Ionic liquid surfactants as multitasking micellar catalysts for epoxidations in water

Article abstract of DOI:10.1039/d0cy00673d

We present a single-component catalyst system for the epoxidation of hydrophobic olefins in aqueous solution. The aggregation of surface-active 1-alkyl-3-methylimidazolium ionic liquids (IL) containing the catalytically active tungstate dianion leads to micelles, which solubilise apolar olefins in aqueous media. The micellisation of tungstate allows for the epoxidation of cyclooctene and other olefins in water, using environmentally benign hydrogen peroxide as oxidant. The structural characterisation of the micelles under catalysis conditions as well as the substrate uptake have been studied by cryo-TEM. 183W-NMR studies revealed that the catalytically active species is a tetraperoxotungstate complex. The addition of organophosphonic acids results in a significant boost of the catalytic activity by formation of a tungstate-phosphonate adduct, investigated using electrospray ionisation mass spectrometry (ESI-MS). The versatile functionalisability of the imidazolium cation enables a covalent linkage of a phosphonate group, which is further increasing the catalytic activity.

Full text of DOI:10.1039/d0cy00673d

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Ionic liquid surfactants as multitasking micellar
catalysts for epoxidations in water†
Cite this: DOI: 10.1039/d0cy00673d
Fabian Schmidt,a Bastian Zehner,b Wolfgang Korth,b
a
Andreas Jessb and Mirza Cokoja
*
We present a single-component catalyst system for the epoxidation of hydrophobic olefins in aqueous
solution. The aggregation of surface-active 1-alkyl-3-methylimidazolium ionic liquids (IL) containing the
catalytically active tungstate dianion leads to micelles, which solubilise apolar olefins in aqueous media.
The micellisation of tungstate allows for the epoxidation of cyclooctene and other olefins in water, using
environmentally benign hydrogen peroxide as oxidant. The structural characterisation of the micelles under
catalysis conditions as well as the substrate uptake have been studied by cryo-TEM. 183W-NMR studies
Received 3rd April 2020,
Accepted 9th June 2020
revealed that the catalytically active species is
a tetraperoxotungstate complex. The addition of
organophosphonic acids results in a significant boost of the catalytic activity by formation of a tungstate–
phosphonate adduct, investigated using electrospray ionisation mass spectrometry (ESI-MS). The versatile
functionalisability of the imidazolium cation enables a covalent linkage of a phosphonate group, which is
further increasing the catalytic activity.
DOI: 10.1039/d0cy00673d
rsc.li/catalysis
transfer the catalyst into the organic phase; the formation of the
active species and the oxygen transfer to the substrate take place
Introduction
The epoxidation of unfunctionalised olefins beyond ethylene
and propylene is usually catalysed by tailor-made metal
complexes in homogeneous phase with good selectivities and
high reaction rates.1–3 However, these approaches are often
expensive and the product separation from the catalyst and
catalyst reusability is in most cases either impossible or not
studied. Furthermore, molecular metal-containing catalysts
often require oxidants, which are ecologically questionable or
difficult to handle, such as organic peroxides or iodosobenzene.
In consequence, these issues limit the application of molecular
epoxidation catalysts beyond the laboratory scale. For these
reasons, simple and recyclable catalysts with the ability to
oxidise nonpolar olefins with environmentally benign hydrogen
peroxide are of high interest. Biphasic liquid–liquid
epoxidations of olefins with H2O2 using phase-transfer additives
are well-known and they have been widely studied with water-
soluble complexes of group 6 transition metals, particularly
molybdenum and tungsten.4–9 The most common additives –
crown ethers, tetraalkylammonium- or alkylpyridinium salts –
at the interface of the organic and the aqueous phase.10,11
Hence, both product and catalyst are in the organic phase and
consequently, separation is intricate and render the
multicomponent catalytic system very complex. Therefore, a
reaction in the aqueous phase is of advantage, as – in the ideal
case – the substrate/product phase is free of the water-soluble
catalyst, thus enabling a more facile product separation. For
this, however, it has to be ensured that the substrate is
solubilised in the catalyst/oxidant phase, which can be realised
by micellar phase transfer agents or surfactants.12 A substrate
amount higher than the solubility inside the micelles leads to a
biphasic system, consisting of the upper substrate/product
phase and the quasi-homogeneous micellar solution, enabling a
a Chair of Inorganic and Metal-Organic Chemistry, Faculty of Chemistry and
Catalysis Research Center, Technical University of Munich, Ernst-Otto-Fischer-Str.
1, D-85747 Garching bei München, Germany. E-mail: mirza.cokoja@tum.de
b Chair of Chemical Engineering, Faculty of Engineering Science, Universitätsstraße
30, D-95447 Bayreuth, Germany
Fig.
1 Simplified concept of micellar biphasic epoxidation. The
† Electronic supplementary information (ESI) available: All synthesis procedures,
full cryo-TEM images, micellar calculations, as well as tensiometric and
conductometric measurements. See DOI: 10.1039/d0cy00673d
micelles are shown as spheres in the aq. phase. Both educt and
product are immiscible in water, apart from the fraction inside the
micelles.
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
simple product separation with the micellar catalyst remaining
in the oxidant phase (Fig. 1).
(Do)) were synthesised via a convenient anion exchange
procedure (for experimental details see the ESI†). The 1-alkyl-
3-methylimidazolium bromide precursors [RMIM]Br were
dissolved in water and converted to the corresponding
[RMIM][OH] via an ion exchange resin. The basic eluate was
subsequently added to a suspension of tungstic acid in water.
