Modular
M
PS
3
‑
Based
Frameworks
for
Superionic
Conduction
of
Monovalent
and
Multivalent
Ions
Zachery
W. B. Iton, Zion Irving-Singh,
Son-Jong
Hwang,
Amit Bhattacharya,
Sammy
Shaker,
Tridip
Das,
Rapha
ë
le
J. Cl
é
ment,
William
A. Goddard
III, and Kimberly
A. See
*
Cite
This:
J. Am.
Chem.
Soc.
2024,
146, 24398−24414
Read
Online
ACCESS
Metrics
& More
Article
Recommendations
*
sı
Supporting
Information
ABSTRACT:
Next-generation
batteries
based
on more
sustainable
working
ions
could
offer
improved
performance,
safety,
and
capacity
over
lithium-ion
batteries
while
also
decreasing
the
cost.
Development
of next-generation
battery
technology
using
“beyond-
Li”
mobile
ions,
especially
multivalent
ions,
is limited
due
to a lack
of understanding
of
solid
state
conduction
of these
ions.
Here,
we
introduce
ligand-coordinated
ions
in
M
PS
3
-based
(M
= Mn,
Cd)
solid
host
crystals
to simultaneously
increase
the
size
of the
interlayer
spacing,
through
which
the
ions
can
migrate,
and
screen
the
charge-dense
ions.
The
ligand-assisted
conduction
mechanism
enables
ambient
temperature
superionic
conductivity
of various
next-generation
mobile
ions
in the
electronically
insulating
M
PS
3
-based
solid.
Without
the
coordinating
ligands,
all of the
compounds
show
little
to no ionic
conductivity.
Pulsed-field
gradient
nuclear
magnetic
resonance
spectroscopy
suggests
that
the
ionic
conduction
occurs
through
a hopping
mechanism,
where
the
cations
are
moving
between
H
2
O molecules,
instead
of a vehicular
mechanism
which
has
been
observed
in other
hydrated
layered
solids.
This
modular
system
not
only
facilitates
tailoring
to different
potential
applications
but
also
enables
us to probe
the
effect
of different
host
structures,
mobile
ions,
and
coordinating
ligands
on the
ionic
conductivity.
This
research
highlights
the
influence
of cation
charge
density,
diffusion
channel
size,
and
effective
charge
screening
on
ligand-assisted
solid
state
ionic
conductivity.
The
insights
gained
can
be applied
in the
design
of other
ligand-assisted
solid
state
ionic
conductors,
which
will
be
especially
impactful
in realizing
solid
state
multivalent
ionic
conductors.
Additionally,
the
ion-intercalated
M
PS
3
-based
frameworks
could
potentially
serve
as a universal
solid
state
electrolyte
for various
next-generation
battery
chemistries.
■
INTRODUCTION
The
landscape
of
modern
battery
technology
has
been
dominated
by lithium-ion
batteries
(LIBs).
However,
material
availability,
scalability,
cost,
and
escalating
global
energy
demands
1
−
3
motivate
the
development
of alternative
battery
chemistries.
“Next-generation”
mobile
ions,
such
as Na
+
, K
+
,
Mg
2+
, Ca
2+
, Zn
2+
, and
Al
3+
, represent
a promising
frontier
in
battery
technology
due
to their
abundance
and
potential
for
high
volumetric
capacities.
4
However,
one
of
the
major
challenges
in developing
battery
technology
based
on
next-
generation
mobile
ions
is the
difficulty
in achieving
the
solid
state
conduction
of those
ions.
Solid
state
conduction
is crucial
for
ion
transport
in
electrodes,
interphases,
and
solid
electrolytes.
The
larger
size
of ions
like
Na
+
, K
+
, and
Ca
2+
restricts
their
movement
through
the
typically
rigid
migration
pathways
available
in solid
materials.
Additionally,
the
mobility
of ions
with
higher
charge
densities,
such
as Mg
2+
, Zn
2+
, and
Al
3+
, is impeded
by
the
strong
electrostatic
interactions
between
these
mobile
ions
and
other
ions
within
the
solid.
These
challenges
are
outlined
in
detail
in
our
recent
perspective.
4
Difficulties
with
solid
state
ionic
conduction
are
less
pronounced
in Li-based
systems
due
to the
relatively
small
size
and
low
charge
density
of Li
+
.
