Chinese Chemical Letters  2020, Vol. 31 Issue (12): 3250-3254   PDF    
Phosphine-phosphonium ylides as ligands in palladium-catalysed C2-H arylation of benzoxazoles
Zhenyu Yaoa, Xing Lina, Remi Chauvina,b,**, Lianhui Wanga, Emmanuel Grasb, Xiuling Cuia,*     
a Engineering Research Centre of Molecular Medicine of Ministry of Education, Key Laboratory of Fujian Molecular Medicine, Key Laboratory of Precision Medicine and Molecular Diagnosis of Fujian Universities, Key Laboratory of Xiamen Marine and Gene Drugs, School of Biomedical Sciences, Huaqiao University, Xiamen 361021, China;
b Laboratory of Coordination Chemistry (LCC), CNRS & Universite' de Toulouse (UPS, INP), Toulouse 31077 Cedex 4, France
Abstract: As balanced electron-rich P, C-chelating ligands, phosphine-phosphonium-ylides are considered for their ability to in situ promote palladium-catalysed direct C(sp2)-H arylation. Using methyl phosphonium salts of 2, 2'-bis(diphenylphosphino)-1, 1'-binaphtyl ("methyl-BINAPIUM") as ylide precursors under optimized reaction conditions, arylation of benzoxazole was found to proceed in moderate to high yield to give functional 2-aryl benzoxazoles. A strong anion effect of the non-salt free ylide was evidenced (TfO- > I- > PF6- ≈ salt-free). This first example of phosphonium ylides as ligands in catalytic C-H activation extends the prospect of their general implementation in homogeneous transition metal catalysis.
Keywords: Electron-rich    Ligands    Phosphine-phosphonium-ylides    Benzixazole    C-H activation    

Transition-metal-catalyzed direct C-H bond functionalization have experienced tremendous developments over the past decade, allowing novel key step approaches in the synthesis of natural products and therapeutic agents [1]. These reactions have required catalysts with high kinetic reactivity to ensure broad substrate scope, lower reaction temperature and catalyst loading, short time and favourable product distributions. Ligand development has also played a central role. While the efficiency of phosphine and N-heterocyclic carbenes (NHCs) as σ-donating ligands of catalytic transition metal centers has been widely exemplified [2, 3], general guidelines for the design of suitable ligand scaffolds to promote C-H arylation are still lacking. In this context, however, beyond electron-balanced monodentate phosphoramidites [4] or imidazolium ylides [4b], bidentate amino/imino-amides, such as mono-protected amino acids (MPAAs), were shown to be behave as LX N, N-ligands in enantioselective C(sp2)-H arylation, or through reverse polarity from aryl boronates [5]. Noteworthy is the recently reported use of a P, C-ligand for Pd-catalysed 2-arylation of benzoxazoles (actually a phosphine-NHC ligand in a labile PNC pincer framework) [6]. Such ligands enable access to a structurally diverse range of valuable products via C-H bond activation with superior chemo-, regio- and enantio-selectivity. But the potential of ligands in improving catalyst efficiency of C-H bond activation remains to be further explored. As a new prospect of our continuing studies on direct C(sp2)-H bond activation [7], we envisaged the implementation of particular P, C-ligands, i.e., phosphino-phosphonium ylides, hereafter referred to as [8] "PhosphYls" (Scheme 1). These ligands were shown to be stronger σ-donating ligands than NHCs [8b].

Download:
Scheme 1. Phosphine-phosphonium salts and ylides there of (PhosphYls) as LX-chelating ligands of late transition metals M (e.g., Ru, Rh, Pd).

