Rhodium-catalyzed regioselective opening of vinyl epoxides with Et₃N·3HF reagent – formation of allylic fluorohydrins

Abstract

A highly regioselective rhodium-catalyzed ring-opening of vinyl epoxides with Et₃N·3HF reagent to form branched allylic fluorohydrins is described. The reaction occurs at room temperature under ambient air and relies on RhCOD₂BF₄ as an effective catalyst, providing the desired 1,2-addition allylic fluorohydrins in moderate to good yields with excellent levels of regioselectivity. Mechanistic studies demonstrate that the regioselective ring-opening of enantiopure vinyl epoxide occurs with inversion of stereochemistry.

---

Introduction

The many important roles that fluorine-containing compounds play in pharmaceuticals, agrochemicals, medical imaging, and high-performance materials have made the synthesis of this class of molecules a major focus in recent years.¹ Consequently, practical and elegant methods have been developed for the construction of aryl–F and alkyl–F bonds.^2,3 In contrast, the synthesis of allylic fluorides remains relatively underdeveloped.⁴ In recent years, a few examples of the transition-metal-catalyzed fluorination of allylic electrophiles with fluoride ions have been reported for the regio- and enantioselective preparation of allylic fluoride products.⁵ We recently reported that [IrClCOD]₂ is an effective catalyst for the regioselective fluorination of allylic trichloroacetimidates with Et₃N·3HF, providing secondary and tertiary allylic fluorides with good yields and excellent selectivity (Scheme 1a).⁶ In an effort to expand the capabilities of our fluorination method, we sought the development of a variant proceeding via the opening of vinyl epoxides, as depicted in Scheme 1b.

Scheme 1 Transition-metal-catalyzed fluorination.

Vinyl epoxides are commonly used building blocks for the synthesis of biologically active targets.⁷ They are quite reactive in the presence of transition-metal catalysts.⁸ Palladium is most frequently used to promote the ring-opening of vinyl epoxides with a wide variety of nucleophiles, leading to 1,4-addition products due to the electron-deficient nature of the vinyl epoxide oxygen.⁹ The 1,2-addition products can be achieved by choice of the ligand and co-catalysts.¹⁰ Rhodium has also been reported as an effective catalyst for the ring-opening of vinyl epoxides with both anilines and alcohols to provide trans-1,2-amino alcohols or alkoxy alcohols, respectively, in good yield and with high levels of diastereo- and regio-selectivity.¹¹ Even though the transition-metal-catalyzed asymmetric ring-opening of terminal epoxides^1e,12 and oxabicyclic olefins¹³ using the nucleophilic fluoride ion has been developed, to the best of our knowledge there are no reports on the transition-metal-catalyzed regioselective ring-opening of vinyl epoxides with nucleophilic fluoride reagents.¹⁴

Results and discussion

Initial studies focused on establishing the regioselectivity of the opening of vinyl epoxide 1 with Et₃N·3HF reagent (Table 1). Using our optimized conditions for the iridium-catalyzed fluorination of allylic trichloroacetimidates,⁶ we began our investigation with 2.5 mol% [IrCl(COD)]₂ (entry 1) in Et₂O at room temperature. The 1,2-addition allylic fluorohydrin 2A was obtained in 37% NMR yield. A number of rhodium(I) catalysts were then investigated. RhCOD₂BF₄ (entry 4) produced the best result to provide an 88% NMR yield of 2A.¹⁵ We subsequently screened a number of solvents, and THF (entry 7) provided 2A with the highest NMR yield (92%). In all cases with rhodium catalysts, the ring-opening of 1 proceeded to completion within 30 min, and the undesired 1,4-addition product was not observed. A control experiment (entries 8 and 9) was performed without the catalyst, and a 23–35% NMR yield of allylic fluorohydrin 2A was observed after 18 h.¹⁶ Olah's reagent performed poorly under rhodium conditions (entry 10) with only 12% conversion after the reaction had been stirring for 14 h. Other nucleophilic fluoride sources (CsF, TBAT, AgF, and KF) showed no reactivity (entries 11–14). For full analysis and isolated yield determination, crude fluorohydrin 2A was treated with 4-fluorobenzoyl chloride (entry 15), and the corresponding allylic fluoride 2 was isolated in 79% yield over two steps.¹⁷ To demonstrate the reproducibility and scalability of this reaction, 10 mmol of epoxide 1 (entry 15) was subjected to similar conditions: fluoride product 2 was obtained in 76% yield, which is comparable to the use of 0.2 mmol of substrate 1.

