Catalytic olefin epoxidation with a carboxylic acid-functionalized cyclopentadienyl molybdenum tricarbonyl complex Dedicated to Prof. Maria José Calhorda on the occasion of her 65th birthday and in recognition of her outstanding contributions to the field of computational organometallic chemistry.

Article abstract of DOI:10.1016/j.jorganchem.2013.10.037

The complex CpMo(CO)₃CH₂COOH (1) (Cp = η⁵-C₅H₅) has been examined as a precatalyst for the epoxidation of cis-cyclooctene and α-pinene using tert-butylhydroperoxide (TBHP) as oxidant. A high turnover frequency of ca. 600 mol mol_Mo^-1 h^-1 was achieved in the epoxidation of cyclooctene, giving the epoxide as the only reaction product. With α-pinene as substrate, the added-value products α-pinene oxide and campholenic aldehyde were obtained. Two different approaches to facilitate catalyst recovery and reuse were explored: (1) use of an ionic liquid (IL) as solvent, and (2) intercalation of 1 in a Zn,Al layered double hydroxide (LDH) by a direct synthesis (coprecipitation) route. Characterization of the LDH by powder X-ray diffraction, thermogravimetric analysis, FT-IR and ¹³C CP MAS NMR spectroscopies showed that the CpMo(CO)₃CH ₂COO^- anions intercalate in a bilayer arrangement, resulting in an interlayer spacing of 20.7 A?. In the epoxidation of cyclooctene, catalytic activity in the first batch run was very high for the catalyst/IL mixture and moderate for the LDH. Characterization of the LDH after catalysis indicated that nearly complete oxidative decarbonylation of supported complexes had occurred (by reaction with TBHP), resulting in the presence of immobilized oxomolybdenum species. However, catalytic activities for both the recovered LDH and catalyst/IL decreased in consecutive runs, due in part to progressive removal of active species during either the catalytic reaction (for the LDH) or the solvent extraction/work-up (for the catalyst/IL mixture).

Full text of DOI:10.1016/j.jorganchem.2013.10.037

Journal of Organometallic Chemistry xxx (2013) 1e7
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Catalytic olefin epoxidation with a carboxylic acid-functionalized
cyclopentadienyl molybdenum tricarbonyl complex
a
a
b
a
Ana C. Gomes , Sofia M. Bruno , Marta Abrantes , Clara I.R. Magalhães ,
a
a
a,
*
Isabel S. Gonçalves , Anabela A. Valente , Martyn Pillinger
Department of Chemistry, CICECO, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal
a
b
Centro de Química Estrutural, Complexo Interdisciplinar, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, Portugal
a r t i c l e i n f o
a b s t r a c t
The complex CpMo(CO) CH COOH (1) (Cp ¼ h5-C H ) has been examined as a precatalyst for the
Article history:
Received 30 September 2013
Received in revised form
3
2
5 5
epoxidation of cis-cyclooctene and
a
-pinene using tert-butylhydroperoxide (TBHP) as oxidant. A high
ꢀ1 ꢀ1
turnover frequency of ca. 600 mol molMo
h
was achieved in the epoxidation of cyclooctene, giving the
-pinene as substrate, the added-value products -pinene
18 October 2013
epoxide as the only reaction product. With
a
a
Accepted 21 October 2013
oxide and campholenic aldehyde were obtained. Two different approaches to facilitate catalyst recovery
and reuse were explored: (1) use of an ionic liquid (IL) as solvent, and (2) intercalation of 1 in a Zn,Al
layered double hydroxide (LDH) by a direct synthesis (coprecipitation) route. Characterization of the LDH
Dedicated to Prof. Maria José Calhorda on
the occasion of her 65th birthday and in
recognition of her outstanding contribu-
tions to the field of computational organo-
metallic chemistry.
13
by powder X-ray diffraction, thermogravimetric analysis, FT-IR and C CP MAS NMR spectroscopies
ꢀ
showed that the CpMo(CO)
3
CH
2
COO anions intercalate in a bilayer arrangement, resulting in an
ꢀ
interlayer spacing of 20.7 A. In the epoxidation of cyclooctene, catalytic activity in the first batch run was
very high for the catalyst/IL mixture and moderate for the LDH. Characterization of the LDH after
catalysis indicated that nearly complete oxidative decarbonylation of supported complexes had occurred
Keywords:
Molybdenum
Cyclopentadienyl ligands
Carbonyl ligands
Layered double hydroxides
Epoxidation
(
by reaction with TBHP), resulting in the presence of immobilized oxomolybdenum species. However,
catalytic activities for both the recovered LDH and catalyst/IL decreased in consecutive runs, due in part
to progressive removal of active species during either the catalytic reaction (for the LDH) or the solvent
extraction/work-up (for the catalyst/IL mixture).
Ionic liquids
Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction
[11,12]. Based on catalytic activities obtained for the epoxidation of
cis-cyclooctene (used as a benchmark substrate), standout pre-
5
The epoxidation of olefins catalyzed by molecular transition
metal compounds continues to be a topic of considerable academic
and industrial interest [1e3]. Molybdenum-based catalysts have
been widely used in this reaction. The well known industrial
example is the Arco-Lyondell process for the epoxidation of pro-
catalysts include (
h
-C
(CH(CH
perature ionic liquid (RTIL) as a solvent) [14], and the fluorinated
complex CpMo(CO) CF (with hexafluoroisopropanol as a solvent;
) [15]. These tricarbonyl complexes undergo oxidative
decarbonylation in situ by reaction with TBHP, forming catalytically
5
Bz
5
)Mo(CO)
3
Cl (Bz ¼ benzyl) [13], the ansa-
5
1
complex [Mo(
h
-C
5
H
4
2 3 3
) )-h -CH)(CO) ] (with a room tem-
3
3
5
Cp ¼
5 5
h -C H
6
pene using Mo(CO) as a precursor, which is oxidized in situ by the
0
0
2
oxidant tert-butylhydroperoxide (TBHP) to give the active molyb-
denum(VI) catalyst [4e6]. During the last decade there has been
increasing interest in the use of cyclopentadienyl molybdenum
active Mo-oxo, Cp MoO X, and Mo-oxo-peroxo, Cp MoO(h -O )X,
2
2
species. More complex species may also be formed as in the anal-
ogous oxidation of (CpBz)Mo(CO) Me that gives [(CpBz)MoO
O) [16]. Computational studies support a mechanism in which
3
2 2
] (m-
0
0
5
carbonyls of the type Cp Mo(CO)
3
X (Cp ¼
h
5
-C R
5
(R ¼ H, alkyl,
0
t
ansa-bridge); X ¼ halide, alkyl, ansa-bridge) as precatalysts for
active intermediates of the type Cp MoO(OH)(OO Bu)X and
0
2
t
olefin epoxidation [7,8]. Great variability by ligand modification
Cp Mo(
2
h -O )(OH)(OO Bu)X are formed by reaction of the respec-
(
both on the Cp-ring and directly at the metal center) is possible,
tive Mo-oxo and Mo-oxo-peroxo complexes with excess TBHP
[17,18].
