Three- and two-site heteropolyoxotungstate anions as catalysts for the epoxidation of allylic alcohols by H2O2 under biphasic conditions: Reactivity and kinetic studies of the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7?, [Co3(OH2)6(A-PW9O34)2]12?, and [M4(OH2)2(B-PW9O34)2]10? anions, where M?=?Mn(II), Co(II), Ni(II), Cu(II) and Zn(II)

Article abstract of DOI:10.1016/j.ica.2019.119178

The trimetallic phosphopolyoxotungstate anions [Ni₃(OH₂)₃(B-PW₉O₃₄){WO₅(H₂O)}]^7? and [Co₃(OH₂)₆(A-PW₉O₃₄)₂]^12? have been studied as epoxidation catalysts for oxygen transfer from 30% H₂O₂ to a range of allylic alcohols under biphasic conditions (1,2-dichloroethane/H₂O) at 15 °C. The reaction mechanism involves coordination of an allylic alcohol at an M(II) site in each case, prior to transfer of a peroxy oxygen from an adjacent W(O₂) site. The latter is formed from a terminal W = O unit by reaction with H₂O₂. Evidence of W(O₂) formation was obtained through IR studies. The W(O₂) group forms the epoxide by transfer of an oxygen atom to the C[dbnd]C bond of the coordinated allylic alcohol. Kinetic studies using 3-methyl-2-buten-1-ol as the allylic alcohol substrate have been modelled with all three metal sites catalytically active. The reaction involves an autocatalysis mechanism involving an induction period, which can be rationalised by proposing not only coordination of the allylic alcohol to M(II), but also the product hydroxy epoxide, both through their –OH groups. The autocatalysis is generated by formation of the W(O₂) group adjacent to a coordinated hydroxy epoxide, which competes with coordination of allylic alcohol. The mechanism requires some twenty-one steps involving just the generic steps listed above, with all three metal sites catalytically active. Temperature-dependent kinetic studies and subsequent Eyring analyses have shown that the Co(II)-containing catalyst is the most active of the two. Analogous studies of the epoxidation of 3-methyl-2-buten-1-ol by the two-site [M₄(OH₂)₂(B-PW₉O₃₄)₂]^10? ions as catalysts, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), at 15 °C gave an order of reactivity of Cu(II) > Ni(II) > Zn(II), Co(II), Mn(II), which mostly mimics the natural order of stability constants (the Irving-Williams series), suggesting that the formation of the allylic alcohol complexes play a dominant role in this series of related complex anions, with greater replacement of water by allylic alcohol leading to greater reactivity.

Full text of DOI:10.1016/j.ica.2019.119178

InorganicaChimicaActa499(2020)119178
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Research paper
Three- and two-site heteropolyoxotungstate anions as catalysts for the
epoxidation of allylic alcohols by H2O2 under biphasic conditions: Reactivity
and kinetic studies of the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−,
[Co3(OH2)6(A-PW9O34)2]12−, and [M4(OH2)2(B-PW9O34)2]10− anions,
where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II)
⁎
Paulus Hengky Abrama,b, Robert C. Burnsa, , Lichun Lia,1
a Discipline of Chemistry, School of Environmental and Life Sciences, The University of Newcastle, Callaghan, New South Wales 2308, Australia
b Department of Chemistry, Faculty of Education, University of Tadulako, Jl. Sukarno Hata Km. 9, Palu, Indonesia
A R T I C L E I N F O
A B S T R A C T
Keywords:
The trimetallic phosphopolyoxotungstate anions [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-
PW9O34)2]12− have been studied as epoxidation catalysts for oxygen transfer from 30% H2O2 to a range of allylic
alcohols under biphasic conditions (1,2-dichloroethane/H2O) at 15 °C. The reaction mechanism involves co-
ordination of an allylic alcohol at an M(II) site in each case, prior to transfer of a peroxy oxygen from an adjacent
W(O2) site. The latter is formed from a terminal W = O unit by reaction with H2O2. Evidence of W(O2) formation
was obtained through IR studies. The W(O2) group forms the epoxide by transfer of an oxygen atom to the C]C
bond of the coordinated allylic alcohol. Kinetic studies using 3-methyl-2-buten-1-ol as the allylic alcohol sub-
strate have been modelled with all three metal sites catalytically active. The reaction involves an autocatalysis
mechanism involving an induction period, which can be rationalised by proposing not only coordination of the
allylic alcohol to M(II), but also the product hydroxy epoxide, both through their –OH groups. The autocatalysis
is generated by formation of the W(O2) group adjacent to a coordinated hydroxy epoxide, which competes with
coordination of allylic alcohol. The mechanism requires some twenty-one steps involving just the generic steps
listed above, with all three metal sites catalytically active. Temperature-dependent kinetic studies and sub-
sequent Eyring analyses have shown that the Co(II)-containing catalyst is the most active of the two. Analogous
studies of the epoxidation of 3-methyl-2-buten-1-ol by the two-site [M4(OH2)2(B-PW9O34)2]10− ions as catalysts,
where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), at 15 °C gave an order of reactivity of Cu(II) > Ni(II) > Zn
(II), Co(II), Mn(II), which mostly mimics the natural order of stability constants (the Irving-Williams series),
suggesting that the formation of the allylic alcohol complexes play a dominant role in this series of related
complex anions, with greater replacement of water by allylic alcohol leading to greater reactivity.
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
[Co3(OH2)6(A-PW9O34)2]12−
[MI4I(OH2)2(B-PW9O34)2]10− (MII = Mn, Co, Ni,
Cu, Zn)
Epoxidation
Allylic alcohols
Autocatalysis
1. Introduction
itself, H2O2 is the best as it provides 47% active oxygen while the re-
duction product is the environmentally friendly H2O [5]. However,
Oxidation of organic substrates such as alkanes, alkenes, as well as
both primary and secondary alcohols has been studied extensively using
both isopoly- and heteropoly-oxometalates as catalysts under homo-
geneous and biphasic reaction conditions with a variety of oxidants.
These oxidants include species such as molecular oxygen, H2O2, iodo-
sylarenes, alkyl peroxides, t-butyl hydroperoxide, periodate, perox-
omonosulfate (HSO5−), etc. [1–5]. Of all these oxidants, other than O2
both O2 and H2O2 generally demonstrate complex oxidation chemistry
with little control over the reactions. A comprehensive review of earlier
work using the above oxidants with transition metal-substituted poly-
oxometalates has been given by Hill and co-workers [1,5]. Most reac-
tions that involved O2 as the oxidant examined alkanes and alkenes and
were carried out at higher temperatures (room temperature to 150 °C),
and generally resulted in several reaction products. Reactions that
⁎ Corresponding author.
E-mail addresses: Robert.Burns@newcastle.edu.au (R.C. Burns), lichunli@zjut.edu.cn (L. Li).
1 Present address: College of Chemical Engineering, Zhejiang University of Technology, 18 Chaowang Road, Hangzhou 310014, PR China.
https://doi.org/10.1016/j.ica.2019.119178
Received 12 July 2019; Received in revised form 27 September 2019; Accepted 27 September 2019
Availableonline28September2019
0020-1693/©2019ElsevierB.V.Allrightsreserved.

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
involved aqueous H2O2 were performed at about room temperature, but
also generated multiple products. Several structural types of hetero-
polyoxometalates are considered to be robust, oxidation-resistant in-
organic metalloporphyrin analogues under the conditions employed for
these reactions [1–5].
The most successful heteropolyoxometalate catalysts are the sand-
wich-type species, such as the [WZn{M(OH2)}2(B-ZnW9O34)2]m− ions,
where M = Mn(II), Co(II), Ni(II), Cu(II), Zn(II), Pd(II) and Pt(II), with
m = 12; Fe(III) and Rh(III), with m = 10 and Ru(III), with m = 11
(there is a coordinated OH− in place of an H2O molecule in this anion)
[6–8]. These species have been used to epoxidise alkenes and/or pri-
mary and secondary allylic alcohols under biphasic conditions using
aqueous H2O2 at about room temperature [7–14]. Other hetero-
polyoxotungstate anions that have functioned as successful catalysts for
the epoxidation of alkenes or allylic alcohols using H2O2 under biphasic
conditions include the mono-substituted Dawson ions [α2-P2W17O61
M
(L)]n−, where M = Co(II), Ni(II), Zn(II), Pd(II), Cr(III), Mn(III), Fe(III)
and Ir(IV) (with L = Br− or H2O) [15], and also the Ni(II)-substituted
9−
quasi-Dawson-type fluoro-oxotungstate [Ni(OH2)H2F6NaW17O55
]
[16]. Related sandwich-type species that have been studied for alkene
oxidation include the [(FeIII)4(OH2)2(B-PW9O34)2]6− and [(FeIII
)
4
(OH2)2(P2W15O56)2]12− ions [17,18], the [(FeII)4(B-PW9O34)2]10− ion
[19], and the three Mn(II)-containing [{MnII(OH2)3}2(WO2)2
(BiW9O33)2]10−, [{MnII(OH2)3}2(SbW9O33)2]12− and [{MnII(OH2)3}2
{MnII(OH2)2}2(TeW9O33)2]18− ions [20]. For allylic alcohols, epox-
idation is highly important as epoxides may be used as raw materials for
epoxy resins as well as for the synthesis of biologically important mo-
lecules, as epoxidation leads to the formation of chiral centers [21–24].
Transition metal sandwich compounds have also been used for other
oxidations. The spontaneous assembly of [WZn{Zn(OH2)}2(B-
ZnW9O34)2]12− from Na2WO4 and Zn(NO3)2 has been reported, where it
was used for the in situ oxidation of ammonia with H2O2 to hydro-
xylamine, which in turn reacted with ketones and aldehydes to yield
oximes [25]. Other organic reactions catalysed by heteropolyoxometalates
have been reviewed by Qiao and Hou [26], and even the oxidative de-
sulfurization of fuel oils can be accomplished by heteropolyoxometalates
among other environmental applications [27,28].
Fig. 1. A representation of the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− anion.
Colours: pink, Ni; green, W; orange, P; red, O; white, H. Note that the two H
atoms of the H2O in the {WO5(H2O)} unit are not included in the diagram and
are assumed to be disordered around the three terminal O atoms. (For inter-
pretation of the references to colour in this figure legend, the reader is referred
to the web version of this article.)
the compound originally described by Knoth et al. as K12[Co3(OH2)3(A-
PW9O34)2].10H2O; see also below for further description of this structure)
[29–31]. Representations of these two anions are shown in Figs. 1 and 2,
respectively. The increase in the number of sites may provide more active
anions for epoxidation reactions of allylic alcohols under biphasic condi-
tions. Other examples of interest are the related two-site [M4(OH2)2(B-
PW9O34)2]10− ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II)
[30,32–35]. A representation of a two-site [M4(OH2)2(B-PW9O34)2]10− ion
is shown in Fig. 3. In these anions the heteropolyoxotungstate “sandwich”
Recently, we have studied the comparative catalytic activities of the
transition metal-substituted [WZn{M(OH2)}2(B-ZnW9O34)2]12− sandwich
anions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), for the epox-
idation of a selected series of allylic alcohols and related species using
aqueous H2O2 as the source of oxygen under biphasic conditions [14]. For
these studies we chose aqueous H2O2 as the oxidant based on our previous
studies using transition metal-substituted Dawson anions as catalysts as the
reactions were controllable and resulted in few side products [15]. Mole-
cular oxygen was unreactive under the chosen experimental conditions. The
reactions were performed under experimental conditions whereby the
substituted metal ions, M(II), in the waist of the anions were able to de-
monstrate their role in the epoxidation reaction and established the order
for these labile first row transition metal ions of Zn(II) > Mn(II) ~ Cu
(II) > Co(II) > Ni(II). This order was related to the rate of exchange of the
aqua groups on the substituted M(II) ions, and implied the coordination of
the allylic alcohol at an M(II) site as part of the reaction mechanism. Other
steps included formation of a W(O2) site adjacent to the coordinated allylic
alcohol and subsequent peroxo oxygen transfer to the coordinated allylic
alcohol. A kinetic study of the epoxidation of a selected allylic alcohol, 3-
methy-2-buten-1-ol, by aqueous H2O2 was also studied under biphasic
conditions using the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion as the cata-
lyst. This study showed that the epoxidation followed an autocatalysis
mechanism, which involved not only coordination of the allylic alcohol, but
also the product hydroxy epoxide to M(II), both through their –OH groups.
The [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion has two Mn(II) sites where
coordination of allylic alcohol can occur, but there are examples of transi-
tion metal-substituted heteropolyoxotungstate anions that have three M(II)
sites, such as the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6
(A-PW9O34)2]12− ions (note that the latter is a reformulation of the anion in
12−
9−
component [B-ZnW9O34
]
has been replaced by a [B-PW9O34
]
ion,
lowering the overall charge of the entire catalyst anion (i.e. from −12 to
−10), as well as introducing structural changes because of the change from
Zn(II) to P(V). None of these two- or three-site anions have been used
previously as oxygen transfer catalysts for the epoxidation of alkenes or
allylic alcohols by H2O2 under biphasic conditions.
The present studies examine the relative catalytic activities of the
three-site [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-
PW9O34)2]12− ions and compares them with that of the previously stu-
died two-site [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion in a series of
epoxidations of selected primary allylic alcohols and related compounds
using aqueous H2O2 as the oxidant, as well as kinetic studies of these
three-site anions in the epoxidation of 3-methyl-2-buten-1-ol by H2O2
under biphasic conditions, i.e. 1,2-dichloroethane (1,2-dce)/aqueous
H2O2. Studies are also reported on the [M4(OH2)2(B-PW9O34)2]10− series
of ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), as epoxida-
tion catalysts for allylic alcohols under identical reaction conditions.
2. Experimental
2.1. General
2.1.1. Materials and equipment
All materials used were of analytical or reagent grade and were
2

