of 35
Supporting Information for: 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,
Raphaële J. Clément,
William A. Goddard III,
and Kimberly A. See
,
Department of Applied Physics and Materials Science, California Institute of Technology,
Pasadena, California 91125, United States
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena,
California 91125, United States
Materials Department and Materials Research Laboratory, University of California, Santa
Barbara, Santa Barbara, California 93106, United States
§
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena,
CA 91125, United States
Materials and Process Simulation Center (MSC), California Institute of Technology, Pasadena,
California 91125, United States
S1
Supplementary Notes
1. Ion exchange Reaction in
M
PS
3
Phases
The ion exchange reaction occurs in a less controlled manner with MnPS
3
than CdPS
3
. With
CdPS
3
, a chelating agent – such as ethylenediaminetetraacetic acid (EDTA) – must be added to
the ion exchange reaction in order for Cd
2+
to leave the CdPS
3
lattice. Whereas, with MnPS
3
the more labile Mn
2+
readily leaves MnPS
3
in the aqueous solution without EDTA. As a result,
the final Mn content of the exchanged materials differs depending on the intercalated ion. For
example, during the second ion exchange to insert Na
+
or Ca
2+
, in addition to replacing the K
+
from the first exchange, some additional Mn
2+
from the metal layer is replaced. However, for the
second exchanges with CdPS
3
no EDTA is added, so no additional Cd
2+
is expected to be leached
from the compound.
Furthermore, the pristine MnPS
3
does not completely react in the conditions used in this study
(small amounts are still present in the XRD, Figure S1). Attempts to use higher molarity KCl
solutions or to add EDTA to facilitate full transformation to
K
0
.
5
Mn
0
.
75
PS
3
often led to exfoliation
or degradation of the material due to loss of excess Mn.
K
0
.
5
Mn
0
.
75
PS
3
could be obtained through a
second successive K
+
exchange, but we found that it did not significantly affect the electrochemical
properties, therefore for simplicity the study was carried out using the product of the single K
+
exchange,
K
0
.
4
Mn
0
.
80
PS
3
.
2. PFG NMR Analysis
The PFG NMR data was processed using TOPSPIN 4.3.0 and fitted using Python 3.12.4 using the
following models:
S2
3D isotropic model
1,2
(Tanner-Stejskal equation):
I
=
I
0
exp
(
D
i
(
γδg
)
2
(
δ
3
))
=
I
0
exp (
D
i
·
b
)
where
b
= (
γδg
)
2
(
δ
3
)
(1)
I
0
is the initial signal intensity, D
i
is the self-diffusion coefficient,
δ
is the gradient duration, and
is the diffusion time. The
δ
,
, and g values were selected to ensure sufficient decay window.
To account for the distribution of diffusion coefficients due to inhomogeneity in sample, random
orientation of grains, geometrical confinement etc., the Tanner-Stejskal equation can be written
as:
3
I
=
I
0
exp (
D
DDC
·
b
)
β
(2)
D
DDC
is distributed diffusion coefficient and
β
is the stretching exponent.
2D anisotropic model:
4,5
I
=
I
0
π/
2
0
sin
θ
exp[
D
eff
(
θ
)
·
b
]
dθ,
where
D
eff
=
D
xy
sin
2
θ
+
D
z
cos
2
θ
(3)
θ
is the angle between the direction of the applied magnetic field gradient and the z axis. D
xy
denotes in-plane diffusivity, and D
z
is the out-of-plane diffusivity. Equation 3 accounts for all pos-
sible grain orientations and their contribution to the observed PFG NMR signal intensity. As we
can observe in Figure S23, the 3D isotropic model using a biexponential fit (two diffusion coeffi-
cients) provides an excellent fit of the
1
H and
7
Li PFG NMR signal attenuation curves. In contrast,
the 2D anisotropic model provides a satisfactory fit of the
7
Li PFG NMR signal attenuation curve
but does not fit the
1
H PFG NMR data well. For the
7
Li PFG NMR data (Figure S23a ), the fit to the
2D diffusional model indicates that the out of plane
1
H diffusion coefficient, D
z
(3.0
×
10
10
m
2
/s),
S3
is significantly gather than D
xy
(1.9
×
10
11
m
2
/s), which is highly unlikely for two-dimensional dif-
fusion for the layered materials. The stretched exponential model exhibits reasonable curve fits but
is not as good as biexponential fit. Though the stretched exponential model shows a better curve fit
than the 2D anisotropic model, it is not due to a distribution of crystallite orientations as the impact
of random orientation is accounted in the 2D model. Taken together, those findings provide strong
evidence for the presence of two distinct
1
H and
7
Li diffusion species in these materials, which
are not significantly impacted by 2D confinement of the layered structure and the distribution of
crystallite orientations. We tentatively attribute those two
1
H/
7
Li diffusional environments to ion
transport at the grain boundaries and in the bulk. Hatz
et al.
also found that
7
Li PFG NMR data
collected on a lithium tin sulfide nanosheets could only be fit using a biexponential 3D diffusional
model and the Tanner-Stejkal equation.
6
S4
Supplementary Figures
Figure S1: XRD patterns of K
+
ion exchange reactions with (a) MnPS
3
and (b) CdPS
3
using dif-
ferent concentrations of KCl
(
aq
)
solutions. For MnPS
3
, XRD reflections pertaining to the pristine
material were still present after a reaction with 3 M KCl. Therefore, under these reaction condi-
tions a two phase mixture of MnPS
3
and K
.
50
Mn
.
75
PS
3
is formed, which corresponds to an overall
stoichiometry of K
.
40
Mn
.
8
PS
3
. The pristine CdPS
3
phase is no longer present after reaction in 2 M
KCl with EDTA and the buffer solution.
S5
Figure S2: A comparison of (a)
σ
RT
and (b) E
a
of pure batches (by elemental analysis) with the
average values of select CdPS
3
-based ion exchange materials. The electrochemical performance
of the pure batches is within error of the overall average, suggesting that the presence of a small
amount of impurities in some batches doesn’t affect the electrochemical performance.
S6