Novel
Synthesis
Pathways
for Highly
Oxidative
Iron Species:
Generation,
Stability,
and Treatment
Applications
of Ferrate(IV/V/VI)
Sean
T. McBeath,
*
,
§
Yi Zhang,
§
and Michael
R. Hoffmann
Cite This:
Environ.
Sci. Technol.
2023,
57, 18700−18709
Read Online
ACCESS
Metrics
& More
Article
Recommendations
*
sı
Supporting
Information
ABSTRACT:
Difficulties
arise
related
to
the
economy-of-scale
and
practicability
in
applying
conventional
water
treatment
technologies
to
small
and
remote
systems.
A
promising
oxidation
technology
better
suited
for
these
applications
is
that
of
electro-oxidation
(EO),
whereby
contaminants
are
degraded
via
direct,
advanced,
and/or
electrosynthesized
oxidant-mediated
reactions.
One
species
of
oxidants
of
particular
interest
includes
ferrates
(Fe(VI)/(V)/
(IV)),
where
only
recently
has
their
circumneutral
synthesis
been
demonstrated,
using
high
oxygen
overpotential
(HOP)
electrodes,
namely
boron-doped
diamond
(BDD).
In
this
study,
the
generation
of
ferrates
using
various
HOP
electrodes
(BDD,
NAT/Ni
−
Sb
−
SnO
2
, and
AT/Sb-SnO
2
)
was
investigated.
Ferrate
synthesis
was
pursued
in
a current
density
range
of
5
−
15
mA
cm
−
2
and
initial
Fe
3+
concentrations
of
10
−
15
mM.
Faradaic
efficiencies
ranged
from
11
−
23%,
depending
on
operating
conditions,
with
BDD
and
NAT
significantly
outperforming
AT
electrodes.
Speciation
tests
revealed
that
NAT
synthesizes
both
ferrate(IV/V)
and
ferrate(VI),
while
the
BDD
and
AT
electrodes
synthesized
only
ferrate(IV/V)
species.
A
number
of
organic
scavenger
probes
were
used
to
test
the
relative
reactivity,
including
nitrobenzene,
carbamazepine,
and
fluconazole,
whereby
ferrate(IV/V)
was
significantly
more
oxidative
than
ferrate(VI).
Finally,
the
ferrate(VI)
synthesis
mechanism
by
NAT
electrolysis
was
elucidated,
where
coproduction
of
ozone
was
found
to
be
a key
phenomenon
for
Fe
3+
oxidation
to
ferrate(VI).
KEYWORDS:
Electro-oxidation,
electro-synthesis,
ferrate,
ozone,
boron-doped
diamond,
nickel-doped
antimony
tin oxide
■
INTRODUCTION
The
prospect
of
electrochemical
technologies
in
water
treatment
processes
has
been
growing
due
to
their
favorability
in
various
niche
applications
when
compared
to
traditional
technologies.
In
general,
electrochemical
technologies
are
suitable
in
nonconventional
water
treatment
applications,
in
part
due
to
their
favorable
economy-of-scale
and
relative
simplicity.
Electrochemical
oxidation
(electro-oxidation,
EO),
for
example,
is
a promising
alternative
technology
for
small
and
decentralized
system
applications,
as
it
can
eliminate
the
chemical
supply
chain
associated
with
conventional
oxidation/
disinfection
options
by
generating
chemicals
on-site
and
on-
demand.
Conventional
EO
processes,
employing
efficient
nonactive
electrode
materials,
typically
proceed
via
two
reaction
pathways:
(1)
direct
electron
transfer
(DET)
at
the
electrode
surface
and
(2)
hydroxyl
radical
(
•
OH)-mediated
oxidation.
The
latter
reaction
pathway
is
possible
during
EO
when
using
high
oxygen
overpotential
(HOP)
electrode
materials,
such
as
boron-doped
diamond
(BDD)
and
other
mixed
metal
oxides
(MMO),
which
possess
a
greatly
increased
potential
range
(reduction
and
oxidation)
of
water
stability
(e.g.,
BDD:
−
1.25
−
2.3
V
SHE
1
).
Although
the
primary
mechanism
for
pollutant
degradation
during
EO
processes
has
been
attributed
to
•
OH-mediated
oxidation,
2
the
reaction
is
limited
at
the
electrode
surface
where
reactive
oxygen
species
are
generated.
3
A
third
EO
reaction
mechanism
exists,
however,
whereby
the
electrosynthesis
of
residual
chemical
oxidants
proceeds
via
ion
oxidation
at
the
electrode
surface,
resulting
in
pollutant
degradation
in
the
bulk
water
solution
(e.g.,
not
limited
to
the
electrode
surface).
Some
examples
of
electro-generated
oxidant
species
include
persulfate
(
E
0
=
1.96
V
SHE
),
4
peroxodiphosphate
(
E
0
=
2.07
V
SHE
),
5
and
various
reactive
chlorine
species
6,7
when
sulfate,
phosphate,
and
chloride
are
present
in
the
water
matrix,
respectively.
