Asymmetric allylic substitution reactions with a xylophos-Pd catalyst

Article abstract of DOI:10.1007/s007060070025

The chiral diphosphine ligand xylophos (1) was tested as an auxiliary in palladium catalyzed allylic substitution reactions. Whereas its activity was found to be generally good only in the case of 1,3-diphenylprop-2-en-1-yl acetate, a fair level of asymmetric induction was achieved with sodium dimethyl malonate (83% ee) and benzylamine (66% ee) as nucleophiles.

Full text of DOI:10.1007/s007060070025

Monatshefte fuÈr Chemie 131, 1173±1179 (2000)
Asymmetric Allylic Substitution Reactions
with a Xylophos-Pd Catalyst
1
1
1
1
Á
Oscar Pamies , Aurora Ruiz , Gemma Net , Carmen Claver ,
Hermann Kalchhauser2, and Michael Widhalm2;Ã
1
Â
Facultat de Quõmica, Universitat Rovira i Virgili, E-43005 Tarragona, Spain
È
Institut fuÈr Organische Chemie, Universitat Wien, A-1090 Wien, Austria
2
Summary. The chiral diphosphine ligand xylophos (1) was tested as an auxiliary in palladium
catalyzed allylic substitution reactions. Whereas its activity was found to be generally good only in
the case of 1,3-diphenylprop-2-en-1-yl acetate, a fair level of asymmetric induction was achieved
with sodium dimethyl malonate (83% ee) and benzylamine (66% ee) as nucleophiles.
Keywords. Allylic substitution; Asymmetric catalysis; Allylic amination.
Introduction
Chiral auxiliaries derived from the chiral pool have attracted much attention,
rendering tedious optical resolution procedures obsolete and being available in large
amounts at reasonable prices. Particulary with phosphorus ligands containing a
carbohydrate backbone impressive results have been obtained in various asymmetric
transformations like hydrogenation, hydrocyanation, and hydroformylation [1].
One of the more recently developed ligands belonging to this group is 1,2-O-
isopropylidene-3,5-bis-diphenylphosphane-ꢀ-D-xylofuranose (xylophos, 1), a dipho-
sphino ligand accessible from xylofuranose in few steps [2] whose high ef®ciency in
asymmetric hydrogenation [2a] and hydroformylation reactions has recently been
demonstrated [2b]. The obtained results encouraged us to investigate its scope and
Ã
Corresponding author

Á
O. Pamies et al.
1174
limitations also in Pd-catalyzed allylic substitution reactions and to compare
reactivity and enantioselectivity to other ligands of similar structure [3]. The
appropriate catalyst was either prepared in situ from 1 and [Pd(allyl)Cl]2, Pd(OAc)2,
or Pd2(dba)3, or added to the reaction mixture as preformed xylophos-Pd(II) complex
(10). As test reactions, alkylations of linear and cyclic allyl acetates with the carbon
nucleophile dimethyl malonate and four N-nucleophiles were investigated.
Results and Discussion
Allylic alkylation with dimethyl malonate
Typical test reactions (Scheme 1) with 1,3-disubstituted allylic substrates and
dimethyl malonate under standard conditions were performed showing good
enantioselectivity of up to 78% ee for substrate 2a, but only 2±13% ee for aliphatic
substrates (Table 1). In order to improve the asymmetric induction, optimization of
the reaction conditions was attempted. Changing the solvent from THF to CH3CN
resulted in a heterogeneous reaction mixture, slow conversion, and side reactions.
Even after three days only 27% of the desired product together with 24% of 1,3-
diphenylpropenol was obtained, and 14% of unreacted starting material were
recovered. In CH2Cl2 at room temperature the reaction was complete within 24 h,
but with a lower level of asymmetric induction (61% ee). Neither variation of the Pd
precursor nor of the ratio of ligand:Pd improved the enantioselectivity. The sharp
drop in enantioselectivity with 1 mol% of ligand and the even more dramatic one
upon application of the preformed xylophos-Pd(II) complex 10 points to the presence
of different catalytically active species coexisting under the reaction conditons
applied; in the latter case, even a reversal of the absolute con®guration was observed.
In order to detect presumably catalytically active intermediates, 31P NMR
experiments were performed. Recording the spectrum of a mixture of 10 and 1 (2:1)
Scheme 1

