Fluorous biphasic catalysis: Synthesis and characterization of copper(I) and copper(II) fluoroponytailed 1,4,7-Rf-TACN and 2,2′-R f-bipyridine complexes - Their catalytic activity in the oxidation of hydrocarbons, olefins, and alcoho

Article abstract of DOI:10.1002/chem.200304771

In this contribution on fluorous biphasic catalysis (FBC), we present the synthesis and characterization of new copper complexes, and define their role, as precatalysts, in the FBC oxidation of hydrocarbons, olefins, and alcohols. Thus the previously repo

Full text of DOI:10.1002/chem.200304771

FULL PAPER
Fluorous Biphasic Catalysis: Synthesis and Characterization of Copper(i) and
Copper(ii) Fluoroponytailed 1,4,7-R -TACN and 2,2'-R -Bipyridine
f
f
Complexes–Their Catalytic Activity in the Oxidation of Hydrocarbons,
Olefins, and Alcohols, Including Mechanistic Implications
[
a]
[a]
[a]
[a]
MarÌa Contel,* Cristina Izuel, Mariano Laguna,* Pedro R. Villuendas,
[
b]
[c]
Pablo J. Alonso, and Richard H. Fish*
Abstract: In this contribution on fluo-
rous biphasic catalysis (FBC), we pres-
ent the synthesis and characterization of
new copper complexes, and define their
role, as precatalysts, in the FBC oxida-
tion of hydrocarbons, olefins, and alco-
hols. Thus the previously reported, but
poorly characterized, fluoroponytailed
action of 1 with [Cu(CH CN) ]PF or
[CuCl] provided new Cu complexes,
which could be isolated and fully char-
bon and olefin functionalization. The
oxidation reactions of these substrates in
the presence of the necessary oxidants,
tert-butyl hydroperoxide (TBHP) and
oxygen gas, proceeded under FBC con-
3
I
4
6
acterized as [Cu(R -tacn)X']X, in which
f
X ˆ PF (6) or X' ˆ Cl (7) (soluble in
6
II
I
I
perfluoroheptane). The Cu and Cu
ditions for 5, 7, and Cu salts with 2.
However, the complexes with ligand 2
complexes, 4 ± 7, were characterized by
elemental analysis, mass spectrometry,
and IR, diffuse reflectance UV/Vis, and
EPR spectroscopies; complex 7 was also
could not be recycled, owing to signifi-
II
ligand,
2,2'-R -bipyridine
(R ˆ -
cant ligand dissociation. The Cu com-
f
f
II
(
CH ) C F ) 2, as well as the new Cu
plex 4, with the ligand 1, provide the
oxidation of 4-nitrobenzyl alcohol to
4-nitrobenzaldehyde under single-phase
FBC conditions at 908C with TEMPO
(2,2,6,6-tetramethylpiperidinyl-1-oxy)
2
3
8
17
1
19
1
fluoroponytailed carboxylate synthon
characterized by H and F{ H} NMR
spectroscopy. Complexes 4 and 5, as well
as 6 and 7 generated in situ, were
evaluated as precatalysts for hydrocar-
complex [Cu(C F (CH ) CO ) ] 3, will
8
17
2
2
2 2
be addressed. Moreover, the reaction of
previously described ligands, 1,4,7-Rf-
TACN 1, or 2,2'-R -bipyridine 2 with 3
and O ; the precatalyst 4, can be utilized
f
2
afforded new perfluoroheptane-soluble
for an additional four catalytic cycles
without loss of activity. Plausible mech-
anisms concerning these FBC oxidation
reactions will be discussed.
Keywords:
copper ¥ mechanisms ¥ N ligands ¥
oxidation
biphasic catalysis
¥
II
Cu
complexes,
[Cu(C F (CH ) -
8 17 2 2
CO ) (R -tacn)] 4 and [Cu(C F (CH ) -
2
2
f
8
17
2 2
CO ) (R -bpy)] 5, respectively. The re-
2
2
f
Introduction
stone of global industrial research for the preparation of a
variety of products for world-wide consumption.[ Homoge-
neous catalytic oxidation is more selective toward the above-
mentioned products and, in addition, the reactions are carried
out at lower temperatures than those used in heterogeneous
catalytic oxidation. Thus, the development of a homogeneous
oxidation process for alkane and alkene functionalization has
been an important goal in numerous industrial and academic
1]
The efficient, catalytic synthesis of alcohols, aldehydes,
ketones, and epoxides from alkanes and alkenes is a corner-
[
a] Dr. M. Contel, Prof. M. Laguna, C. Izuel, P. R. Villuendas
Instituto de Ciencia de Materiales de Arag o¬ n
Departamento de QuÌmica Inorg a¬ nica
Universidad de Zaragoza-C.S.I.C, 50009 Zaragoza (Spain)
Fax : (‡)34-976-761187
[1]
laboratories. In order for this process to be industrially
practicable in view of economical and environmental con-
cerns, new and innovative approaches for the separation of
the homogeneous catalyst from the oxidation products are
still needed. In recent years, Horv a¬ th and R a¬ bai introduced
the concept of fluorous biphasic catalysis (FBC),[ and
demonstrated that by using fluorocarbon solvents and mod-
ifying the homogeneous catalyst×s ligand structure with long-
chain perfluoroalkane derivatives (fluoroponytails) to make
the precatalyst soluble in the fluorocarbon phase, the products
of the catalytic reaction would be soluble in a separate solvent
E-mail: mcontel@posta.unizar.es
mlaguna@posta.unizar.es
[
b] Prof. P. J. Alonso
Instituto de Ciencia de Materiales de Arag o¬ n, Departamento de FÌsica
de la Materia Condensada
Universidad de Zaragoza-C.S.I.C, 50009 Zaragoza (Spain)
2]
[
c] Prof. R. H. Fish
Lawrence Berkeley National Laboratory
University of California
Berkeley, CA 94720 (USA)
E-mail: rhfish@lbl.gov
4168
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304771
Chem. Eur. J. 2003, 9, 4168 ± 4178

4
168±4178
phase that is immiscible in the fluorocarbon phase; by
removing this second phase, the products are easily separated
from the precatalyst. It should also be noted that the
miscibility of perfluoroalkanes and perfluoroethers with the
products of alkane and alkene oxidation, such as alcohols,
aldehydes, ketones, and epoxides, is extremely low or
nonexistent.[ This latter critical characteristic, and the fact
that perfluoroalkanes and perfluoroethers are very interesting
for their nonpolarity and extremely low intermolecular forces,
makes them ideal solvents for homogeneous biphasic catalytic
processes. The FBC concept has now been amply demon-
strated with the conversion of many examples of classic
In this paper, we report on the synthesis and character-
ization of new Cu and Cu complexes with the previously
I
II
described fluorocarbon-soluble 1,4,7-triazacyclononane (R -
f
TACN)[ 1 and bipyridine (R -bpy)
7a]
[10d]
2 ligands that contain
f
the fluoroponytail, C F -(CH ) -. We demonstrate their
8
17
2 3
catalytic activity with hydrocarbons and olefins under FBC
oxidation conditions, with tert-butyl hydroperoxide (TBHP)
and the necessary oxygen gas, at room temperature. Further-
more, the oxidation of alcohols to aldehydes, under FBC
conditions, has also been performed with TEMPO (2,2,6,6-
tetramethylpiperidinyl-1-oxy) and oxygen gas at 908C, in-
volving one phase during the reaction and two phases on
cooling to room temperature. We also provide EPR results to
ascertain the fate of the precatalyst, while presenting plausible
mechanisms for all of these transformations under FBC
conditions.
