Olefin epoxidation in solventless conditions and apolar media catalysed by specialised oxodiperoxomolybdenum complexes

Article abstract of DOI:10.1016/j.molcata.2011.02.004

The epoxidation of olefin substrates, in both apolar organic media and under solventless conditions, with aqueous hydrogen peroxide and catalysed by molybdenum complexes has been investigated. The catalysts compounds employed were the oxodiperoxomolybdenum complexes of several pyridine, 2,2′-bipyridine and pyrazole ligands with apolar functions (alkyl chains, alkyl-trimethylsilyl groups and polydimethylsiloxanyl polymer), which showed enhanced solubility in relatively apolar organic media. Both the isolated complexes and in situ preparations were catalytically active. The solubility of the new catalyst complexes appears to facilitate the catalytic activity in these systems, since activity was not observed for the analogous, insoluble complexes of unfunctionalised ligands. In these systems, the oxidant, aqueous hydrogen peroxide, forms a separate phase and the catalyst resides in the organic phase. From a green chemistry and economic perspective the elimination of organic solvents and co-catalysts from a reaction system would present advantages and, consequently, the epoxidation reaction was also investigated under solventless conditions. The 3-hexyl-5-methylpyrazole and 3-hexyl-5-heptylpyrazole complexes were found to show heightened activities, the latter being particularly efficient in these conditions, whilst bipyridines apparently inhibit the epoxidation. In addition, the mechanism of the epoxidation reaction was studied through DFT calculations for the model olefin substrate ethylene with the oxodiperoxomolybdenum complex of 3-hexyl-5-heptylpyrazole. The oxo-transfer reaction occurred by interaction of the ethylene with the peroxo ligand via the spirocyclic transition state proposed by Sharpless.

Full text of DOI:10.1016/j.molcata.2011.02.004

Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Olefin epoxidation in solventless conditions and apolar media catalysed by
specialised oxodiperoxomolybdenum complexes
Matthew Herbert, Francisco Montilla∗, Agustín Galindo∗
Departamento de Química Inorgánica, Facultad de Química, Universidad de Sevilla, Aptdo 1203, 41071 Sevilla, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
The epoxidation of olefin substrates, in both apolar organic media and under solventless conditions,
with aqueous hydrogen peroxide and catalysed by molybdenum complexes has been investigated. The
catalysts compounds employed were the oxodiperoxomolybdenum complexes of several pyridine, 2,2ꢀ-
bipyridine and pyrazole ligands with apolar functions (alkyl chains, alkyl-trimethylsilyl groups and
polydimethylsiloxanyl polymer), which showed enhanced solubility in relatively apolar organic media.
Both the isolated complexes and in situ preparations were catalytically active. The solubility of the new
catalyst complexes appears to facilitate the catalytic activity in these systems, since activity was not
observed for the analogous, insoluble complexes of unfunctionalised ligands. In these systems, the oxi-
dant, aqueous hydrogen peroxide, forms a separate phase and the catalyst resides in the organic phase.
From a green chemistry and economic perspective the elimination of organic solvents and co-catalysts
from a reaction system would present advantages and, consequently, the epoxidation reaction was also
investigated under solventless conditions. The 3-hexyl-5-methylpyrazole and 3-hexyl-5-heptylpyrazole
complexes were found to show heightened activities, the latter being particularly efficient in these
conditions, whilst bipyridines apparently inhibit the epoxidation. In addition, the mechanism of the
epoxidation reaction was studied through DFT calculations for the model olefin substrate ethylene with
the oxodiperoxomolybdenum complex of 3-hexyl-5-heptylpyrazole. The oxo-transfer reaction occurred
by interaction of the ethylene with the peroxo ligand via the spirocyclic transition state proposed by
Sharpless.
Received 11 November 2010
Received in revised form 1 February 2011
Accepted 6 February 2011
Available online 15 February 2011
Keywords:
Catalysis
Molybdenum
Epoxidation
Hydrogen peroxide
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
such catalysts and the relative scarcity of rhenium, catalysts based
on more common transition metal elements would be preferable
The selective oxidation of olefins to the corresponding epox-
ides is a highly important industrial transformation in the synthesis
of many fine and bulk chemical intermediates. Generally efficient
oxidative technologies have been developed for the epoxidation
of light olefins [1] but for heavier, more complex substrates, stoi-
chiometric quantities of organic oxidants are typically employed
[2], resulting in the production of substantial quantities of haz-
ardous by products and poorer atom economies [3]. For these
reasons the investigation of catalytic epoxidations for heavier
olefins has been an important area of research for several decades.
Epoxidations facilitated by transition metal catalyst compounds
can employ hydroperoxide oxidants including H2O2 as oxidant,
resulting in more benign waste products (small alcohols or water)
and improved atom economies [4]. In homogeneous epoxidations
methyltrioxorhenium based catalysts have been generally found
to give the highest activities [5]. However, due to the expense of
[6]. Molybdenum(VI) compounds have produced some of the more
active catalysts to be studied in this capacity [7] and amongst these
oxodiperoxomolybdenum type complexes stand out as attrac-
tive options due to their facile and cheap preparation, chemical
simplicity and robustness [8,9]. Homogeneous molybdenum catal-
ysed epoxidations typically employ organohydroperoxide reagents
(most commonly tert-butylhydroperoxide) as oxidant but a limited
number of systems have been reported employing “greener” hydro-
gen peroxide [10], most notably including the highly active NaHCO3
co-catalysed systems developed by Bhattacharyya et al. [11].
Continuing our research into metal catalysed oxidations in apo-
lar media [12] we have now investigated oxodiperoxomolybdenum
catalysed epoxidations in solventless systems and in low-polar
organic solvents. A difficulty which is encountered when trying to
use these compounds as catalysts under such conditions is their
poor solubility, particular in more apolar organic media includ-
ing halocarbon, aromatic and hydrocarbon solvents. The role that
appropriate ligands can play in solubilising the catalyst complex
is crucial to achieving an efficient catalytic system [13]. Previously
reported systems in such media typically employed catalyst com-
∗
Corresponding authors. Fax: +34 954 557153.
E-mail addresses: montilla@us.es (F. Montilla), galindo@us.es (A. Galindo).
1381-1169/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2011.02.004

112
M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
plexes incorporating ligands functionalised with long alkyl chains,
such as the alkyl functionalised pyrazolylpyridine ligands used
by Thiel et al. [9] or trialkylphosphine oxide ligands [14]. These
functional groups help to solubilise the catalyst and thus permit
effective homogeneous catalysis. In our own previous studies we
have investigated the use of ionic liquids to solubilise oxodiper-
oxomolybdenum catalysts with simpler unfunctionalised ligands
to investigate and compare their catalytic efficiency in epoxida-
tion reactions [15]. As a result we determined that pyrazole type
ligands produce a significant enhancement in the catalytic activ-
ity of these catalysts [16]. Developing on these precedents the
study to be presented here investigated oxodiperoxomolybdenum
catalysed olefin epoxidations in relatively apolar media employing
specialised alkyl, trimethylsilyl and polydimethylsiloxane func-
tionalised N-donor base ligands in order to solubilise the catalyst
complexes. As well as pyridine and bipyridine based ligands,
3,5-dialkylpyrazole species were investigated under solvent-free
conditions to see if pyrazole species would again induce enhanced
catalytic activity, as previously observed in ionic liquids [16].
umn chromatography employing silica gel and 0.5% NEt3 in 1:1
hexane–ethyl acetate as eluent. The final product C was a yellow
oil (440 mg, 1.75 mmol, 58%). IR (NaCl, cm−1): 500, 685, 755, 774,
839, 981, 1036, 1251, 1415, 1595, 1558, 1598, 2851, 2898, 2952,
3024, 3067. 1H NMR (CDCl3): ı 0.00 (br s, 18H, Si(CH3)3), 0.32 (t,
JHH = 6.8 Hz, 1H, CH), 2.77 (d, JHH = 6.7 Hz, 2H, CH2), 7.13 (br m, 2H,
1
m-NC5H4), 8.46 (br s, 2H, o-NC5H4). 13C{ H} NMR (CDCl3): ı 0.0
(s, Si(CH3)3), 14.6 (s, CH), 31.4 (s, CH2), 123.7 (s, m-NC5H4), 149.4
1
(s, o-NC5H4), 153.7 (s, p-NC5H4). 29Si{ H} NMR (CDCl3): ı 3.8 (s,
Si(CH3)3). Elemental analysis, calculated for C13H25NSi2: C, 62.08;
H, 10.02; N, 5.57. Experimental: C, 58.45; H, 9.79; N, 4.95%.
