Interactions of vanadium(iv) with amidoxime ligands: Redox reactivity

Article abstract of DOI:10.1039/c7dt04069e

The use of amidoxime-functionalized polymer fibers as a sorbent for uranium has attracted recent interest for the extraction of uranium from seawater. Vanadium is one of the main competing ions for uranium sorption as V(v) species, however, vanadium is al

Full text of DOI:10.1039/c7dt04069e

Dalton
Transactions
View Article Online
View Journal
PAPER
Interactions of vanadium(IV) with amidoxime
ligands: redox reactivity†
Cite this: DOI: 10.1039/c7dt04069e
a,b
a,b
a,b,c
b
c
B. F. Parker,
J. Arnold
S. Hohloch,
J. R. Pankhurst,
Z. Zhang, J. B. Love,
a,b
b
*
and L. Rao
*
The use of amidoxime-functionalized polymer fibers as a sorbent for uranium has attracted recent inter-
est for the extraction of uranium from seawater. Vanadium is one of the main competing ions for uranium
sorption as V(V) species, however, vanadium is also present as V(IV) in seawater. In the present study, the
interactions of V(IV) with amidoxime and similar ligands were explored. Attempts were made to synthesize
V(IV) complexes of glutaroimide-dioxime, a molecular analogue of polymer sorbents. However, V(IV) was
found to react irreversibly with glutaroimide-dioxime and other oxime groups, oxidizing to the V(V) oxi-
dation state. We have explored the reactions and propose mechanisms, as well as characterized the redox
behavior of the vanadium–glutaroimide-dioxime complex.
Received 30th October 2017,
Accepted 29th March 2018
DOI: 10.1039/c7dt04069e
rsc.li/dalton
V(IV) by aerated water; nonetheless, approximately 10–15% of
vanadium is present in the V(IV) state. This naturally occurring
Introduction
7
The world’s oceans contain over 4.5 billion tons of uranium, mixture of oxidation states, which is subject to seasonal and
over 1000 times more than all known terrestrial sources, and it geographical variation, is a consequence of biological uptake
7
,9,10
is therefore a promising source of uranium for nuclear and use of vanadium in marine environments.
Vanadium
1
energy. The extraction technique receiving the most interest is an essential element for many organisms, although typically
makes use of polymer sorbents functionalized with amidoxime in very small amounts. In vivo, vanadium is present in several
groups, with significant improvements being made over the chemical environments in oxidation states ranging from V(III)–
past 6 years that have more than doubled the sorbents’ V(V), often V(IV), and when marine organisms die they release
1
,2
11
uranium capacity.
Solution-state studies have generally vanadium in those varying oxidation states.
involved several molecular analogues of these groups (Fig. 1).
Although the coordination chemistry of V(V) has been
A
5
Glutaroimide-dioxime (L ) has received the most attention due studied recently with glutaroimide-dioxime, as well as its
3
to its high affinity for uranium. It does have a significant
drawback, which is its relatively low selectivity, forming strong
4
complexes with several other metals including iron, copper,
and notably vanadium, forming a very unusual, strong non-
5
,6
oxido complex in that case (Fig. 1).
Vanadium is present in seawater at approximately 0.5–2
7
ppb (10–40 nM), which is similar to that of uranium at 3.3
8
ppb (13 nM), and indeed is the second most abundant tran-
8
,9
sition metal in seawater after molybdenum.
It is mainly
present in the V(V) oxidation state due to slow oxidation of
a
Department of Chemistry, University of California – Berkeley, Berkeley, CA 94720,
USA. E-mail: arnold@berkeley.edu
b
Chemical Sciences Division, Lawrence Berkeley National Laboratory, 1 Cyclotron
Road, Berkeley, CA 94720, USA
c
EaStCHEM School of Chemistry, University of Edinburgh, Joseph Black Building,
The King’s Buildings, Edinburgh, EH9 3FJ, UK
†
Electronic supplementary information (ESI) available: NMR spectra, electro-
chemistry data, and X-ray crystallographic tables. CCDC 1582633. For ESI and
crystallographic data in CIF or other electronic format see DOI: 10.1039/
c7dt04069e
Fig. 1 Ligands investigated for V(IV) reactivity and their known V(V)
complexes.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
6
,12
interactions with polymer sorbents,
the analogous chem- at room temperature, during which the mixture turned dark
istry of V(IV) has not yet been investigated. Our initial work brown and the solids completely dissolved. The mixture was
sought to study the interactions of V(IV) with seawater-relevant transferred to a glass Petri dish and warmed at 40 °C over-
amidoxime ligands (Fig. 1) in the same manner as U(VI), Fe(III), night, during which all the solvent evaporated to leave a dark
3
,5,13
V(V) and other metals.