Table 1 Studies with ring-opening of vinyl epoxide^a

Entry	Catalyst	Loading (mol%)	F source	Solvent	Time (h)	NMR yield 2A^b (%)	Isolated yield 2^c (%)
a All reactions were conducted with 0.2 mmol of vinyl epoxide 1. b The NMR yield of allylic fluorohydrin 2A was determined by ^19F NMR analysis of the crude reaction mixture using PhCF3 as an internal standard. c For full analyses, crude product 2A was then transformed into the isolable allylic 4-fluorobenzoate 2 and the isolated yield was determined. d The reaction was performed with 10 mmol of vinyl epoxide 1.
1	[IrCl(COD)]2	2.5	Et3N·3HF	Et2O	2	37
2	[RhCl(COD)]2	2.5	Et3N·3HF	Et2O	2	55
3	RhCOD2OTf	5	Et3N·3HF	Et2O	0.5	70
4	RhCOD2BF4	5	Et3N·3HF	Et2O	0.5	88
5	RhCOD2BF4	5	Et3N·3HF	MTBE	0.5	71
6	RhCOD2BF4	5	Et3N·3HF	CH2Cl2	0.5	79
7	RhCOD2BF4	5	Et3N·3HF	THF	0.5	92
8	None	0	Et3N·3HF	Et2O	18	23
9	None	0	Et3N·3HF	THF	18	35
10	RhCOD2BF4	5	Pyridine·HF	THF	14	12
11	RhCOD2BF4	5	CsF	THF	18	0
12	RhCOD2BF4	5	TBAT	THF	18	0
13	RhCOD2BF4	5	AgF	THF	18	0
14	RhCOD2BF4	5	KF	THF	18	0
15	RhCOD2BF4	5	Et3N·3HF	Et2O	0.5	89 (88)^d	79 (76)^d

---

With the optimized conditions in hand, we proceeded to explore the scope and limitation of the rhodium-catalyzed regioselective ring-opening of vinyl epoxides (Table 2). We discovered that five- and seven-membered ring vinyl epoxides performed poorly under the same reaction protocol. We reasoned that acyclic vinyl epoxides possessing greater conformational flexibility may undergo ring-opening in the presence of RhCOD₂BF₄ and Et₃N·3HF reagent. To our excitement, various acylic vinyl epoxide substrates 3–13 (Table 2) proceeded smoothly under rhodium conditions. Because most epoxide substrates are too volatile for isolation, all allylic fluorohydrins 14A–24A were converted into the 4-fluoro-benzoate derivatives 14–24 for full analyses and isolated yield calculation. Overall, the reactivity of the vinyl epoxides is dependent on both the steric and electronic nature of the substituents at the C(1)- and C(2)-position of the allyl moiety. For instance, the presence of the methyl group at the C(2)-position of vinyl epoxides 3 and 4 (Table 2, entries 1 and 2) provided allylic fluorides 14 and 15 in 60% and 62% yield, respectively, compared to their epoxide counterparts 5–8 (43–58%, entries 3–6). We discovered that vinyl epoxides containing the C(1)-electron-donating groups (e.g. 4-methoxy-phenyl) only resulted in decomposition. In contrast, electron-withdrawing substituents at the C(1)-position of vinyl epoxides 4–7 (entries 2–5) provided allylic fluorohydrins 15A–18A in 50–76% NMR yield (43–62% of their isolable fluoride derivatives). On the other hand, increasing the steric hindrance at the C(2)-position (entry 7) requires a significantly longer reaction time (10 h vs. 1 h) and allylic fluoride 20 was isolated in 59% yield. Vinyl epoxide 10 (entry 8) bearing the primary alkyl functionality also performed adequately to provide the allylic fluorohydrin 21A in 61% NMR yield.