After filtration of the aqueous product solution and
subsequent solvent removal in vacuo, the [RMIM]2ijWO4] ILs
were obtained in nearly quantitative yields (>95%). The
reaction can easily be scaled up to a multi-gram synthesis
without any limitations. For a precise description of the
micellar system, the CMC values have to be determined in
order to know the minimum required catalyst amount. The
CMC is the concentration, which is required to saturate the
aqueous phase with surfactant molecules. Above this
concentration, the solubility of surfactants is exceeded and
micelles are formed. Parameters such as temperature and
solvent/solvent mixtures have a major influence on the
CMC.33,34
The CMCs of the synthesised [RMIM]2ijWO4] ILs in
aqueous hydrogen peroxide (50 wt%) were determined by
tensiometry and conductometry at a temperature of 50 °C,
which is used in the catalysis studies. The CMCs of the
[RMIM]2ijWO4] ILs with an alkyl chain length ranging from C4
to C12 are presented in Table 1. Both methods show nearly
identical results which is in accordance to our previous
studies.35 An increase of the alkyl chain length results in a
lower CMC, while imidazolium cations with short alkyl
chains, such as the butyl moiety dissolve in aq. H2O2 without
micelle formation.36 For micellar catalysis in water, the
solubilisation of the organic substrates inside the micelles is
an essential premise. To investigate this, we analysed the
macroscopic structure of the micelles by cryo-TEM before
and after addition of cis-cyclooctene (COE) to a solution of a
representative sample of 2.5 mol% [OMIM]2ijWO4] in aq. H2O2
(50 wt%). Aliquots of the aqueous phase as used in catalysis
studies (vide infra) including catalyst, substrate and aq. H2O2
(50 wt%) were subjected to a TEM grid (copper) at 50 °C and
immediately frozen below −180 °C (see the Experimental
section for details). This procedure enables the investigation
of dynamic processes in liquid state by rapid solidification
and permanent cooling under high vacuum conditions. The
left image in Fig. 2 shows the micelles of 2.5 mol%
[OMIM]2ijWO4] in aq. H2O2 prior to addition of COE.
The surfactant of choice has to meet certain criteria to allow
for micellar epoxidation catalysis in water: it has to be stable
towards oxidative conditions, it must not decompose the
oxidant, and it must be exclusively soluble in water for product
separation without catalyst loss. Classic tensides, such as
cetyltrimethylammonium bromide (CTAB), which is used on
industrial scale, are soluble in organic media to a certain extent
and therefore, they are ineffective from the point of catalyst
separation and reusability. On top of that, the bromide anion
decomposes H2O2 at elevated temperatures (halide oxidation).
Other common surfactants, like octyltrimethylammonium
chloride do not solubilise hydrocarbons in an aqueous solution
of H2O2. The concept of epoxidation in aqueous phase is,
therefore, rather unexplored and just few studies describe the
application of surfactants with molecular/organometallic
epoxidation catalysts in water.13–16 It is well established that
imidazolium-based ionic liquids (IL) are a class of extremely
versatile designer compounds, which allow for simple, task-
specific functionalisation with a virtually limitless variation of
cations and anions.17 One subclass – the surface-active ILs
(SAILs) – can act as surfactants18,19 being able to transfer
hydrophobic organic compounds into water, without themselves
being soluble in the organic phase.20,21 Few groups have
reported on SAILs acting as both phase transfer reagent and
catalyst,22–24 which shows that the concept of such micellar
multitasking catalysts is still in an early stage of development.
In this regard, we reported on the activation of hydrogen
peroxide and olefin epoxidation in water by imidazolium
perrhenate catalysts.25–28 Yet, the H2O2 activation is rather slow
and delicate; it is highly depending on reaction conditions, and
rhenium is a very expensive and rare metal. This has driven us
to expand our micelle catalyst concept to cheaper and more
abundant metals such as tungsten, which is one of the
archetypical oxidation metals with hydrogen peroxide in organic
media.4,29,30
Hence, in this work, we report on the synthesis of
imidazolium-based ILs containing the simplest tungsten-oxo
anion [WO4]2−. We studied the parameters, which control the
micelle formation in aqueous hydrogen peroxide, the nature of
interaction of the tungstate dianion and H2O2, as well as the
uptake of substrate inside the micelles and the catalytic activity
of the system. Furthermore, we investigated the influence of
organophosphonic acid additives as it is reported to increase
the activity of tungstate catalysts.31,32 As the imidazolium cation
is easily functionalised, we anchored the organophosphonic
group to the IL in order to obtain a single-component catalyst,
which operates in water and is easily recyclable.
The shape of the micelles is nearly spherical with an
average diameter of 16 5 nm (ESI,† Fig. S2), with visible
contrast due to the presence of the heavy metal tungsten.
Table 1 CMCs of [RMIM]2ijWO4] ILs in 50 wt% H2O2 at 50 °C with R =
Oct, De, Do
Results and discussion
CMC [mmol L−1] by
tensiometry
CMC [mmol L−1] by
conductometry
Synthesis of ILs and micelle formation in aqueous hydrogen
peroxide
Catalyst
[OMIM]2ijWO4]
[DeMIM]2ijWO4]
[DoMIM]2ijWO4]
31.9
5.4
1.9
31.1
5.1
2.1
The obtained 1-alkyl-3-methylimidazolium tungstate ILs
[RMIM]2ijWO4] (R = butyl (Bu), octyl (Oct), decyl (De), dodecyl
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
The background is vitrified aq. H2O2. Upon addition of COE,
oxide (COO). For kinetic studies, aliquots were taken from
the organic phase after certain time intervals.
the micelles swell up an average diameter of 21
6 nm.