In this way, we would understand brown solid. The solid was extracted with dichloromethane
the role of V(IV) in determining uranium sorption selectivity (5 × 10 mL), dried over magnesium sulphate, filtered, and the
attempt to synthesize the non-oxido vanadium in the V(IV) oxi- solvent removed to give a brown oil. This was crystallized from
dation state in the same way as V(V). However, while doing so dichloromethane (20 mL) layered with an equal volume of
4
we discovered that V(IV) reacts with amidoxime ligands in con- THF. Large dark brown-black blocks of [NEt ][1] formed over
ditions relevant to seawater, oxidizing to V(V) in the process. several days. It is very soluble in chloroform, dichloromethane,
This type of reactivity has previously been observed with hydro- alcohols, and most other polar solvents, has trace solubility in
xamic acids, a similar class of ligands with N–OH groups, THF, and is insoluble in all low-polarity solvents. The same
where the O atom is abstracted upon complexation with V(III), product can be obtained in similar yield by using sodium
V(IV), and Mo(V), forming the corresponding amide and oxi- orthovanadate instead of ammonium metavanadate, adjusting
1
4
1
dized metal-oxo complexes. One previous report of V(IV) the pH to 6–8. Yield: 230 mg (0.50 mmol, 63%). H NMR
3
reacting with salicylaldoxime leads to the V(V)-salicylaldoxime (CDCl
complex as the major product along with smaller amounts of 1.95 ppm (quintet, 4H, J
other complexes arising from coupled small organic mole- (t, 8H, JHH = 6.3 Hz, CH
cules. Other low oxidation state metal ions have also been Hz, N(CH
observed to react similarly, including Mo(II) and U(IV),
although in seawater uranyl is present exclusively as uranyl(VI) CCH
carbonate complexes.
In our attempts to prepare and characterize V(IV) complexes 739 ppm.
of the model oxime ligands and reagents shown in Fig. 1, we Glutaramidoxime (L )
3
): 1.23 ppm (t, 12H,
J
HH = 7.4 Hz, N(CH
2 3 4
CH ) ),
3
= 6.3 Hz, CH CH CH ), 2.62 ppm
HH
2
2
2
3
3
2
CH
2
CH
2
), 3.18 ppm, (t, 8H, JHH = 7.4
, 126 MHz): 7.7 ppm (4C,
1
5
13
2
CH
3
)
4
); C NMR (CDCl
3
1
6
17
N(CH CH ) ), 21.4 ppm (2C, CCH CH CH C), 21.6 ppm (4C,
2
3 4
2
2
2
1
2
CH
2
2 2 3 4
CH C), 52.5 ppm (4C, t, JCN = 3.0 Hz, N(CH CH ) ),
18
51
154.0 ppm (4C, CvN–O–V); V NMR (CDCl
3
, 132 MHz):
B 19
20
and vanadyl toluenesulfonate
found that they invariably reacted to form V(V) complexes were synthesized according to literature procedures.
A
under the conditions tested. Cyclic voltammetry was per- Glutaroimide-dioxime (L ) was synthesized according to litera-
formed on the 2 : 1 glutaroimide-dioxime V(V) complex 1 and, ture procedure and recrystallized twice from methanol before
2
1
surprisingly, it was found to undergo reduction at a strongly use. Vanadyl (oxovanadium(IV)) solutions were prepared by
negative potential. We also attempted to synthesize a V(IV) reducing a solution of acidified V O with excess hydroxyl-
2
5
2
complex in non-aqueous solvents both directly and through amine hydrochloride. V(IV) was precipitated as VO by the
chemical reduction of the V(V) complex, however, no stable addition of sodium hydroxide, then redissolved in hydro-
complexes could be isolated, suggesting a V(IV) analogue may chloric acid. The concentration of V(IV) was determined by
4
be inherently unstable towards ligand reactivity or dispropor- titration with a solution of KMnO of known concentration.
tionation to V(V) and reduced species.
All other chemicals were obtained from commercial sources
and used as received.
Cyclic voltammetry
Methods
Synthesis
Cyclic voltammetry experiments were performed on a Gamry
Instruments Reference 600 potentiostat and data acquired
Acetamidoxime (HAO). Acetamidoxime (HAO) was syn- with Gamry Framework software. Experiments in acetonitrile
thesized by stirring equal volumes of a 50% aqueous hydroxyl- and DCM were performed in an inert atmosphere glovebox
amine solution, acetonitrile, and ethanol (5 mL each) at room using dry solvent with 0.1 M [NBu ][PF ] electrolyte, with a Pt
4 6
temperature for 24 hours, followed by warming to 40 °C in a disc working electrode, Pt gauze counter electrode and Pt wire
glass Petri dish until all the volatiles evaporated and a white quasi-reference electrode. Experiments in water and methanol
solid remained. It was then dissolved in water (5 mL) and were performed with a Pt disc working electrode, Pt wire
evaporated again to remove any remaining acetonitrile. The counter electrode and Ag/AgCl reference electrode, with 0.2 M
resulting solid was redissolved in approximately 10 mL of KCl as the electrolyte in water and 0.2 M [NEt ][Cl] in metha-
4
1
water and its concentration was determined by H NMR nol. Ferrocene was added as an internal standard to non-
+
spectroscopy and protonation titration.
