Hammett 13C NMR and X-ray studies of π-allylpalladium phosphinooxazoline chiral ligand complexes

Article abstract of DOI:10.1016/j.tetlet.2005.01.032

The transmission of substituent effects across the palladium center to the allyl carbons is monitored by ¹³C NMR and X-ray crystallography as a way to probe electronic asymmetry in chiral ligand design. Based on their similar chemical shift trends and X-ray structures, the π-allylpalladium intermediates provide a good model for early transition state reactions, which are less sensitive to electronic perturbations. Hammett analysis supports an electronic basis for enantioselection that increases as the transition state becomes later.

Full text of DOI:10.1016/j.tetlet.2005.01.032

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 1441–1445
13
Hammett C NMR and X-ray studies of p-allylpalladium
phosphinooxazoline chiral ligand complexes
Paul B. Armstrong,a Lisa M. Bennett,b Ryan N. Constantine,a Jessica L. Fields,a
Jerry P. Jasinski,b Richard J. Staplesc and Richard C. Bunta,*
aDepartment of Chemistry and Biochemistry, Middlebury College, Middlebury, VT 05753, USA
bDepartment of Chemistry, Keene State College, Keene, NH 03435, USA
cX-ray Crystallography Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA 02138, USA
Received 16 August 2004; revised 7 January 2005; accepted 10 January 2005
Available online 23 January 2005
Abstract—The transmission of substituent effects across the palladium center to the allyl carbons is monitored by 13C NMR and
X-ray crystallography as a way to probe electronic asymmetry in chiral ligand design. Based on their similar chemical shift trends
and X-ray structures, the p-allylpalladium intermediates provide a good model for early transition state reactions, which are less
sensitive to electronic perturbations. Hammett analysis supports an electronic basis for enantioselection that increases as the tran-
sition state becomes later.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Phosphinooxazoline (PHOX) chiral ligands have proven
extremely effective at inducing asymmetry in palladium-
catalyzed allylic-substitution reactions.1 These ligands
are proposed to function by controlling the ratio of
the exo to endo p-allylpalladium intermediates and then
directing nucleophilic addition trans to phosphorus (the
better acceptor ligand) in the favored exo intermediate.
The low-temperature NMR structure of the initially
formed alkene–palladium complex has established this
Ôtrans to phosphorus in exoÕ addition as the major path-
way.2 Nonetheless, this does not preclude other explana-
tions for its favorability (e.g., steric interactions with the
nucleophile or in the product alkene complex) nor does
it provide any information about the pathway(s) leading
to the minor enantiomer.
tuent (X) (1a–f: 89–93% ee). In contrast, for reactions
with benzylamine as the nucleophile the eeÕs showed a
great sensitivity to the backbone substituent (X) (1a–f:
16–67% ee). These results suggest that the position of
the transition state also influences the enantioselectivity.
To gain further insight into the basis for enantioselecti-
vity and the position of the transition state, we measured
the 13C NMR chemical shifts and examined several
X-ray structures of the p-allylpalladium intermediates
themselves (2a–f), which are regarded as good models
for an early transition state structure.4 13C NMR chemi-
cal shifts have been successfully correlated to both
positive charge density5 and regioselectivity of
nucleophilic attack with substituted 1,3-diphenylallyl-
palladium complexes with achiral ligands by Hammett
analysis.6 In the case of ligands 1a–f the electron donat-
ing or withdrawing substituent is on the ligand back-
bone and should show Hammett correlations via
electronic transmission across the palladium center to
the allyl termini. Based on the proposed mechanism
for enantioselection, the biggest effect was expected on
C3 in the exo diastereomer.