2.2.2. 2,6-(Bis-trimethylsilanylmethyl)pyridine (D)
The same experimental procedure was followed for this deriva-
tive but in place of 4-picoline, 2,6-lutidine (0.50 equiv., 463 ␮L,
4.0 mmol) was doubly deprotonated by two equivalents of LDA
before reacting with chlorotrimethylsilane (1.00 equiv., 1.02 mL,
8.0 mmol). Compound D was obtained as a pale yellow oil (520 mg,
2.07 mmol, 52%). IR (NaCl, cm−1): 722, 806, 841, 948, 978, 1030,
1093, 1105, 1153, 1184, 1203, 1311, 1378, 1398, 1444, 1466, 1588,
1600, 1622, 2098, 2427, 2489, 2567, 2720, 2760, 2850, 2923, 2960,
3400. 1H NMR (CDCl3): ı 0.00 (br s, 18H, Si(CH3)3), 2.41 (s, 4H,
CH2), 6.68 (d, JHH = 7.5 Hz, 2H, m-NC5H3), 7.27 (t, JHH = 7.5 Hz, 1H,
2. Experimental
2.1. General
1
p-NC5H3). 13C{ H} NMR (CDCl3): ı −0.1 (s, Si(CH3)3), 32.8 (s, CH2),
1
117.3 (s, m-NC5H4), 135.4 (s, p-NC5H4), 157.0 (s, o-NC5H4). 29Si{ H}
Synthetic reactions were performed under a dry, oxygen-free,
nitrogen atmosphere using standard Schlenk techniques. Solvents
were dried and purified appropriately prior to use, using standard
procedures. Chemicals were obtained from commercial sources
and used as supplied or purified by distillation prior to use, as
appropriate. Infrared spectra were recorded on a Perkin-Elmer
Model 883 spectrophotometer (as either liquid film supported on
KBr discs or in pressed KBr pellets). NMR spectra were recorded
NMR (CDCl3): ı 2.0 (s, Si(CH3)3). Elemental analysis, calculated for
C13H25NSi2: C, 62.08; H, 10.02; N, 5.57. Experimental: C, 60.37; H,
10.16; N, 4.97%.
2.2.3. 4,4ꢀ-Di(tridecyl)-2,2ꢀ-bypiridine (E)
To a solution of iPr2NH (1.05 equiv., 950 ␮L, 6.7 mmol) in dry THF
(∼30 mL) at −78 ◦C was added nBuLi 1.6 M in hexanes (1.05 equiv.,
4.0 mL, 6.7 mmol), resulting in the in situ formation of lithium
diisopropylamide (LDA). A solution of 4,4-dimethyl-2,2-bipyridine
(1.00 equiv., 590 mg, 3.20 mmol) in THF (15 mL) was then added
and the solution warmed to 0 ◦C in an ice bath, stirring at this tem-
perature for 5 min. 1-Iodododecane (1.65 mL, 6.7 mmol) was then
added, stirring at 0 ◦C for a further 5 min. The reaction was then
allowed to warm to room temperature before leaving to stir for
36 h. The solution was concentrated to a volume of approximately
10 mL on a rotavap before adding dichloromethane (50 mL) and
washing with distilled water (3× 25 mL). The aqueous wash frac-
tions were then themselves extracted with dichloromethane (3×
25 mL) before combining all organic fractions, drying the solution
over MgSO4, filtering and evaporating the solvent on a rotavap to
leave the crude product as a cream coloured solid. This was recrys-
tallised from methanol to give the final product as a colloidal white
solid (900 mg, 70%). The 1H NMR data correlate to those previously
reported [17]. IR (NaCl, cm−1): 834, 1110, 1547, 1597, 1964. 1H NMR
(CDCl3): ı 0.80 (t, JHH = 6.4 Hz, 6H, CH3), 1.18 (v br m, 40H, interme-
diate CH2), 1.62 (m, 4H, bipy-CH2CH2CH2), 2.61 (t, JHH = 7.8 Hz, 4H,
bipy-CH2CH2), 7.05 (d, JHH = 4.5 Hz, 2H, 5-CH and 5ꢀ-CH of bipy),
8.16 (s, 2H, 3-CH and 3ꢀ-CH of bipy), 8.48 (d, JHH = 4.5 Hz, 2H, 6-
1
using a Bruker AMX-300 spectrometer with 13C{ H} and 1H shifts
referenced to the residual solvent signals. All data are reported in
ppm downfield from Si(CH3)4. Gas chromatography was carried
out using a Varian CP-3800. Elemental analysis (C, H, N) was con-
ducted by the Microanalytical Service of the Universidad de Sevilla
(CITIUS) on an Elemental LECO CHNS 93 analyser. The synthesis
and characterisation of both A and B have been described previ-
ously by us [12b]. The syntheses of compounds E and H have been
communicated previously [17], but the experimental procedures
employed vary slightly and are here described. Although the use
of D has been described previously [18] its synthesis was not been
reported in detail and thus is here included.
2.2. Synthesis of ligands
2.2.1. 4-(2,2-Bis-trimethylsilanyl-ethyl)pyridine (C)
To a solution of iPr2NH (1.05 equiv., 447 ␮L, 3.15 mmol) in
dry THF (∼25 mL) at −78 ◦C was added nBuLi 1.6 M in hexanes
(1.05 equiv., 1.88 mL, 3.15 mmol), resulting in the in situ forma-
tion of lithium diisopropylamide (LDA). A solution of 4-picoline
(1.00 equiv., 291 ␮L, 3.0 mmol) in THF (15 mL) was then added and
the solution warmed to 0 ◦C in an ice bath, stirring at this temper-
ature for 5 min. Next, 1-chloro-2,2-bis(trimethylsilanyl)methane
(1.00 equiv., 650 ␮L, 3.0 mmol) was added, stirring at 0 ◦C for a fur-
ther 5 min before allowing it to warm to room temperature and
leaving it to stir for 36 h. The solution was then concentrated to
approximately 10 mL, dissolved in dichloromethane (50 mL), and
washed with water (3× 25 mL). To prevent significant loss of prod-
uct in the washes, the aqueous fractions were then themselves
extracted with dichloromethane (3× 25 mL), and all of the organic
solutions were then combined. The resulting solution was dried
(MgSO4), filtered, and evaporated on a rotary evaporator (rotavap),
leaving the crude product as an oil. This was purified by col-
1
CH and 6ꢀ-CH of bipy). 13C{ H} NMR (CDCl3): ı 14.1 (s, CH3), 22.7,
29.4, 29.5, 30.5, 31.9 (s, intermediate CH2), 35.6 (s, bipy-CH2), 121.3,
123.9 (s, C-3 and C-5 of bipy), 149.0, 152.9, 156.2 (s, C-2, C-4 and
C-6 of bipy). Elemental analysis, calculated for C36H60N2: C, 83.01;
H, 11.61; N, 5.38. Experimental: C, 84.23; H, 12.25; N, 5.59%.
2.2.4. 4,4ꢀ-Bis-(bistrimethylsilanylmethyl)-2,2ꢀ-bipyridine (F)
A method analogous to that described for compound E was used
with chlorotrimethyilsilane (1.00 equiv., 770 ␮L, 5.9 mmol) in place
of 1-iodododecane and the molar proportions of all other reagents
adjusted accordingly. Compound F was obtained as fine white crys-
tals (310 mg, 24%). IR (KBr, cm−1): 451, 558, 620, 654, 691, 724, 756,
778, 839, 862, 914, 990, 1035, 1069, 1159, 1217, 1247, 1404, 1540,

M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
113
1580, 3047. 1H NMR (CDCl3): ı 0.00 (s, 36H, Si(CH3)3), 1.72 (s, 4H,
CH), 6.88 (d, JHH = 5 Hz, 2H, 5-CH and 5ꢀ-CH of bipy), 8.01 (s, 2H, 3-CH
and 3ꢀ-CH of bipy), 8.42 (d, JHH = 5 Hz, 2H, 6-CH and 6ꢀ-CH of bipy).
and NH2NH2·H2O (1.75 mL, 35 mmol) added. The reaction was left
stirring overnight at room temperature. The solvent was then evap-
orated to leave the crude product as a brown oil. This was purified
by elution through a silica gel column obtaining the product as
a clear, pale yellow oil (2.16 g, 13.0 mmol, 52%). IR (NaCl, cm−1):
725, 792, 1008, 1027, 1148, 1315, 1378, 1467, 1580, 2857, 2927,
3033, 3103, 3134, 3196. 1H NMR (CDCl3): ı 0.88, (t, JHH = 6.3 Hz,
3H, CH2CH3), 1.31 (br, 6H, intermediate CH2), 1.60 (br quintet, 2H,
pz-5-CH2CH2CH2), 2.28 (s, 3H, pz-5-CH3), 2.60 (t, JHH = 7.7 Hz, 2H,
1
13C{ H} NMR (CDCl3): ı 0.0 (s, Si(CH3)3), 30.6 (s, CH), 121.5, 123.4
(s, C-3 and C-5 of bipy), 148.4, 153.9, 155.7 (s, C-2, C-4 and C-6 of
1
bipy). 29Si{ H} NMR CDCl3): ı 2.2 (s, Si(CH3)3). Elemental analysis,
calculated for C24H44N2Si4: C, 60.95; H, 9.38; N, 5.92. Experimental:
C, 59.90; H, 10.12; N, 6.07%.