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
Fig. 3. A representation of an [M4(OH2)2(B-PW9O34)2]10− anion. Colours: pink,
M; green, W; orange, P; red, O; white, H. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this
article.)
Fig. 2. A representation of the [Co3(OH2)6(B-PW9O34)2]12− anion. Colours:
pink, Co; green, W; orange, P; red, O; white, H. (For interpretation of the re-
ferences to colour in this figure legend, the reader is referred to the web version
of this article.)
Diamond TG/DTA instrument. Sample masses of 10–20 mg were rou-
tinely used, with a heating rate of 10 °C/min and were performed in a
flowing air atmosphere (20 mL/min).
2.1.2. Synthesis and characterization of the compounds K6Na[Ni3(OH2)3(B-
PW9O34){WO5(H2O)}].13H2O, K12[Co3(OH2)6(A-PW9O34)2].18H2O and
(cation)10[M4(OH2)2(B-PW9O34)2].(15–23)H2O, where M = Mn(II), Co
(II), Ni(II), Cu(II) and Zn(II)
obtained commercially, as indicated below. All were used without
further purification. The following materials were employed in the
synthesis of the compounds: Na2WO4·2H2O (99%, Sigma-Aldrich), Zn
(NO3)2·6H2O (98%, Sigma-Aldrich), Mn(NO3)2·6H2O (98%, APS), Co
(NO3)2·6H2O (98–102%, Univar), Ni(NO3)2·6H2O (98%, Chem Supply),
Cu(NO3)2.2½H2O (98%, APS), KCl (99.5%, BDH Analar), HNO3 (70%,
Merck).
The trimetallic compounds K6Na[Ni3(OH2)3(B-PW9O34){WO5(H2O)}].
13H2O and K12[Co3(OH2)6(A-PW9O34)2].18H2O were prepared according to
the procedures described by Clemente-Juan et al. and Knoth et al., respec-
tively [30,31]. Compounds of the type (cation)10[M4(OH2)2(B-PW9O34)2].
(15–23)H2O, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) were pre-
pared as described by Gomez-Garcia et al. (Mn compound), Finke et al. (Co,
Cu and Zn compounds) and Clemente-Juan et al. (Ni compound) [30,34,36].
Materials used for the catalyst reactions and kinetic studies included
allyl alcohol (99%, Riedel-de Haën), 2-methyl-2-propen-1-ol (98%,
Fluka), trans-crotyl alcohol (96%, Sigma-Aldrich, 96% trans isomer), 3-
methyl-2-buten-1-ol (99%, Sigma-Aldrich), geraniol (98%, Sigma-
Aldrich), cinnamyl alcohol (98%, Alfa Aesar), trans-2-penten-1-ol (95%,
Sigma-Aldrich), cis-2-penten-1-ol (97%, Alfa Aesar), trans-2-hexen-1-ol
(97%, Alfa Aesar), cis-2-hexen-1-ol (94%, Alfa Aesar), allyl ethyl ether
(95%, Sigma-Aldrich), 1,2-dichloethane (ACS reagent, ≥99%, Sigma-
Aldrich), t-butylbenzene (99%, Sigma-Aldrich), H2O2 (~30% aqueous
solution, Chem Supply), NaI (99.5%, Chem Supply), 2-propanol
(99.7%, Chem Supply), glacial acetic acid (100%, BDH Analar),
Na2S2O3.5H2O (Merck, AR Grade, ≥99.0%) and Aliquat 336 (methyl-
trioctylammonium chloride, 96%, Sigma-Aldrich).
2.2. Catalysis reactions
2.2.1. Catalyst stock solution preparation
Bulk heteropolyoxotungstate catalyst solutions were prepared by
dissolving 0.050 mmol of the Na+/K+ or K+ salt in 10–20 mL of water.
Slight heating (~30–40 °C) was required to achieve dissolution in some
cases. On cooling, each solution was transferred to a 100 mL separatory
funnel, along with 5 mL of water to ensure complete transfer of the
catalyst solution. To this solution was added exactly 10.00 mL by pip-
ette of a 1,2-dichloroethane solution that contained 1.0 mmol of Aliquat
336 (20 equivalents). The amount of added Aliquat 336 was always
kept constant, as quaternary ammonium cations are known to extract
H2O2 into 1,2-dichloroethane [22]. The resulting mixture was shaken
for about 5 min which caused the coloured aqueous phase to deco-
lourise, indicating that the catalyst had transferred to the 1,2-di-
chloroethane phase. The two phases were allowed to separate
Solid state infrared spectra (IR) were obtained using a Perkin Elmer
Spectrum BX instrument as KBr discs (resolution of 2 cm−1).
UV–visible-near infrared spectra were recorded on a Cary 6000i in-
strument from 1800 to 300 nm. Solution spectra were obtained using
1 cm quartz cells, while diffuse reflectance spectra were obtained with a
Harrick praying mantis attachment. Thermogravimetric/differential
thermal analyses (TG/DTA) were performed on
a Perkin Elmer
3