3]
[4]
organic catalytic reactions to the FBC paradigm, amongst
them oxidation reactions.[
5±11]
However, there is still an urgent
need to synthesize and fully characterize new fluoropony-
tailed ligands and their metal complexes for unequivocal
structural identification, and in order to understand the
mechanisms of the oxidation chemistry under FBC reaction
conditions.[
7d]
Results and Discussion
I
II
Synthesis and characterization of R -Cu and R -Cu com-
Abstract in Spanish: En esta contribuci o¬ n en cat a¬ lisis bif a¬ sica
fluorada (FBC) presentamos la sÌntesis y caracterizaci o¬ n de
nuevos complejos de cobre y definimos su papel, como
precatalizadores, en la oxidaci o¬ n FBC de hidrocarburos,
f
f
I
plexes: In the field of FBC, several Cu derivatives with
fluoroponytailed amine ligands have been previously synthe-
sized. These complexes were prepared in situ from either
CuCl or CuBr ¥ SMe and the corresponding ligands, R -
olefinas y alcoholes. AsÌ, se describe la preparacio
¬
n y
2
f
macrocycles[ or R -bipyridine,
7b]
[10d]
and were evaluated in the
caracterizacio
¬
n del ligando fluorado 2,2'-R -bipiridina 2
f
f
[7b]
oxidation of alkanes and alkenes,
or the oxidation of
(
R ˆ -(CH ) C F ), que habÌa sido publicado anteriormente
f
2
3
8
17
alcohols,[ respectively. Several other Cu complexes contain-
ing tri- or tetradentate amine ligands have been prepared in
situ in a similar manner, and have been used as precatalysts
8]
I
pero sin datos experimentales sobre sÌntesis o caracterizaci o¬ n, y
la preparaci o¬ n del compuesto sint o¬ n carboxilato
[
Cu(C F (CH ) CO ) ], 3. La reacci o¬ n de los ligandos
8 17 2 2 2 2
[12]
for the cyclization of unsaturated esters or living radical
polymerizations[ under FBC conditions. However, these
complexes, usually generated in situ, were not isolated or
characterized properly. The synthesis consisted of the addition
previamente descritos 1,4,7-R -TACN, 1, o 2,2'-R -bipyridina,
f
f
13]
II
2
con 3 permiten la formaci o¬ n de compuestos de Cu solubles
en perfluoroheptano como [Cu(C F (CH ) CO ) (R -tacn)],
8
17
2
2
2
2
f
4
, y [Cu(C F (CH ) CO ) (R -bpy)], 5, respectivamente. La
8 17 2 2 2 2 f
I
of an excess of the Cu salt to a perfluorocarbon solution of
the fluoroponytailed amine ligand, and the new fluorocarbon-
reacci o¬ n de 1,4,7-R -TACN, 1, con [Cu(CH CN) ]PF o [CuCl]
f
3
4
6
I
genera nuevos compuestos de Cu que pueden ser aislados y
[
7b]
soluble compound was characterized in some cases by the
UV/Vis spectrum of the resulting green solution. We believe
that not only the green color of the perfluorocarbon solutions,
but also the UV/Vis data, tentatively demonstrated that the
oxidation of the fluoroponytailed Cu complexes to Cu
derivatives was occurring. Thus, we wanted to synthesize
caracterizados completamente como [Cu(R -tacn)X']X, con
f
X ˆ PF , 6, o X' ˆ Cl, 7 (soluble en perfluoroheptano). Los
6
I
II
compuestos de Cu y Cu , 4 ± 7, se han caracterizado por
an a¬ lisis elemental, espectrometrÌa de masas, IR, UV/Vis
I
II
(
reflectancia difusa) y espectroscopÌa de EPR. El derivado 7
ha sido caracterizado tambi e¬ n por espectroscopÌa de RMN de
I
II
1
19
1
and isolate not only the Cu , but also Cu fluoroponytailed
H y F{ H}. Los compuestos 4 y 5, asi como los derivados 6 y
, generados in situ, se evaluaron como precatalizadores en la
I
II
complexes, as both Cu and Cu salts and complexes have
previously displayed catalytic activity in the homogeneous
oxidation of hydrocarbons, olefins, and alcohols.[
The fluoroponytailed ligands employed in this study were
the previously described fluorocarbon soluble 1,4,7-triazacy-
7
funcionalizaci o¬ n de hidrocarburos y olefinas. Las reacciones
de oxidaci o¬ n de estos sustratos, en presencia de los oxidantes
necesarios, hidroper o¬ xido de tert-butilo (TBHP) y oxÌgeno gas,
14]
I
se consiguen en condiciones FBC para 5, 7 y sales de Cu con el
[7a]
[10d]
clononane (R -TACN),
1 and bipyridine (R -bpy),
2
f
f
ligando 2. No obstante, los compuestos con 2 no pudieron
ligands (Figure 1) that contained the C F -(CH ) - group. The
reciclarse debido a una gran disociacio
¬
n del ligando. El
8
17
2 3
II
convenient three carbon spacer in the fluoroponytail was
necessary, not only to insulate the amine from the powerful
electron-withdrawing effect of the perfluoroalkyl group, but
also to avoid a facile elimination reaction of HI when
synthesizing fluoroponytailed amine ligands by alkylation
reactions with C F (CH ) I. The latter compound, along with
ligand 1, have been previously reported by Fish et al.[
Although ligand 2 has been prepared before, its synthesis
was not completely described, and its characterization was not
compuesto de Cu , 4, con el ligando R -TACN, 1, permite la
f
oxidaci o¬ n de alcoholes 4-nitro-bencÌlico a 4-nitro-benzaldehi-
do en condiciones FBC, monofase a 908C, con TEMPO
(
2,2,6,6-tetrametilpiperidina-N-oxilo) y O ; el precatalizador,
2
4, puede ser utilizado durante cuatro ciclos catalÌticos adicio-
nales, sin pe
¬rdida de actividad. Se discutir a¬ n mecanismos
8
17
2 3
7a,d]
posibles para estas reacciones de oxidaci o¬ n en condiciones
FBC.
Chem. Eur. J. 2003, 9, 4168 ± 4178
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4169

M. Contel, M. Laguna, R. H. Fish et al.
FULL PAPER
II
[16]
be directly attributed to a Cu complex. The stoichiometry
of 4 was confirmed by elemental analysis and its LSIS-MS
spectrum displayed a signal at m/z 2063 (75%), which could
‡
be assigned to the [M À {C F (CH ) CO }] ion. The IR
8
17
2
2
2
spectrum had bands corresponding to the carboxylate group,
in addition to those that can be assigned to the R -TACN
f
ligand. It appears that the carboxylate group in 4 (tentative
structure in which R ˆ C F -(CH ) - in Figure 3) was in a
Figure 1. Perfluoroponytailed amine ligands 1 and 2.
f
8
17
2 3
monodentate rather a chelating or bridging coordination
provided. Thus, we have also described the synthesis and
characterization of 2.
À1
mode, resulting in the bigger difference (n ˆ 229 cm )
observed between the asymmetric and symmetric CÀO
In order to insure the fluorocarbon solubility of the CuII
complexes of ligands 1 and 2, we appended fluoroponytailed
groups to the metal complex we were going to utilize as a
synthon, as previously discovered by Fish et al., since metal
complexes with multiple charges and counter anions, such as
À1
stretches (n ˆ 1632 and 1403 cm , respectively) in its IR
spectrum.
À
À
À
ClO , PF , or even Cl would be very difficult to solubilize
4
6
in perfluorocarbons.[
7a,d]
Thus, the new Cu complex
II
[
Cu(C F (CH ) CO ) ] 3, was synthesized by reaction of
8
17
2
2
2 2
CuSO ¥ 5H O in acetone with two equivalents of the triethy-
4
2
lammonium salt of the 3-R -alkylpropionic acid
f
C F (CH ) CO H; elemental analysis of 3 revealed a metal/
8
17
2
2
2
carboxylate ratio of 1:2. Complex 3 was found not to be
soluble in perfluorocarbons, soluble in trifluorotoluene, and
only partly soluble in CH Cl . It further appears that the
Figure 3. Tentative structure of compound 4.
2
2
carboxylate group is chelating rather than being in a mono-
dentate coordination mode, as suggested by the small differ-
Compound 5 was not soluble in cold perfluoroheptane, but
it could be solubilized in hot perfluoroheptane (above 408C).
Again, elemental analysis and LSIMS-MS confirmed the
composition of the compound. The UV/Vis spectrum, re-
corded in perfluoroheptane, showed one absorption band at
l ˆ 246 nm, which can be assigned to the fluorous ligand, but
its diffuse reflectance spectrum (solid state) showed one
intense absorption band at 680 nm in accordance with values
À1
ence (n ˆ 153 cm ) observed between the asymmetric and
À1
symmetric CÀO stretches (n ˆ 1573 and 1420 cm , respec-
tively) in its IR spectrum.[ Its lack of solubility prevented us
from obtaining its UV/Vis spectrum in perfluoroheptane or
CH Cl solutions. However, we were able to record the UV/
15]
2
2
Vis spectrum in the solid state by using diffuse reflectance
measurements. The intense absorption band displayed by 3
(
6
tentative structure given in Figure 2, R ˆ C F -(CH ) -) at
f
8
17
2
3
II
[16]
obtained for Cu compounds. In contrast to compound 4, it
appears that the carboxylate group in 5 is in a bidentate rather
than a monodentate coordination mode, as evidenced by the
II
66 nm was in accordance with values found for other Cu
[16]
complexes.
À1
smaller difference (n ˆ 173 cm ) observed between the
asymmetric and symmetric CÀO stretches (nˆ 1569 and
À1
1
394 cm , respectively) in its IR spectrum. This is in
accordance with the bidentate coordination mode of the Rf-
bpy, which would allow the R -carboxylate groups to bind in a
chelating mode to the Cu center in 5 (Figure 4), while in 4 the
f
Figure 2. Tentative structure of complex 3.
II
three N-donor atoms of the R -TACN appear to limit the
carboxylate coordination to the Cu to a monodentate mode.
f
II
The addition of ligands 1 or 2 to perfluoroheptane
suspensions of 3 allowed the total (for 1) or partial (for 2)
solubilization of the Cu complex at room temperature, giving
II
deep green solutions. In order to isolate these compounds, we
added ligands 1 and 2 to blue solutions of 3 in CH Cl , which
2
2
almost immediately produced green solids that could be
isolated and characterized as [Cu(C F (CH ) CO ) (R -tacn)]
8
17
2
2
2
2
f
4
and [Cu(C F (CH ) CO ) (R -bpy)] 5. Unfortunately, none
8 17 2 2 2 2 f
of the Cu complexes reported gave single crystals suitable for
X-ray analysis.