By design, suitable PhosphYls are of limited structural variety, as they must be derived from monoalkyl-triarylphosphoniums, with a single typeofacidicP+CHunitwith low sterichindrance andmoderate pKa value, such as those of the P+CH3 moiety, to allow selective deprotonation and efficient coordination (Scheme 1). Under these requirements, the simplest PhosphYls are ylides of phosphonium salts of o-bis(diphenylphosphino)benzene (dppbz) such as L1·HX, which were shown to provide catalysts for malonate C-allylation (Pd) [9], asymmetric alkene hydrogenation or ketone hydrosilylation (Rh) [10], and theoretical olefin metathesis (Ru) [11]. Alternative PhosphYls satisfying the above requirement are methylenephosphorane derivatives of 2, 20-bis(diphenylphosphino)-1, 10-binaphtyl (BINAP), such as the methyl-BINAPIUM ylide L2 [12], usedinRh-catalysedhydogenation and hydrosilylation [13], or its methoxycarbonyl-stabilized versions, Yliphos (R = CO2Et, Scheme 1) used in Pd-catalysed C-allylation [14] (not tobe confused with YPhos, ylido-phosphines acting as strongly donating P ligands in Pd or Au catalysts) [15]. As the low pKa values of Yliphos favours partial de-coordination of the stabilized free ylide [17], exploratory investigations prompted the choice of the rac-methyl-BINAPIUM ylide L2 as a benchmark PhosphYl. On the tracks of pioneering works [16] and recent advances in Pd-catalysed C-H arylation of heterocycles [17], the transformation was chosen as a model to further evaluate the ability of PhosphYls to form in situ catalysts.

Indeed, while their electron-donating ability is expected to facilitate the step of the aryl halide oxidative addition, their bulky chelating characters appear appropriate for steric protection of the Pd center.

With the view to accessing functional hetero-aromatic systems without resorting to an ortho-directing group, the reaction of the benzoxazole substrates 1 with various aryl halides 2 was considered [7]. The target 2-aryl-benzoxazole products 3 indeed exhibit versatile properties [18], going from antibacterial [19], cholesterol ester transfer protein inhibition [20], adenosine receptor antagonism [21], amyloidogenesis inhibition [22], and thioflavine-like binding affinity to β-amyloid plaques [23], to fluorescence in the solid state [24].

As a soluble Pd(Ⅱ) salt widely used in C-H activation catalysis also exhibiting propensity to form Pd-C(sp3) bonds [25], Pd(OAc)2 was envisaged as catalyst precursor, to which non-salt-free PhosphYls could be added, after in situ deprotonation of parent phosphoniums. Considering the possible effect of the anion X- in non-salt-free ylides, well known in the Wittig reaction [26], several phosphonium salts L2· HX were compared (X = I, TfO, PF6).

Benzoxazole 1 and bromobenzene 2a, or 4-bromoanisole 2b, were selected as the model reactants. From a preliminary screening under various conditions, the following conditions were adopted: DMSO as a solvent, Cs2CO3 as a base and Pd(OAc)2 as catalyst precursor at 50 ℃ for 14 h (Table 1). At the outset, commercially available monodentate phosphines, PPh3, PCy3, PtBu3, were first selected as standard reference ligands, giving 3a in 48%, 45% and 37% yields, respectively (entries 1-3). Bidentate phosphines BINAP, dppm, o-dppbz, dppe and dppf also gave medium yields of 53%, 44%, 51%, 47% and 36%, respectively (entries 4-8). Under the same conditions, using the methylphosphonium salt L1· HX(X=I, TfO, PF6) of o-dppbz as pro-ligand (nonsalt-free PhosphYls L1/HX or L2/HX were separately prepared by deprotonation of L1·HX or L2·HX with nBuLi in DMSO solution during 15 min at room temperature), the coupling product 3a was isolated in 53%, 59% and 43% yields, respectively (entries 9-11). These results are similar to those obtained by bidentate phosphines, and then it was decided to change the o-dppbz backbone to the BINAP backbone. The use of the BINAPIUM ylide generated from L2·HI gave 3a in an improved yield of 65% (entry 12). The same ylide L2 generated from L2·HOTf gave 3a in a much higher yield of 93%, revealing a dramatic counterion effect (entry 13). This was confirmed by the result of L2·HPF6, giving 3a in a twice lower yield of 50% (entry 14), thus showing a correlation of the yield with the association effect between the anion and the Pd and/or P+ centers [27].

Table 1
Ligand effect in Pd-catalysed 2-phenylation of benzoxazole 1 with bromobenzene 2a.a

A systematic screening of the reaction conditions was then undertaken (Table 1). In the absence of any potential P-ligand, the reaction was found to proceed with 16% yield only (entry 15). It was also found that both the solvent and base are of critical importance. By reducing the amounts of Cs2CO3 from 1.0 equiv. to 0, the yields in 3a are reduced to 41%, 33% and 0% respectively (entries 16-18). Upon replacement of Cs2CO3 by weaker bases, such as Et3N or PPh3, product was not observed (entries 19 and 20). The use of a stronger base, such as NaHMDS, tBuOK or MeONa, gave lower yields in the range of 27%-39% (entries 21-23). When the same bases were used in THF instead of DMSO, the yields dropped to 12%-20% (entries 24-26). Other solvents, such as THF, DMF, MeCN, EtOH or NMP, also resulted in a dramatic decrease of the yield (< 15%, details see the Supporting information).