Tetraethylammonium vanadium(V)
dioxime) ([NEt
glutaroimide-dioxime) ([NEt
taroimide-dioxime (290 mg, 2.1 mmol), ammonium metavana-
date (117 mg, 1.0 mmol), and tetraethylammonium bromide An internal standard of approximately 1 mM tert-butanol (the
168 mg, 0.80 mmol) to a mixture of water (10 mL) and metha- exact concentration calculated per experiment based on stock
nol (5 mL). The mixture was stirred vigorously for 30 minutes solution volumes) in D O/H O mixtures was used as the
aqueous solutions (Fc /Fc at 0 V) and potassium ferricyanide
bis(glutaroimide- was added as an internal standard to aqueous solutions
3
−
4−
4
][1]). Tetraethylammonium vanadium(V) bis ([Fe(CN)
][1]) was prepared by adding glu-
6
]
/[Fe(CN)
6
]
at +0.04 V) to reference experiments.
(
4
NMR
(
2
2
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
1
solvent for quantitative experiments. H NMR spectra in water Cyclic voltammetry of complex 1
or H O/D O mixtures were acquired on a Bruker AV-500 instru-
2
2
In order to solubilize 1 in non-aqueous solvents to perform
nonaqueous reductions, a cation exchange reaction was per-
formed to synthesize the organic-soluble tetraethylammonium
complex. The extraction of 1 in the presence of tetralkylammo-
nium salts from water into immiscible organic solvents failed
and all compounds remained in the aqueous phase. However,
the tetraethylammonium salt of 1 was successfully isolated by
ment (500 MHz) with a WATERGATE solvent suppression pulse
5
1
sequence.
V NMR spectra were acquired on a Bruker
DRX-500 instrument (132 MHz).
Results and discussion
Attempts to prepare a reduced L -vanadium(IV) complex
synthesizing
bromide in a water/methanol mixture, evaporating the solution
to dryness, then extracting into dichloromethane. [NEt ][1] was
1 in the presence of tetraethylammonium
A
4
In order to study the complexation behaviour of vanadium(IV)
obtained as very dark brown-black block crystals by layering a
A
A
with L , we initially attempted a direct synthesis from L and
aqueous vanadyl (oxovanadium(IV)) under a variety of reaction
conditions, varying the pH over the range of approximately
dichloromethane solution with THF. [NEt ][1] was used in all
4
cyclic voltammetry experiments (Fig. 2) due to its exact stoi-
chiometry and higher purity, unlike the sodium salt isolated
4
–10, as well as the stoichiometry, rate, and order of addition
from water which may contain excess reagents and salts.
of reagents. However, in water, only known V(V) species were
c
p
In water (pH 7), one irreversible reduction occurs at E
5
identified as products: 1, 2, and mixtures of free vanadates,
+
−
920 mV vs. Fc /Fc, which we assign to the reduction of V(V) to
with the exact speciation of the latter dependent on concen-
V(IV), followed by subsequent reaction. This is significantly
2
2
trations and final pH. Regardless of the conditions used, V(V)
more negative than free V(V)/V(IV) redox potentials (−600 mV
complexes 1 and 2 and free V(V) species were invariably pro-
V
+
IV 2+
V
3−
for [V O ] /[V O]
[
in acid or −400 mV for [V O ]
/
2
4
5
1
duced, confirmed by V NMR and visually, with the reaction
changing to the amber-brown colour of 1. To preclude the role
of oxygen in these reactions, water degassed by boiling or spar-
ging with nitrogen was used, with no change in outcome. This
indicates that V(IV) oxidation in water is not caused by oxygen
but rather by the ligand itself.
IV
2−
23
H V O ] in strong base, relative to ferrocene), consistent
2 4
with the vanadium atom being more electron rich due to
binding of the strongly electron-donating amidoxime ligand.
Solid was observed plated onto the electrode, likely VO since
2
the V(IV) complex that is formed is not stable. Decomposition
could conceivably proceed through ligand dissociation and
subsequent vanadium hydrolysis to form VO2.
When the same reaction was performed in dry, degassed
methanol, the reaction mixture turned a deep red colour
within seconds, followed by a slow change to dark amber, indi-
cating the formation of 1. If the reaction is carried out under
argon and the solvent removed rapidly just after the onset of
the reaction, colourless crystals of toluenesulphonic acid form,
Cyclic voltammetry was also performed in non-aqueous sol-
vents to investigate if water precludes the isolation of a V(IV)
complex. In methanol, the same irreversible behaviour was
observed, with an irreversible reduction at a similar potential
c
p
of E −1000 mV. The similarity of the electrochemical behav-
and
a deep red oil containing polyoxovanadates (V(V))
iour and chemical reactions (see above) suggests that a similar
2
2
remains. When water is added to the red solution or oil, the
vanadium clusters hydrolyse immediately and form the V(V)
complex 1 with its characteristic dark amber colour. If the reac-
tion is performed in basic methanol (sodium methoxide) or
methanol with a drop of added water, the red colour was not
observed, instead forming 1 directly, due to hydrolysis of
1
0
vanadium clusters.
