Self-assembly between 1,4-diazabicyclo[2.2.2]octane and bis(hexafluoroalcohols): solid/liquid phase switching for catalyst recycling

Abstract

Bis(hexafluoroalcohol) derivatives readily self-assemble with 1,4-diazabicyclo[2.2.2]octane through hydrogen bonding. The DABCO® catalysts can be released from these crystalline adducts to promote organic transformations. Regeneration of H-bonds allows facile recovery and reuse of this “supramolecular supported catalyst” (>10 times).

---

The introduction of fluorine atoms in an organic molecule deeply impacts on its physical and chemical properties.¹ Notably, fluoroalkyl chains largely enhance the hydrogen bond donor ability of neighbouring hydroxyl functions, as exemplified by the associations between hexafluorinated alcohols R(CF₃)₂COH (HFA) and various H-bond acceptors such as ethers, phosphonates, amines,…^2,3 For example, the diamine 1,4-diazabicyclo[2.2.2]octane, also called triethylenediamine (TEDA),§ has been shown to form a solid adduct with n-Pr(CF₃)₂COH where H-bonds have been clearly brought to light by crystallography.⁴

Recently, we developed an original strategy for the recycling of amine-based organocatalysts through a non-covalent association with insoluble fluorous molecules: ionic pairing with a perfluoroalkyl acid,^5a and halogen bonding (X-bonding) with an iodoperfluoroalkane.^6–8 Due to the reversibility of such non-covalent interactions, the catalyst can be released from the fluorous partner in an appropriate reaction medium to promote a transformation. After completion of this reaction, the supramolecular fluorous catalyst precursor is regenerated, precipitated by addition of an organic solvent, and simply recovered by filtration (solvomorphic separation). Among the decisive advantages of this approach, it is worth noting that the fluorous adduct is readily available in a single step, and since there is no modification of the backbone of the catalyst, its efficacy is only depending on the reversibility of the non-covalent association. Moreover, such a strategy avoids acidic work-up commonly used to remove amines and incompatible with sensitive reaction products.

When this concept was applied to TEDA, the corresponding X-bonded adduct smoothly catalysed the Morita–Baylis–Hillman reaction,⁹ and it could be recycled up to 5 times.^6,10 However, the percentage of the recovered catalyst progressively decreased after each run (30% recovery overall after the 5th cycle), with consequences on reaction times and yields. Herein, we report that bis(hexafluorinated alcohols) and TEDA readily self-assemble to form a crystalline H-bond network. The reversibility of the association allows liquid/solid phase switching for effective delivery/recovery of the catalyst in various reactions (Morita–Baylis–Hillman reaction, tandem oxa-Michael addition/aldol condensation reaction, and cyanosilylation of ketones).¹¹

Among commercially available HFA, the bis(hexafluorinated alcohols) 1,3- and 1,4-bisHFAB constitute an original class since they possess two H-bond donor groups (Fig. 1).¹² In the case of TEDA as partner, an association with a bisHFAB could theoretically lead to infinite chains, and offering thus the possibility to trap this organocatalyst in a H-bond network. For this purpose, a solution of TEDA (mp 156 °C) in dichloromethane was added to a solution of 1,3-bisHFAB (mp 22 °C) in the same solvent: the crystalline solid 1a immediately precipitated (91% yield; mp 208 °C). In the same way, using 1,4-bisHFAB (mp 87 °C) led to compound 1b (89% yield; mp 240 °C). In order to determine the nature of the association in adducts 1a and 1b (H-bonding vs. alcoholate/ammonium salt), X-ray diffraction studies were performed (Fig. 2). The corresponding structures were obtained with a residual factor R-factor = 0.07 for 1a and 0.14 for 1b, allowing unambiguous location of hydrogen atoms. These analyses confirmed the expected hydrogen bonding nature of the association between HFAB and TEDA with d(O–H⋯N) ≤ 1.71 Å and ∠(OHN) ≥ 159° in adducts 1a and 1b (see ESI† for details).

Fig. 1 TEDA, 1,3- and 1,4-bisHFAB.

Fig. 2 Single crystal X-ray structures of H-bonded adducts of TEDA with 1,3-bisHFAB (1a, top) and 1,4-bisHFAB (1b, bottom).

