A new class of electrolytes have been reported, hybridizing aqueous with non-aqueous solvents, which combines non-flammability and non-toxicity characteristics of aqueous electrolytes with the superior electrochemical stability of non-aqueous systems. Here, we report measurements of the structure of an electrolyte composed of an equal-mass mixture of 21 m LiTFSI-water and 9 m LiTFSI-dimethyl carbonate using high-energy x-ray diffraction and polarized neutron diffraction with isotope substitution. Neutron structure factors from partially and fully deuterated samples exhibit peaks at low scattering vector Q that we ascribe to long-range correlations involving both solvent molecules and TFSI anions. We compare both sets of measurements with results of molecular dynamics simulations based on a polarizable force field. The structures derived from simulations are generally in agreement with those measured, except that neutron structure factors predicted for two partially deuterated samples show very intense scattering increasing up to the low-Q limit of simulation, indicating a partial segregation between the two solvents not observed in experimental measurements.

The impact of lithium-ion batteries can be felt ubiquitously in our life, from portable electronics and electric vehicles to grid-storage applications, while their limited energy densities often make our daily schedule, accessibility, and mobility inconvenient. The push for new battery chemistries with higher energy density and increased safety has become ever stronger. Lithium-ion batteries based on aqueous electrolytes attract intense attention due to their intrinsic non-flammable nature. However, the electrochemical instability of water, as characterized by its narrow voltage window of 1.23 V, places an upper limit on their energy densities (<70 Wh/kg), excluding them completely from the category of high-energy rechargeable batteries. This upper limit has recently been breached by a new class of electrolytes, concentrated aqueous solutions of lithium bis(trifluoromethanesulfonyl)imide (LiTFSI).1 Such an aqueous electrolyte can support diverse battery chemistries within its expanded voltage window of 3.0 V.

A new class of electrolytes have been reported, hybridizing water-in-salt with organic solvent-in-salt electrolytes, which combines non-flammability and non-toxicity characteristics of aqueous electrolytes with the superior electrochemical stability of non-aqueous systems.2 It is called hybrid aqueous with non-aqueous (HANE) solvents. The secondary interphasial ingredient introduced by the non-aqueous component, dimethyl carbonate (DMC), helps to expand the electrochemical window of the hybridized electrolyte to 4.1 V, which supports the operation of a 3.2 V aqueous Li-ion battery based on Li4Ti5O12 and LiNi0.5Mn1.5O4 to deliver a high energy density of 165 Wh/kg for >1000 cycles.

Molecular Dynamics (MD) simulations provide detailed information on structural and transport properties of electrolytes that are essential for their choice and optimization for use in advanced batteries. Simulations of an equal-mass mixture of 21 m LiTFSI-water and 9 m LiTFSI-DMC reported in Ref. 2 revealed a pronounced ionic network with TFSI, H2O, and DMC all participating in the Li+ cation coordination, with significant concentrations of (TFSI)2H2O aggregates and Li+2(DMC) solvates.

MD results are dependent on the ability of the force field to accurately capture intermolecular interactions. In conjunction with other experimental techniques, such as nuclear magnetic resonance (NMR), Raman, and FTIR, scattering experiments provide a powerful means for validating the models. We have recently reported structural measurements of LiTFSI-H2O (water-in salt) and LiTFSI-LiOTf-H2O (water in bisalt) electrolytes with high-energy x-ray diffraction (HEXRD) and neutron diffraction with isotope substitution (NDIS)3 and dynamical measurements on the same materials with quasielastic neutron scattering (QENS).4 The results were compared with MD simulations of Borodin et al.1,5,6 and Zhang et al.7 These reproduced the general features of the experimental data and their changes with concentration, but differed in detailed respects, suggesting ways in which their force fields might be modified to better represent the actual systems. Here, we report an in-depth neutron and x-ray scattering study of the structure of the HANE using a composition from Ref. 2. It reveals new structural details due to the use of H2O/D2O and h-DMC/d-DMC isotopic substitutions. MD simulations aided in the interpretation of scattering experiments.

