Tris(pyrazolyl)methane molybdenum tricarbonyl complexes as catalyst precursors for olefin epoxidation

Article abstract of DOI:10.1016/j.molcata.2012.12.010

The molybdenum tricarbonyl complexes [Mo(CO)₃(HC(3,5-Me ₂pz)₃)] (1) and [Mo(CO)₃(HC(pz)₃)] (2) (HC(3,5-Me₂pz)₃ = tris(3,5-dimethyl-1-pyrazolyl) methane, HC(pz)₃ = tris(1-pyrazolyl)methane) were obtained in good yields by the microwave-assisted reaction of Mo(CO)₆ with the respective organic ligand. Complete oxidative decarbonylation of 1 and 2 was achieved by reaction with excess tert-butylhydroperoxide (TBHP) in 1,2-dichloroethane at 55 °C. For complex 1, the (μ₂-oxo) bis[dioxomolybdenum(VI)] hexamolybdate of composition [{MoO₂(HC(3,5- Me₂pz)₃)}₂(μ₂-O)][Mo ₆O₁₉] (3) was obtained in good yield, and its structure was determined by single crystal X-ray diffraction. The compound (4) obtained by oxidative decarbonylation of 2 was not unambiguously identified, but may be chemically analogous to 3. Compounds 1-4 were examined for the first time as homogeneous (pre)catalysts for the epoxidation of olefins with TBHP, using different types of cosolvents at 55 °C. During the catalytic reactions 1 and 2 transform in situ into 3 and 4, respectively, and the latter two are fairly stable catalysts. Catalytic tests and characterization studies of the recovered catalysts were carried out in an attempt to understand the kinetic differences observed between the compounds prepared in situ during the catalytic reaction and those prepared prior to the catalytic reaction, from the same precursor complex.

Full text of DOI:10.1016/j.molcata.2012.12.010

Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Tris(pyrazolyl)methane molybdenum tricarbonyl complexes as catalyst
precursors for olefin epoxidation
Ana C. Gomesa, Patrícia Nevesa, Sónia Figueiredob, José A. Fernandesa, Anabela A. Valentea,
Filipe A. Almeida Paza, Martyn Pillingera,∗, André D. Lopesb, Isabel S. Gonc¸ alvesa,∗
a Department of Chemistry, CICECO, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal
b Faculty of Science and Technology, Department of Chemistry, Biochemistry and Pharmacy, Centro de Investigac¸ ão em Química do Algarve, University of the Algarve, Campus de
Gambelas, 8005-139 Faro, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 October 2012
Received in revised form 3 December 2012
Accepted 10 December 2012
Available online 10 January 2013
The molybdenum tricarbonyl complexes [Mo(CO)3(HC(3,5-Me2pz)3)] (1) and [Mo(CO)3(HC(pz)3)]
(2) (HC(3,5-Me2pz)3 = tris(3,5-dimethyl-1-pyrazolyl)methane, HC(pz)3 = tris(1-pyrazolyl)methane) were
obtained in good yields by the microwave-assisted reaction of Mo(CO)6 with the respec-
tive organic ligand. Complete oxidative decarbonylation of
1 and 2 was achieved by reaction
with excess tert-butylhydroperoxide (TBHP) in 1,2-dichloroethane at 55 ◦C. For complex 1, the
(␮2-oxo)bis[dioxomolybdenum(VI)] hexamolybdate of composition [{MoO2(HC(3,5-Me2pz)3)}2(␮2-
O)][Mo6O19] (3) was obtained in good yield, and its structure was determined by single crystal X-ray
diffraction. The compound (4) obtained by oxidative decarbonylation of 2 was not unambiguously iden-
tified, but may be chemically analogous to 3. Compounds 1–4 were examined for the first time as
homogeneous (pre)catalysts for the epoxidation of olefins with TBHP, using different types of cosol-
vents at 55 ◦C. During the catalytic reactions 1 and 2 transform in situ into 3 and 4, respectively, and the
latter two are fairly stable catalysts. Catalytic tests and characterization studies of the recovered catalysts
were carried out in an attempt to understand the kinetic differences observed between the compounds
prepared in situ during the catalytic reaction and those prepared prior to the catalytic reaction, from the
same precursor complex.
Keywords:
Molybdenum
Tris(pyrazolyl)methane ligands
Oxidative decarbonylation
Homogeneous catalysis
Olefin epoxidation
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
photoactivated release of molecular carbon monoxide, a potential
therapeutic agent (M = Mn) [5c–5f], and as radiopharmaceuticals
Tris(1-pyrazolyl)methane (tpm), HC(pz)3, acts as
a
neu-
for diagnostic (99mTc) or therapeutic (186/188Re) applications in
nuclear medicine [6].
tral tridentate ligand, isoelectronic and isosteric with the
tris(pyrazolyl)hydroborate ion, and displays a wealth of coor-
dination chemistry [1]. In his landmark paper [2], Trofimenko
reported the reaction of HC(pz)3 with BrMn(CO)5 to give the cation
[Mn(CO)3(HC(pz)3)]+, and the reaction of the same ligand with
group 6 hexacarbonyls to form the poorly soluble compounds
[M(CO)3(HC(pz)3)] (M = Cr, Mo or W). The reaction of tris(3,5-
dimethyl-1-pyrazolyl)methane, HC(3,5-Me2pz)3, with Mo(CO)6 to
give [Mo(CO)3(HC(3,5-Me2pz)3)] was also reported. In recent years,
the discovery of improved synthetic procedures based on phase-
transfer reactions [3] has led to the preparation of a series of group 6
(Mo) and group 7 (Mn, Tc, and Re) tricarbonyl complexes anchored
by tpm ligands [3b,4–6]. These complexes are of interest as start-
ing materials (M = Mo, Mn) [2,4a–4c,5b], as candidates for the
[Mo(CO)3(tpm)] complexes are useful starting materials for
the preparation of other complexes in higher oxidation states,
including molybdenum(VI) complexes. For example, oxidation
of [Mo(CO)3(HC(3,5-Me2pz)3)] with nitric acid yields (HC(3,5-
Me2pz)3)MoO3 [4a], and the reaction of Li[Mo(CO)3(tpms)]
(tpms = tris(1-pyrazolyl)methanesulfonate) with AgBF4 in air gives
dinuclear [{MoO2(tpms)}2(␮2-O)] [4c]. Oxomolybdenum(VI) com-
plexes such as these are potentially interesting oxidation catalysts.
Kühn and co-workers reported that tpm and its 3,5-dimethyl
derivative form quite active and selective catalysts of the type
[MoO2X(L)]X (X = Cl, Br) for the epoxidation of olefins with tert-
butylhydroperoxide (TBHP) as oxidant [7]. Later on, we carried
out an in-depth study of the performance of [MoO2Cl(HC(3,5-
Me2pz)3)]BF4 in catalytic olefin epoxidation [8], which revealed
that the presence of residual amounts of water in the reac-
tion system could result in the formation of the catalytically
active dinuclear species [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)](BF4)2
and [Mo2O3(O2)2(␮2-O)(H2O)(HC(3,5-Me2pz)3)].
∗
Corresponding authors. Fax: +351 234 370084.
E-mail addresses: mpillinger@ua.pt (M. Pillinger), igoncalves@ua.pt
(I.S. Gonc¸ alves).
1381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.12.010

A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
65
During the last decade, a broad variety of molybdenum car-
bonyl complexes have been applied directly as catalyst precursors
for olefin epoxidation [9]. Complete oxidative decarbonylation
occurs in situ by reaction with the oxidant (usually TBHP), result-
ing in the formation of oxomolybdenum(VI) compounds ranging
from mononuclear species such as CpMoO2Cl from CpMo(CO)3Cl
[9a], dinuclear species such as [{MoO2(DAB)}2(␮2-O)2] from
[Mo(ꢀ3-allyl)Cl(CO)2(DAB)] (DAB = diazabutadiene) [9o], tetranu-
clear species such as [Mo4O12(pypz)4] from cis-[Mo(CO)4(pypz)]
(pypz = 2-[3(5)-pyrazolyl]pyridine) [9p], octanuclear species such
as [Mo8O24(di-tBu-bipy)4] from cis-[Mo(CO)4(di-tBu-bipy)] (di-
tBu-bipy = 4,4ꢀ-di-tert-butyl-2,2ꢀ-bipyridine) [9q], and polymeric
species such as [MoO3(bipy)] from cis-[Mo(CO)4(bipy)] (bipy = 2,2ꢀ-
bipyridine) [9q].
