Kinetics and mechanism of regioselective amination of the 1-phenylallyl group in cationic palladium(II) complexes bearing bidentate ligands

Article abstract of DOI:10.1016/S0020-1693(01)00333-4

The complexes [Pd(η³-1-PhC₃H₄) (L-L′)]⁺ [L-L′ = 2-(PPh₂)C₆H₄-1-CH = NR (R = Me (1a), i-Pr (1b), t-Bu (1c), (R)-bornyl (1d), C₆H₄OMe-4 (1e), C₆H₃Me₂-2,6 (1f), C₆H₃(i-Pr)₂-2,6 (1g)), 6-MeC₅H₃N-2-CH = NC₆H₄OMe-4 (2a), C₅H₄N-2-CH = N-t-Bu (2b) and C₅H₄N-2-CH₂S-t-Bu (3a)] are generally present in solution as two geometrical isomers, the relative abundance of which depends essentially on the steric requirements of the L-L′ ligand. In the presence of fumaronitrile the cationic complexes undergo a regioselective amination by secondary amines HY at the CH₂ allyl terminus, yielding [Pd(η²-fn)(L-L′)] and the allylamines (E)-PhCH = CHCH₂Y. Under pseudo-first-order conditions the amination rates (k_obs) are found to depend on the k₂[HY] term for 2a and 3a, and on the sum k₂[HY]+k₃[HY]² for the other complexes. The second-order term k₂ is related to direct nucleophilic attack on the CH₂ allyl terminus of the substrate whereas the third-order term k₃ is ascribed to parallel attack by a further HY molecule on the intermediate [Pd(1-PhC₃H₄)(L-L′)(HY)]⁺. The k₂ values depend on the steric and electronic properties of both the amine HY and the ligand L-L′. For complexes 1a-1g, the relatively higher k₂ values and their increase with increasing steric crowding at the nitrogen-bonded carbon of substituent R are interpreted in terms of a greater reactivity of the isomer with the CH₂ allyl terminus trans to phosphorus and cis to the NR group. The high amination rate of 2a, as compared with that of 2b, is related to substantial steric interaction of the CH₂ allyl terminus with the 6-Me pyridine group in close proximity in the predominant isomer.

Full text of DOI:10.1016/S0020-1693(01)00333-4

www.elsevier.nl/locate/ica
Inorganica Chimica Acta 315 (2001) 172–182
Kinetics and mechanism of regioselective amination of the
1-phenylallyl group in cationic palladium(II) complexes bearing
bidentate ligands
Bruno Crociani a,*, Simonetta Antonaroli a, Luciano Canovese b, Fabiano Visentin b,
Paolo Uguagliati b
a Dipartimento di Scienze e Tecnologie Chimiche, Uni6ersita’ di Roma ‘‘Tor Vergata’’, Via della Ricerca Scientifica, 00133 Rome, Italy
b Dipartimento di Chimica, Uni6ersita’ di Venezia, Venice, Italy
Received 6 October 2000; accepted 17 January 2001
Abstract
The complexes [Pd(h3-1-PhC3H4)(L–L%)]+ [L–L%=2-(PPh2)C6H4-1-CHꢀNR (R=Me (1a), i-Pr (1b), t-Bu (1c), (R)-bornyl (1d),
C6H4OMe-4 (1e), C6H3Me2-2,6 (1f), C6H3(i-Pr)2-2,6 (1g)), 6-MeC5H3N-2-CHꢀNC6H4OMe-4 (2a), C5H4N-2-CHꢀN-t-Bu (2b) and
C5H4N-2-CH2S-t-Bu (3a)] are generally present in solution as two geometrical isomers, the relative abundance of which depends
essentially on the steric requirements of the L–L% ligand. In the presence of fumaronitrile the cationic complexes undergo a
regioselective amination by secondary amines HY at the CH2 allyl terminus, yielding [Pd(h2-fn)(L–L%)] and the allylamines
(E)-PhCHꢀCHCH2Y. Under pseudo-first-order conditions the amination rates (kobs) are found to depend on the k2[HY] term for
2a and 3a, and on the sum k2[HY]+k3[HY]2 for the other complexes. The second-order term k2 is related to direct nucleophilic
attack on the CH2 allyl terminus of the substrate whereas the third-order term k3 is ascribed to parallel attack by a further HY
molecule on the intermediate [Pd(1-PhC3H4)(L–L%)(HY)]+. The k2 values depend on the steric and electronic properties of both
the amine HY and the ligand L–L%. For complexes 1a–1g, the relatively higher k2 values and their increase with increasing steric
crowding at the nitrogen-bonded carbon of substituent R are interpreted in terms of a greater reactivity of the isomer with the
CH2 allyl terminus trans to phosphorus and cis to the NR group. The high amination rate of 2a, as compared with that of 2b,
is related to substantial steric interaction of the CH2 allyl terminus with the 6-Me pyridine group in close proximity in the
predominant isomer. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Kinetics and mechanism; Amination; Palladium complexes; Allyl complexes
nucleophilic attack by the amine on the proposed
cationic intermediate [Pd(h3-allyl)(L–L%)]+ are not
completely cleared.
We have been carrying out a systematic mechanistic
investigation of allyl amination in the cationic com-
plexes [Pd(h3-allyl)(L–L%)]+, where L–L% is an N–N%
bidentate a-diimine [1,2-bis(imino)ethane or 2-(imi-
nomethyl)pyridine] [4] or an N–S bidentate 2-
1. Introduction
The palladium-catalysed amination of allyl substrates
is an attractive synthetic methodology [1] which has
been the subject of several studies in recent years since
the reaction can be carried out enantioselectively in the
presence of chiral ligands [2]. In particular, high enan-
tioselectivities (ee up to 97–99%) have been obtained
with P,N donor bidentate ligands [3]. Despite all ef-
forts, however, the mechanism details and the role of
steric and electronic factors in controlling the site of
(thiomethyl)pyridine [5] or
a
P–N bidentate
(2-(diphenylphosphino)-benzylidene)amine [6]. The
pseudo-first-order rate constants kobs are found to de-
pend on the term k2[amine] and, in some cases, on the
sum of the terms k2[amine]+k3[amine]2. For the N–N%
and N–S derivatives the allyl amination is accompanied
by a fast equilibrium displacement of the bidentate
* Corresponding author. Tel.: +39-06-725 94337; fax: +39-06-
725 94328.
E-mail address: crociani@stc.uniroma2.it (B. Crociani).
0020-1693/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(01)00333-4

B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
173
ligand producing an inert bis-amino species [Pd(h3-al-
2.1. Preparation of the ligands
2-(Ph2P)C6H4-1-CHꢀNR (R=C6H3Me2-2,6,
C6H3(i-Pr)2-2,6)
lyl)(amine)2]+. In the case of asymmetrically substituted
allyl ligands, such as the 1,1-dimethylallyl group, the
mechanistic pattern is further complicated by the pres-
ence of geometrical isomers in solution, and by the
regiochemistry of the nucleophilic attack which can
occur at both the terminal allyl carbons [5,6].
