Revised-final

Peculiarities of ion transport in confined-in-ceramics concentrated

polymer electrolyte

R. Blanga,a M. Berman,b M. Biton, c F. Tariq,c V. Yufit, c A. Gladkich,d

S. G. Greenbaum,b N. Brandon, c and D. Golodnitsky a,d,*

a- School of Chemistry, Tel Aviv University; b-Department of Physics and Astronomy,

Hunter College of the City University of New York, NY, 10065, and CUNY Graduate

Center, New York, NY 10036, USA; c- Department of Earth Science and Engineering,

Imperial College London, London, UK; d- Applied Materials Research Center, Tel

Aviv University, Tel Aviv, 69978, Israel

*golod@post.tau.ac.il

Abstract

Polyethylene-oxide/lithium-aluminate films were deposited by electrophoretic

deposition. Films impregnated with lithium iodide formed highly concentrated

polymer-in-ceramic solid electrolytes. Solid-state NMR, FIB-SEM tomography, and

EIS studies showed that only a few percent of the interfacial lithium in the sample is

capable of inducing a fast ion-migration path in the system. We suggest that despite

suppressed crystallinity of PEO confined in ceramics the ion transport in the polymer

medium impedes the total conductivity of the composite electrolyte at near-ambient

temperatures. After melting of the polymer and its complexes, the interfacial

conduction through perpendicular LiAlO2/LiI grain boundaries becomes feasible.

This, together with ion transport via molten, confined polymer electrolyte is followed

by the increase of the overall conductivity of the composite system.

Introduction

The field of solid-state electrochemistry was born in the early 20 th century after the

recognition by Warburg of the phenomenon of pure ionic conductance in some solid

materials [1]. However, rapid progress has only been achieved in the last 30 years by

the discovery of solid electrolytes (superionic solids) with high room-temperature

conductivity and chemical and electrochemical stability. At present, there is great

interest in thin-film lithium-ion-conducting solid electrolytes for application in

batteries and hybrid supercapacitors.

Revised-final

The ionic conductivity in solids is caused by the existence of microscopic defects or

disorder. Defects include point imperfections in ionic crystals, vacancies, interstitials,

antisite and Schottky defects which cause distortion of the lattice in their immediate

neighborhood. The structures that permit fast ion transport are generally disordered,

channeled or layered [2]. Many of the known superionic solids, such as alkali-metal-

conducting beta-alumina, perovskite-type lithium lanthanum titanates, NASICON-

type, LiSICON- and thio-LiSICON, garnet-type conducting oxides, are monovalent

cationic solids.

Amorphous Li-ion conductors include oxide- and sulfide-based glasses, like

Li4+x+δ(Ge1−δ−x-Gax)S4, Li10SnP2S12, Li7P2S8Iand LIPON [3-5]. Increased room-

temperature conductivity of lithium by several orders of magnitude was observed

upon replacement of microcrystalline γ-LiAlO2 with nanocrystalline material [6].

Thin-film ceramic electrolytes are typically produced by vacuum evaporation or

sputtering. It has often been observed that they provide different structure,

morphology and composition than electrolytes obtained by thermal annealing [2].

Polymer ionics is a relative latecomer to the field of solid-state ionics. Although the

complexing ability of oligoethers has been known for some time, Wright and co-

workers were the first to measure the ionic conductivity of poly(ethylene oxide)

(PEO)- salt complexes [7]. Intensive research has been carried out by Armand et.al

and Bruce et al [8, 9]. To date, poly(ethylene oxide)-based polymer electrolytes have

been regarded as one of the most suitable electrolytes for lithium batteries. Despite

over 30 years of worldwide research on polymer electrolytes (PEs), the requirement

of sufficiently high room-temperature cationic conductivity remains inaccessible [9-

13]. Furthermore, there is still considerable scientific controversy over the very nature

of the ion-transport mechanism and the factors governing cation-anion interactions in

polymer-salt complexes based on PEO [10]. Polymer electrolytes, which usually

feature multiphase structures at the microscopic and/or macroscopic levels, make ion

transport very complex. First, the coexistence of different phases, such as an

amorphous phase and various crystalline complexes of PEO and Li+, provides

different pathways for ion transport; in addition, the distribution and structure of the

phases themselves are also intricate. It is established that ion transport in polymer

electrolytes includes the local motion of polymer segments and inter- and intra-chain

ion-hopping between coordinating oxygen sites, which are not fixed and vary with

time and temperature. A number of experimental and theoretical studies of ion

Revised-final

transport in polymer electrolytes have resulted in the identification of a variety of

