Article pubs.acs.org/Organometallics
Selective and Nonselective Aza-Michael Additions Catalyzed by a Chiral Zirconium Bis-Diketiminate Complex Ibrahim El-Zoghbi, Myriam Kebdani, Todd J. J. Whitehorne, and Frank Schaper* Centre in Green Chemistry and Catalysis, Department of Chemistry, Université de Montréal, C. P. 6128 Succ. Centre-Ville, Montréal, Québec H3T 3J7, Canada S Supporting Information *
ABSTRACT: Reaction of the chiral bis-diketiminate complex rac- or (R,R)-C6H10(nacnacXyl)2ZrCl2 with AgOTf yielded the corresponding bis-triflate complex. The complex geometry changes from distorted octahedral in the dichloride complex to a pseudotetrahedral coordination involving π coordination of the diketiminate ligands. The bis-triflate complex is highly active for aza-Michael additions with turnover frequencies of 20000/h for the addition of morpholine to acrylonitrile and 1000/h for the addition of morpholine to methacrylonitrile. The enantioselectivities of the latter reaction in various solvents were low, never surpassing 19% ee. The reaction is first-order in olefin concentration and second order in amine concentration, which is explained by its participation as a base in the reaction mechanism. The presence of catalytic amounts of triethylamine slightly increases the observed rate constants and reduces the reaction order in amine to first order. Other activated alkenes such as methacrylonitrile, crotonitrile, methyl acrylate, and cyclohexenone can be employed, but no reactivity is observed toward styrene or vinyl ethers. Primary amines, secondary amines, and anilines can be employed as nucleophiles with activities correlating with their nucleophilicity, but the catalyst is unstable in the presence of alcohols.
■
INTRODUCTION Although diketiminate (“nacnac”) metal complexes were first synthesized in 1968,1 they did not have any notable impact on the field until the late 1990s. Following Brookhart’s seminal work on α-diimine catalysts in late-metal olefin polymerization,2 it was a small step to anionic β-diketiminate ligands, following the same design principle.3 Although not as successful as α-diimine for olefin polymerization, the potential of βdiketiminate ligands as spectator ligands, in particular nacnacXyl and nacnacdipp (Xyl = 2,6-dimethylphenyl, dipp = 2,6diisopropylphenyl), was soon recognized.4 First reports of a zirconium diketiminate complex were made by Lappert and co-workers in 19945 and Mittal et al. in 1995,6 soon followed by more extensive work by others7 that was often oriented toward olefin polymerization.7d−g,l−n,q−t Bisdiketiminate zirconium complexes, nacnac2ZrX2, were first prepared by Collins, who reported a κ2 coordination of diketiminate and a cis-octahedral geometry for these complexes.7a Mixed diketiminate−cyclopentadienyl (or indenyl) complexes, (nacnac)CpZrX 2 , on the other hand, adopt a pseudotetrahedral symmetry with a coordination of the diketiminate ligand via its π system.7a It soon became apparent that bis-diketiminate zirconium complexes did not tolerate additional steric bulk. Substitution was limited to the para position of the N-aryl substituent, and the ubiquitous nacnacdipp ligand afforded only the mono-diketiminate complexes nacnacdippZrX3(L)0,1,7b,c,i but not (nacnac)2ZrX2. Ortho substitution of the N-aryl is possible when the steric demand of the second imine is reduced. Thus, nonsymmetric diketiminates © XXXX American Chemical Society
with one ortho-substituted and one ortho-unsubstituted N-aryl substituent, nacnacdipp,Ar,7l or ketiminate ligands, acnac, in which one imino group is replaced by oxygen,7b readily form bisligand complexes, but the nonsymmetric nature of the ligands gave rise to a variety of isomers. Dulong et al. prepared nonsymmetric, tridentate diketiminate ligands with an N-dipp and an N-aryloxy substituent, which form the corresponding complexes.7w Of course, no reactive coordination site remained available in the (nacnacdipp,Ar−O)2Zr complexes obtained. A different approach was pursued by Gong et al.: bridging of two diketiminates by an ethylene or proypylene bridge generated a tetradentate ligand, which afforded cis-octahedral bis-diketiminate zirconium complexes (CH2)2,3(nacnacAr)2ZrCl2 with a variety of ortho-substituted N-aryl substituents (Ar = 2,6-R2phenyl with R = Me, CF3, Cl, iPr).7m We recently started investigating bis-diketiminate zirconium complexes as potential catalysts for organic transformations. The attraction of complexes of the type nacnacR2ZrX2 (1; Scheme 1) as precatalysts is based on the presence of a chiral metal center with close proximity of the N-R substituents to the reactive coordination sites in cis positions. Initial explorations showed that bis-diketiminate complexes with N-alkyl substituents indeed showed the desired complex geometry.7o Their application, however, was impeded by a fast epimerization of the metal center7a,b as well as a very low reactivity, presumably due to steric crowding.7o Bridging of the diketiminate Received: June 12, 2013
A
dx.doi.org/10.1021/om400544z | Organometallics XXXX, XXX, XXX−XXX