An
additional
group
of
powerful
oxidants
that
has
yet
to
receive
the
same
attention,
as
it
relates
to
circumneutral
electrosynthesis
for
water
treatment
applications,
is
the
generation
of
high
oxidation
state
iron
species
known
as
Special
Issue:
Oxidative
Water
Treatment:
The
Track
Ahead
Received:
December
7,
2022
Revised:
February
1,
2023
Accepted:
February
7,
2023
Published:
February
16,
2023
Article
pubs.acs.org/est
© 2023
The Authors.
Published
by
American
Chemical
Society
18700
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
This article is licensed under CC-BY 4.0
ferrates.
Ferrates
are
particularly
well-suited
for
water
treat-
ment
applications,
8
−
10
as
they
are
not
known
to
form
recalcitrant
oxidation
byproducts
like
chlorinated
disinfection
byproducts
11
and
their
reduced
products
are
nontoxic
hydrolysis
Fe
3+
species,
which
have
also
been
reported
to
effectively
function
as
coagulant
chemicals.
12
−
14
Most
commonly
in
water
treatment
practices,
potassium
ferrate
(K
2
FeO
4
)
is
used
to
form
aqueous
ferrate(VI)
(Fe(VI)/
Fe
VI
O
4
2
−
),
which
is
characteristically
purple
and
has
a
high
redox
potential
(
E
0
=
2.2
V
SHE
).
However,
lesser
reported
high
oxidation
state
iron
species
also
exist,
namely
ferrate(V)
(Fe(V)/Fe
V
O
3
3
−
)
and
ferrate(IV)
(Fe(IV)/Fe
IV
O
4
4
−
).
15,16
Some
studies
have
found
that
Fe(V)
and
Fe(IV)
ferrate
species
yield
degradation
rates
as
much
as
2
−
5
orders
of
magnitude
greater
than
Fe(VI)
for
various
organic
pollutants,
such
as
organosulfur
and
phenolic
compounds,
in
high
pH
conditions.
16,17
In
general,
much
is
still
unknown
about
the
various
ferrate
species
and
their
reactivity
and
stability
in
circumneutral
aqueous
conditions.
While
ferrates
are
conventionally
synthesized
using
a
wet
chemical
method
by
oxidation
of
Fe
3+
in
highly
alkaline
conditions,
15,18
−
20
electrochemical
21,22
and
thermal
chemical
22
methods
also
exist
under
challenging
and
unstable
conditions.
More
recently,
evidence
of
circumneutral
ferrate
electrosyn-
thesis
has
been
reported
through
the
use
of
BDD
electrodes
and
Fe
2+
/Fe
3+
precursors
for
water
treatment
applica-
tions.
23
−
29
In
these
studies,
however,
no
iron
speciation
was
performed
and
ferrate
was
assumed
to
be
in
its
most
stable
Fe(VI)
form.
Although
BDD
has
been
the
preferred
material
for
the
circumneutral
generation
of
ferrate
to
date,
due
to
its
aforementioned
electrocatalytic
properties,
it
also
has
several
limitations.
In
addition
to
prohibitive
costs,
30
BDD
electrodes
require
slow
growth
rates
to
yield
high
quality
films,
they
are
limited
to
substrates
that
are
compatible
with
its
growth
conditions,
and
they
are
generally
size-limited
due
to
the
chemical
vapor
deposition
method
by
which
they
are
synthesized.
31
Another
group
of
promising
materials
includes
substoichiometric
alternatives
to
platinum
group
metal
oxides,
known
as
Magne
́
li
phase
titanium
oxides
(Ti
n
O
2
n
−
1
, 4
≤
n
≤
10),
32,33
which
also
possess
many
favorable
electrocatalytic
properties
for
water
treatment
and
ferrate
synthesis.
However,
similar
to
BDD,
their
applications
are
somewhat
limited
due
to
a
challenging
production
process
requiring
the
reduction
of
TiO
2
in
high
temperature
conditions
and
pure
H
2
.
34
Another
material
considered
as
an
HOP
material
is
antimony-doped
tin(IV)
oxide
(AT/Sb-SnO
2
),
which
has
a comparatively
facile
and
inexpensive
preparation
method
by
dip
or
brush
coating
and
subsequent
annealing
(400
−
600
°
C).
35
Additionally,
several
metal
precursors
can
be
added
to
the
coating
solution
to
generate
metal-doped
AT
electrodes.
Of
particular
interest,
nickel-doped
AT
electrodes
(NAT/Ni
−
Sb
−
SnO
2
) have
been
observed
to
coproduce
O
3
in
addition
to
•
OH,
36
and
facilitate
enhanced
degradation
of
a number
of
organic
pollutants
when
compared
to
AT
electrodes.
37,38
The
fabrication
method
of
both
AT
and
NAT
electrodes
can
facilitate
the
fabrication
of
high-surface-area
flow-through
electrodes,
31,39
which
can
enhance
mass
transport
and
faradaic
efficiency
and
therefore
the
electrochemical
generation
of
ferrate.
In
this
study,
the
circumneutral
electrosynthesis
of
ferrates
was
investigated
using
a
heterojunction
NAT
and
AT
electrodes.
To
date,
the
circumneutral
electrosynthesis
of
ferrates
has
yet
to
be
yielded
with
any
electrode
other
than
BDD.