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
overnight. In the case of a colourless Zn(II)-containing compound,
shaking was continued for 10 min to ensure complete transfer of the
catalyst to the organic phase. After phase separation the 1,2-di-
chloroethane solution containing the catalyst was transferred to a dry
volumetric flask.
3. Results and discussion
3.1. Synthesis and characterization of K6Na[Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}].13H2O, K12[Co3(OH2)6(A-PW9O34)2].18H2O and
(cation)10[M4(OH2)2(B-PW9O34)2].(15–23)H2O, where M = Mn(II), Co
(II), Ni(II), Cu(II) and Zn(II)
2.2.2. Catalysis reactions and analysis of the products of reaction
All catalysis reactions were performed in
a
biphasic 1,2-di-
All heteropolyoxotungstate compounds were prepared according to
literature procedures, including the precursor compounds A-Na8H
(PW9O34).24H2O and B-Na8H(PW9O34).8H2O (water content de-
termined by TG/DTA) [39,40]. The resulting compounds were char-
acterized by TG/DTA, and both IR and UV–visible spectroscopy. The
TG/DTA and spectroscopic data are given in the Supplementary
Material, Tables S1–S5 and Figs. S1–S3, and where available agree with
the published data. In particular, the compound formulated by Knoth
et al. as K12[Co3(OH2)3(A-PW9O34)2].10H2O [31], was found in our
case to have twenty-four thermally-removable water molecules (up to
240 °C, Supplementary Material, Tables s4), and is better formulated as
chloroethane/aqueous H2O2 mixture using a 10 mL glass vial that was
fitted with a plastic screw top and teflon-coated magnetic stir bar, which
was stirred using a Corning immersible stir pad (Model No. 440837). The
screw top had a pinhole to vent any oxygen buildup from dismutation of
H2O2. Typically, a sample of catalyst (1.0 mL, 5 μmol, 20 equivalents of
Aliquat 336) was transferred to a 5 mL volumetric flask, followed by
5.0 mmol of organic substrate (0.35–1.15 mL) and 2.5 mmol (0.39 mL) of
t-butylbenzene (as an internal GC standard) and the solution made up to
5.00 mL with 1,2-dichloroethane. This procedure gave a constant con-
centration of 1 mM of catalyst for each reaction, with a constant substrate
concentration of 1.000 M. The resulting solution was then transferred to
the glass vial, and a 2-fold excess of aqueous H2O2 (~30% v/v,
~10 mmol, 1 mL, pH ~3.2) over organic substrate was added. This al-
ways resulted in a biphasic mixture. The mixture was stirred vigorously
at the designated temperature for 3 h, after which time the phases were
allowed to separate by standing (5 min). The organic phase, both prior to
the addition of the aqueous H2O2 and following reaction after standing,
was analyzed by direct injection into a gas chromatograph. For the lower
molecular weight allylic alcohols allyl alcohol, 2-methyl-2-propen-1-ol
and trans-crotyl alcohol, but less so for 3-methyl-2-buten-1-ol, some
transfer of the alcohol to the aqueous phase took place during the re-
action, which resulted in greater apparent reactivity. This was sig-
nificantly reduced by adding KCl (0.30 g) and stirring for 1–2 min, before
the phase separation stage and subsequent GC analysis. Gas-phase
chromatographic analyses were performed on a Shimadzu 17A gas
chromatograph with a 15 m AT-Aquawax capillary column, coupled to a
Shimadzu C-R8A Chromatopac integrator as described previously [14].
On completion of a reaction the H2O2 consumption was analyzed
according to the following procedure. A 0.10 mL sample of the aqueous
phase was removed using a 0.10 mL gas-tight precision glass syringe
(SGE, 100F-GT) and analyzed by iodometric titration in acidic 2-pro-
panol with standard 0.100 M Na2S2O3 according to the procedure of
Mair and Hall [37].
K
12[Co3(OH2)6(A-PW9O34)2].18H2O. Thus the anion is formally
[Co3(OH2)6(A-PW9O34)2]12−, as shown in Fig. 2, containing three six-
coordinate Co(II) with three externally coordinated water molecules,
one to each of the three Co(II) in the waist of the anion, and three
internally coordinated water molecules (as proposed by Knoth et al.
[31]), again with one attached to each of the three Co(II). In Fig. 2 the
anion structure has been based on that of the related [Cu3(NO3)(A-
PW9O34)2]13− ion [41], but with Co⋯Co distances set at 5.356 Å,
which is the same as the average Cu⋯Cu distance in the Cu3-containing
−
anion. In this last anion, there is an internally coordinated NO3 ion
(rather than three water molecules), but no externally coordinated H2O
molecules (the three Cu(II) are each 5-coordinate), presumably because
the externally directed 6th coordination site around each Cu(II) is va-
cant from weaker bonding as a result of the Jahn-Teller effect, with the
two elongated trans bonds on each Cu(II) directed towards an oxygen
−
atom of the internally coordinated NO3 group as well as (at least
formally) radially outwards. In the [Co3(OH2)6(A-PW9O34)2]12− ion,
the three external Co(II) sites must be accessible by H2O and allylic
alcohol for the reaction mechanism to apply (see below), as the term-
inal W = O sites, which generate the W(O2) sites on reaction with
H2O2, are located on the outside of the anion and are not available to
the internally coordinated Co(II) sites. Also note that in Fig. 2 there are
several long O⋯H bonds between the coordinated water molecules in
the internal cavity (2.35–2.37 Å).
2.3. Kinetic studies
3.2. The trimetallic [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and
[Co3(OH2)6(A-PW9O34)2]12− ions
Kinetic studies of the epoxidation of 3-methyl-2-buten-1-ol by H2O2
under biphasic conditions were performed using [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− as the catalyst an-
ions, allowing comparison with our previous studies on [WZn{Mn
(OH2)}2(B-ZnW9O34)2]12− as the catalyst [14].
3.2.1. Comparative catalytic activity studies of the trimetallic [Ni3(OH2)3(B-
PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− ions
A series of reactions using the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
and [Co3(OH2)6(A-PW9O34)2]12− ions, with a range of allylic alcohols and
related compounds, which has been used previously for a comparison of
the catalytic activities of the related [WZn{M(OH2)}2(B-ZnW9O34)2]12−
ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), was studied under
identical conditions of concentration and temperature, etc., to that re-
ported in the published study [14]. This allows a direct comparison with
the earlier results. The catalytic activity results are shown in Table 1, along
with the results for the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion (the re-
action data for the analogous Co(II), Ni(II), Cu(II) and Zn(II) species can be
found in [14]). Also included in Table 1 are the % H2O2 efficiency and the
turnover frequency (TOF, h−1), which is defined as the number of moles
of product per mole of catalyst per hour. All reactions exhibited less than
100% efficiency in H2O2 consumption, as a result of the competitive dis-
mutation of H2O2 (i.e. 2H2O2 → 2H2O + O2), arising from the hetero-
transition metal ions. The efficiency of H2O2 consumption varied widely,
but there was no apparent relationship between the efficiency of H2O2
consumption with either the metal ion or the allylic alcohol.
A typical organic solution makeup consisted of 5 μmol of catalyst,
20 μmol of Aliquat 336, 5 mmol of 3-methyl-2-buten-1-ol (0.51 mL) and
2.5 mmol of t-butylbenzene (0.39 mL), all made up to 5.00 mL in 1,2-
dichloroethane, and which was subsequently homogenized by shaking.
Reaction was initiated by addition of a mixture of 20 mmol (2.00 mL) of
30% aqueous H2O2 and 2.00 mL of H2O (pH 7). A biphasic mixture
always resulted. For the catalyst and H2O2 concentration dependence
studies the amounts of catalyst and H2O2/H2O were varied accordingly.
The reaction was studied by GC, as described above. Sampling occurred
every 15 min, and the experimental conditions were adjusted to give
sufficient data points for subsequent analysis. Temperature studies were
performed from 15.0 to 35.0 °C for [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− and 1.0 to 15.0 °C for [Co3(OH2)6(A-PW9O34)2]12−
.
The catalyst concentrations were varied from 0.25 to 2.00 mM, while
the aqueous H2O2 concentration was varied from ~ 2.13 to ~ 8.60 M.
The kinetic data were analysed using an in-house extended version of
ReactLab Kinetics [38].
4

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
Table 1
Oxidation of allylic alcohols by 30% aqueous H2O2 using the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− ions in biphasic 1,2-di-
chloroethane/H2O at 15 °C. The corresponding data for the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion under the same conditions are included [14].
Substrate
Aniona
% conversion, % H2O2 efficiency, TOF (h−1), % aldehyde
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
[Co3(OH2)6(A-PW9O34)2]12−
[WZn{Mn(OH2)}2(B-ZnW9O34)2]12−
allyl alcohol
6, 22, 18, < 0.5
13, 43, 43, < 0.5
27, 36, 90, < 0.5
64, 94, 214, < 0.5
36, 98, 133, < 0.5
23, 78, 77, < 0.5
28, 71, 93, 1c
29, 43, 96, < 0.5
51, 44, 171, < 0.5
73, 56, 243, < 0.5
99, 73, 331, < 0.5
94, 64, 375, 1
48, 51, 159, 3
33, 42, 108, 1
38, 32, 126, 1
46, 48, 153, 1
44, 38, 146, < 0.5
NR
26, 68, 86, < 1
40, 72, 134, < 1
76, 82, 252, < 1
94, 74, 314, < 1
77, 97, 306, 3
41, 68, 137, 4
25, 48, 82, 2
3, 30, 10, < 0.5
43, 46, 144, 2
27, 29, 89, 1
NR
2-methyl-2-propen-1-ol
trans-crotyl alcohol
3-methyl-2-buten-1-ol
geraniolb
cinnamyl alcohol
trans-2-hexen-1-ol
cis-2-hexen-1-ol
17, 64, 56, 1c
trans-2-penten-1-old
cis-2-penten-1-old
allyl ethyl ether
37, 78, 123, < 0.5c
45, 74, 150, < 0.5c
NRe
a
Reaction conditions: Reactions were carried out at 15 °C for 3 h using 5 mmol of allylic alcohol, 2.5 mmol of t-butylbenzene (internal standard for gas chro-
matography), 5 μmol of catalyst, 20 μmol of Aliquat 336 and 10 mmol of H2O2.
b
c
d
e
Reaction conditions: as above, but reactions for 2.5 h.
Reaction conditions: as above, but reaction at 25 °C.
Data for [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− from [42].
NR = No reaction.
The relative reactivities of the allylic substrates have been compre-
hensively discussed previously [14,15]. Briefly, increasing substitution
around the C]C by, for example, –CH3 groups (reactions of allyl alcohol,
2-methyl-2-propen-1-ol, trans-crotyl alcohol and 3-methyl-2-buten-1-ol)
demonstrated increases in catalytic activity, which can be correlated
with the decreasing stability of the HOMO [14]. Geraniol (trans-3,7-di-
methyl-2,6-octadien-1-ol) showed epoxidation only at the C]C site
proximate to the –OH group, demonstrating high regioselectivity in these
reactions. Cinnamyl alcohol (trans-3-phenyl-2-propen-1-ol) was less re-
active than the structurally similar trans-crotyl alcohol, and always re-
sulted in a small amount of cinnamaldehyde. A steric effect of the large
phenyl group is likely to play a role in reactions of this substrate.
Of more interest are the reactions of the trans- and cis-pentenols and
the trans- and cis-hexenols. In the absence of a catalyst, epoxidation
usually occurs more readily for the cis-isomer, based on the energies of
the HOMOs [14]. However, for [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− it
was observed previously that the trans-isomer was much more reactive
than the cis-isomer for both the trans- and cis-pentenols and the trans-
and cis-hexenols [14,42], which suggests a steric role for the catalyst at
the Mn(II) site where the allylic alcohol coordinates. Notably, an
identical order of activity was observed for the trans- and cis-hexenols
with the other four [WZn{M(OH2)}2(B-ZnW9O34)2]12− ions, where
M = Co(II), Ni(II), Cu(II) and Zn(II) [14]. However, for [Ni3(OH2)3(B-
PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− the pat-
terns of trans:cis activity were observed to be somewhat different to
those found for the five [WZn{M(OH2)}2(B-ZnW9O34)2]12− ions, as
well as to each other (Table 1). This confirms that the environment
around the M(II) site in these sandwich-type anions, where the allylic
alcohol coordinates, imposes a steric effect on the epoxidation reaction,
assuming the same reaction mechanism (see below). The [Ni3(OH2)3(B-
PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− ions are
both much more open at their M(II) active sites than are the five [WZn
{M(OH2)}2(B-ZnW9O34)2]12− ions (compare Figs. 1–3), such that the
ratio of activity for the cis isomers relative to the trans isomers in-
creases, and in some cases actually exceeds that of the trans isomers.
A substrate that has an oxygen atom in the same relative position to
the C]C of the allylic alcohols but does not contain an –OH group is
allyl ethyl ether. However, this did not show any reaction with H2O2
using either trimetallic catalyst, and demonstrates the importance of
the –OH group to the reaction mechanism. Note that not one of the five
[WZn{M(OH2)}2(B-ZnW9O34)2]12− ions acted as a catalyst for the re-
action of allyl ethyl ether with H2O2 under the same conditions [14].
Of more importance for the present study, it is noted that for the three
anions listed in Table 1, the order of catalytic activity is observed to be
[Co3(OH2)6(A-PW9O34)2]12− > [WZn{Mn(OH2)}2(B-ZnW9O34)2]12−
>
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
.
3.2.2. Kinetic studies of the trimetallic [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− ions in the
epoxidation of 3-methyl-2-buten-1-ol using aqueous H2O2
Kinetic studies of the epoxidation of 3-methyl-2-buten-1-ol by H2O2
under biphasic conditions (in 1,2-dichloroethane/H2O) using the tri-
metallic [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-
PW9O34)2]12− ions as catalysts were undertaken, as these offer the
possibility of a greater number of active sites compared to the pre-
viously examined [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion [14]. As was
found for this last species, all kinetic traces for the two trimetallic an-
ions showed induction periods (see below), suggesting a similar reac-
tion mechanism to that of the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion,
albeit with three active metal sites rather than two. The reaction me-
chanism involves the exchange of the coordinated water on an M(II) by
allylic alcohol (which coordinates through the –OH group), formation
of a W(O2) group at an adjacent site by reaction of a W = O unit with
H2O2, transfer of the peroxy oxygen to the coordinated allylic alcohol,
followed by loss of the coordinated epoxide. Further discussion is given
below. The Ni-OH2 distances in the [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− catalyst range from 2.09(2) to 2.10(2) Å, which are
slightly long for a Ni-OH2 distance (typically 2.06 Å) so that the water
will readily exchange with another ligand. The structure of the
[Co3(OH2)6(A-PW9O34)2]12− ion, while it has not been reported, has
been discussed above and it is assumed that the externally-directed Co-
OH2 bonds will also exchange readily as Co(II) is a highly labile first
row transition metal ion.
As an example of the formation of peroxy species as part of the
reaction mechanism, the IR spectroscopic approach that was used
previously for the detection of peroxo species of [WZn{Mn(OH2)}2(B-
ZnW9O34)2]12− was employed to examine peroxo formation in
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− [14]. The results are shown in
the Supplementary Material, Fig. S4. The figure shows the IR spectra
from 1200 to 600 cm−1 of Aliquat 336, the parent [Ni3(OH2)3(B-
PW9O34){WO5(H2O)}]7− ion, and the latter following treatment with
15% H2O2 after 1 and 2 h. Note that the splitting of the υ(PeO) band
observed for [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− in Fig. S1 cannot
be observed in Fig. S4(b), 1042 cm−1, which is likely the result of the
5