Diffraction measurements were made with HEXRD and NDIS with polarized neutrons, which provides the capability to remove relatively large spin-incoherent scattering, leaving the coherent scattering along with the isotope-incoherent scattering, which in the present case is represented only by a very small6,7 Li contribution. Since this system has two hydrogenous components, water and DMC, samples were prepared with four different combinations of H2O/D2O and h-DMC/d-DMC. X-ray diffraction emphasizes the heavier atoms, while neutron diffraction is more sensitive to correlations involving the hydrogen atoms, but with different weights for H and D. Thus, the measurements provide five independent combinations of correlations between different pairs of atom types. The solutions with different H:D combinations will have the same structure except for quantum differences,8 which will be negligible in the present context, so HEXRD measurements were made only on the samples with H2O and h-DMC.

Dynamical studies on the same samples carried out with QENS and neutron-spin echo scattering will be reported separately.9 

Samples were prepared with four different isotopic compositions: (LiTFSI) (H2O)1.4375(C3O3H6)0.5625, (LiTFSI)(H2O)1.4375(C3O3D6)0.5625, (LiTFSI) (D2O)1.4375(C3O3H6)0.5625, and (LiTFSI)(D2O)1.4375(C3O3D6)0.5625, following the procedures described in Refs. 1 and 2. For convenience, we will refer to these as LiTFSI-H2O-hDMC, LiTFSI-H2O-dDMC, LiTFSI-D2O-hDMC, and LiTFSI-D2O-dDMC, respectively. The components had the highest commercially available purities: LiTFSI: 99.95%, D2O: >99.9%, and D and d-DMC: 99% D. LiTFSI and D2O were obtained from Sigma-Aldrich and d-DMC from CDN Isotopes Inc. The mixing was performed in a glove bag after drying the salts under vacuum at 100 °C. The differences from ideal chemical compositions were all less than 0.1%.

NDIS measurements10,11 with polarized neutrons were made on the D7 instrument12 at the Institut Laue-Langevin (ILL) in Grenoble, France. The incident wavelength was 4.87 Å, giving a limited Q range, 0.1–2.5 Å−1, but data in this range probe intermolecular correlations over relatively long distances, providing a sensitive test of MD predictions. Samples were loaded inside aluminum hollow cylinders, giving a sample thickness of 0.5 mm, and then placed inside a standard orange cryostat and then measured at 315 ± 2 K. Measured transmissions were in the range 0.89–0.91. For each sample, spin-flip and non-spin-flip measurements were taken and used to extract the coherent and incoherent nuclear scattering.12 The correction and data analysis procedures followed those described in Ref. 3, producing the neutron-weighted average structure factor,

SnQ=α,βncαcβbcoh,αbcoh,βSαβQ1α,βncαcβbcoh,αbcoh,β+1,
(1)

where cα and bcoh,α are, respectively, the concentration and coherent scattering length of chemical species α, n is the number of different chemical species, and Sn(Q) and Sαβ(Q) are the Faber–Ziman neutron-weighted average and partial structure factors, respectively.13 

HEXRD measurements were carried out at the BL04B2 beam line14 at the SPring-8 synchrotron radiation facility. The energy of the incident x rays was 61.2 keV. Each sample was loaded into a 2 mm-diameter quartz tube and measured for 2 h. at 299 K. The diffraction data were corrected for polarization, absorption, and background, and the contribution of Compton scattering was subtracted using standard analysis procedures.15 The corrected datasets were normalized following the procedure described in Ref. 3, producing the x-ray-weighted average structure factor Sx(Q) given by Eq. (5) with the neutron scattering lengths replaced by atomic form factors.

The APPLE&P polarizable force field and the simulation methodology followed the procedures described in Refs. 1 and 2. An atomic dipole version of the SWM-4DP water model was used as discussed by Starovoytov et al.16 The Li+ and TFSI charges were +0.94e and −0.94e, respectively, to partially account for the water model polarization reduced from the gas-phase value of 1.44 to 1.0425 Å3.