In
this
study,
we
highlight
the
successful
generation
of
ferrates,
including
the
first
report
of
the
circumneutral
electrosynthesis
of
powerful
intermediate
state
ferrate(IV)
and
ferrate(V)
species,
using
these
MMO
and
BDD
materials.
Moreover,
an
in-depth
mechanistic
study
yields
the
a
novel
reaction
pathway
to
Fe(VI)
from
Fe(III)
via
electrolysis
and
ozone.
The
study
also
includes
oxidation
kinetics,
oxidant
stability,
and
speciation,
as
well
as
application
for
water
treatment
purposes,
presenting
a
potentially
powerful
alter-
native
to
conventional
and
costly
BDD
electrodes
and
wet
chemical
synthesis
processes
for
conventional
ferrate(VI)
generation.
■
MATERIALS
AND METHODS
Electrode
Preparation.
Three
types
of
electrodes
were
used
including
(1)
a single-layer
AT-coated
electrode
on
a Ti
substrate,
(2)
a double-layer
coated
electrode
consisting
of
a
NAT
top
layer
and
an
AT
bottom
layer,
on
a Ti
substrate,
and
(3)
a monocrystalline
BDD
electrode.
The
MMO
electrodes
were
prepared
using
clean
Ti
plates
(2
×
3 cm
2
),
which
were
etched
using
a
1:4
HF:HNO
3
solution
for
1
min.
The
AT
precursor
solution
was
prepared
using
360
mM
SnCl
4
·
5H
2
O
(98%,
Aldrich)
and
40
mM
SbCl
3
(>99.0%,
Aldrich).
The
NAT
precursor
solution
was
prepared
using
360
mM
SnCl
4
·
5H
2
O,
15
mM
SbCl
3
, and
4 mM
Ni(OCOCH
3
)
2
·
4H
2
O
(98%,
Aldrich).
The
respective
metal
oxide
coatings
were
deposited
on
the
Ti
plates
using
a
dip-coater
(MTI
Corporation
Bridgman
Crystal
Growth
Furnace),
which
included
repeated
intervals
of
dipping
the
substrate
into
the
aqueous
metal
oxide
precursor
solution(s)
for
20
s,
drying
at
room
temperature,
followed
by
a calcination
step
by
annealing
at
600
°
C
for
10
min.
This
was
repeated
until
a
desired
mass
loading
was
achieved
(AT
electrode:
1.3
mg
cm
−
2
, NAT
electrode:
1.3
mg
cm
−
2
AT
and
1.3
mg
cm
−
2
NAT),
whereby
a final
annealing
step
at
600
°
C
was
performed
for
1.5
h.
The
BDD
electrode
was
commercially
purchased
from
NeoCoat.
It
was
prepared
by
a
chemical
vapor
deposition
process
and
had
a
thin-film
(2
−
3
μ
m)
monocrystalline
layer
on
a 1 mm
silicon
substrate.
Experimental
Methods
and Procedures.
All
experi-
ments
were
performed
using
an
undivided
electrolysis
cell
(50
mL).
All
electrolyte
solutions
were
continuously
stirred
with
a
magnetic
stirrer
at
a rate
of
400
rpm.
The
anode
(AT,
NAT
and
BDD)
and
cathode
(stainless
steel)
were
6
cm
2
for
all
experiments,
separated
by
a 5 mm
interelectrode
gap.
All
tests
were
performed
in
a
three-electrode
configuration,
using
an
Ag/AgCl
reference
electrode
(BASI
Inc.)
and
Biologic
VSP-
300
potentiostat.
The
base
water
matrix
used
for
all
experiments
was
a
phosphate
buffer
(pH
=
7.0,
0.1
M),
composed
of
Milli-Q
water,
NaH
2
PO
4
,
and
Na
2
HPO
4
(Millipore
Sigma).
The
desired
initial
Fe
3+
concentration
was
attained
by
addition
of
FeCl
3
(Sigma-Aldrich).
Current
densities
of
5,
10,
and
15
mA
cm
−
2
were
investigated,
to
operate
in
a potential
range
avoiding
excess
oxygen
evolution,
while
maximizing
hydroxyl
radical,
DET
and
ozone
formation,
depending
on
each
respective
electrode.
Samples
for
ferrate,
probe
or
pollutant
analysis
were
extracted
from
the
electro-
chemical
cell
at
a maximum
volume
of
0.5
mL
throughout
60
or
90
min
experiments.
After
electrolysis,
ferrate
solutions
were
centrifuged
for
2
min
at
5000
rcf
and
subsequently
filtered
using
a 0.45
μ
m
glass
fiber
syringe
filter
(Tisch)
to
remove
any
nonaqueous
iron
species
(e.g.,
Fe
2+
/Fe
3+
oxides
and
hydrox-
ides).
Linear
sweep
voltammetry
(LSV)
and
cyclic
voltammetry
(CV)
experiments
were
also
conducted
under
a scan
rate
of
50
Environmental
Science
& Technology
pubs.acs.org/est
Article
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
18701
mV
s
−
1
in
relevant
electrolytes
for
electrode
material
characterization
as
well
as
oxidation
mechanism
and
ferrate
speciation
analysis.