Table 2 Survey of secondary and tertiary vinyl epoxides^a

Entry	Vinyl epoxides	Time (h)	Allylic fluorohydrins NMR yield^a	Allylic fluorides isolated yield
a NMR yields of allylic fluorohydrins 14A–24A were determined by ^19F NMR analysis of the reaction mixture using PhCF3 as an internal standard. b Reactions were performed using 0.2 mmol of vinyl epoxides. c Reactions were performed using 0.5 mmol of vinyl epoxides. d An 11 : 1 mixture of 1,2-addition products 14A and 15A and their undesired 1,4-addition products was observed. e A 6 : 1 mixture of 1,2-addition product 20A and its undesired 1,4-addition product was observed. f A 1.7 : 1 mixture of 1,2-addition product 21A and undesired 1,4-addition product was observed.
1	3: R = H	1	14A (82%)^d	14 (60%)^b
2	4: R = Cl	1	15A (76%)^d	15 (62%)^c
3	5: R = 4-Cl	1	16A (64%)	16 (46%)^b
4	6: R = 3-Br	1	17A (63%)	17 (43%)^c
5	7: R = 4-CF3	4	18A (50%)	18 (46%)^b
6	8: R = H	1	19A (69%)	19 (58%)^b
7		10	20A (73%)^e	20 (59%)^b
8		3	21A (61%)^f	21 (48%)^b
9		1	22A (50%)	22 (49%)^b
			22A (50%)	22 (44%)^c
10	12: R = H	0.5	23A (80%)	23 (70%)^b
11	13: R = Me	0.5	24A (65%)	24 (56%)^b

---

The success of butadiene monoepoxide 11 (Table 2, entry 9) to form fluorohydrin 22A (50% NMR yield) and its isolable fluoride 22 (44–49%) led to our studies of isoprene monoepoxide 12 (entry 10) and C(2)-substituted methyl substrate 13 (entry 11). The reactions were fast (0.5 h) and proceeded with complete regioselectivity, providing tertiary allylic fluorohydrins 23A and 24A, respectively, in good NMR yields (65–80%). As expected, the more hindered vinyl epoxide 13 provided allylic fluorohydrin 24A in lower yield than that of substrate 12. These results are consistent with what has been observed when comparing the ring-opening of vinyl epoxide 3 (entry 1) to that of the more hindered 9 (entry 7), where the NMR yield is 82% vs. 73%.

With the ability to access a number of allylic fluorohydrins in moderate to good yields and with excellent regioselectivity, we next sought to establish the utility of this method by transforming tertiary allylic fluorohydrins into the tertiary alkyl fluorides 25 and 26 (Scheme 2), which could be potentially used as structural motifs of anticancer ether phospholipid compounds.¹⁸ One-pot regioselective ring-opening of vinyl epoxides 12 and 13 followed by hydrogenation and benzoylation furnished tertiary fluorides 25 and 26 in 52% and 50% yield, respectively.

Scheme 2 Conversion of allylic fluorohydrins into tertiary alkyl fluorides.

To determine if Lewis acid behavior alone was responsible for reactivity with RhCOD₂BF₄, we conducted control experiments with cyclohexyl epoxide 27 and styrene oxide 28 (Scheme 3) under the same rhodium conditions. No products were observed after 18 h, suggesting that RhCOD₂BF₄ is likely not acting as a Lewis acid and the presence of an alkene unit is required for the reaction to occur. These results are consistent with previously reported results for the rhodium-catalyzed regioselective ring-opening of vinyl epoxides with alcohols and anilines.¹¹

Scheme 3 Control experiments.

To gain some mechanistic insight, enantiopure epoxide (S)-11 was subjected to standard rhodium conditions (Scheme 4a). Allylic fluoride 22 was isolated in 43% yield with 69% ee. When the ring opening of (S)-11 was conducted in the absence of the catalyst, 22 was isolated in only 14% yield, albeit with a higher enantioselectivity (80% ee).^14b,19,20 On the other hand, opening of enantiopure epoxide (R)-8 containing the C(1)-phenyl group, with and without the rhodium catalyst (Scheme 4b), resulted in significant racemization of allylic fluoride 19 (10–13% ee).

Scheme 4 Stereospecific ring-opening of vinyl epoxides.

To establish the absolute stereochemistry of the allylic fluoride product, 22 was subjected to cross-metathesis with 4-bromo-styrene 29 (Scheme 5). The major enantiomer of diene product 30 (67% ee) was shown by X-ray crystallographic analysis to be R-configured.²⁰ Collectively, the data illustrates that the rhodium-catalyzed regioselective opening of vinyl epoxide with Et₃N·3HF proceeds with inversion of stereochemical configuration.²¹ This is opposite to what has been reported for palladium-catalyzed allylic fluorination.^5a,f

Scheme 5 Derivatization of fluoride 22 for X-ray analysis.