In the present work, we report for the first time on the use of the
complexes [Mo(CO)3L] with L = tpm and its 3,5-dimethyl deriva-
tive as catalyst precursors in the epoxidation of different types of
olefins with TBHP using different cosolvents. Attempts to iden-
tify the active species formed and understand the differences in
catalytic performances between the compounds prepared in situ
during the catalytic reaction and those prepared prior to the cat-
alytic reaction, from the same precursor complex, have also been
made and discussed.
was used for temperature measurement. The reaction temper-
ature of 110 ◦C was reached and maintained using a dynamic
control mode in which the power (max. 150 W) was automatically
adjusted.
2.2. Tris(1-pyrazolyl)methane (HC(pz)3))
The synthesis method used for HC(pz)3 was based on the lit-
erature procedure [3a]. A mixture of pyrazole (8.17 g, 0.12 mol),
anhydrous K2CO3 (82.93 g, 0.60 mol), (Bu4N)HSO4 (2.04 g, 6 mmol)
and CHCl3 (100 mL) was stirred and refluxed for 19 h. The resul-
tant brown mixture was filtered and the solid washed with hot
CH2Cl2. The filtrate and washings were combined, and activated
charcoal was added. After filtering again, the filtrate was con-
centrated at reduced pressure and hexane/CH2Cl2 were added.
Storage of the mixture in a fridge overnight gave a precip-
itate which was collected by filtration, washed with diethyl
ether/hexane (1:1), and finally vacuum-dried. Yield: 1.5 g, 17.5%.
The 1H and FT-IR spectra were in agreement with published data
[3a,3b].
2.3. General procedure for the preparation of [Mo(CO)3L]
Mo(CO)6 (0.50 g, 1.89 mmol), HC(pz)3 or HC(3,5-Me2pz)3
(1.89 mmol), and toluene (20 mL) were transferred to a 35 mL glass
vessel and the mixture was heated in a microwave oven at 110 ◦C
for 3 h. The mixture was cooled to room temperature and then
transferred to a Schlenk tube, and the resultant lemon yellow solid
was cannula filtered, washed with n-hexane (3 × 20 mL) and diethyl
ether (2 × 20 mL), and finally vacuum-dried.
2. Experimental
2.1. General considerations
Microanalyses for CHN and ICP-OES analysis for Mo were
performed at the University of Aveiro. Powder X-ray diffraction
(XRD) data were collected on a Philips X’pert MPD diffractometer
equipped with an X’Celerator detector, a graphite monochroma-
2.3.1. [Mo(CO)3(HC(3,5-Me2pz)3)] (1)
Yield: 0.62 g, 69%. Anal. Calcd for C19H22MoN6O3 (478.36): C,
47.71; H, 4.64; N, 17.57. Found: C, 47.75; H, 4.70; N, 17.40. FT-IR
(KBr, cm−1): 3127 (w), 1902 (vs, ꢃ(CO)), 1765 (vs, ꢃ(CO)), 1563 (m),
1458 (m), 1401 (s), 1389 (s), 1372 (m), 1303 (s), 1254 (s), 1042 (m),
1034 (m), 978 (m), 980 (s), 860 (m), 766 (s), 702 (s).
˚
tor (Cu K␣ X-radiation filtered by Ni (ꢁ = 1.5418 A)) and a flat-plate
sample holder, in a Bragg–Brentano para-focusing optics config-
uration (40 kV and 50 mA). Samples were step-scanned in 0.04◦
2ꢂ steps with a counting time of 50 s per step. SEM images were
recorded using a Hitachi SU-70 UHR Schottky FE-SEM instru-
ment.
2.3.2. [Mo(CO)3(HC(pz)3)] (2)
Transmission FT-IR spectra were measured on a Mattson 7000
spectrometer. Attenuated total reflectance (ATR) FT-IR spectra
were measured on a Bruker optics Tensor 27 equipped with a
Specac Golden Gate Mk II ATR accessory having a diamond top-
plate and KRS-5 focusing lenses. FT-Raman spectra were recorded
on a RFS-100 Bruker FT-Spectrometer equipped with a Nd:YAG
laser with an excitation wavelength of 1064 nm. 1H NMR spectra
were measured with a Bruker CXP 300 instrument.
Yield: 0.52 g, 70%. Anal. Calcd for C13H10MoN6O3 (394.20): C,
39.61; H, 2.56; N, 21.32. Found: C, 39.66; H, 2.54; N, 21.21. FT-IR
(KBr, cm−1): 3149 (m), 3134 (m), 2956 (m), 1903 (vs, ꢃ(CO)), 1767
(vs, ꢃ(CO)), 1515 (w), 1439 (s), 1402 (s), 1287 (s), 1244 (s), 1221
(m), 1096 (s), 1054 (s), 980 (m), 856 (m), 789 (m), 766 (s), 607 (m),
521 (m), 418 (m).
2.4. General procedure for the complete oxidative
decarbonylation of [Mo(CO)3L]
All preparations and manipulations were carried out using
standard Schlenk techniques under nitrogen. Solvents were dried
by standard procedures, distilled under nitrogen, and kept over
A solution of 5–6 M TBHP in n-decane (6.93 mL, 0.038 mol, for
the reaction with 1; 8.42 mL, 0.046 mol, for the reaction with 2) was
added dropwise to a magnetically stirred suspension of 1 (0.12 g,
0.25 mmol) or 2 (0.12 g, 0.30 mmol) in dry 1,2-dichloroethane
(20 mL), and the mixture was stirred at 55 ◦C for 24 h. The resul-
tant yellow precipitate was filtered, washed with diethyl ether
(3 × 20 mL), and vacuum-dried.
˚
4 A molecular sieves. 1H-pyrazole (98%, Sigma–Aldrich), 3,5-
dimethylpyrazole (99%, Sigma–Aldrich), Mo(CO)6 (Fluka), K2CO3,
tetrabutylammonium hydrogen sulfate (98%, Sigma–Aldrich),
5–6 M TBHP in n-decane (Sigma–Aldrich), cis-cyclooctene (95%,
Sigma–Aldrich), diethyl ether (99.5%, Sigma–Aldrich), pentane
(98%, Sigma–Aldrich), n-hexane (99%, Sigma–Aldrich), acetonitrile
(CH3CN, 99.5%, Sigma–Aldrich) and toluene (99.5%, Panreac) were
purchased from commercial sources. 1,2-Dichloroethane (DCE,
Sigma–Aldrich, 99%) was dried using CaH2 at room temperature,
2.4.1. Reaction of 1 with TBHP to give
[{MoO2(HC(3,5-Me2pz)3)}2(ꢄ2-O)][Mo6O19] (3)
˚
followed by distillation and storage over activated 4 A molecular
Yield: 0.04 g, 73%. Anal. Calcd for C32H44Mo8N12O24 (1748.28):
C, 21.98; H, 2.54; N, 9.61. Found: C, 21.98; H, 2.65; N, 9.75. FT-
IR (KBr, cm−1): 3142 (w), 2967 (w), 2925 (w), 1563 (m), 1461
(m), 1447 (m), 1415 (m), 1390 (m), 1301 (m), 1265 (m), 1234
(w), 1124 (w), 1097 (w), 1045 (m), 953 (vs, ꢃ(Mo O)), 945 (sh,
sieves. The ligand tris(3,5-dimethyl-1-pyrazolyl)methane was pre-
pared as described in the literature [3b].
The microwave-assisted syntheses were carried out in a Dis-
cover S-Class (CEM Corporation, USA) microwave oven at 2.45 GHz
under stirring and simultaneous cooling with compressed air
(20 psi) to prevent bulk overheating. A vertical focused IR sensor
ꢃ(Mo O)), 917 (vs, ꢃ(Mo O)), 861(m), 797 (vs, ꢃ(Mo
O Mo)), 772
(vs, ꢃ(Mo Mo)), 702 (m), 668 (w), 634 (w), 603 (m), 595 (w),
O

66
A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
490 (w), 433 (br), 394 (w), 355 (w), 314 (w). FT Raman (cm−1):
2927 (s), 1558 (w), 1461 (w), 1377 (w), 1298 (w), 1159 (w), 1120
(w), 982 (vs), 945 (vs), 920 (m), 654 (w), 595 (m), 406 (w), 332 (w).
1H NMR (300 MHz, 25 ◦C, DMSO-d6): ı = 8.30 (s, 1H, CH), 5.95 (s, 3H,
4-H(pz)), 2.07 (s, 9H, 3,5-Me), 1.94 (s, 9H, 3,5-Me). Single crystals of
3 suitable for X-ray diffraction were obtained by recrystallisation
from acetonitrile/diethyl ether.