opening the way to immobilization [9,10] and chirality introduction
In the present study the complex CpMo(CO)
been prepared, characterized, and examined as a precatalyst for the
epoxidation of cis-cyclooctene and -pinene. The recovery and
reuse of the homogeneous catalyst was explored using an ionic
3 2
CH COOH (1) has
a
*
Corresponding author. Fax: þ351 234 401470.
E-mail address: mpillinger@ua.pt (M. Pillinger).
0
022-328X/$ e see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jorganchem.2013.10.037
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

2
A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
liquid as a solvent. Additionally, a new approach for the immobi-
lization of cyclopentadienyl molybdenum tricarbonyl complexes is
reported, based on the intercalation of 1 in a layered double
hydroxide.
residue (NaCp) washed with anhydrous n-pentane (2 ꢂ 15 mL) and
vacuum-dried. NaCp was treated overnight with Mo(CO) (5.40 g,
20.4 mmol) in refluxing THF (60 mL). The resultant reaction
mixture was cooled to room temperature and ClCH CONH (1.91 g,
0.4 mmol) in THF (60 mL) added. After stirring at room temper-
6
2
2
2
2
. Experimental
ature for 2 h, the brown suspension was vacuum-dried and the
resultant residue extracted with chloroform. The chloroform frac-
tions were combined and evaporated to dryness, giving CpMo(-
2.1. Materials and methods
CO)
3
CH
2
COONH
2
, which was subsequently treated with water
ꢁ
Microanalyses for CHN and ICP-OES analysis for Mo were
(34 mL) and HCl (37%, 12 mL, ca. 140 mmol) at 80 C for 1 h.
determined by C.A.C.T.I., University of Vigo. Powder X-Ray diffrac-
tion (XRD) data were collected at room temperature on an X’pert
CpMo(CO)
was filtered, washed with water, vacuum-dried, and recrystallized
from chloroform (1.30 g, 21%, relative to Mo(CO) ). Anal. Calcd for
MoO (304.11): C, 39.50; H, 2.65. Found: C, 39.2; H, 2.65. FT-IR
3 2
CH COOH (1) precipitated as a yellow powder, which
MPD Philips diffractometer with a curved graphite monochromator
6
ꢀ
(
Cu-K
a
radiation,
l
¼ 1.54060 A) and a flat-plate sample holder, in a
C
10
H
8
5
ꢀ
1
BraggeBrentano para-focusing optics configuration (45 kV, 50 mA).
Samples were step-scanned in 0.02 2
(KBr, cm ):
2524 (w), 2025 (vs,
(vs, (CO)), 1647 (vs,
n
¼ 3438 (br), 3120 (w), 2935 (w), 2784 (w), 2624 (w),
(CO)), 1949 (vs, (CO)), 1931 (vs, (CO)), 1908
(C]O)), 1427 (m), 1416 (m), 1283 (vs, (CeO)),
ꢁ
q
steps with a counting time
n
n
n
of 50 s per step. SEM with coupled EDS was carried out on a Hitachi
SU-70 (S-4100) instrument using a 15 kV accelerating voltage.
Thermogravimetric analysis (TGA) was performed using a Shi-
n
n
n
1104 (m), 1043 (m), 1002 (w), 923 (w), 844 (m), 825 (m), 742 (w),
655 (m), 580 (m), 552 (s), 482 (s), 435 (m), 360 (w). FT-Raman
ꢁ
ꢀ1
ꢀ1
madzu TGA-50 system at a heating rate of 5 C min under air. FT-
IR spectra were obtained as KBr pellets using an FTIR Mattson-7000
(cm
)
n
¼ 3129 (m), 3109 (w), 3020 (w), 2959 (w), 2013 (s), 1957
(vs),1935 (w),1903 (s),1110 (s),1058 (w),1043 (s), 465 (w), 440 (m),
ꢀ
1
spectrophotometer and recorded from 4000 to 350 cm . Raman
spectra were recorded on a Bruker RFS100/S FT instrument
417 (m), 394 (m), 362 (s), 341 (s), 156 (w), 140 (s), 124 (vs), 112 (vs).
1
H NMR (300 MHz, CDCl3, 298 K):
d
¼ 5.43 (s, 5H, Cp), 1.80 (s, 2H,
, 298 K):
¼ 239.3 (CO), 226.3
COOH), 93.8 (Cp), ꢀ4.5 (CH COOH).