The 2-(diphenylphospino)benzaldehyde (2 mmol) and
the appropriate amine RNH2 (4 mmol) were dissolved
in 60 cm3 of a MeOH/CH2Cl2 mixture (3:1 v/v) under
nitrogen. When the condensation was complete (4 days
for R=C6H3Me2-2,6; 6 days for R=C6H3(i-Pr)22,6, at
room temperature), the reaction mixture was worked
up as described for the preparation of the other
iminophosphines [7]. The yields were in the range 87–
In order to obtain more information about the fac-
tors which govern the amination rate and regiochem-
istry, we have extended our kinetic studies to the
reactions of a variety of cationic substrates [Pd(h3-al-
lyl)(L–L%)]+ (L–L%=P–N, N–N% and N–S bidentate
ligand; allyl=1-PhC3H4) with an allylic group bearing
an electron-withdrawing phenyl substituent.
1
90% (based on the theoretical amount). H NMR spec-
tra in CH2Cl2-d2: R=C6H3Me2-2,6, l 8.86 [d, 1H,
J(PH)=5.6 Hz, NꢀCH], 1.85 [s, 6H, 2CH3]; R=
C6H3(i-Pr)2-2,6, l 8.88 [d, 1H, J(PH)=5.7 Hz, NꢀCH],
2.74 [spt, 2H, J(HH)=6.9 Hz, 2CH], 1.00 [d, 12H,
4CH3].
2. Experimental
The iminophosphines [7], 2-(iminomethyl)pyridines
[8], 2-(thiomethyl)pyridines [5b] and dimer [Pd(m-
Cl)(h3-1-PhC3H4)]2 [9] were prepared by published
methods. Piperidine, diethylamine and morpholine were
distilled over anhydrous K2CO3 under nitrogen. All the
other chemicals and solvents were commercial grade,
and were purified or dried by standard methods when
required [10]. The solvents were evaporated to small
volume or to dryness at reduced pressure in a rotary
evaporator. All the new ligands and complexes were
characterized by elemental analysis, IR spectra in the
solid and NMR spectra. For brevity, the analytical and
IR data of some representative derivatives for each
class of related compounds are listed in Table 1.
2.2. Preparation of complexes
[Pd(p3-1-PhC3H4)(L–L%)]BF4
The complexes were prepared from the reaction of
[Pd(m-Cl)(h3-1-PhC3H4)]2 (0.25 mmol) with the appro-
priate ligand L–L% (0.5 mmol) in CH2Cl2 (20 cm3),
followed by addition of an excess of NaBF4 (2 mmol)
dissolved in 10 cm3 of MeOH. The reaction mixture
was worked up as previously described for related
derivatives [5,7,11], to give the required products which
were purified by reprecipitation from a CH2Cl2/Et2O
solvent mixture (yields in the range 77–95%). The
complexes were characterized by elemental analysis and
Table 1
Analytical a and IR b data
Compound
Analysis (%)
C
w(CꢀN) (cm−1
)
w(BF) (cm−1
)
H
N
2-(PPh2)C6H4-1-CHꢀNC6H3Me2-2,6
2-(PPh2)C6H4-1-CHꢀNC6H3(i-Pr)2-2,6
81.9 (82.42)
82.5 (82.82)
6.2 (6.15)
7.2 (7.18)
3.6 (3.56)
3.1 (3.12)
1631 ms
1626 ms
[Pd(h3-1-PhC3H4)(L–L%)]BF4
(1b)BF4
(1d)BF4
(1f)BF4
(2a)BF4
(3a)BF4
57.8 (58.01)
62.4 (62.66)
61.1 (61.43)
51.3 (51.47)
46.5 (46.41)
4.9 (4.87)
5.5 (5.61)
4.8 (4.73)
4.4 (4.32)
5.0 (4.92)
2.2 (2.18)
2.0 (1.90)
2.0 (1.99)
5.3 (5.22)
2.8 (2.85)
1624 m
1610 m
1609 m
1620 mw
1075 vs
1053 vs
1055 vs
1070 vs
1061 vs
[Pd(h2-fn)(L–L%)]
L–L%=2-(PPh2)C6H4-1-CHꢀNC6H3(i-Pr)2-2,6
L–L%=6-MeC5H3N-2-CHꢀNC6H4OMe-4
66.5 (66.30)
54.3 (54.49)
5.5 (5.41)
4.1 (4.07)
6.6 (6.63)
10.8 (10.59)
1616 m, 2191 s c
1614 mw, 2195 s c
a Required values are given in parentheses.
b As Nujol mulls.
c w(CꢁN) band.

174
B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
1
IR spectra (Table 1), and by H and 13C NMR spectra
2.5. Kinetic and displacement equilibria measurements
(Table 2).
For the complexes 1a(BF4)–1g(BF4) the kinetics of
allyl amination were studied by adding known aliquots
of amine solution to a solution of the complex and
fumaronitrile in the thermostatted cell compartment of
a Perkin–Elmer Lambda 40 spectrophotometer (25°C).
The reactants were such as to ensure constant excess
over the metal substrate. For the complexes 2a(BF4),
2b(BF4) and 3a(BF4) the initial solution of complex
([Pd]0=1×10−4 mol dm−3) and fumaronitrile ([fn]0=
2–4×10−4 mol dm−3) also contained an excess of the
corresponding free ligand L–L% ([L–L%]0=1×10−3
mol dm−3). The equilibrium constants KE for the
cationic complexes 2b and 3a were determined by spec-
trophotometric titration, as previously described for
analogous displacement reactions involving 2-(imi-
nomethyl)pyridines [4b] or 2-(thiomethyl)pyridines [5b].
Mathematical and statistical data analysis was carried
out on a personal computer by means of a locally
adapted version of Marquardt’s algorithm [13] written
in TURBOBASIC (Borland).
2.3. Preparation of complexes [Pd(p2-fn)(L–L%)]
An excess of diethylamine (1.5 mmol) was added to a
solution of [Pd(h3-1-PhC3H4)(L–L%)]BF4 (0.3 mmol)
and fumaronitrile (fn) (0.4 mmol) in CH2Cl2 (20 cm3).