A multiple time step integrator was employed with three time steps: inner, middle, and outer. An inner time step was set to 0.5 fs for the integration of bonded interactions. A middle time step of 1.5 fs was used for all non-bonded interactions within a truncation distance of 7.5 Å, and an outer time step of 3.0 fs was used for all non-bonded interactions between 7.5 Å and the non-bonded truncation distance Rcutoff = 15 or 20 Å in an additional run. The Ewald summation method was used for electrostatic interactions between permanent charges with other permanent charges or induced dipole moments with k = 83 vectors. The reciprocal part of Ewald was calculated every 3.0 fs. Induced dipoles were found self-consistently with convergence criteria of 5 × 10−14 (electron charge x Å).2 

MD simulations were extended at 333 K from those described in Ref. 2 up to 36 ns total simulation time using Rcutoff = 15 Å. During 36 ns MD simulations, ions exchanged their environments multiple times as the Li+ residence times are as follows: with water (0.88 ns), DMC (4.4 ns), and TFSI (2.9 ns) at 333 K. A second replica was created using the artificially increased repulsion between DMC/DMC before switching back to the original force field, equilibrated at 363 K for 1 ns followed by 22 ns MD simulations using Rcutoff = 15 Å. The composition of the Li+ first coordination shell [the number of O(TFSI), Ow(water), and O(DMC) within 2.8 Å of Li+] for the second replica matched that of the first replica within 0.01. A final configuration of the second replica was used as a starting point of MD simulations using the increased Rcutoff to 20 Å. These simulations were performed for 5 ns.

The x-ray-weighted structure factor for LiTFSI-H2O-hDMC is shown in Fig. 1(a), and the neutron-weighted structure factors for the four isotopic compositions are shown in Fig. 1(b), along with the corresponding MD predictions. The x-ray-weighted structure factor is also shown in Fig. 1(b) over the reduced Q range. Experimental results are represented by symbols, and MD predictions are represented by two curves corresponding to two independent replicas with different initial conditions and Rcutoff = 15 Å. The neutron-weighted structure factors display systematic changes with deuteration. On deuterating the DMC molecules [passing from SA(Q) to SB(Q)], the main peak at 1.0 Å−1 is joined by a second peak at 1.5 Å−1. This second peak also shows up in the x-ray result, Sx(Q). On passing from SB(Q) to SC(Q), when only the water molecules are deuterated, and further to SD(Q) where both DMC and water molecules are deuterated, the peak at 1.5 Å−1 gains intensity and the one at 1.0 Å−1 loses intensity.

FIG. 1.

Measured (symbols) and MD predicted (curves) structure factors: (a) x-ray-weighted average structure factor over the full Q range of the measurement; (b) neutron-weighted average structure factors over the full Q range of measurements, along with the x-ray-weighted average structure factor over the same Q range. The subscripts x, A, B, C, and D denote results from x-ray diffraction from LiTFSI-H2O-hDMC and neutron diffraction from (A) LiTFSI-H2O-hDMC, (B) LiTFSI-H2O-dDMC, (C) LiTFSI-D2O-hDMC, and (D) LiTFSI-D2O-dDMC, respectively; the additional subscripts MD and MDa denote the corresponding MD predictions for two independent replicas with different initial conditions.

FIG. 1.

Measured (symbols) and MD predicted (curves) structure factors: (a) x-ray-weighted average structure factor over the full Q range of the measurement; (b) neutron-weighted average structure factors over the full Q range of measurements, along with the x-ray-weighted average structure factor over the same Q range. The subscripts x, A, B, C, and D denote results from x-ray diffraction from LiTFSI-H2O-hDMC and neutron diffraction from (A) LiTFSI-H2O-hDMC, (B) LiTFSI-H2O-dDMC, (C) LiTFSI-D2O-hDMC, and (D) LiTFSI-D2O-dDMC, respectively; the additional subscripts MD and MDa denote the corresponding MD predictions for two independent replicas with different initial conditions.