Analytical
Methods.
Ferrate
concentration
was
measured
with
an
indirect
spectrophotometric
method
using
an
ABTS
(2,2
′
-azino-bis(3-ethylbenzothiazoline-6-sulfonic
acid))
re-
agent
(Sigma-Aldrich).
In
the
presence
of
excess
ABTS,
ferrate(VI)
oxidizes
ABTS
with
a
1:1
M
ratio,
producing
a
light-absorbing
radical
cation
(ABTS
•
+
)
with
a
visible
UV-
absorption
maxima
at
415
nm.
40
−
42
Ferrate
standards
and
samples
were
analyzed
using
a
UV
−
vis
spectrophotometer
(Thermo
Scientific
Nanodrop
2000c)
and
a
1
cm
quartz
cuvette.
Ferrate
standards
were
prepared
using
potassium
ferrate
(99%)
(Element
26)
in
a concentration
range
of
0
−
10
mM.
Ferrate(VI)
was
also
directly
analyzed
using
direct
UV
−
vis
spectrophotometry
at
530
nm
in
1
and
5
cm
quartz
cuvettes,
in
a
concentration
range
of
0
−
2
mM
with
co-
occurring
Fe
3+
(dosed
with
FeCl
3
),
to
simulate
water
matrix
conditions
during
electrolysis
experiments
and
to
understand
the
effects
on
UV-absorbance
and
ferrate(VI)
stability.
Raman
spectroscopy
(Renishaw
inVia
Qontor)
and
Fourier
transform
infrared
spectroscopy
(FTIR)
(Thermo
Scientific
Nicolet
iS50)
were
also
pursued
for
a
limited
number
of
tests
(operating
methods
are
detailed
in
the
Supporting
Informa-
tion).
Free
chlorine
concentrations
were
also
monitored
using
the
DPD
(
N
,
N
-diethyl-
p
-phenylenediamine)
reagent
(Hach
DPD
method
1012)
and
a DR
300
colorimeter.
Oxidant
probe
species
and
micropollutants,
namely
nitro-
benzene
(NB),
carbamazepine
(CBZ),
and
fluconazole
(FCZ)
(Sigma-Aldrich),
were
quantified
using
high
performance
liquid
chromatography
(HPLC),
equipped
with
a
ZORBAX
Eclipse
XDB-C18
column
(Agilent,
2.1
×
50
mm
2
, 3.5
μ
m
particles)
and
a
UV
detector
at
254,
285,
and
205
nm,
respectively.
The
mobile
phase,
flowing
at
0.5
mL
min
−
1
, was
a
composition
of
water
with
0.1%
formic
acid
and
acetonitrile
(ACN)
under
the
gradient:
0
min,
10%
ACN;
2
min,
10%
ACN;
6 min,
95%
ACN;
8 min,
95%
ACN;
9 min,
10%
ACN;
12
min,
10%
ACN.
An
injection
volume
of
20
μ
L
was
used
with
a total
runtime
of
12
min
for
each
sample.
■
RESULTS
AND DISCUSSION
Electrode
Performance.
Nonactive
electrodes
are
char-
acterized
by
their
high
oxygen
evolution
reaction
(OER)
overpotential.
43
The
NAT
and
AT
electrodes
have
previously
been
shown
to
have
oxygen
evolution
potential
at
∼
2.4
V
RHE
.
38
BDD
demonstrates
even
higher
OER
potential
at
∼
2.7
V
RHE
44
(see
Figure
S1
in
the
Supporting
Information).
One
primary
mechanism
of
oxidation,
with
respect
to
all
three
electrode
materials
(e.g.,
BDD,
NAT,
and
AT),
is
mediated
through
the
generation
of
•
OH,
which
was
previously
demonstrated
by
Zhang
et
al.
through
radical
scavenging
studies
involving
Figure
1.
(A)
Ferrate(IV/V/VI)
generation
(a)
using
BDD,
NAT,
and
AT
electrodes
with
a Fe
3+
0
=
10
mM
and
10
mA
cm
−
2
. Fe(VI)
Eq
=
Equivalent
oxidative
capacity
of
Fe(VI)
with
ABTS
(1:1
molar
ratio).
(B)
Faradaic
efficiency
of
ferrate
(IV/V/VI)
generation
at
5,
10,
and
15
mA
cm
−
2
and
with
a Fe
3+
0
=
10
mM.
Figure
2.
Ferrate(IV/V/VI)
generation
with
a Fe
3+
0
=
10
mM
at
5,
10,
and
20
mA
cm
−
2
using
(A)
NAT
electrode
and
(B)
BDD
electrode.
Inset
figures
are
final
ferrate
solutions
after
90
min
of
electrolysis.
Fe(VI)
Eq
=
Equivalent
oxidative
capacity
of
Fe(VI)
with
ABTS
(1:1
molar
ratio).
Environmental
Science
& Technology
pubs.acs.org/est
Article
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
18702
nitrobenzene
and
benzoic
acid
38
and
widely
reported
for
BDD
electro-oxidation.