Figure 4. Tentative structure of compound 5.
Complex 4 was found to be totally soluble in perfluor-
oheptane, while its UV/Vis spectrum in this solvent had only
one absorption band at 272 nm, which could be assigned to the
fluorous ligand. However, its diffuse reflectance spectrum
II
After synthesizing these paramagnetic Cu fluorous com-
I
pounds, we attempted the preparation of Cu complexes with
the more fluorocarbon soluble ligand 1, at room temperature.
(
solid state) showed two bands at 1013 and 700 nm that could
In this case, we did not use R -carboxylate synthon 3, as it had
f
4170
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4168 ± 4178

Fluorous Biphasic Catalysis
4168±4178
I
been previously shown that the in situ synthesis of Cu -
I
fluorous amine derivatives could be achieved by using the Cu
salt and the corresponding fluoroponytailed ligands. The
addition of 1 to [Cu(CH CN) ]PF or CuCl afforded an off-
3
4
6
white ([Cu(R -tacn)]PF 6) and a white ([Cu(R -tacn)Cl] 7)
f
6
f
solid in moderate (36%), and high (78%) yields, respectively
Figure 5). Satisfactory elemental analyses were obtained for
(
Figure 6. X-band EPR spectra of powdered samples: a) complex 3;
b) complex 4; c) complex 5; d) compound 7; e) isolated complex from 2
and [CuCl].
Figure 5. Tentative structures of 6 and 7.
both compounds, and their LSIMS-MS spectra had signals
that corresponded to the monocationic complexes, [Cu(R -
tacn)] , at m/z 1573 for 6 (60%) and 7 (88%). Complex 6 was
only soluble in trifluorotoluene; however, compound 7 was
f
‡
II
of the Cu complexes, 3 (spectrum a), 4 (spectrum b), and 5
I
(spectrum c), as well as the putative Cu complex from CuCl
totally soluble in cold perfluoroheptane and partly soluble in
and 2 (spectrum e), there were strong indications of inter-
molecular exchange reactions. This was particularly evident
1
19
1
[17]
CHCl , allowing for its characterization by H and F{ H}
3
NMR spectroscopy in CDCl . UV/Vis data for 7 were also
in the EPR spectrum of compound 3 (spectrum a), and the
spectrum recorded with the solid from the reaction of CuCl
and 2 (spectrum e). In these EPR spectra, similar features
3
obtained in perfluoroheptane, revealing two absorption bands
at 212 and 260 nm. However, its solubilility in CH Cl , or even
2
2
in CHCl , was not enough to perform a cyclic voltammetric
appeared at g
1
ˆ 2, g
2
ˆ 2.06 (spectrum a), and g
1
ˆ 2.15, g ˆ
2
3
II
study. In addition, the EPR spectra of 6 and 7 were silent, as
2.06 (spectrum e), suggesting a mixture of Cu complexes. In
I
I
expected for Cu diamagnetic complexes. We also found that 6
accordance with the nature of 7, a Cu complex, no indication
II
and 7 were light sensitive, and that they could be easily
of a Cu complex was detected in its silent EPR spectrum
II
oxidized to Cu species (especially 6) after several hours at
(spectrum d).
room temperature. However, they could also be kept for
longer periods of time at À208C under argon, while being
protected from direct light exposure.
Functionalization of alkenes, alkanes, and alcohols using
‡
copper(ii) complexes and in-situ-generated [Cu(R
-ligand)]
f
We mentioned in the introduction that the combination of
derivatives as precatalysts
CuBr¥ SMe and 2 had produced a deep green solution in
2
[8]
perfluoroheptane, but the characterization of this putative
Oxidation of alkenes and alkanes: The FBC oxidation results
for substrates cyclohexene, cyclohexane, and toluene in the
presence of fluoroponytailed copper complexes are presented
in Table 1. All of the experiments were carried out under
biphasic conditions in perfluoroheptane; the upper phase was
I
Cu complex had not been reported thus far. When we
attempted the reaction between CuCl and 2 in trifluoroto-
luene in order to synthesize the compound analogous to 7, we
immediately obtained a deep green solution, from which a
deep green solid was isolated. Although its LSIMS-MS
spectrum displayed a prominent signal at m/z 1167 (100%),
II
the substrate itself (Scheme 1). Cu complexes 4 and 5 were
dissolved in perfluoroheptane; however, we found it more
‡
which can be assigned to the species [M À Cl] , its elemental
analysis did not fit with a stoichiometry of a complex like
[
Cu(R -bpy)Cl]. This complex was partly soluble in trifluor-
f
otoluene and in cold perfluoroheptane (it was soluble in hot
perfluoroheptane) and thus we were not able to obtain its
NMR or UV/Vis spectra in perfluoroheptane solution. How-
ever, its green color, as well as its EPR spectrum, indicated
II
that a partial oxidation to a Cu complex had occurred.
The EPR spectra of solid, powdered samples of compounds
3
2
, 4, 5, and 7, as well as the solid from the reaction of CuCl and
, were measured at room temperature. Figure 6 shows the
EPR spectra of all the samples studied (spectra a ± e). In all
I
cases except for the Cu compound 7 (spectrum d), the EPR
II
spectra had signals from several Cu complexes. In the cases
Scheme 1. Cyclohexene FBC oxidation with complex 7 as precatalyst.
Chem. Eur. J. 2003, 9, 4168 ± 4178
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4171

M. Contel, M. Laguna, R. H. Fish et al.
FULL PAPER
I
convenient to generate the Cu derivatives 6 and 7 in situ for
catalytic activity decreasing to around 75% in the second run
and to half of the initial value in the third experiment. With
catalytic experiments by the addition of excess (2:1)
[
Cu(CH CN) ]PF or CuCl to the perfluoroheptane solutions
the Cu/R -CYCLAM system, the catalytic activity in a second
3
4
6
f
of 1, with subsequent filtration through celite to remove any
undesired excess of Cu salts. The molar ratio between the
run decreased to half of its initial value (yield 248%, TON
I
155), and the authors mentioned that the hydrocarbon layer
turned increasingly blue.[
7b]
Moreover, in a second run
substrate, the oxidant, and the catalyst was S/O/C ˆ
[7b]
5000:62.5:1 in all experiments.
[Mn(C F (CH ) CO ) (R -tacn)] afforded a yield of 400%
8 17 2 2 2 2 f
(vs 650% in the first run).[
We believed that 6 (formed in situ from [Cu(CH CN) ]PF
6
7a,d]
The oxidation of alkenes and alkanes always provided a
mixture of alcohols and ketones. As previously described,[
Table 1 demonstrates that the olefin with allylic hydrogens,
cyclohexene, provides the highest yield of oxidation products
with TBHP/O . In the absence of O or TBHP, only negligible
7a,b,d]
3
4
and 1) was not an efficient catalyst (entry 2) under FBC
conditions, owing to its lack of solubility in perfluoroheptane,
as shown with the isolated compound. The reaction of
cyclohexene with 7 (formed in situ) in the presence of
oxidants (H O /O ) was also studied (entry 6) but no oxida-
2
2
amounts of cyclohexenol and cyclohexenone were detected,
which is fully indicative of an autoxidation reaction, that is,
both were necessary for oxidation to proceed (Table 1,
entries 4 and 5).[
2
2
2
II
tion products were formed. Although complexes of Cu and
7a,d]
I
Cu with the R -bpy ligand 2 (entries 7 ± 11) did catalyze the
f
All of the complexes derived from the ligand 1, showed
catalytic activity (Table 1, entries 1 ± 3, 7 ± 9), but only 7
oxidation of cyclohexene (TBHP/O ), they could not be
2
recycled (see complex 5, entries 10 and 11; CuCl and ligand 2
(entries 13 and 14). We found that after the reaction had taken
place, the fluorous phase turned colorless, while the organic
(
prepared in situ from CuCl and R -TACN) afforded yields
f
or TON that could be compared with those previously
reported for CuCl and the fluoroponytailed macrocyclic
II
layer became blue, indicating that Cu ions were present in
ligands derived from tetraazacyclotetradecane (R -CY-
this upper layer. We believe that the different behavior
between the compounds with R -TACN and R -bpy ligands
f
[7b]
CLAM)
Co).[
or [M(C F (CH ) CO ) (R -tacn)] (M ˆ Mn,
8
17
2
2
2
2
f
f
f
7a,d]
The first system afforded yields of 514% and TON
might be due to the different coordination modes of these
I
II
of 320 for the oxidation of cyclohexene to cyclohexenol and
ligands to the Cu and Cu centers, which may be related to
the steric and electronic properties of the catalytic active sites,
as well as their stability constants. In all oxidation experi-
ments, it was observed that the TBHP employed as a free
[7b]
cyclohexenone after eight hours. The latter complexes gave
yields of 650% and TON of 131 (M ˆ Mn) after three hours or
II
[7a,d]
7
50%, TON ˆ 156 (M ˆ Co ) after 20 hours.