Based on the results obtained in Schemes 4 and 5, we propose the following mechanistic rationale for the regioselective ring-opening of vinyl epoxides lacking the C(1)-functional group.²² The rhodium catalyst first coordinates to the olefin of (S)-11 to form rhodium–alkene complex 31 (Scheme 6). Subsequent ionization of 31 provides the π-allylrhodium intermediate 32.²³ We hypothesize that hydrogen fluoride may first attack the rhodium center of 32 to form the corresponding fluoride–rhodium complex 33. Fluoride is widely regarded as a hard nucleophile,²⁴ which has been known to bind to the metal center of the allyl–metal complex in allylic substitution.²⁵ Protonation of the alkoxide followed by reductive elimination and benzoylation affords the corresponding (R)-allylic fluoride 22 with retention of stereochemistry, which is confirmed by X-ray crystallographic analysis.²⁰ To rationalize the excellent levels of regioselectivity for the 1,2-addition products, we hypothesize that the fluoride ion selectively attacks the more substituted allyl position of 34. Because (R)-22 was generated in 69% ee (Scheme 4a), we propose that nucleophilic attack by fluoride onto the π-allylrhodium complex 32 (Scheme 6) must occur faster than isomerization of complex 32 to its corresponding ent-32. Alternatively, the erosion of enantioselectivity observed in product 22 could arise if the allylic fluorohydrin intermediate is undergoing a degenerate substitution reaction with the Et₃N·3HF reagent.²⁶

Scheme 6 Proposed mechanism for rhodium-catalyzed ring-opening of vinyl epoxides lacking the C(1)-substituent.

Conclusions

We have developed a rhodium-catalyzed methodology that efficiently facilitates the regioselective ring-opening of a variety of vinyl epoxides with Et₃N·3HF to provide the corresponding 1,2-addition allylic fluorohydrins in moderate to good yields and with excellent regioselectivity. Mechanistic studies show that the opening of vinyl epoxides occurs with inversion of configuration. The operationally simple, mild, and rapid conditions make this method an attractive strategy for the incorporation of fluorine-18 ion into the allyl moiety of vinyl epoxides. Application of the ring-opening of vinyl epoxide procedure to the preparation of fluorine-18 radiotracers will be reported in due course.

Acknowledgements

We thank the University of Iowa for financial support and Dr Dale Swenson for X-ray crystallographic analysis. Q.Z. also thanks the University of Iowa for the Graduate Research Fellowship.