After stirring for 2 h at ambient temperature the result-
ing solution was worked up as described for the prepa-
ration of analogous palladium(0) derivatives [7] (yields
in the range 70–93%). The products were characterized
by elemental analysis and IR spectra (Table 1), and by
1H NMR spectra in CH2Cl2-d2; [Pd(h2-fn){2-
(Ph2P)C6H4-1-CHꢀNR}]: R=i-Pr, l 8.34 [d, 1H,
J(PH)=3.3 Hz, NꢀCH], 3.68 [m, 1H, CH], 3.12 [dd,
1H, J(PH)=3.4 Hz, J(HH)=9.4 Hz, CH (fn) trans to
N], 2.75 [dd, 1H, J(PH)=J(HH)=9.4 Hz, CH (fn)
trans to P], 1.35 [d, 3H, J(HH)=6.3 Hz, CH3], 1.28 [d,
3H, J(HH)=6.8 Hz, CH3]; R=C6H3Me2-2,6, l 8.16
[d, 1H, J(PH)=3.2 Hz, NꢀCH], 3.20 [dd, 1H, J(PH)=
3.4 Hz, J(HH) 9.4 Hz, CH (fn) trans to N], 2.19 [dd,
1H, J(PH)=J(HH)=9.4 Hz, CH (fn) trans to P], 2.07
[s, 3H, CH3], 2.04 [s, 3H, CH3]; R=C6H3(i-Pr)2-2,6, l
8.15 [d, 1H, J(PH)=2.9 Hz, NꢀCH], 3.23 [m, 1H, CH],
2.75 [m, 1H, CH], 3.21 [dd, 1H, J(PH)=3.7 Hz,
J(HH)=9.5 Hz, CH (fn) trans to N], 2.26 [dd, 1H,
J(PH)=J(HH)=9.5 Hz, CH (fn) trans to P], 1.24 [d,
3H, J(HH)=6.9 Hz, CH3], 1.14 [d, 3H, J(HH)=6.6
Hz, CH3], 0.92 [d, 3H, J(HH)=6.9 Hz, CH3], 0.77 [d,
3H, J(HH)=6.6 Hz, CH3]; [Pd(h2-fn)(6-MeC5H3N-2-
CHꢀNC6H4OMe-4)], l 8.63 [s, 1H, NꢀCH], 3.89 [s, 3H,
OCH3], 2.93 [s, 3H, CH3], 3.03 [d, 1H, J(HH)=9.2 Hz,
CH (fn)], 2.89 [d, 1H, J(HH) 9.2 Hz, CH (fn)].
3. Results and discussion
3.1. Isomer distribution and dynamic beha6iour in
solution
Owing to the asymmetric structure of the allyl and
L–L% ligands, the complexes [Pd(h3-1-PhC3H4)(L–L%)]+
may be present in solution with two geometrical iso-
mers I and II, as shown in Scheme 1.
The occurrence of both the isomers I and II in most
1
cases is indicated by the H and 13C{1H} NMR spectra
2.4. NMR measurements
(Table 2).
The magnitude of J(H2ꢂH1a) coupling constants
(14.0–11.5 Hz) suggests a syn position for the 1-Ph
substituent of the allyl group in all the cationic com-
plexes under study, independently of the isomeric struc-
ture I or II and of the nature of the bidentate ligand
L–L%. Only for the two minor diastereoisomers of
structure II of complex 1d, the allyl configuration can-
not be assigned because the H1a resonances are detected
as overlapping and unresolved multiplets.
1
The H and 13C{1H} NMR spectra were recorded on
a Bruker AM 400 spectrometer, operating at 400.13
and 100.61 MHz, respectively. Chemical shifts (ppm)
are given relative to SiMe4. The H 2D ROESY spectra
1
of 1c(BF4) in CHCl3-d and CH2Cl2-d2 were obtained in
the phase-sensitive mode using the TPPI phase cycle
with the ROESY pulse sequence, modified to eliminate
the offset dependence of cross peak intensity [12]. A
total of 128 transients on a size of 2K were accumu-
lated in the phase-sensitive mode for 512 experiments.
A spin-lock period corresponding to a transverse mix-
ing time of 0.3 s was applied. Data were processed with
the 2D NMR program TRITON, licensed by the Univer-
sity of Utrecht, Utrecht, The Netherlands. Elaboration
of the spectrum was carried out on a Digital Graphic
workstation. A sine bell apodization function, shifted
by p/3, was applied in both dimensions. A zero filling
was applied in t1 in order to obtain a real 1K×1K
matrix.
3.2. Complexes [Pd(p3-1-PhC3H4)(P–N)]+ (1a–1g)
In these compounds the I/II isomer ratio generally
decreases with increasing bulkiness of the imino-nitro-
gen substituent R. When R is an alkyl group, the ratio
decreases from 7.3:1 for 1a (R=Me) to 3.3:1 for 1c
(R=t-Bu), with a noticeable exception for 1d [R=(R)-
bornyl] where a 15.7:1 ratio is observed. When R is an
aryl group, the ratio decreases from 13.3:1 for 1e
(R=C6H4OMe-4) to 0.7:1 for 1g [R=C6H3(i-Pr)2-2,6].

B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
175
Table 2
Selected 1H a and 13C{1H} b NMR data for the complexes [Pd(h3-1-PhC3H4)(L–L%)]BF4
c
Complex
Isomer
Allyl protons
H1a
Allyl carbons
C1
H2
H3s
H3a
C2
C3
1a(BF4)
I (88%)
5.56 dd
(13.4) d
(10.0) e
4.40 d
6.34 dt
3.13 d
(8.7) d
102.4 d
(26.4)
115.0 d
(6.8)
53.4 d
(3.9)
II (12%)
6.49 m
5.05 dd
(7.0) d
(7.0) e
4.14 dd
(13.9) d
(9.6) e
n.o.
n.o.
n.o.
(12.8) d
1b(BF4)
1c(BF4)
1d(BF4) g
1e(BF4)
1f(BF4)
I (85%)
5.51 dd
(13.3) d
(10.2) e
4.27 d
6.35 dt
(9.5) d
3.14 br s
102.8 d
(26.4)
114.5 d
(7.2)
52.2 d
(4.5)
II (15%)
mk
5.11 dd
4.1 f
73.7 d
(4.1)
119.9 d
(6.3)
80.7 d
(26.9)
(12.5) d
(7.0) d
(7.0) e
I (77%)
5.61 dd
(13.8) d
(10.6) e
4.23 d
6.25 dt
6.15 m
3.20 br
2.95 br
106.6 d
(26.5)
113.0 d
(7.6)
50.4 d
(5.3)
II (23%)
5.26 dd
(7.5) d
(7.5) e
4.10 dd
(15.0) d
(11.0) e
72.1 d
(6.9)
120.5 d
(5.6)
81.6 d
(25.9)
(14.0) d
I (94%) h
II (6%) j
5.71 dd, 4.98 dd
(13.8) d, (11.5) d
(9.3) e, (11.5) e
4.60 m i
6.25 m
mk
3.30 m i
3.20 d, 2.64 d
(13.5) d, (12.5) d
4.81 dd
(6.7) d
(6.7) e
4.22 dd, 4.16 dd
(14.5) d, (15.0) d
(9.2) e, (10.0) e
I (93%)
II (7%)
5.29 dd
(13.5) d
(10.0) e
4.43 d
6.20 dt
(9.5) d
3.26 br s
104.8 d
(26.3)
114.7 d
(6.8)
53.8 d
(4.3)
6.40 m
4.03 dd
(7.3) d
(7.3) e
3.99 dd
(14.2) d
(9.3) e
n.o.
n.o.
n.o.