relevant transport mechanisms, such as the hopping motion of cations forming a

weak coordination shell between Li+ ions and ether oxygens (EO), free ion motion

along percolating channels in the PEO melt, etc. [8–23]. However, among the

different controlling factors, the segmental mobility of the polymer backbone has

been identified as a key factor in cation and anion mobilities. Hundreds of articles

aimed at suppressing the crystallinity of polymer electrolytes have been published and

extensive research is still in progress. The research is focused on the synthesis of new

block-copolymers, new salts with large anions, plasticizing of polymer matrices by

organic and ceramic additives for the improvement of ion conduction that is strongly

coupled to the segmental relaxation.

The issue of the effect of ceramic fillers in polymer electrolytes is very important [24-

Scrosati] and at present there are several publications that contradict the accumulated

knowledge of the composite-polymer field. Many researchers have found that lithium-

ion hopping occurs in a sequential manner on the skin areas of ceramic fillers [25-27].

Wieczorek et al [28] applied the Lewis acid−base theory to explain conductivity

enhancement in the case of polymer electrolytes filled with inorganic nanoparticles. In

[29] it was found that high-aspect-ratio ceramic nanowires filling a polymer matrix

provide long-range Li+-transfer channels, something not possible in the case of

nanoparticles that are randomly distributed in the polymer matrix. Of particular

interest are electrolytes which contain high concentrations of ion charge carriers. The

main idea of the strategy of polymer-in-salt electrolytes [30-33] is to keep to a

minimum the polymer concentration required for good mechanical properties,

allowing the major ion-conduction path to go through the inorganic salt.

Unfortunately, the cast films are brittle and difficult to handle.

Recently, the simple and inexpensive method of electrophoretic deposition was used

by us for the first time to fabricate novel solid ion-conducting polymer-in-ceramic

electrolytes [34, 35]. EPD provides good conformal deposits on complex and uneven

electrode geometries, something particularly important for the elimination of poor

point contacts, which cause high internal resistance of the battery. Structure,

morphology and ion-conduction properties of composite LiAlO2/PEO films, both

pristine and saturated by LiI, have been characterized by a variety of experimental

techniques. The choice of the materials is based on the literature data and our previous

research. In composite polymer electrolytes, γ-LiAlO2 is combined with polyethylene

Revised-final

oxide to improve the mechanical properties, conductivity and the interfacial

electrode/electrolyte stability. In [36, 37] the authors suggested that nanocrystallites of

γ-LiAlO2 [36] and Al2O3 in concentrated LiI:PEO electrolytes [37] create highly

defective interfacial regions with preferable low-energy conduction paths.

We present here a more detailed study of the ion-conduction mechanism of

concentrated confined-in-ceramic LiAlO2/P(EO)n LiI polymer electrolyte by solid-

state NMR, FIB-SEM tomography, and EIS methods. 3D Comsol model, based on

FIB-SEM tomography and conductivity data, has been originally applied to the

composite polymer-in-ceramic electrolyte system.

Experimental

The particle size of the powder in suspension is an important parameter that affects

the topography of the deposited films. Micro- and nanosize LiAlO2 particles were used

for the deposition of composite films. In order to reduce the size of the particles, wet-

milling of 2.0-2.5µm-size LiAlO2 was carried out in a planetary, high-energy ball mill

(PM100 Retsch). 1.0g LiAlO2, 2.5ml PEGDME 500 (Merck), 0.5ml Triton x-100

(Sigma) and 107g ZrO2 3mm-Ø balls were introduced into a 50ml ZrO2 grinding

bowl. The milling was carried out at 450RPM for different periods of time and the

sample was transferred to 300ml ethanol after separation from the balls. The particle-

size distribution was measured with a Zetasizer Nano ZS. It was found that the

process of wet ball milling markedly reduces the average particle size of LiAlO2 to

about 175-190nm after three hours of milling.

LiAlO2-based composite films were deposited by cathodic EPD, with the use of

acetone-based deposition baths of the following composition:250ml acetone, 0.70–

1.84g LiAlO2 (LAO), 2ml water, 0.4ml acetylacetone, 0.027–0.05g iodine, 0–0.15g

PEI, 0.70g PEO and 1ml Triton X-100. A Keithley SourceMeter, model 2400,

interfaced with LabTracer software and a PC, were used to control the EPD process

and to monitor the current and voltage profiles. Deposition of the membranes was

performed on nickel substrates (working electrode) with the use of a graphite plate as

counter electrode. The distance between the working and counter electrodes was

1.5cm. The thickness of the films was about 70–75μm. Samples for electrochemical-

cycling tests were dried under vacuum for 15 hours at 80◦C, and transferred to an

argon-filled glove box (Brown, ≤10 ppm water) for impregnation with 20%w/w LiI in

acetone solution in order to obtain a LiI:P(EO)3 complex and to deposit the excess salt

Revised-final

on the ceramic particles. The samples were again dried under vacuum for six hours at

50C and transferred to the glove box for cell assembly. Electrochemical-impedance

spectroscopy was used to test the conductivity of the composite electrolyte.