Close modal

The most pronounced effect of deuteration is seen in three neutron-weighted structure factors for partially or fully deuterated samples, all of which show very high levels of coherent scattering at Q values below 1.0 Å−1, with shallow broad peaks around 0.35 Å−1 in partially deuterated samples and around 0.5 Å−1 in the fully deuterated one.

The behaviors of Sx(Q) and SA(Q) generally resemble the corresponding results for the 21 m LiTFSI-H2O electrolyte reported in Ref. 3, but the pronounced shoulder in low-Q peak in the neutron Sn(Q) at 0.5 Å−1 in the latter is barely visible in SA(Q). On the other hand, in the 21 m LiTFSI-D2O electrolyte, the double main peak centered at 2.0 Å−1 is replaced by a broad peak centered at 1.5 Å−1 in SA(Q), and the sharp low-Q peak at 0.5 Å−1 is replaced by the high level of coherent scattering referred to above. Overall, the striking consequences of adding the second solvent (DMC) observed in the present work contrast with the relatively minor effects of adding a second salt (LiOTf) reported in Ref. 3.

Turning now to MD predictions, denoted by curves in Fig. 1, they agree almost perfectly with the measured structure factors in the Q range beyond 2 Å−1, dominated by intramolecular correlations. In the region of the main peaks at 1.0 and 1.5 Å−1, the agreement is also rather good, except that the double peak structure exhibited by Sx(Q) and SB(Q) appears as a single broad peak in MD predictions. On the other hand, the predicted shapes for the main peak in SA(Q), SC(Q), and SD(Q) are in line with experimental results. Increasing the non-nonbonded interaction cutoff from 15 to 20 Å in simulations does not change S(Q) for Q > 0.6 Å−1 but results in slightly smaller peaks for the two partly deuterated samples at Q around 0.4 Å−1, as shown in Fig. S1 of the supplementary material.

The most dramatic discrepancy between MD predictions and experimental results appears in the low-Q behavior for partially deuterated samples. SBMD(Q) and SCMD(Q) both rise from minima around 0.7 Å−1 to high values at the low-Q limit of the MD around 0.2 Å−1. This behavior is largely absent in the measured x-ray and fully hydrogenated and fully deuterated neutron structure factors, in the last case following very well the high level of coherent scattering, at Q values below 1.0 Å−1, including the shallow peak around 0.5 Å−1. These behaviors indicate a partial segregation of the DMC and water solvents that shows up only in predicted structure factors with high contrast between water and DMC.

Molecular/ionic partial structure factors can shed light on the features observed in the measured x-ray- and neutron-weighted structure factors,

FABQ=αεA,βεBnAnBcαcβbcoh,αbcoh,βSαβQ1α,βncαcβbcoh,αbcoh,β.
(2)

MD predictions for the partial structure factors for A, B = TFSI, water, and DMC are shown in Fig. 2 for three cases: the fully hydrogenated sample, the one with only DMC deuterated, and the fully deuterated. The total neutron-weighted structure factors are also reproduced for comparison.

FIG. 2.

Molecular/ionic structure factors [Eq. (2)] for (a) LiTFSI-H2O-hDMC, (b) LiTFSI-H2O-dDMC, and (c) LiTFSI-D2O-dDMC derived from MD simulations with Rcutoff = 15 Å. The totals, shown in black curves, correspond to [SAMD(Q)-1], [SBMD(Q)-1], and [SDMD(Q)-1], respectively, as shown in Fig. 1(b). Correlations involving TFSI are plotted as continuous curved, and those involving only water and DMC are plotted as broken curves. All three panels have the same scales to bring out the variations with partial and total deuteration.

FIG. 2.

Molecular/ionic structure factors [Eq. (2)] for (a) LiTFSI-H2O-hDMC, (b) LiTFSI-H2O-dDMC, and (c) LiTFSI-D2O-dDMC derived from MD simulations with Rcutoff = 15 Å. The totals, shown in black curves, correspond to [SAMD(Q)-1], [SBMD(Q)-1], and [SDMD(Q)-1], respectively, as shown in Fig. 1(b). Correlations involving TFSI are plotted as continuous curved, and those involving only water and DMC are plotted as broken curves. All three panels have the same scales to bring out the variations with partial and total deuteration.