The yield in 3a was decreased from 93% (entry 13) with the salted ylide (L2/LiOTf) to 46% (entry 27) with the salt-free ylide, similar to the 50% yield (entry 14) obtained with the pseudo-salt-free ylide L2/LiPF6 (contrary to TfO-, PF6- is totally inert with respect to the Pd and P+ centers). Under the same conditions, 1 equiv. of LiOTf was added to the salt-free ylide L2, the yield in 3a was restored to 80%, thus demonstrating the key role of LiOTf (entry 28). The effect of LiOTf could be attributed to the interaction ability of the TfO- anion with the P+ and/or the Pd(Ⅱ) centers, by either electrostatic contact or coordination bonding [27].

With the optimized reaction conditions in hand, the scope of the substrates was examined (Scheme 2). Benzoxazole 1 reacted smoothly with bromobenzene 2a and its derivatives 2b-2t to give the anticipated products 3a-3t in moderate to high yields (35%-93%). Aryl bromides with a methyl group at the para-or meta-position(2c and 2d) gave the corresponding products 3c and 3d in 82% and 79% yields, respectively. When the methyl group was attached at the ortho-position, a lower 54% yield was obtained, revealing the steric hindrance of the reacting position. The electron-withdrawing substituent CF3 at the para-position of 2e allowed isolation of the aryl-benzoxazole 3e in a slightly higher yield (84%), while electron-donating OMe counterpart has the opposite effect on the corresponding product 3f (70%). Nevertheless, both weakly electron-withdrawing (F, Cl) and strongly electron-donating (NMe2, tBu) substituents gave higher yields (81%-88%), showing that the limiting step is not, at least not always, an SNAr-like oxidative addition process of the C-Br bond to the electron-rich Pd center (a favored by electron-withdrawing substituents on the aryl ring) [28]. For substituent F, however, the classical trend is restored through the relative reactivity of the para and ortho aryl bromide derivatives 2i and 2m (giving 3i and 3m in 81% and 80% yields) vs. the meta derivative 2k (giving 3k in 35% yield). This trend is also observed for the CF3 substituent, giving 41% yield in 3q for the meta position, vs. 84% yield in 3e for the para position. For a non-fluorinated electron-withdrawing substituent, the position effect of the formyl group remains consistent but much less dramatic, giving 80% yield in 3r for the meta position, vs. 86% yield in 3s for the para position. For the donating OMe substituent, a remarkable counter-effect of the position is observed: while para- and meta-bromoanisoles led to the expected products 3f and 3p in 70% and 62% yields, respectively, the more hindered ortho-bromoanisole 2o gave 3o with a significantly higher yield of 80%, revealing an assistance of the OMe group. The isosteric, but non-coordinating ortho-bromotoluene 2b gives 3b in 54% yield only. 2-Bromonaphthalene and brominated heterocycles such as 2-bromopyridine, 2-bromothiophene, and 3-bromothiophene were also found to be suitable substrates, affording the corresponding products 3t and 3v-3x in 60%-73% yields.

Download:
Scheme 2. Substrate scope. All reactions were carried out using 1 (0.4 mmol), 2 (0.48 mmol), Pd(OAc)2 (5 mol%), L2/LiOTf (5 mol), Cs2CO3 (2 equiv.) in DMSO (2.0 mL) at 50 ℃ for 14 h. Isolated yields.

On the basis of the previous literature reports and above experimental results, a tentative mechanism is proposed in Scheme 3. The PhosphYl ligand L2 could first coordinate in situ to the Pd(Ⅱ) center, prior to or after in situ reduction of Pd(Ⅱ) (e.g., by the phosphine end of L2 or DMSO) to Pd(0) in a zwitterionic palladate complex A similar to previously invoked. [12a, 29].