1
pz-5-CH2CH2), 5.83 (s, 1H, CH of pz). 13C{ H} NMR (CDCl3): ı 12.2
2.2.5. 4,4ꢀ-Bis-(2,2-bis-trimethylsilanyl-ethyl)-2,2ꢀ-bipyridine (G)
A method analogous to that described for compound E was used
with 1-chloro-2,2-bis-trimethylsilanyl-ethane (1.00 equiv., 1.00 g,
5.13 mmol) in place of 1-iodododecane and the molar propor-
tions of all other reagents adjusted accordingly. The product was
obtained as cream coloured crystals (290 mg, 23%). IR (KBr, cm−1):
839, 1036, 1250, 1551, 1590. 1H NMR (CDCl3): ı 0.00 (s, 18H,
Si(CH3)3), 0.47 (t, JHH = 7.0 Hz, 2H, CH2CH), 2.87 (d, JHH = 7.0 Hz, 4H,
CH2CH), 7.14 (d, JHH = 5 Hz, 2H, 5-CH and 5ꢀ-CH of bipy), 8.27 (s, 2H,
3-CH and 3ꢀ-CH of bipy), 8.54 (d, JHH = 5 Hz, 2H, 6-CH and 6ꢀ-CH of
(s, CH3), 13.9 (s, CH3), 22.4, 26.7, 28.9, 29.3, 31.5 (s, intermediate
CH2), 102.8 (s, CH of pz), 144.4 (s, CCH3 of pz), 148.9 (s, CCH2 of pz).
Elemental analysis, calculated for C10H18N2: C, 72.24; H, 10.91; N,
16.85. Experimental: C, 80.24; H, 11.21; N, 14.10%.
2.2.8. 3-Hexyl-5-heptylpyrazole (J)
The synthetic procedure was identical to that performed for I
but using an equal molar quantity of octanoyl chloride in place of
acetyl chloride. This product was obtained as a yellow oil (2.56 g,
10.3 mmol, 41%). IR (NaCl, cm−1): 724, 797, 887, 1003, 1027, 1093,
1163, 1259, 1378, 1465, 1577, 1671, 1737, 2853, 2927, 2953, 3103,
3138, 3200. 1H NMR (CDCl3): ı 0.80 (br t, JHH = 6.5 Hz, 6H, CH3), 1.22
(br, 14H, CH2), 1.55 (m, 4H, pz-CH2CH2CH2), 2.52 (t, JHH = 7.7 Hz, 4H,
1
bipy). 13C{ H} NMR (CDCl3): ı 0.0 (s, Si(CH3)3), 14.3 (s, CH), 31.6
(s, CH2), 121.1 (s, C-5 of bipy), 123.6 (s, C-3 of bipy), 148.7, 154.6,
1
155.8 (s, C-2, C-4 and C-6 of bipy). 29Si{ H} NMR (CDCl3): ı 3.9 (s,
Si(CH3)3). Elemental analysis, calculated for C26H48N2Si4: C, 62.33;
1
pz-CH2CH2), 5.76 (s, 1H, CH). 13C{ H} NMR (CDCl3): ı 13.0 (s, two
H, 9.66; N, 5.59. Experimental: C, 62.28; H, 10.46; N, 5.84%.
alkyl CH3), 21–31 (several s, intermediate CH2), 100.9 (s, CH of pz),
148.2 (s, CCH2 of pz). Elemental analysis, calculated for C16H30N2:
C, 76.74; H, 12.07; N, 11.19. Experimental: C, 80.10; H, 12.76; N,
9.45%.
2.2.6. 4-Methyl-4ꢀ-tridecyl-2,2ꢀ-bipyridine (H)
To a dry solution of diisopropylamine (461 ␮L, 3.26 mmol)
in THF (30 mL) was added nBuLi 1.6 M in hexanes (2.04 mL,
3.26 mmol). This was stirred for 15 min after which a solution of 4,4-
dimethyl-2,2ꢀ-bipyridine (601 mg, 3.26 mmol) in dry THF (25 mL)
was added dropwise over 30 min. This solution was then stirred for
90 min before it was cooled to 0 ◦C and 1-iodododecane (804 ␮L,
3.26 mmol) was added. The solution was stirred for another 90 min
at 0 ◦C and then quenched with water/ice (25 mL) and the mixture
was subsequently extracted with diethyl ether (3× 50 mL) which
was then dried (MgSO4), filtered and evaporated to leave the crude
product. This was recrystallised from ethyl acetate to leave the
product as a white solid (380 mg, 33%). The 1H NMR data correlate to
those previously reported [17]. IR (KBr, cm−1): 534, 583, 729, 824,
898, 992, 1043, 1107, 1245, 1374, 1462, 1551, 1596, 2849, 2916,
3054. 1H NMR (CDCl3): ı 0.81 (t, JHH = 6.6 Hz, 3H, CH2CH3), 1.18 (br
m, 20H, intermediate CH2), 1.62 (quintet, 2H, bipy-4ꢀ-CH2CH2CH2),
2.37 (s, 3H, bipy-4-CH3), 2.62 (t, JHH = 7.8 Hz, 2H, bipy-4ꢀ-CH2CH2),
7.07 (m, 2H, 5-CH and 5ꢀ-CH of bipy), 8.17 (br s, 2H, 3-CH and 3ꢀ-
2.3. Synthesis of metal complexes
2.3.1. [Mo(O)(O2)2(H2O)n] solution in aqueous hydrogen
peroxide
Solutions of the aqua complex of oxodiperoxomolybdenum,
[Mo(O)(O2)2(H2O)n] in aqueous hydrogen peroxide used both in
synthesis and catalytic studies were prepared as follows. A suspen-
sion of MoO3 (2.52 g, 17.5 mmol) in 12 mL 30% aqueous hydrogen
peroxide was heated at 55 ◦C with continuous stirring for approx-
imately 1 h after which complete dissolution of the molybdenum
resulting in a clear yellow solution was observed. At this point the
solution was cooled to 0 ◦C and several drops of hydrogen peroxide
were added and the solution was then made up to 25 mL and stored
in a sealed volumetric flask at 4 ◦C. Occasional venting of this solu-
tion is advised upon prolonged storage due to the accumulation of
pressure following catalytic decomposition of hydrogen peroxide.
A solution of [Mo] with concentration 834 mM is thus obtained.
The solution should actually consist of several molybdenum oxide
species in equilibria which are dependent on factors including the
concentration, pH and temperature of the solution [19]. For the pur-
pose of simplicity the solution is referred to in this work simply as
aqueous [Mo(O)(O2)2(H2O)n].
CH of bipy), 8.48 (m, 2H, 6-CH and 6ꢀ-CH of bipy). 13C{ H} NMR
1
(CDCl3): ı 14.0 (s, CH2CH3), 21.1 (s, bipy-4-CH3), 22.5 (s, CH2CH3),
29.1–29.5 (several s, intermediate CH2), 30.3, 31.8 (s, intermediate
CH2), 35.4 (s, bipy-4ꢀ-CH2), 121.3 (s, C-5ꢀ of bipy), 122.0 (s, C-5 of
bipy), 123.8 (s, C-3ꢀ of bipy), 124.5 (s, C-3 of bipy), 148.6 (s, C-4 and
4ꢀ of bipy), 153.0 (s, C-6 and 6ꢀ of bipy), 155.6 (s, C-2 and 2ꢀ of bipy).
Elemental analysis, calculated for C24H36N2: C, 81.76; H, 10.29; N,
7.95. Experimental: C, 80.94; H, 9.95; N, 7.80%.
2.3.2. [Mo(O)(O2)2(E)] (1e)
2.2.7. 3-Hexyl-5-methylpyrazole (I)
To a solution of 4,4ꢀ-di(tridecyl)-2,2ꢀ-bypiridine (E) (0.19 mmol)
in methanol (15 mL) was added aqueous [Mo(O)(O)2(H2O)n]
(1 equiv., 273 ␮L, 0.191 mmol) and the resulting solution was left
stirring for 30 min. After this time the resulting precipitate was
isolated by filtration, washed with cold distilled water and dried
for several hours under vacuum. The product 1e, the oxodiperox-
omolybdenum complex of E, was obtained as a yellow powder.