13
(
Nd:YAG laser, 1064 nm excitation, InGaAs detector). Liquid-state
CH
2
COOH). C NMR (126 MHz, CDCl
3
d
1
13
H and C NMR spectra were measured with a Bruker Avance
00 instrument. Solid-state NMR spectra were recorded on a
(CO), 188.0 (CH
2
2
3
13
Bruker Avance 400 spectrometer. C cross polarization (CP) magic-
angle spinning (MAS) NMR spectra were recorded at 100.62 MHz
with 3.3
2.2.2. Zn,Al-CpMo
A yellow solution of the sodium salt [CpMo(CO)
3
2
CH COO]Na
1
ꢁ
m
s H 90 pulses, a 2 ms contact time, a spinning rate of
was prepared by addition of 1 equivalent of NaOH to an aqueous
1
2 kHz and 4 s recycle delays. Chemicals shifts are quoted in ppm
suspension (30 mL) of complex 1 (0.46 g, 1.50 mmol). A solution of
Zn(NO ) $6H O (0.59 g, 2.00 mmol) and Al(NO ) $9H O (0.37 g,
3 2 2 3 3 2
27
relative to TMS.
04.26 MHz using a
and 1 s recycle delays. Chemicals shifts are quoted in ppm relative
Al MAS NMR spectra were recorded at
1
p
/12 pulse of 0.78 s, a spinning rate of 14 kHz
m
1.00 mmol) in decarbonated deionized (DD) water (30 mL) was
then added dropwise, and the pH of the mixture was continuously
maintained between 7.5 and 8 using 0.2 M NaOH. After the addition
3
þ
to [Al(H
2
O)
6
]
.
ꢁ
Where appropriate, preparations and manipulations were car-
ried out using standard Schlenk techniques under nitrogen. The
chemicals sodium (Riedel-de-Haën), dicyclopentadiene (BDH),
was completed the reaction mixture was aged at 65 C for 19 h. The
resultant yellow solid was filtered, washed extensively with DD
water and acetone, and finally vacuum-dried. Anal. found: Mo,
ꢀ
1
Mo(CO)
6
(Fluka), 2-chloroacetamide (ClCH
2
CONH
2
, 98%, Sigmae
12.7%. FT-IR (KBr, cm ):
n
¼ 3448 (br), 2017 (vs,
n(CO)), 1950 (vs,
Aldrich), anhydrous n-pentane (95%, SigmaeAldrich), anhydrous
tetrahydrofuran (THF, 99.9%, SigmaeAldrich), anhydrous chloro-
form (99%, SigmaeAldrich), acetone (99%, SigmaeAldrich), hydro-
chloric acid (HCl, 37%, Fluka), zinc nitrate hexahydrate (99%, Fluka),
aluminum nitrate nonahydrate (98.5%, Riedel-de-Haën), 50%
n
(CO)), 1909 (vs, (CO)), 1633 (m), 1579 (m), 1491 (m), 1429 (m),
n
1360 (s), 1350 (s), 1261 (vw), 1113 (w), 1060 (w), 1014 (w), 825 (m),
790 (w), 748 (w), 625 (bd), 580 (w), 553 (w), 484 (m), 428 (m). FT-
ꢀ
1
Raman (cm ):
n
¼ 3119 (m), 2945 (w), 2022 (m), 1944 (m), 1110 (s),
13
1061 (w), 918 (w), 448 (w), 415 (w), 339 (s), 115 (s). C CP MAS
ꢀ
aqueous sodium hydroxide (Fluka), Na
-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide
[bmim]NTf , 99%, io-lo-tec) were obtained from commercial
2
CO
3
(J.M. Vaz Pereira), and
NMR:
d
¼ 242.8 (CO), 232.0 (CO), 227.0 (CO), 191.5 (CH
2
CO
2
), 177.0
2
ꢀ
ꢀ
27
1
(
(CO
3
), 94.0 (Cp), 1.3 (CH
2
CO
2
). Al MAS NMR:
d
¼ 13.5.
2
sources and used as received. The solution of tert-butylhydroper-
oxide (TBHP) in decane (5e6 M, SigmaeAldrich, <4% water) was
dried over activated 4 A molecular sieves prior to use. The Zn,Al-
2.3. Catalytic tests
ꢀ
NO
3þ
coprecipitation of the Zn and Al hydroxides (initial Zn /Al
3
LDH precursor was prepared by the standard method of
The catalytic reactions were carried out under air (autogenous
pressure) and stirred magnetically (1000 rpm) in a closed borosil-
icate reactor (10 mL capacity) equipped with a valve for sampling,
2
þ
3þ
2þ
molar ratio in solution ¼ 2) in the presence of nitrate ions at con-
ꢁ
ꢁ
stant pH (7.5e8) under nitrogen, followed by aging at 80 C for 20 h
and immersed in an oil bath thermostated at 55 C. Typically, the
[
19]. After washing, the material was stored as an aqueous sus-
reactor was loaded with catalyst (18
(1.8 mmol, cis-cyclooctene (Cy) or -pinene (Pin)) and 5e6 M TBHP
in decane (ca. 2.75 mmol). When stated, 150 L of [bmim]NTf was
mmol of molybdenum), olefin
pension in a closed container.
a
m
2
2
2
.2. Synthesis
used as a cosolvent. The olefin, oxidant and (when used) IL were
heated in separate vessels (10 min at the reaction temperature)
prior to addition to the reactor (with preheated walls) containing
the catalyst. Time zero (i.e., the initial instant of the reaction) was
taken as the moment the oxidant was added to the reactor. The
.2.1. CpMo(CO)
3
CH
2
COOH (1)
CpMo(CO) CH COOH (1) was synthesized according to the
3 2
general procedure described by Ariyaratne et al. [20]. A detailed
description of the preparation is given here. Freshly cut sodium
ꢁ
leaching tests were performed by separating the solid (at 55 C)
(
0.78 g, 34.0 mmol) was added to 40 mL of dicyclopentadiene under
from the reaction mixture after 2 h, using a 0.2
membrane; the filtrate was stirred for a further 22 h at 55 C. The
course of the reactions was monitored using a Varian 3800 GC
m
m PTFE GVS
ꢁ
ꢁ
inert atmosphere. The mixture was refluxed overnight at 160 C.
The resultant suspension was filtered and the remaining pink solid
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
3
equipped with a BR-5 (Bruker) capillary column (30 m ꢂ 0.25 mm;
0
2
.25 mm) and a flame ionization detector, using H as the carrier gas.