Conductivity cells (in the coin-cell 2032 setup) comprise the electrolyte sandwiched

between two blocking nickel electrodes. Tests were carried out at 25 to 110C with a

Solartron Impedance Analyzer 1260 at an amplitude of 60mV and over a frequency

range of 15MHz to 1Hz.

Surface morphology was tested with a Quanta 200 FEG Environmental Scanning

Electron Microscope (ESEM) (JEOL Co.). The ESEM is equipped with an HKL-

EBSD and Oxford-EDS integrated analytical system. TOFSIMS tests were performed

under the following operating conditions: In+ primary ions and beam diameter of 1-

10µm with the use of a TRIFT II (Physical Electronics Inc., USA). Wide-line NMR

measurements were carried out with a Varian direct-drive 300MHz spectrometer. A

saturated aqueous solution of lithium trifluoromethanesulfonate (lithium triflate) was

used as reference, set at 0 ppm for 7Li NMR spectra, which were acquired with the

use of a quadrupolar echo sequence. Measurements were carried out at temperatures

from 25◦C to 80◦C. 7Li Pulsed Field Gradient NMR (PFGNMR) diffusion

measurements were also attempted but were not successful because of very short spin-

spin relaxation times. 6,7Li and 27Al Magic-Angle Spinning (MAS) measurements

were carried out on a Chemagnetics 3.2mm MAS probe. Lithium triflate was used as

a reference set to 0ppm for the 6,7Li, while 1M aqueous AlCl3 was used as a reference

set to 0ppm for the 27Al measurements. Measurements were carried out at 25°C, 50°C

and 80°C. Samples were spun at up to 20kHz. All spectral fits were made with the use

of DMFit NMR software.

Three dimensional (3D) imaging of the composite electrolyte was conducted by FIB-

SEM tomography, which sections an exposed face of a specimen with a Ga+ ion beam

prior to imaging the face with a high-resolution electron beam. The process is

sequentially repeated at regularly spaced intervals until the desired 3D volume size

has been acquired for analysis, as detailed in several studies [39, 40]. The composite

polymer-in-ceramic electrolyte sample was analysed with a Zeiss Auriga FIB FEG-

SEM 600 using a 5kV electron-accelerating voltage and 1 nA ion-milling current. A

final volume of 3.13 x 3.13 x 0.82 µm was reconstructed at the voxel resolution of

14.6 nm (x/y) and 20 nm (z). Segmentation was carried out following alignment of the

slices and filtering to reduce the noise with the use of an Avizo Fire (FEI, Bordeaux,

Revised-final

France). Quantification was carried out using IQM QuiQ 3D software (IQM

Elements, London, UK).

Results and Discussion

In our previously published article by using TGA and DSC tests, we have found that

when the EPD is carried out from the acetone-based suspension with 1:1 polymer-to-

ceramic ratio, the film contains 60–72% of LiAlO2, depending on the total solid load

[34]. These films with the maximal porosity of 74% have been chosen for the study of

confined-in-ceramic polymer electrolytes. As mentioned in the Introduction, in [37,

38] it was found that composite LiI-PEO-Al2O3 electrolytes with very high salt

concentration (EO-to-LiI ratio varying from 1.5 to 3) prepared by casting, exhibit a

different ion-conduction mechanism than do more dilute polymer electrolytes; in the

latter case the mechanism is highly dependent on segmental motions of the polymer.

The chosen concentration of LiI salt in the studied composite system was high in

order to ensure a complete complexation of ethylene oxide units in the host polymer

by lithium cation, and to enable the deposition of the excess of LiI on the top of

lithium aluminate particles. Such systems typically exhibit a decoupled from-

segmental-relaxation ion transport. Worthy of mention is the effect of LiI/LiAlO2

grain boundaries on the possibility of the formation of low-energy conduction paths.

