Model intermolecular asymmetric Heck reactions catalyzed by chiral pyridyloxazoline palladium(II) complexes

Inorganica Chimica Acta 359 (2006) 2850–2858 www.elsevier.com/locate/ica

Model intermolecular asymmetric Heck reactions catalyzed by chiral pyridyloxazoline palladium(II) complexes David W. Dodd, Heather E. Toews, Florentino d.S. Carneiro, Michael C. Jennings 1, Nathan D. Jones * Department of Chemistry, The University of Western Ontario, Chemistry Building, London, Ont., Canada N6A 5B7 Received 17 October 2005; accepted 20 November 2005 Available online 10 January 2006 This paper is dedicated to Prof. Brian R. James, teacher, mentor and friend, on the occasion of his 70th birthday.

Abstract The synthesis and characterization of a series of chiral pyridyloxazoline Pd(II) halide complexes, including structural determinations, are described. The use of these compounds, as well as those generated in situ from Pd(OAc)2 and 2 equiv. of a pyridyloxazoline ligand, in the intermolecular asymmetric Heck arylation of 2,3-dihydrofuran is reported. In general, total yields after 24 h at 40 C of the kinetic and thermodynamic products, 2-phenyl-2,5-dihydrofuran and 2-phenyl-2,3-dihydrofuran, respectively, were low (12–28%), while e.e.s (when PhOTf was used as the aryl source) were low (4–29%) for the kinetic and moderate (23–60%) for the thermodynamic product. Both yield and e.e. were compromised by facile catalyst decomposition. The higher e.e.s found for the thermodynamic than for the kinetic product imply a kinetic resolution. When PhI was used as the aryl source, racemic product was invariably generated, which strongly indicated that asymmetric induction was effected through cationic species in the reaction cycle.  2005 Elsevier B.V. All rights reserved. Keywords: Asymmetric Heck reaction; Palladium complex; X-ray crystal structures; Enantioselective catalysis

1. Introduction The palladium-catalyzed Heck coupling of alkenes and aryl- or alkenyl halides is one of the most important methods for the formation of C–C bonds [1,2]. Since the first demonstrations in 1989 of asymmetric intramolecular variants (simultaneously and independently by the groups of Shibasaki [3] and Overman [4]), the Heck reaction has also become a powerful tool for the formation of tertiary and quaternary stereocentres. While the intramolecular asymmetric Heck reaction (AHR) has reached a stage of maturity that allows it to be used with great effect in the synthesis of natural products [5–7], its intermolecular counterpart, first reported by Hayashi and coworkers in 1991 [8] *

1

Corresponding author. Tel.: +1 519 661 2111; fax: +1 519 661 3022. E-mail address: [email protected] (N.D. Jones). X-ray Crystallographic Laboratory.

0020-1693/$ - see front matter  2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ica.2005.11.025

(Scheme 1; catalyst = Pd(OAc)2 + R-BINAP, base = i Pr2EtN, X = OTf), has only been successfully applied in model systems. These frequently involve five- and sevenmembered dihydroheterocycles, e.g., furans [8,9], pyrroles [10], and dioxepins [11]. In part because of its convenient synthesis from readily available chiral amino alcohols, the C2-symmetric bis(oxazoline), or ‘‘box,’’ ligand architecture retains a privileged position in asymmetric catalysis; it has found successful application in a wide variety of transformations, including enantioselective cycloaddition, aldol, Michael, carbonylene and allylic alkylation reactions [12,13]. The tridentate pyridylbis(oxazoline) (‘‘pybox’’) [14] and phenylbis(oxazoline) [15] series have also found a number of asymmetric catalytic applications. Of relevance to this paper is the fact that oxazoline groups have also featured prominently in ligands that have been applied successfully in model intermolecular AHRs,

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858 X O

Chiral Pd catalyst Base - [BaseH]X

+

O

Ph

O

+

13

Ph

14

Scheme 1. The asymmetric Heck arylation of 2,3-dihydrofuran.

N

Ph2P

R

PPh2

N R

2.1. General considerations

N Fe SiMe3

R

PPh2

Chart 1.

particularly in conjunction with phosphine groups. These P,N-donor ligands have included the triphenylphosphine derivatives of Pfaltz and coworkers [16,17], BINAP derivatives of Pregosins group [18], as well as ferrocenyl oxazolines [19] (Chart 1). Our strategy was to use the simplest possible chiral chelating system, and we settled on the pyridyloxazoline, or ‘‘pyox’’, ligand family, originally reported by Bolm et al. [20], that is accessible in one step through the ZnCl2-catalyzed coupling of a b-aminoalcohols and 2cyanopyridine. To our surprise, simple Pd(II)(pyox) dihalide complexes have received little attention [21,22], even though the pyox ligands were reported almost 15 years ago. Some PdCl(Me)(pyox) complexes, on the other hand, have been used as catalysts in asymmetric diene cyclization/hydrosilylation reactions [23–25], and Pd(II) complexes of the pybox and phenylbis(oxazoline) series have found use in the asymmetric Michael reaction [15]. Structural determinations of simple Pd(II)(pyox) dihalide complexes are lacking [26]. More significantly, we are not aware of any use of the general class of chiral bidentate N-donor ligands in the AHR. A box ligand has been used in the related enantioselective annelation of allenes [27], however, and diimines (diazabutadienes) [28], pyridyl-imines [29] and dipyridines [30] have been used as ligands in achiral Heck reactions. This report describes the synthesis and characterization of a series of Pd(II) pyox dichloride complexes (6, 8–11,