Given the nearly spherical shape of the micelles it can
roughly be estimated that the micelle volume doubles after
COE addition, confirming its uptake in the aqueous phase
(Fig. 2, middle). The uptake of COE identified by micelle
swelling in TEM experiments can be quantified by 1H-NMR
spectroscopy. In the absence of micelles, not even traces of
COE were detectable in the aqueous phase. Upon saturation
of 2.5 mol% (178 mM) [OMIM]2ijWO4] in aq. H2O2 with COE,
(Fig. 2, before: left; after: middle) and the CMC into account
(18% of the [OMIM]+ cations do not take part in micelle
formation but are solely dissolved in the aqueous phase), 2.2
molecules per catalyst moiety are taken up inside the
micelles, corresponding to 4.5% of the total COE amount in
a catalysis reaction. The calculations of the COE uptake from
the TEM images results in a solubilisation of 2.5 COE
molecules per [OMIM]2ijWO4], considering a surface excess
concentration of 2.33 μmol m−2 calculated from the Gibbs
adsorption isotherm (see the ESI† for details on the
calculation). The values from the solubility and the COE
uptake are close; however, the calculation from TEM images
neglects the interaction of the imidazolium octyl moiety with
The dissolution of the tungstate catalyst in hydrogen
peroxide leads to a quasi-homogeneous micellar solution.
The addition of COE creates a biphasic system with an
organic and an aqueous phase. While the upper phase
consists of COE and COO only, the oxidant phase contains
the catalytically active micelles, which take up fractions of
the organic overlayer. Stirring enables an exchange of the
micelles' content with the organic bulk phase, ensuring a
steady equilibrium. The surface activity of the catalyst is
dependent on the length of the alkyl chain substituent of the
imidazolium moiety and is investigated in catalytic
experiments. The results are shown in Table 2 and Fig. 3. As
expected, the catalytic activity correlates to the micelle
formation behaviour, resulting in negligible catalytic activity
for Na+ and [BMIM]+ counterions (Table 2, entries 1 and 2) as
a result of their lack of micelle formation in aqueous media37
and hence, a negligible solubility of COE in the aqueous
phase. Remarkably, the alkyl chain length has a significant
influence on the selectivity, since a system of supramolecular
aggregates protects the epoxide from protons by electrostatic
repulsion on the micellar surface diminishing the formation
of the single side product 1,2-cyclooctane diol. The difference
of the catalytic activity of [RMIM]2ijWO4] ILs with varying alkyl
chain lengths at the cation (R = Oct, De, Do) arises from the
different catalyst concentration involved in micelle
formation. As the CMC is smaller for longer alkyl chain
imidazolium cations, the amount of dissolved surfactant is
COE, which might lead to
a somewhat smaller ratio.
Furthermore, the standard deviation of the micelle diameter
is one quarter of the value, leading to an even higher error by
calculation of the volume involving the micellar radius to the
power of three.
reduced and
a higher fraction is involved in micelle
Catalytic epoxidation of cis-cyclooctene with micellar
[RMIM]2ijWO4] catalysts
formation at the applied catalyst concentration. This results
in a higher catalyst amount forming micelles, showing a
higher catalytic activity (Table 2, entries 2–5). Increasing the
catalyst amount from 2.5 mol% to 5.0 mol% [OMIM]2ijWO4]
results in a nearly twofold catalytic activity.
In accordance with the TEM images, the density of
micelles in the aqueous phase is also increased, explaining
the catalytic activities, which arises from the micelles only
(Fig. 2, right TEM image and Table 2, entries 3 and 6). The
cis-Cyclooctene (COE) was used as a case study substrate for
the biphasic epoxidation using aqueous hydrogen peroxide
(50 wt%) at 50 °C (Scheme 1). The concentration of
imidazolium tungstate catalysts was 2.5 to 5.0 mol% vs. COE
(W : COE = 2.5–5.0 : 100); the COE : H2O2 ratio was kept at a
ratio of 100 : 250 to ensure a full conversion to cyclooctene
Fig. 2 Cutouts of cryo-TEM pictures of solutions of [OMIM]2ijWO4] in H2O2 (50 wt%). Left: 178 mM [OMIM]2ijWO4] in H2O2; middle: 178 mM
[OMIM]2ijWO4] in H2O2 and saturated with COE; right: 357 mM [OMIM]2ijWO4] in H2O2 and saturated with COE. For the full images and the plot of
the mean diameter see the ESI,† Fig. S1 and S2.
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Scheme 1 Catalytic epoxidation of COE with [RMIM]2ijWO4] catalysts
and aq. H2O2 as oxidant.
solubility of COE is increased accordingly to 9 mol% of the
total COE amount, maintaining the ratio 2.2 COE molecules
per catalyst moiety.
For a more comprehensive understanding of the catalytic
system, we investigated the nature of the active species by
183W-NMR spectroscopy. The mechanism of perrhenate-based
epoxidation follows an outer-sphere H-bond activation of
hydrogen peroxide.22 Contrasting this, the tetraoxotungstate
dianion forms
a tetra(μ-peroxo)tungstate complex upon
treatment with H2O2.38 The 183W-NMR spectrum of
[OMIM]2ijWO4] (1 M solution in D2O) recorded at room
temperature exhibits a resonance of −6.02 ppm of the central
W-atom (Fig. 4, top spectrum), which is slightly shifted from
the signal of the reference Na2ijWO4] (0.0 ppm, 2 M in D2O).
Upon addition of 5.0 equiv. H2O2 (50 wt% solution in H2O) a
new resonance appears at −1234 ppm, corresponding to the
formation of the tetraperoxo complex [W(μ-O2)4]2−
(Fig. 4, bottom). This is in good agreement with the reported
shift of Na2ijWIJO2)4] at −1247 ppm.39
Fig. 3 Kinetic time-conversion plot of the catalytic epoxidation of
COE with Na2ijWO4] and [RMIM]2ijWO4] as catalyst, aq. H2O2 (50 wt%)
as oxidant. Conditions: molar ratio IL : COE : H2O2
= 2.5 : 100 : 250,
50 °C.
samples were diluted in 50 wt% aq. H2O2 and heated to 50
°C prior to ionisation to achieve reaction-like conditions.
Pure samples of [OMIM]2ijWO4] show
a
variety of
mononuclear peroxotungstate species (Fig. 5, top). Upon
addition of two equivalents PPA an additional signal at m/z =
389 is detected (Fig. 5, bottom), corresponding to
mononuclear P–O–W adduct without peroxo ligand,
therefore not showing the catalytic active species. However,
the formation of
tetraperoxotungstate [W(O2)4]2− is
suppressed and the formation of polyoxotungstates under the
influence of hydrogen peroxide and the acidic phosphonic
acid can be excluded.