Table 1
Crystal and structure refinement data for [{MoO2(HC(3,5-Me2pz)3)}2(␮2-
O)][Mo6O19]·3CH3CN (3·3CH3CN).
3·3CH3CN
Formula
Formula weight
Crystal system
Space group
a (Å)
C38H53Mo8N15O24
1871.47
Monoclinic
P21/n
10.4581(13)
18.335(2)
2.4.2. Reaction of 2 with TBHP to give 4
b (Å)
Yield: 0.06 g. Anal. found: C, 16.96; H, 1.58; N, 10.14. FT-IR
(KBr, cm−1): 3124 (m), 3107 (m), 2981 (m), 1514 (m), 1446 (m),
1408 (s), 1365 (w), 1282 (s), 1257 (m), 1226 (m), 1193 (w), 1099
(m), 1066 (m), 972 (m), 937 (m), 910 (m), 883 (w), 860 (m), 798
(vs), 786 (vs), 753 (m), 646 (w), 607 (w), 586 (w), 575 (w), 501
(w), 401 (w). 1H NMR (300 MHz, 25 ◦C, DMSO-d6): ı = 8.97 (s, 1H,
CH), 7.89 (d, 3H, 3-H(pz)), 7.67 (d, 3H, 5-H(pz)), 6.42 (dd, 3H,
4-H(pz)).
c (Å)
31.623(4)
97.327(5)
6014.1(13)
4
2.067
ˇ (◦)
Volume (Å3)
Z
Dc (g cm−3
)
␮(Mo K␣) (mm−1
Crystal size (mm)
Crystal type
)
1.699
0.10 × 0.07 × 0.03
Yellow block
3.57–29.13
−14 ≤ h ≤ 14
−25 ≤ k ≤ 25
−43 ≤ l ≤ 41
56,527
ꢂ range (◦)
Index ranges
2.5. X-ray crystallography
Reflections collected
Independent reflections
Completeness to ꢂ = 29.13◦
Final R indices [I > 2ꢅ(I)]a,b
Final R indices (all data)a,b
Weighting schemec
15,913 [Rint = 0.0571]
98.3%
Single-crystals
of
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-
O)][Mo6O19]·3(CH3CN) were manually harvested from the
crystallisation vial and immediately immersed in FOMBLIN Y
perfluoropolyether vacuum oil (LVAC 140/13, Sigma–Aldrich) to
avoid degradation caused by evaporation of the solvent. Crystals
were mounted on Hampton Research CryoLoops with the help of
a Stemi 2000 stereomicroscope equipped with Carl Zeiss lenses
[10]. Data were collected on a Bruker X8 Kappa APEX II CCD
area-detector diffractometer (Mo K␣ graphite-monochromated
R1 = 0.0415, wR2 = 0.0780
R1 = 0.0713, wR2 = 0.0888
m = 0.0296, n = 1.4774
−3
˚
Largest diff. peak and hole
0.990 and −0.965 eA
ꢀ
ꢀ
ꢁ
ꢁ
a
R1 =
|Fo| − |Fc|
/
|Fo|.
ꢂ
ꢃ
ꢄ
ꢃ
ꢄ
ꢀ
ꢀ
2
2
wR2 =
w(Fo2 − Fc2
)
/
w(Fo2
)
.
b
c
w = 1/[ꢅ2(Fo2) + (mP)2 + nP] where P = (Fo2 + 2Fc2)/3.
˚
radiation, ꢁ = 0.71073 A) controlled by the APEX2 software
package [11], and equipped with an Oxford Cryosystems Series
700 cryostream monitored remotely using the software inter-
face Cryopad [12]. Images were processed using the software
package SAINT+ [13], and data were corrected for absorption by
the multi-scan semi-empirical method implemented in SADABS
[14].
bar (1000 rpm), immersed in a thermostated oil bath, and oper-
ating under autogenous pressure. Typically, the initial molar
ratios of molybdenum:Cy:TBHP were 1:100:152 and pre-dried 1,2-
dichloroethane (DCE) was used as co-solvent (ca. 1.8 M Cy in DCE).
The reagents, co-solvent and the reactor walls were preheated for
10 min at the reaction temperature of 55 ◦C. The last reactant loaded
to the reactor was TBHP and that instant was taken as time zero for
the kinetic studies.
The progress of the catalytic reactions was monitored by gas
chromatography (Varian 3900 GC equipped with a DB-5 capil-
lary column, 30 m × 0.25 mm × 0.25 ␮m, and a FID detector), using
undecane as internal standard.
For both precursor compounds the reaction mixtures were
biphasic solid–liquid; after 24 h reaction the solids were separated
from reaction mixture by centrifugation, washed with pentane and
dried at room temperature overnight under air atmosphere. The
recovered solids were characterised and/or used in consecutive
batch runs.
The CatFilt assay consisted of separating the solid from the
liquid phase after an established reaction time (at the reaction
temperature), using a 0.2 ␮m nylon membrane, and leaving the
filtrate under stirring until 24 h reaction (at the reaction temper-
ature). GC was used to determine the increment in conversion
(denoted ꢆCatFilt) between the instant the filtration operation
was performed and the final reaction time. The values of ꢆCatFilt
are compared with the increment in conversion (denoted ꢆCat)
observed for the same period of time in the presence of cat-
alyst (i.e., without filtration) under similar reaction conditions;
the reaction is considered homogeneous in nature if the ratio
ꢆCatFilt/ꢆCat = 1. For 0 < ꢆCatFilt/ꢆCat < 1 a homogeneous cat-
alytic contribution exists, but it does not necessarily imply that a
heterogeneous catalytic contribution exists since the dissolution
of the metal species can be rate-limiting to the overall catalytic
process.
The structure was solved using the Patterson synthesis algo-
rithm implemented in SHELXS-97 [15], which allowed the
immediate location of the molybdenum centres and most of the
heaviest atoms. The remaining non-hydrogen atoms were located
from difference Fourier maps calculated from successive full-
matrix least-squares refinement cycles on F2 using SHELXL-97
[15a,16]. All non-hydrogen atoms were successfully refined using
anisotropic displacement parameters. Hydrogen atoms bound to
carbon were located at their idealised positions using appropri-
ate HFIX instructions in SHELXL: 33 for the terminal CH3 methyl
groups, 13 for the CH aliphatic group and 43 for the CH of
the pyrazole rings. All these atoms were included in subsequent
refinement cycles in riding motion approximation with isotropic
thermal displacement parameters (Uiso) fixed at 1.5 (only for CH3)
or 1.2 × Ueq of the parent atoms.
The last differe−n3ce Fourier map synthesis showed the high-
est peak (0.990 eA ) and deepest hole (−0.965 eA−3) located at
˚
˚
˚
0.70 and 0.69 A from Mo5, respectively, which is one of the metal
centres of the polyoxometallate anion. Information concerning
crystallographic data collection and structure refinement details is
summarised in Table 1. CCDC-878880 contains the supplementary
crystallographic data (including structure factors) for this paper.
These data can be obtained free of charge from The Cambridge Crys-
tallographic Data Centre via www.ccdc.cam.ac.uk/data request/cif.
2.6. Catalytic olefin epoxidation
The liquid phase epoxidation reaction of cis-cyclooctene (Cy)
with TBHP was carried out in a 5 mL borosilicate batch reac-
tor equipped with a valve for sampling and a magnetic stirring

A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
67
Fig. 1. Representative SEM images of (a) [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19] (3), (b) the recovered solid 1*, (c) 4, and (d) the recovered solid 2*.
3. Results and discussion
TBHP (152 equiv., 5–6 M in decane) in dry 1,2-dichloroethane at
55 ◦C for 24 h. The resultant beige (for 3, starting from 1) or pale
yellow (for 4, starting from 2) solids were isolated by filtration
and characterised by elemental analysis, powder XRD, SEM, FT-
IR, FT-Raman and 1H NMR spectroscopy. SEM examinations for 3
and 4 revealed irregular plate-like agglomerates 0.2–10 ␮m in size
(Fig. 1).
Recrystallisation of 3 from acetonitrile/diethyl ether gave single
crystals suitable for X-ray diffraction (XRD). As will be described
below, the XRD analyses identified the structure as the (␮2-
oxo)bis[dioxomolybdenum(VI)] hexamolybdate of composition
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19] (Scheme 1).
