its ability to distort, bend, and stretch chemical bonds, is unique in the way it activates chemical reactions. In polymer mechanochemistry, the force is transduced in a directional fashion, and the efficiency of activation depends on how well the force is transduced from the polymer to the scissile bond in the mechanophore (i.e., mechanochemical coupling). We have investigated the effects of regio- and stereochemistry on the rate of force-accelerated retro-Diels−Alder reactions of furan/maleimide adducts. Four adducts, presenting an endo or exo configuration and proximal or distal geometry, were activated in solution by ultrasoundgenerated elongational forces. A combination of structural (1H NMR) and computational (CoGEF) analyses allowed us to interrogate the mechanochemical activation of these adducts. We found that, unlike its thermal counterpart where the reactivity is dictated by the stereochemistry, the mechanical reactivity is mainly dependent on the regiochemistry. Remarkably, the thermally active distal-exo adduct becomes inert under tension due to poor mechanochemical coupling.

■

■

INTRODUCTION

© 2017 American Chemical Society

RESULTS AND DISCUSSION

Our study focuses on four different isomers (1−4, Chart 1) presenting an endo or exo configuration and proximal or distal geometry (referring to the furan handle being close or far from the furan/maleimide junction, respectively; see also Figure 5). These adducts were incorporated into the central region of a

Mechanical force is remarkable in its ability to direct chemical transformations along reaction pathways that are otherwise inaccessible to other methods of activation, such as Moore’s anti-Woodward−Hoffmann electrocyclic ring opening1 or Craig’s transition-state trapping.2 At the origin of this unusual behavior is the way the molecular backbone of a putative forcesensitive moiety (mechanophore) is deformed upon application of a directional force.3a−c Depending on how well the force is transduced to the reactive bond(s) of a particular functional group (i.e., mechanochemical coupling), mechanical activation can lead to an enhancement3a or a reduction in reactivity.4a−e For instance, it has been shown that the reactivity of a mechanophore is affected by stereo-,2,5a−f regio-,6a−c and even topological isomerism7 as well as the nature of the linker joining the mechanophore to the force transducing polymer.8a−e Here, we investigate the regio- and stereochemical effects in the force-accelerated retro-Diels−Alder (rDA) reaction of furan/maleimide adducts. This couple is well suited for the fabrication of thermally mendable materials due to its relatively low exchange temperature.9 Its mechanical resistance has been explored on surfaces,10 in gels11 and in solution,12 but the effects of isomerism have not yet been explored. Here we show that unlike its thermal counterpart, where the reactivity is dictated by the stereochemistry, the mechanical reactivity is mainly dependent on the regiochemistry. Remarkably, the thermally active distal-exo adduct becomes inert under tension.

Chart 1. Adducts Used in This Study

Received: August 20, 2017 Published: October 31, 2017 16768

DOI: 10.1021/jacs.7b08895 J. Am. Chem. Soc. 2017, 139, 16768−16771

Article

Journal of the American Chemical Society

Figure 1. Thermal activation of mechanophores 1c−4c (a) and conversion (b) determined by 1H NMR (Mn = 67−79 kDa). Solid lines correspond to a linear fit (R2 = 0.976, 0.988, 0.976, and 0.977 for 1c to 4c, respectively). See the Supporting Information for details.

Figure 2. Mechanical activation of mechanophores 1c−4c (a) and molecular weight dependence on the rate of activation (b). Each point represents a duplicate. Solid lines correspond to a linear fit (R2 = 0.953, 0.992, 0.928, and 0.952 for 1c−4c, respectively). See Supporting Information for details. Conditions: 1c−4c in CH3CN (∼1 mg/mL), 5−10 °C, US (20 kHz, 13.0 W/cm2, 1 s ON/2 s OFF).

poly(methyl acrylate) (PMA) backbone by single-electrontransfer living radical polymerization (SET-LRP) of methyl acrylate from bifunctional initiators 1b−4b.13a,b A total of 23 polymers 1c−4c, with Mn ranging from 41 to 132 kDa, were obtained with a low polydispersity (D̵ = 1.07−1.19; see the Supporting Information for details). The thermal reactivity of the four isomers was probed by heating polymers 1c−4c at 95 °C in toluene (Figure 1). As expected, both endo stereoisomers 1c and 3c cleave much faster than their exo counterparts 2c and 4c (Figure 1 and Table 1).