Download:
Scheme 3. Putative mechanism for PhosphYl-Pd-catalysed 2-arylation of benzoxazole: alternative to the depicted Pd(0)/Pd(Ⅱ) cycle, a Pd(Ⅱ)/Pd(Ⅳ) cycle might also be envisaged.

Then, oxidative-addition of PhBr would give the Pd(Ⅱ) intermediate B, from which the bromide anion would give the amido Pd(Ⅱ) complex D. Deprotonation of D would afford the zwitterionic palladate intermediate E, which would release the arylbenzoxazole product 3 while regenerating the catalytic pivot A. Alternative processes could also be envisaged [9], e.g., via a Pd(Ⅱ)/ Pd(Ⅳ) cycle instead of the Pd(0)/Pd(Ⅱ) cycle or through a transient isonitrile-phenolate resulting for the opening of the oxazole ring of 1 [7]. Nevertheless, the conversion σ/π-C → 3 + A could also proceed via a Pd-C-oxazolyl intermediate F from which the products would be directly released by reductive elimination, as invoked in the quinoline series [9, 30]. The propositions remain, however, quite speculative, as preliminary attempts did not allow isolation of any pre-catalytic L2/Pd complex (mixture of several unidentified species were observed by 31P/1H NMR spectra).

In conclusion, the disclosed C-H arylation allows 2-aryl-benzoxazoles to be produced in moderate to high yields over a night at 50 ℃. The use of the particular PhosphYl BINAPIUM ylide L2 proves to be superior to the use of classical phosphine ligands. The in situ catalytic system happens to be general for arylbromide substrates and specific hetero-unsaturated C-H substrates, i.e., benzoxazole. It is indeed found to be ineffective for related benzothiazole and imidazole substrates under standard conditions, but is thus selective. Considering the adjustable steric and electronic features of the ylide structure and the variety of ways to generate it from phosphonium precursors, prospects of optimization are open, in particular for extending the substrate scope. In this context, the axially chiral BINAPIUM backbone allows envisaging applications in enantioselective catalysis from prochiral substrates, e.g., in Cammidge-like coupling of 1-halo-naphthalenes with dissymmetric bulky (hetero)cycles [31]. Beyond the scope of the present report of in situ catalysis investigations, further efforts in coordination chemistry will be undertaken with the view to elucidating the actual structure of the catalyst. Finally, further extension of the potentialities of generic PhosphYl ligands in transition metal catalysis of other organic transformations deserves continuing investigations.

Ligand development played a significant role in Transition metal catalyzed C-H bond activations which required catalysts with high kinetic reactivity. The efficiency of phosphine and N-heterocyclic carbenes (NHCs) as σ-donating ligands of catalytic transition metal centers has been widely exemplified. Monodentate phosphoramidites and mono-protected amino acids (MPAAs) showed great enantioselectivity in C(sp3)-H arylation. Phosphine-NHC ligand showed high activity in C(sp3)-H arylation. These ligands enable access to a structurally diverse range of valuable products via C-H bond activation with superior chemo-, regio- and enantioselectivity. But the potential of ligands in improving catalyst efficiency of C-H bond activation remains to be further exploited.

In this context, we designed a particular P, C-ligands, i.e., phosphino-phosphonium ylides, hereafter referred to as "Phos-phYls". The ylide end being extremely σ-donating (more than an NHC end), the P, C-chelated Pd center is anticipated to be highly electron-rich, thus prone to accelerate the rate-determining step of oxidative addition of C-halogen bonds. Then Pd-catalyzed C-H arylation of heterocycles was chosen as model to detect reactivity of these ligands. The BINAPIUM ylide L2, a particular PhosphYl ligand, proves to be superior to phosphine ligands. A various of 2- aryl-benzoxazole products are provided in moderate to high yields after one night at 50 ℃.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We gratefully acknowledge the financial support from the National Natural Science Foundation of China (Nos. 21572072 and 21602064), 111 Project (No. BC2018061), and Postgraduate's (Y.Z. Yao) Innovative Fund in Scientific Research of Huaqiao University.