Allylic Substitution Catalyzed by Xylophos-Pd
1175
Table 1. Allylic alkylations with dimethyl malonate
Substrate
Yield/%
Con®guration
ee/%
Conditions
(HPLC/GC)
2a
2a
2a
2a
86
79
79
78
(S)
(S)
(R)
(R)
78
38
2
‡1 mol% xylophos
‡1 mol% 10
53
‡1 mol% 10
‡1 mol% xylophos
‡1 mol% 10
2a
87
(R)
35
‡1 mol% xylophos
‡AgOAc (excess)
‡Pd(OAc)2
2a
2a
2a
2a
3a
4a
5a
6a
78
85
27
66
85
87
89
93
(S)
(S)
(S)
(S)
(S)
(R)
(R)
(R)
55
61
79
83
13
2
CH2Cl2
CH3CN, 72 h
10ꢁC, 48 h
8
3
in CD2Cl2 gave neither any evidence for ligand exchange nor for the formation of
other complexes in equilibrium (1 ‡ Pdꢀ1†Cl2 „ Pdꢀ1†2Cl2). Upon addition of an
excess of 1,3-diphenylprop-2-en-1-yl acetate the spectrum remained unchanged.
Only after adding a small amount of sodium hydride the solution turned orange, the
signals due to the free ligand 1 disappeared, the intensity of the resonances
attributed to 10 decreased, and several broad lines could be detected between 0 and
15 ppm. These observations are in agreement with the accepted mechanism which
requires the formation of a Pd(0) intermediate to which the allylic substrate is added
oxidatively to start the catalytic cycle. In the NMR experiment, the Pd(0) inter-
mediate is assumed to be formed in two steps from 10. Upon reaction of 10 with
sodium hydride, a chloropalladium(II) hydride is generated, and from this HCl is
(formally) reductively eliminated to give a Pd(0) species. For the catalytic reaction
this means that only at this stage any free ligand being present in excess may take
part in the reaction, changing eventually equilibria between competing catalytically
active species. The in¯uence of the Cl ligands was not further investigated, but the
role of halide in facilitating pseudo-allyl rotation processes at palladium has been
discussed in detail in the literature [4]. Lowering the temperature to 10ꢁC resulted
in a considerably slower reaction rate, but increased ee to 83%. The use of BSA
(N,O-bis-(trimethylsilyl)-acetamide) instead of NaH gave irreproducible results
(40±75% ee for 2a) which can be attributed to partial cleavage of the ketal ring of
xylophos caused by the presence of (CH3)3Si-OAc.
Allylic substitution of 1,3-diphenylpropenyl acetate (2a) with N-nucleophiles
With the more promising substrate 2a, allylic substitution reactions with four N-
nucleophiles were performed as well (Scheme 2); the results are given in Table 2. In

Á
O. Pamies et al.
1176
Scheme 2
Table 2. Allylic substitution of 2a with N-nucleophiles
Nucleophile
Product
Yield/%
ee/%1
(HPLC)
Benzylamine
2c
2d
2e
2f
55
66
59
14
33
K-Phthalimide2
Na-Tosylamide
Na-Benzoylhydrazide
11
49
41
1
2c: Daicel-OD1, 0.20% Et2NH / 0.25% 2-PrOH, 99.55% n-hexane; 2d: Daicel-OD1, 2% 2-PrOH /
98% n-hexane; 2e,f: Daicel-OJ1, 15% 2-PrOH / 85% n-hexane; 56% of starting material were
recovered
2
all cases the reactions proceeded slower compared with dimethyl malonate and
afforded products 2c±f in 11±55% yield with 14±66% ee.
Conclusions
It is demonstrated that the diphosphine ligand xylophos is suited as a chiral
auxiliary in palladium catalyzed allylic substitution reactions. The degree of
asymmetric induction was found to be moderate with a maximum of 83% ee. These
results may be attributed to the medium conformative ¯exibility of both the six-
membered chelate structure and the isopropylidene furanoside moiety, giving rise to

