First copper(I) ferrocenyltetraphosphine complexes: Possible involvement in Sonogashira cross-coupling reaction?

Article abstract of DOI:10.1021/om700700q

Preparation and characterization of the first examples of copper(I) ferrocenylpolyphosphine complexes are reported. The molecular structure of complex {P,P′,P″-[1,1′,2,2′-tetrakis(diphenylphosphino)- 4,4′-di-tert-butylferrocene]iodocopper(I)} (1) was solved by X-ray diffraction studies, and its fluxional behavior in solution was investigated by VT-³¹P NMR; both revealed a net triligated coordination preference of the ferrocenyl tetraphosphine Fc(P)₄^tBu with copper. The tetradentate ligand is an active auxiliary in Sonogashira alkynylation; therefore the general question of copper as a competitive coordination partner in the Pd/Cu-catalyzed Sonogashira reaction was raised and discussed. Electronically neutral, activated, and deactivated aryl bromides were employed for coupling with phenylacetylene with various [(Pd)/(Cu)/(tetraphosphine)] systems. The catalytic investigations shown that 1 mol % of complex 1 in combination with palladium is far more effective and selective for Sonogashira coupling than 5 mol % of CuI and palladium in the coupling to phenylacetylene of the deactivated aryl bromide 4-bromoanisole. This system efficiently avoids the concurrent and deleterious consumption of phenylacetylene by formation of diyne or enynes. To our knowledge, this is the first time that this kind of high selectivity is induced in Sonogashira alkynylation by initial ligand complexation to copper instead of palladium. These results demonstrate that coordination of Cu halide cocatalyst is a factor that should no longer be neglected in mechanistic and applied studies of the Sonogashira reaction.

Full text of DOI:10.1021/om700700q

1506
Organometallics 2008, 27, 1506–1513
First Copper(I) Ferrocenyltetraphosphine Complexes: Possible
Involvement in Sonogashira Cross-Coupling Reaction?
Matthieu Beaupérin, Elie Fayad, Régine Amardeil, Hélène Cattey, Philippe Richard,
Stéphane Brandès, Philippe Meunier, and Jean-Cyrille Hierso*
Institut de Chimie Moléculaire de l’UniVersité de Bourgogne (ICMUB-UMR CNRS 5260), UniVersité de
Bourgogne, 9 AVenue Alain SaVary, 21078 Dijon, France
ReceiVed July 13, 2007
Preparation and characterization of the first examples of copper(I) ferrocenylpolyphosphine complexes
are reported. The molecular structure of complex {P,P′,P′′-[1,1′,2,2′-tetrakis(diphenylphosphino)-4,4′-
di-tert-butylferrocene]iodocopper(I)} (1) was solved by X-ray diffraction studies, and its fluxional behavior
in solution was investigated by VT-31P NMR; both revealed a net triligated coordination preference of
t
the ferrocenyl tetraphosphine Fc(P)4 Bu with copper. The tetradentate ligand is an active auxiliary in
Sonogashira alkynylation; therefore the general question of copper as a competitive coordination partner
in the Pd/Cu-catalyzed Sonogashira reaction was raised and discussed. Electronically neutral, activated,
and deactivated aryl bromides were employed for coupling with phenylacetylene with various [(Pd)/
(Cu)/(tetraphosphine)] systems. The catalytic investigations shown that 1 mol % of complex 1 in
combination with palladium is far more effective and selective for Sonogashira coupling than 5 mol %
of CuI and palladium in the coupling to phenylacetylene of the deactivated aryl bromide 4-bromoanisole.
This system efficiently avoids the concurrent and deleterious consumption of phenylacetylene by formation
of diyne or enynes. To our knowledge, this is the first time that this kind of high selectivity is induced
in Sonogashira alkynylation by initial ligand complexation to copper instead of palladium. These results
demonstrate that coordination of Cu halide cocatalyst is a factor that should no longer be neglected in
mechanistic and applied studies of the Sonogashira reaction.
Scheme 1. Catalytic Cycles Proposed for Pd/Cu-Cocatalyzed
Sonogashira Alkynylation
Introduction
The palladium/copper-cocatalyzed alkynylation of aryl and
vinyl halides, usually identified as Sonogashira-Tohda-
Hagihara cross-coupling, is among the most widely employed
methodologies to yield valuable (aryl)alkynes.1 The exact
mechanism of the homogeneous copper-cocatalyzed Sono-
gashira reaction is not yet well-understood and is generally
supposed to take place through two independent catalytic
cycles, as depicted in Scheme 1.
While the first cycle involving palladium is classical from
C-C cross-coupling formation, the second cycle generally
involving copper(I) iodide is poorly known.2 The base that can
be either organic (amines) or inorganic (carbonates) is believed
to assist the copper acetylide formation, with the help of a
π-alkyne copper complex, which would make the alkyne
terminal proton more acidic. The most efficient catalytic systems
reported up to now involve a myriad of different ligands for
which the coordination properties are commonly reported with
palladium (Ln in Scheme 1).1 Curiously, to the best of our
knowledge, the few mechanistic studies around Sonogashira
cross-coupling2 have neglected the possible interference of
copper as transition metal and the potential concurrent formation
of CuXLn-type species, which appears to us an interesting
parameter that could bring helpful information.3 Therefore, the
issue of the coordination of useful ligands to copper might be
examined in the future on a more regular basis.
For our part, we have recently disclosed that ferrocenyl
polyphosphines can be very efficient as ligands in Sonogashira
t
alkynylations (Scheme 2): the triphosphine ligand Fc(P)2 Bu-
(PiPr) allows aryl alkynylation from organic bromides and
chlorides at 10-1 to 10-4 mol % catalyst loading with TONs
up to 250 000;4 with the diphosphine Fc[P(FuMe)2]2 a TON of
(3) (a) Osakada, K.; Sakata, R.; Yamamoto, T. Organometallics 1997,
16, 5354, and references therein. In this very relevant article the authors
investigated intermolecular alkynyl-ligand transfer between copper(I) and
Pd(II) and showed the intricate influence of PPh3 and PEt3 on the reaction,
as well as the possible formation of large amounts of CuI(PPh3)3. (b) For a
review see: Osakada, K.; Yamamoto, T. Coord. Chem. ReV. 2000, 198,
379.