For examining the effect of PPA on the catalytic activity of
micellar [RMIM]2ijWO4] catalysts, we gradually varied the ratio
PPA : [WO4]2− from 1 : 1 to 3 : 1. Kinetic data in Fig. 6 show an
optimum of the catalytic performance of [OMIM]2ijWO4] in
the presence of 5 mol% PPA. Already after 5 min reaction
time the conversion is at ca. 50% and the entire reaction is
completed after 30 min. This renders our catalyst system
highly active and thus comparable to other phase-transfer
catalysts, which are, however, operating on organic phase
and can therefore not be easily recycled as is the case here
a
Influence of organophosphonic acids on the catalytic activity
a
It is known that organophosphonic acids are powerful
additives, which boost the activity of tungstate-based
epoxidation catalysts in organic media, with an in situ
formed phosphonate-diIJperoxo)tungstate complex being the
catalytically active species.31,32 The increased activity is
explained by the higher Lewis acidity of the tungsten center
and in turn increased oxygen transfer ability compared to the
tetraperoxotungstate species.40 As the literature-known
(aminomethyl)phosphonic acid additive is decomposed to
less active phosphate during the epoxidation, we used
phenylphosphonic acid (PPA) as additive, which is stable
towards oxidation. To get a deeper understanding of the
(peroxo)tungstate/phosphonic acid adducts we performed
ESI-MS measurements, as this is a suitable tool to detect
tungstate species and adducts in hydrogen peroxide.41,42 The
a
Table 2 Comparison of COE epoxidation using different tungstate catalysts and 2.5 equiv. of aq. H2O2 (50 wt%) at 50 °C
Entry
Catalyst
Conc.a [mol%]
Conc.a [mmol L−1
]
Micellar conc.b [mmol L−1
]
Conv. [%] (8 h)
Sel. [%] (8 h)
1
2
3
4
5
6
Na2ijWO4]
2.5
2.5
2.5
2.5
2.5
5.0
178
178
178
178
178
357
0
0
146
173
176
325
4
9
48
72
77
81
80
69
94
96
96
99
[BMIM]2ijWO4]
[OMIM]2ijWO4]
[DeMIM]2ijWO4]
[DoMIM]2ijWO4]
[OMIM]2ijWO4]
a
b
Catalyst concentration in the aqueous phase. Concentration of catalyst involved in micelle-formation; respective CMCs (Table 1) subtracted
from deployed catalyst concentration. Na2ijWO4] and [BMIM]2ijWO4] show no micelle formation.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
Fig. 4 183W-NMR spectra of a 1 M solution of [OMIM]2ijWO4] in D2O
(top) and the reaction of [OMIM]2ijWO4] with 5 equiv. of aq. H2O2 (50
wt%) leading to the [W(O2)4]2− complex in D2O (bottom).
Fig. 6 Kinetic time-conversion plot of the catalytic epoxidation of
COE with [OMIM]2ijWO4] as catalyst, aq. H2O2 (50 wt%) as oxidant and
varying concentrations of PPA. Conditions: molar ratio IL : COE : H2O2
= 2.5 : 100 : 250, 50 °C.
(for a comparison with other catalysts, see Table 3 in ref. 32).
A further increase of the PPA concentration results in a loss
of selectivity towards COO (92%, compared to >99% for 2.5
and 5 mol% PPA) due to higher acidity and concomitant
epoxide ring opening. Therefore, all further catalytic
experiments were performed using a PPA : [WO4]2− ratio of 2 :
1. Notably, decomposition of H2O2 was not observed in any
catalysis experiments. Although we cannot fully exclude that
in negligible quantities the oxidant decomposes, we
performed a catalysis run with 100 equiv. COE, 100 equiv.
H2O2, 2.5 equiv. [OMIM]2ijWO4] and 5 equiv. PPA, which
resulted in a COE conversion of 98%, confirming that
decomposition is indeed negligible.
We determined the activation energy by use of the
Arrhenius plot for the epoxidation of COE with 2.5 mol%
[OMIM]2ijWO4] and 5.0 mol% PPA in a temperature range of
30 to 70 °C (ESI,† Fig. S9 and S10). The activation energy of
about 32 kJ mol−1 is significantly low, speaking for a different
mechanism compared to the PPA-free system. Note that
according to tensiometry the CMC does not change upon
addition of 5 mol% PPA. The substrate solubility of COE
inside the PPA-containing micelles is slightly increased to 2.6
Table 3 Epoxidation of various substrates using [OMIM]2ijWO4]. Molar
ratio catalyst : PPA : COE : H2O2 = 2.5:5:100: 250
Conversion
[%] (time)
Selectvity
[%] (time)
Substrate
T [°C]
1-Octene
1-Octene
1-Dodecene
cis-2-Octene
trans-2-Octene
Allyl alcohol
Allyl alcohol with Na2ijWO4]
50
70
70
50
70
50
50
81 (7 h)
95 (4 h)
86 (4 h)
100 (1 h)
92 (1 h)
46 (7 h)
58 (4 h)
70 (4 h)
84 (1 h)
77 (1 h)
80 (30 min)
34 (30 min)
Fig. 5 ESI-MS spectra of [OMIM]2ijWO4] in 50 wt% H2O2 (top) and
after addition of 2 equiv. PPA (bottom).
83 (30 min)
91 (30 min)
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
mol COE per 1 mol IL catalyst, which corresponds to a
slightly higher COE uptake compared to the micelles without
additive. This effect can be explained by the electrostatic
repulsion of the imidazolium head groups: electrolytes, such
as the additive is, can increase the aggregation number of
1-octyl-3-methylimidazolium ILs.43 Conductometry data show
a slightly decreased CMC of [OMIM]2ijWO4] in the presence of
PPA (25.5 vs. 31.1 mmol L−1) but the similar slopes suggest
that the size and shape of the micelles is rather unaffected
by the additive (see the ESI,† Fig. S8).