2
Moreover,
when
Cl
−
is
present
in
the
water
matrix,
all
three
electrodes
are
observed
to
generate
reactive
chlorines
species
(RCS).
The
NAT
electrode
is
unique
in
its
O
3
production
capacity.
Equilibrium
aqueous
O
3
concentration
can
reach
as
high
as
∼
4.7
mg
L
−
1
in
the
absence
of
chloride
and
goes
down
with
higher
chloride
concentration.
38
The
BDD
electrode
is
the
only
material
of
the
three
that
is
shown
to
facilitate
direct
electron
transfer
(DET)-mediated
oxidation.
Specifically,
DET
is
known
to
contribute
to
FCZ
degradation
during
BDD
electrolysis.
45
Ferrate
Generation.
Ferrate
generation
experiments
were
first
performed
using
an
initial
FeCl
3
concentration
of
10
mM
and
a current
density
of
10
mA
cm
−
2
, with
the
NAT,
AT,
and
BDD
electrodes.
Control
experiments
were
run
in
parallel
with
30
mM
NaCl,
yielding
an
equivalent
Cl
−
concentration,
to
understand
the
oxidative
effect
of
cogenerated
reactive
chlorine
species
(RCS)
during
ferrate
experiments.
Additional
control
experiments
were
conducted
to
determine
whether
phosphate
active
species
could
also
be
generated,
whereby
electrolysis
was
performed
in
the
presence
of
the
base
PBS
water
matrix
and
subsequently
analyzed
for
any
oxidative
species.
For
all
anode
materials,
evidence
of
ferrate
generation
was
observed
using
the
ABTS
quantification
method
40,41
(see
Figures
S2
−
S12
for
all
ABTS
absorbance
data).
While
the
circumneutral
generation
of
ferrate
has
been
recently
reported
using
BDD,
24
this
is
the
first
evidence
of
ferrate
generation
using
both
NAT
and
AT
electrodes.
Faradaic
efficiencies
of
22.5,
21.3,
and
12.0%
were
yielded
for
the
BDD,
NAT,
and
AT
electrodes,
respectively
(see
Figure
1b).
Qualitatively,
the
final
ferrate
solutions
were
significantly
different
(see
Figure
2
inset).
While
the
NAT
electrode
produced
a
purplish/pink
solution
that
is
characteristic
of
ferrate(VI),
9
both
the
BDD
and
AT
electrodes
produced
a
yellowish/white
solution
(see
Figure
2),
indicating
that
ferrate(VI)
was
not
produced.
The
electrosynthesis
of
oxidative
iron
species
was
also
observed
to
increase
with
both
current
density
(20
mA
cm
−
2
)
and
initial
Fe
3+
concentration
(15
mM)
(see
Figure
S13
for
all
ABTS
absorbance
data).
During
NAT
experiments,
ferrate
yields
of
12.8,
63.5,
and
69.4%
were
achieved
during
5,
10,
and
20
mA
cm
−
2
operations.
However,
the
Faradaic
efficiency
was
maximized
during
10
mA
cm
−
2
electrolysis
at
21.3%.
Similarly,
during
BDD
electrolysis,
the
greatest
Faradaic
efficiency
was
yielded
during
10
mA
cm
−
2
(see
Figure
1b),
with
yields
of
28.5,
67.2,
and
135.8%
during
5,
10,
and
20
mA
cm
−
2
operations.
Average
cell
potentials
of
3.5,
3.9,
and
4.7
V
were
yielded
during
NAT
electrosynthesis,
while
BDD
potentials
recorded
were
4.7,
5.6,
and
6.9
V
during
5,
10,
and
20
mA
cm
−
2
operations,
respectively.
At
20
mA
cm
−
2
BDD
operations,
as
well
as
for
all
current
density
conditions
using
an
initial
FeCl
3
concentration
of
15
mM,
ABTS
results
suggested
that
ferrate(VI)
yields
exceeded
100%,
indicating
errors
associated
with
the
ABTS
method
for
ferrate(VI)
quantifica-
tion.
In
theory,
ferrate(VI)
can
facilitate
the
exchange
of
3 mol
of
electrons
for
every
1 mol
of
ferrate(VI)
through
its
subsequent
conversion
to
Fe(V),
Fe(IV),
and
Fe(III).
In
phosphate
buffer,
however,
it
has
previously
been
found
that
ferrate(VI)
has
a
1:1.18
ratio
with
ABTS
to
form
ABTS
•
+
in
circumneutral
pH,
whereby
the
subsequent
oxidative
effect
of
Fe(V)
and
Fe(IV)
are
muted
through
rapid
autodecomposition.
46
Under
the
same
conditions,
roughly
9%
of
Fe(V)
was
found
to
react
with
ABTS
to
form
Fe(IV),
whereby
93%
of
that
Fe(IV)
subsequently
reacting
with
ABTS
to
form
ABTS
•
+46
. Achieving
ferrate
yields
greater
than
100%
using
the
ABTS
method
(which
assumes
a
1:1
reaction
stoichiometry)
for
BDD
electrosynthesis
experiments
suggests
that
Fe(V),
or
some
combination
of
Fe(V)
and
Fe(IV),
is
formed,
resulting
in
a
1:1
−
2
molar
reaction
ratio.