O R

O N

N

Pd X

X

N

N

Pd

Cl

Cl

X= Cl; R = iPr (6), iBu (8), sBu (9), Bz (10) X = I; R = iPr (7)

Chart 2) including structural determinations of 6, 7 and 11, and the use of these compounds as catalysts for asymmetric arylation of 2,3-dihydrofuran (Scheme 1). 2. Experimental

O

O

O

2851

11

The pyox ligands, 2-[(4S)-4-alkyl-4,5-dihydro-1,3-oxazol-2-yl]pyridine (alkyl = iso-propyl, 1; iso-butyl, 2; secbutyl, 3; benzyl, 4) and 2-[(4R)-4-phenyl-4,5-dihydro-1, 3-oxazol-2-yl]pyridine (5) [20] and the metal precursors, trans-PdCl2(PhCN)2 [31] and Pd2(dba)3ÆCHCl3 [32], were made according to the literature procedures. Solvents and all other reagents were obtained commercially and used as supplied. Experiments were conducted under N2 unless otherwise noted. NMR spectra were collected using a Varian 400 spectrometer (400.089 MHz for 1H, 100.613 MHz for 13C) at room temperature (r.t.) in CDCl3 solution unless otherwise noted and were referenced as appropriate either to residual 1H or to 13C in the deuterated solvent. All coupling constants are given in Hz: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet. UV–Vis absorbance spectra were collected using a Cary 300 Bio spectrometer illuminated by a D2/W two-lamp system. Data, reported as kmax (nm) [log10 e], were collected at r.t. using approximately 0.1 mM CH2Cl2 solutions of the Pd complexes contained in 1 cm pathlength quartz cuvettes. Elemental analyses were conducted by Guelph Chemical Laboratories (Ltd). 2.2. Crystallography Data were collected either at 150 K (for 6) or near r.t. (for 7 and 11) using a Nonius Kappa-CCD area detector diffractometer running COLLECT software (Nonius, B.V., 1997–2002). The unit cell parameters were calculated and refined from the full data set. Crystal cell refinements and data reduction were carried out using HKL2000 DENZO-SMN (Otwinowski and Minor, 1997) and absorption corrections were applied using HKL2000 DENZOSMN (SCALEPACK). The SHELXTL/PC (Sheldrick, G.M., 2001) suite of programs (V6.14 for Windows NT) was used to solve the structures by direct methods. Subsequent difference Fourier syntheses allowed the remaining atoms to be located. All non-hydrogen atoms were refined with anisotropic thermal parameters. Hydrogen atom positions were calculated geometrically and were included as riding on their respective carbon atoms. Crystallographic data for 6, 7 and 11 are summarized in Table 1.

O

O

2.3. Heck reactions

N

N

N

I

I

N Pd

Pd

12a

12b Chart 2.

Unless otherwise stated, these were carried out in C6H6 solution at 40 C in a manner similar to that previously described [9]. At the beginning of each reaction, solutions contained either Pd(OAc)2 and the appropriate pyox ligand, or one of the preformed complexes 6, 8–11, and

2852

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

Table 1 Crystal data and structure refinement for compounds 6, 7 and 11 Compound

6

7

11

Empirical formula Formula weight Temperature (K) ˚) Wavelength (A Crystal system Space group Unit cell dimensions ˚) a (A ˚) b (A ˚) c (A a, b, c () ˚ 3) Volume (A Z Density (calculated, g cm 3) Absorption coefficient (mm 1) F(0 0 0) Crystal size (mm) h Range for data collection () Index ranges

C11H14Cl2N2OPd 365.57 150(2) 0.71073 orthorhombic P212121

C11H14I2N2OPd 550.47 301(2) 0.71073 orthorhombic P212121

C14H12Cl2N2OPd 533.63 299(2) 0.71073 orthorhombic P212121

5.9329(3) 11.3698(7) 19.6928(11) 90 1328.40(13) 4 1.838 1.783 728 0.18 · 0.10 · 0.05 2.74–25.02 7 6 h 6 7, 13 6 k 6 13, 23 6 l 6 23 11 276 2336 (0.0840) semi-empirical from equivalents 0.9161 and 0.7396

7.2320(2) 11.4423(2) 18.6966(5) 90 1547.16(7) 4 2.363 5.178 1016 0.50 · 0.18 · 0.13 2.81–25.03 8 6 h 6 8, 13 6 k 6 13, 22 6 l 6 22 14 643 2742 (0.0620) semi-empirical from equivalents 0.5638 and 0.1816

10.5447(3) 11.2342(4) 24.9301(5) 90 2953.25(15) 8 1.806 1.613 1584 0.25 · 0.15 · 0.04 2.65–25.02 12 6 h 6 12, 13 6 k 6 13, 29 6 l 6 29 23 600 5204 (0.0680) semi-empirical from equivalents 0.9420 and 0.6885

full-matrix least-squares on F2 2336/0/155 1.041 R1 = 0.0394, wR2 = 0.0846 R1 = 0.0600, wR2 = 0.0926 0.00(6) 0.503 and 0.643

full-matrix least-squares on F2 2742/0/155 1.057 R1 = 0.0309, wR2 = 0.0708 R1 = 0.0384, wR2 = 0.0735 0.07(4) 0.496 and 0.656

full-matrix least-squares on F2 5204/0/338 1.048 R1 = 0.0512, wR2 = 0.1251 R1 = 0.0729, wR2 = 0.1385 0.03(5) 2.281 and 0.832