IR (KBr, cm−1): 660, 723, 764, 836, 861, 948, 1031, 1420, 1467,
1610, 2850, 2920, 3058, 3117, 3432. 1H NMR (CDCl3): ı 0.81 (br
t, JHH = 6.3 Hz, 6H, CH3), 1.18 (v br, 40H, intermediate CH2), 1.55
(br m, 2H, bipy-CH2CH2CH2), 1.73 (br m, 2H, bipy-CH2CH2CH2),
2.62 (t, JHH = 7.7 Hz, 2H, bipy-CH2CH2), 2.85 (t, JHH = 7.8 Hz, 2H, bipy-
CH2CH2), 7.19 (d, 1H, 5- or 5ꢀ-CH of bipy), 7.61 (d, 1H, 5- or 5ꢀ-CH
1-Octyne (3.8 mL, 25 mmol) was degassed, dissolved in dry THF
(25 mL), cooled to −78 ◦C and 1.6 M n-BuLi in hexanes (18.75 mL,
30 mmol) was added. The solution was allowed to warm to room
temperature and stirred for 5 min. The reaction was again cooled to
−78 ◦C and 1.0 M ZnCl2 in diethyl ether (25 mL, 25 mmol) carefully
added. The reaction was warmed to 0 ◦C. Acetyl chloride (2.0 mL,
27.5 mmol) was then added and the reaction was stirred for 6 h.
Working now under air, the resulting solution was washed with sat-
urated NH4Cl (25 mL) and the phases separated. The aqueous phase
was washed twice with ether (25 mL) and all of the organic phases
were then combined and the solvents evaporated on a rotavap.
The crude intermediate was then dissolved in methanol (50 mL)

114
M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
of bipy), 7.81 (s, 1H, 3- or 3ꢀ-CH of bipy), 8.10 (s, 1H, 3- or 3ꢀ-CH of
bipy), 8.21 (d, 1H, 6- or 6ꢀ-CH of bipy), 9.36 (d, 1H, 6- or 6ꢀ-CH of
1
bipy). 13C{ H} NMR (CDCl3): ı 14.1 (s, CH3), 22.7 (s, intermediate
CH2), 29–30.5 (several s, intermediate CH2), 31.9 (s, intermediate
CH2), 35.6, 36.0 (s, bipy-CH2), 121.0, 122.8, 126.7, 126.8 (s, C-3 and
C-5 of bipy), 147.4, 147.8, 154.1, 155.1, 156.4, 160.8 (s, C-2, C-4 and
C-6 of bipy). Elemental analysis, calculated for MoC36H60O5N2: C,
62.05; H, 8.68; N, 4.02. Experimental: C, 61.56; H, 8.75; N, 3.97%.
2.3.3. [Mo(O)(O2)2(F)] (1f)
This product was synthesised following the same experimen-
tal method as for 1b but with F in place of B. The product 1f, the
oxodiperoxomolybdenum complex of F, was obtained as a yellow
powder. IR (KBr, cm−1): 622, 655, 691, 776, 843, 866, 946, 1030,
1219, 1252, 1420, 1603, 2899, 2954, 3447. 1H NMR (CDCl3): ı 0.00,
0.12 (s, 36H, Si(CH3)3), 1.68, 1.95 (s, 2H, CH), 6.82, 7.22 (d, JHH = 6 Hz,
2H, 5-CH and 5ꢀ-CH of bipy), 7.38, 7.64 (br s, 2H, 3-CH and 3ꢀ-CH of
bipy), 8.00, 9.14 (d, JHH = 6 Hz, 2H, 6-CH and 6ꢀ-CH of bipy). Elemen-
tal analysis, calculated for MoC24H44O5N2Si4: C, 44.42; H, 6.83; N,
4.32. Experimental: C, 43.67; H, 6.66; N, 4.32%.
Scheme 1. Pyridine-based ligands.
ether (3× 3 mL). The resulting solution was dried (MgSO4) and anal-
ysed by GC.
2.5. Computational details
2.3.4. [Mo(O)(O2)2(G)] (1g)
The electronic structure and geometries of the model com-
pounds were computed using density functional theory at the
B3LYP level [20,21]. The Mo atom was described with the LANL2DZ
basis set [22] whilst the 6-31G(d,p) basis set was used for the C,
O, N and H atoms. The DFT calculations were performed using
the Gaussian 03 suite of programs [23]. The nature of the opti-
mised geometries of all the compounds were characterised by the
calculated number of imaginary frequency at the same level of the-
ory, i.e., energy minimum structures without imaginary frequencies
(NImag = 0). Transition states were located using the quadratic syn-
chronous transit (QST2) approach [24] and frequency calculations
were performed in order to check the stationary states (NImag = 1).
This product was synthesised following the same experimental
method as for 1b but with G in place of B. The product 1g, the
oxodiperoxomolybdenum complex of G, was obtained as a yellow
powder. IR (KBr, cm−1): 689, 758, 778, 842, 866, 942, 988, 1036,
1250, 1420, 1610, 2899, 2952, 3447. 1H NMR (CDCl3): ı 0.00, 0.08
(s, 36H, Si(CH3)3), 0.11, 0.28 (t, JHH = 6.3, 6.4 Hz, respectively, 2H,
CH2CH), 2.75, 2.97 (d, JHH = 6.3, 6.4 Hz, respectively, 4H, CH2CH),
7.11, 7.53 (d, JHH = 6 Hz, 2H, 5-CH and 5ꢀ-CH of bipy), 7.69, 7.97 (s, 2H,
3-CH and 3ꢀ-CH of bipy), 8.10, 9.25 (d, JHH = 6 Hz, 2H, 6-CH and 6ꢀ-CH
1
of bipy). 13C{ H} NMR (CDCl3): ı 0.0, 0.1 (s, Si(CH3)3), 14.6, 15.0 (s,
CH), 31.8, 32.4 (s, CH2), 120.6, 122.4 (s, C-5 of bipy), 126.2, 126.5 (s,
C-3 of bipy), 147.3, 147.5, 153.7, 158.6, 154.9, 163.1 (s, C-2, C-4 and
C-6 of bipy). Elemental analysis, calculated for MoC26H48O5N2Si4:
C, 46.13; H, 7.15; N, 4.14. Experimental: C, 47.13; H, 7.14; N, 4.50%.
3. Results and discussion
3.1. Ligand synthesis and preparation of metal complexes
2.3.5. [Mo(O)(O2)2(H)] (1h)
This product was synthesised following the same experimen-
tal method as for 1b but with H in place of B. The product 1h, the
oxodiperoxomolybdenum complex of H, was obtained as a yellow
powder. IR (KBr, cm−1): 661, 723, 836, 863, 947, 1032, 1242, 1308,
1420, 1467, 1610, 2850, 2920, 2956, 3448. 1H NMR (CDCl3): two
isomers, ı 0.88 (t, JHH = 6 Hz, 6H, CH3 terminal), 1.25 (v br, 20H,
intermediate CH2), 1.58 (m, 2H, bipy-4ꢀ-CH2CH2CH2), 1.79 (m, 2H,
bipy-4ꢀ-CH2CH2CH2), 2.46 (s, 3H, py-4-CH3), 2.68 (t, JHH = 7.7 Hz,
2H, bipy-4ꢀ-CH2CH2), 2.70 (s, 3H, py-4-CH3), 2.91 (t, JHH = 7.7 Hz,
2H, bipy-4ꢀ-CH2CH2), 7.18, 7.61 (m, 2H, 5-CH and 5ꢀ-CH of bipy),
7.82, 7.85, 8.10, 8.14 (s, 1H, 3-CH and 3ꢀ-CH of bipy), 8.19, 9.34
(m, 2H, 6-CH and 6ꢀ-CH of bipy). Elemental analysis, calculated for
MoC24H36O5N2: C, 54.54; H, 6.87; N, 5.30. Experimental: C, 55.15;
H, 6.80; N, 5.22%.