In the next step, we have studied the influence of the
organic moiety of PPA on the catalytic activity. Electron-
donating
4-methylphenyl
and
electron-withdrawing
4-fluorophenyl substituents should affect the Lewis basicity
of the additive and hence the catalytic activity. However, the
kinetic plots using 5 mol% of PPA, p-MePPA and p-FPPA do
not show any influence of these substituents. Using
ethylphosphonic acid instead of PPA shows a slower reaction
rate, but still improves the reaction compared to the additive-
free approach (Fig. 7). PPA, with its pKa of 1.83 and 7.07 in
water44 is slightly more acidic in nature than
ethylphosphonic acid (pKa 2.43 and 8.05 in water)45 and leads
Fig. 8 Epoxidation of COE with [RMIM]2ijWO4] catalysts, aq. H2O2 (50
wt%) and PPA (molar ratio IL : PPA : COE : H2O2 of 2.5 : 5 : 100 : 250),
50 °C.
Further experiments were carried out for
a deeper
to
a
more powerful additive, as more molecules are
understanding on the correlation of the catalyst loading and
the CMC. As shown in Fig. 9, the reaction rate correlates
linearly with the catalyst loading. However, the intercept of
the trends does not match with the point of origin of
coordinates, but with cIJcatalyst) = 0.037 mol L−1, which is
close to the CMC for [OMIM]2ijWO4] (0.032 mol L−1)
determined by tensiometry under reaction conditions. This
experiment very clearly underlines the necessity of a micellar
environment for a catalytic activity as well as the knowledge
on the CMC values for a predictability of the catalytic
performance.
deprotonated in solution.
This leads to a higher amount of anions which are able to
coordinate to the tungstate. Furthermore, we investigated the
influence of the alkyl chain length using 2.5 mol% catalyst
[RMIM]2ijWO4] and 5.0 mol% PPA (R = butyl, octyl, decyl,
dodecyl). The kinetic curves in Fig. 8 show that the longer
the aliphatic substituent of the micelle-forming imidazolium
cation the more active the catalyst is. Hence, the differences
between the CMC of C10 and C12 alkyl chains compared to C8
and C10 are smaller, the changes of the catalytic activity are
less pronounced.
Supported by these findings, we claim that only the
micelle-bound tungstate bears catalytic activity, since
Fig. 7 Kinetic time-conversion plots of the catalytic epoxidation of
COE with [OMIM]2ijWO4] as catalyst, aq. H2O2 (50 wt%) as oxidant and
varying additives. Reaction conditions: molar ratio IL : additive : COE :
H2O2 = 2.5 : 5 : 100 : 250, 50 °C.
Fig.
9 Variation of [OMIM]2ijWO4]-concentration and linear
extrapolation. Molar ratios IL : PPA
50 °C.
= 1 : 2, COE : H2O2 = 100 : 250,
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
dissolved non-aggregated imidazolium tungstate shows no
activity. In accordance with the Stern theory, besides the
dissolved non-micellar ILs, the tungstate ions within the
micelles are located in the micellar interface due to
electrostatic interactions. We conclude that the reactivity of
the tungstate ion is achieved in the interface of a micellar
environment exclusively, as both phases are in proximity with
the anion.
The PPA additive is expected to be homogeneously
dissolved in the aqueous phase. Therefore, we have covalently
linked an organophosphonic acid moiety to the imidazolium
cation and therefore the micellar surface, expecting a higher
reaction rate due to the proximity of tungstate and the
additive. In such a system the P–O–W adduct is forced to the
micellar interface. Therefore, we anticipate an increase in
reaction rate for the zwitterionic compounds
1 and 2
(Fig. 10). Compounds 1 and 2 were converted to tungstate-
containing ILs by addition of an equimolar amount of
tungstic acid H2WO4, enabling the in situ formation of
protonated phosphonic acid. It has to be noted that due to
the molar ratio [IM]+ : [WO4]2− of 1 : 1, twice the tungstate
amount is used for catalysis reactions compared to the
additive-free [RMIM]2ijWO4] catalysis tests.
Fig. 11 Kinetic plots for the epoxidation of COE. Conditions: molar
ratio [IM]+ : COE : H2O2 = 5 : 100 : 250; 50 °C.
moiety to the cation for an increased interaction with the
tungstate counterion. The catalytic activity is increased by a
factor of 3–4 compared to the 2.5 mol% [OMIM]2ijWO4]
system with 5 mol% EPA and shows even a better catalytic
performance than the [OMIM]2ijWO4]/PPA system, although
an ethyl phosphonate is inferior to the aryl derivative.
Mechanistic investigations by ESI-MS give hint to the
Surprisingly, using this approach a rate enhancement
expected for structure zwitterion 1 did not take place, but is
as exactly as active as the additive-free run using solely
[OMIM]2ijWO4] (Fig. 11). From this, we conclude that the
benzyl-bridged phosphonate has no co-catalytic effect owing
to the structural rigidity of the phenylphosphonate moiety
and the imidazolium ring, creating an unsuitable
constitution of the tungstate ion on the micellar surface. It is
known for structurally similar benzoate imidazolium SAILs
that these anions are preferably located in the alkyl chain
layer of the micelle rather than in the Stern layer.46 The
substitution pattern of the imidazolium ring forces the
phenylphosphonate moiety of the zwitterion 1 to point
towards the micellar surface, which is against the nature of
the additive, explaining a diminished catalytic activity. Using
ESI-MS we tried to identify the expected complex, but we were
not able to detect any adduct formation besides a zwitterion
1 dimer with 699 m/z (ESI,† Fig. S11). Electronic influences of
the para-linkage can be excluded as we have shown that the
phenyl-ring functionalisation has no influence on the
catalytic activity (Fig. 7). To increase the constitutional
flexibility between the phosphonate and the imidazolium
formation of
a [zwitterion 2]2/tungstate complex with
additional oxygen atoms, likely due to peroxo-groups, at 838
m/z with the corresponding tungstate isotopic pattern (ESI,†
Fig. S11).