* Corresponding author. E-mail: jean-cyrille.hierso@u-bourgogne.fr. Tel:
+33 3 80 39 61 06. Fax: +33 3 80 39 36 82.
(1) Doucet, H.; Hierso, J.-C. Angew. Chem., Int. Ed. 2007, 46, 834.
(2) Chinchilla, R.; Nájera, C. Chem. ReV. 2007, 107, 874.
10.1021/om700700q CCC: $40.75
2008 American Chemical Society
Publication on Web 03/05/2008

First Copper(I) Ferrocenyltetraphosphine Complexes
Organometallics, Vol. 27, No. 7, 2008 1507
t
Table 1. Sonogashira Alkynylation with Pd/Fc(P)4 Bu Catalytic
Scheme 2. Ferrocenylpolyphosphine Ligands Used in
System
C-C and C-N Bond Formation
920 000samong the highest reportedswas obtained in the
coupling of 4-bromoacetophenone to phenylacetylene.5
We report herein that the tetraphosphine ligand Fc(P)4 Bu,
a Nonoptimized catalytic conditions: (alkyne:aryl) 2:1, 2 equiv of
K2CO3, 24 h in 10 mL of DMF at 130 °C; GC average yields from two
or three runs are given. bPhenylacetylene dimerization to enynes, as
described in ref 9, is responsible for lower yields; in all runs
unconverted bromobenzene (entry 2) or bromoanisole is observed (entry
3) in more than 30% yield.
t
1,1′,2,2′-tetrakis(diphenylphosphino)-4,4′-di-tert-butylfer-
rocene, which is a very efficient auxiliary in palladium-catalyzed
Suzuki and Heck reactions,6 also provides active systems in
the Sonogashira reaction in the presence of [Pd(η3-C3H5)Cl]2
and copper. More importantly, the coordination properties of
this tetraphosphine ligand with copper halides were examined,
and the X-ray characterization of the first complexes of group
11 for this class of ferrocenyl polyphosphines is reported.7 The
resulting complexes are air-stable P-triligated tetrahedral cop-
per(I) species, on which X-ray diffraction studies, VT-31P NMR
spectroscopy, exact mass, UV, and EPR measurements were
conducted. In light of the studies devoted to copper complex
preparation, the likelihood and interest of copper-phosphine
coordinative interference in Sonogashira cross-coupling reactions
is discussed for the first time, to our knowledge, concerning
this cocatalyzed reaction.
Complex Preparation and Characterization. On the basis
of the preliminary results obtained in Sonogashira alkynylation
reactions with the system Pd/CuI/ferrocenyltetraphosphine, we
started to explore several synthetic conditions to form copper
coordination complexes with the tetraphosphine metalloligand
t
Fc(P)4 Bu. Our first objective was to disclose the coordination
preference of ferrocenylpolyphosphines toward copper halides.
Additionally, the prediction of the number of copper nucleus
atoms able to coordinate the ligand, as well as the final geometry
of the possible coordination complexes formed, was not a priori
evident since the coordination chemistry of polyphosphines
(three or more phosphorus atoms available) toward copper is
scantily developed.10
The first experimental conditions we tested by reacting
Fc(P)4 Bu with 4-fold equivalent of the simple copper chloride
Results and Discussion
t
Preliminary Catalytic Experiments. The rich coordination
hydrate (ratio Cu:P ≈ 1 is standard conditions in catalytic runs)
at ambient temperature in DMF, toluene, dichloromethane, THF,
methanol (in the presence or without trimethylorthoformate as
dehydrating agent), and acetonitrile were unsuccessful. Under
similar conditions at 50 °C in methanol and THF, 31P NMR
spectroscopy showed the formation of mixtures of coordination
species we failed to isolate and crystallize.11 The same
phenomenon was observed by employing under analogous
conditions refluxing acetonitrile as the solvent. In an attempt
to make things easier, several reactions involving only 1 or 2
equivalents of copper chloride to the ligand were investigated
with analogous results.11 We then turned our attention to
anhydrous copper iodide as inorganic complex precursor, this
particular compound being moreover ubiquitous in Sonogashira
alkynylations. Suitable conditions were finally identified for
reproducible formation of copper-tetraphosphine coordination
complex 1 depicted in Scheme 3. From the mixture of 4 equiv
t
chemistry of Fc(P)4 Bu toward palladium has been correlated
to its performances in C-C cross-coupling catalysis.6,8 Experi-
mental conditions and results in Sonogashira alkynylation for
the catalytic system combining the palladium(II) precursor
complex [PdCl(η3-C3H5)]2 with the tetradentate ligand Fc(P)4 Bu
t
are presented in Table 1. Electronically neutral, activated, and
deactivated aryl bromides were coupled to phenylacetylene. The
nature of the substituent on the organic halide plays an important
role: indeed, in the absence of an electron-withdrawing sub-
stituent on the aryl partner, concomitant phenylacetylene dimeri-
zation (Hay-Glaser copper-catalyzed coupling and/or [Pd/Cu]-
cocatalyzed alkyne dimerization to enynes) hampered a total
aryl alkynylation (due to alkyne consumption; see Table 1,
entries 2 and 3).1,9
(4) Hierso, J.-C.; Fihri, A.; Amardeil, R.; Meunier, P.; Doucet, H.;
Santelli, M.; Ivanov, V. V. Org. Lett. 2004, 6, 3473.
t
of anhydrous CuI with Fc(P)4 Bu in deoxygenated acetonitrile,
(5) Hierso, J.-C.; Fihri, A.; Amardeil, R.; Meunier, P.; Doucet, H.;
Santelli, M. Tetrahedron 2005, 61, 9759.
at reflux for 3 h, a 1,1′,2-P-triligated copper iodide complex is
obtained. The complex is isolated from solution as orange
needles after a night at -18 °C; the excess of CuI cocrystallizes
as white, square-shaped crystals of a polymeric inorganic copper
compound, catena-[(µ3-iodo)(acetonitrile-N)copper(I)].12
In the course of our syntheses and attempts to identify
alternative coordination modes for copper, we obtained a
surprising result, important to mention, regarding cyclopenta-
dienyl-alkali metal chemistry (see below). X-ray quality single
crystals of a copper(I) complex stabilized by the unexpected
ligand 1,1′,2,3′-tetrakis(diphenylphosphino)-4,4′-di-tert-butyl-
(6) Hierso, J.-C.; Fihri, A.; Amardeil, R.; Meunier, P.; Doucet, H.;
Santelli, M.; Donnadieu, B. Organometallics 2003, 22, 4490.