The promoting potential as well as the recyclability of the bisHFAB/TEDA adducts 1 was assessed in the Morita–Baylis–Hillman reaction and compared to that of pure TEDA (Table 1). TEDA (20 mol%) was reacted with 4-nitrobenzaldehyde and methyl acrylate in methanol, and full conversion into the Morita–Baylis–Hillman adduct 2 was attained within 5 h (entry 1).^10a When the catalyst precursor 1a was used under the same conditions, reactants were also fully converted into 2, albeit in a longer reaction time (18 h). After evaporation of the volatiles, catalyst precursor 1a and product 2 were easily separated by addition of dichloromethane: 2 nicely solubilised in this solvent whereas 1a remained fully insoluble.¹³ After filtration, evaporation of the solvent in the filtrate afforded pure product 2 (97% yield; entry 2). It is likely that the slower kinetics of the reaction with 1a is due to the equilibrium between this precatalyst and the real active catalyst TEDA. Pleasingly, the recovered precatalyst 1a could be reused with the same reactants for 6 further iterative cycles without significant decrease in the activity: full conversion of the reactants into 2 within 18 h was required for the runs 1–5 (entries 2–6), while 22 h were required for a complete reaction in the 6th and 7th runs (entries 7 and 8). Moreover, the recovered adduct 1a was re-engaged to perform four further runs with other benzaldehydes (entries 9–12). Thus, 3-nitrobenzaldehyde and 2,4-dichlorobenzaldehyde reacted within 22 h and 24 h to afford 3 and 4 in 85% and 80% yields respectively (entries 9 and 10), whereas 4-chloro-3-nitrobenzaldehyde and 2,4-dinitrobenzaldehyde led to 5 and 6 in 95% and 86% yields after 18–20 h reaction time (entries 11 and 12). Interestingly, when the less reactive benzaldehyde was engaged as substrate with 1a, 80% conversion into product 7 was attained within 5 days, a reaction time close to that reported with pristine TEDA (entry 13).¹⁴ After precipitation, 1a was recovered (98%) whereas product 7 was afforded in 77% yield. Finally, precatalyst 1b was assessed for the synthesis of compound 2: it exhibited the same activity as 1a with excellent yields for the Morita–Baylis–Hillman product and a catalytic efficacy maintained over 7 runs (entries 14–20). Thus, these results show that the H-bonded adducts between bisHFAB and TEDA are effective recoverable catalysts for the Morita–Baylis–Hillman reaction. Moreover, the recyclability of 1a (11 iterative cycles) is far superior to that previously observed with the X-bonded catalyst (C₈F₁₇I)₂·TEDA (5 cycles).⁶

Table 1 Assessment of 1a and 1b as recyclable catalyst precursors in the Morita–Baylis–Hillman reaction between aromatic aldehydes and methyl acrylate^a

Entry	Catalyst	Cycle^b	R	Time/h	Product	Yield (%)
a Reaction conditions: aldehyde (3 mmol), methyl acrylate (6 mmol), MeOH (12 mmol), 1 (0.6 mmol). b Precatalyst 1 was recovered by using CH2Cl2 (2 mL): ≥95% recovery/run. c Reaction performed with 1a recovered from previous reactions. d 98% recovery of 1a.
1	TEDA	—	4-NO2	5	2	96
2	1a	1	4-NO2	18	2	97
3	1a	2	4-NO2	18	2	98
4	1a	3	4-NO2	18	2	98
5	1a	4	4-NO2	18	2	98
6	1a	5	4-NO2	18	2	97
7	1a	6	4-NO2	22	2	95
8	1a	7	4-NO2	22	2	97
9	1a	8^c	3-NO2	22	3	85
10	1a	9^c	2,4-Cl2	24	4	80
11	1a	10^c	4-Cl-3-NO2	18	5	95
12	1a	11^c	2,4-(NO2)2	20	6	86
13	1a	1	H	120	7	77^d
14	1b	1	4-NO2	18	2	99
15	1b	2	4-NO2	18	2	98
16	1b	3	4-NO2	18	2	96
17	1b	4	4-NO2	18	2	97
18	1b	5	4-NO2	18	2	96
19	1b	6	4-NO2	20	2	95
20	1b	7	4-NO2	22	2	95

---

In 2004, Lesch and Bräse reported that TEDA could be used as a catalytic Brønsted base to promote the condensation between salicylaldehydes and cyclic enones in a domino oxa-Michael addition/aldol reaction sequence, providing an elegant access to tetrahydroxanthen-1-ones.^15,16 However, the reaction was sluggish (1–7 days) and required the use of 0.5 equiv. of TEDA under specific conditions (degassed dioxane/water mixture as solvent, under sonication). Due to the significant amount of the catalyst used, its recovery would render the reaction more attractive. For this, our adduct 1a (1 equiv.) was engaged with 5-methoxysalicylaldehyde and cyclohexenone, and the reaction was just conducted in hot dioxane (Scheme 1). While 1a was insoluble at room temperature, the mixture became homogenous at 95 °C and, after 48 h at this temperature, 94% conversion into the target product 8 was reached. The reaction medium was then cooled to room temperature, and 1a crystallised while 3 remained in solution. This thermomorphic separation afforded 1a (92% recovery) whereas xanthenone 8 was obtained in 82% yield after purification.