Catalyst recycling
The use of environmentally benign H2O2 as oxidant for
oxidation reactions is an important aspect for green catalysis.
Nevertheless, the recyclability of the [OMIM]2ijWO4] IL catalyst
together with the additive PPA is an essential premise, as a
non-recyclable catalyst would diminish the advantages of the
use of ‘green’ H2O2. Therefore, we investigated the
recyclability of the catalytic systems for 10 runs. The reaction
was interrupted after 30 min each upon full conversion (cf.
Fig. 11, black graph). The organic phase was decanted and
the aqueous residue washed two times with n-hexane. In
order to separate the IL catalyst and additive from the
aqueous phase, residual H2O2 was decomposed using a
platinum wire prior to evaporation of all volatiles in vacuo.
Next, the IL/PPA mixture is reused as obtained by addition of
fresh H2O2 solution and COE. This process was repeated 9
times and no loss of activity is observed for the
[OMIM]2ijWO4]/PPA system (Fig. 12), speaking for neither a
catalyst nor PPA loss or decomposition. Recycling
experiments using zwitterion 2 as catalyst show a steady
decrease in activity after three runs (Fig. 12). A 31P-NMR
investigation of this catalyst system shows no decomposition
after run 1, but 10% decomposition of the catalyst to
groups, we introduced
corresponding ethylene-bridged zwitterion 2. The reaction
rate is significantly increased, presumably due to
a
C2-linker resulting in the
a
synergistic effect of a covalent linkage of the phosphonic acid
Fig. 10 Structures of zwitterions 1 and 2.
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
rate is rather fast and can be explained by the absence of
mass-transport limitations. Using Na2ijWO4] instead of
[OMIM]2ijWO4] should therefore lead to similar results. The
conversion after 30 min is slightly increased in comparison
to the micellar system, but the selectivity is significantly
higher for the micellar system. This observation correlates to
the previously shown effect of micelles: the epoxide is
stabilised towards ring opening reactions in the presence of
micelles. As diol formation from epoxides under acidic
conditions involve the protonation of the oxirane moiety, it
can be concluded that the positively charged micelle surface
shields oxonium ions from getting in proximity with the
epoxide inside the micelles.
Experimental
All synthesis procedures and catalyst characterisations can be
found in the ESI.†
Fig. 12 Recycling of 2.5 mol% [OMIM]2ijWO4] with 5.0 mol% PPA (dark
grey) and 5.0 mol% zwitterion 2 with 5.0 mol% H2WO4 (light grey) over
10 runs. Selectivities >99%.
CMC determination via tensiometry analysis
The surface tension was measured with a K11 tensiometer
(Krüss GmbH, Germany) according to the Wilhelmy plate
method with an accuracy of 0.01
suspended plate touches the liquid, wets the plate and a
force (F) can be measured. The surface tension is calculated
according to the following equation:
phosphate after 10 runs. In practice, the catalyst can be
recovered easily by decantation of the organic layer only and
is in this study further isolated to enable a true comparison
of the catalytic activity over several runs.
N A vertically
m−2.
F
σ ¼
Substrate variation
L × cosθ
So far, for all catalysis reactions COE was used as a model
substrate. The high stability towards ring opening reactions
of the corresponding epoxide enable a detailed study of the
mechanism, as no follow-up product can interfere with the
catalyst and micelle formation. In the extension of this study,
we tested the epoxidation towards further epoxides and the
applicability on other structural motifs (Table 3). Terminal
olefins such as 1-octene and 1-dodecene are successfully
oxidised to high conversion and moderate selectivities with
the presented [OMIM]2ijWO4]/PPA system. The reaction rates
are reduced compared to COE, which is caused by the
decreased reactivity of terminal olefins with lower electron
density. The selectivity is somewhat lower due to the ring
opening to the corresponding 1,2-diols, which is facilitated
for terminal epoxides. Nevertheless, an increase in
temperature for the epoxidation of 1-octene results in a
higher selectivity as the reaction time is shorter. Internal
olefins are epoxidised faster than terminal olefins. cis-2-
Octene has the same double bond configuration as COE but
shows a decreased selectivity due to ring opening. trans-2-
Octene requires a higher temperature for a comparable
reactivity, as the trans double bond configuration is less
reactive. Nevertheless, no isomerisation is observed for both
2-octenes, which supports the postulated mechanism and
excludes radical processes. Allyl alcohol was used to
investigate the epoxidation of olefins in a one-phasic micellar
environment, as both, the substrate and the product glycidol
are completely soluble in aq. hydrogen peroxide. The reaction
L is the length of the plate and θ the contact angle between
liquid and solid surface of the plate. Due to the
decomposition of hydrogen peroxide on platinum surfaces,
the usually used platinum plates were substituted with glass
plates. The data were drawn in a figure with surface tension
over the logarithmic concentration of the surfactant to see a
sharp break point at the CMC.
CMC determination via conductometry
Conductivity measurements were conducted with a Metrohm
712 conductometer with a modified, homemade conductivity
probe as the decomposition of H2O2 destroyed the
commercially available conductivity probes. Tin (99.99%
purity, supplied by EvoChem) was used as electrode material.
A PT100 thermo couple surrounded by a glass tube was used
to monitor the temperature of the sample. The IL
concentration, at which a stepwise change of the slope of the
conductivity versus the IL concentration occurred, was set as
the CMC.