Close modal

For all three samples, the increase in the total neutron-weighted structure factor from 2.0 to 2.5 Å−1 is clearly dominated by intermolecular TFSI–TFSI correlations and agrees with measurements. The contributions to the main peak in the 1.0–1.5 Å−1 region are more complicated. In the fully hydrogenated sample [Fig. 2(a)], the strong positive TFSI–TFSI correlation is largely canceled by the strong negative TFSI–DMC correlation. However, because (1) the TFSI–TFSI correlation is slightly broader and shifted to higher Q, (2) the TFSI–DMC correlation has a positive overshoot at higher Q, and (3) there is a small contribution from the DMC–DMC, the total neutron-weighted structure factor exhibits a composite peak centered at 1.1 Å−1. As seen in Fig. 1(b), this reproduces the measured SA(Q) quite nicely.

In the sample with deuterated DMC, TFSI–TFSI correlations [item (1) above] naturally carry less weight and those involving DMC (2, 3) peak at higher Q, leading to a composite peak in the total neutron-weighted structure factor around 1.4 Å−1. The peak in the measured SB(Q) [Fig. 1(b)] is also shifted to higher Q but exhibits a double structure, suggesting that contributions (2) and (3) peak at slightly different Q values. In the fully deuterated sample, the peak moves out to 1.5 Å−1 and is in good agreement with the measurements.

The very strong DMC–DMC correlation at low Q in the sample with deuterated DMC confirms the suggestion made in Sec. III A of a partial segregation between water/water and DMC/DMC in the MD prediction. In the sample with only the water deuterated, the water–water correlation has a similar, but slightly smaller, effect due to the lower deuterium loading. In the fully deuterated sample, large positive water–water and DMC–DMC correlations are effectively canceled by the large negative water–DMC correlation. Thus, the segregation between water and DMC in the MD prediction only shows up in the sample with high contrast, i.e., with one, but not both, of the two solvents deuterated. This segregation is visible in the snapshot of an MD configuration shown in Fig. 3.

FIG. 3.

A snapshot of the MD cell. DMC molecules are denoted as blue transparent isosurfaces, and water molecules are denoted as yellow isosurfaces using van der Waals radii around atoms as defined in Jmol.17 The linear dimension of the simulation cell is 46.27 Å using Rcutoff = (a) 15 Å and (b) 20 Å.

FIG. 3.

A snapshot of the MD cell. DMC molecules are denoted as blue transparent isosurfaces, and water molecules are denoted as yellow isosurfaces using van der Waals radii around atoms as defined in Jmol.17 The linear dimension of the simulation cell is 46.27 Å using Rcutoff = (a) 15 Å and (b) 20 Å.

Close modal

It remains to explain the high coherent scattering in the region below 1.0 Å−1 observed in partially and fully deuterated samples, only in the latter case followed by the MD prediction. Peak amplitudes of this scattering that are around 0.9, 0.6, and 0.4 in the two partially deuterated and fully deuterated samples, respectively, suggest an ordering of water and DMC molecules on a length scale of about 12 Å, similar to, but not as drastic as, the segregation observed in MD predictions, combined with a slightly longer-range ordering associated with the TFSI anions.

The addition of a second solvent—dimethyl carbonate—to the water-in-salt electrolyte LiTFSI-H2O is found to have striking effects on the intermediate-range order probed by HEXRD and NDIS, in contrast to the relatively minor effects of the addition of a second salt that we studied previously. These changes are observed both in the main diffraction peak and at smaller scattering vectors, where substantial coherent scattering is observed in both partially and completely deuterated samples. These are ascribed to long-range correlations, on the order of 10–20 Å, involving both solvent molecules and TFSI anions.

The structures derived from MD simulations reported in Ref. 2 are generally in agreement concerning the main peak in the structure factors around 1.0–1.5 Å depending on the type of deuteration. On the other hand, structure factors predicted for the two partially deuterated samples show scattering increasing up to the low-Q limit of simulation, indicating a partial segregation between the two solvents not observed in experimental measurements.