Scheme 1 Tandem oxa-Michael addition/aldol condensation between a salicylaldehyde and cyclohexenone.

As a final example, the catalytic cyanosilylation of ketones with TMSCN (1.05 equiv.) was investigated (Scheme 2).¹⁷ Starting from heptan-2-one, under solvent-free conditions, precatalyst 1a (20 mol%) solubilised very fast and full conversion into silylated cyanohydrin 9 was obtained within 4.5 h. Unfortunately, at the end of the reaction 1a could not be precipitated. We assumed that this was due to a partial silylation of 1,3-bis(HFAB), hampering thus the regeneration of the H-bond network. However, since the O–Si bond in hexafluorinated silyl alcohols is weak, heating the reaction mixture at 60 °C under vacuum was sufficient to regenerate the solid 1a. Then, 1a was satisfyingly recovered (70%), and product 9 was obtained in 92% yield. Under similar conditions, the recycled catalyst precursor 1a was used with the more sterically hindered acetophenone as substrate. The latter was successfully converted within 18 h into the corresponding cyanohydrin 10 (86% yield).

Scheme 2 Cyanosilylation of ketones.

In summary, we have shown that commercially available bis(hexafluoroalcohols) and TEDA readily self-assemble through hydrogen bonds to form a crystalline compound characterized by X-ray diffraction. Due to the reversibility of the supramolecular association, the catalyst can be released and promote reactions: Morita–Baylis–Hillman reaction, tandem oxa-Michael addition/aldol condensation reaction and cyanosilylation of ketone. No tedious work-up procedure is required since reaction products and the H-bonded adduct are easily separated under solvo- or thermomorphic conditions, and this “ready for use supported catalyst” can be recycled at least ten times. As a perspective, this simple technique could be further applied to chiral compounds such as bis(dihydroquinidines) for enantioselective transformations.

Su Chen (undergraduate student, Univ. Paris Sud) is kindly acknowledged for her participation in this work. The authors are grateful to Central Glass Co. Ltd for the kind gift of 1,3- and 1,4-bisHFAB.