Reflections collected Independent reflections [Rint] Absorption correction Maximum and minimum transmission Refinement method Data/restraints/parameters Goodness-of-fit on F2 Final R indices [I > 2r(I)] R indices (all data) Absolute structure parameter Largest difference in peak ˚ 3) and hole (e A

PhI or PhOTf, 2,3-dihydrofuran and iPr2EtN in a molar ratio of 0.03(:0.06):1:5:3, respectively. In a typical experiment, the ligand (0.53 mmol) was added to a C6H6 solution (10 mL) containing Pd(OAc)2 (60.2 mg, 0.27 mmol) and allowed to stir at r.t. for a period of 10 min after which PhI (1.82 g, 8.93 mmol) or PhOTf (2.02 g, 8.93 mmol), 2,3-dihydrofuran (3.13 g, 44.67 mmol) and iPr2EtN (3.46 g, 26.80 mmol) were added in that order. Aliquots (1.5 mL) of the Heck reaction mixtures were removed at 0, 6 and 24 h, and the catalyst was precipitated with hexanes (1.5 mL). Samples were further prepared for GC and NMR analysis by filtering through alumina (5.0 · 0.75 cm).

higher field than those of S-14 [9]. The relative intensities of these two peaks were then correlated with those of the two peaks corresponding to 14 as determined by GC (below). Once these identifications were made, e.e. determination was subsequently based on GC data. An analogous analysis could not be applied to the kinetic product 13 because it was always produced in very small quantities and could not be separated from residual ArX in solution so as to allow unambiguous NMR e.e. determination; the ortho protons in the phenyl group of R-13 should appear at higher field than those of S-13, but in our case these were always obscured by ArX peaks. 2.5. Gas chromatography

2.4. Enantiomeric excess (e.e.) determinations The e.e. for the thermodynamic product, 14, was determined in the first instance from 1H NMR spectra of the reaction products that had been acquired in the presence of the optically pure shift reagent, Eu(hfc)3 (hfc = 3-((heptafluoropropyl)hydroxymethyl)camphorate). In these spectra, peaks for the olefinic protons of R-14 appeared at

Analyses were conducted using an Agilent 6890N instrument equipped with a 20 m · 250 lm Chiraldex capillary with b-cyclodextrin as the stationary phase. A 50:50 split injection was used for a sample of 0.75 lL at a constant pressure of 11 psi and a flow of 1.4 mL min 1 giving a velocity of 0.36 m s 1. An initial temperature of 40 C was held for 2 min and then followed by a ramp of

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

3 C min 1 to a final temperature of 120 C, which was held for a further 8 min. Efficient separation of regioisomers and enantiomers was obtained in this manner. Elution times (min): C6H6 (3.2), iPr2EtN (6.2), PhI (16.7), PhOTf (19.4), (S)-2-phenyl-2,3-dihydrofuran (S-14, 23.2), (R)-2phenyl-2,3-dihydrofuran (R-14, 23.7), 2-phenyl-2,5dihydrofuran (24.3 and 24.8). 3. Syntheses 3.1. Dichloro-2-[(4S)-4-iso-propyl-4,5-dihydro-1,3-oxazol2-yl]pyridinepalladium(II) (6) The title complex has been reported previously, but no synthetic or characterization data were given [22]. It was synthesized in this work by the addition of transPdCl2(PhCN)2 (200 mg, 0.53 mmol) to a CH2Cl2 solution (10 mL) containing a slight excess of pyridyloxazoline ligand 1 (100 mg, 0.54 mmol). The reaction was stirred at r.t. for 2 h during which time it turned from a pale to a slightly darker orange. The solvent was evaporated and the sticky orange residue was triturated with Et2O to give 187 mg (96%) of a fine orange powder. This was washed with additional portions of Et2O and dried under vacuum. Anal. Calc. for C11H14N2Cl2OPd: C, 35.94; H, 3.84; N, 7.62%. Found: C, 36.07; H, 3.93; N, 7.71%. 1H NMR: d 0.81 (d, 3H, CH3, 3JHH = 6.86), 0.94 (d, 3H, CH3, 3JHH = 7.07), 2.87 (dq, 1H, CH, 3JHH = 6.86, 7.07), 4.60 (m, 1H, CH), 4.74 (dd, 1H, CH2, 2JHH = 3.76, 3JHH = 8.83), 4.97 (dd, 1H, CH2, 2JHH = 3.76, 3JHH = 9.51), 7.63 (t, 1H, py, 3JHH = 7.13), 7.81 (d, 1H, py, 3JHH = 7.53), 8.12 (t, 1H, py, 3JHH = 7.54), 9.01 (d, 1H, py, 3JHH = 5.35). 13 C{1H} NMR: d 14.5 (CH3), 18.9 (CH3), 28.9 (CH), 72.6 (CH), 126.1 (py), 129.5 (py), 141.1 (py), 145.1 (OCN), 151.0 (py), 169.8 (py). UV–Vis: 312 [3.746]. Orange plates suitable for X-ray diffraction analysis were grown by slow diffusion of Et2O into a CH2Cl2 solution containing 6.