The FT-IR spectra of 3 and 4 show the absence of ꢃ(CO)
bands, indicating that oxidative decarbonylation was complete.
For 3, very strong bands at 797 and 953 cm−1 are assigned
3.1. Catalyst preparation
The molybdenum tricarbonyl complexes [Mo(CO)3(HC(3,5-
Me2pz)3)] (1) and [Mo(CO)3(HC(pz)3)] (2) were obtained in good
yields of ca. 70% by the microwave-assisted reaction of Mo(CO)6
with the respective tris(1-pyrazolyl)methane ligand. Compared
with the conventional synthesis method, which involves reflux-
ing a mixture of the organic ligand, Mo(CO)6 and DMF [2,4a],
the microwave-assisted method affords the desired compounds in
comparable yields but in shorter reaction times (3 h vs. 8 h). The
solid-state FT-IR spectra of 1 and 2 show strong ꢃ(CO) bands at
1903 and 1766 cm−1, in agreement with that previously reported
[2].
With the aim of obtaining a better understanding of the molyb-
denum species formed upon use of complexes 1 and 2 as catalyst
precursors for the epoxidation of olefins, separate reactions were
performed in which 1 and 2 were treated with a large excess of
to ꢃ(Mo O Mo) and ꢃ(Mo O), respectively, of the hexamolyb-
date anion [17]. In accordance with that previously reported
for the (␮2-oxo)bis[dioxomolybdenum(VI)] tetrafluoroborate salt
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)](BF4)2 [8], a strong band at
Scheme 1. Synthesis of 3 by oxidative decarbonylation of 1.

68
A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
Table 2
Table 3
Selected bond lengths (Å) and angle ranges (◦; for all Mo6+ environments in each moi-
ety) for the {MoN3O3} and {MoO6} octahedral coordination environments present
in [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19]·3CH3CN (3·3CH3CN).
Weak hydrogen bonding geometry (intranuclear distances in Å and interaction
angles in degrees) for the strongest contacts present in [{MoO2(HC(3,5-
Me2pz)3)}2(␮2-O)][Mo6O19]·3CH3CN (3·3CH3CN).a
For [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)]2+
D
H· · ·A
dD· · ·A
∠(DHA)
C1 H1A· · ·O3†
3.343(5)
3.352(6)
3.320(6)
3.337(5)
3.286(5)
3.232(5)
3.185(5)
3.138(5)
3.488(6)
3.421(8)
3.194(5)
3.211(6)
3.513(7)
3.432(7)
3.297(7)
3.390(7)
136
140
132
132
135
141
134
137
157
164
166
157
157
167
138
140
Mo1 O1
Mo1 O2
Mo1 O3
Mo1 N1
Mo1 N3
Mo1 N5
Mo2 O1
Mo2 O4
Mo2 O5
Mo2 N7
Mo2 N9
Mo2 N11
1.891(3)
1.686(3)
1.687(3)
2.203(3)
2.328(3)
2.306(3)
1.886(3)
1.691(3)
1.689(3)
2.346(3)
2.230(3)
2.303(3)
102.92(13)–104.53(15)
74.11(11)–76.93(11)
84.91(11)–91.78(12)
157.15(12)–163.95(12)
C6 H6A· · ·O2†
C6 H6C· · ·O5†
C16 H16A· · ·O1†
C21 H21C· · ·O4†
C26 H26A· · ·O4†
C26 H26C· · ·O3†
C3 H3· · ·O7i
C6 H6B· · ·O12ii
C10 H10B· · ·N15
C13 H13· · ·O21iii
C28 H28· · ·O20iv
C30 H30C· · ·O24iv
C34 H34C· · ·O23iii
C36 H36C· · ·O10v
C38 H38B· · ·O2ii
O
N
O
O
Mo
Mo
Mo N (cis)
O
N
Mo N (trans)
a
Symmetry operations used to generate equivalent atoms: (i) −x, −y − 1, −z; (ii)
−x + 1, −y + 1, −z; (iii) −x + 1/2, y + 1/2, −z + 1/2; (iv) x, y + 1, z; (v) x + 1, y, z.
2−
]
For the Lindqvist [Mo6O19
†
Intramolecular (cation) weak hydrogen bonds.
Mo (␮6-O)
2.309(2)–2.326(2)
Mo (␮2-O)
Mo Ot
1.853(3)–1.999(3)
1.674(3)–1.680(3)
discussed here (Table 2 summarises the most relevant geometrical
parameters for this anion).
(␮2-O) Mo (␮6-O)
(␮2-O) Mo (␮2-O) (cis)
(␮2-O) Mo (␮2-O) (trans)
Ot Mo (␮2-O)
75.08(10)–78.20(10)
81.91(12)–92.08(13)
152.41(12)–153.49(12)
101.49(14)–105.12(15)
176.41(13)–177.95(13)
The dicationic entity, [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)]2+, has
already been described by some of us [8] and constitutes another
example belonging to the series of compounds of general formula
[Mo2O4(␮2-O)(L)2]n+ (n = 0, 2), in which L is either a tris-pyrazolyl
methane/hydroborate [4c,8,20] or a triazacyclononane [21]. This
cation comprises two crystallographically independent Mo6+ cen-
Ot Mo (␮6-O)
917 cm−1 and a shoulder at ca. 945 cm−1 are assigned to the
asymmetric and symmetric stretching vibrations, respectively, of
the cis-[MoO2]2+ core, and the very strong band at 772 cm−1 is
tres bridged by a ␮2-oxido ligand, with a Mo
O
Mo kink angle of
˚
164.90(16)◦ and a Mo· · ·Mo intermetallic distance of 3.7442(7) A.
These values correspond to the smallest angle and shortest distance
when compared to the analogous measurements in the aforemen-
tioned series of compounds (values found in the ca. 167–180◦
assigned as the asymmetric stretching mode, ꢃ(Mo
O Mo)asym,
associated with the Mo2O bridge. Overall, these Mo O stretching
vibrations are in agreement with those reported by Pedrosa et al.
for several (␮2-oxo)bis[dioxomolybdenum(VI)] hexamolybdates
of composition [Mo2O5L6][Mo6O19] (L = H2O, dimethylformam-
ide, dimethylacetamide, dimethyl sulfoxide, dibutyl sulfoxide,
tributylphosphane oxide, triphenylphosphane oxide, hexam-
ethylphosphoramide) [18]. Several intense bands between 1200
and 1600 cm−1 for 3 are assigned to internal ligand modes from
the tridentate tris(3,5-dimethylpyrazolyl)methane ligand (e.g., the
characteristic pyrazolyl ring stretching vibration at 1563 cm−1).
The 1H NMR spectrum of 3 in (CD3)2SO exhibits only one set of
signals for the organic ligand, indicating the presence of C2 sym-
metry, with each tpm ligand coordinated to the Mo atom through
the three equivalent pyrazolyl groups.
The characterisation data for 4 are somewhat similar to those for
3 in that several bands due to tpm and Mo O modes are observed
in the IR spectrum in the 1200–1600 cm−1 and 700–1000 cm−1
ranges, respectively, and the 1H NMR spectrum contains only
one set of tpm signals. However, at the time of writing, it is not
conclusively clear that 4 is a (␮2-oxo)bis[dioxomolybdenum(VI)]
hexamolybdate analogous to 3. This will be further discussed in
Section 3.3.
˚
and ca. 3.75–3.80 A ranges, respectively) [4c,8,20,21]. Besides the
␮2-oxido ligand, each metal centre is further coordinated to two
mutually cis terminal oxido moieties and to the facially tricoor-
dinated tris(3,5-dimethylpyrazolyl)methane ligand. For the two
{MoN3O3} environments the Mo O and Mo N distances were
˚
˚
found in the 1.686(3)–1.891(3) A and 2.203(3)–2.346(3) A ranges,
respectively. The coordination geometry of the metal centres can be
envisaged as distorted octahedra, with the cis and trans octahedral
angles ranging from 74.71(11) to 104.53(15)◦ and from 157.15(12)
to 163.95(12)◦, respectively (Table 2).
The crystal packing is governed by the simultaneous need to fill
the available space and to minimise the interionic distances (Fig. 3).