The reaction of V(IV) with glutaroimide-dioxime was also
tested in aprotic solvents. Vanadyl acetylacetonate was dis-
A
solved in dry acetonitrile and excess L . Due to the low solubi-
A
lity of L , the reaction proceeded slowly, and the solution first
turned green and then amber, also forming a brown-black pre-
cipitate. Small amounts of 1 were present in solution (deter-
5
1
mined by V NMR spectroscopy), and the insoluble black
product could not be extracted into any solvent including
methanol and water, likely caused by some irreversible
A
decomposition process. Disproportionation of
a V(IV)-L
species formed initially could be occurring to form insoluble,
reduced metal oxides. Based on these results, cyclic voltamme-
try was performed to determine the accessibility of the V(IV)
oxidation state from 1 and establish possible conditions for
this reaction.
Fig. 2 Cyclic voltammograms of 1 in several solvents, all measured at a
scan rate of 100 mV s− against a Pt working electrode.
1
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
2
7
process may be taking place in both cases, with the V(IV) vents (discussed below). Since the ligands are strongly elec-
complex that is transiently formed reacting with the ligand tron-donating, such that they displace all oxido groups on the
and then dissociating and forming VO
polyoxovanadates.
2
or kinetically inert metal upon coordination, they are poor ligands with respect to
stabilizing the lower oxidation state on vanadium.
Different electrochemical behaviour was observed in the
The unusual redox behaviour of 1 and its reluctance to be
aprotic solvents acetonitrile and dichloromethane. In the reduced is not a direct consequence of the rare non-oxo
c
p
former, an irreversible reduction occurs at E −1.5 V, a more environment, as several other air-stable non-oxo V(IV) com-
negative potential than in water or methanol. The initial plexes are known. One of them is Amavadin, a natural product
reduction likely occurs in the same manner as in the protic that is found in the hyperaccumulating fungus Amanita mus-
2
8
solvents, but without a proton source for the ligand to dis- caria (Fig. 3, left). In contrast to 1, vanadium is present exclu-
sociate completely and the V(IV) complex to hydrolyse, the sively in the V(IV) oxidation state in naturally-occurring
acetonitrile coordinates then reacts, leading to a V(IV) complex Amavadin as well as synthetic derivatives, and although the
that remains soluble. An oxidation occurs on the return scan, V(V) form is readily accessible by chemical oxidizing agents
c
but is a different process and not a reverse wave (also at E
such as hydrogen peroxide, it will slowly revert to the stable
p
−
1.5 V with no peak-to-peak separation), likely this new V(IV) V(IV) oxidation state over time with no apparent decompo-
1
0
complex reoxidizing to V(V), with other oxidations of reacted sition. In some organic solvents, however, the V(V) form is
ligands or other complexes observed at more positive poten- favoured, turning red, indicating an easily accessible redox
tials as well. In dichloromethane, the same reduction occurs at couple. This is notable because the ligand in Amavadin, 2,2′-
c
p
E −1.41 V (E1/2 −1.32 V), but shows reversible behaviour, with (hydroxyimino)dipropionate, contains
a
central N–OH
a peak-to-peak separation of 190 mV. Although this value is hydroxylamine group that is deprotonated in the complex,
relatively large, the square root law is linear (see ESI, Fig. S1 similar to the four N–O groups in 1, yet it is inert when directly
and S2†) and the large separation is therefore attributed to bonded to V(IV). A limited number of other non-oxido V(IV)
uncompensated resistance from the solvent. We attribute this complexes are known, including a complex with a cis-inositol
process to the same initial reduction as in acetonitrile. derived ligand and other similar compounds (Fig. 3, right)
2
6,29
However, the V(IV) complex is stable in DCM, since the solvent that are also unstable when oxidized to V(V).
is noncoordinating and generally less reactive, making it
reversible on the CV timescale.
We attempted the chemical reduction of 1 in both aqueous
and nonaqueous solvents; to target a reduction at E1/2
Vanadium(IV) reactivity with amidoximes and oximes
A
(
[
−1.32 V) we used cobaltocene as a reducing agent (−1.33 V for In addition to L (glutaroimide-dioxime), the reactions of
+
Cp
2 2
Co] /[Cp
Co]) which would be a mild reaction if it were other amidoximes and oximes with V(IV) were also monitored
2
4
1
51
irreversible. This was attempted in both acetonitrile and by H and V NMR spectroscopy. We tested glutaramidoxime
B 19
30
DCM, and in both cases, no reaction was observed initially, (amidoxime B, L ) and acetamidoxime (HAO) as models for
but within an hour an intractable black solid started to form other functional groups on uranium sorbents. Acetone oxime
in the same manner as the direct reaction with vanadyl acetyl- was also tested as it is the simplest ketoxime which should
acetonate (above), and no products could be isolated, implying react to form a single product, acetone (substrates shown in
that a possible V(IV) complex is inherently unstable in these Fig. 1). For these substrates, only free V(V) species were
conditions, with this decomposition in aprotic solvents being observed in the products instead of metal complexes, as the
caused by a lack of hydrolysis reaction pathways.
substrates and probable products are expected to be poor
6
,19
51
The reduction of 1 in protic solvents was also attempted ligands (see ESI, Fig. S6–S11†).