the distal-exo adduct reacts more slowly than the distal-endo isomer, which could suggest an alternative reaction pathway. The 1H NMR of sonicated polymers provides structural information on the chemoselectivity of the force-induced cleavage of polymers 1c−4c (Figure 3). For the proximal isomers, the appearance of characteristic protons associated with the maleimide double bond (Ha, 6.72 ppm) and the aromatic protons of furan 6c (H1, 7.4 ppm; H2/H3, 6.35 ppm) as well as the shift and coalescence of the protons of the methylene unit linking furan 6 to the polymer (ΔδH5 = 0.20 and 0.46 ppm for 1c and 0.15 and 0.61 ppm for 2c) are indicative of a retro-Diels−Alder process (Figure 3a−f). Similarly, sonication of distal-endo adduct 3c regenerates furan 5c (H1, 7.45 ppm; H4, 7.39 ppm; H2, 6.41 ppm; ΔδH5 = 0.35 ppm) and maleimide 7c (Ha, 6.72 ppm, Figure 3h−j). However, unlike the other isomers, distal-exo adduct 4c is left completely unchanged after sonication (Figure 3i-l). The elongation/energy curves of adducts 1a−4a, obtained by CoGEF calculations (DFT B3LYP/6-31G*),16a−c show the dramatic difference in energy (Emax) and force (Fmax) required to reach the state of maximal deformation before the eventual rupture of a covalent bond (Figure 4a and Table 1). These simulations show that proximal isomers 1a and 2a as well as distal-endo 3a (albeit at higher Fmax) should follow a retroDiels−Alder pathway (see Figures S5−7). This is not the case for the distal-exo isomer 4a for which the simulation leads to an unselective cleavage, in agreement with what we observe experimentally (see Figure S8). The origin of the mechanoresistance of the distal-exo isomer can be found by looking at how the “scissile” bond responds to the elongation of the mechanophore (i.e., the mechanochemical coupling, Figure 4b). One can see that in the distal-endo and both proximal isomers, the “scissile” bond d elongates exponentially upon stretching of the mechanophore until scission occurs. In contrast, that same bond elongates very modestly in distal-exo and completely relaxes at the end, which again indicates that another bond breaks in the process.17

Table 1. Calculated and Experimental Activation Parameters of the Four Mechanophores Investigated adduct

Emax (kJ/mol)a

Fmax (nN)a

Mlim (kDa)

t1/2 at 95 °Cb (min)

1 2 3 4

261 324 545 866

3.6 3.8 4.3 5.8

25.6 24.8 30.0 33.3

25 200 25 600

a

For 1a−4a, determined from CoGEF calculations (see the Supporting Information for details). bFor 1a−4c (Mn = 67−79 kDa).

The mechanochemical activation of adducts 1−4 was performed in solution using high-intensity ultrasound. Typically, a dilute solution of polymers 1c−4c in acetonitrile (∼1 mg/mL) was sonicated at low temperature (5−10 °C) with a high-intensity probe (20 kHz, 13.0 W/cm2, 1 s ON/2 s OFF), and the evolution of the reaction was monitored by GPC (see the Supporting Information for details). The relative rates of degradation were determined using the Nalepa method14 and plotted against the polymer molecular weight15a,b to determine the limiting mass (Mlim) below which no activation occurs (Figure 2). In stark contrast with the thermally activated reaction, in which the endo adducts cleave 8−24× faster than their exo counterparts (Table 1), a change of selectivity is observed in the force-accelerated reaction where the proximal isomers react faster than the distal isomers (Figure 2). Moreover, in addition to the clear difference in reactivity between the proximal and the distal adducts, it also appears that 16769

DOI: 10.1021/jacs.7b08895 J. Am. Chem. Soc. 2017, 139, 16768−16771

Article

Journal of the American Chemical Society

Figure 3. Partial 1H NMR (400/500 MHz, CDCl3) of proximal-endo polymer 1c (Mn = 90 kDa) before (a) and after (b) sonication (60 min), and proximal-exo polymer 2c (Mn = 96 kDa) before (f) and after (e) sonication (120 min), along with reference polymers 6c (c) and 7c (d), and of distal-endo polymer 3c (Mn = 102 kDa) before (g) and after (h) sonication (120 min), distal-exo polymer 4c (Mn = 88 kDa) before (l) and after (k) sonication (240 min), along with reference polymers 5c (i) and 7c (j).

shows a linear relationship with Mlim and can be used to predict the mechanical activity of each isomer (Figure S10). Qualitatively, the difference in mechanochemical coupling can be rationalized by looking at how the “scissile” bonds align with the force vector in the stretched intermediates of the four isomers (Figure S11). It is obvious from Figure 5 that the “scissile” bond (shown in red) is relatively well aligned with the force vector in the three mechanophores undergoing a retrocycloaddition but not in the distal-exo isomer where it is almost orthogonal to the force vector. The alignment of a putative scissile bond with the force vector is a useful tool to

Figure 4. CoGEF of adducts 1a−4a (DFT B3LYP/6-31G*) (a) and mechanochemical coupling extracted from CoGEF calculations (b).