This occurs with the concomitant formation of supramolecular con-
tacts, in particular C H· · ·␲ (only a handful; not shown) and a
large number of weak hydrogen bonds of the C H· · ·N and C H· · ·O
types (see Table 3 for the geometrical details of the most relevant;
not shown). Most of these weak interactions are located within
the [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)]2+ dication itself (i.e., are of
intramolecular nature), connecting the coordinated organic ligand
to the neighbouring oxido groups (Table 3). We note that these
intramolecular contacts may be the structural reason why the com-
plex herein reported exhibits the smallest Mo
O Mo angle and
3.2. Crystal structure description of
[{MoO2(HC(3,5-Me2pz)3)}2(ꢄ2-O)][Mo6O19]·3CH3CN
shortest Mo· · ·Mo intermetallic distance among all known related
compounds, including the analogous cations previously reported
by us [8].
Compound 3 crystallised in the monoclinic space group P21/n,
with the asymmetric unit being composed of a bimetallic dication,
a hexametallic dianion and three crystallisation solvent molecules
3.3. Catalytic epoxidation of cis-cyclooctene
(3·3CH3CN, Fig. 2). The dianion, [Mo6O19
]
2−, is the typical Lindqvist
polyoxometallate commonly found in a wide variety of compounds
[19]. As a consequence, its crystallographic features will not be
Compounds 1–4 were explored as (pre)catalysts for the liquid-
phase epoxidation of olefins by using cis-cyclooctene (Cy) as a

A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
69
Fig. 2. Ionic entities composing the asymmetric unit of compound 3·3CH3CN. Atoms are represented as thermal ellipsoids drawn at the 50% probability level. Hydrogen
atoms have been omitted for clarity. For selected bond lengths and angle ranges see Table 2.
model substrate, tert-butylhydroperoxide (TBHP) as oxygen donor,
and DCE as co-solvent, at 55 ◦C (Table 4). To the best of our knowl-
edge, there are no prior reports on catalytic olefin epoxidation using
these compounds. For all tested compounds, the reaction mixtures
were biphasic solid–liquid, and 1,2-epoxy-cyclooctane (CyO) was
always the only reaction product. Blank catalytic tests using equiv-
alent amounts of the (soluble) neutral tpm ligands gave less than
2% conversion under similar reaction conditions, which indicates
that the active species contain molybdenum.
The reaction of Cy with TBHP in the presence of 1 gave
96% CyO yield at 6 h reaction and 55 ◦C (Table 4 and Fig. 4).
The ATR FT-IR spectrum of the beige coloured solid recovered
at the end of the reaction (1*) matches that for the (␮2-
oxo)bis[dioxomolybdenum(VI)] hexamolybdate 3 (Fig. 5). The
conversion of 1 to the high-oxidation-state molybdenum com-
pound 3 involves oxidation, release of CO and, to a certain extent,
the dissociation of the neutral tridentate ligand from the metal
centre. The reaction of Cy in the presence of 1* (using the same
initial molar ratios of Mo:Cy:TBHP) led to similar catalytic results
to those obtained with 1 (Table 4 and Fig. 4). No induction period
was observed for 1, which together with the roughly coinci-
dent kinetic curves for 1 and 1*, suggests that the conversion
of 1 to the active oxidising species is not a rate-limiting step
(Fig. 4), in contrast to that reported previously for the com-
plex [Mo(CO)3(1,1,1-tris(methylaminomethyl)ethane)] (<1%/27%
conversion at ∼80 min/∼8 h reaction) and the resultant oxidised
compound (2%/31% conversion at 35 min/∼8 h reaction), used
as pre(catalysts) in the same reaction under similar conditions
Table 4
Catalytic epoxidation of Cy with TBHP (ca. 5.5 M in decane) in the presence of molybdenum complexes possessing the ligands HC(3,5-Me2pz)3 or HC(pz)3, at 55 ◦C.
Molybdenum complex
Reaction conditions (Mo:Cy:TBHP, co-solvent)a
CyO yield (%)b, reaction time
Ref.
[Mo(CO)3(HC(3,5-Me2pz)3)] (1)
[Mo(CO)3(HC(pz)3)] (2)
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19] (3)
4
1:100:152, DCE
1:100:152, DCE
1:100:152, DCE
1:100:152, DCE
1:100:200, none
96/100(92/100), 6 h/24 h
99(100), 2 h
46/87(45/86), 6 h/24 h
This work
This work
This work
This work
[7]
94/99(95/99), 2 h/4 h
∼
X
X
=
Cl: 56/86, 4 h/24 h
[MoO2X(HC(pz)3)]X
=
∼
Br: 60/100, 4 h/24 h
=
∼
[MoO2Cl(HC(3,5-Me2pz)3)]Cl
[MoO2Cl(HC(3,5-Me2pz)3)]BF4
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)](BF4)2
[Mo2O3(O2)2(␮2-O)(H2O)(HC(3,5-Me2pz)3)]
[MoO3(HC(3,5-Me2pz)3)]
1:100:200, none
1:100:150, none
1:100:150, none
1:100:150, none
1:100:150, none
50/75, 4 h/24 h
[7]
[8]
[8]
[8]
[8]
87/97(86/98), 6 h/24 h
88/100(88/99), 6 h/24 h
88/99, 6 h/24 h
46/76, 6 h/24 h
a
Initial molar ratios of Mo:Cy:TBHP, and co-solvent used (none = no co-solvent was used, DCE = 1,2-dichloroethane).
Values in brackets are for a second 24 h-batch run using the solids recovered from the first run.
b

70
A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
Fig. 3. Crystal packing of 3·3CH3CN viewed in perspective along the [1 0 0] direc-
tion of the unit cell. Acetonitrile molecules are represented in space-filling mode.
For clarity, supramolecular contacts between adjacent molecular units have been
omitted (see Table 3 for tabulated geometrical data).
(without a co-solvent) [22]. The ATR FT-IR spectrum of the solid
recovered from this second run (1**) is similar to that for 3 (and
1*), indicating that this compound is fairly stable under the reac-
tion conditions used (Fig. 5). For 1, the liquid phase of the reaction
mixture for run 1 was pale yellow in colour whereas for run 2 it
was colourless; the pale yellow colour may be attributed to the
dissociated ligand HC(3,5-Me2pz)3, which is originally pale yellow.
For comparison, the reaction of Cy was carried out in the pres-
ence of 3 for two consecutive 24 h-batch runs (Table 4 and Fig. 4).
The kinetic curves for the two runs are essentially coincident, which
correlates with the similar ATR FT-IR spectra of 3 and the solids
recovered after runs 1 (3*) and 2 (3**) (Fig. 5). These results parallel
those discussed above for 1 in that the recovered solids are always
identified as 3. Table 4 compares catalytic data for 3 with literature
Fig. 5. ATR FT-IR spectra of HC(3,5-Me2pz)3, [Mo(CO)3(HC(3,5-Me2pz)3)] (1),
[{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19] (3), and solids recovered after one (N*)
or two (N**) batch catalytic runs of 24 h at 55 ◦C (N = 1, 3).
data for related mononuclear or dinuclear complexes containing
the same organic ligand. The catalytic results for 3 are either on a par
with or somewhat inferior to those found for the related complexes.
Based on the similar ATR FT-IR spectra of 1* and 3 (or 3*) (Fig. 5),
the kinetic differences between these compounds (Table 4 and
Fig. 4) do not seem to be due to differences in the types of active
species. A CatFilt assay (details in Section 2) was carried out for 3
to assess the homo/heterogeneous nature of the catalytic reaction,
and involved filtering the catalyst at 2 h reaction and monitoring
the evolution of the reaction for a further 22 h; this assay gave
∼
ꢆCatFilt/ꢆCat 1, indicating that the catalytic reaction was homo-
=
geneous in nature. Compound 3 and the corresponding recovered
solids were cream in colour and the liquid phases were apparently
colourless. ICP-OES of the liquid phase (separated by membrane
filtration, at 55 ◦C) of the reaction of Cy in the presence of 3 dur-
ing 24 h, at 55 ◦C, indicated that the amount of Mo was less than
0.3% of the initial amount of Mo added to the reactor (correspond-
ing to <0.03 mM molybdenum). Assuming a concentration of active
species of 0.03 mM, one may postulate that the turnover frequency
is at least 2700 mol molMo−1 h−1 (based on conversion at ca. 5 h
reaction). In contrast to that observed for 1*, for 3 (and 3*) an induc-
tion period was observed, which was longer when using one fourth
of the typical initial amount of 3 loaded to the reactor (1%/21% at
2 h/6 h reaction, 55 ◦C) (Fig. 4). Since the reaction mixtures were
always biphasic solid–liquid and the catalytic reaction is homo-
geneous in nature, the dependence of the reaction rate of Cy on
the initial amount of 3 is consistent with the overall rate of the
homogeneous catalytic process being governed by the dissolution
of the metal species. If the saturation concentration is reached rel-
atively fast and the dissolution process is not rate-limiting, then
Fig. 4. Kinetic curves for the reaction of Cy with TBHP in the presence of 1 (+ run
1, ꢁ run 2) or 3 (× run 1, ꢂ run 2) at 55 ◦C under typical reaction conditions (initial
molar ratios of Mo:Cy:TBHP of 1:100:152). The curves obtained at 55 ◦C with either
the 3-contact experiment (—) or using a 4-fold decrease in the initial amount of 3
(ꢀ) are also shown.