(11.7) d
I (88%)
5.06 dd
(13.6) d
(10.1) e
4.52 d
6.17 m
6.32 m
3.41 d
(6.0) d
3.03 d
104.3 d
(26.3)
114.5 d
(6.9)
54.5 d
(4.1)
(12.0) d
II (12%)
3.56 dd
(7.1) d
(7.1) e
3.93 dd
(14.0) d
(9.4) e
75.1 d
(5.6)
118.9 d
(5.6)
83.6 d
(25.8)
(11.6) d
1g(BF4) g
I (42%)
4.91 dd
(13.0) d
(11.0) e
4.34 d
6.20 m
6.25 m
3.22 d
(6.0) d
3.04 d
106.8 d
(26.4)
113.9 d
(6.8)
53.1 d
(4.1)
(11.7) d
II (58%)
3.59 dd
(6.9) d
(6.9) e
3.94 dd
(14.0) d
(9.1) e
74.7 d
(5.9)
119.0 d
(5.6)
84.9 d
(25.3)
(11.7) d
2a(BF4)
2b(BF4)
3a(BF4) g
I (\95%)
II (\95%)
4.90 d
6.03 m
6.15 m
4.50 d
(6.2) d
3.57 d
83.8
81.8
112.4
111.3
60.8
59.9
(11.9) d
(12.5) d
4.86 d
4.42 d
(7.1) d
3.42 d
(11.9) d
(13.0) d
I (12%) k
II (88%) l
4.82 d
6.35 m
6.25 m
4.63 d
(7.5) d
mk m
3.96 d
(12.5) d
(13.5) d
5.20 d, 5.17 d
3.31 d, 3.15 d
(13.2) d, (11.8) d
(11.8) d, (11.8) d
a In CH2Cl2-d2 at 28°C unless otherwise stated. Satisfactory integration values are obtained; coupling constants in Hz. Abbreviations: s, singlet;
d, doublet; t, triplet; dd, doublet of doublets; m, multiplet; mk, masked; n.o., not observed.
b In CHCl3-d at 28°C; J(PC) coupling constants (Hz) in parentheses.
c See Scheme 1 for isomer structure and numbering scheme; the percentages of isomer are given in parentheses.
d J(HH).
e J(PH).
f Overlapping with the CH signal of the i-Pr group.
g At −70°C.
h Two diastereoisomers in 1.3:1 molar ratio.
i Unresolved multiplet.
j Two diastereoisomers in 1.6:1 molar ratio.
k Two isomers in 2:1 molar ratio.
l Two isomers in 1.3:1 molar ratio.
m Masked by signals of the CH2 group in the range 4.1–4.5 ppm.

176
B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
In the latter case, the large steric requirements of the
substituent R destabilize isomer I in favour of II. Like
the analogues [Pd(h3-allyl)(P–N)]+ (allyl=C3H5 and
1,1-Me2C3H3) [6,14], the complexes 1a–1g undergo var-
ious dynamic processes in solution, involving either
inversion of the P–N chelate ring relative to the
PꢂPdꢂN coordination plane or h3–h1–h3 rearrange-
ment of the allyl ligand or interconversion of the iso-
mers I and II through the P–N ligand site exchange.
The P–N ring inversion brings about the intercon-
version of conformers A and A% (see Fig. 1), which
result from the fact that the six-membered chelating
unit of the iminophosphine is not coplanar with the
PꢂPdꢂN coordination plane [14].
In spite of the larger size of the 1-phenyl allyl group
and the presence of bulky imino-nitrogen substituents
(e.g. in 1c, 1d and 1g), such a dynamic process is very
fast (on the NMR time scale) and cannot be frozen
even at the lowest explored temperature (−70°C).
Thus, the observed NMR spectra are time averaged
between those of the two conformers A and A% for each
geometrical isomer I and II. In this context, it is worth
noting that: (i) rotation around the NꢂR bond is hin-
dered for the C6H3Me2-2,6 and C6H3(i-Pr)2-2,6 sub-
Scheme 1. [(R)-Bornyl=endo-(1R)-1,7,7-trimethylbiciclo[2.2.1]hept-2-yl].

B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
177
Fig. 1. Side view from the allyl ligand toward the Pd atom of the conformers A and A% for each geometrical isomer I or II. The curved line
represents the ꢀCHC6H4ꢂ unit lying above or below the NꢂPdꢂP plane.
The mutual interconversion of isomers I and II of
each complex 1a–1g is rather slow on the NMR time
scale. For complex 1c such interconversion is suggested
by the observation of weak exchange cross peaks be-
tween the NꢀCH and t-Bu signals of isomer I, and the
corresponding signals of isomer II in the phase-sensitive
2D 1H NMR ROESY spectra in CH2Cl2-d2 and
CHCl3-d.
stituents in 1f and 1g, respectively, as suggested by the
multiplicity of the Me and i-Pr proton resonances (e.g.
two Me singlets are observed for each isomer of 1f: at
1.49 and 1.97 ppm for I and at 1.97 and 2.22 ppm for
II); and (ii) the inversion rate is greatly reduced in the
corresponding complexes [Pd(h3-1,3-Ph2C3H3)(P–N)]+
containing the much bulkier 1,3-diphenylallyl ligand
[15]. Similar conformational changes involving chelate
ring inversion have been reported for the d8 metal ion
complexes with P,N- and P,S-bidentate ligands [16,17].