Notes and references

1. (a) J.-P. Bégué and D. Bonnet-Delpon, Bioorganic and Medicinal Chemistry of Fluorine, Wiley, Hoboken, 2008; (b) K. Uneyama, Organofluorine Chemistry, Blackwell Publishing, Oxford, 2006; (c) P. Kirsch, Modern Fluoroorganic Chemistry, Wiley, Weinheim, 2004; (d) R. E. Banks, B. E. Smart and J. C. Tatlow, Organofluorine Chemistry, Principles and Commercial Applications, Plenum, New York, 1994.
2. For some relevant articles, see: (a) L. Kong, J. Wang, X. Fu, Y. Zhong, F. Meng, T. Luo and J. Liu, Carbon, 2010, 48, 1262; (b) J. W. Grate, Chem. Rev., 2008, 108, 726; (c) A. Berkessel, J. A. Adrio, D. Hüttenhain and J. M. Neudörfl, J. Am. Chem. Soc., 2006, 128, 8421; (d) B. Juršić, M. Ladika and D. E. Sunko, Tetrahedron Lett., 1985, 26, 5323; (e) W. J. Middleton and R. V. Lindsey Jr., J. Am. Chem. Soc., 1964, 86, 4948.
3. (a) D. Vuluga, J. Legros, B. Crousse and D. Bonnet-Delpon, Chem. Commun., 2008, 4954; (b) A. Demeter, V. Mile and T. Bérces, J. Phys. Chem. A, 2007, 111, 8942; (c) J.-F. Berrien, M. Ourévitch, G. Morgant, N. E. Ghermani, B. Crousse and D. Bonnet-Delpon, J. Fluorine Chem., 2007, 128, 839; (d) N. C. Maiti, Y. Zhu, I. Carmichael, A. S. Serianni and V. E. Anderson, J. Org. Chem., 2006, 71, 2878.
4. D. Vuluga, J. Legros, B. Crousse, A. M. Z. Slawin, C. Laurence, P. Nicolet and D. Bonnet-Delpon, J. Org. Chem., 2011, 76, 1126.
5. (a) D. Vuluga, J. Legros, B. Crousse and D. Bonnet-Delpon, Chem.–Eur. J., 2010, 16, 1776; (b) a very close approach has been further reported for the same purpose: N. Lu, W.-H. Chang, W.-H. Tu and C.-K. Li, Chem. Commun., 2011, 47, 7227.
6. S. Dordonne, B. Crousse, D. Bonnet-Delpon and J. Legros, Chem. Commun., 2011, 47, 5855.
7. For interesting examples of fluorous/organic liquid phase switching through non-covalent interactions with fluorous acids or iodoperfluoroalkanes, see: (a) R. Corrêa da Costa, T. Buffeteau, A. Del Guerzo, N. D. McClenaghan and J.-M. Vincent, Chem. Commun., 2011, 47, 8250; (b) G. Gattuso, A. Pappalardo, M. F. Parisi, I. Pisagatti, F. Crea, R. Liantonio, P. Metrangolo, W. Navarrini, G. Resnati, T. Pilati and S. Pappalardo, Tetrahedron, 2007, 63, 4951; (c) C. Palomo, J. M. Aizpurua, I. Loinaz, M. J. Fernandez-Berridi and L. Irusta, Org. Lett., 2001, 3, 2361; (d) V. Chechik and R. M. Crooks, J. Am. Chem. Soc., 2000, 122, 1243; (e) for a review, see: J.-M. Vincent, J. Fluorine Chem., 2008, 129, 903.
8. For reviews on supported organocatalysts, see: (a) F. Cozzi, Adv. Synth. Catal., 2006, 348, 1367; (b) M. Benaglia, A. Puglisi and F. Cozzi, Chem. Rev., 2003, 103, 3401.
9. For pioneer works and reviews on the Morita–Baylis–Hillman reaction, see: (a) K.-i. Morita, Z. Suzuki and H. Hirose, Bull. Chem. Soc. Jpn., 1968, 41, 2815; (b) A. B. Baylis and M. E. D. Hillman, Chem. Abstr., 1972, 77, 34174q; Ger. Offen. 2155113, 1972; for reviews, see: (c) V. Declerck, J. Martinez and F. Lamaty, Chem. Rev., 2009, 109, 1; (d) D. Basavaiah, A. J. Rao and T. Satyanarayana, Chem. Rev., 2003, 103, 811; (e) D. Basavaiah, K. V. Rao and R. J. Reddy, Chem. Soc. Rev., 2007, 36, 1581.
10. Immobilised quinuclidines have also been reported for the same purpose: (a) on magnetic nanoparticles: S. Luo, X. Zheng, H. Xu, X. Mi, L. Zhang and J.-P. Cheng, Adv. Synth. Catal., 2007, 349, 2431; (b) on ionic liquids: X. Mi, S. Luo and J.-P. Cheng, J. Org. Chem., 2005, 70, 2338.
11. It is worth noting that outstanding results on the self-assembly of metals and ligands to afford self-supported catalysts have been reported by Ding's group: (a) L. Yu, Z. Wang, J. Wu, S. Tu and K. Ding, Angew. Chem., Int. Ed., 2010, 49, 3627; (b) K. Ding, Z. Wang, X. Wang, Y. Liang and X. Wang, Chem.–Eur. J., 2006, 12, 5188; (c) X. Wang, L. Shi, M. Li and K. Ding, Angew. Chem., Int. Ed., 2005, 44, 6362; (d) Y. Liang, Q. Jing, X. Li, L. Shi and K. Ding, J. Am. Chem. Soc., 2005, 127, 7694; (e) X. Wang and K. Ding, J. Am. Chem. Soc., 2004, 126, 10524.
12. It can be noted that bisHFABs have been used as additives in catalysis: (a) O. Coulembier, D. P. Sanders, A. Nelson, A. N. Hollenbeck, H. W. Horn, J. E. Rice, M. Fujiwara, P. Dubois and J. L. Hedrick, Angew. Chem., Int. Ed., 2009, 48, 5170; (b) M. M. Kabat, L. M. Garofalo, A. R. Daniewski, S. D. Hutchings, W. Liu, M. Okabe, R. Radinov and Y. Zhou, J. Org. Chem., 2001, 66, 6141.
13. Alternatively, warm toluene (ca. 50 °C) can be used instead of dichloromethane.
14. D. Basavaiah, K. R. Reddy and N. Kumaragurubaran, Nat. Protoc., 2007, 2, 2665.
15. B. Lesch and S. Bräse, Angew. Chem., Int. Ed., 2004, 43, 115.
16. While this domino reaction could occur either through a Morita–Baylis–Hillman reaction/Michael addition, or via a Michael addition/aldol condensation (where the diamine acted as a nucleophile in the first case, and as a base in the second), several experimental results by Lesch and Bräse pleaded in favour of the second path (see ref. 15).
17. S.-K. Tian, R. Hong and L. Deng, J. Am. Chem. Soc., 2003, 125, 9900.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental procedures and analytical data. CCDC reference numbers 855431–855432. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2cy00528j

‡ New address: Institut de Recherche en Chimie Organique Fine, Laboratoire COBRA, CNRS/Univ. Rouen/INSA Rouen, F-76130 Mont-Saint-Aignan, France. E-mail: julien.legros@univ-rouen.fr

§ 1,4-Diazabicyclo[2.2.2]octane is also commonly called DABCO® catalyst. DABCO® catalyst is a trademark owned by Air Products and Chemicals Inc.

---