The
ABTS
technique,
and
the
resulting
molar
reaction
ratio,
has
been
previously
used
to
identify
the
role
of
intermediate
ferrate
species
(e.g.,
Fe(IV/
V))
in
water
treatment
processes.
47
The
generation
and
use
of
Fe(IV)
and
Fe(V)
species,
particularly
for
water
treatment
processes,
has
been
a recent
topic
of
much
research
interest
and
has
primarily
been
reported
through
the
activation
of
ferrate(VI).
47
The
activation
of
ferrate(VI),
to
exploit
the
high
oxidation
capabilities
of
Fe(IV/V),
has
been
investigated
using
a number
of
strategies
including
the
use
of
acid,
48
UV,
49
and
metal
cations
like
Fe(III),
50
among
many
other
organic
and
inorganic
activators.
47
Of
particular
relevance,
the
role
of
Fe(III)
in
Fe(IV/V)
generation
is
notable
due
to
its
efficacy
toward
Fe(IV)
generation
and
its
co-occurrence
during
ferrate
electrosynthesis.
Ferrate
Speciation.
Several
difficulties
arise
when
conducting
a
speciation
study
on
ferrate(IV/V/VI),
partic-
ularly
for
direct
detection
and
quantification
of
low
concentration
aqueous
solutions.
Both
Raman
spectroscopy
and
Fourier
transform
infrared
(FTIR)
spectroscopy
were
used
Figure
3.
UV-spectra
(left)
and
cyclic
voltammograms
(right)
of
final
BDD
and
NAT
ferrate
solutions
after
90
min
of
electrolysis
at
10
mA
cm
−
2
and
an
initial
FeCl
3
concentration
of
10
mM
(CV
condition:
scan
rate
50
mV
s
−
1
, 0.0
V
vs
E
OC
to
−
1.3
V
vs
ref.)
Environmental
Science
& Technology
pubs.acs.org/est
Article
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
18703
to
analyze
the
ferrate
solutions
obtained
during
BDD
and
NAT
electrolysis,
but
no
differentiating
peaks
were
observed
(see
Supporting
Information
for
Raman
and
FTIR
spectra
graphs,
Figures
S14
and
S15,
respectively).
Moreover,
high-valent
iron-
oxo
complexes
can
exist
in
various
forms
and
the
exact
structures
are
still
largely
unknown.
47
For
example,
the
synthesis
of
ferrate(IV),
by
hydroxyl
radical-mediated
oxidation
of
Fe(III),
in
phosphate,
pyrophosphate,
and
carbonate
solutions
has
been
reported.
The
resulting
ferrate-
(IV)
species
include
one
or
more
hydroxide,
pyrophosphate,
or
carbonate
ligand
(L
m
).
However,
the
exact
Fe
IV
-L
m
structure
and
number
of
ligands
attached
to
the
central
iron
atom
are
not
known.
51,52
It
should
also
be
noted
that
phosphate
has
previously
been
observed
to
function
as
a
Fe(IV/V)
ligand
(e.g.,
Fe
V
-L
m
and
Fe
IV
-L
m
),
47
which
may
play
a crucial
role
in
ferrate
speciation
in
this
study
due
to
the
presence
of
phosphate
in
the
base
PBS
water
matrix.
The
stability
of
ferrate
species
has
also
been
observed
to
vary
depending
on
exact
structure
and
the
ligand(s)
coordinated
to
the
iron
atom.
52
Direct
colorimetry
can
also
be
used
to
identify
ferrate(VI)
in
solution,
53,54
at
a much
lower
molar
absorption
coefficient
than
that
of
ABTS;
therefore,
higher
concentrations
and/or
longer
UV
path
lengths
are
required.
Using
a
5
cm
quartz
cuvette,
both
the
BDD
and
NAT
electrogenerated
ferrates
were
analyzed
(see
Figure
3).
The
UV
absorbance
spectra
confirmed
the
qualitative
observations,
whereby
the
solution
produced
with
the
NAT
electrode
showed
evidence
of
ferrate(VI)
with
an
absorbance
maximum
at
526
nm.
40
This
peak
is
slightly
higher
than
that
of
ferrate(VI)
observed
with
the
chemical
standard
(using
K
2
FeO
4
)
and
the
peak
shapes
are
notably
different,
as
seen
in
Figure
S35
with
a maximum
at
524
nm.
When
Fe(III)
was
added
to
the
K
2
FeO
4
mixture,
however,
the
same
absorbance
maximum
was
achieved
(see
Figure
S35).
Moreover,
while
the
NAT-produced
ferrate(VI)
solution
was
more
pink
(compared
to
a purple
color
traditionally
associated
with
a high
concentration
of
ferrate(VI)),
it
was
similar
to
the
color
yielded
using
K
2
FeO
4
and
Fe(III).
These
results
indicate
that
the
lighter
color
observed
during
NAT
electrolysis
is
due
to
either
the
PBS
and/or
co-occurring
Fe(III)
cations
in
solution.
A
secondary
shoulder
between
275
and
310
nm
was
observed
for
both
the
BDD
and
NAT
derived
ferrates,
which
is
also
characteristic
of
ferrate(VI)
UV-spectra.