3
. Results and discussion
3.1. Synthesis and characterization
The reaction of NaCp with Mo(CO)
6
, followed by addition of
2
CO)
-chloroacetamide, gave the tricarbonyl complex CpMo(-
CH COONH , which was subsequently hydrolyzed to give yel-
low CpMo(CO) CH COOH (1), as illustrated in Scheme 1. FT-IR and
3
2
2
3
2
1
H NMR data for 1 were in agreement with literature results [20].
In recent work we successfully heterogenized the complexes
3
2ꢀ
2ꢀ
[
2
Mo(
h
-C
3
H
5
)Cl(CO)
2
(bpdc)] and cis-[Mo(CO)
4
(bpdc)] (bpdc ¼
0
0
,2 -bipyridine-5,5 -dicarboxylate) by using layered double hy-
droxides (LDH) as supports [19,21]. The former complex was
intercalated in a one-pot direct synthesis approach [19], while the
latter complex was intercalated by ion exchange of a Zn,Al-NO
3
LDH
precursor [21]. Initial attempts to intercalate complex 1 by ion
exchange of a Zn,Al-NO
3
LDH with a Zn/Al molar ratio of 2 were
ꢁ
unsuccessful (reaction conditions: water as solvent, 50 C, 6 h):
while the IR spectra of the resultant material indicated the presence
of complex 1, the powder XRD pattern was unchanged from that for
the starting material, suggesting that the tricarbonyl complex was
mainly adsorbed on the external surface of the LDH particles. These
materials were not further characterized. Instead, a one-pot
coprecipitation method at constant pH was found to give an
intercalated LDH (abbreviated as Zn,Al-CpMo) as evidenced by
powder XRD, elemental analysis, thermogravimetric analysis,
13
27
2þ
3þ
vibrational spectroscopy, and C and Al MAS NMR. Zn and Al
2
þ
3þ
were chosen as the M /M combination since Zn,Al-LDHs are
Fig. 1. FT-IR spectra (KBr) of (a) Zn,Al-NO
and (d) the solid recovered after one batch run of Cy epoxidation using Zn,Al-CpMo as
the precatalyst. Selected bands are highlighted.
3
, (b) complex 1, (c) the material Zn,Al-CpMo,
prepared (and are stable) at near-neutral pH, thereby providing the
ꢀ
best compatibility with the anionic complex CpMo(CO)
3
CH
2
COO .
A comparison of the FT-IR spectra for Zn,Al-CpMo and complex
ꢀ
1
confirms the presence of intact CpMo(CO)
3
CH
2
COO anions
resultant solid was unchanged from that for the freshly prepared
sample, indicating that the material is neither oxygen- nor mois-
ture-sensitive.
(Fig. 1). Zn,Al-CpMo displays three very strong
n
(CO) bands at 1909,
950 and 2017 cm , similar to those displayed by 1 (1908,1949 and
025 cm ) and in agreement with that reported for other com-
ꢀ
1
1
2
ꢀ
1
Fig. 2 compares the 13C CP MAS NMR spectrum of Zn,Al-CpMo
0
plexes of the type Cp Mo(CO)
3
X [13,20,22,23]. Complex 1 exhibits
ꢀ1
with the solution (CDCl
3
) NMR spectrum of complex 1. The inter-
two very strong bands at 1283 and 1647 cm , which are attributed
to the CeO and C]O stretching vibrations of the carboxylic acid
group. These bands are absent in the spectrum of Zn,Al-CpMo.
ꢀ
calated LDH shows signals at 1.3 (CH
2
CO
2
), 94.0 (Cp), 191.5
ꢀ
(CH
2
CO
2
) and 227e243 ppm (CO) for the organometallic guest, and
ꢀ1
a weak signal at 177.0 ppm that is probably due to carbonate anions.
The Cp resonance for Zn,Al-CpMo is unshifted when compared
Instead, bands at 1350 and 1579 cm (both absent in the spec-
trum of 1) appear and are assigned to (OCO) and as(OCO),
respectively, of the deprotonated carboxylate group. Whereas the
n
s
n
3 2
with that for 1 in CDCl , while the signals for the CH and
2
7
ꢀ
1
carboxylate groups are shifted slightly downfield. The Al MAS
NMR spectrum of Zn,Al-CpMo presents a broad, somewhat asym-
metric signal centered at 13.5 ppm that indicates the presence of
octahedral aluminum only (Fig. 2).
material Zn,Al-NO
the antisymmetric stretching mode
3
contains a very strong band at 1383 cm due to
of nitrate ions, this band is
n
3
not evident in the spectrum of Zn,Al-CpMo, indicating that the
material is essentially free of interference from this anion (which is
present in the synthesis mixture due to the use of Zn(NO
Al(NO ) ). Below 750 cm , the IR spectrum of Zn,Al-CpMo contains
several bands due to the guest species (484, 553 and 580 cm ) in
addition to Zn/Al-OH lattice translation modes that are character-
The thermal decomposition behavior of Zn,Al-CpMo was stud-
ied by thermogravimetric analysis under air (Fig. 3). Three weight
3
)
2
and
ꢀ1
3 3
ꢁ
ꢀ
1
loss processes can be distinguished up to 325 C: loss of coin-
ꢁ
ꢁ
tercalated water (20e105 C, DTGmax ¼ 85 C, 6.6%), loss of CO
ꢁ
ꢀ
1
molecules from the intercalated tricarbonyl complex (105e200 C,
istic of Zn,Al-LDHs (428, 553 and 625 cm ) [24].
ꢁ
DTGmax ¼ 150 C, 11.7%) and partial dehydroxylation (loss of water)
To assess the stability of Zn,Al-CpMo, the sample was deliber-
ately exposed to air for one month. The FT-IR spectrum of the
ꢁ
ꢁ
of the hydroxide layers (200e325 C, DTGmax ¼ 265 C, 9.7%).