A
series of ligand species consisting of pyridine, 2,2ꢀ-
bipyridine or pyrazole functionalised with various alkyl
chains, trimethylsilyl groups or polydimethylsiloxane poly-
mer was synthesised (Schemes 1–3). The preparations of
4-polydimethylsiloxanylethylpyridine (A) and 4-tridecylpyridine
(B) have been previously reported by us and are not repeated
here [12b]. The synthesis of the other pyridine ligands (Scheme 1),
4-(2,2-bis-trimethylsilanyl-ethyl)pyridine (C) and 2,6-(bis-
trimethylsilanylmethyl)pyridine (D), were based on a previously
described procedure for the synthesis of related substituted
bipyridines [25] and the synthetic method was analogous to
that which we previously described for the preparation of B
[12b]. The methyl groups of 4-picoline and 2,6-lutidine were
deprotonated by lithium diisopropylamide and reaction with
chloro-bistrimethylsilanyl-methane or chlorotrimethylsilane as
appropriate subsequently gave the products (see details in Section
2). The use of compound D in experimental studies has been pre-
viously described elsewhere, though without details concerning
its synthesis [18].
2.4. General procedure for catalytic olefin epoxidation
The reactor (a 50 mL vial equipped with
a Young valve
and containing a stirrer flea) was charged with 0.5 M aqueous
[Mo(O)(O)2(H2O)n] (50 ␮L, 0.025 mmol), the base additive as spec-
ified (typically 0.1 mmol monodentate, 0.05 mmol bidentate), the
reaction solvent if required (2 mL), the oxidant (30% aqueous H2O2,
350 ␮L, 3 mmol) and the olefin substrate (1 mmol), in the afore-
mentioned order. The reactor was sealed and heated at 60 ◦C,
maintaining constant stirring in a thermostatted oil bath for the
duration of the reaction. Upon completion the reactor was imme-
diately cooled to 0 ◦C and the products extracted with petroleum
In
the
same
manner,
the
functionalised
2,2-
bipyridines (Scheme 2), 4,4ꢀ-di(tridecyl)-2,2ꢀ-bipyridine (E),
4,4ꢀ-bis-(bistrimethylsilanylmethyl)-2,2ꢀ-bipyridine
bis-(2,2-bis-trimethylsilanyl-ethyl)-2,2ꢀ-bipyridine
(F),
(G)
4,4ꢀ-
and
4-methyl-4ꢀ-tridecyl-2,2ꢀ-bipyridine (H) were prepared by depro-
tonation of 4,4ꢀ-dimethyl-2,2ꢀ-bipyridine and subsequent reaction
with appropriate organo or silano halides (see details in Sec-
tion 2). The synthesis and 1H NMR data of compounds E and

M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
115
Scheme 2. 2,2ꢀ-Bipyridine-based ligands.
Scheme 3. Synthesis of alkyl substituted pyrazole ligands.
H have been communicated previously [17]. Complete spec-
troscopic characterisation of these and of the new substituted
bipyridines F and G has been performed (see Section 2 and
Figure S1 in Supplementary data).
Functionalised pyrazoles were synthesised by first coupling
an alkyne, as an organometallic intermediate, with the appropri-
ate acid chloride, followed by reaction with hydrazine to form
the pyrazole ring (Scheme 3). In this manner both 3-hexyl-5-
methylpyrazole (I) and 3-heptyl-5-hexylpyrazole (J) were prepared
and characterised.
CH hydrogen atom of the –CH2CH(SiMe3)2 groups, the signals
shifted slightly upfield with respect to the corresponding signal
on the spectrum of the free ligand G. In the same way, the CH
signals of the aromatic hydrogen atoms, which for the free G
ligand display a doublet + singlet + doublet pattern typical of 4,4ꢀ-
substituted-2,2ꢀ-bipyridines, are clearly split into two set of signals
due to coordination to the {Mo(O)(O2)2}, demonstrating the non-
equivalence of the pyridine rings. The 13C{ H} NMR spectrum of
1
1g affords complimentary information, with ten distinct singlets
observable for the aromatic carbon atoms in the pyridine rings,
rather than the five that would be expected if the rings were equiv-
alent. Analogous observations were made in the NMR spectra of
all of the complexes of the symmetrical bipyridine ligands E–G
(compounds 1e–g), with two distinct sets of signals visible for
each aromatic ring. These observations would seem to indicate that
these complexes adopt the structure shown in Scheme 4, with the
split signals resulting from the electronic difference between the
axially and equatorially coordinated rings in the complex. An anal-
ogous structure has previously been ascertained by X-ray analysis
of [Mo(O)(O2)2(bipy)] [26] and has also been seen in related com-
plexes with substituted bipyridine ligands [27] and other bidentate
N-ligands [9b,c,28].
For the complex of the unsymmetrically substituted bipyridine
H, [Mo(O)(O2)2(H)] (1h), the 1H NMR spectra indicate that both of
the two possible isomers corresponding to the two alternative coor-
dination modes H can adopt with the {Mo(O)(O2)2} core (Scheme 4)
were present, in a ratio at least approximately 1:1, indicating that
neither was formed more favourably. Illustrating this, the 1H NMR
The corresponding complexes of oxodiperoxomolybdenum
were subsequently prepared. Generally, this was achieved fairly
easily by dissolving molybdenum trioxide in excess aqueous hydro-
gen peroxide and reacting this solution appropriately with the
ligands. The monodentate complexes of the pyridines took the
form of viscous yellow coloured oils whilst the bipyridines gave
pale yellow, colloidal solids. The isolation and spectroscopic char-
acterisation of the complexes of the bidentate bipyridine ligands,
1e–h, was both significant and interesting as oxodiperoxomolyb-
denum complexes of bipyridines are usually very insoluble in
apolar organic solvents making it difficult to record their NMR
spectra in solution. The solubilising functions of these species
render them highly soluble in CDCl3 such that well defined spec-
1
tra were easily achieved. The 1H and 13C{ H} NMR spectra of
these complexes are relatively simple since the only signals arise
exclusively from the ligands. As an example, in the spectra of
[Mo(O)(O2)2(G)] (1g), in the 1H NMR spectrum two separate triplet
signals at 0.11 and 0.28 ppm are observed corresponding to the
Scheme 4. Structures of molybdenum complexes 1e–g and the two isomers of 1h.

116
M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
Table 1
Oxobisperoxomolybdenum catalysed cis-cyclooctene epoxidations in apolar organic solvents with a selection of base additivesa.
Entry
Solvent
Hexane
Ligand
Conversion
Yield
1
2
3
4
5
6
7
8
9
Pyridine
5
4
9
0
0
1
2
1
0
0
4-(Polydimethylsiloxanyl-ethyl)-pyridine (A)
2,6-Bis-trimethylsilanylmethyl-pyridine (D)
2,2ꢀ-Bipyridine
4,4ꢀ-Ditridecyl-2,2ꢀ-bipyridine (E)
–
Pyridine
0
2
0
2
CHCl3
4-(Polydimethylsiloxanyl-ethyl)-pyridine (A)
4-Tridecylpyridine (B)
86
99
72
38
12
17
17
86
99
71
38
2
10
4-(2,2-Bis-trimethylsilanyl-ethyl)-pyridine (C)
2,6-Bis-trimethylsilanylmethyl-pyridine (D)
2,2ꢀ-Bipyridine
11
12
13
14
4,4ꢀ-Ditridecyl-2,2ꢀ-bipyridine (E)
4,4ꢀ-Bis(bistrimethylsilanylmethyl)-2,2ꢀ-bipyridine (F)
17
17
a
Experimental conditions: Aqueous [Mo(O)(O2)2(H2O)n] 0.025 mmol, ligand 0.1 mmol for monodentate and 0.05 mmol for bidentate, 30% H2O2(aq) 3.0 mmol, cis-cyclooctene
1.0 mmol, solvent 10 mL, t = 18 h, T = 60 ◦C. Yields and conversions calculated by GC.
of 1h shows two pairs of distinct singlets and triplets corresponding
to the 4-CH3 and 4ꢀ-CH2 groups respectively, both shifted slightly
downfield with respect to the free H ligand. Similarly, a pair of mul-
tiplets corresponding to the 4-CH2CH2 groups and pairs of signals
are apparent for the aromatic protons.
odentate counterparts (entries 8–11, Table 1). Previously we had
made a similar observation in epoxidation reactions in ionic liquids
[16b,c].
The
two
4-substituted
monopyridyl
ligands
4-
(polydimethylsiloxanyl-ethyl)pyridine (A) (entry 8, Table 1)
and 4-tridecylpyridine (B) (entry 9, Table 1) afforded high yields
within the 18 h, with complete conversion to the epoxide observed
in the latter.1 Use of the solubilising ligands seems to result in
complete dissolution of the metal catalyst complex in the organic
phase of a biphasic system and the oxodiperoxomolybdenum
complexes of the monodentate pyridine ligands are apparently
reasonably efficient catalysts. The other monodentate pyridine
ligands tested, the structural isomers 4-(2,2-bis-trimethylsilanyl-
ethyl)pyridine (C) and 2,6-bis(trimethylsilanylmethyl)pyridine
(D) (entries 10 and 11, Table 1), afforded significant epoxidation,
though with a marked difference in the observed yields indicating
that steric hindrance due to the proximity of the 2,6-functions
to the metal centre in the former may influence the epoxidation,
resulting in slower oxidation. However, the benefits from the
solubilising effect of the methyl-trimethylsilanyl substituents still
clearly outweigh this hindrance compared with unsubstituted
pyridine (entry 7, Table 1), showing that solubilisation of the
catalyst complexes is a crucial factor.