The h3–h1–h3 rearrangement occurs at an apprecia-
ble rate only for isomer I and brings about a selective
syn–anti exchange of the H3s and H3a protons (cis to
phosphorus) through breaking of the PdꢂC1 bond
(trans to phosphorus), rotation around the C2ꢂC3 bond
and re-formation of the h3 complex, as proposed for
several h3-allylpalladium(II) derivatives with various
bidentate ligands containing a P-donor group [18]. As a
3.3. Complexes [Pd(p3-1-PhC3H4)(N–N%)]+
In solution complex 2a is present almost exclusively
as the isomer I and complex 2b as the isomer II. In
1
both cases, the H NMR spectra show the presence of
trace amounts of a second compound, presumably the
isomer II for 2a and the isomer I for 2b, the structure
of which cannot be rigorously determined for the weak-
ness of the observed signals. The structural assignment
of the predominant isomer is essentially based on chem-
ical shift changes caused by the phenyl ring currents of
h3-bound 1-phenylallyl ligand. Accordingly, for 2a the
cis position of the allyl Ph group relative to the imine
CHꢀNC6H4OMe-4 unit is suggested by the high field
shift of the ortho and meta protons signals of the
C6H4OMe-4 substituent at 6.79 and 6.55 ppm, respec-
tively [cf. the value of approximately 7.5 ppm for the
ortho protons resonance found in [Pd(h3-2-
1
consequence, in the H NMR spectra at the ambient
temperature the H3s and H3a protons of isomers I are
generally detected as two broad signals, which in some
cases coalesce into a broad resonance or a time-aver-
aged doublet (as for complex 1a). The fine structure of
the signals is observed at lower temperature (cf. the
spectra of 1g at −70°C). The rate of the h3–h1–h3
process increases in the presence of coordinating species
such as amines (NEt3) or trace amounts of Cl− ions in
CHCl3-d. For complex 1d containing the chiral (R)-
MeC3H4)(N–N%)]+
(N–N%=C5H4N-2-CHꢀNC6H4-
OMe-4) [11]]. For 2b the cis position of the allyl Ph
group relative to the pyridine nitrogen is clearly indi-
cated by the remarkable upfield shift of the pyridine
6-H proton resonance at 6.73 ppm [cf. the value of 8.78
ppm for the same signal in [Pd(h3-2-MeC3H4)(N–N%)]+
(N–N%=C5H4N-2-CHꢀN-t-Bu) [19]]. In contrast to the
dynamic behaviour of the related complexes [Pd(h3-1-
MeC3H4)(N–N%)]+ [19], the I ? II isomer interconver-
sion for 2a and 2b, if it occurs, is slow at ambient
1
bornyl substituent the H NMR spectrum at −70°C
shows the presence of two diastereoisomers for each
geometrical isomer I and II. At ambient temperature,
however, the diastereoisomers of structure I intercon-
vert rapidly into each other (as indicated by the large
broadening of the corresponding signals) for the con-
comitant occurrence of the h3–h1–h3 rearrangement of
the allyl ligand and the chelate ring inversion of the
P–N ligand.
1
temperature. On the other hand, the H NMR spectra

178
B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
show no evidence for any h3–h1–h3 rearrangement of
tectable on the NMR time scale. However, preliminary
results on the reactions of the cationic complexes
[Pd(h3-1,3-Ph2C3H3)(L–L%)]+ with secondary amines
under comparable experimental conditions indicate that
the nucleophilic attack at the phenyl-substituted allyl
terminus does occur, but at such a lower rate that its
contribution to Reaction 1 can be confidently ne-
glected. On the other hand, the regioselectivity does not
appear to be affected by the nature and concentration
of the activated olefin, as shown by experiments carried
out in the presence of a less-accepting olefin, such as
dimethyl fumarate, and at different olefin concentra-
tions. A regioselective, palladium-catalysed alkylation
of 1-arylprop-2-enyl acetates with sodium enolates as
soft carbon nucleophiles in the presence of tertiary
phosphines has been reported to be controlled by cata-
lytic amounts of lithium iodide to give linear olefins
(E)-ArCHꢀCHCH2Nu [20]. Furthermore, a recent pa-
per deals with the kinetics and mechanism of the reac-
tions of [Pd(h3-1-PhC3H4)(L)2]+ (L=PPh3, P(OPh)3)
with uncharged nucleophiles which yield only the linear
olefins (E)-PhCHꢀCHCH2Nu [21].
For kinetic measurements the reactions of the
iminophosphine complexes 1a–1g are followed by mon-
itoring the UV–Vis spectral changes (u=500–200 nm)
of CHCl3 solutions of the complexes ([Pd]0 1×10−4
mol dm−3) in the presence of fn ([fn]0 2–8×10−4
mol dm–3) upon addition of variable aliquots of HY
([HY]0 1×10−3–0.08 mol dm−3) at 25°C. Under such
pseudo-first-order conditions the reactions go to com-
pletion as indicated by comparison of the solution
spectra after 7–8 half-lives with those of the final
products [Pd(h2-fn)(P–N)] independently prepared [7].
The conversion to zerovalent complexes appears
to obey the customary monoexponential absorbance
(A) vs. time (t) relationship At=Aꢀ+(A0−Aꢀ)×
exp(−kobst). The pseudo-first-order rate constants
kobs are determined by non-linear regression of absor-
bance At data to time. No dependence of the rates
on the fn concentration can be detected in the range in-
vestigated. The kobs values fit the two-term second and
third-order rate law (Fig. 2)
the 1-phenylallyl ligand.
3.4. Complex [Pd(p3-1-PhC3H4)(N–S)]+ (3a)
1
The H NMR spectrum of 3a at −70°C shows the
presence of four isomeric species: two of structure II in
approximately 1.3:1 molar ratio and two of structure I
in approximately 2:1 molar ratio. As previously re-
ported for related complexes [Pd(h3-allyl)(N–S)]+ (al-
lyl=C3H5, 1-MeC3H4, 2-MeC3H4, 1,1-Me2C3H3;
N–S=2-(thiomethyl)pyridine) [5], the occurrence of
two isomers for each structure I and II of 3a results
from the two opposite orientations of the S-t-Bu sub-
stituent relative to the h3-bound allyl ligand. The iso-
mers II are more abundant (approximately 88% of the
total isomeric population): in both of them the cis
position of the allyl Ph group relative to the pyridine
nitrogen is demonstrated by the large shielding of the
pyridine 6-H proton caused by the phenyl ring currents
(cf. the 6-H resonances at 8.72 and 8.67 ppm for
isomers I, and those at 7.23 and 6.68 ppm for isomers
II). At ambient temperature the 1H NMR spectrum
exhibits broad, time-averaged signals due to the simul-
taneous occurrence of dynamic processes involving a
fast inversion at sulphur (which interconverts the two
isomers for each structure I and II) and a slower N–S
ligand site exchange (which interconverts the structures
I and II) [5].
3.5. Kinetics
The cationic complexes [Pd(h3-1-PhC3H4)(L–L%)]+
react smoothly with an excess of secondary amine HY
(HY=piperidine, diethylamine, morpholine) in CHCl3
(or CHCl3-d) in the presence of fumaronitrile to give
the zerovalent complexes [Pd(h2-fn)(L–L%)] [5b,7] and
the corresponding allylamines, according to Eq. (1):
+fn
[Pd(h3-1-PhC3H4)(L−L%)]++2HYꢂꢂꢂꢂꢂꢂꢃ
+
−H Y
2
[Pd(h2-fn)(L−L%)]+(E)-PhCH=CHCH2Y
(1)
1
According to H NMR spectra recorded at different
times for complex/fn/HY reaction mixtures in a 1:1.2:5
molar ratio in CHCl3-d at 25 and −30°C, the nucle-
ophilic attack occurs at the less-substituted terminus of
the allyl group independently of the nature of the L–L%
bidentate ligand. The CHꢀCHCH2 moiety of the ally-
lamines is detected as an ABX2 spin system, with J(AB)
coupling constants between the olefinic protons in the
range 15.1–15.9 Hz, indicating a trans geometry
around the CꢀC bond. In principle, the nucleophilic
attack may also take place at the more substituted allyl
carbon, as observed for the amination of analogous
1,1-Me2C3H3 derivatives [6]. In this case, the primary
product should undergo a fast isomerization unde-
k
obs=k2[HY]+k3[HY]2
(2)
The values of constants k2 and k3 are listed in Table 3.