See the supplementary material for a figure comparing the neutron- and x-ray-weighted structure factors with the corresponding MD predictions with Rcutoff = 15 and 20 Å and an archive containing the force field and MD simulation.

The modeling at the Army Research Laboratory was supported by the Joint Center for Energy Storage Research (JCESR), a Department of Energy, Energy Innovation Hub. The HEXRD experiments were carried out with the approval of the Japan Synchrotron Radiation Research Institute (Proposal No. 2020A2053). The neutron diffraction measurements were performed through the Easy Access program of the Institut Laue-Langevin.

The authors have no conflicts to disclose.

Marie-Louise Saboungi: Conceptualization (equal); Project administration (equal); Writing – original draft (equal); Writing – review & editing (equal). Oleg Borodin: Investigation (equal). David L. Price: Writing – original draft (equal); Writing – review & editing (supporting). Bela Farago: Investigation (supporting). Miguel A. González: Investigation (equal); Writing – review & editing (equal). Shinji Kohara: Investigation (supporting). Lucile Mangin-Thro: Investigation (supporting). Andrew Wildes: Investigation (supporting). Osamu Yamamuro: Investigation (supporting).

The neutron diffraction data are available at https://doi:10.5291/ILL-DATA.EASY-920 and https://doi: 10.5291/ILL-DATA.EASY-1061. The x-ray diffraction and MD simulation data are available from the corresponding author upon reasonable request. An archive containing the force field and MD simulation files is available as the supplementary material.