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
type of sample preparation. All bands in the spectrum arise from the
anion or Aliquat 336. Bands at 1059 (strong), 894 (medium) and 722
(strong) cm−1 may be observed for Aliquat 336 in Fig. S4(a). The peaks
at 1059 and 722 cm−1 are observed also in the spectra of the anion
treated with H2O2, Fig. S4(c) and (d), with the band at 722 cm−1
overlapping with the broad υ(W–O–W) band from the edge-shared WO6
octahedra of the anion. On treatment with H2O2, formation of a peroxo
band, υ(OeO), can be observed at 824 cm−1, which grows with time.
This is likely a W(η2-O2) site, based on the frequency [14]. There is also
some shift in the positions of the υ(W]O) bands found at 954 and
937 cm−1 in the parent compound, to 962 and 947 cm−1, with the
higher frequency peak becoming more intense than the lower frequency
peak. However, it cannot be determined as to where the peroxidation is
9−
occurring. This could occur at W]O sites of the [B-PW9O34
]
anion,
or at the formal [WO5(H2O)]4− capping site, as the latter has two
terminal W]O bonds.
Fig. 5. H2O2 concentration dependence: plot of residual 3-methyl-2-buten-1-ol
The kinetic traces for the dependence of the reaction on the
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and H2O2 concentrations (both
at 25 °C), as well as the temperature dependence study (15–35 °C) are
shown in Figs. 4–6, respectively. For [Co3(OH2)6(A-PW9O34)2]12− the
traces for catalyst concentration dependence (5 °C) and temperature
dependence (1–15 °C) are shown in the Supplementary Material, Figs.
S5 and S6. The figures show the dependence of the reactions as plots of
the residual 3-methyl-2-buten-1-ol concentration vs time. In general, all
reactions were followed until about 50–60% conversion of the 3-me-
thyl-2-buten-1-ol. This was imposed on the study, as it has been shown
that, in related systems, the build-up of product epoxide can lead to
decomposition of the catalyst [43]. A similar restriction was likewise
applied in the study of the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion as
catalyst [14].
concentration versus time for the epoxidation of 3-methyl-2-buten-1-ol at 25 °C:
[3-methyl-2-buten-1-ol] = 1.000 M,
catalyst
[Ni3(OH2)3(B-PW9O34){WO5
(H2O)}7−] = 1.00 mM, [Aliquat 336] = 0.020 M.
The induction period depended on the concentrations of the catalyst
and H2O2 (the latter was only studied for the [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− ion), as well as the temperature. Similar findings were
found for the analogous studies of the [WZn{Mn(OH2)}2(B-
ZnW9O34)2]12− ion as catalyst [14]. The reactions involving this latter
anion were interpreted in terms of an autocatalysis reaction, as this
anion was known to be stable under the reaction conditions. In the
present case a comprehensive set of reaction steps that account for such
a mechanism is given below, equations (1) to (21), but unlike the two-
site [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion, three sites were used for
Fig. 6. Temperature dependence (15–35 °C): plot of residual 3-methyl-2-buten-1-
ol concentration versus time for the epoxidation of 3-methyl-2-buten-1-ol:
[3-methyl-2-buten-1-ol] = 1.000 M,
catalyst
[Ni3(OH2)3(B-PW9O34){WO5
(H2O)}7−] = 1.00 mM, [Aliquat 336] = 0.020 M.
the
trimetallic
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
and
[Co3(OH2)6(A-PW9O34)2]12− ions, which extended the number of re-
action steps from fourteen to twenty-one. The twenty-one steps in the
mechanism can be classified according to the following scheme.
(a) Preliminary equilibria (coordination of C3 by A):
C3 + A ⇋ C3A (K1)
(1)
(2)
(3)
C3A + A ⇋ C3A2 (K2)
C3A2 + A ⇋ C3A3 (K3)
(b) Peroxo formation; oxygen transfer and associated equilibria:
C3A3 + O → C3A3O (k4)
(4)
(5)
(6)
(7)
(8)
(9)
C3A3O → C3A2E (k5)
C3A2 + E ⇋ C3A2E (K6)
C3A2 + O → C3A2O (k7)
C3A2O → C3AE (k8)
C3A + E ⇋ C3AE (K9)
(c) Autocatalysis and sundry reactions:
Fig. 4. Catalyst ([Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
) concentration de-
C3A2E + O → C3A2ES (k10
C3A2S + E ⇋ C3A2ES (K11
)
(10)
(11)
pendence: plot of residual 3-methyl-2-buten-1-ol concentration versus time for
the epoxidation of 3-methyl-2-buten-1-ol by H2O2 at 25 °C: [3-methyl-2-buten-
1-ol] = 1.000 M, [H2O2] = 4.50 M, [Aliquat 336] = 0.020 M.
)
6

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
C3A2S + A ⇋ C3A3O (K12
)
(12)
two-site [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion as the catalyst [14].
Thus, in the present case, the four species C3A, C3A2, C3AE and C3A2E
comprise a 4-species cyclic system, whereby C3A can react either with A
first and then E, or with E first and then A, with both pathways resulting
in C3A2E formation. ReactLab Kinetics only requires three sets of
equilibrium equations (equations (2), (6) and (9)), along with their
equilibrium constants, to be defined for a 4-species cyclic system. The
program automatically generates the fourth equilibrium, in this case the
reaction between C3AE and A to generate C3A2E. The equilibrium
constant for this reaction is calculated by microscopic reversibility/
detailed balancing from the three defined equilibrium constants
[14,44]. Also, using the chosen equilibrium constants (see below), any
reaction of C3 with O to give C3O can be neglected. The concentrations
of C3 in both the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− system and
the [Co3(OH2)6(A-PW9O34)2]12− system are 1.1 × 10−3 and
2.9 × 10−3 mM, respectively (both for an initial [C3] = 1.00 mM and
[A] = 1.000 M), which are much lower than fully or partially com-
plexed C3. In addition, as the equilibria can be regarded as “in-
stantaneous” compared to the kinetically observable rate constants, the
reaction of C3 with O to give C3O prior to complexation with A or E may
be neglected. Another reaction that was neglected was the reaction of
C3A with O, as the concentration of C3A is much less than C3A2 and
C3A3 (the concentrations of C3A are only 2.0 × 10−2 mM and
3.7 × 10−2 mM in the Ni- and Co-trimetallic systems, respectively), so
that the major reaction pathways proceed through C3A3 and C3A2. The
reactions of C3E3, C3AE2 and C3E2 with O to give C3E3S, C3AE2S and
C3E2S (recall that S is an O that generates a peroxo group adjacent to E)
have also been neglected. Reaction can only occur if there is loss of E
and addition of A. Because the catalysis reactions were only followed
until a maximum of 50–60% conversion, the concentrations of C3E2 and
C3AE2 only reach a maximum of ~0.14 and ~0.09 mM, respectively,
with that of C3E3 much lower still, in the two trimetallic systems. Thus
these reactions were not included in the reaction scheme above. Note,
however, that if the catalysis studies were to be followed to much
higher % conversions, then these reactions, along with their reactions
for loss of E and addition of A, together with subsequent reactions
(transfer of O to A and product formation) would have to be taken into
account. The same approach was also applied successfully in our pre-
vious study using the two-site [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion
as catalyst [14]. Note also that C3E, C3AE, C3E2 and C3AE2 will form a
4-species cyclic system and due account of only three of the four pos-
sible reactions would need to be considered.
(Note, C3A3O formation in equation (12) is followed by equation (5)
above, while an S oxygen becomes an O oxygen when the coordinated E
is lost and replaced with an A);
C3AE + O → C3AES (k13
C3AS + E ⇋ C3AES (K14
C3AS + A ⇋ C3A2O (K15
)
(13)
(14)
(15)
)
)
(Note, C3A2O formation in equation (15) is followed by equation (8)
above, while an S oxygen becomes an O oxygen when the coordinated E
is lost and replaced with an A);
C3A2E + O → C3A2EO (k16
C3A2EO → C3AE2 (k17
C3AE + E ⇋ C3AE2 (K18
C3AE + O → C3AEO (k19
C3AEO → C3E2 (k20
C3E + E ⇋ C3E2 (K21
)
(16)
(17)
(18)
(19)
(20)
(21)
)
)
)
)
)
The equilibrium and rate constants are numbered successively from
1 to 21, with the equilibrium constants designated by K and the rate
constants by k. These are shown in parentheses to the right of the re-
action in each step of the mechanism in equations (1) to (21). A double
arrow (⇋) represents an equilibrium that is regarded as kinetically very
fast on the scale of the overall epoxidation reaction, while a single
arrow (→) is a kinetically observable reaction.
In the above equations C3 represents the three metal sites of the
anion (with C representing a single metal site) and A is the allylic al-
cohol. Thus C3A represents coordination of an allylic alcohol to a single
metal site (so that C3A2 and C3A3 represent coordination of 2 and 3
metal sites, respectively). O is the “peroxo” atom of an H2O2 molecule,
and addition of O indicates formation of a W(O2) peroxo site in the
anion that is adjacent to a coordinated allylic alcohol, which is then
capable of transferring an O atom to the C]C bond of the coordinated
allylic alcohol in a separate step (note that the H2O2 reduction product
H2O is not represented in these equations). The product hydroxy ep-
oxide E is also able to coordinate to a metal center C through its –OH
group, as does the allylic alcohol. Note that E is different from AO, as
AO represents separated A and O entities that have yet to combine to
give the product hydroxy epoxide E. S represents formation of a W(O2)
peroxo site that is adjacent to a coordinated hydroxy epoxide, and so
cannot react until the hydroxy epoxide has been replaced by an allylic
alcohol. The formation of C3A3, C3A2 and C3A are required as at the
concentrations used, the three M(II) sites (i.e. C3) are almost fully co-
ordinated by A at the beginning of the reaction (~99.8%, with a [C3A3]:
[C3A2]:[C3A] ratio of 39.6:8.9:1 for the trimetallic Ni(II) system, and
19.8:6.3:1 for the trimetallic Co(II) system, using the chosen values of
K1, K2 and K3 for each of the trimetallic systems; see below). At the end
of the reaction, equations for the formation of C3E3, C3E2 and C3E
would also be required, but as all catalysis reactions were followed only
to 50–60% conversion the formation of C3E3 was neglected (also see the
discussion below). The steps in the mechanism consist only of equilibria
based on coordination of allylic alcohol and hydroxy epoxide to Ni(II)
or Co(II), formation of tungsten(VI)-peroxo sites, and transfer of oxygen
from a peroxo site to the C]C bond of the coordinated allylic alcohol.
The twenty-one steps listed above comprise only a subset of all
possible reactions that can be written in a three-site system. However,
many of the full set of equations are either unnecessary or are not re-
quired to successfully describe the observed kinetics. This was also
found for the kinetic study of the epoxidation of 3-methyl-2-buten-1-ol
by H2O2 under biphasic conditions (1,2-dichloroethane/H2O) using the
Because of the symmetrical nature of the trimetallic [Ni3(OH2)3(B-
PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12− ions
[29,30,41], several formation and rate constants can be regarded as
being effectively related or even equivalent, as was also invoked in the
analogous study of the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion as
catalyst [14]. In the case of the rate constants for peroxo formation, k4,
k7, k16 and k19 were assumed to be equal, as were k10 and k13. For
peroxo O transfer to an adjacent coordinated allylic alcohol, k5, k8, k17
and k20 were assumed to be equal.
No experimental values for the stepwise coordination of 3-methyl-2-
buten-1-ol to [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− could be ob-
tained. The molar absorptivities of the three observable Ni(II) d-d
transitions were very low (see the Supplementary Material, Table S3)
and the magnitudes of the stability constants were also expected to be
low as well, based on our previous studies [14,15]. This makes their
spectrophotometric determination effectively impossible. However,
previously obtained experimental data is available that allows us to
estimate a likely value. In this we follow a similar procedure to that
used for the evaluation of the stability constant for the addition of 3-
methyl-2-buten-1-ol to [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− as used in
our previous study [14]. Stability constants for allylic alcohols bound to
the substituted Co(II) in the related Dawson anion [P2W17O61Co(Cl)]9−
in 1,2-dichloroethane at 25 °C vary from 19.0 M−1 for allyl alcohol,
13.9 M−1 for 2-methyl-2-propen-1-ol, 12.2 M−1 for trans-crotyl alcohol,
7