Figures 1-3 show the results of NMR studies of concentrated solid LiI:P(EO)2 PE

confined in LiAlO2. Because of the inability to perform NMR diffusion measurements

(because of short spin-spin relaxation times), the mobility of lithium ions was probed

by measuring the full-width at half-maximum (FWHM) in the static NMR-linewidth

measurements. From observation of the 7Li NMR spectra, it can be seen that as the

temperature increases, the FWHM of the 7Li peak decreases and, in addition, there is

clear emergence of a more mobile Li+ fraction from the less mobile lithium-ion

environments, like structural LiAlO2 and Li:P(EO)3 complex (Fig. 1). Motional

narrowing of NMR spectra with increasing temperature in ion conducting solids is a

long-known phenomenon and still is a useful screening tool [41]. In the present case

the narrowing is confined to the polymer fraction of the composite, while the ceramic

linewidth does not change over the present temperature range.

Revised-final

As there is a small chemical shift in the 7Li NMR spectra, it can be assumed that in

the composite sample, the spectral components of mobile lithium cation species

overlap those of the immobile lithium in the LiAlO2. In order to resolve the two sites,

deconvolution was performed with the use of DMfit software. Figure 2 shows the

results of such a deconvolution at 70◦C. Though the choice of this temperature is

somewhat arbitrary, it clearly shows the partially narrowed polymer component

superimposed on the ceramic component. The fraction of the more mobile lithium

ions was about 2% between 25–65C, increasing to 20% at 70C. Further increase in

the mobile fraction is limited by the high LAO content, which exhibits a temperature-

independent linewidth over the range studied.

The appearance of the second, more mobile peak is more resolved in the 7Li MAS

spectrum (Fig. 3). The sample shows the presence of a shoulder at 25°C and 50°C,

and then the emergence of a second peak is clearly seen at 80°C. De-convolution

performed with the use of DMfit shows three resolvable peaks present in the

spectrum, as shown in Figure 4. These are identified as immobile lithium in LiAlO 2

(green), partially mobile lithium in PEO (purple) and mobile interfacial lithium (sharp

line in black). On the basis of prior work on PEO-LiI-Al2O3 composite and milled

mixtures of LiI and Al2O3 [21, 36, 37], we assign the sharp peak to mobile lithium

ions at the ceramic-particle interface with the salt. The 6Li spectrum is shown on the

same stack plot, which reveals two main peaks from the LiAlO2 and the PEO:LiI

complex, with perhaps a hint of the interfacial Li in the noise. The much lower signal-

to-noise ratio of the 6Li spectrum is due to its 6% natural abundance, though

broadening from quadrupole and dipole-dipole effects are much smaller, offering

superior resolution compared to 7Li.

Because NMR is quantitative, it can be seen that the interfacial lithium represents

only a few percent of the total lithium in the sample. An exchange-spectroscopy

experiment was attempted in order to examine the possibility of exchange between

interfacial Li and the PEO:LiI phase. Though we were unsuccessful in performing the

exchange experiment, the resolved difference in resonance frequencies implies that if

exchange between these two species is taking place, it is slower than a few tens of Hz. 27Al MAS measurements were performed on pristine LiAlO2 and solid LiI:P(EO)x:

LiAlO2 electrolyte. No changes were found in the spectra of the electrolyte as

compared to pristine lithium aluminate. The DMfit confirmed the γ-LiAlO2 structure

Revised-final

characterized by a chemical shift of 84ppm, a quadrupolar coupling constant of

3230kHz and chemical-shift anisotropy of η=0.58, which is in agreement with the

literature [37]. The spectrum displayed in Figure 5 reveals a small impurity peak at

12ppm, which appears in the starting material as well. Thus there is no observable

change in the LAO structure resulting from either the EPD process or surface

interactions with lithium ions. Thus the presence of surface reactions between the LiI

and LAO does not produce a sufficient quantity of other aluminum compounds to be

observable by NMR.

XRD and DSC tests of electrophoretically deposited LiAlO2-PEO films, containing 5

to 40% PEO showed that the crystallinity of the confined-in-ceramics polymer is

suppressed [34, 35]. Therefore, improved ionic conductivity of concentrated

composite electrolytes saturated with LiI salt was expected over the wide temperature

range. Figure 6 shows a typical Nyquist plot of the cells comprising composite

LiI:P(EO)1.5:LiAlO2 electrolyte, with a PEO-to-ceramics weight ratio of 3:7

sandwiched between two blocking electrodes. The plot for this electrolyte is

represented by a depressed semicircle with a maximum frequency of 0.5MHz and a

capacitance of 100pF/cm2 at 60◦C. The interpretation of the impedance spectrum is

often based on equivalent-circuit-type models that are used to approximate the

physicochemical processes that occur in the cell. It is generally accepted that the

conductivity of polymer electrolytes, calculated from the high-frequency intercept of

the arc with the X-axis is attributed to the bulk conductivity of the electrolyte and it

obeys Arrhenius or VTF-temperature dependence. In a confined-in-ceramic polymer

electrolyte, the near-ambient conductivity (1) calculated from the high-frequency

intercept of the Nyquist plot with the X-axis is about 0.2-0.3mS/cm for the

LiI:P(EO)3-x:LiAlO2 electrolyte, and is almost unaffected by heating up to 110℃ (Fig.