2853

X-ray diffraction analysis were grown by slow diffusion of Et2O into a CH2Cl2 solution of 7. 3.3. Dichloro-2-[(4S)-4-iso-butyl-4,5-dihydro-1,3-oxazol-2yl]pyridinepalladium(II) (8) This complex was prepared in the same manner as 6. Thus reaction of trans-PdCl2(PhCN)2 (200 mg, 0.53 mmol) and an excess of 2 (120 mg, 0.60 mmol) yielded 196 mg (96%) of a fine orange powder. Anal. Calc. for C12H16N2Cl2OPd: C, 37.77; H, 4.23; N, 7.34%. Found: C, 37.73; H, 4.45; N, 7.42%. 1H NMR: d 0.90 (d, 6H, CH3, 3 JHH = 6.37), 1.39 (t, 1H, CH2, 3JHH = 11.38), 1.48 (m, 1H, CH), 2.27 (t, 1H, CH2, 3JHH = 11.32), 4.57 (m, 2H, CH2), 5.15 (t, 1H, CH, 3JHH = 8.22), 7.66 (t, 1H, py, 3 JHH = 6.31), 7.79 (d, 1H, py, 3JHH = 7.52), 8.14 (t, 1H, py, 3JHH = 7.51), 8.77 (d, 1H, py, 3JHH = 5.02). 13C{1H} NMR: d 21.8 (CH3), 23.8 (CH2), 25.8 (CH3), 44.7 (CH), 62.4 (CH2), 79.0 (CH), 126.2 (py), 129.8 (py), 141.2 (py), 145.0 (NCO), 150.8 (py), 169.7 (py). UV–Vis: 313 [3.692]. 3.4. Dichloro-2-[(4S)-4-sec-butyl-4,5-dihydro-1,3-oxazol-2yl]pyridinepalladium(II) (9) This complex was prepared in the same manner as 6. Thus, reaction of trans-PdCl2(PhCN)2 (200 mg, 0.53 mmol) and an excess 3 (0.12 g, 0.60 mmol) yielded 197 mg (94%) of a fine orange powder. Anal. Calc. for C12H16N2Cl2OPd: C, 37.77; H, 4.23; N, 7.34%. Found: C, 38.17; H, 4.54; N, 7.52%. 1H NMR: d 0.78 (d, 3H, CH3, 3JHH = 6.87), 0.93 (t, 3H, CH3, 3JHH = 7.35), 1.23 (m, 2H, CH2), 2.52 (m, 1H, CH), 4.70 (m, 2H, CH2), 5.02 (t, 1H, CH, 3JHH = 9.26), 7.63 (m, 1H, py), 7.80 (m, 1H, py), 8.13 (m, 1H, py), 8.85 (m, 1H, py). 13C{1H} NMR: d 12.00 (CH3), 12.03 (CH3), 26.1 (CH2), 35.8 (CH), 66.5 (CH2), 72.8 (CH), 126.0 (py), 129.1 (py), 140.9 (py), 145.1 (NCO), 150.1 (py), 169.7 (py). UV–Vis: 313 [3.673].

3.2. Diiodo-2-[(4S)-4-iso-propyl-4,5-dihydro-1,3-oxazol-2yl]pyridinepalladium(II) (7)

3.5. Dichloro-2-[(4S)-4-benzyl-4,5-dihydro-1,3-oxazol-2yl]pyridinepalladium(II) (10)

The dichloride 6 (25 mg, 0.08 mmol) was stirred in acetone (20 mL) at r.t. in the presence of an excess of NaI (74 mg, 0.40 mmol) for a period of 2 h after which the solvent was removed under vacuum and the residue taken up in CH2Cl2 (30 mL). After filtration through Celite 545, the solvent was removed in vacuo to yield 40 mg (97%) of a deep red-purple microcrystalline powder. Anal. Calc. for C11H14N2I2OPd: C, 24.00; H, 2.56; N, 5.09%. Found: C, 23.61; H, 2.91; N, 4.96%. 1H NMR: d 0.77 (d, 3H, CH3, 3 JHH = 6.89), 0.96 (d, 3H, CH3, 3JHH = 7.11), 2.19 (dq, 1H, CH, 3JHH = 6.89, 7.11), 4.67 (m, 1H, CH), 4.79 (m, 2H, overlapping CH2 and CH), 7.74 (m, 2H, py), 7.8 (m, 1H, py), 9.78 (m, 1H, py). 13C{1H} NMR: d 14.0 (CH3), 18.8 (CH3), 29.8 (CH), 68.9 (CH), 72.4 (CH2), 125.9 (py), 129.9 (py), 139.7 (py), 144.4 (NCO), 154.3 (py), 170.8 (py). UV–Vis: 320 [3.878]. Dark red prisms suitable for

The title complex was prepared in the same manner as 6. Thus, reaction of trans-PdCl2(PhCN)2 (200 mg, 0.53 mmol) and an excess of 4 (140 mg, 0.60 mmol) yielded 205 mg (95%) of a fine orange powder. The 1H NMR data for this complex were the same as those reported previously [21]. UV–Vis: 313 [3.714]. 3.6. Dichloro-2-[(4R)-4-phenyl-4,5-dihydro-1,3-oxazol-2yl]pyridinepalladium(II) (11) This complex was prepared in the same manner as 6. Thus, reaction of trans-PdCl2(PhCN)2 (200 mg, 0.53 mmol) and an excess of 4 (130 mg, 0.60 mmol) yielded 195 mg (94%) of a fine orange powder. 1H NMR (DMSO-d6): d 4.85 (dd, 1H, CH2, 2JHH = 5.17, 3JHH = 8.89), 5.35 (dd, 1H, CH, 3JHH = 8.84, 10.02), 5.71 (dd, 1H, CH2,