Organometallics
Article
trile, and crotonitrile (Scheme 2, Table 1). GC/MS and NMR analysis of all reaction mixtures showed only the presence of
Scheme 1
Scheme 2
Table 1. Aza-Michael Addition of Morpholine to Unsaturated Nitrilesa
ligands7m,8 was then investigated to increase reactivity at the ancillary coordination sites. The structure of bridged bisdiketiminate complexes (±)-C6H10(nacnacXyl)2ZrX2 depends on the size of the ancillary ligand: complexes with X = Cl (2; Scheme 1) or Me have a trans-octahedral geometry, and with sterically more demanding ligands (X = OEt (3), NMe2) the complex adopts a cis-octahedral geometry.7x The deformation caused by the cyclohexanediamine bridge increases the reactivity of the ancillary zirconium substituents significantly and allows easy ligand exchange. The corresponding bisethoxide complex 3 (Scheme 1) proved to be highly active in lactide polymerization, surpassing the activity of any other group 4 catalyst by several orders of magnitude.7x Here we report the synthesis of the Lewis acidic chiral bis-triflate complex 4 (Scheme 1) employing the same ligand. Zirconium triflates, in particular zirconocene triflate, have been used to catalyze a variety of organic reactions.9 However, for aza-Michael additions only a limited number of studies reported the use of group 4 compounds as catalysts, in most cases in the form of simple zirconium salts such as ZrCl4 and its derivatives.10 We found it thus of interest to test the chiral Lewis acid 4 in aza-Michael additions of amines to activated olefins, such as acrylonitrile and acrylate. Enantioselective azaMichael additions are a challenging target,11 and we are not aware of any applications of group 4 metal catalysts in asymmetric aza-Michael additions. Zirconium has been used in enantioselective conjugate additions other than aza-Michael additions,12 and both are widely employed in asymmetric hydroaminations of alkynes (Mg) or olefins (Zr).13 Complex 4 was thus tested in enantioselective aza-Michael additions, and structural reasons for its rather disappointing performance will be discussed.
entry
alkene
concentration/M
reaction time/min
conversion/%
1 2 3 4 5 6 7 8 9 10 11
acrylonitrile methacrylonitrile (E/Z)-crotonitrile methacrylonitrile (E/Z)-crotonitrile acrylonitrile methacrylonitrile (E/Z)-crotonitrile acrylonitrileb methacrylonitrileb (E/Z)-crotonitrileb
0.25 0.25 0.25 0.50 0.50 1.0 1.0 1.0 1.0 1.0 1.0
5 40 80 5 10 5 5 5 80 80 80
100 50 40 95 60 100 95 90 40b 0b 0b
a Reaction conditions: C6D6, ambient temperature, [morpholine]: [alkene]:[4] = 100:100:1, yield determined by NMR. bControl experiments without addition of 4.
the anti-Markovnikov product andin the case of uncomplete reactionsstarting material. With a catalyst loading of 1%, 4 proved to be highly efficient for the addition of morpholine to acrylonitrile in benzene-d6, the reaction reaching full conversion in less than 5 min. As expected, the sterically hindered alkenes methacrylonitrile and crotonitrile were less reactive. Addition of 1 equiv (per morpholine) of NEt3 significantly reduced the reactivity. When substance concentrations were raised from 0.25 to 1.0 M, all three olefins displayed conversions of >90% after 5 min of reaction at room temperature (Table 1, entries 6−8). Under the same conditions, the noncatalyzed reaction of morpholine with acrylonitrile proceeded slowly with 40% conversion after 80 min, while the more hindered alkenes did not react in the absence of catalyst (entries 9−11). To test catalyst efficiency, reactions of acrylonitrile with morpholine were run at reduced catalyst concentrations. Efficient catalysis was still observed at 500 ppm of catalyst loading (Table 2, entry 3, 70% after 5 min; the reaction did not surpass 90% conversion, probably due to catalyst decom-
■
RESULTS AND DISCUSSION Attempts to prepare mono- or bis-triflate complexes by reaction of bis-ethoxide complex 3 with triflic acid or [PhN(H)Me2][OTf] at different temperatures and reaction conditions yielded reaction mixtures in which the main product was the protonated ligand (±)-C6H10(nacnacXyl)2H2. Reaction of dichloride complex 2 with 2 equiv of AgOTf, on the other hand, cleanly yielded the bis-triflate complex (±)-C6H10(nacnacXyl)2Zr(OTf)2 (4; Scheme 1). The first and second chlorides are exchanged with similar reactivities, and reactions with 1 equiv of AgOTf yielded mixtures of 2, 4, and very small amounts of a product putatively assigned as (±)-C6H10(nacnacXyl)2Zr(Cl)(OTf). Aza-Michael Additions. Initial screening of the performance of 4 as a catalyst for aza-Michael additions was undertaken for the addition of morpholine to acrylonitrile, methacryloni-
Table 2. Aza-Michael Addition of Morpholine to Acrylonitrile at Different Catalyst Concentrationsa [Zr]:[acrylonitrile]/%
reaction time/min
conversion/%
TOFb
1 0.1 0.05 0.02
5 5 5/40 45
100 100 70/90 9
>20 >200 ∼300 ∼10
a
Reaction conditions: C6D6, ambient temperature, [morpholine] = [acrylonitrile] = 1 M. bIn units of (mol of alkene/((mol of 4) min).