The presence of any
V
with a variety of reducing agents in water (pH 6–8) or water– NMR signals is indicative of oxidation from V(IV) to V(V) since
2
2
2
methanol mixtures before performing cyclic voltammetry V(IV) is NMR-silent. Small amounts of VO precipitation is
B
experiments. Several common reducing agents were tested, observed, especially in the reactions of L , although no
including hydroxylamine, hydrazine, sodium thiosulfate, and
zinc metal, spanning reduction potentials between −0.29 V
2
4,25
and −1.16 V vs. ferrocene,
and in all cases, no reaction was
observed. Although any V(IV) products would be paramagnetic
1
and not necessarily observable by NMR spectroscopy, all
H
5
1
and V NMR spectra as well as absorption spectra remained
unchanged. This lack of reactivity suggests that the presence of
the two amidoxime ligands on vanadium greatly stabilizes the
complex against reduction compared to other V(V) complexes,
which have reduction potentials in the range of +0.6 V vs.
ferrocene (for [VO] /[VO ] ), +0.1 to +0.75 V for non-oxido cis-
2
inositol complexes or −0.25 to +0.49 V for Amavadin deriva-
2
+
+
2
6
tives, with the lower end of the range occurring in organic sol- Fig. 3 Stable non-oxido V(IV) complexes.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
Table 1 Product mixtures of V(IV) and amidoxime and oxime substrates
its reactivity, the colloidal material was isolated as a dark blue
solid by centrifugation and decanting of the solution, after
allowing a reaction involving acetone oxime to proceed for one
hour. When a fresh solution of oxime was added to the solid
the reaction proceeded, producing the expected amount of
acetone, almost 0.5 equivalents.
Substrate (S)
[S] [V] : [S] Unreacted [S] [Product] [Decomp.]
1
6
2
6 : 1
1 : 1
1 : 3
0
2.4
11.0
0.85
3.2
9.0
B
L
0
2
3 : 1
0
6.1
32
1.2
2.4
1.8
0.2
0.01
4.3
2
VO prepared in the absence of acetone oxime was also
HAO
12 1 : 2
3
tested for its reactivity, either as a wet solid after isolation by
centrifugation and decanting of the solution, or after it had
been heated to dryness. Both forms did react, although much
smaller amounts of acetone were produced than when the VO2
6
1 : 6
2
3 : 1
1.1
9.3
32
0.97
5.6
2.9
Acetone oxime 12 1 : 2
1 : 6
3
6
was formed in the presence of acetone oxime. Solid VO
still present in the reactions, especially when the VO
2
was
was
2
+
Note: [VO ] = 6 mM for all reactions. All concentrations are in mM.
Product is acetamide for HAO, acetone for acetone oxime.
2
Decomposition (decomp.) is all other signals that are not product or dried (Table 2). The large difference in reactivity can be attribu-
B
unreacted substrate. Precipitate was present in the reactions of L and
ted to the nature of the colloid and its surface. When the
small amounts were present in reactions of HAO.
colloid is formed, acetone oxime or other substrate may act as
a surfactant enhancing its reactivity, while hydrated VO2
formed in the absence of substrate shows much lower reactiv-
ity and dried VO lower still.
When the reaction of L was performed at higher concen-
trations and on a larger scale in an attempt to isolate products,
attempts were made to quantify this. Table 1 shows the pro-
ducts produced at different vanadium/oxime ratios.
2
B
For these substrates, the formation of a heterogeneous
mixture by the addition of base was necessary for any reaction
to occur. This could indicate that the formation of a hydrated
or colloidal vanadium(IV) oxide VO or VO(OH) is needed and
2
VO precipitated which slowly settled over one week, during
which the solution turned from blue to green to yellow-brown.
The supernatant was then decanted and allowed to slowly
evaporate, upon which brown crystals of glutaramidinium dec-
2
2
3
1
the free vanadyl ion is inert. These reactions transiently turn
a blue-grey colour with the formation of a colloid, as observed
by the Tyndall effect (see ESI Images 1 and 2†) which then
reacts with the substrates (Scheme 1). The heterogeneous
nature of this reaction also explains its slow and variable kine-
tics, taking hours to days to react fully. The reaction mixtures
remain slightly acidic, around pH 4–6, so vanadium oxidation
is not caused by strong base, and the substrates are not reac-
avanadate, [(H N) C(CH ) C(NH ) ] ·[V O ], were obtained.
2
2
2 3
2 2 3
10 28
The structure was confirmed by single crystal X-ray diffraction
see ESI Fig. S14, Table S1†). Decavanadate formation is
(
observed as a consequence of an acidic solution, high
vanadium concentration, and a poor ligand that does not
2
2
interact directly with vanadium at all. Although the amidi-
nium ion was isolated in the crystal structure, hydrolysis can
occur to form the more stable amide, as seen previously in
3
2
tive towards hydrolysis due to low pH.
The formation of the vanadium dioxide colloid in situ
3
4
amidoxime reactivity.