7). The apparent activation energy of 1 conduction (1Ea) is 0.02eV. Such behavior is

unusual for polymer electrolytes and, as in the case of high-concentration

LiI:P(EO)3:Al2O3 electrolyte [36-38], may indicate a strong contribution of the

interfacial grain boundaries. While realizing that the ceramic matrix is polycrystalline,

in the manuscript we refer the term “grain-boundary” to the interface, which forms

between the excess of lithium iodide in polymer electrolyte and ceramic matrix. This

is because the conductivity of ceramics, both of bulk and inter-particles, is much

lower than that of other components of the composite electrolyte and its contribution

is negligible to the overall conductivity of the composite system (see Table 1). The

Revised-final

conductivity, calculated from the resistance of the arc, shows Arrhenius behavior,

with 2 of 2.7x10-5mS/cm at 30C, four orders of magnitude lower than 1. The

apparent activation energy of this conduction process is 1.9eV, which is close to the

data published for composite PEO-based electrolytes, at temperatures below the

melting point of the polymer. 2 increases with temperature and approaches the value

of 1 only at 110C, near the melting points of some Li:P(EO)x complexes (3<x<6).

Contrary to our expectations, the value of 1 of nanosize LiAlO2-based electrolytes

was lower by a factor of 1.7 than that of micron-size electrolytes. However, the low

value of 1Ea (0.2eV) was measured for nanosize-LAO as well. The 2 of the nano-

LAO electrolyte is 1.7x10-4mS/cm at 70C, two orders of magnitude lower than 2 of

the micro-LAO. We attribute this to the very non-homogeneous morphology and

cracking of the film obtained by electrophoretic deposition (Fig. 8), which is caused

by coagulation of PEO micelles. This, in turn, is reflected in non-homogeneous

distribution of lithium salt in the composite system, as shown in the TOFSIMS image

(Fig. 9).

All characteristic peaks of PEO and PEI, i.e. CxHyO and CnHmN fragments with the most

intense peaks of CH3O (m=31), C2H5O (m=45) and C3H7O (m=59) for PEO, and

CH4N (m=30), C2H6N (m=44), C3H6N (m=56) for PEI were detected in the TOFSIMS

mass spectra of the electrophoretically deposited microsize and nanosize LAO-based

composite films. CH2-CH2O species disappear from the spectra in the sample

saturated by LiI. Instead, the spectra show complex ions, LiOCH2 (m=37) and

LiOC2H5 (m=52), indicating the formation of Li-EO complexes. In the micro-LAO

images (not shown here), the LiI2- species, detected in the negative-ion mode, are

arranged in the patterns coinciding with the confined polymer. To the best of our

knowledge, this is the first TOFSIMS detection of the formation of alkali metal-

polyethylene oxide complexes. The fragments of mass 170 and 177 associated with

AlIO and LiIAlO species, respectively, point toward probable lithium salt - ceramics

bonding. The signals, however, were insufficiently strong to be detected by NMR.

The FIB-SEM tomography-segmented data of concentrated polymer electrolyte

confined in ceramics was reconstructed to 3D images (Fig.10a). These were then

used as geometric inputs for modelling. Prior to modelling the real 3D microstructure

data was converted into a volumetric finite-element (FEM) mesh with the use of

Simpleware ScanIP combined with FE (Simpleware, Exeter, UK).outputted to

Revised-final

Comsol. A fine mesh size was utilised for the volume, which was further refined,

becoming 64 times volumetrically smaller at all boundary interfaces. The definition of

the boundaries in the model has been made by a watershed-based separation program,

in which the separation of objects occurs at their narrowest point. The program

determines the labels and volume of the objects to be separated, the position of the

central point, the number of voxels, the maximum opening, and the equivalent circle

diameter, which is the diameter of a circle with the same surface as the watershed

[42].