40
No
absorbance
maxima
was
observed
for
the
BDD
solution,
including
at
530
nm
(e.g.,
ferrate(VI)
absorbance
peak
in
the
presence
of
Fe
3+
),
conclusively
indicating
the
absence
of
ferrate(VI).
Cyclic
voltammetry
provided
further
direct
evidence
of
two
different
iron
species
in
the
BDD
and
NAT
solutions,
as
seen
by
the
different
reduction
peaks
in
Figure
3.
To
further
understand
ferrate
speciation
between
the
BDD
and
NAT
derived
solutions,
the
use
of
selective
organic
probes
was
pursued.
The
recalcitrance
of
nitrobenzene
(NB)
to
ferrate(VI)
oxidation
has
previously
been
reported.
15,55,56
Using
an
initial
concentration
of
0.1
mM,
no
detectable
NB
degradation
was
observed
to
occur
as
a
result
of
the
NAT
produced
ferrate(VI),
when
compared
to
the
control
study
containing
RCS
produced
during
electrolysis
of
the
NaCl
control
solution
(see
Figures
S16
and
S17).
During
parallel
experiments
using
the
BDD-derived
solution,
NB
degradation
was
observed
to
increase
in
the
presence
of
ferrate(IV/V),
when
compared
to
the
control
solution
containing
only
RCS,
with
pseudo
first-order
reaction
rate
constants
of
0.0055
and
0.0047
min
−
1
, respectively
(see
Figure
S18).
Carbamazepine
(CBZ)
was
also
selected
as
a suitable
probe,
particularly
in
the
water
matrix
postelectrolysis,
which
contains
a high
concentration
of
RCS,
due
to
its
relative
persistence
in
highly
chlorinated
waters.
57,58
With
all
electrodes
(NAT,
AT,
and
BDD),
minimal
CBZ
degradation
was
observed
during
control
(NaCl)
experiments.
Ferrate(VI),
produced
using
the
NAT
electrode,
facilitated
the
slow
degradation
of
CBZ
over
∼
15
min
of
mixing,
yielding
a
second-order
reaction
rate
constant
of
1.4
M
−
1
s
−
1
. The
degradation
rate
of
CBZ
with
ferrate(VI)
has
been
previously
observed
to
be
highly
dependent
on
pH,
with
k
2
constants
ranging
from
0.1
to
70
M
−
1
s
−
1
between
pH
conditions
of
8
to
6,
respectively,
59,60
which
is
in
agreement
with
the
present
study.
Significantly
faster
CBZ
oxidation
was
observed
using
the
BDD-generated
ferrate(IV/V),
whereby
83%
of
CBZ
degradation
took
place
in
the
first
20
s of
mixing,
yielding
an
apparent
first-order
reaction
rate
constant
of
0.08
s
−
1
. This
reaction
rate
is
similar
to
that
which
has
been
observed
for
CBZ
degradation
with
permanganate/Mn(VII)/MnO
4
−
,
60
an
analogous
highly
oxida-
tive
manganese
species
also
used
for
aqueous
micropollutant
degradation.
61
−
64
The
pseudo
first-order
reaction
rate
constant
of
permanganate
with
CBZ
was
found
to
increase
with
the
initial
permanganate
concentration,
whereby
a
k
′
≈
0.045
s
−
1
was
yielded
with
an
initial
permanganate
concentration
of
160
μ
M.
Although,
in
general,
ferrate(VI)
is
more
highly
oxidative
Figure
4.
ABTS
(415
nm)
and
direct
(530
nm)
UV
−
vis
spectrophotometric
results
from
NAT
electrolysis
at
10
mA
cm
−
2
and
FeCl
3
of
15
mM
(left),
and
O
3
oxidation
with
an
initial
FeCl
3
of
15
mM
(right).
Environmental
Science
& Technology
pubs.acs.org/est
Article
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
18704
than
the
permanganate
ion,
the
latter
was
reported
to
degrade
CBZ
more
readily
due
to
its
high
reactivity
with
olefin
groups.
65,66
At
present,
no
previous
studies
have
been
published
investigating
the
used
of
ferrate(V)
or
ferrate(IV)
on
the
degradation
of
CBZ,
but
is
observed
to
perform
similarly
to
permanganate.
CBZ
degradation
was
also
observed
using
the
ferrate(IV/V)
generated
using
the
AT
electrode.
Similar
to
the
BDD
derived
ferrate
solution,
rapid
degradation
was
observed
in
the
first
20
s
(see
Figure
S22),
with
an
absolute
CBZ
removal
much
less
than
BDD
due
to
the
significantly
lower
initial
concentration
of
ferrate(IV/V).
Fluconazole
(FCZ)
was
used
as
a probe,
as
it
is
recalcitrant
to
both
chlorine
and
chlorine
radicals
(Cl
2
•−
).
Similar
to
CBZ,
no
FCZ
degradation
was
observed
during
control
experiments
with
both
the
NAT
and
BDD
producing
ferrate
solutions.
Moreover,
ferrate(VI),
synthesized
using
the
NAT
electrode,
did
not
yield
any
detectable
oxidation
of
FCZ.