2854

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

2

JHH = 5.17, 3JHH = 10.02), 7.37 (m, 5H, Ph), 7.67 (m, 1H, py), 7.89 (m, 1H, py), 8.15 (m, 1H, py), 9.14 (m, 1H, py). 13 C{1H} NMR (DMSO-d6): d 65.6 (CH2), 80.6 (CH), 127.2 (Ph), 127.8 (Ph), 128.7 (Ph), 128.8 (Ph), 129.2 (Ph), 130.7 (Ph), 140.4 (py), 142.1 (NCO), 144.7 (py), 150.7 (py), 171.0 (py). UV–Vis: 313 [3.852]. 3.7. Iodophenyl-2-[(4S)-4-iso-propyl-4,5-dihydro-1,3oxazol-2-yl]pyridinepalladium(II) (12) Iodobenzene (310 mg, 1.50 mmol) was added to a C6H6 solution (20 mL) containing Pd2(dba)3 (200 mg, 0.19 mmol) and 1 (10 mg, 0.50 mmol). The purple solution was stirred at 40 C for a period of 3 h during which time it became orange in colour. The solvent was evaporated in vacuo and the residue was triturated with Et2O and washed with several additional portions of Et2O. Residual solvent was removed to yield 88 mg (92%) of a dark orange powder. Two different stereoisomers were visible by 1H NMR in a 1:3 ratio. Anal. Calc. for C17H19ON2IPd: C, 40.78; H, 3.83; N, 5.60%. Found: C, 40.91; H, 3.71; N, 5.60%. Major isomer (12b): 1H NMR: d 0.50 (d, 3H, CH3, 3 JHH = 7.12), 0.69 (d, 3H, CH3, 3JHH = 6.89), 1.20 (dq, 1H, CH, 3JHH = 6.89, 7.12), 4.09 (m, 1H, CH), 4.60 (m, 3H, CH2, CH,), 6.83 (m, 1H, p-Ph), 6.93 (m, 2H, m-Ph), 7.33 (m, 2H, o-Ph), 7.65 (m, 1H, py), 7.79 (m, 1H, py), 7.99 (m, 1H, py), 9.43 (m, 1H, py). Minor isomer (12a): 1 H NMR: d 0.83 (d, 3H, CH3, 3JHH = 7.18), 0.96 (d, 3H, CH3, 3JHH = 6.97), 2.91 (dq, 1H, CH, 3JHH = 6.97, 7.18), 4.10 (m, 1H, CH), 4.58 (m, 3H, CH2, CH,), 6.84 (m, 1H, p-Ph), 7.01 (m, 2H, m-Ph), 7.39 (m, 2H, o-Ph), 7.46 (m, 1H, py), 7.58 (m, 1H, py), 7.95 (m, 1H, py), 9.41 (m, 1H, py). 13C{1H} NMR (12b only): d 13.9 (CH3), 19.0 (CH3), 28.3 (CH), 68.0 (CH2), 71.4 (CH), 123.4 (Ph), 123.5 (Ph), 124.4 (Ph), 124.6 (NCO), 127.6 (Ph), 128.6 (Ph), 129.2 (Ph), 137.0 (py), 138.9 (py), 143.7 (py), 153.3 (py), 168.3 (py). UV–Vis: 326 [3.692]. 4. Results and discussion

deep red-purple diiodide complex 7 was made by reacting 6 with an excess of NaI in acetone. The iodophenyl complex 12 (Chart 2) was made by the oxidative addition of PhI to Pd2(dba)3 (dba = dibenzylidene acetone) in the presence of 1; a mixture of phenyltrans-to-pyridine (12a) and phenyl-trans-to-oxazoline (12b) isomers was generated in this reaction in an approximate molar ratio of 1:3, as determined by NMR spectroscopy. Based on the assumption of unfavourable steric interactions between the phenyl ring and the isopropyl substituent at the 4-position of the oxazoline ring, we postulate that 12b was the dominant isomer. 4.2. Structures The molecular structures of 6, 7 and 11 are given in Figs. 1–3, respectively; selected bond distances and bond angles appear in Table 2. Compound 11 has two crystallographically independent molecules in its unit cell; only one is shown because the bond distances and bond angles of the two forms differ only marginally. The structures of all three compounds are very similar to that of a related Pd(II)Cl2 (pyox) complex incorporating a CH2OH group at the

Fig. 1. ORTEP representation (30% ellipsoids) of the molecular structure of 6. Hydrogen atoms have been omitted for clarity. All unlabeled ellipsoids represent carbon atoms. The absolute configuration at the C2 centre is S.

4.1. Syntheses The pyridyloxazoline ligands 1–5 were made by the ZnCl2-catalyzed coupling of 2-cyanopyridine with the bamino alcohols, S-valinol, S-isoleucinol, S-leucinol, S-phenylalaninol and R-phenylglycinol, respectively, using the procedure developed by Bolm et al. [20]. The Pd(II) complexes 6 and 8–11 (Chart 2) were made in excellent yields by the reaction of slight molar excesses of the ligands 1– 5, respectively, with trans-PdCl2(PhCN)2 in CH2Cl2 solution at r.t. Compounds 6 [22] and 10 [21] have been reported previously, but neither synthetic nor characterization data were given for the former. The complexes were yellow–orange, microcrystalline solids that were soluble in polar organic solvents, such as CHCl3 and CH2Cl2, and insoluble in Et2O and hexanes. They were air- and water-stable, both in the solid state and in solution. The

Fig. 2. ORTEP representation (30% ellipsoids) of the molecular structure of 7. Hydrogen atoms have been omitted for clarity. All unlabeled ellipsoids represent carbon atoms. The absolute configuration at the C2 centre is S.