(7) Hierso, J.-C.; Smaliy, R.; Meunier, P. Chem. Soc. ReV. 2007, 36,
1754; from the library of ferrocenyl polyphosphine ligands we developed,
only coordination complexes of groups 6 to 10 were reported up to now.
(8) Hierso, J.-C.; Fihri, A.; Ivanov, V. V.; Hanquet, B.; Pirio, N.;
Donnadieu, B.; Rebière, B.; Amardeil, R.; Meunier, P. J. Am. Chem. Soc.
2004, 126, 11077.
(9) Batanov, A. S.; Collings, J. C.; Fairlamb, I. J. S.; Holland, J. P.;
Howard, J. A. K.; Lin, Z.; Marder, T. B.; Parsons, A. C.; Ward, R. M.;
Zhu, J. J. Org. Chem. 2005, 70, 703, and references therein. The nature of
the copper species possibly involved in alkyne dimerization is also a very
interesting issue.

1508 Organometallics, Vol. 27, No. 7, 2008
Beaupérin et al.
Scheme 3. Copper Coordination Chemistry of
As shown in Figure 1, phosphorus P(3) almost lies in the
bisecting plane of the P(1)-Ct(1)-P(2) angle [dihedral angle
P(2)Ct(1)Ct(2)P(3) ) -27.1(3)°]; therefore a rotation of the
tetraphosphine relative to the free ligand is observed since
originally phosphorus corresponding to P(2) and P(3) adopt a
pseudoeclipsed conformation [equivalent dihedral angle -5.6(1)°].
t
Ferrocenyltetraphosphine Fc(P)4 Bu
t
In the mononuclear palladium complex [PdCl2{Fc(P)4 Bu}] the
mutual rotation of Cp rings was found to be much less
pronounced, and thus its geometry was found closer to that of
the free ligand [dihedral angle -11.2°]. In 1 the phosphorus
atoms bonded to copper do not exhibit noticeable deviation from
the average plane of the Cp rings; the angle between the planes
encompassing the Cp rings is 5(1)°.
Scheme 4. Mononuclear Copper Complex 1b
As shown in Table 3 the geometrical parameters regarding the
copper coordination sphere for 1b (Figure 2) are very similar to
those found for 1. The most noticeable difference in 1b is the
unexpected position of phosphorus P(4), adjacent to the tBu group.
31P{1H} NMR spectroscopy studies were conducted on 1 in
deuterated dichloromethane: at ambient temperature two signals
at -23.4 and -24.6 ppm are detected (Figure 3, top); the first
one is a clear singlet, whereas the second one is broader and
ill-defined. Low-temperature NMR (VT, Figure 3) conducted
down to -94 °C remarkably confirmed the X-ray structural
studies. Upon cooling the temperature, the fluxional behavior
of the copper complex is evidenced through coalescence (0 to
-40 °C) and decoalescence (-60 °C) of the broader signal.
Three new signals appear at -75 °C, and the occurrence of
four different phosphorus from which three are in a dynamic
exchange (centered at -15.1, -22.1, and -36.6 ppm) is
ascertained at -94 °C. The appearance and chemical shift of
the singlet are almost not modified by the variation of tempera-
ture (shift from -23.4 ppm at 27 °C to -26.7 ppm at -94 °C).
The dynamic phenomenon is reversible since upon increasing the
temperature again above the coalescence temperature, the
spectrum exhibits again the two characteristic signals at -23.4
and -24.6 ppm. These signals are only slightly shifted downfield
relative to the free ligand (multiplets at -28.6 and -32.2 ppm
in CDCl3).6 The 1H NMR at ambient temperature did not show
excessively broad signals, so that they could be properly
assigned; relative to the free ligand all the signals are shifted
downfield: Cp protons give two signals at 4.25 and 4.14 ppm,
ferrocene (1b Scheme 4) were fortuitously formed in quite
significant quantity (about 20% from H NMR integration of
1
CpH relative to 1). The X-ray diffraction studies were repeated
on several crystals and confirmed the first analysis.
Details of the structure determinations are given in the
Experimental Section. Table 2 lists crystallographic data for
complexes 1 and 1b; selected bond distances and angles are
given in Table 3.