The third-order k3 term is unlikely to originate from
a bimolecular attack by a hydrogen-bonded amine
dimer in equilibrium with the monomer under the
prevailing solvent, temperature and concentration con-
ditions, as we have shown in previous studies [6].
We therefore propose the mechanism of Scheme 2, in
which an overall representation of the allyl amina-
tion of complexes containing labile (a-diimines,
2-(thiomethyl)pyridines) and hemilabile (iminophos-
phines) ancillary ligand is depicted.

B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
179
Scheme 2.
In the case of L–L%=iminophosphine, the starting
cationic metal substrate undergoes nucleophilic attack
by the entering amine at the allyl moiety (step k2) or at
the central metal (fast equilibrium K1), forming the
intermediate [Pd(1-PhC3H4)(P–N)(HY)]+ in which the
allyl group is attacked by one further amine molecule
As can be seen in Table 3, the rate of nucleophilic
attack at the allyl group of the starting cationic sub-
strate (k2) increases with increasing basicity and de-
creasing steric requirements of the entering amine, in
line with previous findings [5,6,22] (pKa=11.12, 10.84
and 8.50 for piperidine, diethylamine and morpholine,
respectively).
(step k%). Such an intermediate, which goes undetected
2
1
in any H NMR spectrum of the reaction mixture, may
Diethylamine and morpholine display very similar
reactivity since the higher basicity of the former is
offset by its higher steric hindrance at the reacting
centre. These rate effects are in line with the associative
bimolecular nature of the k2 path, and appear to be
virtually independent of the metal substrate. A remark-
able accelerating effect upon increasing the steric bulki-
ness of the imino nitrogen substituent R of the P–N
ligand is observed. In Fig. 3 is shown a plot of log k2
vs. the front strain parameter S of the substituent R
[23]. Such steric influence can be rationalized by invok-
ing some sort of distortion of the h3-bound allyl ligand
leading to an early product-like transition state. This
ground-state allylic distortion is believed to play a
major role in palladium-catalysed asymmetric allyl ami-
nations [24]. Such interpretation is further supported by
X-ray structure determinations [6]. The k2 values do not
appear to be markedly affected by substituents in the
aryl ring of the NR group in the complexes 1e (R=
C6H4OMe-4), 1f (R=C6H3Me2-2,6) and 1g (R=
C6H3(i-Pr)2-2,6) (see Table 3). Apparently, the
possibility for the aryl ring to rotate around the NꢂC
be conceived of as a labile five-coordinate transient in
equilibrium with four-coordinate species bearing either
an h1-allyl ligand or a monodentate iminophosphine
ligand with a dangling imino arm. Formation of a
low-concentration intermediate is suggested by the rate
increase of the h3–h1–h3 process observed in the pres-
ence of amines, as discussed earlier. The k2 and k% steps
2
yield palladium(0) products with an h2-bound allylam-
monium group, which are rapidly converted to the final
complexes [Pd(h2-fn)(P–N)] by the more p-accepting
fumaronitrile ligand. According to this mechanism, the
rate law should be
k
obs=(k2[HY]+k%K1[HY]2)/(1+K1[HY])
(4)
2
which will reduced to Eq. (2) with k%K1=k3 if the term
2
K1[HY] is much lower than 1, as is expected from our
inability to detect appreciable concentrations of the
reactive intermediate [Pd(1-PhC3H4)(P–N)(HY)]+ (for
these ligands which are firmly bound to the metal
through the strongly coordinating tertiary phosphorus
end, K2 and thus KE=0).

180
B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
This finding is not unprecedented as it was also ob-
served during the reactions of [Pd(h3-1,1-Me2C3H3)(P–
N)]+ with diethylamine [6]. Since it seems very unlikely
that both the isomers react at the same rate, the above
observation is better explained by different amination
rates which are lower than the I ? II interconversion
rate. Thus, the experimental k2 values can be considered
as the sum of the two contributions, unless one isomer
is by far preferred for electronic and/or steric reasons.
Inspection of the 13C NMR data of Table 2 shows that
the unsubstituted allyl carbon C3 for isomer II (C3 trans
to P, lC 80.7–84.9 ppm) is more deshielded than the
corresponding carbon of isomer I (C3 trans to N, lC
50.4–54.5 ppm). On this basis, the lower electron den-
sity at C3 suggests a higher susceptibility to nucleophilic
attack for isomers II [25]. The observed steric influence
on the rates (Fig. 3) is also rationalized by a greater
contribution of isomer II to the amination step k2: in
these isomers the steric interaction between the reacting
CH2 allyl terminus and the cis NR group increases with
increase in the steric crowding at the nitrogen-bonded
carbon atom of the substituent R. In this context, it is
interesting to note that the similar k2 values for com-
plexes 1b, 1f and 1g are paralleled by rather close C3
carbon resonances whereas this is not the case for
complexes 1b (R=i-Pr) and 1c (R=t-Bu) which have
similar C3 carbon resonances, but different k2 values [k2
(1c)\k2 (1b)] owing to the different steric requirements
of the nitrogen substituents.
bond allows it to minimize the steric demand, which
becomes comparable to that of the i-Pr group in 1b.
Because of the presence of two geometrical isomers I
and II for complexes 1a–1g, the nucleophilic attack
may occur at the unsubstituted carbon of the allyl
1
ligand of both isomers. H NMR data on the course of
the reactions of 1a with diethylamine (at 25°C) and
with piperidine (at −30°C) indicate that the isomer
ratio I/II remains constant throughout the reaction.
Fig. 2. Fit of kobs to [HY] for the reaction of 1a with piperidine in
CHCl3 at 25°C, in the presence of fumaronitrile ([fn]0=2×10−4
mol dm−3).