Investigations of epoxidation in systems employing A and
B were subsequently extended to other olefin substrates (see
Table 2). Six substrates were tested in this study, cyclohexene
which is relatively active to epoxidation, the inactive terminal
alkene 1-octene, one cis and two trans secondary alkenes which
should be more activated and also styrene, the oxide of which is
very vulnerable to decomposition through hydrolysis.
A brief examination of the results is sufficient to conclude that
neither of the base additives investigated induced a catalytic sys-
tem which was active in the epoxidation of this wider range of
substrates. Thus, the high activities observed in the preliminary
investigation with cis-cyclooctene seem to be limited to only this
sterically and electronically activated substrate. In the epoxida-
tion of cyclohexene, A produced the most active system, (entry 7,
Table 2) with a reduced level of conversion observed for B (entry
3.2. Mo-catalysed epoxidation reactions
Subsequently, the catalytic activity of these complexes in the
epoxidation of olefins in apolar media was investigated, studying
first of all the oxidation of cis-cyclooctene (Scheme 5) in organic sol-
vent media; chloroform and hexane (see Table 1). In these reactions
the molybdenum catalyst complex formed in situ from an aqueous
solution of oxodiperoxomolybdenum, [Mo(O)(O2)2(H2O)n], and
the respective base additive, mixing them together at the start of
the reaction. Other experimental details are included in the corre-
sponding table footnotes.
This preliminary study focussed only on the functionalised pyri-
dine and bipyridine ligands, which were compared with their
unfunctionalised analogues. In chloroform, all of the functionalised
ligands solubilised the catalyst sufficiently to result in reaction sys-
tems wherein the catalyst and olefin substrate were dissolved in
the reaction solvent with the aqueous hydrogen peroxide oxidant
forming a separate phase. When the unfunctionalised analogues
were used however, the molybdenum complexes were always
observed to precipitate as a yellow solid, leaving no detectable
colouration in the organic phase. In hexane there was never any
observable solubility even for the functionalised ligands.
Significant conversion in hexane was never achieved (entries
1–5, Table 1), probably attributable to the insolubility of the cat-
alysts in such a highly apolar solvent. Following these results we
discontinued studies in hydrocarbon solvents. In chloroform, when
no base or the unsubstituted analogue bases (entries 6, 7 and 12,
Table 1) were employed there was never any significant conversion
to the epoxide product, again probably due to the insolubility of the
catalyst complexes. Low, but significant conversions were observed
when 4,4ꢀ-dialkyl substituted bipyridines were used (entries 13 and
14, Table 1). However, the yields with these bidentate ligands were
found to be markedly lower than those obtained with their mon-
1
It should be noted that certainly some of the pyridine coordinating species that
were investigated here probably converted in situ to their corresponding N-oxides,
a reaction that we have demonstrated for B. In fact, IR and elemental analysis of
the product from the reaction of B with oxodiperoxomolybdenum indicated that
the N-oxide complex, [Mo(O)(O2)2(BO)(H2O)], had been isolated [16c]. Additionally,
oxidation of 4-tridecylpyridine by oxodiperoxomolybdenum species in ionic liquid
medium has been also demonstrated [16b].
Scheme 5. Selected test reaction.

M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
117
Table 2
Oxobisperoxomolybdenum catalysed olefin epoxidations in chloroforma.
Entry
Ligand
Olefin
Conversion
Yield
1
2
Cyclohexene
1-Octene
23
0
5
0
B
3
4
5
6
7
8
trans-2-Octene
trans-4-Octene
cis-2-Heptene
Styrene
Cyclohexene
1-Octene
15
25
29
60
75
0
13
22
29
0
75
0
N
C13H27
A
9
trans-2-Octene
trans-4-Octene
cis-2-Heptene
Styrene
2
2
1
100
3
2
1
0
N
C2H4 Si
10
11
12
O
n
a
Experimental conditions: Aqueous [Mo(O)(O2)2(H2O)n] 0.025 mmol, base additive 0.10 mmol, 30% H2O2(aq) 3.0 mmol, olefin substrate 1.0 mmol, chloroform 10 mL, t = 18 h,
T = 60 ◦C. Yields and conversions calculated by GC.
1, Table 2). However, given the long reaction time the incomplete
conversion for such an active substrate still indicates fairly limited
efficiency. In neither system was any trace of epoxidation detected
for the terminal olefin 1-octene (entries 2 and 8, Table 2). In spite of
its markedly elevated efficiency in the epoxidation of cyclohexene,
the system using A generated only trace yields for all of the other
substrates tested (entries 9–12, Table 2). In the conversion of the
secondary alkenes B did show a moderate level of activity however
(entries 3–5, Table 2). In neither system was any unreacted styrene
detected after reaction (entries 6 and 12, Table 2), although the high
conversions indicate that whilst the styrene was probably oxidised
to the epoxide it was subsequently transformed. In both cases GC
analysis showed low levels of benzaldehyde which may therefore
be an intermediate breakdown product.
The study subsequently moved to solventless epoxidations, sim-
ilar to those we previously studied using methyltrioxorhenium
catalysts [12b]. In the same way as in the solvent based studies this
began with an investigation of the influence of the various ligands in
the epoxidation of cis-cyclooctene (see Table 3). The bipyridine type
ligands were not investigated in this study but we introduced the
alkyl functionalised pyrazoles 3-hexyl-5-methylpyrazole (I) and 3-
hexyl-5-heptylpyrazole (J), having witnessed in related studies that
pyrazole ligands can produce more activated oxodiperoxomolyb-
denum catalysts for epoxidations [16], sulphide oxidation [29], and
alkylbenzene oxidation [30].
this may also be a factor. Moving on to the functionalised ligands,
both A and B gave enhanced yields when compared to unfunc-
tionalised pyridine (entries 5 and 6, Table 3), but yields were well
short of complete unlike in chloroform. When the functionalised
pyrazoles I and J were tested however, complete conversion of
the olefin substrate was observed within 18 h (entries 7 and 8,
Table 3). Selectivity was lower for I (63%) than J (95%) however,
indicating that there was a greater extent of epoxide hydrolysis in
the former. By reducing the reaction time to 4 h these selectivities
could be improved so that with J a 100% conversion with complete
selectivity for the epoxide was achieved. The reactions were also
analysed after 66 h, finding that the hydrolysis slowly continues
over time, although in the case of J the yield still remained above
90%. The heightened activity of these latter two pyrazole based
systems when compared to unfunctionalised pyrazole analogues
is presumably due to the enhanced solubility of the metal com-
plexes, whilst compared to the functionalised pyridines the use of
a pyrazole type ligand as opposed to a pyridine results in a more
active catalyst. As previously mentioned we have made similar
observations in ionic liquid based systems and the resistance
of the pyrazole species to oxidation may well be the reason for
this enhancement [16]. Pyridines are fairly quickly converted to
their corresponding N-oxides under the oxidising conditions but
pyrazoles do not undergo any such oxidation with the result that,
in situ, pyrazoles are the stronger donor ligands, being N- rather
than O-donor ligands. This may result in an electronically enhanced
catalyst complex and subsequently more efficient epoxidation.
The epoxidation systems utilising pyrazoles I and J were mon-
itored over a 2 h period to observe the profile and progress of the
epoxidation (see data in Table S1 in Supplementary data). Fig. 1
shows the solventless epoxidation reaction where J is the base
additive used and Figure S2 displays the same reaction for I. These
experiments demonstrated that epoxidation in these systems pro-
ceeds relatively rapidly, with almost all of the substrate consumed
Only minor epoxide yields (less than 10%) were generated in
the absence of any ligand or where unfunctionalised pyridine was
used (entries 1 and 2 respectively, Table 3). Pyrazole was found to
produce a slightly more active system and 3,5-dimethylpyrazole
gave an even higher yield of 31% (entries 3 and 4, Table 3). This
additional enhancement may be due to the solubilising influence of
the methyl groups but we have previously observed that the more
basic 3,5-dimethylpyrazole seems to produce a greater activating
effect than unfunctionalised pyrazole in similar systems [16] and
Table 3
Oxobisperoxomolybdenum catalysed solventless epoxidation of cis-cyclooctenea.