Catalysis studies
In a standard catalytic experiment a 20 mL screw neck vial
equipped with an air-tight cap was loaded with 2.5 mol% of
the imidazolium tungstate catalyst (and 5.0 mol% of
phosphonic acid additive if required). Subsequently, 50 wt%
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
aqueous hydrogen peroxide (1.4 mL, 25 mmol, 2.5 equiv.)
and the internal standard mesitylene (69 μL, 5.0 mol%) were
added and heated to reaction temperature using a tight-
fitting temperature-adjustable metal vial holder. After
complete dissolution of the components, the organic
substrate (10 mmol, 1.0 equiv.) was added and the biphasic
reaction mixture stirred at 500 min−1 using a 20 mm
magnetic stirring bar. For kinetic data, samples were taken
after certain time intervals from the organic top layer and
solubilised large amounts of substrates in the aqueous phase,
but also allows for further functionalisation. Short-chain
covalent linking of the phosphonate group to the surfactants
head group leads to a further increase of the reaction rate,
due to the high additive concentration on the micellar
surface.
The micellar tungstate catalyst operating in aqueous
media is somewhat lower in activity compared to tungstate
catalysed reactions in the organic phase,40 but it enables a
facile product separation without the need of further
purification. The catalyst remains in the aqueous phase and
can thus easily be recovered or maintained in solution for a
new reaction upon addition of H2O2 and substrate. The
recycling studies demonstrate that such micellar catalysts
have the perspective to be used in a continuous reactor
process, which is so far still rare for molecular epoxidation
catalysts.
The multitasking character of such surface-active ILs – (i)
micellisation and substrate solubilisation in water and (ii)
catalytic activity combined with simple optional task-specific
functionalisation – highlights the potential of these systems
well beyond the herein presented exemplary epoxidation of
olefins. In this regard, especially ionic liquids exhibit a huge
playground for a multitude of reactions, given their nearly
unlimited potential in terms of permutations of cations and
anions, as well as functionalisation towards tailor-made
multitask, single component catalysts.
1
investigated by H-NMR.
Catalyst recycling
After the first run, the organic overlayer was decanted and
the residual product extracted with 3 × 10 mL n-hexane.
Then, H2O2 was decomposed using a platinum wire for 30
min at 50 °C prior to evaporation of water in vacuo.
Afterwards, the residue is freshly charged with H2O2 and COE
according to the procedure descripted above.
Cryo-transmission electron microscopy
Cryo-TEM measurements were carried out with a Jeol 1400
plus microscope at 120 kV equipped with a Ruby CCD
detector. Blotting of the [OMIM]2ijWO4]/H2O2 and
[OMIM]2ijWO4]/COE/H2O2 was performed with a FEI Vitrobot
MK3 at 50 °C and 80% humidity on Quantifoil 1.2/1.3 grids
(Cu 300 mesh). For this, a solution of the sample (3 μL) was
coated onto a TEM copper grid, and the grid was blotted with
two pieces of filter paper for approximately 4 s to form a
solution thin film. Then, the grid was quickly plunged into a
reservoir of liquid ethane (cooled by liquid nitrogen) and
kept in liquid nitrogen until the transfer of the grid to a
cryogenic sample holder.
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgements
The authors thank the Deutsche Forschungsgemeinschaft
(DFG) for financial support of this work in the frame of the
DFG-project ‘Hydrophobic Ionic Liquids as Micellar Catalysts
for the Epoxidation of Olefins’ (project no. 326649877, grant
no. Co 1543/1-1 and Je 257/24-1). The authors further thank
Dr. Carsten Peters (Electron Microscopy Group at the Faculty
of Chemistry of the Technical University of Munich) for the
preparation and recording of cryo-TEM samples.
Electrospray ionisation mass spectrometry
ESI-MS measurements were performed on a Varian 500 LC/
MS ion trap spectrometer using 50 wt% aqueous H2O2 as
carrier at 20 μL min−1 with ionisation at 50 °C.
NMR spectroscopy
1H-,13C- (1024 scans, 101 MHz) and 31P- (32 scans, 161 MHz)
NMR spectra were recorded on a Bruker Avance III 400 MHz
(US). 183W-NMR spectra (16.6 MHz, 16 384 scans) were References
recorded on a Bruker Avance 400 MHz.
1 S. Huber, M. Cokoja and F. E. Kühn, J. Organomet. Chem.,
2014, 751, 25–32.
2 G. Grigoropoulou, J. H. Clark and J. A. Elings, Green Chem.,
2003, 5, 1–7.
3 S. A. Hauser, M. Cokoja and F. E. Kühn, Catal. Sci. Technol.,
2013, 3, 552–561.
4 C. Venturello, E. Alneri and M. Ricci, J. Org. Chem., 1983, 48,
3831–3833.
5 J.-M. Brégeault, M. Vennat, L. Salles, J.-Y. Piquemal, Y.
Mahha, E. Briot, P. C. Bakala, A. Atlamsani and R.
Thouvenot, J. Mol. Catal. A: Chem., 2006, 250,
177–189.
Conclusions
We present the micellar epoxidation of olefins with surface-
active ILs. This is the first report on tungsten-based
epoxidation catalysts operating in water. The organic uptake
of the water-insoluble substrate by micelles is shown by cryo-
TEM under reaction-like conditions. The addition of
phenylphosphonic acid to the imidazolium tungstate
catalysts results in an enormous increase of the reaction rate,
caused by formation of
tungstate species. The imidazolium structural motif not only
a highly active phosphonate–
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
6 N. Mizuno, K. Yamaguchi and K. Kamata, Coord. Chem. Rev.,
2005, 249, 1944–1956.
7 S.-S. Wang and G.-Y. Yang, Chem. Rev., 2015, 115,
4893–4962.
8 Y. Zhou, Z. Guo, W. Hou, Q. Wang and J. Wang, Catal. Sci.
Technol., 2015, 5, 4324–4335.