A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
71
increasing the initial amount of 3 beyond the saturation concen-
tration (the liquid phase is saturated) should not affect the rate of
the homogeneous catalytic reaction, which was not the case for 3.
In contrast, if the initial dissolution of the metal species is relatively
slow (as postulated for 3) and saturation concentration is reached
only after a specific time t, then the liquid phase of the reaction
mixture is not saturated until that instant t, even if the reaction
mixtures are solid–liquid (with solid in excess); in this case, the
reaction of Cy may not be under kinetic regime with the overall
reaction rate depending on the initial amount of metal compound
added, as effectively observed for 3.
A separate experiment (denoted 3-contact) was performed for
pre-dissolving the active species prior to the addition of the olefin:
3 was mixed with the solvent and oxidant (without olefin; the ini-
tial molar ratio Mo:TBHP was the same as that used for the typical
reaction conditions) during 24 h at 55 ◦C (1000 rpm), followed by
separation of the solid by membrane filtration, and finally adding
the olefin to the solution (the initial molar ratio Cy:TBHP was the
same as that used for the typical reaction conditions) and monitor-
ing the catalytic reaction for a further 24 h at 55 ◦C. The catalytic
reaction was much faster than that observed under typical con-
ditions (Fig. 4), which supports that (i) the catalytic reaction is
homogeneous in nature, (ii) the active species are fairly stable, and
(iii) the dissolution process of the metal species is rate-limiting.
Properties of the solids such as particle/crystallite sizes and
crystallinity can affect their dissolution processes. Hence, the mor-
phologies of the solids 1* and 3 were investigated. The SEM images
of 1* and 3 show comparable morphologies comprising irregular
plate-like agglomerates and heterogeneous particle size distribu-
tions, making it difficult to assess differences in texture properties
between 1* and 3 (Fig. 1). The powder XRD patterns of 3 and 3* are
very similar indicating that the crystalline structure remains essen-
tially the same during the catalytic reaction (Fig. 6), consistent with
the above discussion regarding the stability of 3 under the catalytic
reaction conditions (similar catalytic results in recycling runs and
similar ATR FT-IR spectra of the fresh and recovered solids). The
powder XRD patterns of 3 and 3* are similar to that of 1* with
the exception of an additional peak at ca. 12.8◦ 2ꢂ for 1* (Fig. 6);
these results were reproducible in separate experiments for cat-
alyst recovery, and may be due to (i) differences in crystalline
structures of the obtained powdered solids 1* and 3 consisting
of (␮2-oxo)bis[dioxomolybdenum(VI)] hexamolybdate, and/or (ii)
the presence of a crystalline impurity in 1* and not in 3. Attempts
to isolate metal species from the liquid phases were unsuccessful,
partly due to their very low concentration. The kinetic differences
between 1* and 3 can be explained by hypothesis (i) considering
differences in the dissolution rates of the solids (discussed above),
or hypothesis (ii). Since the ATR FT-IR spectra of 1* and 3 are similar
(discussed above), hypothesis (i) seems more plausible. It is worth
summarising that the rate-limiting dissolution of the metal species
in the case of 3 is supported by the dependence of the initial rate
of the reaction of Cy (takes place in homogeneous phase, based
on the CatFilt test) on the initial amount of (stable) 3; the liquid
phase of the reaction mixture is not saturated until the (thermody-
namic) solubility has been reached, even if the reaction mixtures
are solid–liquid (with solid in excess).
Fig. 6. Powder XRD patterns of [{MoO2(HC(3,5-Me2pz)3)}2(␮2-O)][Mo6O19] (3) and
4, and solids recovered after one batch catalytic run of 24 h at 55 ◦C starting with
1–4 (1*–4*).
within 4 h reaction. The ATR FT-IR spectra of the recovered solids
(4* and 4** for runs 1 and 2, respectively) were very similar to
that for 4, and essentially identical to that for 2* and 2** (Fig. 8).
The catalytic results were reproducible using different batches of
The reaction of Cy with TBHP in the presence of 2 gave 99%
CyO yield at 2 h reaction (Table 4 and Fig. 7). Comparable catalytic
results were obtained for the resultant recovered solid (2*) used
in a second batch run (using the same initial mass ratios of metal
compound:Cy:TBHP). The ATR FT-IR spectra of the recovered
solids (2* and 2** for runs 1 and 2, respectively) were similar, and
comparable to that of 4 (Fig. 8). For comparison, the reaction of Cy
was carried out in the presence of 4 and the resultant recovered
solid (4*) (Table 4 and Fig. 7). The kinetic curves were essentially
coincident for the two runs, and >99% CyO yield was reached
Fig. 7. Kinetic curves for the reaction of Cy with TBHP in the presence of 2 (+ run
1, ꢁ run 2) or 4 (× run 1, ꢆ run 2) at 55 ◦C, or in the presence of 4 at 35 ◦C (—),
under typical reaction conditions (initial molar ratios of Mo:Cy:TBHP of 1:100:152).
The curve obtained at 55 ◦C using a 7-fold decrease in the initial amount of 4 is also
shown (ꢀ).

72
A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
Table 5
Olefin epoxidation with TBHP in the presence of 2 at 55 ◦C.
Olefin
Co-solvent
Conversiona (%)
Epoxide
selectivityb (%)
None
DCE
Acetonitrile
Toluene
DCE
DCE
DCE
DCE
DCE
86/100
100/–
87/100
80/100
21/50
59/94
85/99
68/88
74/90
100/100
100/–
Cis-cyclooctene
100/100
100/100
100/100
100/100
100/100
100/100
100/95c
1-octene
Trans-2-octene
Cyclododecene
3-Carene
4-Vinyl-1-cyclohexene
a
Olefin conversion at 6 h/24 h reaction.
Selectivity to the corresponding monoepoxide product (for 4-vinyl-1-
b
cyclohexene as substrate the monoepoxide was 4-vinylcyclohexene-1,2-epoxide).
c
4-Vinylcyclohexene diepoxide was a by-product detected at 24 h.
solubility of the metal compound had not been reached at the time
of the filtration step (i.e., the dissolution of metal compound is a rel-
atively slow process). A 7-fold decrease in the initial amount of 4
resulted in a lower reaction rate at 55 ◦C (17% conversion at 15 min
compared with 42% for 4 under typical reaction conditions, Fig. 7).
These results, together with the differences in reaction rate for 2*
and 4/4* (conversion at 15 min was 69% and 42%, respectively), par-
allel those observed for 1* and 3, and may be due to differences
in the rates of dissolution of the metal compounds, affecting the
overall catalytic process. The crystalline structures of 4 and 4* are
similar, whereas that for 2* is different (Fig. 6). Nevertheless, the
corresponding ATR FT-IR spectra are similar, suggesting that these
compounds are chemically analogous. Major differences in particle
morphology exist between 2* and 4 (Fig. 1), which may account for
differences in the dissolution rates of these compounds.
The performance of 2 was further investigated in the epoxida-
tion of different olefins (1-octene, trans-2-octene, cyclododecene,
3-carene and 4-vinyl-1-cyclohexene) with TBHP, using DCE as co-
solvent at 55 ◦C (Table 5). For all substrates, only epoxide products
were formed, indicating excellent selectivity of the catalyst. A com-
parative study of the reactions of the linear C8 olefins, namely
1-octene and trans-2-octene, indicates higher reactivity of the latter
in which the C C double bond is internal and possesses higher elec-
tron density. Somewhat consistent with these results are those for
the diene substrate 4-vinyl-1-cyclohexene in that regioselectivity is
high towards epoxidation of the endocyclic C C bond (with higher
electron density) relative to the exocyclic one (4-vinylcyclohexene
diepoxide was formed in 5% yield at 90% conversion, reached at
24 h reaction, Table 5). These electronic effects are in agreement
with the epoxidation step involving an oxygen atom transfer of
an electrophilic oxidative species to the olefin, as proposed in the
literature for oxomolybdenum(VI) complexes used as epoxidation
catalysts with hydroperoxide oxidants [9h,9l,23]. A comparison of
the results for the cyclic olefins Cy and cyclododecene suggests that
steric effects are important since the latter (bulkier) olefin was less
reactive than the former. Steric effects may also contribute to the
comparable reactivities of the substrates 4-vinyl-1-cyclohexene
Fig. 8. ATR FT-IR spectra of HC(pz)3, [Mo(CO)3(HC(pz)3)] (2), 4, and solids recovered
after one (N*, NLiq) or two (N**) batch catalytic runs of 24 h at 55 ◦C (N = 2, 4).
pre-synthesised 4. Compared with the literature data for related
systems (Table 4), the catalytic results for 2 and 4 are superior.