Therefore, in many cases the
appears to be necessary for reactivity of the substrates other
1
A
ammonium ion was observed in the H NMR spectra of these
than L . With multiple vanadium centres in close proximity,
reactions as a distinctive 1 : 1 : 1 triplet at δ = 7.16 ppm due to
H– N coupling ( JHN = 52 Hz), along with the formation of
amides or carboxylic acids.
this can allow for two-electron processes and oxygen atom
abstraction. This type of reactivity is well known with
vanadium oxides, notably in the contact process that is used
1
14
1
3
3
This type of reductive deoxygenation of amidoximes has
been reported previously with transfer hydrogenation reac-
on an industrial scale to produce sulfuric acid. To confirm
3
5
36
tions and enzyme activity. In the latter case, O-substituted
amidoximes were used as prodrugs for relatively reactive and
less bioavailable amidines in vitro and in vivo after deoxygena-
tion by liver enzymes. A somewhat similar deoxygenation reac-
tion of amidoximes and imide-dioximes has been reported
Table 2 Vanadium dioxide reactivity resulting from preparation method
Equivalents of acetone
produced (relative to V)
VO
2
precipitation
With acetone oxime, decanted
Without acetone oxime
Without acetone oxime, dried at 120 °C
0.49
0.16
0.12
Scheme 1 Reaction of acetone oxime (upper) or acetamidoxime
B
(
lower) with V(IV) and base. Similar reactions take place with L . The
2
+
exact composition of the V(IV) colloid is unknown, and the product V(V) Note: All with [acetone oxime] = 20 mM, [VO2] = 20 mM. Ideal = 0.5
species depend on concentration and pH.
equivalents.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
A
with nitrous acid, although in that case the amine portion of Table 3 Product mixtures of V(IV) and glutaroimide-dioxime (L )
the amidoxime also reacts, as the reaction produces nitrous
A
A
A
A
A
3
4,37,38
[L ] V : L ratio Unreacted [L ] [L ] in 1 [L ] in 2 [Decomp.]
oxide and the amide.
Similar reactivity has been
observed with hydroxamic acids, although only when
1
1
18
6 : 1
1 : 2
1 : 3.3
0
0.23
1.0
0
1.7
3.0
0.24
2.0
4.1
0.69
8.6
9.1
2
N-substituted, reacting with reduced V(III), V(IV), Mo(IV), and
1
4
Mo(V). In contrast, stable hydroxamate complexes are known
2
+
with all of these oxidation states without N-substitution. Note: All concentrations are in mM. [VO ] = 6 mM for all reactions.
Limited examples of V(IV) reacting with salicylaldoxime and
Multiple decomposition products ([decomp.]) were observed that could
not be identified or quantified separately. Spectra acquired 6 hours
salicylamidoxime have also been reported, where the V(V)
after mixing.
1
5
complex was isolated in low yield. The low yield reported is
consistent with 2 equivalents of V(IV) reacting with each
ligand, in a one-electron process on the metal and two-electron
process abstracting an oxygen atom from the ligand, resulting
in at least half of the ligand remaining intact.
Glutaroimidoxime was identified as one decomposition product
by comparison with an NMR spectrum of the pure compound
(
see ESI Fig. S4 and S5†) although the major decomposition
product, likely arising from further reaction of glutaroimidox-
ime, was not able to be identified, not matching any known
product or intermediate of hydrolytic decomposition.
A
Vanadium(IV) reactivity with glutaroimide-dioxime (L )
3
2
After exploring the reactivity of V(IV) with simple oximes and
amidoximes, we then looked the reaction of L with V(IV) by
NMR to explain our observations described earlier. Unlike the
other substrates, the reactions remained homogeneous
throughout the reaction and no added base was needed. We
A
The concentrations of products vary depending on the ratio of
reagents used, shown in Table 3. Here we can see that more than
A
one equivalent of vanadyl reacts with one molecule of L , as the
concentration of decomposition products exceeds the vanadium
A
A
attribute this difference in reactivity between L and the other
substrates to the effectiveness of L as a ligand, where a transi-
concentration when excess L is used. Additionally, if excess
A
ligand remains after the V(IV) is oxidized to V(V), it forms the
ent V(IV) complex forms in solution.
complexes 1 and 2, which are the products observed previously.
A
In order to determine if O-atom transfer or other reactivity
was occurring between V(IV) vanadyl and L , quantitative NMR
We propose that the rapid homogeneous reaction of L with
A
vanadyl occurs through the formation of a transient V(IV)
complex followed by reaction with another equivalent of V(IV)
experiments were performed. When the two reagents were
mixed, the reaction mixtures turned dark brown and reactions
started within seconds as observed previously. Spectra were
acquired after approximately 6 hours to allow the reaction
mixture to equilibrate; a sample spectrum is shown below
Fig. 4). In these reactions, the V NMR spectra contained the
known 1 : 1 and 2 : 1 complexes as well as free vanadium when
excess vanadium is present. The initial pH of the vanadium
(
Scheme 2). This effectively amounts to disproportionation of
the vanadium, but rather than being reduced to V(III), the
ligand is reduced instead to form a second equivalent of V(V).
This mechanism is also consistent with the lack of sensitivity
to oxygen (see above), making a free radical mechanism very
5
1
(
A
unlikely. Without complexation to form a transient L -V(IV)
5
complex or the presence of organic radicals, a third-order reac-
tion would be needed, which is not consistent with the rapid
reaction and is relatively rare. There is still more to learn about
the mechanism, reactivity, and unusual complexation behavior
solutions was 3–4 and glutaroimide-dioxime solution 8, and
when mixed the final pH was in the range of 6–7 due to
buffering by the ligand and complexes formed.