1.
L.
Suo
,
O.
Borodin
,
T.
Gao
,
M.
Olguin
,
J.
Ho
,
X.
Fan
,
C.
Luo
,
C.
Wang
, and
K.
Xu
, “
‘Water-in-salt’ electrolyte enables high-voltage aqueous lithium-ion chemistries
,”
Science
350
,
938
943
(
2015
).
2.
F.
Wang
,
O.
Borodin
,
M. S.
Ding
,
M.
Gobet
,
J.
Vatamanu
,
X.
Fan
,
T.
Gao
,
N.
Edison
,
Y.
Liang
,
W.
Sun
,
S.
Greenbaum
,
K.
Xu
, and
C.
Wang
, “
Hybrid aqueous/non-aqueous electrolyte for safe and high-energy Li-ion batteries
,”
Joule
2
,
927
937
(
2018
).
3.
M. A.
González
,
H.
Akiba
,
O.
Borodin
,
G. J.
Cuello
,
L.
Hennet
,
S.
Kohara
,
E. J.
Maginn
,
L.
Mangin-Thro
,
O.
Yamamuro
,
Y.
Zhang
,
D. L.
Price
, and
M.-L.
Saboungi
, “
Structure of water-in-salt and water-in-bisalt electrolytes
,”
Phys. Chem. Chem. Phys.
24
,
10727
(
2022
).
4.
M. A.
González
,
O.
Borodin
,
M.
Kofu
,
K.
Shibata
,
T.
Yamada
,
O.
Yamamuro
,
K.
Xu
,
D. L.
Price
, and
M.-L.
Saboungi
, “
Nanoscale relaxation in ‘water-in-salt’ and “water-in-bisalt” electrolytes
,”
J. Phys. Chem. Lett.
11
,
7279
7284
(
2020
).
5.
L.
Suo
,
O.
Borodin
,
W.
Sun
,
X.
Fan
,
C.
Yang
,
F.
Wang
,
T.
Gao
,
Z.
Ma
,
M.
Schroeder
,
A.
von Cresce
,
S. M.
Russell
,
M.
Armand
,
A.
Angell
,
K.
Xu
, and
C.
Wang
, “
Advanced high-voltage aqueous lithium-ion battery enabled by ‘water-in-bisalt’ electrolyte
,”
Angew. Chem., Int. Ed.
55
,
7136
7141
(
2016
).
6.
O.
Borodin
,
L.
Suo
,
M.
Gobet
,
X.
Ren
,
F.
Wang
,
A.
Faraone
,
J.
Peng
,
M.
Olguin
,
M.
Schroeder
,
M. S.
Ding
,
E.
Gobrogge
,
A.
von Wald Cresce
,
S.
Munoz
,
J. A.
Dura
,
S.
Greenbaum
,
C.
Wang
, and
K.
Xu
, “
Liquid structure with nano-heterogeneity promotes cationic transport in concentrated electrolytes
,”
ACS Nano
11
,
10462
10471
(
2017
).
7.
Y.
Zhang
,
N. H. C.
Lewis
,
J.
Mars
,
G.
Wan
,
N. J.
Weadock
,
C. J.
Takacs
,
M. R.
Lukatskaya
,
H.-G.
Steinrück
,
M. F.
Toney
,
A.
Tokmakoff
, and
E. J.
Maginn
, “
Water-in-salt LiTFSi aqueous electrolytes. 1. Liquid structure from combined molecular dynamics simulation and experimental studies
,”
J. Phys. Chem. B
125
,
4501
4513
(
2021
).
8.
Y. S.
Badyal
,
D. L.
Price
,
M.-L.
Saboungi
,
D. R.
Haeffner
, and
S. D.
Shastri
, “
Quantum effects on the structure of water at constant temperature and constant atomic density
,”
J. Chem. Phys.
116
,
10833
10837
(
2002
).
9.
O.
Borodin
,
B.
Farago
,
G.
Guenther
,
M. A.
González
,
D. L.
Price
,
M.
Russina
,
M.-L.
Saboungi
,
T.
Yamada
, and
O.
Yamamuro
(to be published).
10.
B.
Farago
and
A.
Wildes
, Qualifying Coherent Incoherent separation WASP vs D7,
Institut Laue-Langevin (ILL)
,
2021
.
11.
M. A.
González
and
L.
Mangin-Thro
, Quantifying coherent/incoherent contribution in LiTFSI-DMC-D2O mixtures (related to WASP experiment 6-03-454 and previous EASY-920),
Institut Laue-Langevin (ILL)
,
2021
.
12.
J. R.
Stewart
,
P. P.
Deen
,
K. H.
Andersen
,
H.
Schober
,
J.-F.
Barthélémy
,
J. M.
Hillier
,
A. P.
Murani
,
T.
Hayes
, and
B.
Lindenau
, “
Disordered materials studied using neutron polarization analysis on the multi-detector spectrometer, D7
,”
J. Appl. Cryst.
42
,
69
84
(
2009
).
13.
T. E.
Faber
and
J. M.
Ziman
, “
A theory of the electrical properties of liquid metals, III. The resistivity of binary alloys
,”
Philos. Mag. A
11
,
153
173
(
1965
).
14.
S.
Kohara
,
M.
Itou
,
K.
Suzuya
,
Y.
Inamura
,
Y.
Sakurai
,
Y.
Ohishi
, and
M.
Takata
, “
Structural studies of disordered materials using high-energy x-ray diffraction from ambient to extreme conditions
,”
J. Phys.: Condens. Matter
19
,
506101
(
2007
).
15.
K.
Ohara
,
Y.
Onodera
,
M.
Murakami
, and
S.
Kohara
, “
Structure of disordered materials under ambient to extreme conditions revealed by synchrotron x-ray diffraction techniques at SPring-8—Recent instrumentation and synergic collaboration with modelling and topological analyses
,”
J. Phys.: Condens. Matter
33
,
383001
(
2021
).
16.
O. N.
Starovoytov
,
O.
Borodin
,
D.
Bedrov
, and
G. D.
Smith
, “
Development of a polarizable force field for molecular dynamics simulations of poly (ethylene oxide) in aqueous solution
,”
J. Chem. Theory Comput.
7
(
6
),
1902
1915
(
2011
).
17.
See http://www.jmol.org/ for Jmol: an open-source Java viewer for chemical structures in 3D.

Supplementary Material