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
to 12.7 M−1 in 3-methyl-2-buten-1-ol. For the same four allylic alcohols
the K1 values for the analogous Mn(III)-substituted Dawson anion range
from 14.2 to 10.4 M−1 [15]. These values do not involve deprotonation
of the –OH group, and hence any formation of an M(n+)-alcoholate
type species.
systems. The average values are given in both tables. In general, the
rate constants for the concentration dependence studies are fairly
constant for both the Ni and Co trimetallic systems. Thus k4 is relatively
constant, while some slight variations are observed for k5 and k10 for
the [catalyst] and [H2O2] (Ni3 system) and [catalyst] (Co3 system)
dependence studies. The variations may have several possible causes.
Note that the studies were carried out in 1,2-dichloroethane solvent,
and involved significant concentrations of ionic species (both catalyst
and Aliquat 336, although the latter is actually constant throughout all
studies), and high concentrations of neutral species (3-methyl-2-buten-
1-ol, H2O2 and product hydroxy epoxide) in 1,2-dichloroethane. No
activity coefficients could be applied during the numerical fitting pro-
cess so that no account could be made for ionic strength which will vary
in the [catalyst] studies or for the actual activities of the neutral solvent
in all studies. This will likely have some effect on the fitted values of the
rate constants and are likely the cause of the observed variations. Si-
milar variations were observed previously for the related [WZn{Mn
(OH2)}2(B-ZnW9O34)2]12− catalyst system. While k4 and k5 are fairly
similar for both the Ni3 and Co3 catalyst systems, even allowing for the
temperature difference, k10 is much faster in the Co3 system. This likely
arises from the ease of formation of the transition state in the auto-
catalytic pathway. Further comparisons are evident from the enthalpies
of activation (ΔH≠), entropies of activation (ΔS≠) and free energies of
activation at 25 °C (ΔG≠), see below.
For the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− ion, we have chosen to
use a value of K1 = 17.7 M−1 (i.e. log K1 = 1.25) for a Ni(II) site co-
ordinating a 3-methyl-2-buten-1-ol molecule at 25 °C in 1,2-dichloroethane
for the present study, which was evaluated by taking the stability constant
value for complexation of 3-methyl-2-buten-1-ol to Co(II) in the (structu-
rally similar) Co(II)-substituted Dawson anion and scaling it upwards by
40%, so as to account for the natural order of stability constants (the Irving-
Williams series). The scaling was based on relative stability constants for
coordination of a neutral O donor (in this case H2O) to Ni(II) relative to Co
(II), made with the same experimental techniques and under the same
experimental conditions [45,46]. This generates the value for K1. However,
it is unlikely that K1, K2 and K3 are equal for all three Ni(II) sites, given that
the Ni···Ni distances average only 3.176 Å in the [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− ion [29,30]. Some reduction in Kn values for successive
complexation would likely arise from steric effects between the co-
ordinating allylic alcohol molecules. Moreover, in the case of the two-site
species [Co4(OH2)2(B-PW9O34)2]10−, for the N-donor N-methylimidazole,
which has much greater stability constant values than the O-donor 3-me-
thyl-2-buten-1-ol used in the present study, two ligand complexation steps
have been observed with K2 < K1 [47]. We have thus chosen to succes-
sively decrease a Kn+1 value relative to Kn by the factor 0.5, giving values
of K2 = 8.9 M−1 and K3 = 4.4 M−1 for complexation of the three 3-me-
thyl-2-buten-1-ol molecules to Ni(II). Using the same approach for
[Co3(OH2)6(A-PW9O34)2]12− (despite the fact that the Co⋯Co distance is
certain to be larger, see above) the initial K1 value was assumed to be the
same as that for complexation of 3-methyl-2-buten-1-ol to Co(II) in the Co
(II)-substituted Dawson anion, i.e. K1 = 12.7 M−1, with K2 = 6.4 M−1 and
K3 = 3.2 M−1. In the case of the Co(II) trimetallic system, no attempt was
made to account for the difference in temperature at which the kinetic
dependence on the catalyst concentration was determined and the stability
constant data. However, this temperature difference should only lead to a
small effect on the value of the stability constants [48]. While the final
values of K1, K2 and K3 in both trimetallic systems may be approximate in
each case, some variation in their magnitudes will be discussed below. No
analogous stability constants for complexation of the product hydroxy
epoxide to Ni(II) or Co(II) are known (i.e. K6, K9, K11, K14, K18 and K21), but
these are likely to be similar to the analogous values for 3-methyl-2-buten-
1-ol, so that the stability constants for 3-methyl-2-buten-1-ol and its ep-
oxide were assumed equivalent for the respective values of K1, K2 and K3
for both trimetallic systems. This is the same procedure that was used in the
case of our previous study of the kinetics of the epoxidation of 3-methyl-2-
buten-1-ol by H2O2 under biphasic conditions with [WZn{Mn(OH2)}2(B-
ZnW9O34)2]12− as the catalyst [14].
As discussed previously [14], the autocatalysis is caused by an in-
crease in the rate of formation of a peroxo site that is adjacent to a co-
ordinated hydroxy epoxide (k10 and k13) relative to one adjacent to a
coordinated allylic alcohol (k4 and k7). The steps that are involved in the
autocatalysis (k5, k10, K11, K12, k8, k13, K14 and K15) result in the accel-
eration of the overall rate with time following the induction period, as
shown in Figs. 4–6, S5 and S6 and the increase in k10 (and k13) relative to
k4 (and k7), respectively, may again be attributed to an electronic (pos-
sibly inductive) effect from the coordinated hydroxy epoxide. Alter-
natively, the increase may be steric in nature, with the coordinated hy-
droxy epoxide actually directing formation of the adjacent peroxo group.
The latter might be expected to depend on the environment surrounding
the site where the allylic alcohol coordinates to the M(II), but no firm
conclusions can be made considering the significantly different en-
vironments of the [Co3(OH2)6(A-PW9O34)2]12−, [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− and [WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ions.
Variation of the K1 value (and thus K2, K3, K6, K9, K11, K12, K14, K15, K18
and K21) for the eleven equilibria by increasing its value (×2) or decreasing
its value (×0.5) produced only relatively small changes in the values of k4,
k5 and k10. For example, for the trimetallic Ni(II) system with [catalyst] =
1.00 mM, [3-methyl-2-buten-1-ol] = 1.000 M, [H2O2] = 4.50 M and
T = 25 °C, increasing and decreasing K1 (and K2 and K3) by the above
factors gave variations for k4 of −6 and +19%, respectively, k5 of +9 and
−15%, respectively, and k10 of +1 and −12%, respectively. Likewise, for
the trimetallic Co(II) system with [catalyst] = 1.00 mM, [3-methyl-2-buten-
1-ol] = 1.000 M, [H2O2] = 2.125 M and T = 5 °C, the variations were for
k4 of −2 and −3%, respectively, k5 of +6 and −23%, respectively, and k10
of +15 and −6%, respectively. In each case the choice made for K1 does
not have a strong bearing on the fitted rate constants.
With these simplifications in place, this leaves just k4, k5 and k10 to
be determined for each trimetallic system. The results of the fitting of
the GC data for both the catalyst and H2O2 concentration dependence
studies are given in Tables 2 and 3 for the [Ni3(OH2)3(B-PW9O34
)
{WO5(H2O)}]7− ion and for the concentration dependence study for
[Co3(OH2)6(A-PW9O34)2]12−, respectively. Also included in Tables 2
and 3 are the temperature dependence data for each trimetallic anion.
For [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− a typical fitting of the cal-
culated and experimental data for [catalyst] = 1.50 mM, [H2O2] =
4.50 M and [3-methyl-2-buten-1-ol] = 1.000 M at 25.0 °C is shown in
Fig. 7. The sum of the differences squared between the experimental
and fitted data points (ssq) is 0.0008. The ssq values for the reactions in
Table 2 varied from 0.0001 to 0.0010, and for the reactions in Table 3
from 0.0001 to 0.0004.
Setting the three values of K1, K2 and K3 equal within each of the
trimetallic systems, with values for all log Kn = 1.25 for the Ni(II)
system and all log Kn = 1.10 for the Co(II) system resulted in variations
for k4 of −9%, k5 of +14% and k10 of +6%, for the former, and k4 of
−5% k5 of +11% and k10 of +16% for the latter. Again the variations
are not large, indicating that the choice of K1 does not have a strong
influence on the fitted rate constants. Attempts to fit the values of K1, K2
and K3 along with the rate constants were unsuccessful, giving incon-
sistent results. It is assumed that the data are not sensitive (or extensive)
enough to be able to refine both the equilibrium and rate constants.
For the temperature dependence data for both trimetallic systems
(Tables 2 and 3) no variations in the values of K1 (and K2 and K3) were
made in either case, as the temperature dependence data of K1 for
The three unique rate constants k4, k5 and k10 reported in Tables 2
and 3 for the eight [catalyst] and [H2O2] dependence studies in Table 2
and the three [catalyst] studies in Table 3 should be individually con-
stant, with differences evident between the Ni and Co trimetallic
8