Our Hammett studies of the enantioselectivity with elec-
tronically modified PHOX ligands (1a–f) provided addi-
tional support for the electronic basis of the trans to
phosphorus addition mechanism, but they also demon-
strate that the nucleophile plays an important role in
determining the enantioselectivity.3 In particular, for
reactions with sodiodimethyl malonate as the nucleo-
phile the eeÕs were less sensitive to the backbone substi-
Palladium complexes 2a–f were prepared from the corre-
sponding chiral ligands (1a–f)3 and 1,3-diphenylallylpal-
ladium chloride dimer by reaction with silver
tetrafluoroborate in acetone (Scheme 1).6b,7 The com-
plexes were obtained in essentially quantitative yield
Keywords: Asymmetric catalysis; Hammett analysis; N,P-ligands;
Palladium; p-Allyl complexes.
Corresponding author. Tel.: +1 802 443 2559; fax: +1 802 443
2072; e-mail: rbunt@middlebury.edu
*
0040-4039/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.01.032

1442
P. B. Armstrong et al. / Tetrahedron Letters 46 (2005) 1441–1445
correlate with higher observed enantioselectivity
BF4
X
(increasing 1a to 1f).3 Contrary to our expectations
though, no pronounced changes in the chemical shifts
of C3 are observed in the exo diastereomer relative to
the others; the chemical shifts of C1 and C3 change by
about the same amount for both diastereomers. Like-
wise, the differences between the two allyl termini
(D(dC3ꢀdC1)) remain constant at 30.1 ppm for the exo
complex and 20.4 ppm for the endo complex. The larger
Dd for the exo diastereomer has been reported in 1,3-
dialkylallyl complexes of 2d and related ligands.13 In
addition, this Dd difference has been proposed to corre-
late to diastereomer reactivity for several S,N- and P,N-
ferrocene ligand palladium complexes.14 Therefore, it
seems reasonable to conclude that the exo complexes
of 2a–f are either more reactive or more selective (or
both) than the endo complexes.15 This provides an
explanation for how the enantioselectivity obtained with
these ligands (>98% ee in THF1) can exceed the exo:
endo ratio (ꢁ8:1 in THF12).16
X
Ph
Ph
O
AgBF4
O
Pd
+
N
Pd
PPh2
Ph
Cl
N
PPh2
i-Pr
Ph
2
i-Pr
1a-f
2a-f
Scheme 1. Synthesis of p-allylpalladium complexes 2a–f: X = NMe2,
OMe, Me, H, F, Cl (only exo diastereomer shown).
and sufficient purity for NMR spectroscopy. Analyti-
cally pure samples and crystals suitable for X-ray crystal
diffraction (see below) were grown by slow vapor diffu-
sion of diethyl ether into a solution of the complex in
acetonitrile. The 1H and 13C NMR assignments for
2a–f were made based on the literature data for the anal-
ogous hexafluoroantimonate salt of 2d (X = H)8 and
confirmed by H NMR and 13C HMQC NMR experi-
1
ments (Table 1). The exo:endo ratio was determined by
integration of the 1H NMR spectra. The signals for
the major diastereomer were assigned to the exo com-
plex, which is known to predominate in solution.1a,8,9
These dD trends can also be seen in the similar slopes of
the Hammett plots of rp and rm versus the 13C NMR
shifts of C1 (trans to N) and C3 (trans to P) in both
the exo (solid symbols) and endo (open symbols) com-
plexes (Fig. 1 and 2).17 The data fits better to rp than
rm, as judged by R2.18 The better fit to rp could suggest
a more significant substituent influence via N; however,
the uniformity of the chemical shift changes of both C1
and C3 indicates a similarity in the effects transmitted via
N (para to X and trans to C1) and P (meta to X and trans
to C3). Thus, the Hammett fit to rp is perhaps best inter-
preted as a general electron donating/withdrawing effect
with a strong resonance component. This makes sense
as the allyl fragment is p-conjugated to the ligand
via Pd and rp is known to have a strong resonance