Solid
state
conduction
of larger
cations
has
typically
been
achieved
in structures
based
on open
frameworks
with
large
migration
bottlenecks,
like
Prussian
blue
analogues,
5
−
7
β
-
Alumina,
8
−
11
or NASICON
phases,
12
−
14
whereas
solid
state
conduction
of charge-dense
ions
has
been
mostly
limited
to
electronically
conductive
materials.
In systems
like
Mo
6
S
8
,
thiospinel
Ti
2
S
4
, and
MgSc
2
S4,
the
mobile
electrons
are
hypothesized
to facilitate
ionic
mobility
by
screening
the
charge-dense
ions,
decreasing
the
strength
of electrostatic
interactions
within
the
solid.
15
−
19
For
electronically
insulating
materials,
such
as solid
electrolytes
or interphases,
there
is an
absence
of mobile
electrons
to screen
the
charge
of the
targeted
ions.
Only
a handful
of electronically
insulating
inorganic
solids
have
been
shown
to conduct
charge-dense
ions,
such
as Zn
2+
in ZnPS
3
,
20
or Mg
2+
in borohydrides,
21
−
26
but
each
has
its
own
challenges
associated
with
low
RT
conductivities
(
σ
RT
) or high
activation
energy
(
E
a
).
Received:
May
8, 2024
Revised:
August
7, 2024
Accepted:
August
8, 2024
Published:
August
20,
2024
Article
pubs.acs.org/JACS
© 2024
The Authors.
Published
by
American
Chemical
Society
24398
https://doi.org/10.1021/jacs.4c06263
J. Am.
Chem.
Soc.
2024,
146, 24398
−
24414
This article is licensed under CC-BY-NC-ND 4.0
Downloaded via CALIFORNIA INST OF TECHNOLOGY on November 23, 2024 at 00:22:30 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
One
path
to enable
the
conduction
of large
or charge-dense
cations
in electronic
insulators
is to introduce
ligands,
such
as
H
2
O, that
can
coordinate
with
the
targeted
cations
within
the
host
crystal.
Solid
state
mobility
of ligand-coordinated
ions
was
extensively
studied
in mica
clays
such
as vermiculite
and
montmorillonite.
27
−
31
However,
in clays
the
σ
is generally
<0.1
mS
cm
−
1
. In battery
systems,
the
addition
of H
2
O to increase
multivalent
ionic
conduction
has
been
primarily
attempted
for
cathode
materials,
such
as MnO
2
and
V
2
O
5
.
32
−
35
Recently,
this
concept
has
been
extended
to
electronically
insulating
inorganic
solids,
like
Li
2
Sn
2
S
5
36
and
ZnPS
3
,
37
as well
as
various
MOFs
and
COFs,
38
−
42
for
solid
state
electrolyte
applications,
however
in most
cases
the
conductivity
is still
below
practically
useful
values
or the
excess
solvent
leads
to
detrimental
reactivity
in the
system.
Here,
we leverage
coordinated
ligand
molecules
within
solid
frameworks
based
on
M
PS
3
materials
(M
= Mn,
Cd)
to
achieve
two
primary
goals:
(1)
increasing
the
size
of the
interlayer
spacing,
through
which
the
ions
can
migrate,
and
(2)
screening
charge-dense
ions
to decrease
electrostatic
inter-
actions
within
the
lattice.
By combining
the
advantages
of large
interlayer
spacing
and
charge
screening,
we
can
devise
a
universal
framework
for
ambient
temperature
superionic
conduction
of various
ions
within
an inorganic,
electronically
insulating
solid.
CdPS
3
and
MnPS
3
exhibit
a peculiar
mechanism
for
the
intercalation
of hydrated
cations
into
the
van
der
Waals
gap.
43
Unlike
the
redox-based
intercalation
in common
battery
electrodes
in which
incorporation
of a cation
coincides
with
a Faradaic
reduction
of
the
host,
44
when
cations
are
intercalated
into
M
PS
3
materials,
charge
balance
is maintained
by
M
2+
loss
in an
ion
exchange
mechanism.
43,45,46
This
mechanism
is similar
to that
observed
in clays.
27,28,47,48
CdPS
3
and
MnPS
3
form
monoclinic
(
C
2/
m
)
layered
compounds
with
a slightly
distorted
hexagonal
network
of edge-sharing
Cd
2+
or
Mn
2+
octahedra.