The gradient of the potential V in the 1:1 binary electrolyte can expressed as (1):

∇V =− j❑ +2¿¿¿

Where i is current, c concentration of electrolyte, t+ transference number of the

lithium ion and f± mean activity coefficient.

When trying to build a high fidelity model of conduction in solid polymer electrolyte

at microscale, additional phenomena should be taken into account such as nonlinear

polarizability, ion-ion interactions, crystallinity, effects of fillers and others

Development of such a sophisticated model is beyond the scope of this paper.

Furthermore, in order to simplify the calculation on real 3D dataset the current was

assumed to be transferred by lithium ions only i.e t+ = 1. This assumption is valid in

solid ceramic lithium or sodium conductors as well as in solid polymer electrolytes

when the anion is rigid and large and therefore cannot be transported along the

polymer film [43, 44]

3-D imaging techniques, such as tomography, allow for the direct quantification of the

individual particles. A novel algorithm for particle separation and advanced

quantification to break down the complex 3D structure of the Li-PEO phase (Fig. 10b)

was applied. It can be seen that the confined LiI:P(EO)3+x particle size distribution in

LiAlO2 matrix is not uniform (Fig 10c). The modal radial size of confined PE entity

is ca. 40nm (Fig.10c), in agreement with the XRD data [34]. For modelling of the

polymer-in-ceramic electrolyte, the meshed 3D microstructure was then transferred

into COMSOL (COMSOL Multiphysics, Stockholm, Sweden). The following

lithium-ion-conducting phases were considered:

Revised-final

Table 1: The conductivity of composite electrolyte phases (the data for composite

electrolyte are extracted from Fig.7)

Phases RT conductivity,

S/cm

Conductivity

at 70C, S/cm

-LiAlO2 <10-12 10-10

LiI:P(EO)3+x confined 3*10-8 2*10-5

LiI:P(EO)3+x free-standing 2*10-8 5*10-6

LiI/-LiAlO2

grain boundaries

7*10-5 1*10-4

As mentioned above, for the model, grain boundaries (GB) were randomly introduced

to the structure by applying a watershed segmentation using Avizo fire software. It is

important to note that the GBs are not structured and oriented as in the real volume. In

the model, we apply a potential of 50mV between two sides of the electrode, one side

with the potential and the other side grounded. In this way, the effect of the different

phases on the distribution of the electric potential can be observed. We use the

COMSOL electric currents (ec) module, based on Ohm’s law in a stationary study

with symmetric boundary conditions:

∇ ∙ J=Q j

J=σE+Je

E=−∇V

Where J is current density, Qj is current source Je is external current density, is

conductivity, E is electric field and V is electric potential.

Four different phases, according to the real data which are acquired as: pore (electric

insulator), ceramic bulk (the matrix), LiI:P(EO)3 electrolyte and randomly oriented

GBs, are introduced in the model.

In Figure 10d it is clearly seen that the overall potential gradient is relatively normal

as defined, and the presence of confined electrolyte particles has a minor effect. While

the overall trends in Figures 10d and 10e are similar, the presence of the GB changes

the potential gradient within the structure. It appears that the GBs have a much greater

effect on the electric potential than does any other phase, including the confined

Revised-final

LiI:P(EO)3+x electrolyte. The presence of the GB-phase affects the initial gradient;

some regions now have a different potential than they had without the GB, at the same

distance from the terminal. The potential profile looks a lot more "stepped", with most

of the current passing through the GBs. The same trend in electric potential and the

same effect of GBs compared to the case of a structure without GB, which was

observed in the models at room temperature, is also obtained in the models at 70°C

(Fig. 10f, 10g). The increase in temperature does not dramatically affect the potential

distribution, since the major contribution to the conductivity, the GB conductivity, is

not influenced by temperature. Since the GB conductivity is significantly higher than

that of the other phases, GB conduction dominates.

On the basis of experimental data and with a simplified “brick-wall” approach, we

suggest that the confined-in-ceramic concentrated polymer electrolyte creates four

possible conduction pathways (Fig.11a). The first, particularly pronounced at near-

ambient temperatures, is the low-resistance path along the LiAlO2/LiI grain

boundaries, which are parallel to the current flow. The conductivity of LiAlO2 at RT

is about 10−10S/cm and, therefore, the both second conduction path and the possible

third one - crossing LAO and confined PE grains, are highly much less significant.