References
[1]
(a) R.A. Hughes, C.J. Moody, Angew. Chem. Int. Ed. 46(2007) 7930-7954;
(b) I. Osaka, R.D. McCullough, Acc. Chem. Res. 41(2008) 1202-1241.
[2]
(a) C.A. Tolman, Chem. Rev. 77(1977) 313-348;
(b) A.F. Littke, G.C. Fu, Angew. Chem. Int. Ed. 37(1998) 3387-3388.
[3]
(a) M. Scholl, S. Ding, C.W. Lee, et al., Org. Lett. 1(1999) 953-956;
(b) W.A. Herrmann, Angew. Chem. Int. Ed. 41(2002) 1290-1309;
(c) M.C. Jahnke, F.E. Hahn, Top. Organomet. Chem. 30(2010) 95-129.
[4]
J. Pedroni, N. Cramer, Chem. Commun. 51 (2015) 17647-17657. DOI:10.1039/C5CC07929B
[5]
(a) K. Xiao, D. Lin, M. Miura, et al., J. Am. Chem. Soc. 136(2014) 8138-8142;
(b) P. Shen, L. Hu, Q. Shao, et al., J. Am. Chem. Soc. 140(2018) 6545-6549;
(c) Q. Shao, J. He, Q. Wu, et al., ACS Catal. 7(2017) 7777-7782;
(d) C. Pi, Y. Li, X.L. Cui, et al., Chem. Sci. 4(2013) 2675-2679;
(e) D. Gao, Q. Gu, S.L. You, J. Am. Chem. Soc. 138(2016) 2544-2547;
(f) J. Wu, X.L. Cui, L. Chen, et al., J. Am. Chem. Soc. 131(2009) 13888-13889.
[6]
Y. Li, X. Yu, Y. Wang, et al., Organometallics 37 (2018) 979-988. DOI:10.1021/acs.organomet.8b00005
[7]
(a) Z. Yang, X. Lin, L.H. Wang, et al., Org. Chem. Front. 4(2017) 2179-2183;
(b) C.S. Kuai, L.H. Wang, H. Cui, et al., ACS Catal. 6(2015) 186-190;
(c) L.H. Wang, D. Xiong, L.H. Jie, et al., Chin. Chem. Lett. 29(2018) 907-910;
(d) T. Wan, S.D. Du, C. Pi, et al., ChemCatChem 11(2019) 3791-3796;
(e) X. Mi, Y. Kong, J. Zhang, et al., Chin. Chem. Lett. 30(2019) 2295-2298;
(f) Y. He, C. Pi, Y. Wu, et al., Chin. Chem. Lett. 31(2020) 396-400.
[8]
(a) H. Schmidbaur, Angew. Chem. Int. Ed. 22(1983) 907-927;
(b) Y. Canac, C. Lepetit, M. Abdalilah, et al., J. Am. Chem. Soc.130(2008) 8406-8413;
(c) E.P. Urriolabeitia, Top. Organomet. Chem. 130(2010) 15-48;
(d) T. Scherpf, R. Wirth, K.S. Feichtner, et al., Angew. Chem. Int. Ed. 54(2015) 8542-8546;
(e) C. Lepetit, V. Maraval, Y. Canac, et al., Coord. Chem. Rev. 308(2016) 59-75;
(f) E. Serrano, T. Soler, E.P. Urriolabeitia, Eur. J. Inorg. Chem. 2017(2017) 2220-2230;
(g) L.T. Scharf, V.H. Gessner, Inorg. Chem. 56(2017) 8599-8607;
(h) V.H. Gessner, Struct. Bond. 177(2018) 117-157.
[9]
(a) R. Zurawinski, B. Donnadieu, M. Mikolajczyk, et al., J. Organomet. Chem. 689(2004) 380-386;
(b) Y. Canac, C. Duhayon, R. Chauvin, Angew. Chem. Int. Ed. 46(2007) 6313-6315.
[10]
R. Zurawinski, R. Donnadieu, M. Mikolajczyk, et al., Organometallics 22 (2003) 4810-4817. DOI:10.1021/om030343g
[11]
L. Arnedo, R. Chauvin, A. Poater, Catalysts 7 (2017) 1-12.
[12]
(a) P. Leglaye, B. Donnadieu, J.J. Brunet, R. Chauvin, Tetrahedron Lett. 39(1998) 9179-9182;
(b) M. Soleilhavoup, L. Viau, G. Commenges, et al., Eur. J. Inorg. Chem. (2003) 207-212.