ꢁ
A further mass loss of 7.3% takes place up to 500 C, which is
attributed to a combination of the further dehydroxylation of the
hydroxide layers and the decomposition of the remaining organic
moieties. TGA curves were also measured for Zn,Al-NO
3
and com-
plex 1. Loss of cointercalated water from Zn,Al-NO
3
takes place at a
ꢁ
slightly higher temperature (DTGmax ¼ 125 C) than that for Zn,Al-
CpMo, while the initial dehydroxylation takes place at the same
ꢁ
temperature (DTGmax ¼ 265 C). Complex 1 loses CO molecules
ꢁ
ꢁ
Scheme 1. Synthesis of complex 1.
abruptly in the temperature range 150e190 C (DTGmax ¼ 170 C).
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

4
A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
the pure complex 1. Taking into account the mass losses noted
ꢁ
above for Zn,Al-CpMo up to 200 C, together with the Mo content of
2.7% determined by ICP-OES and the solid state NMR evidence for
some carbonate interference, the following formula may be pro-
posed for Zn,Al-CpMo: Zn Al (OH)12(CpMo(CO) CH COO)1.5(-
CO O (calcd: Mo, 13.5%; CO, 11.9%; H O, 6.8%).
1
4
2
3
2
3
)0.25$4H
2
2
The powder XRD pattern of Zn,Al-CpMo, recorded at ambient
ꢁ
temperature in the range of 3e70 2
q, presents several reflections
that are characteristic of hydrotalcite-like materials (Fig. 4). In
particular, the six symmetric, equally spaced peaks found between 3
ꢁ
and 30
2q can be assigned as the 00l basal reflections for an
ꢀ
expanded phase with an interlayer spacing (d003) of 20.7 A. Sub-
traction of 4.8 A for the thickness of the brucite-like layer gives an
interlayer distance (gallery height) of 15.9 A. This distance is
3 2
compatible with a bilayer-like arrangement of CpMo(CO) CH COO
ꢀ
ꢀ
ꢀ
anions within the interlayer region. Indeed, a similar gallery height
ꢀ
of 15.2 A was previously found for a Zn,Al-LDH intercalated by fer-
rocenecarboxylate anions (also prepared by coprecipitation from
aqueous solution) [25], which were proposed to adopt a bilayer-like
arrangement. A bilayer-like arrangement is necessary since these
organometallic anions are spatially incapable of balancing the host
layer charge of a Zn
a possible arrangement of CpMo(CO)
2
Al-LDH, if arranged in a monolayer. Fig. 5 shows
ꢀ
3 2
CH COO anions for which
the calculated basal spacing exactly matches the observed value. It is
likely that the guest species will be oriented in such a way as to
maximize hydrogen-bonding interactions between the carboxylate
groups and the layer hydroxyl groups, and position the hydrophobic
aromatic groups towards the center of the galleries, as far apart as
possible from the hydrophilic layer hydroxyls. In addition to the
basal reflections, the powder XRD pattern of Zn,Al-CpMo contains
ꢁ
very broad, asymmetric (hkl) reflections above ca. 2
q
¼ 30 , char-
acteristic of turbostratic layered structures and/or an intergrowth of
the rhombohedral and hexagonal polytypes [26]. The turbostratic
effect is caused by a decrease in the ordering along the stacking axis
Fig. 2. A. (a) 13C NMR spectrum of 1 in CDCl
CpMo. Spinning sidebands are indicated by asterisks. B. Al MAS NMR spectrum of
Zn,Al-CpMo.
. (b) 13C CP MAS NMR spectrum of Zn,Al-
3
27
ꢁ
The second decomposition step occurs between 350 and 500
C
ꢁ
(
DTGmax ¼ 470 C), leaving a residual mass of 52.3%. After that, a
ꢁ
third mass loss takes place above 650 C, which could correspond to
sublimation of a Mo phase. Overall, the TGA results indicate that
the thermal behavior of the intercalated complex parallels that of
x y
O
Fig. 4. Powder XRD patterns of (a) Zn,Al-NO
Selected reflections are assigned.
3 3
, (b) Zn,Al-CpMo, and (c) Zn,Al-ExCO .
3
Fig. 3. TGA curves of Zn,Al-CpMo (dd), Zn,Al-NO ($ $ $ $ $) and complex 1 (e e e).
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
5
ꢁ
CyO) yield within 30 min, at 55 C (Fig. 7, Table 1). Under the
(
conditions used the reaction mixture was always homogeneous. In
the absence of the molybdenum complex or without TBHP, Cy
epoxidation did not take place to a measurable extent throughout
24 h (less than 5% CyO yield), indicating that the simultaneous
presence of the molybdenum complex and TBHP is required for
catalytic epoxidation. The catalytic activity remained high at nearly
ambient temperature, giving 100% conversion within 90 min at
ꢁ
3
5 C. The catalytic results for 1 are superior to those reported for
other molybdenum tricarbonyl complexes bearing
a cyclo-
pentadienyl (Cp) ligand, used as precatalysts in the same reaction
under similar conditions. Some reported conversions at 1 h are: 91%
for CpMo(CO)
3
(CH
2
epC
6
H
4
eCO
2
CH
3
) [28]; ca. 85% for CpMo(-
Cl [30]; ca. 65% for CpMo(-
h -cyclopentadienyl molybdenum
CO)
CO)
3
Me [29]; ca. 95% for CpMo(CO)
3
5
3
Et [31]. Ansa-bridged
tricarbonyl complexes led to conversions in the range of 80e100%
at 4 h reaction of Cy [31e35].