A
small
amount
of
FCZ
degradation
was
observed
using
the
ferrate(IV/V),
produced
with
the
BDD
electrode,
once
again
highlighting
the
greater
oxidation
potential
of
these
iron
species
compared
to
ferrate(VI).
In
agreement
with
previous
researchers,
the
BDD
produced
ferrate(IV/V)
consistently
yielded
greater
degrada-
tion
rates
when
compared
to
ferrate(VI)
for
all
organic
micropollutant
tested.
16,17
A
complete
set
of
the
CBZ
and
FCZ
degradation
data
can
be
found
in
the
Supporting
Information
(Figures
S19
−
S24).
To
better
understand
the
ferrate
speciation
during
NAT
electrolysis,
a
combination
of
ABTS
and
direct
UV-
spectrophotometer
analysis
was
conducted
in
parallel
at
415
and
530
nm
to
evaluate
the
relative
ferrate(IV/V)
and
ferrate(VI)
synthesis,
respectively.
The
absorbance
related
to
the
generation
of
ferrate(VI)
increased
over
the
initial
10
min
of
electrolysis;
thereafter,
it
stabilized
for
the
remainder
of
the
experiment.
The
ABTS
absorbance,
which
reflects
oxidation
by
ferrate(IV),
ferrate(V),
and
ferrate(VI),
continued
to
increased
significantly
throughout
the
entirety
of
electrolysis
(see
Figure
4).
These
results
indicate
that
only
a small
fraction
of
the
initial
Fe
3+
is
converted
to
ferrate(VI),
1.65
mM
(
±
0.2
mM)
under
these
conditions
(10
mA
cm
−
2
and
[Fe
3+
]
0
= 15
mM),
in
a fast
reaction
within
10
min
of
starting
electrolysis,
and
remained
constant
thereafter.
Although
ferrate(VI)
generation
reaches
a
plateau,
the
formation
of
ferrate(IV/V)
continues
to
be
facilitated
throughout
the
90
min
of
electrolysis.
Each
of
the
techniques
used
in
this
study,
however,
is
indirect
analysis
for
ferrate
speciation.
In
future
studies,
it
would
be
important
to
incorporate
direct
analysis,
particularly
the
use
of
Mo
̈
ssbauer
spectroscopy.
In
order
to
use
Mo
̈
ssbauer
spectroscopy,
reactor
design
and
current
efficiency
limitations
of
the
current
setup
would
need
to
be
improved,
as
a higher
concentration
of
ferrates
in
a solid
sample
would
be
required.
47
Ferrate
Stability.
A
long-term
stability
study
was
performed
on
both
the
ferrate(VI)
(e.g.,
NAT
produced
ferrate)
and
ferrate(IV/V)
(e.g.,
BDD
produced
ferrates)
to
understand
their
relative
self-decomposition
in
ambient
temperature
(20
°
C)
and
light
conditions.
Similar
to
the
previously
described
ferrate
generation
and
speciation
tests,
all
stability
tests
were
performed
in
parallel
with
control
experiments
using
NaCl
solutions
and
monitoring
the
degradation
of
cogenerated
RCS.
Using
the
same
initial
FeCl
3
(15
mM)
concentration
with
both
the
NAT
and
BDD
electrodes,
after
60
min
of
electrolysis,
initial
Fe(VI)
Eq
concentrations
of
3.9
and
13.6
mM
were
achieved
after
centrifugation
and
filtration,
respectively.
The
ABTS
absor-
bances
were
subsequently
monitored
for
the
ferrate
and
RCS
control
solutions
over
70
days.
The
oxidant
species
stability
of
all
solutions
(e.g.,
ferrate
and
control)
degraded
similarly;
however,
some
outlier
ABTS
absorbances
were
observed
between
days
3
−
11
for
the
BDD
ferrate(IV/V).
Unlike
the
remainder
of
the
stability
data,
these
outliers
included
large
variations
in
ferrate(IV/V)
concen-
trations,
and
no
conclusive
evidence
had
been
gathered
to
explain
these
outliers.
One
possible
explanation
may
be
related
to
the
decomposition
of
Fe(V)
to
Fe(IV)
and/or
Fe(III)
during
sampling
by
oxidation
with
nonaqueous
iron
hydr-
(oxides)
(e.g.,
reduced
ferrate
products).
All
ABTS
stability
data
are
included
in
the
Supporting
Information
(Figures
S25
−
S28).
When
accounting
for
the
effect
of
RCS,
zero-order
degradation
was
observed
for
both
the
NAT
produced
ferrate(VI)
and
BDD
produced
ferrate(IV/V).
Excluding
the
Figure
5.
Normalized
ferrate
degradation
with
(A)
BDD
derived
ferrate(IV/V)
(Fe(VI)
Eq
=
13.6
mM)
and
(B)
NAT
derived
ferrate(VI)
(Fe(VI)
=
3.9
mM).
Environmental
Science
& Technology
pubs.acs.org/est
Article
https://doi.org/10.1021/acs.est.2c09237
Environ.
Sci. Technol.
2023,
57, 18700
−
18709
18705