Table 3
As far as the third-order k3 path is concerned, the
values in Table 3 are the highest observed so far in the
amination of the cationic complexes [Pd(h3-allyl)(P–
N)]+ [6]. It thus appears that the k3 terms are strongly
dependent on the nature of the allyl ligand in the order
1-PhC3H4\C3H5ꢁ1,1-Me2C3H3. This explains why
Rate constants k2 and k3 for Reaction 1 with complexes 1a–1g in
CHCl3 at 25°C
a
Complex
Amine
k2
k3
(mol−1 dm3 s−1
)
(mol−2 dm6 s−1
)
1a
pip
dea
morph
0.1790.01
0.02590.002
0.02890.004
678947
2.4393.41
36.194.8
1b
1c
1d
1e
1f
pip
dea
morph
0.990.1
0.07290.007
0.1990.05
94099130
160918
3749129
pip
dea
morph
8.290.3
0.4290.04
0.4490.08
84691620
9489114
18369222
pip
dea
morph
2.8390.08
0.2390.02
0.1090.09
9069324
293952.2
11649252
pip
dea
morph
0.6890.01
0.0590.01
0.1190.01
376952
58919.2
47931
pip
dea
morph
0.990.1
0.0690.01
0.0890.03
41979125
62.1928.6
167986
1g
pip
dea
morph
1.0990.21
0.1790.07
0.0690.02
20639129
3609225
418970
Fig. 3. Correlation of log k2 to the front-strain parameter S of the
nitrogen substituent R in the iminophosphine ligands of complexes
1a–1d. For the (R)-bornyl substituent, the same parameter as the
2-endo-norbornyl group [23] was used.
a Pip=piperidine; dea=diethylamine; morph=morpholine.

B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
181
Table 4
calculated values of k2, k3 and KE are listed in Table 4.
As can be seen the k2 value of 2a is approximately 50
times greater than the corresponding value of 2b. Since
the complexes are present in solution almost entirely as
isomer I (2a) or as isomer II (2b), the CH2 allyl
terminus of 2a (in close proximity to the 6-Me pyridine
group) appears to be much more reactive than the CH2
allyl terminus of 2b (cis to the N-t-Bu group). Such a
large difference in reactivity is better rationalized by
steric factors (a much greater distortion of the allyl
ligand caused by the 6-Me group on the pyridine ring)
rather than by electronic factors (the C3 carbons reso-
nances being at close chemical shifts of 60.8 ppm for 2a
and 59.9 ppm for 2b in the 13C NMR spectra). Accord-
ingly, the amination rates (k2) of complexes [Pd(h3-al-
Rate constants (k2 and k3) and equilibrium constants (KE) for Reac-
tion 1, of complexes 2a, 2b and 3a with piperidine in CHCl3 at 25°C
Complex
k2
k3
KE
(mol−1 dm3 s−1
)
(mol−2 dm6 s−1
)
2a
2b
3.5490.19
0.0790.02
126910
2.191.3
0.2790.15
(0.2590.08) a
3a
0.6490.02
0.6790.04
(0.8490.02) a
a Evaluated from direct spectrophotometric titration (see Section
2).
the 1-phenylallyl complexes 1a–1g show the highest
overall rates of amination under comparable experi-
mental conditions. Owing to the composite nature of
lyl)(N–S)]+
(N–S=2-(thiomethyl)pyridine)
were
found to increase with increasing bulkiness of the pyri-
dine ligand [5]. It is worth noting that the isomer II,
predominant (88%) in the isomeric mixture of complex
3a, and the isomer II of 2b are structurally similar, with
the CH2 allyl terminus trans to the pyridine nitrogen
and cis to a coordinated atom bearing a t-Bu sub-
stituent. Even if isomer II of 3a undergoes a dynamic
process with inversion at sulphur, it is reasonable to
assume that it will react with piperidine at a rate
comparable to that of 2b. On this assumption, the
higher reaction rate observed for 3a (k2 value approxi-
mately nine times greater than the corresponding value
of 2b) would originate from a faster nucleophilic attack
at the CH2 allyl terminus trans to sulphur in isomer I.
the k3 constants (k3=k%K1 in the proposed mechanism)
2
and their large errors of estimate, it would be worthless
trying to interpret the observed variation of k3 values
for complexes 1a–1g in terms of the electronic and
steric effects of the P–N ligand and the attacking
amine, although gross changes reflect those related to
the k2 values (see earlier).
On the basis of the above results and the finding that
all the complexes [Pd(h3-1-PhC3H4)(L–L%)]+ undergo a
regioselective allyl amination yielding the linear ally-
lamines (E)-PhCHꢀCHCH2Y, we have carried out a
kinetic study of the reactions of piperidine with the
complexes 2a, 2b and 3a containing weakly bound
ancillary ligands such as 2-(iminomethyl)pyridines and
2-(thiomethyl)pyridines, in the presence of fumaroni-
trile and an excess of free bidentate ligand. In fact, with
these substrates a rapid pre-equilibrium displacement of
the labile L–L% ligand by the entering amine takes place
(KE=K1K2 in the proposed mechanism), yielding a
virtually unreactive bis-amino species [Pd(h3-al-
lyl)(HY)2]+ [4,5], while the unaltered starting complex
[Pd(h3-allyl)(L–L%)]+ undergoes rate-determining nu-
cleophilic attack at the unsubstituted allyl carbon by
the amine (k2 step). In the case of 2b (L–L%=C5H4N-2-
CHꢀN-t-Bu), the mono-amine intermediate [Pd(al-
lyl)(L–L%)(HY)]+ also undergoes a slow allyl amination
4. Conclusions
The cationic complexes [Pd(h3-1-PhC3H4)(L–L%)]+
undergo regioselective amination by secondary amines
at the unsubstituted allyl terminus, independently of the
nature of the bidentate ligand L–L% and of the attack-
ing amine. When L–L% is an iminophosphine, the over-
all reaction rates are the highest observed so far for
analogous complexes having different allyl ligands. This
is essentially due to the high k3 values of the quadratic
term k3[amine]2 in the experimental rate low. When
L–L% is a sterically hindered 2-(iminomethyl)pyridine,
the cationic complexes are present in solution almost
exclusively in one isomeric form. This finding, com-
pounded with the regioselectivity of nucleophilic attack,
allows us to draw significant conclusions about the
electronic and steric effects of the L–L% ligand in these
reactions.
by a further amine molecule (step k%), giving rise to the
observed third-order term k3 in the experimental rate
law 5:
2
k
obs=(k2[HY]+k3[HY]2)/(1+KE[HY]2/[L−L%]0) (5)
In the case of 2a (L–L%=6-MeC5H3N-2-
CHꢀNC6H4OMe-4) and 3a (L–L%=C5H4N-2-CH2S-t-
Bu), the observed rate law is
Acknowledgements
k
obs=k2[HY]/(1+KE[HY]2/[L−L%]0)
(6)
Both rate laws directly originate from Eq. (3) in
Financial support by the Italian Ministero per l’Uni-
versita’ e la Ricerca Scientifica e Tecnologica is grate-
fully acknowledged.
Scheme 2 with k3=k%K1 and K1[HY]ꢂ1 (Eq. (5)), and
2
with k3[HY]2ꢂk2[HY] and K1[HY]ꢂ1 (Eq. (6)). The

182
B. Crociani et al. / Inorganica Chimica Acta 315 (2001) 172–182
[11] B. Crociani, F. Di Bianca, A. Giovenco, T. Boschi, Inorg. Chim.
References
Acta 127 (1987) 169.
[12] (a) D. Marion, K. Wuthrich, Biochem. Biophys. Res. Commun.