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

2855

we postulate that this is because there is a significant contribution arising from the resonance form that places a formal positive charge on the oxygen atom and a formal negative charge on the nitrogen atom. 4.3. Heck reactions

Fig. 3. ORTEP representation (30% ellipsoids) of the molecular structure of 11. Hydrogen atoms have been omitted for clarity. All unlabeled ellipsoids represent carbon atoms. The absolute configuration at the C2 centre is R.

Table 2 Selected bond distances and bond angles for complexes 6, 7, and 11

Selected Pd–N1 Pd–N2 Pd–X1 Pd–X2

6 (X = Cl) bond distances (A˚) 2.017(5) 2.048(6) 2.285(2) 2.298(2)

Selected angles () N1–Pd–N2 80.9(3) X1–Pd–X2 92.08(7) N1–Pd–X1 93.2(2) N2–Pd–X2 93.8(2)

7 (X = I) 2.052(5) 2.102(6) 2.5694(7) 2.5705(7) 79.0(2) 91.23(2) 94.3(2) 95.4(2)

11 (X = Cl) 2.011(7) 2.027(7) 2.277(2) 2.276(2) 79.9(3) 91.75(9) 94.5(2) 94.1(2)

Estimated standard deviations are given in parentheses.

6-position of the pyridine ring [26]. In each of the complexes 6, 7 and 11, the Pd(II) centre adopts the expected square planar coordination geometry, and the pyox ligand is approximately planar except for the alkyl or aryl substituent at the 4-position of the oxazoline ring, which protrudes prominently from one face of each complex. The absolute configuration at this centre, being fixed ultimately by the chirality of the corresponding amino acid, is S for 6 and 7, and R for 11. The respective Flack parameters are 0.00(6), 0.07(4) and 0.03(5), indicating that that the correct hand was refined in each case and that there was no sign of racemic twinning. Metrical parameters are unsurprising: the bite angle of the pyox ligands range from 79 to 81, which spans those found for the CH2OH-modified pyox complex (two crystallographically independent forms: 80.2(2) and 80.4(1)) [26]. The Pd–N and Pd–Cl bond distances in 6 and 11 are also similar to those found previously [26], while the corresponding lengths in 7 are significantly longer. It is interesting to note that in all cases, particularly in 6 and 7, the Pd–N(oxazoline) bond is shorter than the Pd–N(pyridine) bond. In addition, the Pd–Cl bond trans to the oxazoline donor in 6 is longer than that trans to the pyridine donor. These data strongly suggest that oxazoline is a stronger field ligand than pyridine;

Results for the asymmetric Heck arylation of 2,3dihydrofuran (Scheme 1) using the preformed complexes 6 and 8–11 as catalysts, as well as catalysts generated in situ by the reaction of Pd(OAc)2 and 2 equiv. of the ligands 1–5, are given in Table 3. The reaction conditions (3 mol% catalyst, C6H6 solvent, iPr2EtN base, 40 C) were modeled after those of Hayashi and coworkers [9]. We stress that our intention in choosing these conditions was not to optimize, but rather to enable comparison to other published reports. Our results are consistent with those of Hayashi and others [7–9]: 2-phenyl-2,5-dihydrofuran (13) was the kinetic product of the reaction, while 2-phenyl-2, 3-dihydrofuran (14) was the thermodynamic product. In all runs, the concentration of 13 was lower after 24 h of reaction than after 6 h, while the concentration of 14 showed the opposite trend. In general, total yields after 24 h of 13 and 14 (12–28%) were disappointing, while e.e.s (when PhOTf was used as the aryl source) were low (4–29%) for 13 (indeterminate enantiomer, see Section 2) and moderate (23–46%) for R-14, although when a combination of Pd(OAc)2 and 5, which incorporates an R-stereocentre, was used as the catalyst, the e.e. for S-14 was a somewhat higher 60%. By comparison, the Pd(OAc)2 + 2 R-BINAP catalytic system of Hayashi and coworkers gave an isolated total yield of 63% and e.e.s of 60 and 82% for S13 and R-14 under conditions identical to ours [9]. The single enantiodetermining step in the AHR cycle is the insertion of the cyclic alkene into the Pd–Ph bond; bhydride elimination from this species gives the kinetic product. Therefore, a higher e.e. for 14 than for 13 implies that double bond migration, which converts 13–14, is faster in one enantiomer of 13 than it is in the other [9]. Consequently, the enantiopurity of 14 is enriched over time at the expense of that of 13. In some cases, this kinetic discrimination is so effective as to invert the enantioselectivity of 13 (as in Hayashis case, above). Although we we did not determine which enantiomer of 13 was produced in excess (see Section 2), absolute e.e.s for 13 were consistently lower in our trials than those for 14, and thus a kinetic enrichment process analogous to that observed in the BINAP systems was clearly implicated. In contrast, Pfaltz and coworkers reported that 13 was the sole product of reactions catalyzed by Pd complexes incorporating oxazolinylphosphine ligands [17]. We found that when PhI was used as the substrate, there was no asymmetric induction (Table 3). This result is consistent with those of Hayashi, who proposed a requisite cationic pathway for the formation of 13 in high e.e. [9]. This cationic pathway features dissociation of halide or pseudohalide prior to coordination of the alkene,

2856

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

Table 3 Summary of catalytic data for the asymmetric Heck arylation of 2,3-dihydrofuran (Scheme 1)a Catalyst Aryl source Yield (%)b 14:13 ratio (6 h) 14:13 ratio (24 h) e.e. 14 (%)b,c e.e. 13 (%)b,c