The
Cd
2+
or Mn
2+
bonds
are
coordinated
by
[P
2
S
6
]
4
−
polyanions.
The
layers
stack
along
the
c
-axis
separated
by a van
der
Waals
gap
of
∼
3.5
Å.
49
The
basal
spacing
(d
(001)
)
of CdPS
3
is slightly
larger
than
that
of MnPS
3
(6.55
vs 6.49
Å)
due
to the
larger
size
of Cd
2+
over
Mn
2+
(0.95
vs 0.83
Å).
50
Here,
we
exploit
the
ability
of CdPS
3
and
MnPS
3
to host
cations
within
the
van
der
Waals
gap
to generate
materials
that
contain
various
ions
of interest
through
a sequential
ion
exchange
strategy
that
is illustrated
in Figure
1. First,
a cation
with
a small
hydrated
radius
(e.g.,
K
+
) is intercalated
into
the
van
der
Waals
gap
of an
M
PS
3
material.
45
Charge
neutrality
is
maintained
through
the
loss
of the
labile
metal
(Cd
2+
or Mn
2+
)
from
the
metal
layer
(eq
1),
resulting
in negatively
charged
sheets
of
M
PS
3
sandwiching
positively
charged
hydrated
ions.
The
hydrated
K
+
ions
can
be further
exchanged
to introduce
a
larger
hydrated
cation
(
A
),
like
Li
+
, Na
+
, Mg
2+
, Ca
2+
, Zn
2+
, and
Al
3+
(eq
2).
Direct
insertion
of large
hydrated
cations
is
kinetically
limited;
as such,
this
sequential
exchange
method
is
preferred:
M
M
y
M
PS
KCl
K
PS
H
O
Cl
x
x
3
(aq)
2
1
3
2
2(aq)
+
·
+
(1)
M
y
A
A
M
z
K
PS
H
O
Cl
PS
H
O
KCl
x
x
n
x
n
x
2
1
3
2
(aq)
2
/
1
3
2
(aq)
·
+
·
+
(2)
Additionally,
we demonstrate
that
the
H
2
O molecules
can
be
exchanged
for
aprotic
ligands,
e.g.,
acetonitrile
(MeCN),
and
tetrahydrofuran
(THF).
The
ligand
exchange
serves
several
purposes:
(1)
to probe
the
effect
of different
ligand
molecules
on ionic
mobilities,
(2)
to demonstrate
that
the
mobile
ions
conduct
in the
absence
of H
+
, and
(3)
to broaden
the
applicability
of
these
frameworks
to
H
2
O-incompatible
systems.
The
ability
of
M
PS
3
materials
to intercalate
a wide
variety
of
guest
ions
and
molecules
has
been
useful
for
diverse
Figure
1.
Schematic
of the
ion
and
ligand
exchange
used
in the
present
study.
In an aqueous
KCl
solution,
CdPS
3
intercalates
hydrated
K
+
ions
into
the
van
der
Waals
gap
and
maintains
charge
neutrality
by losing
Cd
2+
ions
from
the
metal
layer.
After
K
0.5
Cd
0.75
PS
3
is formed
one
can
perform
either:
(1)
second
ion
exchange
to introduce
desired
monovalent
or multivalent
mobile
ions,
or (2)
ligand
exchange
to replace
H
2
O in the
system.
Journal
of
the
American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.4c06263
J. Am.
Chem.
Soc.
2024,
146, 24398
−
24414
24399
applications.
The
areas
of research
span
from
nonlinear
optics,
51,52
photoluminescence,
53
hosts
for
biomolecules
54
or
polymers,
55,56
and
superconducting
magnets.
57
However,
the
ionic
conductivity
of these
materials
has
been
less
rigorously
explored.
Some
studies
have
investigated
the
mobility
of
hydrated
Na
+
, K
+
, and
Cs
+
in CdPS
3
.
58
−
61
These
studies
concluded
that
Na
+
was
slightly
mobile
(10
−
5
S cm
−
1
) but
that
K
+
and
Cs
+
were
immobile.
Another
study
investigated
the
conductivity
of trivalent
cations
in
M
PS
3
in the
dried
state,
finding
no significant
ionic
conduction
(10
−
8
−
10
−
10
S cm
−
1
).