Complex 1 was further investigated as a precatalyst in the
ꢁ
oxidation of
a
-pinene (Pin) with TBHP, at 55 C. To the best of our
knowledge, there is only one report in the literature on the epox-
idation of Pin in the presence of complexes of the type CpMo(CO)
3
X
(
(
X ¼ Cl [30]). For 1, the main reaction products were
a-pinene oxide
PinOx) and campholenic aldehyde (CPA), and their selectivities
depended on the conversion. These two products are used as in-
termediates or final products in the fragrance [36,37], aroma and
pharmaceutical industries [38]. PinOx and CPA were formed in 43%/
2% and 5%/35% selectivity, respectively, at 62%/64% conversion,
reached at 4 h/24 h reaction. Other reaction products included
trans-carveol, iso-pinocamphone and trans-pinocarveol, formed in
less than 10% total yield. These products, as well as CPA, may be
formed via the isomerization of PinOx [37e39]. The previously
ꢀ
3 2
Fig. 5. Model for the interlayer arrangement of CpMo(CO) CH COO anions in the
material Zn,Al-CpMo. Van der Waals spheres are drawn for the guest molecules. For
simplicity interlayer water molecules have been omitted.
3
studied complex CpMo(CO) Cl led to poorer catalytic results: PinOx
and CPA selectivities were ca. 20% at 40% conversion, reached at 6 h
reaction [30].
due to the loss of van der Waals interactions between adjacent
layers and the absence of a densely packed interlayer space formed
by high charge density anions such as carbonate.
Preliminary catalyst stability tests carried out for 1 were very
promising. The catalytic activity remained very high when the re-
agents Cy and TBHP were recharged (in the same amounts as those
The powder XRD pattern of Zn,Al-CpMo contains one additional
ꢀ
reflection at d ¼ 6.37 A which is unexpected for an LDH with
ꢁ
added initially) to the reactor after the first 24 h run, at 55 C,
ꢀ
d
003 ¼ 20.7 A. We initially assumed that this reflection could be due
leading to 98% CyO yield at 30 min reaction. Hence, the recovery
and reuse of the homogeneous catalyst was explored using an ionic
liquid (IL) as solvent, namely 1-butyl-3-methylimidazolium bis(-
to an impurity that was not removed during the washing step. To
further investigate, a separate experiment was performed in which
the material Zn,Al-CpMo (100 mg) was treated with a large excess
2
trifluoromethylsulfonyl)imide ([bmim]NTf ). The catalyst/IL
(
ca. 35ꢂ) of Na
2
CO
3
in water, with the intention of fully exchanging
mixture was recovered and reused after separating Cy and CyO by
solvent extraction, and subsequently adding the reagents Cy and
TBHP in amounts equal to those used in the first batch. The catalytic
activity was very high, leading to 99% conversion at 10 min reac-
tion, similar to that observed without a cosolvent (Fig. 7, Table 1).
ꢀ
intercalated CpMo(CO)
3 2
CH COO anions for carbonate anions. After
stirring overnight at room temperature, the resultant white solid
referred to as Zn,Al-ExCO ) was separated from the yellow solution
byfiltration, washed with waterand acetone, andvacuum-dried. The
powder XRD pattern of Zn,Al-ExCO is in agreement with patterns
reported for Zn,Al-CO
(
3
3
These results are superior to those reported for other CpMo(CO)
IL systems, tested in the same reaction under similar reaction
conditions: CpMo(CO) CF /[bmim]NTf led to 93% conversion at
4 h [15]; CpMo(CO) Me/[bmim]BF led to ca. 55% conversion at 1 h
29]; CpMo(CO) Me/[bmim]NTf led to ca. 13% conversion at 1 h,
Cl/[bmim]NTf led to 58% conversion at 4 h [40]. For
system, the reaction rate decreased in consecu-
X/
3
ꢀ
3
LDHs (d003 ¼ 7.64 A) [24,27], and neither
ꢀ
contains the 00l basal reflections for the 20.7 A phase nor the addi-
3
3
2
ꢀ
tional reflection at d ¼ 6.37 A (Fig. 4). Hence we tentatively conclude
2
3
4
that the latter reflection may arise due to the periodic distribution of
[
3
2
ꢀ
CpMo(CO) CH
3 2
COO anions within the gallery region.
and CpMo(CO)
3
2
Fig. 6 shows representative SEM images of Zn,Al-CpMo at
different magnifications. The powder comprises irregularly shaped
particles which, under higher magnification, revealed an ill-defined
morphology that in some regions approximated the classic “sand
rose” microstructure. Spatial elemental profiles for Zn, Al and Mo
the 1/[bmim]NTf
2
tive batch runs (Table 1). These results may be partly due to the
gradual removal of active species during the solvent extraction
process. Similar effects have been reported for the catalytic system
CpMo(CO) Me/[bmim]BF (removal of active species during solvent
3 4
extraction of the reaction products, accounting for a drop in reac-
tion rate in consecutive batch runs) [29]. The system CpMo(-
(
generated by employing an elemental mapping technique) showed
a homogeneous distribution of the elements across the particles.
3 3 2
CO) CF /[bmim]NTf led to similar CyO yield for 8 consecutive
3.2. Catalytic olefin epoxidation
runs; no description of the recycling procedure was given, making
it difficult to carry out comparative studies [15].
The epoxidation of cis-cyclooctene (Cy) with TBHP in the pres-
ence of 1 was very fast and selective, giving 100% cyclooctene oxide
In a different approach to improve the catalyst recovery and
reuse, complex 1 was immobilized in a Zn,Al layered double
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

6
A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
Fig. 6. SEM images of Zn,Al-CpMo. The EDS mapping microimages for the elements Zn, Al and Mo correspond to the SEM image labeled “SEM”.
hydroxide, as described above. Under similar reaction conditions to
ꢁ
those used for 1 (no cosolvent, 55 C), CyO was again the only
product formed, but the reaction was slower than that observed for
the homogeneous catalyst (Fig. 7, Table 1). The initial turnover
frequency (determined at 10 min reaction) was 592 and
ꢀ
1
ꢀ1
5
mol molMo
h
for 1 and Zn,Al-CpMo, respectively, and the latter
led to 65%/97% conversion at 6 h/24 h. The different catalytic results
for 1 and Zn,Al-CpMo may be partly due to differences in the
chemical nature of the metal species and/or diffusion limitations in
the case of the intercalated complex (which is a solideliquid
Table 1
Catalytic epoxidation of cis-cyclooctene with TBHP at 55 C.a
ꢁ
Precatalyst
Cosolventb
Catalytic
run
Reaction
time (h)
Conversion
(%)
Epoxide
selectivity
(
%)
1
ns
1
1
2
3
1
2
3
0.5
100
99
100
100
[bmim]NTf
[bmim]NTf
[bmim]NTf
2
2
2
0.17
6/24
6/24
6/24
6/24
6/24
65/85
26/56
65/97
10/34
7/25
100/100
100/100
100/100
100/100
100/100
Zn,Al-CpMo
ns
ns
ns
a
Reaction conditions: Mo:Cy:TBHP molar ratio ¼ 1:100:150, without additional
ꢁ
Fig. 7. Kinetic profiles of the catalytic epoxidation of cis-cyclooctene (Cy) with TBHP at
cosolvent, 55 C.