6

8

9

10

11

12 0.97 1.50

25 0.84 1.10

18 0.33 1.32 0 0

19 0.66 0.91

17 0.42 0.98

Pd(OAc)2 + 1

Pd(OAc)2 + 2

Pd(OAc)2 + 3

Pd(OAc)2 + 4

Pd(OAc)2 + 5

23 0.41 0.76 39 4

28 0.98 2.33 42 26

30 1.67 2.56 23 9

22 0.62 0.98 46 20

15 0.42 0.52 60d 29

PhI

Aryl source Yield (%)b 14:13 ratio (6 h) 14:13 ratio (24 h) e.e. 14 (%)b,c e.e. 13 (%)b,e

PhOTf

a All reactions were conducted under N2 in C6H6 solution at 40 C. Reaction solutions contained 2,3-dihydrofuran, PhX (X = I, OTf) and iPr2EtN in a 1:5:3 molar ratio and employed 3 mol% of Pd. When the catalysts were formed in situ, a 2:1 molar ratio of ligand to Pd was used. Reported yields are for 13 and 14 combined, as determined by GC; e.e.s were determined by chiral GC. b After 24 h. c For the R-enantiomer, unless otherwise stated. d For the S-enantiomer. e Indeterminate enantiomer (see Section 2).

which is not favoured when the halide is the relatively strongly coordinating iodide. In the corresponding neutral pathway, partial dissociation of the chiral chelating ligand is required for alkene coordination, necessarily lowering alkene enantioface discrimination. Our results strongly suggested, therefore, a cationic pathway for the AHR catalyzed by pyox complexes when PhOTf was used as the aryl source. We envision the production of R-14 via the partial mechanism shown in Scheme 2. Steric obstruction by the R-group on the 4-position of the oxazoline ring encourages the incoming alkene to bind preferentially through its re face to the three-coordinate, cationic Pd(Ph)(pyox) intermediate, I. Migratory insertion of the alkene into the Pd– Ph bond fixes the absolute configuration at the 2-position of the ring in III. Rotation about the Pd–C bond, and bhydride elimination from the 4-position generates R-13. Reinsertion of R-13 with opposite regiochemistry into the resulting Pd–H bond, followed by a second b-hydride elimination from the 5-position finally generates the thermodynamic product, R-14. It is important to note that the cation I in Scheme 2 is that which would result from iodide abstraction from 12b (for R = iPr), which we have postulated to be the major isomer of 12. Presumably, the minor isomer 12b, which would give rise to a three-coordinate cation with a vacant site cis to the achiral pyridine ring, would

R

O N

O Pd N Ph

I

R

O N O

Pd N Ph

II

not discriminate between the enantiotopic faces of 2,3dihydrofuran. When the preformed dihalide complexes 6, 8–11 were used as the catalysts and PhOTf as the aryl source (not shown in Table 3), Pd black deposited immediately, and in every case there was no reaction whatsoever. When contrasted to the data given in Table 3, this result suggested that either iodide or acetate was important in stabilizing the Pd centre during the reaction cycle. The disappointing yields and only moderate enantioselectivity in our system must have been due in part to catalyst decomposition: in all runs, Pd black deposited from solution within 0.5 h of beginning the reaction. Moreover, when the reactions were followed closely over time, it was apparent that approximately 50% of the product observed after 24 h had been generated within the first 2 h. This finding was surprising to us because Pd complexes of monodentate oxazoline ligands have been used to great effect by Gossage et al. in achiral Heck couplings, even in the presence of air and moisture, giving moderate turn over numbers and no reported metal deposits [33]. Also, no metallic deposit was reported for Pd(diazabutadiene)catalyzed reactions [28], and only after all aryl halide had been consumed was it noted to appear in reactions catalyzed by complexes of pyridyl-imines [29]. At high catalyst loadings, Pd black did deposit in reactions catalyzed by

R

O N

Pd N O

Ph III

Scheme 2. Production of R-14 from 2,3 dihydrofuran via [Pd(Ph)(S-pyox)]+.

O

R-14

Ph

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

dipyridyl complexes, however [30]. It is likely that the possibility for a trans orientation of N-donors, which is impossible in our case, lent stability to Gossages catalyst, but the relative instability of our catalysts compared to those incorporating diazabutadiene and pyridyl-imine ligands, which are also constrained by cis N-donors, remains unexplained. When two drops of Hg metal were added to our reaction mixtures, yields were somewhat lower than in its absence (e.g., 15% versus 23% for Pd(OAc)2 + 1) indicating that a second, heterogeneous pathway was responsible for a portion of the couplings. Presumably, this pathway was also achiral, which necessarily translated into lower e.e.s in our system and complicated our determination of the efficacy of pyox ligands as asymmetric inducers. The decomposition problem was not alleviated through addition of 1 equiv. of PPh3 to the system, nor through use of Pd2(dba)3 as the metal source. At the end of a typical catalytic reaction, one of the metal complexes in solution could be identified as PdI2 (pyox). In the case of catalysis by 6, this was determined by comparing the NMR spectrum of the crude reaction solution to that of the diiodide, 7, which had been synthesized and characterized independently. We also identified 7 in the following stoichiometric experiment: A mixture of the two iodophenyl isomers 12a,b, which are presumably intermediates in the catalytic cycle, was combined in CDCl3 with a molar equivalent of 2,3-dihydrofuran, and the reaction was followed in situ by 1H NMR spectroscopy at r.t. Within minutes, Pd black had deposited in the tube. At this point, the solution contained free 1, the diiodide 7, unreacted 2,3-dihydrofuran and both of the products 13 and 14. Upfield peaks characteristic of Pd-hydride species were not observed, even though no base was added to this system. We are currently working to understand the inherent instability in the system, to improve the resilience of the catalysts, and to increase the rate of reaction and enhance enantioselectivity by optimizing reaction conditions.