62
Notably,
nanosheet-based
membranes
using
Cd
0.85
PS
3
Li
0.15
H
0.15
were
found
to have
exceedingly
high
2D
H
+
conductivity
(300
mS
cm
−
1
at RT
and
98%
relative
humidity
(RH)).
63
Recently,
Yu and
Ren
reported
on CdPS
3
-based
nanosheet
membranes
with
various
intercalated
cations.
64
The
mem-
branes
boasted
impressive
2D
ionic
conductivity
but
importantly,
the
conductivity
was
similar
irrespective
of the
intercalated
cation
(170
−
330
mS
cm
−
1
at RT
and
98%
RH).
In fact,
the
ion-intercalated
membranes
also
displayed
similar
behavior
with
regard
to layer
expansion
and
activation
energy.
This
is likely
due
to excess
H
2
O content
within
the
restacked
membranes
which
could
lead
to significant
contribution
from
H
+
conduction
through
a Grotthus
mechanism,
63,65
or direct
diffusion
of solvated
ions
in a confined
liquid
electrolyte.
Such
a mechanism
is distinct
from
conduction
of ligand-coordinated
ions
within
a host
crystal.
The
transition
from
ligand-
coordinated
solid
state
conduction
to
the
conduction
of
solvated
ions
within
a confined
liquid
in the
solid
framework
at
excess
ligand/solvent
concentrations
has
been
noted
in
previous
studies
of hydrated
Li
2
Sn
2
S
5
, MOFs,
and
clays.
36,66
−
68
To
minimize
contributions
from
H
+
conduction
and
to
reveal
inherent
structure
−
property
relationships
that
can
be
drowned
out
in the
presence
of
excess
H
2
O,
here
we
investigate
bulk,
ion-intercalated
CdPS
3
and
MnPS
3
materials
primarily
at ambient
RH
(
∼
40
to 55%),
and
with
coordinating
ligands
that
range
from
H
2
O to
aprotic,
neutral
solvent
molecules.
This
allows
us
to
understand
the
nuanced
differences
between
the
effects
of
various
frameworks,
intercalated
ions,
and
coordinating
ligands
on
the
ionics,
thereby
deepening
our
understanding
of ionic
conduction
in
solids.
Additionally,
the
effective
bulk
pellet
conduction
enabled
by the
inherent
2D
conduction
channels
of randomly
oriented
particles
in the
polycrystalline
samples
in this
study
can
be directly
applied
to practical
battery
applications.
We
also
introduce
MnPS
3
as an environmental
friendly
alternative
with
comparable
performance,
and
demonstrate
nonaqueous
analogues
that
would
be
more
compatible
with
desirable
electrode
materials.
We
employ
several
characterization
techniques
to study
the
structural
and
chemical
changes
after
ion
exchange
as well
as
the
resulting
ionic
mobility.
These
techniques
include
electrochemical
impedance
spectroscopy
(EIS),
X-ray
diffrac-
tion
(XRD),
thermogravimetric
analysis
(TGA),
inductively
coupled
plasma
mass
spectrometry
(ICP-MS),
scanning
electron
microscopy
(SEM),
energy
dispersive
X-ray
spectros-
copy
(EDS),
quantum
mechanics
simulations,
solid
state
magic
angle
spinning
nuclear
magnetic
resonance
spectroscopy
(MAS
NMR),
and
pulsed-field
gradient
nuclear
magnetic
resonance
spectroscopy
(PFG
NMR).
At ambient
temperature
and
RH,
all of the
H
2
O-coordinated
interlayer
cations,
barring
Al
3+
-
intercalated
MnPS
3
, exhibit
“superionic”,
or practically
useful,
bulk
conductivity
(>0.1
mS
cm
−
1
) and
relatively
low
E
a
.
Notably,
in the
polycrystalline
samples
studied
here,
both
σ
RT
and
E
a
vary
significantly
depending
on
the
identity
of the
intercalated
cation.
The
ligand
exchange
to aprotic
molecules
generally
results
in decreased
σ
RT
and
increased
E
a
, but
achieving
practically
useful
conductivity
is still
possible
using
ligands
like
MeCN.
■
RESULTS
AND
DISCUSSION
Chemical
and
Structural
Characterization
after
Ion
Exchange.
Elemental
Analysis.