Allylic Substitution Catalyzed by Xylophos-Pd
1177
the coexistence of several catalytically active species in a labile equilibrium. From
this point of view xylophos behaves according to expectations: its ef®ciency ranges
between the excellent one of diphosphines forming ®ve-membered chelates and the
poor one of diop analogues involving seven-membered chelates [1g].
Experimental
General
NMR spectra were recorded on a Varian Gemini 300 NMR spectrometer at 300 (1H), 75.4 (13C), and
1
121.4 (31P) MHz in CD2Cl2 in 5 mm sample tubes at 300 K. H and 13C spectra were referenced to
internal TMS, 31P chemical shifts are quoted relative to external 85% H3PO4. The numbering of
carbon and phosphorus atoms refers to that given in the formula of 10; interchangeable assignments
are marked by an asterisk.
[PdCl2(xylophos)] (10; C32H32Cl2O3P2Pd)
55.0 mg (1.05 mmol) of xylophos were added to a solution of 38.3 mg (1 mmol) of [PdCl2(PhCN)2]
in 2 cm3 of CH2Cl2. After 10 min, the product was precipitated by addition of diethyl ether; ®ltration
afforded 65 mg (91%) of 10.
1H NMR: ꢁ ˆ 1:19 (s, 3H, CH3), 1.26 (s, 3H, CH3), 2.42 (ddd, 1H, H-50, 3J(H-50, H-5) ˆ 15.8 Hz,
2
3J(H-50, H-4) ˆ 8.6 Hz, J(H-50, P-1) ˆ 5.8 Hz), 2.80 (m, 1H, H-5), 2.91 (t, 1H, H-3, 3J(H-3, H-
3
4) ˆ 2J(H-3, P-2) ˆ 6.9 Hz), 4.49 (m, 1H, H-4), 4.74 (dd, 1H, H-2, J(H-2, H-1) ˆ 4.1 Hz, 3J(H-2, P-
3
3
2) ˆ 6.9 Hz), 5.60 (dd, 1H, H-1, J(H-1, H-2) ˆ 4.1 Hz, J(H-1, P-1) ˆ 2.1 Hz), 7.30±8.20 (m, 10H,
13
Ã
Ã
CH ˆ ) ppm; C NMR: ꢁ ˆ 26:1 (C-7 ), 26.3 (C-8 ), 26.8 (dd, C-5, J(C-5, P-2) ˆ 19.5 Hz, J(C-5,
P-1) ˆ 32.5 Hz), 45.2 (dd, C-3, J(C-3, P-2) ˆ 23.9 Hz, J(C-3, P-1) ˆ 10.8 Hz), 73.3 (d, C-4, J(C-4,
P)1 ˆ 4.5 Hz), 82.4 (t, C-2, J(C-2, P-2) ˆ 2.9 Hz), 103.6 (C-1), 111.8 (C-6), 124.3, 125.0, 125.2, 125.8
(Carom), 127.4±135.3 (CHarom) ppm; 31P NMR: ꢁ ˆ 18:2 (d, P-1, J(P, P) ˆ 7.4 Hz), 22.8 (d, P-2)
2
ppm; calcd.: C 54.60, H 4.58; found: C 54.47, H 4.65.
Allylic alkylations with dimethyl malonate [5]
4 cm3 of dry THF were degassed, [Pd(ꢂ3-allyl)Cl]2 (1.8 mg, 0.005 mmol, 1 mol% Pd), xylophos
(10.5 mg, 0.02 mmol, 2 mol%), and 1 mmol of substrate were added in sequence, and the solution was
stirred for 15 min. A solution of malonate anion was prepared by adding NaH dispersion (63 mg,
1.5 mmol) in portions to a degassed solution of dimethyl malonate (198 mg, 171mm3, 1.5 mmol) in
4 cm3 of THF. After 30 min, the resulting turbid solution was added via a te¯on tube to the ice-cooled
solution containing allyl acetate 2a±6a and the catalyst, and the mixture was stirred overnight at room
temperature. The reaction was quenched by addition of a small amount of 2 N HCl, and the mixture
was extracted with Et2O (3Â10 cm3). The combined extracts were washed with brine until neutral and
dried over Na2SO4. After removal of the solvent the crude products were puri®ed by column
1
chromatography (silica gel); the purity of the isolated products was checked by H NMR spectro-
scopy.
2b: Column: 10Â2.2 cm, PE:CH2Cl2 ˆ 50:50; UV/Vis: ꢃmax ˆ 290 nm; ee: chiral HPLC (Chiralcel
OD-H1, 250Â4.6 mm, 2% 2-PrOH / 98% n-hexane); speci®c rotation calculated for optically pure
20
D
(S)-2b: ‰ꢀŠ
ˆ
22:4 (c ˆ 1.8, CHCl3) [6].
3b: Column: 25Â2.5 cm, PE:Et2O ˆ 75:25; UV/Vis: ꢃmax ˆ 220 nm; ee: enantioselective GC:
1
(50% octakis-(6-O-methyl-2,3-di-O-pentyl)-ꢄ-cyclodextrin, FS, 0.25Â25 m, 0.5 atm H2, 55ꢁC); H
1
Coupling phosphorus atom not identified