The X-ray molecular structure obtained for 1 (Figure 1)
indicated that the copper(I) atom is bonded to three phosphorus
atoms of the tetraphosphine in a distorted tetrahedral environ-
ment completed by a iodide [Cu-I 2.5848(14) Å]. The three
metal-phosphorus bonds are substantially different [Cu-P(1)
2.310(3) Å, Cu-P(2) 2.355(2) Å, Cu-P(3) 2.273(2) Å], with
a relative variation around 3.5% observed between the shortest
and the longest one. In general, for the unsolvated complexes
[(PPh3)3CuX] (X ) halide) X-ray structural data have shown
P-Cu-P and P-Cu-X angles close to the ideal tetrahedral
angle of 109.5°.13 In 1 the major deviation is observed for the
closed angle P(1)-Cu-P(2) ) 85.9(1)°, originating from the
1,2-P chelating pattern, and P-Cu-I angles lie between 116°
and 121°; several molecules of acetonitrile are present in the
crystals. To accommodate its fac tricoordination, the ferroce-
nyltetraphosphine ligand undergoes severe constraints worth
detailing in light of the previously reported X-ray diffraction
structures of the free ligand and of the mononuclear palladium
t
the methyl groups from Bu are detected at 0.82 ppm, and the
signals corresponding to phenyl protons of the PPh2 groups give
large multiplets between 6.80 and 8.10 ppm.11
The VT-31P NMR data evidence the lability of the copper-
iodide fragment, for which the interaction with three of the four
phosphorus atoms is undoubtedly observed whatever the tem-
perature. The phosphorus corresponding to P(4) in the molecular
structure in the solid state is certainly kept away from the
coordination sphere of copper, even in solution at ambient
temperature. The nature of the fluxional behavior visibly reveals
that the three phosphorus stay bonded to copper (three signals
at -94 °C, which coalesce to form one signal at rt);14 this
suggests that at higher temperature the mutual locations of
phosphorus toward the copper atom (clearly nonequivalent in
the solid state) are exchanged at high rate, leading to an average
signal at the NMR scale rate. VT-31P NMR studies on 1b
revealed a similar behavior, in which the atom P(4) give a singlet
detected at -20.7 ppm and is not involved in the dynamic
exchange of the three other phosphorus atoms toward CuI; they
gave two broad signals at -28.3 and -22.9 ppm (with 1:2
integration ratio, respectively).11 The 1H NMR at ambient
t
complex [PdCl2{Fc(P)4 Bu}].6
(10) (a) Hierso, J.-C.; Amardeil, R.; Bentabet, E.; Broussier, R.;
Gautheron, B.; Meunier, P.; Kalck, P. Coord. Chem. ReV. 2003, 236, 143.
A check in the Cambridge Structural Database (06/2007) revealed that X-ray
diffraction studies for copper complexes with a tricoordinated polyphosphine
ligand were found only for the flexible tripodal ligands 1,1,1-tris(2-
diphenylphosphinomethyl)ethane and 1,1,1-tris(2-(diphenylphosphino)eth-
yl)amine; see for instance: (b) Garcia-Seijo, M. I.; Sevillano, P.; Gould,
R. O.; Fernandez-Anca, D.; Garcia-Fernandez, M. E. Inorg. Chim. Acta
2003, 353, 206. (c) Ghilardi, C. A.; Midollini, S.; Orlandini, A. Inorg. Chem.
1982, 21, 4096.
(11) NMR spectra are available as Supporting Information.
(12) (a) Arkhireeva, T. M.; Bulychev, B. M.; Sizov, A. I.; Sokolova,
T. A.; Belsky, V. K.; Soloveichik, G. L. Inorg. Chim. Acta 1990, 169, 109.
(b) Healy, P. C.; Kildea, J. D.; Skelton, B. W.; White, A. H. Aust. J. Chem.
1989, 42, 79. (c) Nilsson, K.; Oskarsson, A. Acta Chem. Scand. A 1985,
39, 663. (d) Jasinski, J. P.; Rath, N. P.; Holt, E. M. Inorg. Chim. Acta
1985, 97, 91. A PovRay illustration of the staggered conformation of the
catena complex is available as Supporting Information.
(13) Hanna, J. V.; Boyd, S. E.; Healy, P. C.; Bowmaker, G. A.; Skelton,
B. W.; White, A. H. Dalton Trans. 2005, 2547.
(14) If one phosphorus atom was alternatively released from copper
coordination, at low temperature it should appear as a clear singlet.

First Copper(I) Ferrocenyltetraphosphine Complexes
Organometallics, Vol. 27, No. 7, 2008 1509
Table 2. Crystal Data and Structure Refinement Details for 1 and 1b
1
1b
formula
M
T, K
C66H62FeCuIP4 · 3(CH3CN)
1348.49
115(2)
C66H62FeCuIP4 · 3(CH3CN)
1348.49
115(2)
cryst syst
space group
a, Å
triclinic
P1
12.553(5)
triclinic
P1
11.7939(2)
j
j
b, Å
13.193(5)
15.6829(3)
c, Å
20.198(5)
18.4282(3)
R, deg
86.425(5)
76.582(5)
79.054(5)
3193.9(19)
91.007(1)
103.969(1)
103.612(1)
3204.8(1)
ꢀ, deg
γ, deg
V, Å3
Z
2
2
F(000)
1384
1.402
Nonius KappaCCD
ꢁ rot and ω scans
0.71073
1384
1.397
Nonius KappaCCD
ꢁ rot and ω scans
0.71073
Dcalc, g/cm3
diffractometer
scan type
λ, Å
µ, mm-1
cryst size, mm3
sin(θ)/λmax, Å-1
index ranges
1.184
1.184
0.30 × 0.30 × 0.175
0.45 × 0.37 × 0.25
0.65
0.65
h -16; 16
h -15; 15
k -14; 17
k -17; 20
l -26; 24
l -23; 23
RC ) reflns collected
IRC ) indep RC
IRCGT ) IRC and [I > 2σ(I)]
refinement method
no. of data/restraints/params
R for IRCGT
17674
21864
11 378 [R(int) ) 0.0627]
14 411 [R(int) ) 0.0215]
7174
13176
full-matrix LS on F2
11 378/108/746
R1a ) 0.080, wR2b ) 0.164
R1a ) 0.146, wR2b ) 0.199
1.067
full-matrix LS on F2
14 411/12/797
R1a ) 0.059, wR2b ) 0.134
R1a ) 0.064, wR2b ) 0.136
1.218
R for IRC
goodness-of-fitc
∆F, max, min, e Å-3
1.787 and -1.167
2.709 and -1.946
2
2 2
2
2
a R1 ) ∑(|Fo| - Fc|)/∑|Fo|. b wR2 ) [∑w(Fo - Fc ) /∑[w(Fo2)2]1/2 where w ) 1/[σ2(Fo + (0.0552P)2 + 28.7358P] for 1 and w ) 1/[σ2(Fo
+
18.4318P] for 1b where P ) (Max(Fo2,0) + 2Fc )/3. c Goodness of fit ) [∑w(Fo - Fc ) /(No - Nv)]1/2
.