26 M. Cokoja, R. M. Reich, M. E. Wilhelm, M. Kaposi, J.
Schäffer, D. S. Morris, C. J. Münchmeyer, M. H. Anthofer,
I. I. E. Markovits, F. E. Kühn, W. A. Herrmann, A. Jess and
J. B. Love, ChemSusChem, 2016, 9, 1773–1776.
27 J. Schäffer, M. Alber, W. Korth, M. Cokoja and A. Jess,
ChemistrySelect, 2017, 2, 11891–11898.
9 Y. Ishii, K. Yamawaki, T. Ura, H. Yamada, T. Yoshida and M.
Ogawa, J. Org. Chem., 1988, 53, 3587–3593.
10 E. V. Dehmlow, Angew. Chem., Int. Ed. Engl., 1977, 16,
493–505.
28 J. Schäffer, B. Zehner, W. Korth, M. Cokoja and A. Jess,
Chem. Eng. Technol., 2019, 42, 232–240.
29 Y. Ishii, K. Yamawaki, T. Yoshida, T. Ura and M. Ogawa,
J. Org. Chem., 1987, 52, 1868–1870.
11 M. Mąkosza and M. Fedoryński, Adv. Catal., 1987, 35,
375–422.
30 K. Sato, M. Aoki and R. Noyori, Science, 1998, 281,
1646–1647.
12 G. La Sorella, G. Strukul and A. Scarso, Green Chem.,
2015, 17, 644–683.
31 M. Quenard, V. Bonmarin and G. Gelbard, Tetrahedron Lett.,
1987, 28, 2237–2238.
13 D. Monti, P. Tagliatesta, G. Mancini and T. Boschi, Angew.
Chem., Int. Ed., 1998, 37, 1131–1133.
14 J. H. M. Heijnen, V. G. de Bruijn, L. J. P. van den
Broeke and J. T. F. Keurentjes, Chem. Eng. Process.,
2003, 42, 223–230.
32 K. Sato, M. Aoki, M. Ogawa, T. Hashimoto, D. Panyella and
R. Noyori, Bull. Chem. Soc. Jpn., 1997, 70, 905–915.
33 T. Inoue, H. Ebina, B. Dong and L. Zheng, J. Colloid Interface
Sci., 2007, 314, 236–241.
34 M. Tariq, M. G. Freire, B. Saramago, J. A. P. Coutinho,
J. N. C. Lopes and L. P. N. Rebelo, Chem. Soc. Rev., 2012, 41,
829–868.
15 Y. Zhang, R. Tan, G. Zhao, X. Luo and D. Yin, Catal. Sci.
Technol., 2016, 6, 488–496.
16 F. P. Ballistreri, C. M. A. Gangemi, A. Pappalardo, G. A.
Tomaselli, R. M. Toscano and G. Trusso Sfrazzetto, Int. J.
Mol. Sci., 2016, 17, 1112.
17 Ionic Liquids, ed. B. Kirchner, Springer, Berlin Heidelberg,
2010.
18 C. Jungnickel, J. Łuczak, J. Ranke, J. F. Fernández, A. Müller
and J. Thöming, Colloids Surf., A, 2008, 316, 278–284.
19 J. Łuczak, J. Hupka, J. Thöming and C. Jungnickel, Colloids
Surf., A, 2008, 329, 125–133.
20 J. Łuczak, C. Jungnickel, M. Markiewicz and J. Hupka,
J. Phys. Chem. B, 2013, 117, 5653–5658.
35 B. Zehner, F. Schmidt, W. Korth, M. Cokoja and A. Jess,
Langmuir, 2019, 35, 16297–16303.
36 M. Blesic, M. H. Marques, N. V. Plechkova, K. R. Seddon,
L. P. N. Rebelo and A. Lopes, Green Chem., 2007, 9, 481.
37 J. Wang, H. Wang, S. Zhang, H. Zhang and Y. Zhao, J. Phys.
Chem. B, 2007, 111, 6181–6188.
38 M. Grzywa, W. Łasocha and D. Rutkowska-Żbik, J. Solid State
Chem., 2009, 182, 973–982.
39 V. Nardello, J. Marko, G. Vermeersch and J. M. Aubry, Inorg.
Chem., 1998, 37, 5418–5423.
40 R. Noyori, M. Aoki and K. Sato, Chem. Commun., 2003, 1977.
41 M. J. Deery, O. W. Howarth and K. R. Jennings, J. Chem. Soc.,
Dalton Trans., 1997, 4783–4788.
21 S. Dorbritz, W. Ruth and U. Kragl, Adv. Synth. Catal.,
2005, 347, 1273–1279.
22 T. Dwars, E. Paetzold and G. Oehme, Angew. Chem., Int. Ed.,
2005, 44, 7174–7199.
23 H. Morawetz, Adv. Catal., 1969, 20, 341–371.
24 A. Cognigni, P. Gaertner, R. Zirbs, H. Peterlik, K. Prochazka,
C. Schröder and K. Bica, Phys. Chem. Chem. Phys., 2016, 18,
13375–13384.
42 P. U. Maheswari, P. de Hoog, R. Hage, P. Gamez and J.
Reedijk, Adv. Synth. Catal., 2005, 347, 1759–1764.
43 N. M. Vaghela, N. V. Sastry and V. K. Aswal, Colloids Surf., A,
2011, 373, 101–109.
44 H. H. Jaffé, L. D. Freedman and G. O. Doak, J. Am. Chem.
Soc., 1953, 75, 2209–2211.
25 I. I. E. Markovits, W. A. Eger, S. Yue, M. Cokoja, C. J.
Münchmeyer, B. Zhang, M.-D. Zhou, A. Genest, J. Mink, S.-L.
Zang, N. Rösch and F. E. Kühn, Chem. – Eur. J., 2013, 19,
5972–5979.
45 P. C. Crofts and G. M. Kosolapoff, J. Am. Chem. Soc.,
1953, 75, 3379–3383.
46 O. Singh, P. Singla, V. K. Aswal and R. K. Mahajan, Langmuir,
2019, 35, 14586–14595.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020