Notes and references

1. (a) K. Müller, C. Faeh and F. Diederich, Science, 2007, 317, 1881; (b) V. Gouverneur, Science, 2009, 325, 1630; (c) X.-L. Qui, X.-H. Xu and F.-L. Qing, Tetrahedron, 2010, 66, 789; (d) T. Furuya, A. S. Kamlet and T. Ritter, Nature, 2011, 473, 470; (e) C. Hollingworth and V. Gouverneur, Chem. Commun., 2012, 48, 2929.
2. Representative examples of non-radioactive aryl–fluorine bonds: (a) D. V. Yandulov and N. T. Tran, J. Am. Chem. Soc., 2007, 129, 1342; (b) D. A. Watson, M. Su, G. Teverovskiy, Y. Zhang, J. García-Fortanet, T. Kinzel and S. L. Buchwald, Science, 2009, 325, 1661; (c) N. D. Ball and M. S. Sanford, J. Am. Chem. Soc., 2009, 131, 3796; (d) X. Wang, T.-S. Mei and J.-Q. Yu, J. Am. Chem. Soc., 2009, 131, 7520; (e) T. Furuya, A. E. Strom and T. Ritter, J. Am. Chem. Soc., 2009, 131, 1662; (f) P. Tang, T. Furuya and T. Ritter, J. Am. Chem. Soc., 2010, 132, 12150; (g) P. Tang, W. Wang and T. Ritter, J. Am. Chem. Soc., 2011, 133, 11482; (h) T. Xu, X. Mu, H. Peng and G. Liu, Angew. Chem., Int. Ed., 2011, 50, 8176; (i) P. S. Fier and Hartwig, J. Am. Chem. Soc., 2012, 134, 10795; (j) T. Truong, K. Klimovica and O. Daugulis, J. Am. Chem. Soc., 2013, 135, 9342; (k) P. S. Fier, J. Luo and J. F. Hartwig, J. Am. Chem. Soc., 2013, 135, 2552.
3. Representative examples of non-radioactive alkyl–fluorine bonds: (a) D. D. Steiner, N. Mase and C. F. Barbas, Angew. Chem., Int. Ed., 2005, 44, 3706; (b) M. Marigo, D. Fielenbach, A. Braunton, A. Kjærsgaard and K. A. Jørgensen, Angew. Chem., Int. Ed., 2005, 44, 3703; (c) H. Jiang, A. Falcicchio, K. L. Jensen, M. W. Paixão, S. Bertelsen and K. A. Jørgensen, J. Am. Chem. Soc., 2009, 131, 7153; (d) T. Wu, J. Yin and G. Liu, J. Am. Chem. Soc., 2009, 131, 16354; (e) J. Erb, D. H. Paull, T. Dudding, L. Belding and T. Lectka, J. Am. Chem. Soc., 2011, 133, 7536; (f) P. Kwiatkowski, T. D. Beeson, J. C. Conrad and D. W. C. MacMillan, J. Am. Chem. Soc., 2011, 133, 1738; (g) V. Rauniyar, A. D. Lackner, G. L. Hamilton and F. D. Toste, Science, 2011, 334, 1681; (h) R. J. Phillips, K. Hiramatsu and F. D. Toste, J. Am. Chem. Soc., 2012, 134, 8376; (i) N. A. Cochrane, H. Nguyen and M. R. Gagne, J. Am. Chem. Soc., 2013, 135, 628; (j) F. Sladojevich, S. I. Arlow, P. Tang and T. Ritter, J. Am. Chem. Soc., 2013, 135, 2470.
4. (a) M. C. Pacheco, S. Purser and V. Gouverneur, Chem. Rev., 2008, 108, 1943; (b) S. Thibaudeau and V. Gouverneur, Org. Lett., 2003, 5, 4891; (c) B. Greedy, J. M. Paris, T. Vidal and V. Gouverneur, Angew. Chem., Int. Ed., 2003, 42, 3291; (d) S. Thibaudeau, R. Fuller and V. Gouverneur, Org. Biomol. Chem., 2004, 2, 1110; (e) S. Boldon, J. E. Moore and V. Gouverneur, Chem. Commun., 2008, 3622.
5. (a) M. H. Katcher and A. G. Doyle, J. Am. Chem. Soc., 2010, 132, 17402; (b) M. H. Katcher, A. Sha and A. G. Doyle, J. Am. Chem. Soc., 2011, 133, 15902; (c) C. Hollingworth, A. Hazari, M. N. Hopkinson, M. Tredwell, E. Benedetto, M. Huiban, A. D. Gee, J. M. Brown and V. Gouverneur, Angew. Chem., Int. Ed., 2011, 50, 2613; (d) A. M. Lauer and J. Wu, Org. Lett., 2012, 14, 5138; (e) E. Benedetto, M. Tredwell, C. Hollingworth, T. Khotavivattana, J. M. Brown and V. Gouverneur, Chem. Sci., 2013, 4, 89; (f) C. Xue, X. Jiang, C. Fu and S. Ma, Chem. Commun., 2013, 49, 5651; (g) M.-G. Braun, M. H. Katcher and A. G. Doyle, Chem. Sci., 2013, 4, 1216.
6. J. J. Topczewski, T. J. Tewson and H. M. Nguyen, J. Am. Chem. Soc., 2011, 133, 19318.
7. Comprehensive Organic Synthesis, ed. B. M. Trost and I. Fleming, Pergamon Press, New York, 1991, vol. 6.
8. B. M. Trost and M. L. Crawley, Chem. Rev., 2003, 103, 2921.