We found
that for the same substrate, yields of 136% (with a TON of 85)
radical initiator was converted to tert-butyl alcohol and
acetone, indicative of tert-butyloxy radical formation.[
7a,d]
We
were obtained after eight hours (entry 3) for 7, while after
24 hours (same conditions, entry 7) the yields and TON were
also attempted the oxidation of cyclohexane and toluene with
increased to 425% and 276, respectively. More importantly,
we were able to recover the fluorous phase and recycle it for a
second (entry 8) or a third (entry 9) catalytic experiment. The
overall yield was then 938% and the TON ˆ 609, with the
the best catalytic system, 7, generated in situ, and, as described
II
II
for the [M(C F (CH ) CO ) (R -tacn)] {M ˆ Mn , Co } sys-
8
17
2
2
2
2
f
tems,[
7a,d]
the yields of oxidation products were much lower
than for cyclohexene (entries 15 and 16, Table 1).
Table 1. Alkene and alkane functionalization with fluorous CuII and Cu complexes under FBC conditions.[a]
I
Entry Substrate
Oxidant
Precatalyst
t [h] Substrate
Selectivity
Ketone/
aldehyde [%] yield [%]
Total
TON[e]
converted[ [mmol] alcohol [%]
b]
[c]
[d]
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
2
2
2
4
8
8
8
8
8
175
86
678
12
13
8
2210
1601
1068
1023
13
681
590
62
3.2
1.3
27
±
±
96.8
98.7
73
100
100
100
35
17
136
2.2
3
22
11
85
[Cu(CH
3
CN)
4
]PF
6
‡ R
f
-TACN (6)
CuCl ‡ R
f
-TACN (7)
cyclohexene
cyclohexene TBHP
cyclohexene
O
2
7
7
7
7
1.5
1.6
0.9
276
200
133
128
8
85
74
8
5
H
2
O
2
8
24
±
1.5
425
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexene TBHP/O
cyclohexane TBHP/O
2
2
2
2
2
2
2
2
2
2
21.8
29.3
18.6
15.6
±
12.1
13.8
±
78.2
fluorous recovered phase from entry 7 24
fluorous recovered phase from entry 8 24
5
fluorous recovered phase from entry 10
[Cu(CH
70.7
81.4
84.4
100
87.9
86.2
100
56.7
51.6
308
205
204
12
136
113
12
1
1
1
1
1
1
1
8
8
8
8
8
3
CN)
4
]PF
6
‡ R
f
-bipy
CuCl ‡ R
f
-bipy
fluorous recovered phase from entry 13
7
7
24
24
39
87
43.3
48.4
7.5
17
toluene
TBHP/O
13
[
a] Conditions: Hydrocarbon phase: Substrate (S) ˆ 40 mmol; o-dichlorobenzene (internal standard) ˆ 2 mmol; 70% tBuOOH (O) ˆ 500 mmol. Fluoro-
carbon phase: precatalyst (8 mmol, see experimental section) in perfluoroheptane (2 mL). Molar ratio S/O/C 5000:62.5:1. Tˆ 298 K. Reactions were carried
out in the dark in the presence of molecular oxygen. Products were identified by comparison with authentic samples by GC. [b] Determined by GC using the
internal standard method. [c] mmol product/mmol converted substrate. [d] Based on the added tBuOOH. [e] mmol converted substrate/mmol precatalyst.
4172
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4168 ± 4178

Fluorous Biphasic Catalysis
4168±4178
Oxidation of alcohols: The oxidation of alcohols to aldehydes
or ketones under FBC conditions has been performed under
an oxygen atmosphere both by Knochel et al. with a catalytic
Table 2. Oxidation of 4-nitrobenzyl alcohol to 4-nitrobenzaldehyde under
FBC conditions using 4 as precatalyst.[
a]
t [h] Conversion TON[c]
Entry Catalyst
% mol % mol
catalyst TEMPO
[%][
b]
system consisting of CuBr¥ SMe and 2, prepared in situ, at
2
8]
[9]
9
08C[ and by Uemura et al. with Pd(OAc) and fluoropo-
2
1
2
3
4
5
6
7
8
9
0
4
2
±
3.5
3.5
±
±
±
±
±
3.4
3.4
6
6
6
6
6
6
6
4
4
4
8
8
8
8
8
8
8
63
61
65
96
94
92
90
92
51
8
31
30
19
27
27
26
26
26
14
2
nytailed pyridines prepared in situ at 808C. In the first
example, the initiator was TEMPO and in the second case,
they used 3 ä molecular sieves. However, while the Cu system
used a small amount of fluorous ligand (2 mol%),[ the Pd
RFP[d] from entry 1
4
4
RFP from entry 4
RFP from entry 5
RFP from entry 6
RFP from entry 7
RFP from entry 8
RFP from entry 9
8]
system needed a much larger amount of fluorous ligand
9]
(
20 mol%).[ Benzylic and allylic alcohols reacted faster than
aliphatic alcohols for both systems and, in general, high yields
were obtained.
1
±
6
The recyclability of the Cu system was demonstrated for the
oxidation of 4-nitrobenzyl alcohol to 4-nitrobenzaldehyde.
The reaction was performed eight times without a major loss
of catalytic activity. In the Pd system, the recyclability of the
precatalyst was demonstrated for six runs. It should also be
noted that no mechanistic information was available for this
transformation nor for the role of TEMPO in this FBC
oxidation reaction. More importantly, we believed that the
[a] Conditions: hydrocarbon phase: 4-nitrobenzyl alcohol, TEMPO; fluo-
rous phase: 4 in perfluoroheptane (2 mL). Reactions were carried out at
1
9
08C in the presence of molecular oxygen. [b] Measured by H NMR.
[
c] mmol converted substrate/mmol catalyst. [d] Recovered fluorous phase.
hours, conversion to the aldehyde was at 63% (entry 1) and
1% (entry 2) in the first and second runs, respectively,
6
showing that catalytic activity did not change during the
second run. This preliminary experiment helped to optimize
the FBC reaction conditions (entries 3 ± 5). We found it was
necessary to increase the amount of catalyst to 3.5 mol% and
the reaction time from 4 to 8 hours to obtain a 96%
conversion to the benzaldehyde product.
I
II
reported reaction between Cu salts and 2 also gave Cu
complexes during the reaction, while the oxidation of alcohols
I
II
with TEMPO/O with Cu and Cu salts in a homogeneous
2
II
system also suggested that Cu ions were responsible for
initiating the oxidation. We then decided to study the Cu
complex 4 under single-phase FBC reaction conditions[4f] at
[18]
II
9
4
08C with 4-nitrobenzyl alcohol in the FBC oxidation to
-nitrobenzaldehyde (Scheme 2).
‡
Autoxidation mechanism and the fate of the [Cu(R -tacn)]
f
precatalyst in the oxidation of alkenes and alkanes
The TBHP/O initiated oxidation reactions presented here
2
are in agreement with an autoxidation mechanism involving
.
.À
alkoxy (RO ) or alkoxyperoxy (ROO ) radicals, in which the
.
.
reaction is initiated by tBuO or tBuO radicals produced
from redox reactions (Haber± Weiss process with Cu /Cu ).
This mechanism is similar to that proposed by Fish et al. with
2
I
II
[7a,d]
the [Mn(C F (CH ) CO ) (R -tacn)] catalytic system.
It
8
17
2
2
2
2
f
has been previously reported that in the case of the substrate
cyclohexene, the allylic radical that forms is then trapped by
9
À1 À1
O (k > 1 Â 10 m s ) to provide cyclohexenylperoxy radi-
2
À1
cals, which are able (ROOÀH ˆ 90 kcalmol ) to homolyti-
cally remove a hydrogen from benzylic or allylic CÀH bonds
À1
(
85 kcalmol ), and hence propagate the radical reac-
[19, 20]
tions.
The secondary cyclohexenyl hydroperoxide that is
formed then decomposes catalytically in the presence of the
Cu(R -tacn)Cl] precatalyst to give the alcohol and the ketone
[
f
products. Scheme 3 defines the initiation, propagation, termi-
nation, and product-forming steps for the sequence of
Scheme 2. FBC oxidation of 4-nitrobenzylacohol with precatalyst, com-
plex 4.
reactions [Eqs. (1) ± (8)].[
7a,d]
Moreover, the lower yields
observed in the case of cyclohexane are consistent with the
À1 [19]
stronger CÀH bond (95 kcalmol ), which implicates the
The initial FBC mode at room temperature (Scheme 2)
includes 4 in the lower perfluoroheptane layer, while the
substrate alcohol, chlorobenzene, and TEMPO were in the
upper organic layer (Table 2). At 908C, in a now single
homogeneous phase, under an oxygen atmosphere for four
cyclohexylperoxy radical and thus causes a lowering of the
rate of the propagation step. The chain termination step
presumably comes mainly from the coupling of two cyclo-
hexenylperoxy radicals to give alcohol, ketone, and O
2
[Russell-type mechanism, Scheme 3, Eq. (7)].[
7a,d]
Chem. Eur. J. 2003, 9, 4168 ± 4178
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4173

M. Contel, M. Laguna, R. H. Fish et al.
FULL PAPER
Figure 7. X-band EPR spectra of the perfluoroheptane layer during the
course of the oxidation of cyclohexene with TBHP/O , 7 as precatalyst,
2
measured at liquid nitrogen temperature. a) 2 h; b) 3.5 h; c) 6 h; d) 11 h;
e) 23 h.
after 23 hours (spectrum e). After 3.5 h (spectrum b), the
II
EPR spectrum showed some contribution of Cu , but the
spectrum was not the same as the one after six hours. This fact
II
could mean that a Cu complex was formed after addition of
I
TBHP/O . What was clear was that it takes ꢀ3 h for the Cu
2
II
precatalyst 7 to be oxidized to the Cu complex, which is
consistent with a Haber± Weiss mechanism.[
7a,d]
The catalytic results (Table 1, entry 1) have shown that the
CuII complex with the R -TACN ligand, 4, catalyzes the
f
oxidation of cyclohexene, but to a smaller extent than 7, or the
Cu complex 5. The reason for which 7 was a recoverable
Scheme 3. Haber± Weiss mechanism for the oxidation of cyclohexene with
precatalyst, complex 7.