2
2
2 2
Table 3. Selected Bond Distances and Angles for Crystals of 1 (left)
and 1b (right)
Distances (Å)
Cu-I
2.5848(14)
2.310(3)
2.355(2)
2.273(2)
2.5799(6)
2.2988(12)
2.3342(12)
2.2771(11)
Cu-P(1)
Cu-P(2)
Cu-P(3)
Angles (deg)
120.34(6)
P(1)-Cu-I
113.40(3)
P(2)-Cu-I
118.42(7)
116.57(7)
85.93(8)
106.92(9)
103.81(8)
36.2(3)
125.33(4)
118.15(3)
86.26(4)
105.06(4)
102.99(4)
36.8(2)
P(3)-Cu-I
P(1)-Cu-P(2)
P(1)-Cu-P(3)
P(2)-Cu-P(3)
P(1)-Ct(1)-Ct(2)-P(3)a
P(2)-Ct(1)-Ct(2)-P(3)
-27.1(3)
-26.4(2)
Figure 1. View of the molecular structure of copper complex 1.
Hydrogen atoms and solvate molecules are omitted for clarity.
Thermal ellipsoids have been drawn at the 50% probability level.
a Ct labels refer to Cp centroids.
temperature gave four signals at 3.95, 4.09, 4.47, and 4.58 ppm
for Cp protons, the methyl groups from tBu are detected at 0.80
ppm, and the signals corresponding to phenyl protons of the
PPh2 groups give signals between 6.0 and 8.25 ppm.11
Thus, in 1b the position of a PPh2 group adjacent to the tBu
group was to us rather enigmatic. A more accurate examination
of synthetic conditions led to a plausible explanation (illustrated
in Scheme 6) for the formation of the 1,1′,2,3′-tetrakis(diphe-
nylphosphino)-4,4′-di-tert-butylferrocene as a byproduct, under
certain conditions. It has been previously observed in the alkali-
metal chemistry of substituted phosphino-cyclopentadiene spe-
As mentioned above, the formation of 1b and especially of
its related ligand 1,1′,2,3′-tetrakis(diphenylphosphino)-4,4′-di-
tert-butylferrocene was unanticipated. We examined therefore
more carefully the pool of ligands we had routinely employed,
and 31P NMR spectroscopy confirmed the formation, in addition
t
to Fc(P)4 Bu, of this new tetraphosphine.11 The reported three-
t
step synthesis of Fc(P)4 Bu from the cyclopentadienyl (Cp)
(15) (a) Broussier, R.; Bentabet, E.; Mellet, P.; Blacque, O.; Boyer, P.;
Kubicki, M. M.; Gautheron, B. J. Organomet. Chem. 2000, 598, 365. (b)
Broussier, R.; Bentabet, E.; Amardeil, R.; Richard, P.; Meunier, P.; Kalck,
P.; Gautheron, B. J. Organomet. Chem. 2001, 637–639, 126.
(16) Broussier, R.; Ninoreille, S.; Legrand, C.; Gautheron, B. J.
Organomet. Chem. 1997, 532, 55.
t
lithium derivative BuCpLi has been initially optimized to
quantitatively induce the successive phosphorylations of the Cp
ring in adjacent positions, i.e., in ꢀ-position from the hindered
tBu group (Scheme 5).15

1510 Organometallics, Vol. 27, No. 7, 2008
Beaupérin et al.
t
Scheme 5. Synthesis Pathway for Fc(P)4 Bu
nylphosphino)-4-tert-butylcyclopentadienyllithium in the as-
sembly step 3 in Scheme 5. The major change in synthetic
procedures relative to the originally published conditions15a was
clearly the concentration in cyclopentadienyllithium salts.
Indeed, concentrations around 1.5 × 10-3 mol · mL-1 in toluene
were employed, 4-fold higher than the usual conditions reported
(about 3.5 × 10-4 mol · mL-1). Deprotonation of quickly formed
(diphenylphosphino)cyclopentadienes by some remaining cy-
clopentadienyllithium salt reactant7 might help such rearrange-
ments. Further studies are necessary to investigate this side-
product formation in cyclopentadienyl-alkali metal salt chemistry;
in particular calculations on proton acidity of phosphorylated
cyclopentadienes might help in this purpose. Nevertheless the
fact that a phosphorylation on a position adjacent to a hindered
group on the Cp rings is feasible is a result worth mentioning.
The difficulty encountered in forming tetraphosphine/copper
coordination complexes at ambient temperature, whatever the
solvent we employed (in priority solvents classically used in
catalytic studies), clearly indicates that, as expected, affinity of
Figure 2. View of the molecular structure of copper complex 1b.
Hydrogen atoms and solvate molecules are omitted for clarity.
Thermal ellipsoids have been drawn at the 50% probability level.
t
polyphosphine Fc(P)4 Bu toward palladium is greater (at rt
palladium/tetraphosphine complexes immediately form).6 How-
ever, when the temperature is raised to 80 °C, several coordina-
tion species with copper salts are observed, and under certain
conditions very stable complexes such 1 are even formed. The
excess of copper (relative to palladium) usually employed under
Sonogashira catalytic conditions statistically augments the
opportunity of ligand transfer from palladium to copper.
Additionally, the fluxional behavior of CuI toward phosphorus
ligation evidenced herein could also be a favorable factor for
copper interference in palladium/polyphosphine coordination.
Therefore, the opening study reported herein is, from a
coordination viewpoint, rather in favor of a possible role of CuI/
ligand interaction in the multipart mixture (Pd/Cu/ligand/aryl/
alkyne/base/solvent) that characterizes Sonogashira alkynyla-
tions. Investigations combining spectroscopic and kinetic methods,
like has been done in the evaluation of factors contributing to
the accelerating effect of CuI in Stille palladium-catalyzed cross-
coupling of aryl iodides with organostannanes,17 would be of
interest on this topic. With the view to directly get valuable
insights into the influence of copper/ligand coordination com-
plexes in Sonogashira reactions, a more in-depth study of the
catalytic performances obtained under various conditions was
then carried out and is described in the following.