[13]
(a) L. Viau, C. Lepetit, G. Commenges, et al., Organometallics 20(2001) 808-810;
(b) C. Canal, C. Lepetit, M. Soleilhavoup, et al., Affinidad 61(2004) 298-303.
[14]
(a) T. Ohta, T. Fujii, N. Kurahashi, et al., Sci. Eng. Rev. Doshisha Univ. 39(1998) 133-141;
(b) T. Ohta, H. Sasayama, O. Nakajima, et al., TetrahedronAsymmetry 14(2003) 537-542.
[15]
(a) T. Scherpf, R. Wirth, S. Molitor, et al., Angew. Chem. Int. Ed. 54(2015) 8542-8546;
(b) L.T. Scharf, V.H. Gessner, Inorg. Chem. 56(2017) 8599-8607.
[16]
(a) R.F. Heck, J. Am. Chem. Soc. 90(1968) 5518-5526;
(b) T. Mizoroki, K. Mori, A. Ozaki, Bull. Chem. Soc. Jpn. 44(1971) 581-583;
(c) A. Jutand, Introduction, in: M. Oestreich (Ed.), The Mizoroki-Heck Reaction, John Wiley & Sons, 2009, pp. 1-50;
(d) Y.P. Wang, H.M. Lee, J. Organomet. Chem. 791(2015) 90-98.
[17]
(a) J.C. Lewis, J. Wu, R.G. Bergman, et al., Angew. Chem. Int. Ed. 45(2006) 1589-1591;
(b) H.Q. Do, O. Daugulis, J. Am. Chem. Soc. 129(2007) 12404-12405;
(c) J.C. Lewis, A.M. Berman, R.G. Bergman, et al., J. Am. Chem. Soc. 130(2008) 2493-2500;
(d) D. Zhao, W. Wang, F. Yang, et al., Angew. Chem. Int. Ed. 48(2009) 3296-3300.
[18]
C.S. Demmer, L. Bunch, J. Med. Chem. 97 (2015) 778-785. DOI:10.1016/j.ejmech.2014.11.064
[19]
S. Abdeen, T. Kunkle, N. Salim, et al., J. Med. Chem. 61 (2018) 7345-7357. DOI:10.1021/acs.jmedchem.8b00989
[20]
C.J. Smith, A. Ali, L. Chen, et al., Bioorg. Med. Chem. Lett. 20 (2010) 346-349. DOI:10.1016/j.bmcl.2009.10.099
[21]
R. Duroux, L. Agouridas, N. Renault, et al., Eur. J. Med. Chem. 144 (2018) 151-163. DOI:10.1016/j.ejmech.2017.12.007
[22]
S.M. Johnson, S. Connelly, I.A. Wilson, et al., J. Med. Chem. 51 (2008) 260-270. DOI:10.1021/jm0708735
[23]
X. Wang, M. Cui, J. Jia, et al., Eur. J. Med. Chem. 89 (2015) 331-339. DOI:10.1016/j.ejmech.2014.10.046
[24]
C. Carayon, S. Fery-Forgues, Photochem. Photobiol. Sci. 16 (2017) 1020-1035. DOI:10.1039/C7PP00112F
[25]
A. Poveda, I. Alonso, M.A. Fernandez-Ibanez, Chem. Sci. 5 (2014) 3873-3882. DOI:10.1039/C4SC00848K
[26]
P.A. Byrne, D.G. Gilheany, J. Am. Chem. Soc. 134 (2012) 9225-9239. DOI:10.1021/ja300943z
[27]
(a) T. Hayashida, H. Kondo, J.I. Terasawa, et al., J. Organomet. Chem. 692(2007) 382-394;
(b) V.J. Argyle, L.M. Woods, M. Roxburgh, et al., CrystEngComm 15(2013) 120-134.
[28]
(a) H.M. Senn, T. Ziegler, Organometallics 23(2004) 2980-2988;
(b)L.J. Gooßen, D. Koley, H.L.Hermann, etal., Organometallics 24(2005)2398-2410.
[29]
R. Chauvin, Eur. J. Inorg. Chem. 2000 (2000) 577-591.
[30]
X. Ren, P. Wen, X. Shi, et al., Org. Lett. 15 (2013) 5194-5197. DOI:10.1021/ol402262c
[31]
A.N. Cammidge, K.V.L. Crepy, Chem. Commun. 36 (2000) 1723-1724.