II
catalyst, while its catalytic activity in the second and third run
decreases to 75 and 50%, respectively, could possibly be the
II
In order to determine the redox chemistry occurring with
formation of a Cu complex that could be easily reduced to a
I
the starting complex, [Cu(R -tacn)Cl] 7, the best catalytic
Cu complex, as designated by the Haber± Weiss process. In
f
II
system we found for the oxidation of cyclohexene under FBC
conditions, we carried out the following EPR experiments
contrast, the low catalytic activity of the Cu catalytic system
with 4 might be a consequence of not being able to efficiently
I
(
Figure 7, spectra a ± e). Compound 7 was dissolved in per-
generate the Cu complex [Scheme 3, Eq. (2)], because of a
7a,d]
fluoroheptane and then cyclohexene, o-dichlorobenzene (as
internal standard), and TBHP were added. The reaction
started in the presence of O2 and an aliquot from the
perfluoroheptane phase was removed and immediately frozen
at liquid nitrogen temperature (LNT) in an EPR tube, after
two hours of reaction. The EPR spectrum (measured at 77 K)
is shown in Figure 7 (spectrum a). There was no indication of
presumed kinetic effect [Scheme 3, Eq. (2)].[
The fact that
Equation (1) (Scheme 3) involves the generation of the tert-
I
II
butoxy radical implicates the Cu /Cu redox couple.
II
More importantly, the behavior of the Cu complex 5, with
ligand 2, was quite different (Table 1, entries 10 and 11). In
I
this example, the reduction to the Cu complex appears to be
more facile, possibly because of the different electronic
II
a Cu signal in this initial spectrum. This was in accordance
properties of the fluoroponytailed ligand 2. Unfortunately,
I
I
with the oxidation state of the precatalyst 7, whose Cu center
we were not able to isolate the corresponding Cu compound
had also been previously confirmed by an EPR spectrum of a
solid sample (Figure 6, spectrum d).
The reaction was further followed by EPR spectroscopy
analogous to 7. We also determined that 5 was not recoverable
under FBC conditions and that after the reaction had taken
II
place, the free ligand 2, as well as Cu ions, could be observed
taking aliquots at different reaction times and after six hours
in the organic phase, as analyzed by mass spectrometry.
II
(
spectrum c), a Cu signal was observed. It could be described
by an axial spin-Hamiltonian with a g tensor, g ˆ 2.26(1) and
Mechanism and fate of the [Cu(C F (CH ) CO ) (R -tacn)]
(4) precatalyst in the oxidation of 4-nitrobenzyl alcohol
k
8
17
2
2
2
2
f
II
g ˆ 2.05(1), as well as a Cu nuclei hyperfine interaction
?
given by A ˆ 520(5) MHz and A <50 MHz. These EPR
k
?
II
II
spectra are typical of a nearly square planar Cu complex.
We have shown that an isolated R -Cu complex 4 can be a
f
This EPR signal remains after 11 hours (spectrum d), and
recyclable and efficient precatalyst for the oxidation of
4174
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4168 ± 4178

Fluorous Biphasic Catalysis
4168±4178
alcohols under FBC conditions at 908C with TEMPO/O . The
2
mechanism of this reaction was thus of interest. Moreover,
several years ago Semmelhack et al.[ reported on the
17]
II
oxidation of alcohols to aldehydes with oxygen and Cu ions,
mediated by a nitrosonium ion that was generated from the
TEMPO radical. The catalytic cycle proposed by these
II
authors is shown in Scheme 4. The Cu ions are reduced to
Figure 8. X-band EPR spectra of the perfluoroheptane layer during the
course of the oxidation of 4-nitrobenzyl alcohol with TEMPO/O
2
, at 908C,
4
as precatalyst measured at liquid nitrogen. a) 0.5 h; b) 3.5 h. c) 6 h; d) 8 h.
7
7 K. The spectrum of the reaction mixture also has a CuII
signal, with g ˆ 2.26(1) and g ˆ 2.06(1), A ˆ 520(5) MHz
k
?
k
[18]
and A <50 MHz.
After 3.5 h, both signals decrease
?
(
(
spectrum b), and after 6 h the signals have decreased further
spectrum c). After 8 h, we observed a recovery of the signal
Scheme 4. Semmelhack mechanism for the oxidation of alcohols to
II
corresponding to the Cu species (spectrum d). An EPR
spectrum of the upper phase (chlorobenzene layer), after the
reaction had taken place, showed a EPR signal that can be
assigned to the TEMPO radical. Thus, TEMPO appears to be
regenerated as was proposed in Scheme 4. It is clear from the
silent EPR spectrum after 3.5 h (Figure 8, spectrum c) that the
initial Cu complex 4 is reduced to the Cu complex and is
then regenerated to a Cu complex (spectrum d). This is in
accordance with the mechanism proposed by Semmelhack
et al. (Scheme 4) and with the catalytic results reported in
this paper (Table 2). After 4 h, a 65% yield of aldehyde was
obtained; however, by leaving the reaction for longer periods
of time, a nearly 100% yield of aldehyde (Table 2) was
formed, concomitant with a full recovery of 4. We believe that
the TEMPO was also regenerated, but probably ended up in
the organic phase where it was destroyed during the work-up
procedure; TEMPO was then replenished during the next
catalytic run.
We also wish to propose an alternative mechanism to
Semmelhack×s, described above for the oxidation of alcohols
to aldehydes, that can reasonably fit the experimental data
presented, and is favored by Sheldon et al.[ This is shown in
II
aldehydes mediated by Cu , TEMPO, and O
2
.
I
Cu in a one-electron oxidation of TEMPO 8 to the nitro-
sonium ion 9 [Scheme 4, Eq. (9)]. The alcohol is then oxidized
by 9 [Eq. (10)], generating the aldehyde and hydroxylamine
II
I
II
1
0; rapid syn-proportionation of 10 with 9 regenerates 8
II
[Eq. (11)]. Finally, Cu is regenerated by oxygen gas, in a
II
[17]
process that consumes protons and provides Cu and water
II
[Eq. (4)], following a general method for recycling Cu
[17]
ions. This catalytic cycle was based on the reactions and
electrochemical studies that indicated that Cu could indeed
II
oxidize 8 to 9.
In order to verify the Semmelhack mechanism,[17] we used
EPR spectroscopy to follow the redox chemistry of the
starting fluorous catalyst 4 during the oxidation of 4-nitro-
benzylic alcohol with TEMPO/O2 at 908C under FBC
conditions. To reiterate, precatalyst 4 was recoverable and
recyclable, while its catalytic activity did not decrease over the
course of five runs. We carried out an EPR catalytic experi-
ment (Figure 3, spectra a ± d) as was described above for the
oxidation of cyclohexene. Precatalyst 4 was dissolved in
perfluoroheptane, and then chlorobenzene, TEMPO, and
19]
II
Scheme 5. It entails the coordination of the alcohol to the Cu
center, followed by a b-hydrogen elimination, which forms the
Cu hydride, the aldehyde, and a proton. The regeneration
of the Cu center is now the role of the TEMPO radical, which
I
[19]
4-nitrobenzyl alcohol were added to the reaction mixture. The
II
reaction started at 908C in the presence of O , and after
2
I
30 min, followed by cooling to room temperature, an aliquot
abstracts hydrogen from Cu H to give TEMPOH, while the
was removed from the perfluoroheptane phase and immedi-
ately frozen at 77 K in an EPR tube. The EPR spectrum is
shown in Figure 8 (spectrum a, at LNT). The spectrum shows
a narrow central signal at about g ˆ 2.006, which can be
associated with the TEMPO radical. This was further
confirmed by measuring this radical in perfluoroheptane at
TEMPO radical can be regenerated by O
2
. Both mechanistic
II
Schemes 4 and 5 indicate that the presence of Cu is critical
for the oxidation of the alcohol to the aldehyde. The
mechanistic differences are in the role of the TEMPO radical,
and we are in the process of attempting to differentiate
between the two pathways.