Entry
Ligand
Conversion (yield)
t = 4 h
t = 18 h
12 (5)
t = 66 h
1
2
3
4
5
6
7
8
None
Pyridine
Pyrazole
3,5-Dimethylpyrazole
A
B
I
–
–
–
–
–
–
–
–
–
–
–
8 (8)
16 (12)
31 (31)
25 (23)
31 (31)
100 (63)
100 (95)
–
99 (88)
100 (100)
100 (29)
100 (91)
J
a
Experimental conditions: Aqueous [Mo(O)(O2)2(H2O)n] 0.025 mmol, base additive 0.10 mmol, 30% H2O2(aq) 3.0 mmol, olefin substrate 1.0 mmol, T = 60 ◦C. Conversions
and yields calculated by GC.

118
M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
Fig. 2. Calculated Gibbs free energy profile of the model epoxidation reaction with
[Mo(O)(O2)2(J)] (kcal mol−1).
Fig. 1. Reaction profile of the oxobisperoxomolybdenum catalysed solventless
epoxidation of cis-cyclooctene with J as ligand. Experimental details in Table S1.
range of substrates was limited. Only barely higher than trace yields
were recorded for 1-octene in all of the systems. For the secondary
trans and cis olefin substrates the systems using dimethylpyrazole
and A demonstrated generally lowered activities compared to those
using B, I and J. There was no obvious difference in reactivity toward
the cis and trans substrates indicating that neither is more inhib-
ited nor activated than the other. Traces of epoxide were obtained
in the epoxidation of styrene, though with significant levels of con-
version in all cases, indicating that the epoxidation may well have
proceeded but that subsequently the product was largely lost to
hydrolysis. In the systems using I and J a 7% yield of benzaldehyde
was detected.
within little over half an hour. In the system utilising I the epoxide
subsequently begins to hydrolyse whilst where J was used there
was no substantial evidence of this within the 2 h. It is not clear
why the hydrolysis reaction was so much more rapid in the sys-
tem using I, it may be attributable to differences in the solubilities
of the respective complexes of I and J but this is entirely spec-
ulative. The TON and TOF calculated for the reaction with J are
9.6 and 115 h−1 (for 5 min), respectively. These values are lower
than those previously reported by us for a rhenium catalysed sol-
ventless epoxidation system [12b] though this is unremarkable as
methyltrioxorhenium derivatives are typically much more active
epoxidation catalysts than oxomolybdenum compounds [31]. The
values are also far lower than those reported for the Mo catal-
ysed olefin epoxidation system developed by Bhattacharyya and
co-workers, which constitutes one of the optimal examples of
an oxomolybdenum catalysed epoxidation [11,32]. However the
solventless systems reported here are advantageous in that nei-
ther a co-catalyst, such as NaHCO3, nor any organic solvent is
required. Finally, a further study was conducted using a much lower
catalyst:substrate ratio (0.25 mol% catalyst), under the same exper-
imental conditions, raising the achievable TON to 1820, but with a
lower TOF value.
In the same way as for the study of epoxidations in chloro-
form, the efficiency of the solventless systems in the oxidation of
alternative olefin substrates was investigated. Dimethylpyrazole,
the functionalised pyridines, A and B, and the functionalised pyra-
zoles, I and J, were all studied in this capacity. These results are
shown in Table 4. A quick overview of the results reveals that, in
common with the study in chloroform, none of the systems inves-
tigated had a particularly high activity toward the less activated
substrates. Somewhat disappointingly the activity of even the sys-
tems using the functionalised pyrazoles I and J toward the wider
3.3. DFT study of the mechanism of the stoichiometric
epoxidation reaction
Density functional calculations were carried out to obtain
further insight into the mechanism of the epoxidation reac-
tion. Ethylene, the simplest olefin model was selected as the
substrate with the oxodiperoxomolybdenum complex of 3-hexyl-
5-heptylpyrazole as the oxidising molybdenum species, this
complex being the most active under solventless conditions. The
starting model compound [Mo(O)(O2)2(J)] (1), as well as other con-
sidered intermediates (see Scheme 6), were optimised without
symmetry restrictions at the B3LYP level.
The reaction profile for the epoxidation is shown by Fig. 2, which
contains the optimised structures for compounds 1–3 and the tran-
sition state TS. The reaction is exergonic, in common with related
studies [33]. The first step is the oxo-transfer reaction from the per-
oxo ligand of complex A to the ethylene substrate which affords
the dioxoperoxo complex 2, wherein the generated epoxide is ini-
tially coordinated to the molybdenum centre. This intermediate
is characterised by the transition structure TS which takes the
Table 4
Oxobisperoxomolybdenum catalysed solventless epoxidationsa.
Ligand
Conversion (yield)
1-Octene
trans-2-Octene
trans-4-Octene
cis-2-Heptene
Styrene
3,5-Dimethylpyrazole
3(1)
1(1)
5(1)
1(1)
14(1)
5(5)
4(4)
44(22)
73(24)
65(23)
17(10)
3(3)
21(18)
48 (25)
68 (32)
87(1)
13(6)
51(12)
71 (30)
49 (36)
15(3)
23(3)
62(8)
87(1)b
83(1)b
A
B
I
J
a
Experimental conditions: Aqueous [Mo(O)(O2)2(H2O)n] 0.025 mmol, base additive 0.10 mmol, 30% H2O2(aq) 3.0 mmol, olefin substrate 1.0 mmol, t = 18 h, T = 60 ◦C. Con-
versions and yields calculated by GC.
b
7% yield of benzaldehyde quantified.

M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
119
Scheme 6. Catalytic cycle for the epoxidation reaction.
form of the spirocycle originally proposed by Sharpless [34] and
Acknowledgements
theoretically analysed for other [Mo(O)(O2)2(L)] models by sev-
eral groups [35]. The structural parameters of the TS here agree
well with those of similar species as previously reported [16a,35].
The calculated energy barrier for TS (28 kcal mol−1) is also simi-
lar to other computed values and is equivalent to that found for
the closely related [Mo(O)(O2)2(dmpz)] compound (dmpz = 3,5-
dimethylpyrazol) [16a]. The dissociation of the epoxide from the
intermediate 2 is slightly favourable (ca. −3 kcal mol−1) and yields
the dioxoperoxo complex 3. The catalytic cycle is closed by the
energetically favourable (−9.4 kcal mol−1) reaction of hydrogen
peroxide with 3 which regenerates the starting complex 1. Solvent
effects were included in the calculations and no significant influ-
ence in the energetic reaction profile was detected (see Table S2 in
the Supplementary data).
This research was supported by the Spanish Ministerio de Cien-
cia e Innovación (CTQ2007-61037 and CTQ2010-15515) and Junta
de Andalucía (Proyecto de Excelencia, FQM-2474).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcata.2011.02.004.
References
[1] (a) R.P. Nielsen, J.H. La Rochelle, US Pat., 3962136 (1976);
(b) D. Kahlich, U. Wiechern, J. Lindner, Propylene Oxide; Ullmann’s Encyclope-
dia of Industrial Chemistry, Wiley-VCH, Weinheim, 2002, pp. 4–9.
[2] A.S. Rao, in: B.M. Trost, I. Fleming, S.V. Ley (Eds.), Comprehensive Organic Syn-
thesis, vol. 7, Pergamon Press, Oxford, 1991, p. 357.
4. Conclusions
[3] Corresponding to principles outlined in:
(a) P.T. Anastas, J.C. Warner, Green Chemistry: Theory and Practice, Oxford
University Press, Oxford, 1998;
(b) M. Lancaster, Green Chemistry: An Introductory Text, RSC Paperbacks, Royal
Society, 2002.
It was shown that low polarity substituents (alkyl chains,
trimethylsilane bearing functions, polydimethylsiloxane poly-
mers) attached to the coordinated ligands of oxodiperoxomolyb-
denum complexes enhance their solubility in low polar organic
solvents. In epoxidation reactions the solubility enhancements
induced by appropriate ligands lead to high yields in the epox-
idation of cis-cyclooctene, not attainable with unfunctionalised
analogue catalysts due to their insolubility. Pyrazoles were found
to induce significantly more efficient epoxidation than pyridines,
whilst 2,2ꢀ-bipyridines seem to slow the reaction. Of particular
interest were the observations made in the epoxidations conducted
under solventless conditions, where the oxodiperoxomolybdenum
complex of 3-hexyl-5-heptylpyrazole showed markedly enhanced
catalytic efficiency in the epoxidation of cis-cyclooctene with
hydrogen peroxide. The elimination of the necessity for any organic
solvent and the absence of any co-catalyst makes this system sig-
nificantly advantageous in green chemistry terms. However, the
substrate range of these systems is limited, whilst the complexes
are active in the epoxidation of less active substrates the oxida-
tions proceed too slowly to be of practical use. Surmounting this
limitation is a key objective of our continuing research.