component.17
The ratio of exo to endo diastereomers is believed to play
a role in the enantioselectivity of PHOX ligands despite
their rapid equilibration in solution.1,10 Since nucleo-
philic attack trans to phosphorus in the endo complex
gives the opposite (minor) enantiomer and would, pre-
sumably, also be electronically favored, this pathway
must be disfavored by some other means. A change in
the exo:endo ratio as a function of substituent would
be one possibility; however, complexes 2a–f show a con-
stant, approximately 6:1 exo:endo ratio for all the sub-
stituents investigated. Consequently, the substituentÕs
effect on the ligandÕs enantioselectivity must have
another origin.11,12
The similarities of 2a–f by 13C NMR in terms of chemi-
cal shift changes and Dd led us to examine some of their
X-ray structures to see if any bond length or angle differ-
ences could be detected that might explain their differing
enantioselectivities. Suitable crystals were obtained for
In contrast to the exo:endo ratios, the chemical shifts of
the allyl carbons do change with different substituents
and show the expected downfield trend with electron
withdrawing groups. The increased deshielding does
Table 1. 13C NMR chemical shifts of p-allylpalladium complexes 2a–f
X
X
O
O
Ph
Ph
Ph
N
P
N
P
Ph
Ph
~ 6 : 1
Pd
Pd
i-Pr
Ph
i-Pr
C3
C1
Ph
Ph
C3
C1
Exo
Endo
X
exo:endoa
C1 (exo)
C3 (exo)
Ddb (exo)
C1 (endo)
C3 (endo)
Ddb (endo)
NMe2
OMe
Me
H
85:15
86:14
85:15
86:14
86:14
83:17
69.87
70.57
70.58
70.74
71.19
71.28
100.03
100.65
100.80
100.85
101.26
101.26
30.16
30.08
30.22
30.11
30.07
29.98
73.91
74.62
74.61
74.73
75.32
75.32
94.22
95.02
95.18
95.14
95.63
95.81
20.31
20.40
20.57
20.41
20.31
20.49
F
Cl
a Ratio determined by integration of 1H NMR spectra taken in CD2Cl2 and based on the average of all resolved peaks for each complex.
b Dd = dC3ꢀdC1.

P. B. Armstrong et al. / Tetrahedron Letters 46 (2005) 1441–1445
1443
105
100
95
90
85
80
75
70
65
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
σ
p
y = 101.00 + 1.1951x R2= 0.943
y = 95.404 + 1.4536x R2= 0.938
y = 74.971 + 1.3433x R2= 0.891
y = 70.922 + 1.3274x R2= 0.933
C3 exo
C3 endo
C1 endo
C1 exo
Figure 4. ORTEP drawing (40% ellipsoids) of 2f with the hydrogens
ꢀ
and BF4 omitted for clarity.
Figure 1. Hammett plot of d 13C versus rp for complexes 2a–f.
pared with the known structure of 2d8 (Table 2). While
some variation in bond lengths and angles between the
three structures are apparent, no striking differences
were found that suggest any structural basis for the sub-
stituent effects. In addition, small perturbations in twist
angle (s) and other ligand and allyl conformations have
been attributed to crystal packing forces.8 Thus, it seems
that by NMR and X-ray crystallography these allylpal-
ladium complexes are remarkably similar despite the dif-
ferences obtained with the ligands in some asymmetric
reactions.
105
100
95
90
85
80
75
70
65
-0.2
-0.1
0
0.1
0.2
0.3
0.4
Based on the similarity of both the 13C NMR data and
the X-ray structure data for the allylpalladium com-
plexes, very little difference in enantioselectivity is
expected with ligands 1a–f for reactions proceeding
though an early transition state. This is precisely what
was observed for reactions with sodiodimethyl malonate
as the nucleophile.3 In contrast, reactions with benzyl-
amine as the nucleophile in which large differences in
enantioselectivity were observed with ligands 1a–f likely
proceed through a later transition state. With more
bond formation between the nucleophile and C3 (for
the major enantiomer via exo diastereomer) in a later
transition state, the palladium–ligand complex accepts
more electron density back from the allyl fragment as
it rotates to form the palladium(0)–alkene complex.