Comparative Catalytic Performances of Pd/Cu Tetra-
phosphine Systems. The three electronically neutral, activated,
and deactivated aryl bromides initially employed for coupling
with phenylacetylene and the temperature and solvent/base
conditions described below were used to reinvestigate the
performances obtained in aryl alkynylation with various [(Pd)/
(Cu)/(tetraphosphine)] systems. Table 4 summarizes the condi-
tions and results obtained under strict absence of oxygen.9
We first checked the origin of phenylacetylene dimerization
and enyne formation that hampered the coupling of the more
demanding substrate 4-bromoanisole in the preliminary catalytic
reactions (Table 1). Experiments conducted with phenylacety-
Figure 3. VT-31P{1H} NMR of crystals of 1 solubilized in CD2Cl2.
cies the occurrence of phosphino group migration due to
sigmatropic rearrangement (Scheme 6b).16
This kind of transposition is probably at the origin of the
formation (Scheme 6a) of 1,3-bis(diphenylphosphino)-4-tert-
n
butylcyclopentadienes, which upon deprotonation with BuLi
(step 2 in Scheme 6) would yield 1,3-bis(diphenylphosphino)-
4-tert-butylcyclopentadienyllithium; the many isomeric forms
existing for the corresponding 1,3-bis(phosphino)dienes might
favor 1,5-sigmatropic transposition. The formation of 1,1′,2,3′-
tetrakis(diphenylphosphino)-4,4′-di-tert-butylferrocene would
then originate from contamination of 1,2-bis(diphenylphos-
phino)-4-tert-butylcyclopentadienyllithium by 1,3-bis(diphe-
(17) Casado, A. L.; Espinet, P. Organometallics 2003, 22, 1305, and
references therein. In this study the authors found that CuI captures part of
the free ligand (PPh3, AsPh3) and consequently moderates their deleterious
effect on the rate-determining Sn/Pd transmetalation step.

First Copper(I) Ferrocenyltetraphosphine Complexes
Organometallics, Vol. 27, No. 7, 2008 1511
Scheme 6. Rearrangement in Alkali-Metal Chemistry of Substituted Phosphinocyclopentadienes
Table 4. Investigation of Aryl Alkynylation with [(Cu)/(Pd)/(Tetraphosphine)] Systemsi
yield in
coupling product (%)
t
entry
palladium source
copper source
additional Fc(P)4 Bu
aryl bromide
1
2
3
4
5
6
7
8
5 mol % CuI
5 mol % CuI
1 mol % complex 1
5 mol % CuI
1 mol % complex 1
5 mol % CuI
0a
1 mol %
0a
0a
4-bromoanisole
4-bromoanisole
4-bromoanisole
4-bromoacetophenone
4-bromobenzene
4-bromoanisole
4-bromoacetophenone
4-bromobenzene
4-bromoanisole
0a
0a
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
0.5 mol % [PdCl(allyl)]2
5b
1 mol %
1 mol %
1 mol %
100c
89d
22e
100f
98g
91h
9
10
11
12
1 mol % complex 1
1 mol % complex 1
1 mol % complex 1
a No reactions. b 95% unreacted ArBr, phenylacetylene dimerization diyne, and enyne products are obtained.9 c Phenylacetylene dimerization and
unidentified heavier products are obtained from phenylacetylene excess. d 5% of unreacted ArBr and 5% of heavier benzene products. e 70% of
unreacted ArBr and 7% of anisole. f Traces of diyne from phenylacetylene dimerization. g 2% of unreacted ArBr. h 9% of unreacted ArBr. i Catalytic
conditions: (alkyne:aryl) 2:1, 2 equiv of K2CO3, 24 h in 10 mL of DMF at 130 °C, GC average yields from two or three runs are given.
lene in the presence of only a copper source (CuI, complex 1,
or mixture of CuI/Fc(P)4 Bu) were conducted (entries 1 to 5).
The apparent reason for this is that the system {[PdCl(allyl)]2/
1} does not lead to diyne and enyne formation, contrary to the
other [Pd/Cu] systems we investigated. To our knowledge, this
is the first time that this kind of high selectivity is induced in
Sonogashira alkynylation by initial ligand complexation to
copper instead of palladium. These results demonstrate that
coordination of Cu halide cocatalyst in the Sonogashira reaction
is a factor that should no longer be neglected. In particular
Fairlamb and co-workers have proposed that for terminal alkynes
homocoupling Cu(I)/Cu(II) redox systems are necessary to
reoxidize Pd(0) to Pd(II).9 To go further on this idea, it is
reasonable to think that the coordination of a ligand on CuI
should strongly modify any Cu(I)/Cu(II) redox systems and,
therefore, should influence the alkyne dimerization side reactions
of the fascinating Sonogashira coupling.
t
In these experiments the starting alkyne and aryl halides were
recovered unchanged. It clearly appeared that palladium is
absolutely necessary; otherwise no reactions occur with the
alkyne in the presence or in the absence of 4-bromoanisole.
The pivotal role of palladium in alkyne dimerization was
confirmed in the following runs (entry 6), in which the coupling
of bromoanisole to phenylacetylene in the presence of CuI and
in the absence of tetraphosphine ligand was attempted. Under
these conditions only 5% of coupled (aryl)alkyne was obtained,
while diyne and enynes are formed in appreciable amounts.
In the next runs (entries 7 to 9) copper-free alkynylation
conditions were used, showing that 4-bromoacetophenone and
bromobenzene can visibly be coupled to phenylacetylene (100%
and 89% yield, respectively) in the absence of copper. More
importantly, it was shown that copper has a deleterious effect
for bromobenzene coupling when compared to the preliminary
results reported in Table 1 (entry 2). Conversely, the coupling
of 4-bromoanisole to phenylacetylene was not very successful
(22% of coupling product) in the absence of copper (entry 9).
Finally, we checked the usefulness of complex 1 as a source of
copper (entries 10 to 12). Much to our surprise the three
electronically neutral, activated and deactivated aryl bromides
were very efficiently coupled to phenylacetylene. In particular,
the 4-bromoanisole substrate gave 80 to 90% yield of coupling
product in the several runs conducted, while under the previous
conditions explored this substrate was very reluctant to couple.