The fourth ion pathway is likely to proceed through the PE medium, coupled with and

restricted capacitatively by the GBs perpendicular to the current flow. On being

heated to above 65C, the polymer electrolyte softens. In the case of heterogeneous

salt distribution and incomplete complexation of lithium by ether oxygens, the

remains of the host PEO constitutes a melt. This promotes inter-lamellae and inter-

chain ion transport in confined polymer electrolytes. As a result, the path through

perpendicular LiAlO2/LiI GBs, which actually have the same conductivity as the

parallel LiAlO2/LiI GBs, becomes feasible in agreement with our Comsol model

based upon the tomography data. Here we would like to emphasize that only a few

percent of the interfacial lithium in the sample is capable of inducing the fast ion-

migration path in the system. This is reflected in the increased ionic conductivity of

the system and can be approximated by the equivalent-circuit model (Fig.11b).

Based on the data collected so far, we suggest that the way to lower the negative

effect of the low conductivity of confined-in-ceramic PEO-based polymer electrolyte,

is to produce a film with high content of core/shell LiAlO2/LiI particles and very little

PE, the former to be created by high-energy ball milling. For the electrophoretic

deposition of a film composed of such core/shell particles, a thorough search for the

Revised-final

solvent and/or mixture of solvents is essential, to prevent dissolution of the lithium

iodide shell in the suspension. Experiments are in progress and will be reported

elsewhere. Finally, we would like to mention that polymer-free ceramic electrolytes

are prepared either as thick pellets by hot pressing, or as thin films by physical vapor-

deposition techniques. These electrolytes, as well as cast high-ceramic-content

polymer composites, are very brittle. With the purpose of minimization of the internal

impedance of batteries of different shapes and dimensions, we suggest fabrication of

solid electrolytes of broad-spectrum composition on uneven and complex-geometry

electrodes by the relatively simple and inexpensive electrophoretic deposition process.

Summary 

PEO-LiAlO2 films, containing about 30% polymer and 70% micron-sized lithium

aluminate, have been prepared by a simple electrophoretic-deposition method. The

films were impregnated with lithium iodide in order to obtain the LiI:P(EO)3 complex,

and the excess salt was deposited on top of ceramic particles. TOFSIMS spectra

indicated formation of lithium-polymer complexes. The mechanism of ion transport in

high-concentration polymer-in-ceramic solid electrolyte was studied by solid-state

NMR, FIB-SEM tomography, and EIS methods. Deconvolution of the static and MAS

NMR spectra showed three resolvable peaks, identified as immobile Li+ in LiAlO2,

partially mobile Li+ in PEO, and mobile Li+ at the ceramic-particle interface. The

near-ambient-temperature conductivity of the micron-size LiAlO2 solid electrolyte,

calculated from the high-frequency intercept of the Nyquist plot with the X-axis, was

about 0.2-0.3mS/cm, and the apparent activation energy of conduction was 0.02eV.

Non-homogeneous morphology of nanosize LiAlO2-based films and non-

homogeneous salt distribution, resulted in lower conductivity values of the solid

electrolyte. The studies showed that, while the interfacial lithium represents only a

few percent of the total lithium in the sample, its contribution governs the fast ion

migration in the system. A 3D Comsol model emphasised the strong effect of grain

boundary conductivity on the potential distribution within the composite electrolyte

and supported the experimental data. On the basis of experimental tests, we suggest

an additional line of attack on the problem of low conductivity of confined-in-ceramic

PEO-based polymer electrolyte, which would combine the mechanochemistry (to

produce core/shell LiAlO2/LiI nanoparticles) with EPD to produce a film with high-

ceramic and very low polymer contents.

Revised-final

Acknowledgements

The authors thank the EU FP7 Project “Materials for Aging Resistant Li-ion High

Energy Storage for Electric Vehicle” and are grateful from the Israel Science

Foundation (ISF grant 2797/11 as part of the INREP center) for participation in

funding. The NMR program at Hunter College is supported in part by the U.S. Office

of Naval Research.

Figure captions

Fig.1. Static linewidth 7Li NMR spectra of the composite confined-in ceramic

LiI:P(EO) 3+x:LiAlO2 electrolyte, comprising µm-size lithium aluminate

Fig.2. 7Li NMR spectra and fit of the composite confined-in ceramic LiI:P(EO)3+x:

µm-LiAlO2 electrolyte at 70C

Fig.3. 7Li MAS NMR spectrum of the composite confined-in ceramic LiI:P(EO)3+x :

µm-LiAlO2 electrolyte at different temperatures,

Fig.4. 7Li MAS and 6Li NMR spectra and fit of the composite confined-in ceramic

LiI:P(EO)3+x: : µm-LiAlO2 electrolyte at 80C

Fig. 5. 27Al MAS NMR spectra of polymer ceramic composite at several temperatures

as the as-received µm-size LAO material.