The ability of the ligand (presumably primarily through
σ
m
y = 100.63 + 1.7978x R2= 0.737
y = 94.944 + 2.2262x R2= 0.760
y = 74.527 + 2.2422x R2= 0.857
y = 70.493 + 2.1222x R2= 0.824
C3 exo
C3 endo
C1 endo
C1 exo
Figure 2. Hammett plot of d 13C versus rm for complexes 2a–f.
one electron donating complex (ent-2b, X = OMe) and
one electron withdrawing complex (2f, X = Cl). Their
structures were solved19 and ORTEP plots20 are shown
in Figures 3 and 4. Key structural parameters are com-
Table 2. Selected bond distances and angles from the X-ray crystal
structures of 2b,d,f
Parameter
2b (X = OMe) 2d (X = H)a 2f (X = Cl)
˚
Pd–C1 distance (A)
2.142(4)
2.288(4)
2.113(4)
2.2828(9)
87.46(9)
65.76(19)
113.9
2.137(2)
2.258(2)
2.096(2)
2.2668(5)
88.92(5)
66.75(8)
118.5
2.16(1)
2.250(11)
2.111(9)
2.285(3)
87.3(3)
65.6(4)
115.0
˚
Pd–C3 distance (A)
˚
Pd–N distance (A)
˚
Pd–P distance (A)
N–Pd–P (ꢁ)
C1–Pd–C3 (ꢁ)
Tilt angle, a (ꢁ)b
Twist angle, s (ꢁ)c
5.5
5.3
3.0
Figure 3. ORTEP drawing (30% ellipsoids) of ent-2b (enantiomer of
ligand 1b was used for crystal structure of 2b) with hydrogens and
a Data for complex 2d (SbF6 salt) taken from Ref. 8.
ꢀ
b Tilt angle defined between the [N, Pd, P] and [C1, C2, C3] planes.
c Twist angle defined as dihedral [N, P, C1, C3].
ꢀ
BF4 anion omitted for clarity.

1444
P. B. Armstrong et al. / Tetrahedron Letters 46 (2005) 1441–1445
2. (a) Junker, J.; Reif, B.; Junker, B.; Felli, I. C.; Reggelin,
M.; Griesinger, C. Chem. Eur. J. 2000, 6, 3281–3286; (b)
Steinhargen, H.; Reggelin, M.; Helmchen, G. Angew.
Chem., Int. Ed. 1997, 36, 2108–2110.
3. Constantine, R. N.; Kim, N.; Bunt, R. C. Org. Lett. 2003,
5, 2279–2282.
4. Auburn, P. R.; Mackenzie, P. B.; Bosnich, B. J. Am.
Chem. Soc. 1985, 107, 2033–2046.
5. Akermark, B.; Krakenbergen, B.; Hansson, S.; Vitagliano,
A. Organometallics 1987, 6, 620–628.
6. (a) Branchadell, V.; Moreno-Manas, M.; Pajuelo, F.;
Pleixats, R. Organometallics 1999, 18, 4934–4941; (b)
Moreno-Manas, M.; Pajuelo, F.; Parella, T.; Pleixats, R.
Organometallics 1997, 16, 205–209; (c) Malet, R.; Moreno-
Manas, M.; Parella, T.; Pleixats, R. J. Org. Chem. 1996,
61, 758–763.
7. See Supplementary material for experimental procedures
and full characterization data.
8. Kollmar, M.; Steinhagen, H.; Janssen, J.; Goldfuss, B.;
Malinovskaya, S. A.; Vazquez, J.; Rominger, F.; Helm-
chen, G. Chem. Eur. J. 2002, 8, 3103–3114.
phosphorus) to accept this electron density must then be
modulated by the substituents (X).
Combined with the increased reactivity or selectivity of
the exo complex (Dd = 30.1 ppm vs 20.4 ppm for endo),
the variable position of the transition state provides a
general explanation for the electronic basis for enantio-
selection. The exo complex either reacts faster (trans to
P) or more selectively (trans to P vs trans to N) than
the endo complex.15 This inherent bias is not easily per-
turbed by the backbone substituents (X) for reactions
with early transition states as the similar trends in the
13C NMR chemical shifts and X-ray structures show.
Surprisingly, these results suggest that as the reaction
coordinate shifts toward a later transition state, where
steric effects are often invoked to explain enantioselec-
tivity, the electronic perturbations become more impor-
tant, not less. These electronic effects may be
superimposed on any underlying steric influences
though.
9. Baltzer, N.; Macko, L.; Schaffner, S.; Zehnder, M. Helv.
Chim. Acta 1996, 79, 803–812.
10. Kudis, S.; Helmchen, G. Angew. Chem., Int. Ed. 1998, 37,
3047–3050.
In summary, despite the potential for ligands 1a–f to
show different enantioselectivity, their 1,3-diphenylallyl-
palladium complexes (2a–f) show similar trends in
their 13C NMR chemical shifts and complexes 2b,d,f
have strikingly similar X-ray crystal structures. Hammett
analysis suggests a general correlation with the reso-
nance acceptor/donor ability of the substituent (via fit
to rp). Thus, we conclude that reactions which show
enantiomeric variation with ligands 1a–f proceed by a
transition state that is later relative to the allylpalladium
intermediates. With this better understanding of the
electronic basis for enantioselection with PHOX ligands,
the design of new chiral ligands of this and other types,
as well, should be greatly improved. We are continuing
our investigations of these ligands by synthesizing iso-
meric 5-substituted PHOX ligands to better understand
the Hammett analysis of rm and rp and the influence of
the substituent position.
11. The exo:endo ratio for the chloro-substituted complex (2f)
was determined from only two resolved peaks to due to
increased spectral overlap (data not shown). Furthermore,
the 83:17 ratio is not statistically different (based on a
standard deviation of the average of 2.0%) from the other
ratios.
12. The exo:endo ratio is solvent dependent and has been
reported as 8:1 in THF-d8, 11:1 in DMSO-d6, and 6:1 in
CDCl3 see Refs. 1 and 8.
13. Kollmar, M.; Goldfuss, B.; Reggelin, M.; Rominger, F.;
Helmchen, G. Chem. Eur. J. 2001, 7, 4913–4927.
14. (a) You, S. L.; Hou, X. L.; Dai, L. X.; Yu, Y. H.; Xia, W.;
Sun, J. J. Org. Chem. 2002, 67, 4684–4695; (b) Deng, W.
P.; You, S. L.; Hou, X. L.; Dai, L. X.; Yu, Y. H.; Xia, W.
J. Am. Chem. Soc. 2001, 123, 6508–6519.
15. The pathway leading to the minor enantiomer is not
known. If the preference for nucleophilic attack trans to
phosphorus is essentially complete as suggested by work
with unsymmetrical allyl substrates (see Ref. 8), then it
becomes an issue of reactivity alone (exo vs endo) rather
than selectivity (attack trans to N vs P).
Acknowledgements
16. The difference in dD between the exo and endo diastereo-
mers is likely the result of steric interactions between the
allyl substituent and the equatorial phenyl ring on
phosphorus that result in a decreased influence of the
trans effect in the endo complex. In complexes of 1a–f with
an unsubstituted allyl fragment lacking these steric inter-
actions the endo complex is favored (55–60%) in solution
and shows a greater dD in the 13C NMR shifts than the
minor exo complex (24 ppm vs 18 ppm). For similar
examples see Ref. 9.
This research was supported by an award from Research
Corporation, and acknowledgment is also made to the
Donors of The Petroleum Research Fund, administered
by the American Chemical Society, for support of this
research.