Acknowledgements We thank the Natural Sciences and Engineering Research Council of Canada (NSERC) and the University of Western Ontario for financial support of our Research Program. HET acknowledges the support of an NSERC University Summer Research Assistantship. Appendix A. Supplementary data Crystallographic data for the complexes 6, 7 and 11 as CIF. These are available upon request from the Cambridge Crystallographic Data Centre (http://www.ccdc.cam. ac.uk). Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ica. 2005.11.025. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

5. Conclusions The reported Pd(II)-pyox compounds are attractive as potential catalysts for the AHR because they are accessible in a modular manner from readily available chiral sources using high-yielding reactions. When compared to BINAP and its derivatives, the pyox ligands are cheaper and simpler to make in enantiopure form, and even though they are far less demanding sterically still give rise to moderate e.e.s for the thermodynamic product in the asymmetric Heck arylation of 2,3-dihydrofuran. Moreover, the compounds are air- and water-stable both in solution and in the solid state. However, they are inherently unstable under Heck reaction conditions, possibly because the hard Ndonors are poor ligands for Pd(0) and because they are constrained to be cis. This instability, which compromises both yield and e.e., would need to be overcome before Pd(II)-pyox complexes could be used more widely in asymmetric Heck and related coupling reactions.

2857

[16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

A. de Meijere, F.E. Meyer, Angew. Chem. Int. Ed. 33 (1994) 2379. V. Farina, Adv. Synth. Catal. 346 (2004) 1553. Y. Sato, M. Sodeoka, M. Shibasaki, J. Org. Chem. 54 (1989) 4738. N.E. Carpenter, D.J. Kucera, L.E. Overman, J. Org. Chem. 54 (1989) 5846. A.B. Dounay, L.E. Overman, Chem. Rev. 103 (2003) 2945. L.F. Tietze, H. Ila, H.P. Bell, Chem. Rev. 104 (2004) 3453. M. Shibasaki, M.V. Erasmus, T. Ohshima, Adv. Synth. Catal. 346 (2004) 1533. F. Ozawa, A. Kubo, T. Hayashi, J. Am. Chem. Soc. 113 (1991) 1417. F. Ozawa, A. Kubo, Y. Matsumoto, T. Hayashi, Organometallics 12 (1993) 4188. F. Ozawa, T. Hayashi, J. Organomet. Chem. 428 (1992) 267. Y. Koga, M. Sodeoka, M. Shibasaki, Tetrahedron Lett. 35 (1994) 1227. J.S. Johnson, D.A. Evans, Acc. Chem. Res. 33 (2000) 325. K.A. Jørgensen, M. Johannsen, S. Yao, H. Audrain, J. Thorhauge, Acc. Chem. Res. 32 (1999) 605. J. Bayardon, D. Sinou, M. Guala, G. Desimoni, Tetrahedron: Asymm. 15 (2004) 3195. M.A. Stark, G. Jones, C.J. Richards, Organometallics 19 (2000) 1282. G. Helmchen, A. Pfaltz, Acc. Chem. Res. 33 (2000) 336. O. Loiseleur, M. Hayashi, M. Keenan, N. Schmees, A. Pfaltz, J. Organomet. Chem. 576 (1999) 16. K. Salvekumar, M. Valentini, P.S. Pregosin, A. Albinati, F. Eisentraeger, Organometallics 19 (2000) 1299. W.-P. Deng, X.-L. Hou, L.-X. Dai, X.-W. Dong, Chem. Commun. (2000) 1483. C. Bolm, K. Weickhardt, M. Zehnder, T. Ranff, Chem. Ber. 124 (1991) 1173. A. Gsponer, T.M. Schmid, G. Consiglio, Helv. Chim. Acta 84 (2001) 2986. K.J. Miller, M.M. Abu-Omar, Eur. J. Org. Chem. (2003) 1294. T. Pei, R.A. Widenhoefer, Org. Lett. 2 (2000) 1469. N.S. Perch, A. Widenhoefer, J. Am. Chem. Soc. 121 (1999) 6960. X. Wang, S.Z. Stankovich, R.A. Widenhoefer, Organometallics 21 (2002) 901. M. Svensson, U. Bremberg, K. Hallman, I. Csoregh, C. Moberg, Organometallics 18 (1999) 4900. R.C. Larock, J.M. Zenner, J. Org. Chem. 60 (1995) 482. G.A. Grasa, R. Singh, E.D. Stevens, S.P. Nolan, J. Organomet. Chem. 687 (2003) 269.

2858

D.W. Dodd et al. / Inorganica Chimica Acta 359 (2006) 2850–2858

[29] P. Pelagatti, M. Carcelli, M. Costa, S. Ianelli, C. Pelizzi, D. Rogolino, J. Mol. Cat. A 226 (2005) 107. [30] T. Kawano, T. Shinomaru, I. Ueda, Org. Lett. 4 (2002) 2545. [31] F.R. Hartley, Organomet. Rev. A 6 (1976) 119.

[32] T. Ukai, H. Kawazura, Y. Ishii, J.J. Bonnett, J.A. Ibers, J. Organomet. Chem 65 (1974) 253. [33] R.A. Gossage, H.A. Jenkins, P.N. Yadav, Tetrahedron Lett. 45 (2003) 7689.