During the 24 h reaction with 4, the liquid phase developed a
pale yellow colour, possibly due to dissolved metal species (the
recovered solids were cream coloured). ICP-OES of the liquid phase
(separated by membrane filtration at the reaction temperature) of
the reaction of Cy in the presence of 4 during 6 h at 55 ◦C indicated
that the amount of dissolved Mo was 11% of the initial amount of Mo
loaded to the reactor, which corresponds to ca. 1 mM of dissolved
molybdenum. The CatFilt assay for the reaction of Cy in the presence
of 4 was carried out at 35 ◦C (since at 55 ◦C the reaction was very
fast), with catalyst filtration at 2 h and monitoring of the evolution
of the reaction for a further 22 h, which gave ꢆCatFilt/ꢆCat of 0.6,
suggesting that the filtrate contained active metal species. Without
the filtration step, 69%/93% CyO yield was reached at 6 h/24 h reac-
tion at 35 ◦C, indicating that the catalyst is quite active even under
such mild reaction conditions (Fig. 7).
and 3-carene (while the latter possesses a more substituted C
C
In an attempt to isolate the dissolved metal species after 24 h
reaction at 55 ◦C using 4, the reaction mixture was membrane fil-
tered, and a solid (4Liq) precipitated upon addition of pentane to
the filtrate. The ATR FT-IR spectrum of 4Liq was similar to that of 4
and all the other recovered solids starting from 2 or 4 (Fig. 8). Hence,
the dissolved and undissolved metal species are chemically anal-
ogous, and are most likely responsible for the catalytic reaction.
Metal species (denoted 2Liq) with an ATR FT-IR spectrum simi-
lar to that for 4Liq could also be isolated from the liquid phase of
the reaction of Cy with TBHP in the presence of 2. The fact that
ꢆCatFilt/ꢆCat was less than unity may be because the equilibrium
endocyclic double bond (higher electron density), the correspond-
ing epoxide retains the imposing gem-dimethyl group in the fused
cyclopropyl ring (bulkier olefin)).
The catalytic system 2/Cy/TBHP/55 ◦C was further explored
without using a cosolvent or using toluene or acetonitrile as cosol-
vent, and the results are compared with those for DCE (Table 5).
When the cosolvent is toluene or MeCN the reaction of Cy is slower
(based on conversion at 6 h reaction) than that for DCE. These
results, taken together with the fact that the reaction was also
slower without a cosolvent, suggest that DCE is a favourable sol-
vent for this catalytic system. The slower reaction in the case of

A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
73
toluene is possibly due to the lower solubility of the metal species
in this solvent; similar results have been reported in the literature
for molybdenum carbonyl complexes, tested as pre-catalysts in the
epoxidation of olefins with TBHP, in that apolar cosolvents such as
toluene and n-hexane lead to poor catalytic performances [9j,9o].
The metal species are apparently more soluble when MeCN is the
cosolvent, which does not correlate with the catalytic results. The
enhanced catalyst solubility in MeCN is possibly counterbalanced
by the Lewis basicity of these solvent molecules, levelling off the
Lewis acidity of the active species.
[2] S. Trofimenko, J. Am. Chem. Soc. 92 (1970) 5118–5126.
[3] (a) S. Julia, J.M. del Mazo, L. Avila, J. Elguero, Org. Prep. Proced. Int. 16 (1984)
299–307;
(b) D.L. Reger, T.C. Grattan, K.J. Brown, C.A. Little, J.J.S. Lamba, A.L. Rheingold,
R.D. Sommer, J. Organomet. Chem. 607 (2000) 120–128;
(c) M.S. Goodman, M.A. Bateman, Tetrahedron Lett. 42 (2001) 5–7.
[4] (a) I.K. Dhawan, M.A. Bruck, B. Schilling, C. Grittini, J.H. Enemark, Inorg. Chem.
34 (1995) 3801–3808;
(b) S. Dilsky, J. Organomet. Chem. 692 (2007) 2887–2896;
(c) C. Dinoi, M.F.C.G. da Silva, E.C.B.A. Alegria, P. Smolen´ ski, L.M.D.R.S. Martins,
R. Poli, A.J.L. Pombeiro, Eur. J. Inorg. Chem. (2010) 2415–2424;
(d) A. Sánchez-Méndez, A.M. Ortiz, E. de Jesús, J.C. Flores, P. Gómez-Sal, Dalton
Trans. 36 (2007) 5658–5669.
[5] (a) D.L. Reger, R.F. Semeniuc, M.D. Smith, J. Organomet. Chem. 666 (2003)
87–101;
4. Conclusions
(b) A.J. Hallet, R.A. Baber, A.G. Orpen, B.D. Ward, Dalton Trans. 40 (2011)
9276–9283;
(c) J. Niesel, A. Pinto, H.W.P. N’Dongo, K. Merz, I. Ott, R. Gust, U. Schatzschneider,
Chem. Commun. (2008) 1798–1800;
(d) H. Pfeiffer, A. Rojas, J. Niesel, U. Schatzschneider, Dalton Trans. 38 (2009)
4292–4298;
(e) G. Dördelmann, H. Pfeiffer, A. Birkner, U. Schatzschneider, Inorg. Chem. 50
(2011) 4362–4367;
The tris(pyrazolyl)methane molybdenum(0) tricarbonyl com-
plexes 1 and 2 have never been examined as (pre)catalysts for
any type of organic transformation. In fact, this is the first report
examining tricarbonyl complexes of the type [Mo(CO)3(L)] as
catalyst precursors for olefin epoxidation. An important advan-
tage of these precursors is that they are readily prepared in
one step from Mo(CO)6. As mentioned in the introduction,
previous work by Enemark and co-workers showed that oxi-
dation of [Mo(CO)3(HC(3,5-Me2pz)3)] with nitric acid yielded
(HC(3,5-Me2pz)3)MoO3 [4a]. We anticipated that the same
type of trioxo complex would form upon complete oxidative
(f) H.-M. Berends, P. Kurz, Inorg. Chim. Acta 380 (2012) 141–147.
[6] (a) R. Garcia, A. Paulo, I. Santos, Inorg. Chim. Acta 362 (2009) 4315–4327;
(b) L.R. Goethals, I. Santos, V. Caveliers, A. Paulo, F.D. Geeter, P.G. Lurdes, C.
Fernandes, T. Lahoutte, Contrast Media Mol. Imaging 6 (2011) 178–188.
[7] A.M. Santos, F.E. Kühn, K. Bruus-Jensen, I. Lucas, C.C. Romão, E. Herdtweck, J.
Chem. Soc. Dalton Trans. (2001) 1332–1337.
[8] P. Neves, S. Gago, S.S. Balula, A.D. Lopes, A.A. Valente, L. Cunha-Silva, F.A.A.
Paz, M. Pillinger, J. Rocha, C.M. Silva, I.S. Gonc¸ alves, Inorg. Chem. 50 (2011)
3490–3500.
decarbonylation of
1 and 2 with TBHP. The isolation of a
[9] (a) M. Abrantes, A.M. Santos, J. Mink, F.E. Kühn, C.C. Romão, Organometallics
22 (2003) 2112–2118;
(␮2-oxo)bis[dioxomolybdenum(VI)] hexamolybdate from 1 (and
possibly also from 2) instead of the trioxo complex was, therefore,
unexpected.