9. (a) J. Tsuji, H. Kataoka and Y. Kobaysahi, Tetrahedron Lett., 1981, 22, 2575; (b) B. M. Trost and G. A. Molander, J. Am. Chem. Soc., 1981, 103, 5969; (c) B. M. Trost and S. R. Angle, J. Am. Chem. Soc., 1985, 107, 6123.
10. Several representative examples, see: (a) B. M. Trost and A. Tenaglia, Tetrahedron Lett., 1988, 29, 2931; (b) B. M. Trost, E. J. McEachern and F. D. Toste, J. Am. Chem. Soc., 1998, 120, 12702–12703; (c) B. M. Trost and E. J. McEachern, J. Am. Chem. Soc., 1999, 121, 8649–8650; (d) X.-Q. Yu, F. Yoshimura, F. Ito, M. Sasaki, A. Hirai, K. Tanino and M. Miyashita, Angew. Chem., Int. Ed., 2008, 47, 750–754.
11. K. Fagnou and M. Lautens, Org. Lett., 2000, 2, 2319.
12. (a) S. Bruns and G. Haufe, J. Fluorine Chem., 2000, 104, 247; (b) G. Haufe, S. Bruns and M. Runge, J. Fluorine Chem., 2001, 112, 55; (c) G. Hauge and S. Bruns, Adv. Synth. Catal., 2002, 344, 165; (d) J. A. Kalow and A. G. Doyle, J. Am. Chem. Soc., 2010, 132, 3268; (e) J. A. Kalow and A. G. Doyle, J. Am. Chem. Soc., 2011, 133, 16001.
13. J. Zhu, G. C. Tsui and M. Lautens, Angew. Chem., Int. Ed., 2012, 51, 12353.
14. There are two examples of regioselective opening of vinyl epoxides with triethylamine trihydrofluoride in the absence of the metal catalyst, see: (a) A. Hedhli and A. Baklouti, J. Fluorine Chem., 1995, 70, 141; (b) L. Hunter, D. O'Hagan and A. M. Z. Sawin, J. Am. Chem. Soc., 2006, 128, 16422–16423.
15. To confirm the relative stereochemistry of 2A, it was converted into allylic benzoate, which underwent hydrogenation to afford the trans-1,2-fluoride whose relative stereochemistry was confirmed by ¹H NMR and COSY analyses. See ESI† for details.
16. Heating the reaction to 40 °C and 60 °C in THF resulted in 37% and 30% NMR yield, respectively.
17. Although THF provides fluorohydrin 2A in higher NMR yield than Et₂O, use of Et₂O as a solvent in rhodium-catalyzed ring-opening of vinyl epoxide 1 resulted in a biphasic mixture. This phase separation eliminates the need for an aqueous workup, and the crude product 2A was extracted from a biphasic mixture and subsequently treated with 4-flurobenzoyl chloride to form isolable allylic fluoride 2.
18. (a) R. Bittman, N. M. Witzke, T.-C. Lee, M. L. Blank and F. Snyder, J. Lipid Res., 1987, 28, 733–738; (b) G. Haufe and A. Burchardt, Eur. J. Org. Chem., 2001, 4501–4507; (c) A. Burchardt, T. Takahashi, Y. Takeuchi and G. Haufe, J. Org. Chem., 2001, 66, 2078–2084; (d) R. Bittman, H.-S. Byun, K. C. Reddy, P. Samadder and G. Arthur, J. Med. Chem., 1997, 40, 1391–1395.
19. It is known that regioselective opening of enantiopure vinyl epoxide 11 provided allylic fluorohydrin with opposite stereochemistry.^14a.
20. See the ESI† for details.
21. This result is consistent with what has been reported for the rhodium-catalyzed regioselective opening of vinyl epoxides with alcohols and anilines.¹¹.
22. Based on the result obtained with epoxide R-(8) in Scheme 4b, we hypothesized that vinyl epoxides bearing the C(1)-substituted group are likely to occur via S_N1 mechanism.
23. For an alternative enylrhodium intermediate, see: P. A. Evans and J. D. Nelson, J. Am. Chem. Soc., 1998, 120, 5581.
24. R. G. Parr and R. G. Pearson, J. Am. Chem. Soc., 1983, 105, 7512–7516.
25. B. M. Trost, T. Zhang and J. D. Sieber, Chem. Sci., 2010, 1, 427.
26. A. Hazari, V. Gouverneur and J. M. Brown, Angew. Chem., Int. Ed., 2009, 48, 1296.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedures, characterization data for all new compounds. CCDC 964273. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c3sc51949j

---