113 (1983) 964. (b) A.A. Bothner-By, R.L. Stephens, J. Lee,
C.D. Warren, R.W. Jeanloz, J. Am. Chem. Soc. 105 (1984) 811.
(c) C. Griesinger, R.R. Ernst, J. Magn. Reson. 75 (1987) 261.
[13] D.W. Marquardt, SIAM J. Appl. Math. 11 (1963) 431.
[14] G. Bandoli, A. Dolmella, L. Crociani, S. Antonaroli, B. Cro-
ciani, Transition Met. Chem. 25 (2000) 17.
[15] B. Crociani, S. Antonaroli, R. Scarpelli, unpublished.
[16] (a) T.B. Rauchfuss, F.T. Patino, D.M. Roundhill, Inorg. Chem.
14 (1975) 652. (b) G.M. Kapteijn, M.P.R. Spee, D.M. Grove, H.
Kooijman, A. Spek, G. van Koten, Organometallics 15 (1996)
1405.
[1] (a) B.M. Trost, Angew. Chem., Int. Ed. Engl. 28 (1989) 1173. (b)
S.A. Godleski, in: B.M. Trost, I. Fleming, M.F. Semmelbach
(Eds.), Comprehensive Organic Synthesis, vol. 4, Pergamon,
Oxford, 1991, 585pp. (c) J. Tsuji, Palladium Reagents and
Catalysts. Innovation in Organic Synthesis, Wiley, Chichester,
1995, pp. 290.
[2] (a) C.G. Frost, J. Howarth, J.M.J. Williams, Tetrahedron:
Asymmetry 3 (1992) 1089. (b) J.M.J. Williams, Synlett (1995)
705. (b) B.M. Trost, D.L. Van Vranken, Chem. Rev. 96 (1996)
395. (c) H. Tye, D. Smyth, C. Eldred, M.J. Wills, J. Chem. Soc.,
Chem. Commun. (1997) 1053.
[3] (a) P. von Matt, O. Loiseleur, G. Kock, A. Pfaltz, C. Lefeber, T.
Feucht, G. Helmchen, Tetrahedron: Asymmetry 5 (1994) 573. (b)
R. Jumnah, A.C. Williams, J.M.J. Williams, Synlett (1995) 821.
(c) U. Burckhardt, V. Gramlich, P. Hoffmann, R. Nesper, P.S.
Pregosin, R. Salzmann, A. Togni, Organometallics 15 (1996)
3496. (d) U. Burckhardt, M. Baumann, A. Togni, Tetrahedron:
Asymmetry 8 (1997) 155. (e) K. Selvakumar, M. Valentini, M.
Wo¨rle, P.S. Pregosin, A. Albinati, Organometallics 18 (1999)
1207.
[4] (a) B. Crociani, S. Antonaroli, F. Di Bianca, L. Canovese, F.
Visentin, P. Uguagliati, J. Chem. Soc., Dalton Trans. (1994)
1145. (b) L. Canovese, F. Visentin, P. Uguagliati, F. Di Bianca,
S. Antonaroli, B. Crociani, J. Chem. Soc., Dalton Trans. (1994)
3113. (c) L. Canovese, F. Visentin, P. Uguagliati, B. Crociani, F.
Di Bianca, Inorg. Chim. Acta 235 (1995) 45.
[17] A. Albinati, J. Eckert, P.S. Pregosin, H. Ru¨egger, R. Salzmann,
C. Sto¨ssel, Organometallics 16 (1997) 579.
[18] (a) J. Sprinz, M. Kiefer, G. Helmchen, M. Reggelin, G. Huttner,
O. Walter, L. Zsolnai, Tetrahedron Lett. 35 (1994) 1523. (b) R.
Ferna`ndez-Galan, F.A. Jalo`n, B.R. Manzano, J. Rodr`ıguez-de la
Fuente, M. Vrahami, B. Jedlicka, W. Weissensteiner, G. Jogl,
Organometallics 16 (1997) 3758. (c) C. Breutel, P.S. Pregosin, R.
Salzmann, A. Togni, J. Am. Chem. Soc. 116 (1994) 4067. (d) P.S.
Pregosin, R. Salzmann, A. Togni, Organometallics 14 (1995)
842. (e) P. Barbaro, P.S. Pregosin, R. Salzmann, A. Albinati,
R.W. Kunz, Organometallics 14 (1995) 5160. (f) P.S. Pregosin,
R. Salzmann, Coord. Chem. Rev. 155 (1996) 35. (g) J. Hermann,
P.S. Pregosin, R. Salzmann, A. Albinati, Organometallics 14
(1995) 3311. (h) T. Hosokawa, Y. Wakabayashi, K. Hosokawa,
T. Tsuji, S.I. Murabashi, J. Chem. Soc., Chem. Commun. (1996)
859.
[5] (a) L. Canovese, F. Visentin, P. Uguagliati, G. Chessa, V.
Lucchini, G. Bandoli, Inorg. Chim. Acta 275 (1998) 385. (b) L.
Canovese, F. Visentin, P. Uguagliati, G. Chessa, A. Pesce, J.
Organomet. Chem. 566 (1998) 61.
[19] B. Crociani, S. Antonaroli, M. Paci, F. Di Bianca, L. Canovese,
Organometallics 16 (1997) 384.
[20] M. Kawatsura, Y. Uozumi, T. Hayashi, J. Chem. Soc., Chem.
Commun. (1998) 217.
[6] B. Crociani, S. Antonaroli, G. Bandoli, L. Canovese, F.
Visentin, P. Uguagliati, Organometallics 18 (1999) 1137.
[7] S. Antonaroli, B. Crociani, J. Organomet. Chem. 560 (1998) 137.
[8] G. Matsubayashi, M. Okumoka, T. Tanaka, J. Organomet.
Chem. 66 (1974) 465.
[21] O. Kuhn, H. Mayr, Angew. Chem., Int. Ed. Engl. 38 (1999) 343.
[22] L. Canovese, F. Visentin, G. Chessa, A. Niero, Inorg. Chim.
Acta 293 (1999) 44.
[23] H.D. Beckhaus, Angew. Chem., Int. Ed. Engl. 17 (1978) 593.
[24] (a) A. Togni, U. Burckhardt, V. Gramlich, P.S. Pregosin, R.
Salzmznn, J. Am. Chem. Soc. 118 (1996) 1031. (b) P.E. Blo¨chl,
A. Togni, Organometallics 15 (1996) 4125.
[9] M.C. Ashraf, T.G. Burrowes, W.R. Jackson, Aust. J. Chem. 29
(1976) 2643.
[10] W.L.F. Armarego, D.D. Perrin, Purification of Laboratory
Chemicals, 3rd ed., Pergamon, New York, 1988.
,
[25] B. Akermark, B. Krakenberger, S. Hansson, A. Vitagliano,
Organometallics 6 (1987) 620.
.