Chem. Eur. J. 2003, 9, 4168 ± 4178
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4175

M. Contel, M. Laguna, R. H. Fish et al.
FULL PAPER
producing the aldehyde. The other mechanism involves the
II
I
Cu mediated b-hydrogen elimination providing a Cu H,
II
which reacts with TEMPO to recycle Cu and TEMPO-H,
while regenerating TEMPO with O . Future FBC studies will
2
focus on gas to liquid technology, as well as further studies on
FBC oxidation mechanisms.
Experimental Section
All manipulations involving organolithium reagents and hygroscopic
TACN were performed under an inert atmosphere of dry argon, using
Schlenk techniques, and all solvents were dried and degassed before use.
Scheme 5. Alternative mechanism for the oxidation of alcohols to
II
aldehydes mediated by Cu , TEMPO, and O
2
.
f
R -TACN 1 was prepared as described previously from a nitrogen atom
[
7a,d]
alkylation reaction between commercial TACN and C
8 2 3
F17(CH ) I.
Carboxylic acid, CF (CF CH CH CO H, was generously donated by Elf
3
2
)
7
2
2
2
Atochem. The fluorocarbon solvents and some starting perfluoro deriva-
tives were purchased from ABCR., while 4,4'-dimethyl-2,2'-bipyridine,
alkenes, alkanes, and alcohols were from Aldrich. tert-Butyl hydroperoxide
Conclusions
(
70% TBHP) was purchased from Aldrich, TEMPO (2,2,6,6-tetramethyl-
In summary, we have prepared and fully characterized by
piperidinyl-1-oxy) from Avocado, and were used without further purifica-
I
II
[21]
1
several standard spectroscopic techniques new Cu and Cu
complexes with fluoroponytailed amine ligands. To the best of
tion. [Cu(CH
3
19
CN)
4
]PF
6
was prepared as previously reported.
H,
1
3
1
1
C{ H}, and F{ H} NMR spectra were obtained on a 300 MHz Varian-
I
II
GEMINI 2000 apparatus in CDCl
relative to SiMe
3
solutions; chemical shifts are quoted
our knowledge, the reported Cu and Cu fluoroponytailed
derivatives were previously unknown, as the previously
1
13
1
19
1
4
(external, H and C{ H}) and CFCl
3
(external F{ H}).
GC analyses were performed on a Hewlett Packard (HP) 5890 Series II
instrument. C,H,N elemental analyses were performed with a Perkin Elmer
2400 microanalyzer, while mass spectra (LSIMS, 3-nitrobenzyl alcohol as
matrix) were obtained on a VG Autospec spectrometer using a Cesium
gun. The IR spectra were recorded on a Perkin ± Elmer Spectrum One FT-
IR instrument in a Universal ATR Sampling Accessory. The UV/Vis
spectra were recorded on a Unicam-Vision 32 instrument with accompa-
nying V1.02 software. The diffuse reflectance spectra were recorded on a
Cary 500 scan UV/Vis-nir spectrophotometer. The Electron Paramagnetic
Resonance (EPR) spectra were measured on a Bruker ESP380E spec-
trometer working at X-band using a PPPP cavity. Powdered samples as well
as solutions were introduced in a standard EPR quartz tube (707-SQ from
Wilmad). The spectra of powdered samples were recorded at room
temperature (RT). To follow catalytic experiments, aliquots of the
perfluoroheptane layer were removed and immediately frozen at liquid
nitrogen temperature (LNT). These frozen solutions were measured at
LNT using an immersion quartz Dewar. The microwave frequency was
determined with a Hewlett Packard HP5350B frequency counter. For
I
reported fluoroponytailed Cu complexes had been prepared
in situ and their characterization was incomplete or nonexis-
tent. More importantly, in many previously reported Cu
complexes that were studied, in their synthesis or catalytic
reactivity, there was no information concerning the oxidation
state of the resulting Cu complexes. We have now rectified
I
this situation by isolating and fully characterizing all of the Cu
II
and Cu compounds reported in either synthesis or under
catalytic reactivity, in this FBC contribution.
I
II
All of the reported novel Cu and Cu complexes were
evaluated in the FBC oxidations of hydrocarbons and olefins
with TBHP/O . The most promising catalytic systems were
2
I
II
the [Cu(R -tacn)Cl] 7, as well as the Cu and Cu complexes
f
with the R -bpy ligand 2. However, only 7 could be totally
f
II
separated from the organic phase by a simple decantation
process and thus recycled for three runs. EPR experiments
showed that the oxidation process occurs by an autoxidation
mechanism, with formation of alkenyl or alkyl hydroperox-
detecting Cu complexes, the conditions were as follows: 3320 Æ 125 mT
scan range, 2 mW microwave power, 9.51 GHz microwave frecuency, RTor
LNT temperature, 0.4 mT modulation amplitude, 100 KHz modulation
frequency, conversion time, 163.84 msg (see below).
Preparation of fluoroponytailed ligands and their copper complexes
ides as the key intermediates, which are then catalytically
4
f
,4'-Diheptadecafluoroundecyl-2,2'-bipyridine (R -bpy, 2): LDA, prepared
‡
decomposed by the [Cu(R -tacn] catalyst, probably at the
f
in situ from bisisopropylamine (2.37 mmol, 0.33 mL) and nBuLi
(2.42 mmol, 1.21 mL) in dry THF (60 mL) at À788C, was treated with
4,4'-dimethyl-2,2'-bipyridine (0.95 mmol, 0.17 g). The temperature was
allowed to rise from À78 to 08C over 1 h. After cooling down again to
solvent interface, to provide the alcohol and ketone prod-
ucts.[
7a,d]
However, this precatalyst was found to be less
efficient after the third run, possibly because of the concen-
À788C, C
8 2 3
F17(CH ) I (2.08 mmol, 1.21 g) in dry THF (5 mL) was added.
II
tration of the Cu complex that could not be readily reduced
The reaction mixture was stirred at this temperature for 5 h and then
warmed up to room temperature overnight. After total removal of the
I
to the initial Cu complex under the normal Haber± Weiss
reaction conditions.
Furthermore, the Cu complex [Cu(C F (CH ) CO ) (R -
tacn)] 4 was an efficient and recoverable precatalyst for the
oxidation of 4-nitrobenzyl alcohol to 4-nitrobenzaldehyde
2 2 2
THF, the residue was extracted with CH Cl /Et O (50 mL, 1:1) and washed
II
with H O (3 Â 15 mL). The organic layer was dried over Na SO , and all
2
2
4
8
17
2
2
2
2
f
solvents were concentrated to ꢀ1 mL. Subsequent addition of cold
n-pentane (6 mL) afforded pure 2 as a beige powder (0.361 g, 35%).
1
2
H NMR (CDCl
3
, 258C): d ˆ 8.57 (d, J(H,H) ˆ 5.1 Hz, 2H; H1,1'), 8.24 (s,
with TEMPO/O , under single phase FBC conditions at 908C,
1H; H3,3'), 7.13 (d,
4H; -CH
-
1
2
J(H,H) ˆ 5.1 Hz, 2H; H ), 2.74 (t, J(H,H) ˆ 7.5 Hz,
2
2
2,2'
with catalytic activity being maintained even after five runs.
The EPR experiments can be interpreted in two ways. In one
mechanism, this Cu complex was reduced to Cu by the
2
CH
2
-), 2.08 (m, 4H; -CH
2
CH
2
CH
2
-), 1.73 (dd, 8H;
1
CH
2
CH
2
CH CH
2
2
-C
8
F
17); 13C{ H} NMR (CDCl
3
, 258C): d ˆ 156.2, 155.6,
49.15, 123.8 (d), 121.1 (d), 120 ± 108 (4m, vbr), 34.8, 30.4 (t), 29.6, 19.7, 0.7
II
I
1
9
1
(
t); F{ H} NMR (CDCl
À123.0 (6F; CF ), À123.8 (2F; CF
LSIMS-MS: m/z (%): 1105 (5) [M] , 645 (100) [M À {(CH
3
, 258C): d ˆ À81.7 (3F; CF
3
), À115.4 (2F; CF
2
),
II
TEMPO radical and then oxidized again to the initial Cu
2
2
), À124.6 (2F; CF
2
), À127.3 (2F; CF
2
);
‡
‡
complex by O , while also regenerating TEMPO, and
2
)
3
C
8
F
17}] ;
2
4176
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4168 ± 4178

Fluorous Biphasic Catalysis
4168±4178
elemental analysis calcd (%) for C34
C 36.64, H 1.95, N 2.69.
H
22
F
34
N
2
: C 36.97, H 2.00, N 2.54; found
The addition of Et
2
O (10 mL) afforded a deep green powder (0.065 g,
54%). LSIMS-MS: m/z (%): 2271 (29) [M‡R
f
À bpy À Cl À
‡
‡
(
(
3
CH
CH
2
)
)
3
C
C
8
F
F
17 À F] , 1167 (100) [M À Cl] , 707 (100) [M À Cl À
[
Cu(C
added to a solution of CF
acetone (15 mL). This solution was added dropwise to a suspension of
CuSO ¥ 5H O (0.342 g, 1.37 mmol) in acetone (30 mL). After 24h stirring
at room temperature, complex 3 precipitated as a blue powder, which was
collected by filtration and washed with Et O to give the title compound
0.098 g, 69%). Complex 3 was totally soluble in trifluorotoluene, partly
soluble in dichloromethane, and insoluble in acetone, Et O, n-hexane,
8
F
17(CH
2
)
2
CO
2
)
2
]
(3): Triethylamine (0.39 mL, 2.80 mmol) was
‡
2
3
8
17 À F] ); elemental analysis calcd (%) for C34
22 34 2
H F N ClCu: C
3
(CF CH CH CO H (1.35 g, 2.80 mmol) in
2
)
7
2
2
2
3.93, H 1.84, N 2.33; found C 32.51, H 3.61, N 1.41.