1
A
Fig. 4 H NMR spectrum of product mixture of L and V(IV) after
hours. Major decomposition product at δ = 2.91 ppm unidentified.
Scheme 2 Proposed reaction mechanism of glutaroimide-dioxime
with V(IV). The mixture of V(V) product species formed varies depending
on concentrations and pH.
6
A
2+
Conditions. 18 mM L , 6 mM VO
.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
of glutaroimide-dioxime with vanadium, and further studies
are forthcoming.
Acknowledgements
This work is supported by the Fuel Cycle Research and
Development Campaign (FCRD)/Fuel Resources Program,
Office of Nuclear Energy, the U.S. Department of Energy
Conclusions
(
USDOE) under Contract No. DE-AC02-05CH11231 at Lawrence
In this study we have investigated the reactivity taking place Berkeley National Laboratory (LBNL). The collection of NMR
between vanadium(IV) and amidoxime ligands. Rather than data is supported by the UC Berkeley NMR facility, funded in
A
form stable V(IV) complexes with glutaroimide-dioxime (L ), part by NSF grants CHE 9633007, CHE 82-08992, and NIH
the ligand reacts to oxidize the vanadium to V(V), resulting in grants 1S10RR016634-01, RR 02424A-01. S. H. acknowledges
A
the formation of the known non-oxido V(V) complex of L , 1. the German Academic Exchange Service (DAAD) for a postdoc-
Based on the observed stoichiometry of the reaction and pro- toral scholarship. J. R. P. thanks the University of Edinburgh
ducts observed, we propose that a V(IV) complex is formed tran- for funding a research visit to UC Berkeley and LBNL through
A
siently due to the strong chelating ability of L , but this then a ScotCHEM International Graduate School Scholarship.
reacts with more reduced vanadium to form V(V) in solution.
The reductive reaction of V(IV) was also explored with ligands
A
B
and substrates that are similar to L ; glutaramidoxime (L ),
Notes and references
acetamidoxime, and acetone oxime. These three substrates
required the precipitation of VO
same oxygen transfer reaction to form V(V) with an observed
vanadium/oxime stoichiometry.
2
as a colloid, followed by the
1 J. Kim, C. Tsouris, R. T. Mayes, Y. Oyola, T. Saito,
C. J. Janke, S. Dai, E. Schneider and D. Sachde, Sep. Sci.
Technol., 2013, 48, 367–387.
Cyclic voltammetry was performed on the non-oxido V(V)
2 G. A. Gill, L. Kuo, C. J. Janke, J. Park, R. T. Jeters,
G. T. Bonheyo, H.B. Pan, C. Wai, T. Khangaonkar,
L. Bianucci, J. R. Wood, M. G. Warner, S. Peterson,
D. G. Abrecht, R. T. Mayes, C. Tsouris, Y. Oyola,
J. E. Strivens, N. J. Schlafer, R. S. Addleman, W. Chouyyok,
S. Das, J. Kim, K. Buesseler, C. Breier and E. D’Alessandro,
Ind. Eng. Chem. Res., 2016, 55, 4264–4277.
3 G. Tian, S. J. Teat, Z. Zhang and L. Rao, Dalton Trans., 2012,
41, 11579.
4 X. Sun, C. Xu, G. Tian and L. Rao, Dalton Trans., 2013, 42,
14621–14627.
5 C. J. Leggett, B. F. Parker, S. J. Teat, Z. Zhang, P. D. Dau,
W. W. Lukens, S. J. M. Peterson, A. J. P. Cardenas,
M. G. Warner, J. K. Gibson, J. Arnold and L. Rao, Chem.
Sci., 2016, 7, 2775–2786.
6 A. S. Ivanov, C. J. Leggett, B. F. Parker, Z. Zhang, J. Arnold,
S. Dai, C. W. Abney, V. S. Bryantsev and L. Rao, Nat.
Commun., 2017, 8, 1560.
7 D. Wang and S. A. Sañudo Wilhelmy, Mar. Chem., 2009,
117, 52–58.
8 K. K. Turekian, Oceans, Prentice-Hall, Englewood Cliffs,
N. J., 1968.
complex of glutaroimide-dioxime, [NEt ][1], to investigate
4
whether reduction of this complex is possible, and we found
that reduction is possible electrochemically but at a very redu-
cing potential (E ≈ −1.0 to −1.5 V vs. ferrocene), and is gener-
ally irreversible. Only in dichloromethane was it found to be
reversible electrochemically, but upon attempted chemical
reduction no products could be isolated. We attribute this to
slow reaction of a transient V(IV) complex and lack of available
hydrolysis pathways in an aprotic, relatively inert solvent. In
protic solvents, no V(IV) complex could be formed either
A
through reduction of 1 or direct reaction of L with V(IV)
sources, and only V(V) complexes were obtained.
A
B
The ligands explored in this study (L , L , and HAO) are
analogues of functional groups on polymer sorbents for the
extraction of uranium from seawater. The extraction is gener-
ally performed with polymer sorbents, with the amidoxime
functional groups resembling the binding sites used.