[4] See examples in:
(a) S.T. Oyama (Ed.), Mechanisms in Homogeneous and Heterogeneous Epoxi-
dation Catalysis, Elsevier Science, 2008;
(b) W.J. Mijs, C.H.R.I. de Jonge (Eds.), Organic Synthesis by Oxidation with Metal
Compounds, Plenum Press, New York, 1986.
[5] For example see one of the first reports on this reaction: W.A. Herrmann, R.W.
Fischer, M.U. Rauch, W. Scherer, J. Mol. Catal. 86 (1994) 243–266.
[6] (a) J. Hartwig, Organotransition Metal Chemistry: From Bonding to Catalysis,
University Science Books, Sausalito, California, 2010;
(b) V. Conte, O. Bortolini, Transition Metal Peroxides, Synthesis and Role in
Oxidation Reactions in “The Chemistry of Peroxides”, vol. 2, Part 1, Wiley, 2006,
p. 1053.
[7] See for instance: F.E. Kühn, A.M. Santos, M. Abrantes, Chem. Rev. 106 (2006)
2455.
[8] For commonly cited examples see:
(a) H. Mimoun, I. Seree de Roch, L. Sajus, Tetrahedron 26 (1970) 37–50;
(b) K.B. Sharpless, J.M. Townsend, D.R. Williams, J. Am. Chem. Soc. 94 (1972)
295–296.
[9] (a) W.R. Thiel, M. Angstl, N. Hansen, J. Mol. Catal. A: Chem. 103 (1995) 5–10;
(b) W.R. Thiel, M. Angstl, T. Priemeier, Chem. Ber. 127 (1994) 2373–2379;
(c) W.R. Thiel, Angew. Chem. Int. Ed. 34 (1995) 1737–1738;
(d) W.R. Thiel, Chem. Ber. 129 (1996) 575–580.
[10] (a) G. Grigoropoulos, J.H. Clark, J.A. Elings, Green Chem. 5 (2003) 1–7;
(b) B.S. Lane, K. Burguess, Chem. Rev. 103 (2003) 2457–2473.

120
M. Herbert et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 111–120
[11] N. Gharah, S. Chakraborty, A.K. Mukherjee, R. Bhattacharyya, Chem. Commun.
(2004) 2630–2632.
[12] For our related work in this area see:
mann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K.
Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich,
A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari,
J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Ste-
fanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith,
M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. John-
son, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision C.02,
Gaussian, Inc., Wallingford, CT, 2004.
(a) M. Herbert, F. Montilla, A. Galindo, Dalton Trans. 3 (2010) 900–907;
(b) M. Herbert, A. Galindo, F. Montilla, Organometallics 28 (2009) 2855–2863;
(c) M. Herbert, F. Montilla, A. Galindo, Inorg. Chem. Commun. 10 (2007)
735–737.
[13] W.R. Thiel, J. Eppinger, Chem. Eur. J. 3 (1997) 696–705.
[14] (a) C.I. Altinis Kiraz, L. Mora, L.S. Jimenez, Synthesis 1 (2007) 92–96;
(b) G. Wahl, D. Kleinhenz, A. Schorm, J. Sundermeyer, R. Stowasser, C. Rummey,
G. Bringmann, C. Fickert, W. Kiefer, Chem. Eur. J. 5 (1999) 3237–3251.
[15] (a) M. Herbert, F. Montilla, R. Moyano, A. Pastor, E. Álvarez, A. Galindo, Polyhe-
dron 28 (2009) 3929–3934;
[24] C. Peng, P. Ayala, H. Schlegel, M. Frisch, J. Comput. Chem. 17 (1996) 49.
[25] C.L. Fraser, N.R. Anastasi, J.J.S. Lamba, J. Org. Chem. 62 (1997) 9314–9317.
[26] E.O. Schlemper, G.N. Schrauzer, L.A. Hughes, Polyhedron 3 (1984) 377–380.
[27] (a) T.R. Amarante, F.A.A. Paz, S. Gago, I.S. Gonc¸ alves, M. Pillinger, A.E. Rodrigues,
M. Abrantes, Molecules 14 (2009) 3610–3620;
(b) M. Herbert, A. Galindo, F. Montilla, Catal. Commun. 8 (2007) 987–990.
[16] (a) M. Herbert, E. Álvarez, D.J. Cole-Hamilton, F. Montilla, A. Galindo, Chem.
Commun. 46 (2010) 5933–5935;
(b) M. Herbert, F. Montilla, A. Galindo, R. Moyano, A. Pastor, E. Álvarez, Dalton
Trans., accepted for publication.;
(c) M. Herbert, Ph.D. Thesis, Universidad de Sevilla, 2010.
[17] C.-H. Fischer, J. Photochem. 24 (1984) 415–418.
[18] R. Hacker, P. von, R. Schleyer, G. Reber, G. Mueller, L. Brandsma, J. Organomet.
Chem. 316 (1986) C4–C8.
[19] For example see:
(b) W.A. Herrmann, W.R. Thiel, J.G. Kuchler, J. Behm, E. Herdtweck, Chem. Ber.
123 (1990) 1963–1970;
(c) E.O. Schlemper, G.N. Schrauzer, L.A. Hughes, Polyhedron 3 (1984) 377.
[28] (a) E. da Palma Carreiro, Y.-E. Guo, A.J. Burke, Inorg. Chim. Acta 359 (2006) 1519;
(b) J.A. Brito, M. Gomez, G. Muller, H. Teruel, J.-C. Clinet, E. Dunach, M.A. Maestro,
Eur. J. Inorg. Chem. (2004) 4278;
(c) M.J. Hinner, M. Grosche, E. Herdtweck, W.R. Thiel, Z. Anorg, Allg. Chem. 629
(2003) 2251;
(d) H. Glas, M. Spiegler, W.R. Thiel, Eur. J. Inorg. Chem. (1998) 275.
[29] F. Batigalhia, M. Zaldini-Hernandes, A.G. Ferreira, I. Malvestiti, Q.B. Cass, Tetra-
hedron 57 (2001) 9669–9676.
(a) L.J. Csáinyi, Transit. Metal Chem. 14 (1989) 298–302;
(b) L.J. Csanyi, I. Horvath, Z.M. Galbacs, Transit. Metal Chem. 14 (1989) 90–94;
(c) E. Richardson, J. Less-Common Met. 2 (1960) 360;
[30] S. Das, T. Bhowmick, T. Punniyamurthy, D. Dey, J. Nath, M.K. Chaudhuri, Tetra-
hedron Lett. 44 (2003) 4915.
(d) G.M. Vol’dman, E.A. Mironova, L.V. Bystrov, Zh. Neorg. Khim+ 35 (1990)
1306–1309.
[31] (a) D. Betz, W.A. Herrmann, F.E. Kühn, J. Organomet. Chem. 694 (2009)
3320–3324;
[20] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
(b) F.E. Kühn, A.M. Santos, M. Abrantes, Chem. Rev. 106 (2006) 2455;
(c) F.E. Kühn, A.M. Santos, W.A. Herrmann, Dalton Trans. (2005) 2483.
[32] (a) S.K. Maiti, S. Dinda, R. Bhattacharyya, Tetrahedron Lett. 49 (2008) 6205;
(b) S.K. Maiti, K.M.A. Malik, S. Gupta, S. Chakraborty, A.K. Ganguli, A.K. Mukher-
jee, R. Bhattacharyya, Inorg. Chem. 45 (2006) 9843.
[21] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[22] (a) T.H. Dunning Jr., P.J. Hay, Modern Theoretical Chemistry, Plenum, New York,
1976, p. 1;
(b) P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299.
[23] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheese-
man, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P.
Hratchian, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Strat-
[33] (a) P. Gisdakis, I.V. Yudanov, N. Rösch, Inorg. Chem. 40 (2001) 3755;
(b) D.V. Deubel, J. Sundermeyer, G. Frenking, Inorg. Chem. 39 (2000) 2314;
(c) D.V. Deubel, J. Sundermeyer, G. Frenking, J. Am. Chem. Soc. 122 (2000) 10101.
[34] K.B. Sharpless, J.M. Townsend, J. Am. Chem. Soc. 94 (1972) 295.
[35] (a) D.V. Deubel, J. Phys. Chem. A 105 (2001) 4765;
(b) D.V. Deubel, J. Sundermeyer, Eur. J. Inorg. Chem. (2001) 1819.