ꢁ
b
55
C in the presence of 1 (B), 1/IL (þ) and Zn,Al-CpMo (-).
ns ¼ no cosolvent was added.
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037

A.C. Gomes et al. / Journal of Organometallic Chemistry xxx (2013) 1e7
7
biphasic reaction system). After a 24 h batch run, the solid was
separated from the reaction mixture by centrifugation, washed
with n-hexane, dried at room temperature overnight, and charac-
terized by FT-IR spectroscopy (Fig.1). The bands associated with the
carbonyl ligands are reduced to very weak intensity, indicating that
extensive decarbonylation of the intercalated complex had taken
R&D Project PTDC/QEQ-SUP/1906/2012 and the research grant
BPD/UI89/4864/2013 (A.C.G.) provided within this project. The
NMR spectrometers are part of The National NMR Network
(REDE/1517/RMN/2005), supported by POCI 2010 and the FCT. The
authors also thank the Portuguese NMR Network (IST-UTL Center)
(RECI/QEQ-QIN/0189/2012) for providing access to the NMR fa-
cilities. The FCT and the European Union are acknowledged for a
post-doctoral grant to S.M.B (SFRH/BPD/46473/2008) cofunded
by MEC and the European Social Fund through the program POPH
of QREN.
ꢀ
1
place. Additionally, one new band at 895 cm and a shoulder at ca.
ꢀ1
9
30 cm are present, which are assigned to Mo]O stretching
vibrations, probably consistent with the formation of a dioxomo-
lybdenum(VI) species. However, despite the formation of oxidized
metal species, the reaction rate decreased in consecutive batch runs
(
Table 1). These results may be partly due to metal leaching, which
References
was checked by carrying out the leaching test (as described in detail
in the Experimental section). An increment in conversion between
2
reaction of 79% was observed for the leaching test, which is similar
to the 77% increment observed for the typical catalytic run (without
filtration) in the same time interval. Hence, active species were
leached from the solid into solution during the catalytic reaction.
[1] S.A. Hauser, M. Cokoja, F.E. Kühn, Catal. Sci. Technol. 3 (2013) 552.
[
[
2] S. Huber, M. Cokoja, F.E. Kühn, J. Organomet. Chem. (2013) in press.
3] H. Srour, P.L. Maux, S. Chevance, G. Simonneaux, Coord. Chem. Rev. 257 (2013)
h (instant of the separation of the solid by filtration) and 24 h
3030.
[4] J.-M. Brégeault, J. Chem. Soc. Dalton Trans. (2003) 3289.
[
[
[
[
5] M.N. Sheng, J.G. Zajacek, ARCO, GB Patent 1,136,923, 1968.
6] J. Kollar, Halcon, US Patent 3,351,635, 1967.
7] N. Grover, F.E. Kühn, Curr. Org. Chem. 16 (2012) 16.
8] T. Szyma nꢁ ska-Buzar, Curr. Org. Chem. 16 (2012) 3.
4
. Concluding remarks
[9] K.R. Jain, F.E. Kühn, Dalton Trans. (2008) 2221.
10] P.D. Vaz, C.D. Nunes, Curr. Org. Chem. 16 (2012) 89.
11] M. Abrantes, A. Sakthivel, C.C. Romão, F.E. Kühn, J. Organomet. Chem. 691
(2006) 3137.
[12] M. Abrantes, F.A.A. Paz, A.A. Valente, C.C.L. Pereira, S. Gago, A.E. Rodrigues,
J. Klinowski, M. Pillinger, I.S. Gonçalves, J. Organomet. Chem. 694 (2009) 1826.
[
[
Monomeric cyclopentadienyl molybdenum carbonyl complexes
5
of general formula (
catalysts for olefin epoxidation. In this work, we have found that
the complex CpMo(CO) CH COOH (1) is one of the most active
compounds in this class, using cyclooctene as a benchmark
substrate. Interesting catalytic results were also obtained with
a
h
-C
5
R
5
)Mo(CO)
3
X are known to be pre-
[
13] M. Abrantes, A.M. Santos, J. Mink, F.E. Kühn, C.C. Romão, Organometallics 22
2003) 2112.
[14] D. Betz, A. Raith, M. Cokoja, F.E. Kühn, ChemSusChem 3 (2010) 559.
3
2
(
[
15] S.A. Hauser, M. Cokoja, M. Drees, F.E. Kühn, J. Mol. Catal. A Chem. 363e364
2012) 237.
(
-pinene as substrate, giving the added-value products a-pinene
[
16] A.M. Martins, C.C. Romão, M. Abrantes, M.C. Azevedo, J. Cui, A.R. Dias,
M.T. Duarte, M.A. Lemos, T. Lourenço, R. Poli, Organometallics 24 (2005) 2582.
17] P.J. Costa, M.J. Calhorda, F.E. Kühn, Organometallics 29 (2010) 303.
18] M.J. Calhorda, P.J. Costa, Curr. Org. Chem. 16 (2012) 65.
19] A.C. Gomes, S.M. Bruno, C.A. Gamelas, A.A. Valente, M. Abrantes,
I.S. Gonçalves, C.C. Romão, M. Pillinger, Dalton Trans. 42 (2013) 8231.