Standard catalytic oxidation of alkanes and alkenes under FBC conditions:
A Schlenk tube (previously purged with Ar) was filled with a perfluoro-
heptane solution (2 mL) of the precatalyst (8 mmol). The isolated Cu
complexes 4 (0.020 g) and 5 (0.017 g) were used for the oxidation
experiments. For the Cu complexes, the preferred method was preparation
in situ from excess [CuCl] (16 mmol, 0.017 g) or [Cu(CH CN) ]PF
6
(16 mmol, 0.059 g) and 1 (0.012 g, 8 mmol), or 2 (0.009 g), in perfluorohep-
tane (2 mL). After stirring for 1 h at RT, these suspensions were filtered
4
2
II
2
(
I
2
À1
MeOH, H
Vis (diffuse reflectance): l ˆ 666 nm (28%); elemental analysis calcd (%)
for C22 Cu: C 25.27, H 0.77; found C 25.44, H 0.81.
2
O, and perfluoroheptane. IR: n ˆ 1573, 1420 cm (CˆO); UV/
3
4
H
8
F
34
O
4
I
through Celite (to remove the excess Cu salts). The resulting colourless (1)
[
Cu(C CO (R -tacn)] (4): R -TACN (0.218 g, 0.14 mmol) was
added to a suspension of 3, (0.15 g, 0.14 mmol) in dichloromethane
10 mL). A green solid precipitated immediately and was collected by
filtration. After drying under vacuum, 4 was obtained as a green powder
0.255 g, 80%). Complex 4 was totally soluble in trifluorotoluene and
8
F
17(CH
2
)
2
2
)
2
f
f
or green (2) solutions were placed in the Schlenk tube and used as
precatalysts. The hydrocarbon phase (upper layer of the biphasic system)
consists of 40 mmol of substrate (cyclohexene 4.05 mL, cyclohexane
(
4
.32 mL, toluene 4.26 mL) and 2 mmol (0.23 mL) of o-dichlorobenzene
as internal standard). Then tBuOOH (70%, 500 mmol, 69.2 mL) was added
and the reaction mixture (protected from light exposure) was stirred under
an O atmosphere at ambient temperature (for 8 or 24 h, see Table 1). After
(
(
perfluoroheptane, while being insoluble in acetone, dichloromethane,
À1
chloroform, Et
2
O, n-hexane, MeOH, and H
2
O. IR: n ˆ 1632, 1403 cm
2
(
1
{
CˆO); UV/Vis (perfluoroheptane): l ˆ 272 nm, (diffuse reflectance): l ˆ
the reaction had taken place, the two phases were separated by decantation
and analyzed by GC. In some cases, the fluorous phase could be reused for
further runs (see Table 1), in which case perfluoroheptane (0.5 mL) and a
new load of substrate/oxidant/internal standard were added, and the
oxidation reaction proceeded under the conditions described above.
013 (62%), 700 nm (48%); LSIMS-MS: m/z (%): 2063 (75) [M À
‡
‡
C
8
F
17(CH
2
)
2
CO
2
}] , 1572 (30) [M À {C
8
F
17(CH
2
)
2
CO
2
}
2
] , 1508 (58) [R
f
-
‡
TACN] ; elemental analysis calcd (%) for C61
38 85 4 3
H F O N Cu: C 28.67, H
1.49, N 1.64; found C 29.01, H 1.60, N 2.03.
[
Cu(C
8
F
17(CH
2
)
2
CO
2
)
2
(R
f
-bpy)] (5): R -bpy (0.110 g, 0.1 mmol) was added
2 2
CO ) ] (0.104 g, 0.1 mmol) in dichloro-
f
Standard catalytic oxidation of alcohols under FBC conditions: For the
preparation of 4-nitrobenzaldehyde, a 25 mL Schlenk tube (previously
purged with Ar) was charged with a perfluoroheptane solution (2 mL) of
to a suspension of [Cu(C
8
F
17(CH
2
)
2
methane (10 mL). A green solid precipitated immediately. This precipitate
was collected by filtration, and after drying under vacuum, 5 was obtained
as a pale green powder (0.086 g, 40%). Complex 5 was totally soluble in
trifluorotoluene, hot perfluoroheptane, or hot perfluoro-1,3-dimethylcy-
3
.5 mol% of 4 (0.027 mmol, 0.070 g). The hydrocarbon phase (upper layer
of the biphasic system) consisted of 4-nitrobenzylalcohol (0.120 g,
0
0
.78 mmol) in chlorobenzene (2 mL). TEMPO (6 mol%, 0.04 mmol,
.069 g) was then added, and the biphasic reaction mixture was stirred
clohexane (PP3), partly soluble in dichloromethane, and insoluble in
À1
acetone, Et
2
O, n-hexane, MeOH, and H
2
O. IR: n ˆ 1569, 1394 cm (CˆO);
2
under an O atmosphere at 908C (for 4 or 8 h, see Table 2). After the
UV/Vis (perfluoroheptane): l ˆ 246 nm, (diffuse reflectance): l ˆ 680 nm
reaction, the Schlenk tube was cooled to room temperature and the two
phases were separated by decantation. The fluorous phase was washed with
chlorobenzene (3 Â 2 mL), and the combined organic layers were diluted
‡
(
1
42%); LSIMS-MS: m/z (%): 1541 (5) [M À {C
8
‡
F
17(CH
2
)
2
CO
2
}-{C
2
F
5
}] ,
351 (33) [M À {C
CO
8
F
17(CH
2
)
2
CO
2
} À {C 13}] , 707 (100) [M À
5
F
‡
{
C
8
F
H
17(CH
2
)
2
2
} À {C
6
F
14}] ; elemental analysis calcd (%) for
with Et
2
O (30 mL) and washed successively with brine (40 mL), and water
C
56
30
F
68
O
4
N
2
Cu: C 31.28, H 1.40, N 1.30; found C 31.54, H 1.60, N 1.73.
-tacn)]PF (6): R -TACN (0.151 g, 0.1 mmol) was added to a
solution of [Cu(CH CN) ]PF (0.037 g, 0.1 mmol) in dry dichloromethane
10 mL) under argon. After 2 h of stirring at room temperature, the
solution became pale green. After removal of the CH Cl and addition of
Et O, a very pale green powder precipitated. This precipitate was collected
(
40 mL). After drying (MgSO ), filtration, and complete evaporation of the
4
[
Cu(R
f
6
f
solvent under reduced pressure, the crude product was weighted and
1
3
4
6
analysed by H NMR. The fluorous phase was separated and used directly
(
(after adding 0.5 mL of perfluoroheptane) for further reaction runs.
2
2
EPR-FBC experiment A: Using standard oxidation conditions for cyclo-
hexene, as outlined in Table 1, with precatalyst 7, generated in situ, aliquots
of the perfluoroheptane layer were removed and immediately frozen at
LNT in an EPR tube (Figure 7, EPR spectra a ± e).
2
by filtration, and after drying under vacuum, 6 was obtained as an almost
off white powder (0.061 g, 36%). Complex 6 was soluble in trifluorotoluene
and insoluble in dichloromethane acetone, Et
2
O, n-hexane, MeOH, H
2
O,
EPR-FBC experiment B: Using standard oxidation conditions for 4-nitro-
benzylic alcohol, as outlined in the experimental section using 4 as the
precatalyst, aliquots of the perfluoroheptane layer were removed (after
cooling down the reaction mixture at RT) and immediately frozen at LNT
in an EPR tube (Figure 8, EPR spectra a ± d).
‡
and perfluoroheptane. LSIMS-MS: m/z (%): 1753 (15) [M‡2H
2
O] , 1573
‡
‡
(
60) [M À PF
for C39
Cu(R
suspension of excess [CuCl] (0.050 g, 0.5 mmol) in trifluorotoluene
5 mL). After 3 h of stirring at room temperature, the suspension was
6
] , 1554 (100) [M À PF
6
À F] ; elemental analysis calcd (%)
30 57 3
H F N PCu: C 27.26, H 1.76, N 2.44; found C 27.66, H 1.82, N 2.09.