(b) C. Freund, M. Abrantes, F.E. Kühn, J. Organomet. Chem. 691 (2006)
3718–3729, and references cited therein;
(c) F.E. Kühn, A.M. Santos, M. Abrantes, Chem. Rev. 106 (2006) 2455–2475, and
references cited therein;
Compounds 3 and 4 promote the homogeneous phase epox-
idation of cis-cyclooctene with TBHP under quite mild reaction
conditions. The compounds prepared in situ are advantageous in
that they tend to lead to a faster catalytic reaction than the corre-
sponding pre-prepared compounds. Nevertheless, 3 and 4 are fairly
stable catalysts based on catalyst recycling and characterization
of the recovered catalysts. The catalytic reactions may be limited
by the rate of dissolution of the metal compounds and there-
fore differences in texture properties may influence the overall
catalytic process. For example, in the case of 2, complete oxida-
tive decarbonylation, achieved either in situ during the catalytic
reaction or prior to the catalytic reaction, leads to chemically anal-
ogous compounds but with different crystalline structures, particle
morphologies and sizes, which correlates with the kinetic dif-
ferences between the two compounds, rationalized on the basis
of differences in the rates of dissolution. Care should be taken
when comparing the catalytic activities of different compounds,
by checking if the overall reaction of the olefin is determined by
the activity of the catalysts themselves or by other rate-limiting
regimes.
(d) J. Zhao, A.M. Santos, E. Herdtweck, F.E. Kühn, J. Mol. Catal. A: Chem. 222
(2004) 265–271;
(e) A.A. Valente, J.D. Seixas, I.S. Gonc¸ alves, M. Abrantes, M. Pillinger, C.C. Romão,
Catal. Lett. 101 (2005) 127–130;
(f) A.M. Martins, C.C. Romão, M. Abrantes, M.C. Azevedo, J. Cui, A.R. Dias,
M.T. Duarte, M.A. Lemos, T. Lourenc¸ o, R. Poli, Organometallics 24 (2005)
2582–2589;
(g) A. Capapé, A. Raith, F.E. Kühn, Adv. Synth. Catal. 351 (2009) 66–70;
(h) A.M. Al-Ajlouni, D. Veljanovski, A. Capapé, J. Zhao, E. Herdtweck, M.J. Cal-
horda, F.E. Kühn, Organometallics 28 (2009) 639–645;
(i) M. Abrantes, F.A.A. Paz, A.A. Valente, C.C.L. Pereira, S. Gago, A.E. Rodrigues,
J. Klinowski, M. Pillinger, I.S. Gonc¸ alves, J. Organomet. Chem. 694 (2009)
1826–1833;
(j) M. Abrantes, P. Neves, M.M. Antunes, S. Gago, F.A.A. Paz, A.E. Rodrigues, M.
Pillinger, I.S. Gonc¸ alves, C.M. Silva, A.A. Valente, J. Mol. Catal. A: Chem. 320
(2010) 19–26;
(k) A. Capapé, A. Raith, E. Herdtweck, M. Cokoja, F.E. Kühn, Adv. Synth. Catal.
351 (2010) 547–556;
(l) P.J. Costa, M.J. Calhorda, F.E. Kühn, Organometallics 29 (2010) 303–311;
(m) V.V.K.M. Kandepi, A.P. da Costa, E. Peris, B. Royo, Organometallics 28 (2009)
4544–4549;
(n) P. Neves, C.C.L. Pereira, F.A.A. Paz, S. Gago, M. Pillinger, C.M. Silva,
A.A. Valente, C.C. Romão, I.S. Gonc¸ alves, J. Organomet. Chem. 695 (2010)
2311–2319;
(o) J.C. Alonso, P. Neves, M.J.P. da Silva, S. Quintal, P.D. Vaz, C. Silva, A.A. Valente,
P. Ferreira, M.J. Calhorda, V. Félix, M.G.B. Drew, Organometallics 26 (2007)
5548–5556;
Acknowledgements
(p) P. Neves, T.R. Amarante, A.C. Gomes, A.C. Coelho, S. Gago, M. Pillinger, I.S.
Gonc¸ alves, C.M. Silva, A.A. Valente, Appl. Catal. A: Gen. 395 (2011) 71–77;
(q) T.R. Amarante, P. Neves, A.C. Coelho, S. Gago, A.A. Valente, F.A.A. Paz, M.
Pillinger, I.S. Gonc¸ alves, Organometallics 29 (2010) 883–892.
[10] T. Kottke, D. Stalke, J. Appl. Crystallogr. 26 (1993) 615–619.
[11] APEX2, Data Collection Software (Version 2.1-RC13), Bruker AXS, Delft, The
Netherlands, 2006.
[12] Cryopad, Remote Monitoring and Control (Version 1.451), Oxford Cryosystems,
Oxford, UK, 2006.
[13] SAINT+, Data Integration Engine (version 7.23a), Bruker AXS, Madison, WI, USA,
2005.
We are grateful to the Fundac¸ ão para a Ciência e a Tecnologia
(FCT), QREN, FEDER, COMPETE, the European Union and the Asso-
ciate Laboratory CICECO (Pest-C/CTM/LA0011/2011) for continued
support and funding. The FCT is acknowledged for financial support
towards the purchase of the single-crystal diffractometer, for a PhD
grant to SF (SFRH/BD/45116/2008), and for post-doctoral grants to
PN (SFRH/BPD/73540/2010) and JAF (SFRH/BPD/63736/2009). The
authors thank Tatiana R. Amarante for help in obtaining the SEM
images.
[14] G.M. Sheldrick, SADABS Version 2.01, Bruker/Siemens Area Detector Absorp-
tion Correction Program, Bruker AXS, Madison, WI, USA, 1998.
[15] (a) G.M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr. 64 (2008)
112–122;
(b) G.M. Sheldrick, SHELXS-97, Program for Crystal Structure Solution, Univer-
sity of Göttingen, Göttingen, Germany, 1997.
[16] G.M. Sheldrick, SHELXL-97, Program for Crystal Structure Refinement, Univer-
sity of Göttingen, Göttingen, Germany, 1997.
[17] C. Rocchiccioli-Deltcheff, R. Thouvenot, M. Fouassier, Inorg. Chem. 21 (1982)
30–35.
References
[1] (a) H.R. Bigmore, S.C. Lawrence, P. Mountford, C.S. Tredget, Dalton Trans. 34
(2005) 635–651;
(b) C. Pettinari, R. Pettinari, Coord. Chem. Rev. 249 (2005) 525–543.

74
A.C. Gomes et al. / Journal of Molecular Catalysis A: Chemical 370 (2013) 64–74
[18] M.R. Pedrosa, R. Aguado, V. Díez, J. Escribano, R. Sanz, F.J. Arnáiz, Eur. J. Inorg.
Chem. (2007) 3952–3954.
[19] (a) V. Vrdoljak, B. Prugovecˇki, D. Matkovic´-Calogovic´, J. Pisk, CrystEngComm 13
(c) K.S. Burger, G. Haselhorst, S. Stotzel, T. Weyhermuller, K. Wieghardt, B.
Nuber, Dalton Trans. (1993) 1987–1997.
[22] Z. Petrovski, A.A. Valente, M. Pillinger, A.S. Dias, S.S. Rodrigues, C.C. Romão, I.S.
ˇ
ˇ
(2011) 4382–4390;
Gonc¸ alves, J. Mol. Catal. A: Chem. 249 (2006) 166–171.
[23] (a) F.E. Kühn, M. Groarke, E. Bencze, E. Herdtweck, A. Prazeres, A.M. Santos, M.J.
Calhorda, C.C. Romão, I.S. Gonc¸ alves, A.D. Lopes, M. Pillinger, Chem. Eng. J. 8
(2002) 2370–2383;
(b) M. Sarma, T. Chatterjee, S.K. Das, Dalton Trans. 40 (2011) 2954–2966;
(c) T.R. Amarante, J.A. Fernandes, I.S. Gonc¸ alves, F.A.A. Paz, Acta Crystallogr.
Sect. E: Struct. Rep. Online 67 (2011) m1828–m1829.
[20] K.M. Barnhart, J.H. Enemark, Acta Crystallogr., Sect. C: Cryst. Struct. Commun.
40 (1984) 1362–1364.
(b) L.F. Veiros, A. Prazeres, P.J. Costa, C.C. Romão, F.E. Kühn, M.J. Calhorda, Dalton
Trans. (2006) 1383–1389;
[21] (a) K. Wieghardt, G. Backes-Dahmann, W. Herrmann, J. Weiss, Angew. Chem.
Int. Ed. 23 (1984) 899–900;
(c) A. Comas-Vives, A. Lledós, R. Poli, Chem. Eng. J. 16 (2010) 2147–2158;
(d) C. Dinoi, M. Ciclosi, E. Manoury, L. Maron, L. Perrin, R. Poli, Chem. Eur. J. 16
(2010) 9572–9584.
(b) D.V. Partyka, R.J. Staples, R.H. Holm, Inorg. Chem. 42 (2003) 7877–7886;