Á
O. Pamies et al.
1178
NMR: shift of a doublet of one of the terminal CH3 groups upon addition of Eu(hfc)3; speci®c rotation
20
calculated for optically pure (S)-3b: ‰ꢀŠ
ˆ
31:9 (c ˆ 1.1, CHCl3) [7].
D
4b: Column: 25Â2.5 cm, PE:Et2O ˆ 75:25; UV/Vis: ꢃmax ˆ 220 nm; ee: enantioselective GC:
(50% 6T-2,3-diacetyl-ꢅ-cyclodextrin, FS, 0.25Â25 m, 0.5 atm H2, 90ꢁC); speci®c rotation calculated
20
D
for optically pure (S)-4b: ‰ꢀŠ
ˆ
98:7 (c ˆ 2.27, CHCl3) [8].
5b: Column: 25Â2.5 cm, PE:Et2O ˆ 75:25; UV/Vis: ꢃmax ˆ 220 nm; ee: enantioselective GC:
(50% 6T-2,3-dimethyl-ꢅ-cyclodextrin, FS, 0.25Â25 m, 0.5 atm H2, 110ꢁC); speci®c rotation
20
D
calculated for optically pure (S)-5b: ‰ꢀŠ
ˆ
46:1 (c ˆ 2.86, CHCl3) [8].
6b: Column: 25Â2.5 cm, PE:Et2O ˆ 75:25; UV/Vis: ꢃmax ˆ 220 nm; ee: see 5b; speci®c rotation
20
D
calculated for optically pure (S)-6b: ‰ꢀŠ
ˆ
7:8 (c ˆ 3.04, CHCl3) [8].
Allylic substitution of 2a with N-nucleophiles
2c [9]: The reaction was conducted in 3 cm3 of dry CH2Cl2 at room temperature on a 1 mmol scale
similar as described above for allylic alkylations with dimethyl malonate. As nucleophile,
benzylamine (2 mmol, 214 mg) was added, and the solution was stirred for 48 h. After dilution with
20 cm3 of Et2O, 5 cm3 of saturated NH4Cl solution were added. The organic phase was separated,
washed with NH4Cl, dried over MgSO4, and evaporated. After chromatography on silica gel (column
25Â2 cm, EE:PE ˆ 10:90) the pure product was obtained as a yellow oil.
2d: The reaction was run in THF on a 1 mmol scale. As nucleophile, K-phthalimide (2 mmol,
370 mg) was added as a solid at room temperature, and the mixture was stirred at 50ꢁC for 72 h. For
work-up, cf. 2c.
2e,f: The reactions were performed exactly as described in Ref. [9] (nucleophiles: Na-tosylamide
(2e), Na-benzoylhydrazide (2f)).
Acknowledgments
Â
We thank the Ministerio de Educacion y Cultura for ®nancial support (PB97-0407-C05-01) and the
Á
Generalitat de Catalunya (CIRIT) for awarding a research grant to O. Pamies.
References
[1] For the synthesis and application of carbohydrate auxiliaries, see (i) with phosphinite function-
alities: a) RajanBabu TV, Ayers TA, Casalnuovo AL (1994) J Am Chem Soc 116: 4101; b)
RajanBabu TV, Ayers TA (1994) Tetrahedron Lett 35: 4295; c) RajanBabu TV, Casalnuovo AL
(1996) J Am Chem Soc 118: 6325; d) RajanBabu TV, Ayers TA, Halliday GA, You KK, Calabrese
JC (1997) J Org Chem 62: 6012; e) RajanBabu TV, Radetich B, You KK, Ayers TA, Casalnuovo
AL, Calabrese JC (1999) J Org Chem 64: 3429; f) Yan Y-Y, RajanBabu TV (2000) J Org Chem 65:
900; g) Yan Y-Y, RajanBabu TV (1999) Org Lett 2: 199; (ii) with phosphane functionalities: h)
Lafont D, Sinou D, Descotes G (1979) J Organomet Chem 169: 87; i) Johnson TH, Rangarajan G
(1980) J Org Chem 45: 62; j) Habus I, Raza Z, Sunjic V (1988) Croat Chem Acta 61: 857; k) Sunjic
V, Habus I, Snatzke G (1989) J Organomet Chem 370: 295; l) Brown JM, Cook SJ, Khan R (1986)
Tetrahedron 42: 5105; m) Li C, Bernet B, Vasella A, Broger EA, Meili A (1991) Carbohydr Res
216: 149; n) Sawamura M, Kitayama K, Ito Y (1993) Tetrahedron Asymm 4: 1829; o) Gilbertson
SR, Chang WT (1995) J Org Chem 60: 6226; (iii) with phosphite functionalities: p) Reetz M,
Waldvogel SR (1997) Angew Chem Int Ed Engl 36: 865; q) Buisman GJH, Martin ME, Vos EJ,
Á
Klootwijk A, Kamer PC, van Leeuwen PWNM (1995) Tetrahedron Asymm 6: 719; r) Pamies O,
Á
Net G, Ruiz A, Claver C. (2000) Eur J Inorg Chem 6: 1287; s) Pamies O, Net G, Ruiz A, Claver C
(2000) Tetrahedron Asymm 11: 1097
[2] a) Pamies O, Net G, Ruiz A, Claver C (2000) Eur J Inorg Chem 6: 1287; b) Pamies O (1999) PhD
Thesis, Tarragona, Spain