Conclusions
The first examples of copper(I) ferrocenylpolyphosphine
complexes were reported. The molecular structures of complexes
{P,P′,P′′-[1,1′,2,2′-tetrakis(diphenylphosphino)-4,4′-di-tert-bu-
tylferrocene]iodocopper(I) and {P,P′,P′′-[1,1′,2,3′-tetrakis(diphe-
nylphosphino)-4,4′-di-tert-butylferrocene]iodocopper(I)} were
solved by X-ray diffraction studies. Copper complexes incor-
porating phosphine ligands are less common than copper
complexes incorporating nitrogen ligands (X-ray structures based
on CuP3 motif are 30 times less numerous than CuN3 in CSD
06/2007); structures of copper complexes incorporating tetra-

1512 Organometallics, Vol. 27, No. 7, 2008
Beaupérin et al.
t
of Fc(P)4 Bu (3.38 × 10-5 mol, 35 mg) and [PdCl(η3-C3H5)]2 (1.69
× 10-5 mol, 6.18 mg), under argon, affords the coupling products.
Pure products were obtained after addition of water, extraction with
organic solvents, separation, drying, concentration, and chroma-
tography on silica gel.
phosphine ligands are even more rare.10a The dynamic behavior
of these species in solution was investigated by VT-31P NMR
and showed that the CuI fragment has a fluxional behavior over
three of the four phosphorus atoms. This type of dynamic
t
activity was interestingly exemplified before with Fc(P)4 Bu in
t
palladium coordination chemistry.6 The synthetic and charac-
terization studies were put in relation with the role of the ligand
Tetraphosphine Fc(P)4 Bu. The ligand 1,1′,2,2′-tetrakis(diphe-
t
nylphosphino)-4,4′-di-tert-butylferrocene, Fc(P)4 Bu, was prepared
following the synthesis reported in the literature.6,15 Chemical shifts
t
Fc(P)4 Bu in Pd/Cu-cocatalyzed alkynylation reactions to suggest
of pure product are as follow:8 1H NMR (CDCl3): δ 6.40–8.90 (m,
that under the conditions we used interferences of CuI with the
ligand cannot be totally excluded.
40 H, Ph), 4.14, 4.05 (s, 2 H each, Cp), 0.70 (s, 18H, tBu). 31P{1H}
(CDCl3): δ -28.6, -32.2 (AA′BB′, 3JAB ) 3JA′B′ ) 74 Hz,
J ′
AA
TS
From the catalytic performances it was clearly shown that 1
mol % of complex 1 is far more efficient and selective for
Sonogashira coupling than 5 mol % CuI in the coupling of a
set of electronically neutral, activated, and deactivated aryl
bromides to phenylacetylene. It avoids the concurrent and
deleterious consumption of phenylacetylene by formation of
diyne or enynes. Several experiments showed that the use of
CuI (or complex 1) in the absence of a palladium source does
not produce any reaction (either aryl alkynylation or pheny-
lacetylene dimerization). Additionally, the use of complex 1
offers practical advantages: it has been used in only 1 mol %
against 5 mol % for CuI, and its enhanced stability to air allows
a much easier weighing and stocking compared to strictly
anhydrous CuI.
Finally, these results suggest that a more systematic examina-
tion of the interaction of copper(I) halides with ligands primarily
devoted to palladium stabilization in reactions containing the
two metals would be of high mechanistic interest. This is
especially true for ligands incorporating nitrogen atoms, which
are susceptible to transfer to copper,18 as well as for labile
monophosphines such as PPh3,3a,17 and for alkynylations
conducted in acetonitrile solvent,19 in which robust copper-
phosphine complexes might be formed. Finally, enhanced
performances could be expected from a ligand chemistry
directed to copper in complement to palladium or even from
well-defined Pd/Cu bimetallic complexes.
) 60 Hz). UV–visible (nm, CH2Cl2): 228, 279 (shoulder), 466. In
the ground state, the electronic structure has a strong metal character
and is described as being (3e2g)4(5a1g)2. The two lower energy
absorption bands are assigned to the 5a1g f 4e1g and 3e2g f 4e1g
transition, respectively; these transitions are metal-centered. The
intense LMCT band near 220 nm is assigned to transitions from a
π-orbital ligand to the 4e1g metal level.
Copper Complexes Synthesis: 1. Anhydrous copper iodide
conserved under an argon atmosphere (0.28 g, 1.470 mmol, 4 equiv)
t
and Fc(P)4 Bu (0.4 g, 0.387 mmol) were dissolved in degassed
acetonitrile (20 mL). The mixture was stirred at 80 °C for 3 h,
after which time a pale yellow precipitate was visible. After cooling,
the mixture was filtered over Celite and the resulting solution
concentrated in Vacuo. After two days in a freezer at -18 °C, 1
crystallizes as air-stable orange needles in 54% yield (0.25 g, 0.21
mmol), easily separable from cocrystallized square-shaped, white,
transparent crystals of the inorganic complex catena-[(µ3-iodo)(ac-
etonitrile-N)copper(I)], arising from excess CuI. C66H62P4FeCuI
(MW 1225.39, exact mass 1224.149): m/z 1113.236 (M+ - I +
O); 1097.243 (M+ - I), simulated 1097.244 (σ ) 0.058). Anal.