Moreover, an adjacent bond (bond c, Figure S9), which is meant to contract in the transition state of the rDA as the furan aromaticity is restored, elongates twice as much. Both factors contribute to explain the inhibition of retrocycloaddition in the distal-exo mechanophore. The mechanochemical coupling (defined as the slope of the exponential fit in Figure 4b)

Figure 5. Alignment of stretched intermediates of the four isomers investigated (from left to right: proximal-endo, proximal-exo, distalendo, distal-exo) with the corresponding force vector (double-headed arrows). “Scissile” bonds are shown in red (hydrogen atoms omitted for clarity). 16770

DOI: 10.1021/jacs.7b08895 J. Am. Chem. Soc. 2017, 139, 16768−16771

Article

Journal of the American Chemical Society

(5) (a) Lenhardt, J. M.; Ogle, J. W.; Ong, M. T.; Choe, R.; Martinez, T. J.; Craig, S. L. J. Am. Chem. Soc. 2011, 133, 3222−3225. (b) Kryger, M. J.; Munaretto, A. M.; Moore, J. S. J. Am. Chem. Soc. 2011, 133, 18992−18998. (c) Groote, R.; Szyja, B. M.; Leibfarth, F. A.; Hawker, C. J.; Doltsinis, N. L.; Sijbesma, R. P. Macromolecules 2014, 47, 1187− 1192. (d) Wang, J.; Kouznetsova, T. B.; Niu, Z.; Ong, M. T.; Klukovich, H. M.; Rheingold, A. L.; Martinez, T. J.; Craig, S. L. Nat. Chem. 2015, 7, 323−327. (e) Wang, J.; Kouznetsova, T. B.; Craig, S. L. J. Am. Chem. Soc. 2015, 137, 11554−11557. (f) Wang, J.; Kouznetsova, T. B.; Craig, S. L. J. Am. Chem. Soc. 2016, 138, 10410−10412. (6) (a) Konda, S. S. M.; Brantley, J. N.; Varghese, B. T.; Wiggins, K. M.; Bielawski, C. W.; Makarov, D. E. J. Am. Chem. Soc. 2013, 135, 12722−12729. (b) Gossweiler, G. R.; Kouznetsova, T. B.; Craig, S. L. J. Am. Chem. Soc. 2015, 137, 6148−6151. (c) Robb, M. J.; Kim, T. A.; Halmes, A. J.; White, S. R.; Sottos, N. R.; Moore, J. S. J. Am. Chem. Soc. 2016, 138, 12328−12331. (7) Lee, B.; Niu, Z.; Craig, S. L. Angew. Chem., Int. Ed. 2016, 55, 13086−13089. (8) (a) Ribas Arino, J.; Shiga, M.; Marx, D. J. J. Am. Chem. Soc. 2010, 132, 10609−10614. (b) Klukovich, H. M.; Kean, Z. S.; Ramirez, A. L. B.; Lenhardt, J. M.; Lin, J.; Hu, X.; Craig, S. L. J. Am. Chem. Soc. 2012, 134, 9577−9580. (c) Tian, Y.; Boulatov, R. ChemPhysChem 2012, 13, 2277−2281. (d) Klukovich, H. M.; Kouznetsova, T. B.; Kean, Z. S.; Lenhardt, J. M.; Craig, S. L. Nat. Chem. 2012, 5, 110−114. (e) Wang, J.; Kouznetsova, T. B.; Kean, Z. S.; Fan, L.; Mar, B. D.; Martinez, T. J.; Craig, S. L. J. Am. Chem. Soc. 2014, 136, 15162−15165. (9) Chen, X.; Dam, M. A.; Ono, K.; Mal, A.; Shen, H.; Nutt, S. R.; Sheran, K.; Wudl, F. Science 2002, 295, 1698−1702. (10) Lyu, B.; Cha, W.; Mao, T.; Wu, Y.; Qian, H.; Zhou, Y.; Chen, X.; Zhang, S.; Liu, L.; Yang, G.; Lu, Z.; Zhu, Q.; Ma, H. ACS Appl. Mater. Interfaces 2015, 7, 6254−6259. (11) Min, Y.; Huang, S.; Wang, Y.; Zhang, Z.; Du, B.; Zhang, X.; Fan, Z. Macromolecules 2015, 48, 316−322. (12) Duan, H.-Y.; Wang, Y.-X.; Wang, L.-J.; Min, Y.-Q.; Zhang, X.-H.; Du, B.-Y. Macromolecules 2017, 50, 1353−1361. (13) For recent reviews see: (a) Rosen, B. M.; Percec, V. Chem. Rev. 2009, 109, 5069−5119. (b) Anastasaki, A.; Nikolaou, V.; Nurumbetov, G.; Wilson, P.; Kempe, K.; Quinn, J. F.; Davis, T. P.; Whittaker, M. R.; Haddleton, D. M. Chem. Rev. 2016, 116, 835−877. (14) (a) Sato, T.; Nalepa, D. E. J. Appl. Polym. Sci. 1978, 22, 865− 867. Although this method is known to underestimate the rate of degradation of mechanophore-centered polymers. See: (b) Akbulatov, S.; Boulatov, R. ChemPhysChem 2017, 18, 1422−1450. However, it has proved effective to compare the rate of activation of structurally similar mechanophores (for a recent example, see ref 5b). (15) It has been shown recently that the length of the chain is a better descriptor of the mechanical activation than its molecular weight. However, since only PMA was used in this study, we decided to use Mn as a proxy for the chain length. See: (a) Schaefer, M.; Icli, B.; Weder, C.; Lattuada, M.; Kilbinger, A. F. M.; Simon, Y. C. Macromolecules 2016, 49, 1630−1636. (b) May, P. A.; Munaretto, N. F.; Hamoy, M. B.; Robb, M. J.; Moore, J. S. ACS Macro Lett. 2016, 5, 177−180. (16) (a) Beyer, M. J. Chem. Phys. 2000, 112 (17), 7307−7312 For recent reviews, see:. (b) Ribas-Arino, J.; Marx, D. Chem. Rev. 2012, 112, 5412−5487. (c) Stauch, T.; Dreuw, A. Chem. Rev. 2016, 116, 14137−14180. (17) Rupture of the terminal bond is observed instead, which is usually encountered in CoGEF during a nonselective process (see the Supporting Information for details).