Allylic Substitution Catalyzed by Xylophos-Pd
1179
[3] For allylic substitution reactions cf. a) Tsuji J (1997) In: Palladium Reagents and Catalysis, Wiley,
p 290; b) Trost BM, van Vranken DL (1996) Chem Rev 96: 395; c) Pfaltz A (1999) Allylic
Substitution Reactions. In: Jacobsen EN, Pfaltz A, Yamamoto H (eds) Comprehensive Asym-
metric Catalysis, vol 2. Springer, Heidelberg, p 833; d) see Ref. [1g]
[4] a) Hansson S, Norrby PA, Sjogren MPT, Akermark B, Cucciolito ME, Giordano F, Vitagliano A
È
(1993) Organometallics 12: 4940; b) Gogoll A, Ornebro J, Grennberg H, Backvall JE (1994) J Am
Chem Soc 116: 3631; c) Andersson PG, Harden A, Tanner D, Norby PO (1995) Chem Eur J 1: 12
[5] Widhalm M, Wimmer P, Klintschar G (1996) J Organomet Chem 523: 167
[6] Sprinz J, Helmchen G (1993) Tetrahedron Lett 34: 1769
[7] Widhalm M, Bourghida M (unpublished results) see also: Trost BM, Krueger AC, Bunt RC,
Zambrano J (1996) J Am Chem Soc 118: 6520
[8] Sennhenn P, Gabler B, Helmchen G (1994) Tetrahedron Lett 35: 8595
[9] von Matt P, Loiseleur O, Koch G, Pfaltz A, Lefeber C, Feucht T, Helmchen G (1994) Tetrahedron
Asymm 5: 573
Received June 13, 2000. Accepted July 3, 2000