B
dx.doi.org/10.1021/om400544z | Organometallics XXXX, XXX, XXX−XXX
Organometallics
Article
crotonitrile showed low or no reactivity toward aniline under the conditions investigated (entry 10). With acrylonitrile, 4methoxyaniline proved to be more reactive and 4-chloroaniline to be less reactive than unsubstituted aniline. Thus, the reactivity decreases in the order secondary amines > primary amines > 4-methoxyaniline > aniline > 4-chloroaniline, in parallel with the nucleophilicity of the amine employed. As typically observed for aza-Michael reactions, benzylamine and anilines yielded either only the mono-hydroamination product (entry 4, entry 7, entry 10) or a mixture of mono- and bis-hydroamination products which favors the former (entry 2). Given that the mono-hydroamination products should be more nucleophilic than the starting amines, the reason for their low reactivity to undergo further hydroamination is unclear at the moment. On the basis of the higher reactivity of secondary alkyls in comparison to primary alkyls, simple steric reasons can be excluded as an explanation for the reduced reactivity. The observed kinetics (vide infra) also did not indicate the possibility of product poisoning. No reactivity was observed when phenols or ethanol were used as nucleophiles (Table 3, entries 13−16). The presence of free bis-diketimine in the NMR spectra indicates that the impeding factor is not the nucleophilicity of the alcohols but protonation of the bisdiketiminate ligand. Consequently, ethanolamine likewise did not yield any hydroamination product (entry 17). Complex 4 is thus not stable in the presence of alcohols. To further test the scope of the catalytic reaction and to gain the first insights into a probable reaction mechanism, amines were reacted with alkenes other than conjugated nitriles (Scheme 3, Table 4). Cyclohexenone and methyl acrylate,
position). Below this concentration, reactions turned sluggish and did not reach completion even at prolonged reaction times, consistent with catalyst poisoning by impurities in the reaction mixture. The obtained turnover frequency of ∼300/min at 500 ppm catalyst loading should thus be considered a lower estimate of catalyst activity. It is likely that a notable fraction of the catalyst was deactivated by impurities already at this concentration. The scope of the reaction with regard to the nucleophile was investigated with 1% 4. In all cases, NMR and GC-MS analysis confirmed that hydroamination yielded the expected antiMarkovnikov product and that the uncatalyzed reaction gave less than 5% yield in 3−24 h of reaction time (Table S1, Supporting Information). Benzylamine proved to be likewise highly reactive, and reactions with acrylonitrile reached completion in 5 min at room temperature (Table 3, entry 1). Table 3. Aza-Michael Addition of Different Nucleophilesa entry
nucleophile
alkene
reaction time
conversion/%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
benzylamine benzylamineb benzylamine benzylaminec diethylamine aniline anilinec aniline + NEt3e aniline, 60 °Cf anilinec 4-methoxyaniline 4-chloroaniline 4-methoxyphenol 4-methylphenol 4-chlorophenol ethanol ethanolamine
acrylonitrile acrylonitrile methacrylonitrile methacrylonitrile methacrylonitrile acrylonitrile acrylonitrile acrylonitrile acrylonitrile methacrylonitrile acrylonitrile acrylonitrile acrylonitrile acrylonitrile acrylonitrile acrylonitrile acrylonitrile
5 min 5 min 5 min 10/24 h 5 min 5/60 24 h 5 min 5 min 5/24 h 5 min 5 min 24 h 24 h 24 h 24 h 24 h
97 90b 20 52/70c,d 100 35/80 100c 5e 63f