Calcd: C 64.69, H 5.10. Found: C 64.62, H 4.97. From single
1
crystals (see spectrum in the Supporting Information), H NMR
(CDCl3 at 27 °C): δ 6.80–8.10 (m, 40 H, Ph), 4.25, 4.14 (s, 2 H
t
each, Cp), 0.82 (s, 18H, Bu), residual uncoordinated acetonitrile
gives an intense singlet at 2.02 ppm. 31P{1H} (CDCl3, at 27 °C):
δ -23.4 (s), -24.6 (m, br). UV–visible (nm, CH2Cl2): 232, 298
(shoulder), 467. Since copper(I) has fully occupied d-orbitals, no
d f d transition occurs for the complexes in the visible region and
the UV–visible is similar to the spectrum of the starting ferrocenyl
ligand. The electronic spectrum for 1 is only the signature of the
ligand; no metal–ligand charge transfer (MLCT) occurs. As evoked
for the UV–visible analysis of Cu(I), fully occupied d-orbitals give
a diamagnetic EPR-silent complex at 115 K, which is in total
agreement with NMR and X-ray diffraction analyses, and shows
that no oxidation to Cu(II) occurred. 1b was eventually obtained
Experimental Section
General Considerations. The reactions were carried out in oven-
dried (115 °C) glassware under an argon atmosphere using Schlenk
and vacuum-line techniques. Except for methanol and acetonitrile
(which were deoxygenated by nitrogen freezing/vacuum) the
solvents were distilled over appropriate drying and deoxygenating
agents prior to use. 1H (300.13 and 600.13 MHz) and 31P (121.49
and 242.94 MHz) including variable-temperature NMR experiments
were performed in CDCl3 or CD2Cl2 on a 600 MHz Bruker Avance
II and a 300 MHz Bruker Avance. NMR, UV (Varian Cary 50
spectrophotometer), elemental analyses, and electrospray mass
spectrometry (on a Bruker microOTOF-Q instrument) were per-
formed at the Centre de Spectrométrie Moléculaire (CSM) of the
Institut de Chimie Moléculaire de l’Université de Bourgogne
(ICMUB-UMR CNRS 5260). The UV-visible characterization of
the copper complexes was performed in dichloromethane. Continu-
ous wave (CW) EPR experiments were recorded in frozen solution
(115 K) on a Bruker ELEXSYS 500 spectrometer equipped with a
4122 SHQE/0405 X-band resonant cavity operating at 9.30 GHz
(6 mW power, 100 kHz modulation).
t
concomitantly with 1 from a pool of Fc(P)4 Bu contaminated with
the unforeseen ligand 1,1′,2,3′-tetrakis(diphenylphosphino)-4,4′-di-
tert-butylferrocene. 1,1′,2,3′-Tetrakis(diphenylphosphino)-4,4′-di-
tert-butylferrocene is formed as a byproduct upon modification of
t
the synthetic procedure15a to produce Fc(P)4 Bu (as explained in
t
Schemes 5 and 6). In particular, a high concentration of BuCpLi
(1.43 × 10-3 mol · mL-1 in toluene) was used in the first step of
the synthesis. C66H62P4FeCuI (MW 1225.39, exact mass 1224.149):
m/z 1113.246 (M+ - I + O), simulated 1113.239 (σ ) 0.388);
1097.254 (M+ - I). From single crystals (see spectrum in the
Supporting Information), 1H NMR (CDCl3 at 27 °C): δ 6.50–8.25
(m, 40H, Ph), 3.95, 4.09, 4.47, 4.58 (s, 1H each, Cp), 0.80 (s, 18H,
tBu). 31P{1H} (CDCl3, at 27 °C): δ -20.7 (s), -22.9 (m, br), -28.3
(m, br).
Sonogashira Reactions. The reaction of aryl halide (3.38 × 10-3
mol), phenylacetylene (2 equiv, 0.75 mL), K2CO3 (2 equiv, 0.935
g), and CuI (1.69 × 10-4 mol, 32 mg, or complex 1, 3.38 × 10-5
mol, 41 mg) at 130 °C during 24 h in DMF (10 mL) in the presence
X-ray Crystallographic Structure Determination of 1 and
1b. Intensity data were collected on a Nonius Kappa CCD at 115
K. The structures were solved by direct methods (SIR92)20 and
refined with full-matrix least-squares methods based on F2 (SHELXL-
(18) Batey, R. A.; Shen, M.; Lough, A. J. Org. Lett. 2002, 4, 1411.
(19) DeVasher, R. B.; Moore, L. R.; Shaughnessy, K. H. J. Org. Chem.
2004, 69, 7919.
(20) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J. Appl.
Crystallogr. 1993, 26, 343.

First Copper(I) Ferrocenyltetraphosphine Complexes
Organometallics, Vol. 27, No. 7, 2008 1513
97)21 with the aid of the WINGX program.22 All non-hydrogen
atoms were refined with anisotropic thermal parameters. Hydrogen
atoms were included in their calculated positions and refined with
a riding model. One phenyl group of the 1b isomer was found to
be disordered over two positions (occupation factors to 0.50:0.50).
As for 1b, one phenyl group of the complex 1 is disordered over
two positions (occupation factors: converged to 0.56:0.44). Both
components of this disordered phenyl group were refined as rigid
groups with anisotropic thermal factors Uij restrained to approximate
an isotropic behavior (ISOR). The last Fourier difference maps
indicate the presence of a third badly disordered acetonitrile solvate
molecule in the asymmetric unit, but any tentative model of this
disorder failed. Additionally, the presence of a third solvent
molecule is corroborated by the fact that the unit cell volume (3194
Å3) is very similar to those observed for the isomer 1b (3205 Å3),
which contains three acetonitrile molecules per asymmetric unit.
Crystallographic data are detailed in Table 2.
Acknowledgment. We are thankful to Dr. J. Andrieu
(Université de Bourgogne, ICMUB-SYMS) for helpful
discussions on copper coordination. Thanks are due to G.
Delmas and S. Royer for technical help in ligand synthesis.
We are also thankful to M.-J. Penouilh for exact mass
measurements. We thank the “Conseil Régional de Bour-
gogne” and the CNRS for a BDI grant to M.B.
Supporting Information Available: Crystal structure data for
1 and 1b as a CIF file (CCDC 653478 and 653479), NMR spectra
corresponding to ref 11, exact mass ES analysis for 1 and 1b,
PovRay illustration of inorganic complex catena[(µ3-iodo)(aceto-
nitrile-N)copper(I)], GC conditions, and analysis from aryl alky-
nylation reactions. This material is available free of charge via the
Internet at http://pub.acs.org.
(21) Sheldrick, G. M. SHELXL-97; Institüt für Anorganische Chemie
der Universität Göttingen: Germany, 1998.
(22) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837.
OM700700Q