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
Table 2
Rate constants for the epoxidation of 3-methyl-2-buten-1-ol by H2O2 using the [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− ion as the catalyst under biphasic conditions
(Aliquot 336/1,2-dichloroethane/H2O).a
[catalyst] (M)
[H2O2] (M)
Temperature (°C)
Rate Constants (standard deviations in the last figure in parentheses)
k4 (=k7 = k16 = k19)(M−1 s−1
)
k5 (=k8 = k17 = k20)(s−1
)
k10 (=k13) (M−1 s−1
)
A. [catalyst] and [H2O2] dependence: [3-methyl-2-buten-1-ol] = 1.000 M
0.00200
0.00150
0.00100
0.00050
0.00025
0.00100
0.00100
0.00100
Average:
4.30
4.30
4.30
4.30
4.30
2.15
6.45
8.60
25
25
25
25
25
25
25
25
2.8 (2) × 10−4
2.8 (1) × 10−4
2.9 (1) × 10−4
2.8 (1) × 10−4
2.9 (2) × 10−4
3.1 (1) × 10−4
2.8 (1) × 10−4
2.7 (2) × 10−4
2.9 (1) × 10−4
0.15 (1)
0.17 (1)
0.21 (1)
0.25 (1)
0.34 (1)
0.09 (1)
0.28 (1)
0.39 (1)
0.23 (9)
0.21 (1)
0.26 (1)
0.57 (1)
0.92 (2)
1.23 (1)
0.47 (1)
0.37 (1)
0.42 (1)
0.6 (3)
B. Temperature dependence: [3-methyl-2-buten-1-ol] = 1.000 M
0.00100
0.00100
0.00100
0.00100
4.30
4.30
4.30
4.30
15
20
30
35
2.6 (1) × 10−4
2.8 (1) × 10−4
3.1 (5) × 10−4
3.2 (4) × 10−4
0.13 (1)
0.16 (1)
0.29 (1)
0.38 (1)
0.12 (1)
0.31 (1)
0.86 (1)
1.60 (1)
a
Equilibrium constants: log K1 = 1.25, log K2 = log K9 = log K14 = log K15 = log K21 = 0.95, log K3 = log K6 = log K11 = log K12 = log K18 = 0.65.
coordination of both 3-methyl-2-buten-1-ol and the hydroxy epoxide
are unknown in each system. However, as the values of K1 are small, the
variations of this value for each system will only vary slightly [48] and,
as shown above, the variations in the values of K1, K2 and K3 have only
a small effect on the fitted rate constants.
The resulting Eyring plots (ln k/T vs 1/T) for k4, k5 and k10 for the
trimetallic Ni(II) and Co(II) systems are given in the Supplementary
Material, Figs. S7 to S9 and S10 to S12, respectively (with all
R2 > 0.96). The resulting thermodynamic data are summarised in
Table 4, along with the analogous data for [WZn{Mn(OH2)}2(B-
ZnW9O34)2]12− as catalyst, which were obtained under identical con-
ditions to those used in the current study.
The large negative values of ΔS≠ for k4 and k5 (k3 and k4 for the
[WZn{Mn(OH2)}2(B-ZnW9O34)2]12− ion) likely indicate significant
changes in solvation during these steps. The values of ΔG≠ at 25 °C for
all three anions show the reduction in k10 (k9 for the two-site system),
the autocatalytic pathway, relative to k4 (k3 for the two-site system), the
non-autocatalytic pathway. The thermodynamic data are very similar
for all three anions, which is not surprising as the reaction mechanism
is the same in each case, albeit with a variation in the number of active
sites. However, the ΔS≠ value of k10 for [Co3(OH2)6(A-PW9O34)2]12− is
negative rather than positive, which may indicate significant differ-
ences in solvent redistribution during this step. This results in a smaller
ΔG≠ value at 25 °C for this anion, making it the most active catalyst of
the three heteropolyoxotungstate anions that have been studied.
Fig. 7. Calculated (×) and experimental (•) GC results for the epoxidation of 3-
methyl-2-buten-1-ol at 25 °C: [3-methyl-2-buten-1-ol] = 1.000 M, catalyst
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}7−] = 1.50 mM, [H2O2] = 4.50 M, [Aliquat
336] = 0.020 M. The ssq = 0.0008.
3.3. Comparative catalytic activity studies of the [M4(OH2)2(B-
PW9O34)2]10− ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II)
The series of sandwich-type heteropolyoxotungstate anions of
composition [M4(OH2)2(B-PW9O34)2]10− is directly related to the
Table 3
Rate constants for the epoxidation of 3-methyl-2-buten-1-ol by H2O2 using the [Co3(OH2)6(A-PW9O34)2]12− ion as the catalyst under biphasic conditions (Aliquot
336/1,2-dichloroethane/H2O).a
[catalyst] (M)
[H2O2] (M)
Temperature (°C)
Rate Constants (standard deviations in the last figure in parentheses)
k4 (=k7 = k16 = k19) (M−1 s−1
)
k5 (=k8 = k17 = k20) (s−1
)
k10 (=k13) (M−1 s−1
)
A. [catalyst] dependence: [3-methyl-2-buten-1-ol] = 1.000 M
0.00100
0.00050
0.00025
Average:
2.125
2.125
2.125
5
5
5
3.0 (1) × 10−4
2.9 (1) × 10−4
2.9 (1) × 10−4
2.9 (1) × 10−4
0.140 (1)
0.161 (1)
0.192 (1)
0.16 (2)
3.43 (9)
4.14 (2)
5.46 (5)
4.3 (8)
B. Temperature dependence: [3-methyl-2-buten-1-ol] = 1.000 M
0.00100
0.00100
0.00100
0.00100
2.125
2.125
2.125
2.125
1
2.9 (1) × 10−4
3.0 (1) × 10−4
3.1 (4) × 10−4
3.2 (9) × 10−4
0.098 (1)
0.140 (1)
0.206 (2)
0.248 (1)
2.29 (5)
3.43 (9)
5.4 (2)
6.4 (5)
5
10
15
a
Equilibrium constants: log K1 = 1.10, log K2 = log K9 = log K14 = log K15 = log K21 = 0.80, log K3 = log K6 = log K11 = log K12 = log K18 = 0.50.
9

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
Table 4
Enthalpies of activation (ΔH≠), entropies of activation (ΔS≠) and free energies of activation at 25 °C (ΔG≠) for the epoxidation of 3-methyl-2-buten-1-ol by aqueous
H2O2 under biphasic conditions (1,2-dichloroethane/H2O) using [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−, [Co3(OH2)6(A-PW9O34)2]12− and [WZn{Mn(OH2)}2(B-
ZnW9O34)2]12− ions as catalysts.
Function
Rate constant
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−
[Co3(OH2)6(A-PW9O34)2]12−
[WZn{Mn(OH2)}2(B-ZnW9O34)2]12−a
(ΔH≠) (kJ/mol)
k4
5 (8)b
38 (5)
89 (4)
2 (14)
42 (1)
47 (5)
15 (5)
50 (6)
72 (3)
k5
k10
(ΔS≠) (J/K/mol)
(ΔG≠) (kJ/mol)
k4
−295 (24)
−130 (15)
49 (11)
93
77
74
−304 (57)
−111 (3)
−66 (19)
93
75
67
−258 (15)
−86 (20)
2 (9)
92
76
71
k5
k10
k4
k5
k10
a
b
Data from [14]. (Note, k4, k5 and k10 in the current study correspond to k3, k4 and k9, respectively, for a two-site system; see [14].)
Standard deviations are given in parentheses and include fitting errors in the original rate constants.
previously studied [WZn{M(OH2)}2(B-ZnW9O34)2]12− series of ions,
where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) for both series. The
latter have been comprehensively studied as catalysts for the epoxida-
tion of allylic alcohols under biphasic conditions [14], and this previous
investigation formed the basis for the current studies. The two-site
[M4(OH2)2(B-PW9O34)2]10−-type ions were therefore examined using
identical reaction conditions (temperature, concentrations, etc.) to all
of these previous studies to allow direct comparisons of catalytic ac-
tivity.
The observed order suggests that catalytic activity in this series of
anions is controlled by thermodynamic considerations, such as the
magnitude of the equilibrium constant between the external aqua li-
gand and M(II). If this is large, then replacement of water by allylic
alcohol will be much reduced (or cannot proceed) leading to this step in
the mechanism being inaccessible so that no reaction will occur. Thus a
stronger M–OH2 linkage (a shorter bond) will lead to a reduced cata-
lytic activity. This will obviously depend on temperature.
There are two lines of evidence that support this suggestion. The first
comes from TG/DTA studies, which indicates the strength of retention of
the two water molecules bound to the M(II) in the waist of the anions.
The data have been presented previously in the Supplementary Material,
Table S5, and the TG/DTA study for K10[Co4(OH2)2(B-PW9O34)2]·20H2O
is shown in Fig. S3. The TG/DTA results show loss of both water of
crystallization and constitutional (i.e. M(II)-coordinated) water and in-
volves several steps, which at times overlap to a greater or lesser extent
(see Fig. S3). For all five [M4(OH2)2(B-PW9O34)2]10−-containing com-
pounds, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), there are, in
general, three identifiable stages from room temperature to the tem-
perature of complete loss of water, although in the case of the
[Cu4(OH2)2(B-PW9O34)2]10− ion only two stages are observable. The
highest temperature thermal step, based on the % of water lost can be
assigned to loss of the two water molecules attached at the two M(II)-
OH2 sites in the waist of the anion, although there is some overlap with
the second thermal stage, and complete overlap in the case of the Cu(II)-
containing anion. Arranging the five [M4(OH2)2(B-PW9O34)2]10− ions in
the order of the temperature required for complete water loss from
lowest to highest yields the order Ni(II) (198 °C) < Cu(II)
(225 °C) < Co(II) (240 °C) < Mn(II) (250 °C) < Zn(II) (273 °C). It is
noted that the two most catalytically active anions are also the ones that
lose their constitutional water at the lowest temperatures. The actual
order for the Ni(II)- and Cu(II)-containing anions (which is the reverse of
their relative catalytic activities) may reflect the different cation com-
Somewhat surprisingly, these anions were found to be much less active
as epoxidation catalysts. Some comparative data on the epoxidation of the
most reactive allylic alcohol used in this study, 3-methyl-2-buten-1-ol, is
given in Table 5. Only the Cu(II)- and Ni(II)-substituted anions acted as
oxygen transfer catalysts in these reactions at the temperature used in the
studies (15 °C). It is apparent that from a comparison of the catalytic ac-
tivity of the most active catalyst of this type, [Cu4(OH2)2(B-PW9O34)2]10−
,
with those of the three catalysts in Table 1 and the previously published
data on the other four [WZn{M(OH2)}2(B-ZnW9O34)2]12− ions, where
M = Co(II), Ni(II), Cu(II) and Zn(II) [14], that the [M4(OH2)2(B-
PW9O34)2]10− ions are the least active of all of these heteropolyoxotung-
state anions. Indeed the most active [M4(OH2)2(B-PW9O34)2]10− catalyst,
that is [Cu4(OH2)2(B-PW9O34)2]10−, is less active than the least active of
the five [WZn{M(OH2)}2(B-ZnW9O34)2]12− catalysts, that is [WZn{Ni
(OH2)}2(B-ZnW9O34)2]12−, under identical experimental conditions. Thus
further studies of the [M4(OH2)2(B-PW9O34)2]10− ions were not under-
taken.
Unlike the [WZn{M(OH2)}2(B-ZnW9O34)2]12− series of ions, the cata-
lytic activities of the [M4(OH2)2(B-PW9O34)2]10− ions could not be related
to the rate of exchange of the aqua groups on the M(II) ions in the waist of
the sandwich anion [14]. However, the observed order of Cu(II) > Ni
(II) > Zn(II) ~ Co(II) ~ Mn(II), although incomplete, is reminiscent of the
Irving-Williams series or natural order of stability constants for the same
first row (high-spin) transition metal ions, i.e. Mn(II) < Co(II) < Ni
(II) < Cu(II) > Zn(II) [49]. Note that depending on the ligand, Zn(II) is to
be found in a comparable position to Co(II) in this series.
positions of these two compounds in the TG/DTA studies (K+ vs Na+
/
K+). Of course, it should be recognized that TG/DTA is a dynamic
Table 5
Oxidation of 3-methyl-2-buten-1-ol by 30% aqueous H2O2 using the [M4(OH2)2(B-PW9O34)2]10− ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) in biphasic
1,2-dichloroethane/H2O at 15 °C.a
Anion
% conversion
% H2O2 efficiency
TOF (h−1
)
% aldehyde formation
[Mn4(OH2)2(B-PW9O34)2]10−
[Co4(OH2)2(B-PW9O34)2]10−
[Ni4(OH2)2(B-PW9O34)2]10−
[Cu4(OH2)2(B-PW9O34)2]10−
[Zn4(OH2)2(B-PW9O34)2]10−
< 3
< 3
14
–
–
Not detected
Not detected
Not detected
Not detected
Not detected
–
–
79
70
–
47
135
–
41
< 3
a
Reaction conditions: Reactions were carried out at 15 °C for 3 h using 5 mmol of allylic alcohol, 2.5 mmol of t-butylbenzene (internal standard for gas chro-
matography), 5 μmol of catalyst, 20 μmol of Aliquat 336 and 10 mmol of H2O2.
10