quickly assess the reactivity of a potential mechanophore upon inspection of its molecular model. Quantitatively, the angle reporting the relative orientation of the two linkers in the reactive adducts can be used to predict their mechanical activity (see the Supporting Information for details).

■

CONCLUSIONS In conclusion, we have shown that the reactivity of four isomers of furan/maleimide Diels−Alder adducts varies greatly from thermal to mechanical activation. Singularly, the thermally active distal-exo isomer becomes mechanoresistant due to an ineffective mechanochemical coupling originating in the misalignment of the “scissile” bonds with the force vector. These results are particularly significant for the future development of thermally mendable materials, in which the mechanical properties could be tuned by mixing proximal and distal isomers in various proportions, and as a new addition to the dynamic covalent chemistry toolbox.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b08895. Detailed descriptions of CoGEF calculations, synthetic procedures, characterization of new compounds, and spectroscopic data (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Guillaume De Bo: 0000-0003-2670-6370 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the School of Chemistry of the University of Manchester for their support. We thank the EPSRC for a studentship to R.S. (NOWNANO-CDT) and the Royal Society for a University Research Fellowship to G.D.B.

■ ■

DEDICATION This paper is dedicated to the memory of Professor István E. Markó. REFERENCES

(1) Hickenboth, C. R.; Moore, J. S.; White, S. R.; Sottos, N. R.; Baudry, J.; Wilson, S. R. Nature 2007, 446, 423−427. (2) Lenhardt, J. M.; Ong, M. T.; Choe, R.; Evenhuis, C. R.; Martinez, T.; Craig, S. L. Science 2010, 329, 1057−1060. (3) (a) Beyer, M. K.; Clausen-Schaumann, H. Chem. Rev. 2005, 105, 2921−2948. (b) Caruso, M. M.; Davis, D. A.; Shen, Q.; Odom, S. A.; Sottos, N. R.; White, S. R.; Moore, J. S. Chem. Rev. 2009, 109, 5755− 5798. (c) Polymer Mechanochemistry; Boulatov, R., Ed.; Springer: Berlin, 2015; Vol. 369. (4) (a) Garcia-Manyes, S.; Liang, J.; Szoszkiewicz, R.; Kuo, T.-L.; Fernández, J. M. Nat. Chem. 2009, 1, 236−242. (b) Akbulatov, S.; Tian, Y.; Kapustin, E.; Boulatov, R. Angew. Chem., Int. Ed. 2013, 52, 6992−6995. (c) Krupička, M.; Sander, W.; Marx, D. J. Phys. Chem. Lett. 2014, 5, 905−909. (d) Lee, B.; Niu, Z.; Wang, J.; Slebodnick, C.; Craig, S. L. J. Am. Chem. Soc. 2015, 137, 10826−10832. (e) Akbulatov, S.; Tian, Y.; Huang, Z.; Kucharski, T. J.; Yang, Q.-Z.; Boulatov, R. Science 2017, 357, 299−303. 16771

DOI: 10.1021/jacs.7b08895 J. Am. Chem. Soc. 2017, 139, 16768−16771