Vanadium in particular is a major competitor for binding
sites, but past work has focused on V(V), which makes up the
majority of vanadium in seawater. However, vanadium also
naturally occurs as V(IV) due to biological redox activity, and
V(IV) composes a small but non-negligible amount of total dis-
9 D. Rehder, Bioinorganic Vanadium Chemistry, John Wiley &
Sons, Ltd, Chichester, UK, 2008.
solved vanadium in the ocean. Due to the prompt and irrevers-
A
ible reaction of V(IV) with L and other analogues in seawater- 10 Vanadium: Biochemical and Molecular Biological Approaches,
relevant conditions, V(IV) will react to permanently damage
ed. H. Michibata, 2014.
amidoxime-functionalized polymer sorbents, which is a poten- 11 N. Dennis Chasteen, in Copper, Molybdenum, and Vanadium
tially greater concern than the reversible, albeit strong,
binding of V(V) which simply competes with U(VI) for binding
and extraction.
in Biological Systems, Springer Berlin Heidelberg, Berlin,
Heidelberg, 1983, pp. 105–138.
12 H. Pan, L. Kuo, C. M. Wai, N. Miyamoto, R. Joshi,
J. R. Wood, J. E. Strivens, C. J. Janke, Y. Oyola, S. Das,
R. T. Mayes and G. A. Gill, Ind. Eng. Chem. Res., 2016, 55,
4
313–4320.
Conflicts of interest
1
3 X. Sun, C. Xu, G. Tian and L. Rao, Dalton Trans., 2013, 42,
The authors declare no conflicts of interest.
14621.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
1
1
1
1
1
1
2
4 D. A. Brown, H. Bögge, R. Coogan, D. Doocey, T. J. Kemp, 27 P. D. Smith, R. E. Berry, S. M. Harben, R. L. Beddoes,
A. Müller and B. Neumann, Inorg. Chem., 1996, 35, 1674–1679.
5 V. Zerbib, F. Robert and P. Gouzerh, J. Chem. Soc., Chem.
Commun., 1994, 98, 2179.
6 V. Chilou, P. Gouzerh, Y. Jeannin and C. B. Knobler, Inorg.
Chim. Acta, 1990, 178, 23–25.
M. Helliwell, D. Collison and C. D. Garner, J. Chem. Soc.,
Dalton Trans., 1997, 4509–4516.
28 J. A. L. da Silva, J. J. R. Fraústo da Silva and
A. J. L. Pombeiro, Coord. Chem. Rev., 2013, 257, 2388–2400.
29 B. Morgenstern, B. Kutzky, C. Neis, S. Stucky,
K. Hegetschweiler, E. Garribba and G. Micera, Inorg. Chem.,
2007, 46, 3903–3915.
7 W. L. Smith and K. N. Raymond, J. Inorg. Nucl. Chem.,
1
979, 41, 1431–1436.
8 F. Endrizzi and L. Rao, Chem. – Eur. J., 2014, 20, 14499– 30 S. Vukovic, L. A. Watson, S. O. Kang, R. Custelcean and
4506. B. P. Hay, Inorg. Chem., 2012, 51, 3855–3859.
9 G. Tian, S. J. Teat and L. Rao, Dalton Trans., 2013, 42, 5690– 31 K. Kustin, G. C. McLeod, T. R. Gilbert and B. R. Le Briggs,
1
5
696.
in Structure and Bonding. Copper, Molybdenum, and
Vanadium in Biological Systems, 1983, vol. 53, pp. 139–160.
32 S. O. Kang, S. Vukovic, R. Custelcean and B. P. Hay, Ind.
Eng. Chem. Res., 2012, 51, 6619–6624.
0 Inorganic Syntheses, ed. D. Coucouvanis, John Wiley & Sons,
Inc., New York, USA, 2002, vol. 33, pp. 75–121.
1 C. J. Leggett and L. Rao, Polyhedron, 2015, 95, 54–59.
2
2
2 O. W. Howarth, Prog. Nucl. Magn. Reson. Spectrosc., 1990, 33 H. N. Holmes and A. L. Elder, Ind. Eng. Chem., 1930, 22,
2
2, 453–485.
471–473.
2
2
2
3 S. G. Bratsch, J. Phys. Chem. Ref. Data, 1989, 18, 1–21.
4 N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877–910.
5 CRC Handbook of Chemistry and Physics, ed. J. R. Rumble, 35 U. S. Mahajan, R. R. Godinde and P. N. Mandhare, Synth.
CRC Press/Taylor & Francis, Boca Raton, FL, 98th edn, 2017. Commun., 2011, 41, 2195–2199.
34 J. A. Elvidge, R. P. Linstead and A. M. Salaman, J. Chem.
Soc., 1959, 208–215.
2
6 E. Lodyga-Chruscinska, A. Szebesczyk, D. Sanna, 36 B. Clement, Drug Metab. Rev., 2002, 34, 565–579.
K. Hegetschweiler, G. Micera and E. Garribba, Dalton 37 F. Eloy and R. Lenaers, Chem. Rev., 1962, 62, 155–183.
Trans., 2013, 42, 13404–13416.
38 F. Garny, Ber. Dtsch. Chem. Ges., 1891, 24, 3426–3437.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018