Fig.6. Typical Nyquist plot of the cells comprising composite confined-in ceramic

LiI:P(EO)3+x: µm-LiAlO2 electrolyte (the Nyquist plots are shown only at 70-110C

temperature range in order to demonstrate the way of calculations of the two

conductivities)

Fig.7. Arrhenius plots of of the cells comprising composite confined-in ceramic

LiI:P(EO)3+x: µm-LiAlO2 electrolyte

Fig.8. ESEM images of composite LiI:P(EO)3+x:LiAlO2 electrolyte comprising micro-

and nanosize lithium aluminate

Fig.9. TOFSIMS images of composite LiI:P(EO)3+x:LiAlO2 electrolyte comprising

nanosize lithium aluminate

Fig.10. 3D images of LiI:P(EO)3+x electrolyte-in-ceramic electrolyte of a final

~4.5µm3 volume (a, b); LiI-PEO Particle Equivalent Spherical Radius Size (c);

Electric potential gradient at room temperature (d, e) and 70C (f, g) of the 3D

composite system free-of-GB (d, f) and containing-2% GB(e, g)

Revised-final

Fig.11 Simplified brick-wall model schematics (a) and equivalent circuit of ion

conduction paths in composite confined-in ceramic LiI:P(EO):LiAlO2 electrolyte

References

1 P. Birke and W. Wepner Solid electrolytes in Handbook of Battery Materials,

(Ed.,: J.O.Besenhard) Verlag Chemie (VCH), 1998, 410

2 S. R. Narayanan and Aswin K. Manohar, B. V. Ratnakumar, in Handbook of

Solid State Batteries 2nd Edition Edited by: Nancy J Dudney, William C West

and Jagjit Nanda (Oak Ridge National Laboratory, USA) 2015, pp. 3-50

3 A. Sakuda, A. Hayashi, T. Ohtomo, S. Hama, M. Tatsumisago, Electrochem.

Solid-State Lett. 13 (2010) A73

4 E. Rangasamy, Z. Liu, M. Gobet,K. Pilar, G. Sahu, W. Zhou, H. Wu, S.

Greenbaum, and C. Liang, J. Am. Chem. Soc. 137 (2015) 1384

5 N. J. Dudney, Materials Science and Engineering B 116 (2005) 245

6 D. Wohlmuth, V. Epp, P. Bottke, I. Hanzu, B. Bitschnau, I. Letofsky-Papst,

M. Kriechbaum, H. Amenitsch, F. Hofer, and M. Wilkening, J. Mater. Chem.

A., 2, (2014) 20295

7 D.E. Fenton, J. M. Parker, P. V. Wright, Polymer 14 (1973) 589

8 M. B. Armand, J. M. Chabagno and M. J., D. in Fast Ion Transport in Solids,

eds J. N. Mundy, P.Vashishta and G.KShenow, (1979) 131, (Elsevier Science

Ltd, New York)

9 M. B. Armand, P. G. Bruce, M. Forsyth and S. Bruno, in Energy Materials,

(eds. D. O’Hare, D. W. Bruce R. I. Walton), Ch. 1, (John Wiley & Sons, Ltd.,

New York, 2011)

10 S. Abbrent, S. Greenbaum, E. Peled and D. Golodnitsky, in Handbook of

Solid State Batteries 2nd Edition, Edited by: Nancy J Dudney, William C West

and Jagjit Nanda, World Scientific Publishing Co. Pte. Ltd.5 Toh Tuck Link,

Singapore 596224, (2015) 523

11 G. Ceder, G. Hautier, A. Jain and S. P. Ong, MRS Bull. 36 (2011) 185

12 B. Scrosati and Garche, J., J. Power Sources 195 (2010) 2419

13 V. Di Noto, S. Lavina, G. A. Giffin, E. Negro and B. Scrosati, Electrochim.

Acta 57 (2011) 4

14 D. Golodnitsky, E. Strauss, E. Peled, and S. Greenbaum Journal of The

Revised-final

Electrochemical Society, 162 (14) (2015) A2551

15 S. Mogurampelly and V. Ganesan, Macromolecules, 48, 2773 (2015). C. A.

Angell, Solid State Ionics, 18 (1986) 72

16 C. A. Angell, Solid State Ionics, 18 (1986) 72

17 B. L. Papke, M.A. Ratner and D. F. Shriver, J. Phys. Chem. Solids, 42 (1981)

493

18 S. D. Druger, M .A. