Novel binding modes and hemilability in atropisomeric phosphino-amino palladium complexes

Article abstract of DOI:10.1021/om030681h

Two new allyl complexes of palladium bearing the atropisomeric liganda 2-(diphenylphosphino)-2′-(dimethylamino)biphenyl (L¹) and 2-(dicyclohexylphosphino)-2′-(dimethylamino)-biphenyl (L²) have been prepared and structurally characterized by X-ray crystallography. Their dynamic behavior as well as their structural characteristics in solution have been elucidated by the use of NMR techniques. Of particular interest, Pd-N bond rupture occurs with a low barrier ( < 10 kcal/mol) in the L¹ complex; however, this does not provide a facile pathway for atropisomerization of the ligand. A coordination mode involving η² binding of carbon atoms in the (dimethylamino)phenyl moiety was also established for the dicyclo-hexylphosphino analogue L².

Full text of DOI:10.1021/om030681h

2008
Organometallics 2004, 23, 2008-2014
Novel Bin d in g Mod es a n d Hem ila bility in Atr op isom er ic
P h osp h in o-Am in o P a lla d iu m Com p lexes
J . W. Faller* and Nikos Sarantopoulos
Department of Chemistry, Yale University, 225 Prospect Street, New Haven, Connecticut 06520
Received December 4, 2003
Two new allyl complexes of palladium bearing the atropisomeric ligands 2-(diphenylphos-
phino)-2′-(dimethylamino)biphenyl (L1) and 2-(dicyclohexylphosphino)-2′-(dimethylamino)-
biphenyl (L2) have been prepared and structurally characterized by X-ray crystallography.
Their dynamic behavior as well as their structural characteristics in solution have been
elucidated by the use of NMR techniques. Of particular interest, Pd-N bond rupture occurs
with a low barrier (<10 kcal/mol) in the L1 complex; however, this does not provide a facile
pathway for atropisomerization of the ligand. A coordination mode involving η2 binding of
carbon atoms in the (dimethylamino)phenyl moiety was also established for the dicyclo-
hexylphosphino analogue L2.
Ch a r t 1. MAP a n d Its Atr op isom er ic An a logu es
In tr od u ction
The use of chiral nonracemic heterobidentate ligands
serves the primary goal of introducing steric as well as
electronic asymmetry into a transition-metal complex.
Although the rational design of such ligands begins from
consideration of monodentate or C2-symmetric ex-
amples, introduction of two different types of donors in
a bidentate ligand does not necessarily lead to the
anticipated κ2 coordination mode. Examples of this kind
of deviation from expected behavior have been found in
complexes of BINAPO,1 MOP,2,3 and MAP2,3 as well as
for complexes of the C2-symmetric ligands BINAP,4
MeO-BIPHEP,5,6 NUPHOS,7 BINOL,8 and Me2BINOL.9,10
Biphenyl analogues11 of MAP or its partially hydro-
genated counterpart H8-MAP,11 2-(diphenylphosphino)-
2′-(dimethylamino)biphenyl (L1), and 2-(dicyclohexyl-
phosphino)-2′-(dimethylamino)biphenyl (L2) are axially
chiral but have a relatively low barrier to interconver-
sion of the enantiomeric atropisomers (Chart 1).
Therefore, one might anticipate that, in contrast to
MAP, a racemization pathway would be available for
the biphenyl backbones L1 and L2 on coordination to a
metal, since the hardness of the NMe2 donor would
typically render those ligands hemilabile. In particular,
a previous12 report comparing MAP and H8-MAP com-
plexes has shown that H8-MAP prefers the P,N coordi-
nation mode. This was attributed to its larger bite angle
with respect to MAP. The binaphthyl backbone also
provides a beneficial aromatic stabilization in the cases
of MAP and MOP. Motivated by the ambiguous NMR
data in that report, i.e., the observation of a single
resonance for the methyl groups in a given isomer,
where the diastereotopicity of the Me groups in each
coordinated NMe2 moiety should have given rise to two
resonances, we have investigated the behavior of ligands
L1 and L2. For complexes of L1 and L2 with (allyl)-
palladium, we found that the coordination mode de-
pends on the nature of the PR2 group, independent of
the presence or absence of a naphthyl backbone.
(1) Cyr, P. W.; Rettig, S. J .; Patrick, B. O.; J ames, B. R. Organo-
metallics 2002, 21, 4672-4679.
(2) (a) Kocovsky, P.; Vyskocil, S.; Cisarova, I.; Sejbal, J .; Tislerova,
I.; Smrcina, M.; Lloyd-J ones, G. C.; Stephen, S. C.; Butts, C. P.; Murray,
M.; Langer, V. J . Am. Chem. Soc. 1999, 121, 7714-7715. (b) Lloyd-
J ones, G. C.; Stephen, S. C.; Murray, M.; Butts, C. P.; Vyskocil, S.;
Kocovsky, P. Chem. Eur. J . 2000, 6, 4348-4357. (c) Kocovsky, P.;
Vyskocil, S.; Smrcina, M. Chem. Rev. 2003, 103, 3213-3245.
(3) Dotta, P.; Kumar, A.; Pregosin, P. S.; Albinati, A.; Rizzato, S.
Organometallics 2003, 22, 5345-5349.
(4) Pathak, D. D.; Adams, H.; Bailey, N. A.; King, P. J .; White, C.
J . Organomet. Chem. 1994, 479, 237-245.
(5) Feiken, N.; Pregosin, P. S.; Trabesinger, G.; Scalone, M. Orga-
nometallics 1997, 16, 537-543.
(6) Feiken, N.; Pregosin, P. S.; Trabesinger, G.; Albinati, A.; Evoli,
G. L. Organometallics 1997, 16, 5756-5762.
(7) Doherty, S.; Newman, C. R.; Hardacre, C.; Nieuwenhuyzen, M.;
Knight, J . G. Organometallics 2003, 22, 1452-1462.
(8) Bergens, S. H.; Leung, P.-H.; Bosnich, B.; Rheingold, A. L.
Organometallics 1990, 9, 2406-2408.
Resu lts a n d Discu ssion
Hem ila bility in a P ,N Com p lex. The X-ray crystal
structure of (η3-allyl)Pd(L1)SbF6 (1; Figure 1) clearly
shows a P,N coordination mode in the solid state that
has a single configuration of the allyl relative to the
(9) Brunkan, N. M.; White, P. S.; Gagne, M. R. J . Am. Chem. Soc.
1998, 120, 11002-11003.
(10) Brunkan, N. M.; Gagne, M. R. Organometallics 2002, 21, 4711-
4717.
(11) (a) Tomori, H.; Fox, J . M.; Buchwald, S. L. J . Org. Chem. 2000,
65, 5334-5341. (b) Shen, X. Q.; Guo, H.; Ding, K. L. Tetrahedron:
Asymmetry 2000, 11, 4321-4327.
(12) Wang, Y.; Li, X.; Sun, J .; Ding, K. Organometallics 2003, 22,
1856-1862.
10.1021/om030681h CCC: $27.50 © 2004 American Chemical Society
Publication on Web 03/31/2004

Atropisomeric Phosphino-Amino Palladium Complexes
Organometallics, Vol. 23, No. 9, 2004 2009
F igu r e 1. ORTEP diagram of (η3-allyl)Pd(L1)SbF6 (1). The
crystal is a racemate. The enantiomer shown here has the
S,RPd configuration.
Ch a r t 2. En a n tiom er s of (a llyl)P d XYa
F igu r e 2. 1H NMR spectrum of an equilibrium mixture
of (R*,RPd*)-1 and (R*,SPd*)-1. Syn and anti resonances are
noted in upper case for the major isomer (R*,SPd*)-1. The
20 °C spectrum was taken in CDCl3, which shows better
resolution of resonances at δ ∼4.4 than CD2Cl2. The -90
°C spectrum was taken in CD2Cl2. Note that a terminal
allyl resonance, S′, has shifted under the methyl group at
δ ∼3.0, making its intensity appear anomalously large.
a
Chirality descriptors assume that the priorities are allyl
the solid-state structure of 1, would imply that four
singlet resonances (of relative intensity 3 each) should
be present. The potential hemilability of ligand L1 led
> X > Y and the centroid of the allyl is out of the plane in the
direction of the central carbon.
1
us to record the low-temperature H NMR spectra of 1
axial chirality of the ligand. Meso ligands, such as allyl,
which are not in the square plane of the palladium
produce a potential element of chirality at the pal-
ladium. Thus, as shown in Chart 2, there are two
enantiomers if no other elements of chirality are
present.13-18
When yellow crystals of 1 are dissolved in CDCl3 at
room temperature, two sets of sharp allyl resonances
typical of an η3-allyl group are observed in a ratio of
approximately 1:1. An interesting feature is that there
are only two singlet resonances (of relative intensity 6
each) which correspond to the NMe2 groups of the
R*,RPd* and R*,SPd* diastereomers that are present in
solution. The diastereotopic nature of the two methyl
groups in each of the diastereomers, as inferred from
in CD2Cl2, since a rapid dissociation-reassociation of
the NMe2 group could average the methyl environments.
Indeed, at -90 °C, the methyl group resonances of the
NMe2 moieties in both R*,RPd* and R*,SPd* became
nonequivalent, as shown in Figure 2, although the two
overlap and they were still somewhat broad, owing to
the very low barrier associated with this exchange
process (∆Gq < 10 kcal/mol). We suggest that the
coalescence of the methyl resonances of the NMe2 group
arises from Pd-N bond rupture followed by a rapid
umbrella inversion at the nitrogen and subsequent bond
rotation of the free aryl-NMe2 prior to reassociation,
as shown in Scheme 1.
As the breaking of the Pd-N bond slows at low
temperatures, the pathway allowing equilibration of
methyl sites is less accessible and decoalescence of the
methyl signals occurs. Although the analogue H8-MAP
has been previously investigated by both X-ray crystal-
lography and NMR spectroscopy, the presence of a
singlet NMe2 resonance for each of the exo and endo
diastereomers, which are analogous to the R*,RPd* and
R*,SPd* forms of 1, has received little comment.12 The
use of the coalescence of NMe2 groups to detect hemi-
(13) These isomers are often designated as exo and endo when there
is an obvious reference point out of the plane of the square-planar metal
moiety.14,15 This description, however, is ambiguous for a planar PdXY
set or in cases where a reference point is not clear. One might view
this as a case of planar chirality; however, a nomenclature based on
axial chirality was suggested by Hayashi16 and has been followed by
Kocovsky.2b This can be awkward when dealing with “axial chirality”
of both the ligand and the metal. We have chosen to view the Pd system
as one would a chiral pyramidal phosphorus atom, considering the allyl
as an 18-electron pseudoatom17 placed at the centroid of the allyl.18
(14) Sprinz, J .; Kiefer, M.; Helmchen, G.; Huttner, G.; Walter, O.;
Zsolnai, L.; Reggelin, M. Tetrahedron Lett. 1994, 35, 1523-1526.
(15) Steinhagen, H.; Reggelin, M.; Helmchen, G. Angew. Chem., Int.
Ed. 1997, 36, 2108-2110.
(17) Sloan, T. E. Stereochemical Nomenclature and Notation in
Inorganic Chemistry. Top. Stereochem. 1981, 12, 1-35.
(18) Faller, J . W.; Stokes-Huby, H. L.; Albrizzio, M. A. Helv. Chim.
Acta 2001, 84, 3031-3042.
(16) Hayashi, T.; Kawatsura, M.; Uozumi, Y. J . Am. Chem. Soc.
1998, 120, 1681-1687.

2010 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
Sch em e 1. In ter con ver sion of Dia ster eotop ic
Meth yl Gr ou p s
20 kcal/mol for the process. This suggested that the
process could be observed via line shape or saturation
transfer techniques, which could also provide informa-
tion on the mechanism of epimerization in addition to
rate data.
Having established the assignments of the R*,RPd
*
and R*,SPd* isomers, NMR methods were able to
provide information on the details of the mechanism
through which the epimerization at the metal center
occurred. This was accomplished by performing SST
(spin saturation transfer) experiments on the allylic
proton resonances at 50 °C, the temperature at which
the R*,SPd* proton Ha′ broadens, owing to exchange (∆Gq
) 18.2 kcal/mol). In particular, it was found that at 50
°C in CDCl3, saturation of the R*,SPd* proton Ha′ results
in the saturation of the R*,RPd* proton Hs′ while
saturation of R*,SPd* Ha resulted in saturation of the
R*,RPd* proton Ha. Finally, an EXSY spectrum showed
cross-peaks between the Hs′ and Ha′ (Ha′ and Hs′)
resonances as well as cross-peaks between the Hs and
Hs (Ha and Ha) resonances of different diastereomers.
These observations are in accordance with an η3-η1-
η3 hapticity process of the allyl, where the Pd-C bond
trans to the Ph2P moiety is broken and the palladium
F-bonded intermediate allows for epimerization at the
metal center (see Scheme 2). This leads, for example,
to interconversion of R,SPd and R,RPd wherein the
chirality in the atropisomeric ligand retains its stereo-
chemistry. The saturation transfer pathway that was
found at 50 °C in CDCl3 clearly indicates that epimer-
ization through atropisomerization of the biphenyl-
based ligand L1 is not occurring at a significant rate.
This ligand epimerization could occur either through a
monodentate (P) coordination mode or through a seven-
membered-ring inversion. Atropisomerism of the ligand
would only involve exchange of anti with anti protons
or, alternatively, syn protons with syn protons. One
should note that (a) there was not observable saturation
lability, owing to metal-N bond rupture, has been
previously investigated by Elguero and co-workers.19,20
Since the diastereomers of 1 are not interconverted
rapidly in this process (vide infra), it appears that there
is insufficient time between bond rupture and re-
formation of the Pd-N bond for atropisomerization to
occur at a rapid rate.21,22
By dissolution of several crystals of 1 in CD2Cl2 at
-90 °C and recording of the 1H NMR spectra of the
resulting solution as it slowly warmed to -20 °C in the
NMR instrument, identification of the allylic resonances
corresponding to the R*,SPd* (“crystal”) diastereomer
was possible. The growth of the second set of allylic
resonances that correspond to the R*,RPd* isomer oc-
curred over the course of the experiment, and upon
1
equilibration the H NMR of the warmed solution was
(22) The barrier for the free ligands L1 and L2 have not been
reported. The barrier for 2,2′-bis(diphenylphosphino)biphenyl has been
reported as 22(1) kcal/mol.23 The 13C{1H} NMR spectrum of the ligand
L1 at 40 °C in THF-d8 showed three sharp doublet resonances that
correspond to the three ipso carbons at δ 141.0 (J (P-Cphenyl) ) 15.7
Hz), 140.1 (J (P-Cphenyl) ) 15.1 Hz), and 137.7 (J (P-Cbiphenyl) ) 14.0
Hz). The ipso carbons in the two phenyls are diastereotopic, owing to
slow interconversion of the atropisomers. At 50 °C the resonances at
δ 141.0 and 140.1 became broad whereas the resonance at δ 137.7
remained sharp. Since the resonance of the ipso Cbiphenyl is not expected
to broaden with increasing rate of atropisomerization, it was used as
an internal standard for the broadness at the given temperature. When
the widths of those resonances at half-height (W1/2) were compared, it
was found that the broadening of the Cphenyl resonances was ∼3 Hz at
50 °C. This indicates ∆G* ) 17.5 kcal/mol for the atropisomerization
of the free L1. The protons attached to the R-cyclohexyl carbons in L2
are diastereotopic at room temperature, indicating a barrier >16 kcal/
mol. There are eight carbon atoms in the cyclohexyl groups that are
diastereotopic as a consequence of the pyramidal phosphorus; however,
the R-carbons directly attached directly to phosphorus as well as the
δ carbons are diastereotopic, owing to the axial chirality in the
atropisomers, for a total of 12 cyclohexyl 13C resonances. The R-carbon
resonances (in THF-d8 at 50 °C) at δ 39.4 and 36.1 are readily
identified, owing to the 17 Hz J PC coupling. These resonances show a
small broadening at this temperature, indicating that the barrier is
18.8 kcal/mol. In the 1H spectrum there are 22 diastereotopic protons,
of which one attached to an R-carbon appears at δ 2.04, to low field of
the others, which are overlapped. This proton is also slightly broadened
at 50 °C, owing to exchange. 1H EXSY spectra at this temperature al-
so show 11 cross-peaks, as would be anticipated for atropisomerizations
in particular, a cross-peak was observed between protons at δ 2.04
and 1.49, which also correlate with the 13C signals δ 39.4 and 36.1.
We were unable to measure the barrier for atropisomerization for the
Pd-L1 complex but presume it would be ∼22 kcal/mol, as found for
that with Pd-L2.
identical with that obtained by simply dissolving crys-
tals of 1 in CD2Cl2 at room temperature. Owing to
partial overlap between two resonances of a syn proton,
Hs, and an anti proton, Ha, it was difficult to assign
1
those resonances from the H NMR spectrum alone. An
HMQC spectrum was used for the unambiguous assign-
ments of those resonances. As shown in Figure 2, the
resonances for terminal protons trans to the phosphorus
(Hs, Ha) are coupled to phosphorus and are at lower field
than those that are cis to the phosphorus. Usually
resonances for anti protons are upfield of syn protons;
however, in the R*,RPd* isomer Hs′ is at higher field than
Ha′. This could be attributable to the ring current of the
biphenyl influencing the anti proton. The time required
for epimerization suggested a barrier of between 15 and
(19) Elguero, J .; Fruchier, A.; de la Hoz, A.; J alon, F. A.; Manzano,
B. R.; Otero, A.; de la Torre, F. G. Chem. Ber. 1996, 129, 589-594.
(20) Elguero, J .; Guerrero, A.; de la Torre, F. G.; de la Hoz, A.; J alon,
F. A.; Manzano, B. R.; Rodriguez, A. New J . Chem. 2001, 25, 1050-
1060.
(21) The simplest consideration of the intermediate formed upon
Pd-N bond rupture is that κ1P ligand binding is involved. It is possible,
however, that there is slippage of the metal to a P,C type of bonding
as observed with L2. Since there is little CdN character in that
complex, interconversion of the methyl environments by umbrella
inversion could occur via this intermediate, in which the metal would
continually be bound to some portion of the dimethylaniline fragment.
This would also inhibit atropisomerization in this intermediate.

Atropisomeric Phosphino-Amino Palladium Complexes
Organometallics, Vol. 23, No. 9, 2004 2011
F igu r e 4. EXSY spectrum of (η3-allyl)Pd(L1)SbF6 (1) in
CDCl3 at 50 °C, showing the region of the terminal allylic
protons cis to the phosphorus atom. The important feature
here is that there are observable cross-peaks between syn
protons of one diastereomer and anti protons of the other.
F igu r e 3. EXSY spectrum of (η3-allyl)Pd(L1)SbF6 (1) in
CDCl3 at 50 °C, showing the region of the terminal allylic
protons trans to the phosphorus atom. Note that the order
of some resonances at δ ∼4.4 has crossed over relative to
those observed at 20 °C. The important feature here is that
there are no observable cross-peaks between syn and anti
protons.
Sch em e 2. η3-η1-η3 Mech a n ism Con sisten t w ith
th e Obser ved Site Exch a n ges
F igu r e 5. ORTEP diagram of the triclinic (η3-allyl)Pd(L2)-
SbF6 (2). The crystals contain a racemate. The isomer
shown here is S,SPd
.
Novel Bin d in g in (η3-a llyl)P d (L2)SbF 6 (2). Crys-
tallographic analysis of the complex (η3-allyl)Pd(L2)SbF6
(2) shows not only that L2 has a different coordination
mode than L1 in 1 but also that crystallization occurs
in two different phases, one of which has a chiral space
group and may therefore lead to spontaneous resolution.
As seen in both crystal forms (Figures 5 and 6), which
contain different diastereomers, a P,η2(C1′-C6′) olefin
coordination mode is preferred over the anticipated P,N
coordination mode. This contrasts with the known
structures of the binaphthyl analogue MAP, which
of multiple allylic resonances upon selective saturation
of the exo Ha′ or exo Ha resonances and (b) the EXSY
spectrum (Figures 3 and 4) did not show cross-peaks
that would suggest such a transformation. This strongly
suggests that the absolute configuration of the biphenyl
backbone is effectively maintained at room temperature,
even though the averaging of the methyl groups of the
NMe2 suggests that the ligand is hemilabile.21-23
(23) Desponds, O.; Schlosser, M. Tetrahedron Lett. 1996, 37, 47-
48.

2012 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
Ta ble 1. Cr ysta llogr a p h ic Da ta for 1 a n d 2
2
2
1
(triclinic)
(orthorhombic)
color, shape
yellow, needle yellow, needle yellow, block
empirical formula
C
29H29F6NP-
PdSb
C
29H41F6NP-
PdSb
C
29H41F6NP-
PdSb
formula wt
radiation, λ/Å
T/K
764.67
776.76
776.76
Mo KR (monochr), 0.710 73
173
cryst syst
monoclinic
triclinic
orthorhombic
space group (No.) C2/ c (15)
unit cell dimens
P1h (2)
P212121 (19)
a/Å
b/Å
c/Å
36.8345(4)
9.9269(1)
35.4672(4)
90
115.0758(6)
90
11746.3(2)
16
1.729
11.1900(2)
11.8170(3)
12.4058(3)
98.3213(9)
105.6071(9)
91.0168(9)
1560.56(5)
4
10.9117(2)
16.5774(3)
17.1390(4)
90
90
90
3100.23(9)
4
1.664
R/deg
â/deg
γ/deg
V/Å3
Z
D
calcd/g cm-3
1.653
15.47
µ(Mo KR)/cm-1
16.43
15.57
cryst size/mm
0.05 × 0.19 × 0.10 × 0.14 × 0.24 × 0.24 ×
0.22 0.19
0.27
6675
0.018
total no. of rflns
Rint
24 552
0.045
12 388
0.028
F igu r e 6. ORTEP diagram of orthorhombic (η3-allyl)Pd-
(L2)SbF6 (2). A given crystal contains a single enantiomer.
no. of observns
7474 (n ) 3)
4611 (n ) 3)
3597 (n ) 2)
(I > nσ(I))
no. of params,
constraints
R1,a wR2,b GOF
739, 0
388, 0
352, 0
The isomer found in this crystal was S,RPd
.
Ch a r t 3. Com p a r ison of Bin d in g in MAP a n d L2 a
0.038, 0.037,
1.32
0.039, 0.047,
1.50
0.040, 0.044,
2.22
min, max resid
density/e Å-3
-0.73, 1.53
-1.56, 1.24
-1.06, 1.13
a
b
R1 ) ∑||Fo| - |Fc||/∑|Fo|, for all I > nσ(I). wR2 ) [∑[w(|Fo| -
|Fc|)2]/∑[w(Fo)2]]1/2
.
Ta ble 2. Selected Bon d Dista n ces (Å)a n d An gles
(d eg) for 1 a n d 2
1a
2 (triclinic)
2 (orthorhombic)
a
In the L2 complex the Pd-C1′ and Pd-C6′ bond lengths
Bond Distances
are 2.441 and 2.330 Å in the triclinic isomer but 2.361 and
2.383 Å in the orthorhombic isomer.
Pd1-P1
Pd1-N1
Pd1-C1
Pd1-C2
Pd1-C3
Pd1-C7
Pd1-C8
N1-C4
N1-C5
N1-C6
C1-C2
2.297(1)
2.211(4)
2.097(5)
2.163(5)
2.240(5)
2.281(1)
2.289(1)
2.100(5)
2.144(6)
2.235(6)
2.442(5)
2.332(4)
1.463(6)
1.466(6)
1.391(6)
1.340(8)
1.332(9)
1.441(6)
1.422(6)
1.405(6)
1.384(7)
1.382(7)
1.387(7)
1.505(6)
2.126(6)
2.161(7)
2.214(6)
2.361(6)
2.383(5)
1.455(8)
1.468(8)
1.355(8)
1.42 (1)
1.30 (1)
1.498(7)
1.396(8)
1.408(8)
1.375(9)
1.37 (1)
1.388(9)
1.493(7)
although sometimes bound as a P,C ligand, is generally
considered to bind via an η1 mode12 (Pd-C ) 2.338(11)
Å, which is relatively long) (Chart 3).
1.495(6)
1.493(6)
1.444(6)
1.409(7)
1.394(7)
1.410(4)
1.405(7)
1.377(7)
1.366(7)
1.364(7)
1.405(7)
1.494(7)
The NMe2 group is not planar, and equivalency of the
two methyl groups within the same diastereomer is
anticipated in the 1H NMR spectrum, since the umbrella
inversion at the nitrogen atom has a very low barrier,
particularly if there is little double-bond character in
the N-aryl bond. Indeed, when a CD2Cl2 solution of 2
is cooled to -90 °C, only temperature-induced shifting
of the two NMe2 singlet resonances arising from the two
diastereomers present was observed in the NMR. This
contrasts with the behavior of 1, which indicates that
the NMe2 group is rapidly dissociating from palladium.
An interesting feature of complex 2 is the existence of
a relatively large coupling constant (J s′-a′ ≈ 2.0 Hz)
between the protons Hs′ and Ha′ in both diastereomers.
This presumably arises from the fact that the terminal
allylic CH2 allylic moiety trans to the η2(C1′-C6′) group
is more twisted relative to the rest of the allyl fragment.
Again, by dissolving a triclinic single crystal at -90 °C
in CD2Cl2, we were able to observe its isomerization
toward the diastereomer present in the orthorhombic
crystal, as well as to assign the allylic resonances for
each of the isomers.
C2-C3
C6-C7
C7-C8
C8-C9
C9-C10
C10-C11
C11-C6
C7-C13
Bond Angles
P1-Pd1-N1
P1-Pd1-C7
P1-Pd1-C8
97.9(1)
84.0(1)
93.2(1)
84.3(1)
92.9(1)
a
Distances for only one molecule given.
aspects of the 13C spectrum suggest this, and details for
all shifts are included in the Experimental Section. In
general, the biphenyl carbons attached to P or N do not
show a large coordination shift upon binding of the P
or N to the metal. Thus, for 1, in which L1 binds in the
normal P,N mode, the 2-carbon attached to P shifts from
δ 149.0 to δ 149.1 and 148.9, whereas the 2′-carbon
attached to N shifts from δ 152.2 to δ 151.9 and 151.8.
The double-bond character in the C2′-N bond for L2 in
The metrical parameters for the two isomers can be
compared by reference to Table 2. It appears that the
P,C binding mode is maintained in solution. Several

Atropisomeric Phosphino-Amino Palladium Complexes
Organometallics, Vol. 23, No. 9, 2004 2013
2, however, leads to a downfield shift of 5.7 or 3.3 ppm
(a ∆(13C) value of 3.6 ppm was observed in the binaph-
thyl analogue2). The largest effects would be expected
for the carbon atoms bound to palladium and previously
upfield shifts were observed for the binaphthyl ana-
logue,2 where ∆(13C-1′) ) -55 ppm and ∆(13C-6′) ) -3
ppm. The binding is primarily to C-1′ in the binaphthyl
analogue, whereas there is significant binding to C-6′
in 2, so that a smaller ∆(13C-1′) value and a larger ∆(13C-
6′) value would be expected for 2. The observed coordi-
nation shifts for the major and minor isomers of 2 were
-26.6 and -22.6 ppm for the carbon ∆(13C-1′) and -11.6
and -13.6 ppm for ∆(13C-6′). Therefore, we conclude
that the P,C binding mode is maintained in solution.24
We were also able to observe the racemization of an
orthorhombic single crystal using CD (circular dichro-
ism) spectroscopy. A half-life was observed of ap-
proximately 4000 s at 20 °C, which corresponds to a
value of ∆Gq ) 22.5 kcal/mol. The crystallization of the
orthorhombic phase represents a crystallization-induced
asymmetric transformation24 in which one diastereomer
is formed as a conglomerate and each crystal should
contain only one enantiomer, since the space group is
P212121. In principle, this could provide a method of
separating complexes containing the enantiomeric forms
of the ligand by manual sorting of the crystals.26
In conclusion, we have prepared and characterized
two new allyl complexes of palladium which demon-
strate that, when a Pd-N bond is present, the ligand
is hemilabile and there is a low barrier (<10 kcal/mol)
for the Pd-N bond rupture. Nevertheless, a significantly
higher barrier (>18 kcal/mol) is required for atropi-
somerization of the ligand, and diastereomers intercon-
vert via a η3-η1-η3 allyl rearrangement. The preference
for P,N vs P,C binding is controlled by subtle electronic
and steric effects, and P,N bonding is preferred in the
Ph2P case, whereas P,C bonding is preferred in the Cy2P
analogue.
F igu r e 7. 13C NMR chemical shifts (δ) for L2 in CD2Cl2.
J (13C-31P) values are shown in parentheses with accura-
cies of (0.2 Hz.
were recorded on Bruker 400 or Bruker 500 and Omega 300
or Omega 500 NMR instruments, respectively. The 1H VT
NMR experiments were performed on an Omega 300 NMR or
a Bruker 500 instrument. SST (spin saturation transfer) and
EXSY experiments were performed on a Bruker 500 NMR
instrument. Elemental analyses were carried out by Atlantic
Microlab, Inc., Norcross, GA.
Complexes 1 and 2 were prepared by the same procedures.
Typical conditions are as follows for complex 1.
(η3-a llyl)P d (L1)SbF 6 (1). In a flame-dried Schlenk flask,
5 mL of freshly distilled methylene chloride, [(η3-allyl)PdCl]
(11.9 mg, 0.032 mmol), and the ligand L1 (25 mg, 0.065 mmol)
were stirred at room temperature until the solids dissolved
completely (∼1 min). Silver hexafluoroantimonate (22.3 mg,
0.065 mmol) was added, and AgCl precipitated almost im-
mediately. The resulting mixture was stirred for an additional
1
/
h and then filtered through Celite. Yellow crystals of
2
analytical purity were obtained in high yield (85%) by slow
diffusion of pentane into a methylene chloride solution of the
complex. In solution the NMR spectra indicated there was a
53:47 ratio of diastereomers present in CD2Cl2.
(a ) Cr ysta l Isom er R*,SP d * (Ma jor ; F ou n d in a Cr ysta l
by VT NMR Sta r tin g fr om -90 °C). 1H NMR (500 MHz,
CD2Cl2, δ): 7.67-6.35 (18H, complex, aromatics); 5.89 (1H,
m, Hc); 4.37 (1H, br, Hs); 4.33 (1H, dd, Ha, J c-a ) 14.5 Hz, J (P-
Ha) ) 9.0 Hz); 3.32 (1H, dd, Hs’, J c-s′ ) 6.5 Hz, J s-s′ ) 1.5 Hz);
3.27 (6H, s, N(CH3)2); 2.30 (1H, d, Ha′, J c-a′ ) 12.5 Hz). 31P-
{1H} NMR (202 MHz, CDCl3, δ): 24.1 (s). 13C{1H} NMR (125
MHz, CD2Cl2, δ, selected resonances): 135-118 (all aromatic
C that bear an H showed HMQC cross-peaks in the region);
91.2 (1C, d, J PC ) 25 Hz, terminal allylic C trans to P); 55.5
(1C, br, terminal allylic C cis to P).
Exp er im en ta l Section
All reactions were carried out under a nitrogen atmosphere
using standard Schlenk techniques. Methylene chloride was
distilled over CaH2. The ligands L1 and L2 were purchased
from Strem Chemicals. Silver hexafluoroantimonate was
purchased from Aldrich Chemicals. The dimers [(η3-allyl)PdCl]
and [(η3-crotyl)PdCl] were prepared according to a previously
published procedure. Standard 1H and 31P{1H} NMR spectra
1
(b) Non cr ysta l Isom er , R*,RP d * (Min or ). H NMR (500
MHz, CD2Cl2, δ): 7.67-6.44 (18H, complex, aromatics); 5.89
(1H, m, Hc); 4.44 (1H, br, Hs, partially overlapped with Ha);
3.87 (1H, dd, Ha, J c-a ) 14.0 Hz, J (P-Ha) ) 10.5 Hz); 3.19
(6H, s, N(CH3)2); 3.14 (1H, dd, Ha′, J c-a′ ) 12.0 Hz); 2.93 (1H,
d, Hs′, J c-s′ ) 6.5 Hz, J s-s′ ) 2.0 Hz). 31P{1H} NMR (202 MHz,
CDCl3, δ): 22.7 (s). 13C{1H} NMR (125 MHz, CD2Cl2, δ,
selected resonances): 135-118 (all aromatic C that bear an
H showed HMQC cross-peaks in the region); 87.9 (1C, d, J PC
) 25 Hz terminal allylic C trans to P); 60.1 (1C, br, terminal
allylic C cis to P). Anal. Calcd for C29H29NF6PPdSb: C, 45.55;
H, 3.82; N, 1.83. Found: C, 45.63; H, 3.75; N, 1.83.
(η3-a llyl)P d (L2)SbF 6 (2). Assignments were made with the
assistance of J values, HMQC, and COSY. In solution the
NMR spectra indicated there was a 58:42 ratio of diastereo-
mers present at equilibrium in CD2Cl2. Comparisons of 13C
shifts are most easily followed by reference to Figures 7 and
8.
(24) We have assumed that the type of binding found in the solid
persists in solution. The much lower barrier for methyl interconversion
for the NMe2 group in 2 relative to that in 1 also supports this
assumption. There are also (1) shifts of anti protons to lower field than
syn protons and (2) protons cis to phosphine that are to lower field
than those trans to phosphine, which are not normally observed.
(25) Vedejs, E.; Donde, Y. J . Org. Chem. 2000, 65, 2337-2343.
(26) We have found resolutions of this type useful in a number of
cases, but this one is not ideal. The formation as
a triclinic or
orthorhombic phase is random, although it appears that seeding can
induce one form. A more critical problem is that the orthorhombic
crystals have a propensity for forming racemic twins. Thus, a given
crystal will have a nonracemic mixture, but the fraction can vary. For
example, a large orthorhombic crystal was cut and the larger portion
of the crystal was used for a planned CD experiment; however, the
remaining portion which was used for X-ray analysis was a ∼55:45
racemic twin and the crystal showed a very weak dichroism, indicating
a nearly racemic mixture. Testing several large crystals by CD gave
results which indicated that racemic twinning was common; however,
in some crystals the twinning was minimal and the CD could be
determined and used as a diagnostic for racemization rate. We are
still investigating these phenomena.
(a ) R*,SP d *, th e Ma jor Isom er a t Equ ilibr iu m . This is
the isomer found in the orthorhombic crystals. These reso-
nances did not appear after triclinic crystals were dissolved

2014 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
38.0 (1C, d, cyclohexyl P-C-H, J CP ) 24 Hz); 37.0 (1C, d,
cyclohexyl P-C-H, J CP ) 24 Hz). Anal. Calcd for C29H41NF6-
PPdSb: C, 44.84; H, 5.32; N, 1.80. Found: C, 44.59; H, 5.26;
N, 1.93.
Str u ctu r e Deter m in a tion a n d Refin em en t. Crystals
were obtained by slow diffusion of pentane into methylene
chloride solution of complexes 1 and 2. Data were collected on
a Nonius KappaCCD (Mo KR radiation) diffractometer and
were not specifically corrected for absorption, other than the
inherent corrections provided by Scalepack.27 The structures
were solved by direct methods (SIR92)28 and refined on F for
all reflections.29 Non-hydrogen atoms were refined with aniso-
tropic displacement parameters. Hydrogen atoms were in-
cluded at calculated positions. Relevant crystal and data
parameters are presented in Table 1.
Str u ctu r e Deter m in a tion of 1. This structure determi-
nation was straightforward, and the compound had crystal-
lized in a monoclinic cell with Z ) 16, indicating a space group
of C2/c, which requires two molecules in the asymmetric unit.
The molecules are nearly identical, and only one is shown in
Figure 1.
F igu r e 8. 13C NMR chemical shifts (δ) for (η3-allyl)Pd-
(L2)SbF6 (2) in CD2Cl2. J (13C-31P) values are shown in
parentheses with accuracies of (0.2 Hz. The upper shifts
are for the major isomer. Shifts in italics are tentative,
owing to overlap in the HMQC spectra.
in CD2Cl2 at low temperature but were observed after the
solution was warmed to room temperature.
Str u ctu r e Deter m in a tion of Tr iclin ic 2. This structure
determination was straightforward, and the compound had
crystallized in a triclinic cell with Z ) 2, indicating a space
group of P1h.
1H NMR (400 MHz, CD2Cl2, δ): 7.78-6.62 (8H, complex,
aromatics); 5.51 (1H, m, Hc); 4.05 (1H, ddd, Hs′, J c-s′ ) 6.4 Hz,
J
s-s′. J s′-a′ . 2.0 Hz); 3.97 (1H, dd, Ha, J c-a ) 14.0 Hz, J (P-Ha)
) 8.0 Hz); 2.90 (6H, s, N(CH3)2); 2.71 (2H, Hs and Ha′
overlapped). 31P{1H} NMR (122 MHz, CD2Cl2, δ): 46.9 (s). 13C-
{1H} NMR (125 MHz, CD2Cl2, δ, selected resonances): 135-
115 (all aromatic C that bear an H showed HMQC cross-peaks
in the region); 119.6 (1C, d, J PC ) 6 Hz, central allylic C); 100.5
(1C, d, J PC ) 27 Hz, terminal allylic C trans to P); 55.5 (1C, s,
terminal allylic C cis to P); 37.9 (1C, d, cyclohexyl P-C-H, J CP
) 24 Hz); 37.4 (1C, d, cyclohexyl P-C-H, J CP ) 24 Hz).
A difference NOE showed that irradiation of Me-N at δ 2.90
enhanced a doublet at δ 7.11, indicating that it should be
assigned to the 3′-proton. This implied that the remaining
upfield doublet at δ 6.62 was the 6′-proton. COSY data allowed
assignments of the protons in the rings, assuming the largest
J PH value was to H-3. HMQC data then allowed the carbon
assignments shown in Figure 8. 13C data for carbon atoms not
bearing a proton were relatively weak, particularly the reso-
nance at δ 155.4, but positions were verified with HMQC data
optimized for long-range couplings.
(b) Min or Isom er (F ou n d in a Tr iclin ic Cr ysta l by VT
NMR Sta r tin g fr om -90 °C). 1H NMR (400 MHz, CD2Cl2,
δ): 7.78-6.79 (8H, complex, aromatics); 5.51 (1H, m, Hc); 3.91
(1H, ddd, Hs′, J c-s′ ) 6.8 Hz, J s-s′. ≈ J s′-a′ ≈ 2.2 Hz); 3.56 (1H,
dd, Ha, J c-a ) 13.6 Hz, J (P-Ha) ) 8.8 Hz); 3.22 (1H, ddd, Hs,
J c-s ≈ J (P-Hs) ≈ 8.4 Hz, J s-s′ ≈ 2.2 Hz); 2.93 (1H, d (br), Ha′,
J c-a′ ) 12.8 Hz); 2.81 (6H, s, N(CH3)2). 31P{1H} NMR (122 MHz,
CD2Cl2, δ): 43.1 (s). 13C{1H} NMR (125 MHz, CD2Cl2, δ,
selected resonances): 135-115 (all aromatic C that bear an
H showed HMQC cross-peaks in the region); 119.9 (1C, d, J PC
) 6 Hz, central allylic C); 96.4 (1C, d, J PC ) 27 Hz, terminal
allylic C trans to P); 53.5 (1C, s, terminal allylic C cis to P);
Different crystallization vials tended to contain an excess
of either triclinic or orthorhombic crystals, suggesting a
seeding phenomenon. SbF6 showed a 2:1 disorder, which was
modeled as a rotation about one F-Sb-F axis of ∼45°.
Str u ctu r e Deter m in a tion of Or th or h om bic 2. The data
showed an orthorhombic cell with absences consistent with
the space group P212121. The Flack parameter was ambiguous,
and the opposite hands were refined with only a minor
difference in R factors (R, Rw, GOF of 4.08, 4.61, 2.31 vs 3.99,
4.43, 2.22). Refinement with SHELX suggested a racemic twin
with a ratio of ∼45:55. The enantiomer shown in the ORTEP
diagram gave the lowest R factor on refinement on F.
Ack n ow led gm en t. This work was supported by the
National Science Foundation (Grant No. CHE0092222).
Su p p or tin g In for m a tion Ava ila ble: Listings of atomic
positions, thermal parameters, and bond distances and angles
for 1 and 2. This material is available free of charge via the
Internet at http://pubs.acs.org.
OM030681H
(27) HKL2000 (Denzo-SMN) software package: Minor, W.; Otwi-
nowski, Z. Processing of X-ray Diffraction Data Collected in Oscillation
Mode. In Methods in Enzymology; Carter, C. W., J r., Sweet, R. M.,
Eds.; Academic Press: New York, 1997; Vol. 276A (Macromolecular
Crystallography), pp 307-326.
(28) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J .
Appl. Crystallogr. 1993, 26, 343.
(29) TEXSAN for Windows version 1.06: Crystal Structure Analysis
Package; Molecular Structure Corp., The Woodlands, TX, 1997-1999.