Supplementary data
17. Substituent constants (rp and rm) taken from: Hansch, C.;
Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165–195.
18. Analysis with dual-parameter equations (e.g., F/R, rI/rR,
etc.) gave slightly better fits (R2 > 0.96, data not shown) as
would be expected with an added degree of freedom but
provided no additional insights.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.
2005.01.032. Synthesis and characterization data for 1c
and 2a–c,e–f can be found, in the online version.
19. CCDC-239501 (2b) and CCDC-239500 (2f) contain the
supplementary crystallographic data for this paper. These
data can be obtained free of charge via http://
www.ccdc.cam.ac.uk/data_request/cif, or by emailing
data_request@ccdc.cam.ac.uk, or by contacting The Cam-
bridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033. Com-
References and notes
1. (a) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33,
336–345; (b) Helmchen, G. J. Organmet. Chem. 1999, 576,
203–214; (c) Williams, J. M. J. Synlett 1996, 705–710.

P. B. Armstrong et al. / Tetrahedron Letters 46 (2005) 1441–1445
1445
pound 2b: crystallized from CH3CN/Et2O, colorless
prisms 0.50 · 0.52 · 0.52 mm, tetragonal, space group
P41212, a = 13.6171(9), b = 13.6208(9), c = 42.281(2) A,
T = 293 K, 26,154 reflections collected, 9301 independent
reflections, 6061 reflections [I > 3r(I)], l = 0.637 mmꢀ1
Tmax/Tmin = 1.00/1.00, 469 refined parameters, hydrogen
,
˚
3
˚
a = 90, b = 90, c = 90ꢁ, V = 7842.2(8) A , Z = 8,
atoms placed geometrically after each cycle, R = 0.072,
Rw = 0.088, refinement on jFj, residual electron density
q
calcd = 1.40 g cmꢀ3, 2hmax = 60.82 A, Mo Ka radiation
˚
(k = 0.710730 A, graphite monochromoter),
H
scan,
ꢀ3.05(2.46) eꢀ
A
ꢀ3. Programs used: structure solution:
˚
˚
T = 294 K, 12,907 reflections collected, 11,888 indepen-
dent reflections, 7371 reflections [I > 2r(I)],
l = 0.570 mmꢀ1
Tmax/Tmin = 0.75/0.74, 474 refined
SIR92 (Altomare, A.; Cascarano, G.; Giacovazzo, G.;
Guagliardi, A.; Burla, M. C.; Polidori, G.; Camalli, M. J.
Appl. Cryst. 1994, 27, 435), structure refinement: CRYS-
TALS (Watkin, D. J.; Prout, C. K.; Carruthers, J. R.;
Betteridge, P. W.; Cooper, R. I. Crystals, Issue 11,
Chemical Crystallography Laboratory, Oxford, UK,
2001), weighting scheme: Prince modified Chebychev
polynomial (Watkin, D. J. Acta Crystallogr. Sect. A,
1994, 50, 411; Prince, E. Mathematical Techniques in
Crystallography and Materials Science; Springer: New
York, 1982).
,
parameters, hydrogen atoms placed geometrically after
each cycle, R = 0.047, Rw = 0.052, refinement on jFj,
residual electron density ꢀ0.80(1.02) eꢀ
A
ꢀ3. Compound
˚
2f: crystallized from CH3CN/Et2O, white prisms
0.50 · 0.52 · 0.52 mm, orthorhombic, space group
P212121, a = 13.679(6), b = 13.766(7), c = 20.76(1) A,
˚
3
˚
a = 90, b = 90, c = 90ꢁ, V = 3908.4(31) A , Z = 4,
calcd = 1.41 g cmꢀ3, 2hmax = 55.88 A, Mo Ka radiation
˚
q
˚
(k = 0.710730 A, graphite monochromoter), 2H scan,
20. Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565–566.