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
experiment with a temperature ramp and also deals with solids as op-
posed to anions in solution. The same temperature ramp was used for all
TG/DTA studies and the analyses were performed immediately following
powdering of the solids, to minimize any differences in these studies.
Nevertheless, the TG/DTA studies do suggest that water is not as easily
removed from the Co(II), Mn(II) and Zn(II)-containing anions.
as discussed above. The BV value for the Zn(II)–OH2 bond in [WZn{Zn
(OH2)}2(B-ZnW9O34)2]12− indicates a weak bond, which is consistent
with its known catalytic activity [14]. The BV value for the Co(II)–OH2
bond in [Co4(OH2)2(B-PW9O34)2]10− is weaker than expected (BV <
0.33), and not in agreement with the observed catalytic activity, as
noted above.
The second line of evidence comes from published crystallographic
data on the five [M4(OH2)2(B-PW9O34)2]10−-containing compounds.
Table 6 provides some structural data on the [M4(OH2)2(B-PW9O34)2]10−
ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) (with the PV
replaced by AsV for the Zn(II) species), as well as on [FeI4II(OH2)2(B-
PW9O34)2]6−, the two [WZn{M(OH2)}2(B-ZnW9O34)2]n− ions, where
M = Zn(II) and Rh(III), and [WCu{Cu(OH2)}2(B-ZnW9O34)2]12− for
comparison [6,8,17,30,33–35]. Also included in Table 6 is the average
M–OH2 distance in the corresponding Tutton salts, (NH4)2[M(OH2)6]
(SO4)2, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) (all have three
crystallographically unique bonds, with standard deviations given in
parentheses, except for the Jahn-Teller distorted bond in the Cu(II)-
containing anion) [50], as well as the bond valences (BV) of the M
(II)–OH2 bonds in the heteropolyoxotungstate anions [51].
The one question that remains to be answered is why is it that the
[M4(OH2)2(B-PW9O34)2]10− series of anions is much less reactive than
the [WZn{M(OH2)}2(B-ZnW9O34)2]12− series of anions, where M = Mn
(II), Co(II), Ni(II), Cu(II) and Zn(II) in both series? The major difference
12−
between the two series is that the Zn(II) in the [B-ZnW9O34
]
ion has
9−
been replaced by P(V) in the [B-PW9O34
]
ion. The P(V) has a much
smaller tetracoordinated ionic radius than Zn(II), with Zn(II) = 0.74 Å
and P(V) = 0.31 Å (note As(V) = 0.59 Å) [52], so that a PeO bond
(about 1.53 Å) is much shorter than a Zn-O bond (1.96 Å). This difference
has an effect on the length of the M−O bond that is trans to the ex-
ternally directed M−OH2 bond. The oxygen atom of the M−O bond
trans to the M−OH2 bond is attached to the centrally located P (or As)
9−
12−
and Zn atom of the [B-P(or As)W9O34
]
and [B-ZnW9O34
]
ions.
However, the crystallographic data show that the P (or As) atoms are all
about the same distance from the center of inversion in these anions,
which is located in the plane of the four metal atoms in the waist of the
structures (2.81–2.94 Å). Likewise, in the [WZn{M(OH2)}2(B-
ZnW9O34)2]12− series of anions the Zn is about the same distance from
the center of inversion (2.91–2.94 Å). This means that the distance from
the M(II) to the oxygen atom trans to the coordinated H2O, i.e. M−O(Z),
where Z = P, As or Zn, will be necessarily longer in the [M4(OH2)2(B-
PW9O34)2]10− series of anions compared to those in the [WZn{M
(OH2)}2(B-ZnW9O34)2]12− series. In order to maintain the bond valence
sum around the M(II) atoms, the M−OH2 distance is thus affected, by
shortening, thereby making it more difficult to lose coordinated water
and replace it by an allylic alcohol. The effect of changing the Zn(II) to a
P(V) or As(V) is better seen by comparing the {M–O(Z)} – {M–OH2}
distance in the anions. In the [M4(OH2)2(B-PW9O34)2]10− series of an-
ions this value is positive, while in the [WZn{M(OH2)}2(B-
ZnW9O34)2]12− and related series of anions (Table 6) this value is ne-
Although the errors in the crystallographic data are in some cases
large, the results, with the exception of the Co(II)-containing poly-
oxotungstate anion are consistent with the observed catalytic activities;
that is, bonds longer than expected are weaker, and thus will undergo
H2O-allylic alcohol exchange more readily, and vice versa. Note that all
the 1st row transition metal M(II) ions (where M = Mn, Co, Ni, Cu and
Zn) are normally labile. The Jahn-Teller distorted Cu-OH2 bond is of
course long (and hence weak), the Ni-OH2 bond is normal in length, and
the Mn-OH2 and Zn-OH2 bonds are short (and thus strong). The excep-
tion is the Co-OH2 bond, which is long, but this has a large experimental
error. It may be that a redetermination of the structure of this compound
would show a much shorter bond than previously reported. We note that
in the original crystallographic paper Evans et al. did report that the H2O
ligands in [Co4(OH2)2(B-PW9O34)2]10− could not be replaced with strong
donors, such as pyridine and anions such as SCN− [33]. However, Zhang
et al. have reported that in acetonitrile solution the water ligands in
[Co4(OH2)2(B-PW9O34)2]10− can be replaced by strong donors such as N-
methylimidazole, and the three N-oxides, 4-picoline N-oxide, 4-cyano-
pyridine N-oxide and N-methylmorpholine N-oxide [47]. Note that the
bond valences (defined as cation charge/coordination) for the five MII-
(OH2) bonds of the [M4(OH2)2(B-PW9O34)2]10− series of ions (Table 6),
gative, showing the lengthening of the bond trans to
M in the
[M4(OH2)2(B-PW9O34)2]10− series of anions and reduction in the length
of the M−OH2 bond. This results in a reduction in catalytic activity for
the [M4(OH2)2(B-PW9O34)2]10− series of anions. Finally, IR studies of the
type performed on [Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7−, as described
above, showed that W(O2) formation easily occurred for [Ni4(OH2)2(B-
PW9O34)2]10− and [Zn4(OH2)2(B-PW9O34)2]10− with bands at 827 and
826 cm−1, respectively. The former acts as a catalyst under the reaction
conditions (Table 5), while the latter does not. Thus peroxo formation is
which is + 2/6 = 0.33 for
a standard M(II)–OH2 bond in an [M
(OH2)6]2+ ion, show that the bonds are very strong for Zn(II) and Mn(II)
(both with BV > 0.33), about as expected for Ni(II) (BV ~ 0.33) and
quite weak for Cu(II) (BV < 0.33). This parallels the catalytic activities
Table 6
Relevant structural data concerning the M−OH2 bonds in the [M4(OH2)2(B-PW9O34)2]10− ions, where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) (for the last PV is
replaced by AsV), [Fe4III(OH2)2(B-PW9O34)2]6−, the [WZn{M(OH2)}2(B-ZnW9O34)2]n− ions, where M = Zn(II) and Rh(III), and [WCu{Cu(OH2)}2(B-ZnW9O34)2]12−
,
together with data on the M−OH2 distances in the Tutton salts, (NH4)2[M(OH2)6](SO4)2, and bond valences (BV) of the M−OH2 bonds in the hetero-
polyoxotungstate anions.
Anion
M−OH2 (Å)
M−O(Z) (Å)a
Diff. (Å)b
Av. MII-OH2 for T.S. (Å)c
BV of MII-OH2 bond
[Mn4(OH2)2(PW9O34)2]10−
2.15(2)
2.15(4)
2.05(2)
2.34(2)
2.04(3)
2.002(9)
2.17(2)
2.14e
2.27(1)
2.17(3)
2.14(2)
2.55(1)
2.18(2)
2.169(2)
2.07(2)
2.09e
0.12
2.18(2)
0.38
0.29
0.34
0.17
0.40
–
[Co4(OH2)2(PW9O34)2]10−
0.02
2.09(2)
[Ni4(OH2)2(PW9O34)2]10−
0.09
2.06(1)
[Cu4(OH2)2(PW9O34)2]10−
[Zn4(OH2)2(AsW9O34)2]10−
[Fe4III(OH2)2(PW9O34)2]6−
0.21
2.222(1)d
0.14
2.09(2)
0.17
–
–
–
–
[WZn{Zn(OH2)}2(ZnW9O34)2]12−
[WZn{RhIII(OH2)}2(ZnW9O34)2]10−
[WCu{Cu(OH2)}2(ZnW9O34)2]12−
−0.10
−0.05
−0.08
0.28
–
2.08(2)
2.00(2)
0.34
a
b
c
d
e
n−
Z is the heteroatom in the [ZW9O34
]
anion, i.e. P(V), As(V) or Zn(II).
Diff. = {M−O(Z)} – {M−OH2} distance. These two bonds are trans to each other.
T.S. is the equivalent Tutton salt, (NH4)2[M(OH2)6](SO4)2. The data are taken from [50].
Long M−OH2 bond in Jahn-Teller-distorted structure rather than the average M−OH2 bond, which is 2.09(11) Å for the Cu(II)-containing Tutton salt.
No standard deviations provided in original manuscript [8].
11

P.H. Abram, et al.
InorganicaChimicaActa499(2020)119178
not an impediment to the reaction mechanism in this case, with allylic
alcohol coordination to M(II) the actual determining step in the reaction.
References
[1] C.L. Hill, Comprehensive Coordination Chemistry II: From Biology to
Nanotechnology, Elsevier, Oxford, 2004, pp. 679–759.
4. Conclusions
[2] R. Neumann, Transition Metals for Organic Synthesis, second ed., Wiley-VCH,
Weinheim, 2004, pp. 415–426.
[3] I.V. Kozhevnikov (Ed.), Catalysis by Polyoxometalates, Wiley, Chichester, England,
2002Vol. 2.
The epoxidation of a range of allylic alcohols using the three-site
[Ni3(OH2)3(B-PW9O34){WO5(H2O)}]7− and [Co3(OH2)6(A-PW9O34)2]12−
ions as catalysts with aqueous H2O2 as the oxidant has been studied and
compared with previous data on the [WZn{Mn(OH2)}2(B-ZnW9O34)2]12−
ion. The Co3-containing anion proved to be a more active catalyst under
identical reaction conditions, while the Ni3-containing anion was less ac-
tive than the two-site system. Both the Ni3- and Co3-containing anions
operate by an identical reaction mechanism to that previously established
for [WZn{Mn(OH2)}2(B-ZnW9O34)2]12−, except with three active M(II)
sites rather than two M(II) sites. Kinetic studies of the epoxidation of 3-
methyl-2-buten-1-ol by aqueous H2O2 showed the presence of induction
steps which could be interpreted in terms of an autocatalysis mechanism.
The mechanism involves coordination of an allylic alcohol to a M(II),
along with formation of an adjacent W(O2) from a W]O. Subsequent
transfer of the O from the W(O2) to the allylic alcohol generates the hy-
droxy epoxide. The mechanism involves some twenty-one steps for the
three M(II) sites, with the autocatalysis arising because coordination of the
product hydroxy epoxide competes with coordination of the allylic alcohol
and must be replaced by the latter for reaction to occur, while W(O2)
formation occurs much more readily when the M(II) is coordinated by
hydroxy epoxide than by allylic alcohol.
[4] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199–218.
[5] C.L. Hill, C.M. Prosser-McCartha, Coordination Chem. Rev. 143 (1995) 407–455.
[6] C.M. Tourné, G.F. Tourné, F. Zonnevijlle, J. Chem. Soc., Dalton Trans. (1991)
143–155.
[7] R. Neumann, A.M. Khenkin, Inorg. Chem. 34 (1995) 5753–5760.
[8] R. Neumann, A.M. Khenkin, J. Mol. Catal. A: Chem. 114 (1996) 169–180.
[9] R. Neumann, M. Gara, J. Am. Chem. Soc. 116 (1994) 5509–5510.
[10] R. Neumann, M. Gara, J. Am. Chem. Soc. 117 (1995) 5066–5074.
[11] W. Adam, P.L. Alsters, R. Neumann, C.R. Saha-Möller, D. Sloboda-Rozner, R. Zhang,
J. Org. Chem. 68 (2003) 1721–1728.
[12] R. Neumann, D. Juwiler, Tetrahedron 52 (1996) 8781–8788.
[13] R. Neumann, A.M. Khenkin, D. Juwiler, H. Miller, M. Gara, J. Mol. Catal. A: Chem.
117 (1997) 169–183.
[14] R.C. Burns, P.H. Abram, Trans. Metal Chem. 43 (2018) 705–717.
[15] A.J. Stapleton, M.E. Sloan, N.J. Napper, R.C. Burns, Dalton Trans. 38 (2009)
9603–9615.
[16] R. Ben-Daniel, A.M. Khenkin, R. Neumann, Chem. Eur. J. 6 (2000) 3722–3728.
[17] X. Zhang, Q. Chen, D.C. Duncan, R.J. Lachicotte, C.L. Hill, Inorg. Chem. 36 (1997)
4381–4386.
[18] X. Zhang, Q. Chen, D.C. Duncan, C.F. Campana, C.L. Hill, Inorg. Chem. 36 (1997)
4208–4215.
[19] A.M. Khenkin, C.L. Hill, Mendeleev Commun. (1993) 140–141.
[20] M. Bösing, A. Nöh, I. Loose, B. Krebs, J. Am. Chem. Soc. 120 (1998) 7252–7259.
[21] K.A. Jørgensen, Chem. Rev. 89 (1989) 431–458.
[22] W. Adam, T. Wirth, Acc. Chem. Res. 32 (1999) 703–710.
[23] P.A. Bartlett, Tetrahedron 36 (1980) 2–72.
Studies of the related [M4(OH2)2(B-PW9O34)2]10− series of anions,
where M = Mn(II), Co(II), Ni(II), Cu(II) and Zn(II), indicates that all are
much less catalytically active than the [WZn{M(OH2)}2(B-
[24] P. Besse, H. Veschambre, Tetrahedron 50 (1994) 8885–8927.
[25] D. Sloboda-Rozner, R. Neumann, Green Chem. 8 (2006) 679–681.
[26] Y. Qiao, Z. Hou, Curr. Org. Chem. 13 (2009) 1347–1365.
[27] S. Omwoma, C.T. Gore, Y. Ji, C. Hu, Coord. Chem. Rev. 286 (2015) 17–29.
[28] S.-Y. Dou, R. Wang, Curr. Org. Chem. 21 (2017) 1019–1036.
[29] C.J. Gómez-Garcia, E. Coronado, L. Ouahab, Angew. Chem. Int. Ed. Engl. 31 (1992)
649–651.
ZnW9O34)2]12− series of anions, and this is attributed to the replace-
12−
ment of [B-ZnW9O34
]
by [B-PW9O34
]
9−, and more specifically as
the replacement of Zn(II) by P(V) affects the adjoining M−O(Z) dis-
tance, which in turn affects the (externally directed) trans M−OH2
distance by shortening, making these anions less catalytically active.
This, of course, assumes that an identical reaction mechanism also
operates for these [M4(OH2)2(B-PW9O34)2]10− series of anions, which is
most likely.
[30] J.M. Clemente-Juan, E. Coronado, J.R. Galán-Mascarós, C.J. Gómez-Garcia, Inorg.
Chem. 38 (1999) 55–63.
[31] W.H. Knoth, P.J. Domaille, R.D. Farlee, Organometallics 4 (1985) 62–68.
[32] T.J.R. Weakley, H.T. Evans Jr, J.S. Showell, G.F. Tourné, C.M. Tourné, J. Chem.
Soc., Chem. Commun. (1973) 139–140.
[33] H.T. Evans, C.M. Tourné, G.F. Tourné, T.J.R. Weakley, J. Chem. Soc., Dalton Trans.
(1986) 2699–2705.
[34] C.J. Gómez-Garcia, E. Coronado, P. Gómez-Romero, N. Casañ-Pastor, Inorg. Chem.
32 (1993) 3378–3381.
[35] T.J.R. Weakley, R.G. Finke, Inorg. Chem. 29 (1990) 1235–1241.
[36] R.G. Finke, M. Droege, J.R. Hutchinson, O. Gansow, J. Am. Chem. Soc. 103 (1981)
1587–1589.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
[37] R.D. Mair, R.T. Hall, in: I.M. Kolthoff, P.J. Elving (Eds.) Treatise on Analytical
Chemistry. Part II. Analytical Chemistry of Inorganic and Organic Compounds,
Wiley-Interscience, 1971, Vol. 14, p 411.
[38] http://www.jplusconsulting.com/products.
[39] R. Massart, R. Contant, J.M. Fruchart, J.P. Ciabrini, M. Fournier, Inorg. Chem. 16
(1977) 2916–2921.
Acknowledgments
[40] R.G. Finke, M.W. Droege, P.M. Domaille, Inorg. Chem. 26 (1987) 3886–3896.
[41] W.H. Knoth, P.J. Domaille, R.L. Harlow, Inorg. Chem. 25 (1986) 1577–1584.
[42] P.H. Abram, Ph. D. Thesis, The University of Newcastle, NSW, Australia, 2014.
[43] C. Dean, R. Duncan, C. Chambers, E. Hecht, C.L. Hill, J. Am. Chem. Soc. 117 (1995)
681–691.
P. H. A. would like to acknowledge a scholarship and financial
support from the Ministry of Research, Technology and Higher
Education, Indonesia (DIKTI-Scholarship), as well as a UNRS and OPRS
from the University of Newcastle. Prof. M. Maeder is thanked for access
to the ReactLab programs.
[44] R.B. Jordan, Reaction Mechanisms of Inorganic and Organometallic Systems,
Oxford University Press, New York, 1991.
[45] P.C. Harris, T.E. Moore, Inorg. Chem. 7 (1968) 656–660.
[46] R.F. Pasternack, R.A. Plane, Inorg. Chem. 4 (1965) 1171–1173.
[47] X. Zhang, K. Sasaki, C.L. Hill, J. Am. Chem. Soc. 118 (1996) 4809–4816.
[48] IUPAC Stability Constants Database, Academic Software, 2006. See the section on
the “K v T Temp. Dependence”, under “Definitions”.
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
[49] C.E. Housecroft, A.G. Sharpe, Inorganic Chemistry, fourth ed., Pearson Education
Ltd, England, 2012 Sec. 20.12, pp 709-710.
Appendix A. Supplementary data
[50] F.A. Cotton, L.M. Daniels, C.A. Murillo, J.F. Quesada, Inorg. Chem. 32 (1993)
4861–4867.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.ica.2019.119178.
[51] I.D. Brown, D. Altermatt, Acta Crystallogr. B 41 (1985) 244–247.
[52] R.D. Shannon, Acta Crystallogr. A 32 (1976) 751–767.
12