M
PS
3
materials
can
undergo
ion
exchange
processes
as outlined
in eqs
1 and
2. The
ion
exchanges
have
been
well
characterized
in
previous
re-
ports.
45,52
However,
the
amount
of
M
that
is replaced
by
A
,
described
by
“
x
”
in eqs
1 and
2, differs
between
various
studies
�
ranging
from
0.1
≤
x
≤
0.25.
43,52,60,64
Intuitively,
x
can
be
controlled
by
varying
the
M
PS
3
:
A
Cl
n
(aq)
ratio,
as
illustrated
in Figure
S1.
Here,
we aimed
for high
A
content
(
x
≃
0.25
for CdPS
3
-based
samples
and
≃
0.20
for MnPS
3
-based
samples)
to incorporate
the
largest
number
of charge
carriers,
Table
1. Stochiometry
of Each
A
2
x
/
n
M
y
PS
3
Compound,
x
and
y
Are Measured
by ICP-MS
and Normalized
to P, Which
Is
Assumed
To Be Constant
a
compound
x
y
H
2
O/f.u.
H
2
O/
A
nominal
formula
K
x
Cd
y
PS
3
0.50
±
0.02
0.79
±
0.02
1.0
±
0.1
2
K
0.5
Cd
0.75
PS
3
·
H
2
O
Li
x
Cd
y
PS
3
0.47
±
0.01
0.85
±
0.03
2.0
±
0.7
4.4
Li
0.5
Cd
0.75
PS
3
·
2H
2
O
Na
x
Cd
y
PS
3
0.51
±
0.02
0.81
±
0.02
2.0
±
0.5
3.9
Na
0.5
Cd
0.75
PS
3
·
2H
2
O
Mg
x
Cd
y
PS
3
0.24
±
0.01
0.78
±
0.05
1.9
±
0.0
7.9
Mg
0.25
Cd
0.75
PS
3
·
1.9H
2
O
Ca
x
Cd
y
PS
3
0.24
±
0.02
0.81
±
0.02
2.0
±
0.5
8.3
Ca
0.25
Cd
0.75
PS
3
·
2H
2
O
Zn
x
Cd
y
PS
3
0.41
±
0.09
0.76
±
0.24
0.25
0.63
Zn
0.4
Cd
0.6
PS
3
·
0.25H
2
O
Al
x
Cd
y
PS
3
0.13
±
0.01
0.81
±
0.08
2.5
±
0.6
14.5
Al
0.17
Cd
0.75
PS
3
·
2.3H
2
O
K
x
Mn
y
PS
3
0.40
±
0.05
0.80
±
0.03
0.8
±
0.2
2.1
K
0.4
Mn
0.8
PS
3
·
0.8H
2
O
Li
x
Mn
y
PS
3
0.34
±
0.03
0.84
±
0.03
1.4
±
0.2
4.0
Li
0.4
Mn
0.8
PS
3
·
1.4H
2
O
Na
x
Mn
y
PS
3
0.59
±
0.08
0.71
±
0.04
1.7
±
0.6
2.7
Na
0.6
Mn
0.7
PS
3
·
1.7H
2
O
Mg
x
Mn
y
PS
3
0.19
±
0.01
0.82
±
0.01
1.8
±
0.1
9.5
Mg
0.2
Mn
0.8
PS
3
·
1.8H
2
O
Ca
x
Mn
y
PS
3
0.25
±
0.02
0.76
±
0.02
1.5
±
0.1
6
Ca
0.25
Mn
0.75
PS
3
·
1.5H
2
O
Zn
x
Mn
y
PS
3
0.46
±
0.06
0.62
±
0.06
1.6
±
0.1
2.6
Zn
0.4
Mn
0.6
PS
3
·
1.6H
2
O
Al
x
Mn
y
PS
3
0.12
±
0.00
0.84
±
0.01
2.4
±
0.4
16
Al
0.13
Mn
0.8
PS
3
·
2.2H
2
O
a
H
2
O content
of each
compound,
which
is determined
with
a combination
of TGA
and
measured
mass
pre-
and
postdrying
using
an analytical
balance.
The
amount
of H
2
O per
intercalated
ion
in each
material
and
the
nominal
formula
of each
compound
were
based
on the
measured
cation
content.
Journal
of
the
American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.4c06263
J. Am.
Chem.
Soc.
2024,
146, 24398
−
24414
24400