20] J.K.P. Ariyaratne, A.M. Bierrum, M.L.H. Green, M. Ishaq, C.K. Prout,
M.G. Swanwick, J. Chem. Soc. A (1969) 1309.
21] A.C. Gomes, S.M. Bruno, A.A. Valente, I.S. Gonçalves, M. Pillinger, J. Organomet.
Chem. 744 (2013) 53.
[22] M. Abrantes, S. Gago, A.A. Valente, M. Pillinger, I.S. Gonçalves, T.M. Santos,
J. Rocha, C.C. Romão, Eur. J. Inorg. Chem. (2004) 4914.
oxide and campholenic aldehyde. To date, various approaches have
been reported to facilitate catalyst recycling and reuse using these
complexes, such as immobilization in ionic liquids [15,29,40], cy-
clodextrins [23,28,41] and porous inorganic supports [9,10]. In this
work we describe a completely new approach based on the inter-
calation of the anionic complex in a layered double hydroxide. The
resultant inorganiceorganometallic hybrid material is of significant
interest from a structural point of view due to the high loading of
complexes achieved in a bilayer-type arrangement with some ev-
idence for periodic ordering of the guests within the interlayer
region. Catalytic results for the epoxidation of cyclooctene were
disappointing, despite the fact that the in situ oxidative decarbon-
ylation of intercalated complexes (by reaction with TBHP) pro-
ceeded almost to completion during the first batch run, resulting in
the formation of oxidized metal species. Partial leaching of these
species is mainly responsible for the observed catalytic activity, and
we may assume that the performance of the material as a hetero-
geneous catalyst is conditioned by limited access of the substrate to
the active sites. Better results in terms of catalytic activity were
[
[
[
[
[
[
23] S.S. Braga, S. Gago, J.D. Seixas, A.A. Valente, M. Pillinger, T.M. Santos,
I.S. Gonçalves, C.C. Romão, Inorg. Chim. Acta 359 (2006) 4757.
[24] J.T. Kloprogge, L. Hickey, R.L. Frost, J. Solid State Chem. 177 (2004) 4047.
[
25] S. Gago, M. Pillinger, T.M. Santos, J. Rocha, I.S. Gonçalves, Eur. J. Inorg. Chem.
2004) 1389.
(
[
26] M. Bellotto, B. Rebours, O. Clause, J. Lynch, D. Bazin, E. Elkaïm, J. Phys. Chem.
100 (1996) 8527.
27] F. Thevenot, R. Szymanski, P. Chaumette, Clays Clay Miner. 37 (1989) 396.
28] A.C. Gomes, S.M. Bruno, C. Tomé, A.A. Valente, M. Pillinger, M. Abrantes,
I.S. Gonçalves, J. Organomet. Chem. 730 (2013) 116.
[
[
[29] M. Abrantes, P. Neves, M.M. Antunes, S. Gago, F.A.A. Paz, A.E. Rodrigues,
M. Pillinger, I.S. Gonçalves, C.M. Silva, A.A. Valente, J. Mol. Catal. A Chem. 320
(2010) 19.
[
[
[
30] A.A. Valente, J.D. Seixas, I.S. Gonçalves, M. Abrantes, M. Pillinger, C.C. Romão,
Catal. Lett. 101 (2005) 127.
31] J. Zhao, A.M. Santos, E. Herdtweck, F.E. Kühn, J. Mol. Catal. A Chem. 222 (2004)
2
obtained using the ionic liquid [bmim]NTf as solvent. However, the
reaction rate decreased in consecutive batch runs, most likely due
to the gradual removal of active species during the solvent
extraction process. In future work a wider range of ILs will be
studied with the aim of achieving better catalytic results with the
265.
32] J. Zhao, E. Herdtweck, F.E. Kühn, J. Organomet. Chem. 691 (2006) 2199.
[33] J. Zhao, K.R. Jain, E. Herdtweck, F.E. Kühn, Dalton Trans. (2007) 5567.
[34] A. Capapé, A. Raith, F.E. Kühn, Adv. Synth. Catal. 351 (2009) 66.
[
35] A. Capapé, A. Raith, E. Herdtweck, M. Cokoja, F.E. Kühn, Adv. Synth. Catal. 352
2010) 547.
36] I.O. Jawonisi, G.I. Adoga, Br. J. Pharmacol. Toxicol. 4 (2013) 155.
3 2
title compound CpMo(CO) CH COOH.
(
[
Acknowledgments
[37] V.V. Costa, K.A. da Silva Rocha, L.F. de Sousa, P.A. Robles-Dutenhefner,
E.V. Gusevskay, J. Mol. Catal. A Chem. 345 (2011) 69.
[
[
38] I.V. Il’ina, K.P. Volcho, N.F. Salakhutdinov, Russ. J. Org. Chem. 44 (2008) 1.
39] W.F. Holderich, J. Roseler, G. Heitmann, A.T. Liebens, Catal. Today 37 (1997)
353.
We are grateful to the Fundação para a Ciência e a Tecnologia
FCT), QREN, Fundo Europeu de Desenvolvimento Regional
FEDER), COMPETE, the European Union, the Associate Laboratory
CICECO (PEst-C/CTM/LA0011/2013) and CQE (PEst-OE/QUI/
UI0100/2013) for continued support and funding, including the
(
(
[
40] F.E. Kühn, J. Zhao, M. Abrantes, W. Sun, C.A.M. Afonso, L.C. Branco,
I.S. Gonçalves, M. Pillinger, C.C. Romão, Tetrahedron Lett. 46 (2005) 47.
41] S.S. Balula, A.C. Coelho, S.S. Braga, A. Hazell, A.A. Valente, M. Pillinger,
J.D. Seixas, C.C. Romão, I.S. Gonçalves, Organometallics 26 (2007) 6857.
[
Please cite this article in press as: A.C. Gomes, et al., Journal of Organometallic Chemistry (2013), http://dx.doi.org/10.1016/
j.jorganchem.2013.10.037