[
f f
-tacn)Cl] (7): R -TACN (0.151 g, 0.1 mmol) was added to a
(
filtered through celite (to remove the excess CuCl), and the colourless
solution was evaporated to 1 mL. The addition of n-hexane afforded 7 as an
off-white powder (0.125 g, 78%). This complex was totally soluble in
trifluorotoluene and perfluoroheptane, partly soluble in dichloromethane
Acknowledgments
The University of Zaragoza authors would like to thank the MCYT for
financial support (BQU2002-04090-C02-01). M.C. wishes to thank the
MCYT–Universidad de Zaragoza for her Research Contract ™Ram o¬ n y
Cajal∫. We thank Dr. Victor Orera for the diffuse reflectance measure-
ments. R.H.F. wishes to thank the Ministerio de Educaci o¬ n y Cultura,
Madrid, Spain, for a Visiting Professorship in 2000 at the University of
Zaragoza, and Elf Aquitaine Inc. for support of the LBNL FBC program,
and the Department of Energy under Contract No.DE-AC03-76SF00098.
and chloroform, and insoluble in acetone, Et
2
O, n-hexane, MeOH, and
-CH -CH -C
-CH -C 17), 1.20 (m,
, 258C): d ˆ À80.5,
O. 1H NMR (CDCl
-CH
8
-C F17); F{ H} NMR (CDCl
H
2
3
, 258C): d ˆ 2.25 (m, 6H; -CH
2
2
2
8
F17),
2
6
.10 (s, 12H; -N-CH
H; -CH
2
2
-N), 1.57 (m, 6H; -CH
2
-CH
2
2
8
F
1
9
1
2
-CH
2
-CH
2
3
1
9
1
À110.8, À111.2, À121.1, À121.7, À122.6, À125.9 (the F{ H} NMR
spectrum in CDCl of R
3
f
-TACN (1) has signals at d ˆ À81.7, À115.2,
À122.9, À123.9, À124.7, À127.3); UV/Vis (perfluoroheptane): l ˆ 212,
‡
2
60 nm; LSIMS-MS: m/z (%): 1573 (88) [M À Cl] , 1095 (100) [M À Cl À
‡
(
CH
2
)
3
C
8
F
17 À F] ; elemental analysis calcd (%) for C39
30 51 3
H F N ClCu: C
2
9.12, H 1.88, N 2.61; found C 29.45, H 1.97, N 2.43.
[
1] a) G. W. Parshall, S. D. Ittel, Homogeneous Catalysis, Wiley, New
York, 1992, pp 138 ± 61, p. 172 ± 174, p. 237 ± 264 and references there-
in; b) R. A. Sheldon, J. K. Kochi, Metal-Catalyzed Oxidation of
Organic Compounds, Academic press, New York, 1981, p. 249 ± 271.
Reaction of CuCl and R
f
-bpy (2): R -bpy (2; 0.110 g, 0.1 mmol) was added
f
to a suspension of [CuCl] (0.010 g, 0.1 mmol) in trifluorotoluene (5 mL). A
green solid started to precipitate immediately. After 1 h of stirring at room
temperature, the trifluorotoluene was removed under vacuum to ꢀ1 mL.
[2] I. T. Horv a¬ th, J. Rabai, Science 1994, 266, 72 and references therein.
Chem. Eur. J. 2003, 9, 4168 ± 4178
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4177

M. Contel, M. Laguna, R. H. Fish et al.
FULL PAPER
[
3] a) R. L. Scott, J. Am. Chem. Soc. 1948, 70, 4090; b) J. H. Hildebrand,
D. R. F. Cocharn, J. Am. Chem. Soc. 1949, 71, 22; c) D. L. Dorset,
Macromolecules 1990, 23, 894; d) D. W. Zhu, Synthesis 1993, 953.
4] Recent reviews: a) J.-M. Vincent, A. Rabion, R. H. Fish, ACS
Symposium Series 767, Green Chemical Syntheses and Processes
[11] a) G. Pozzi, F. Cinato, F. Montanari, S. Quici Chem. Commun. 1998,
877; b) G. Pozzi, M. Cavazzini, F. Cinato, F. Montanari, S. Quici, Eur. J.
Org. Chem. 1999, 1947; c) M. Cavazzini, A. Manfredi, F. Montanari, S.
Quici, G. Pozzi, Chem. Commun. 2000, 2171; d) D. Maillard, C.
Nguefack, G. Pozzi, S. Quici, G. Valad e¬ , D. Sinou, Tetrahedron:
Asymmetry 2000, 11, 2881; e) M. Cavazzini, A. Manfredi, F. Mon-
tanari, S. Quici, G. Pozzi, Eur. J. Org. Chem. 2001, 4639.
[12] F. De Campo, D. Last e¬ cou e¡ res, J.-M. Vincent, J.-B. Verlhac, J. Org.
Chem. 1999, 64, 4969.
[
(Eds.: P. T. Anastas, L. C. Heine, T. C. Williamson), Chapter 14, p. 172,
2
000; b) T. J. Kitazume, Fluorine Chem. 2000, 105, 265; c) L. P.
Barthel-Rosa, J. A. Gladysz, Coord. Chem. Rev. 1999, 190 ± 192, 587;
d) M. Cavazzini, F. Montanari, G. Pozzi, S. J. Quici, Fluorine Chem.
1
1
999, 94, 183; e) E. G. Hope, A. M. Stuart, J. Fluorine Chem. 1999,
00, 75; f) R. H. Fish, Chem. Eur. J. 1999, 5, 1677; g) E. De Wolf, G.
[13] D. M. Haddleton, S. G. Jackson, S. A. F. Bon, J. Am. Chem. Soc. 2000,
122, 1542.
Van Koten, B.-J. Deelman, Chem. Soc. Rev. 1999, 28, 37; h) I. T.
Horv a¬ th, Acc. Chem. Res. 1998, 31, 641; i) B. Cornils, Angew. Chem.
[14] Selected examples: a) M. S. Kharasch, G. Sosnovsky, J. Am. Chem.
Soc. 1958, 80, 756; b) J. K. Kochi, J. Am. Chem. Soc. 1962, 84, 774;
c) A. L. J. Beckwith, A. A. Zavitsas, J. Am. Chem. Soc. 1986, 108,
8230; d) N. Komiya, T. Naota, S.-I. Murahashi, Tetrahedron Lett. 1996,
37, 1633; e) G. Sekar, A. DattaGupta, V. K. Singh, Tetrahedron Lett.
1996, 46, 8435.
[15] K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Complexes, 4th ed., Wiley, New York. 1986, pp. 232;
5th ed., 1997, part B pp. 57.
1
997, 109, 2147; Angew. Chem. Int. Ed. Engl. 1997, 36, 2057.
[
[
[
5] I. Klement, H. L¸tjens, P. Knochel, Angew. Chem. 1997, 109, 1605;
Angew. Chem. Int. Engl. 1997, 36, 1454.
6] S. Colonna, N. Gaggero, F. Montanari, G. Pozzi, S. Quici, Eur. J. Org.
Chem. 2001, 181.
7] a) J.-M. Vincent, A. Rabion, V. K. Yachandra, R. H. Fish, Angew.
Chem. 1997, 109, 2438; Angew. Chem. Int. Ed. Engl. 1997, 36, 2346;
b) G. Pozzi, M. Cavazzinni, S. Quici, S. Fontana, Tetrahedron Lett.
[16] F. A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 5th ed.,
1
997, 38, 7605; c) B. Betzemeier, F. Lhermitte, P. Knochel, Tetrahedron
Wiley, New York, 1988, pp. 755.
Lett. 1998, 39, 6667; d) J.-M. Vincent, A. Rabion, V. K. Yachandra,
R. H. Fish, Can. J. Chem. 2001, 79, 888.
[17] M. F. Semmelhack, C. R. Schmid, D. A. Cort e¬ s, C. S. Chou, J. Am.
Chem. Soc. 1984, 106, 3374.
[
[
8] B. Betzemeier, M. Cavazzini, S. Quici, P. Knochel, Tetrahedron Lett.
[18] J. I. MartÌnez, M. Marcos, J. L. Serrano, V. M. Orera, P. J. Alonso, Liq.
2
000, 41, 4343.
Cryst. 1995, 19, 603 ± 613.
9] T. Nishimura, Y. Maeda, N. Kakiuchi, S. J. Uemura, Chem. Soc. Perkin
Trans. 1 2000, 4301.
[19] A. Dijksman, A. Marino-Gonz a¬ lez, A. Mairata i Payeras, I. W. C. E.
Arends, R. A. Sheldon, J. Am. Chem. Soc. 2001, 123, 6826.
[20] R. H. Fish, M. S. Konings, K. J. Oberhausen, R. H. Fong, W. M. Yu, G.
Christou, J. B. Vincent, D. K. Coggin, R. M. Bunachan, Inorg. Chem.
1991, 30, 3002.
[
10] a) G. Pozzi, S. Banfi, A. Manfredi, F. Montanari, S. Quici, Tetrahedron
1996, 52, 11879; b) G. Pozzi, F. Montanari, S. Quici, Chem. Commun.
1997, 69; c) B. Betzemeier, F. Lhermitte, P. Knochel, Synlett 1999, 4,
489; d) S. Quici, M. Cavazzini, S. Ceragioli, F. Montanari, G. Pozzi,
[21] G. K. Kubas, Inorg. Synth. 1990, 28, 68.
Tetrahedron Lett. 1999, 40, 3647.
Received: January 27, 2003 [F4771]
4178
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4168 ± 4178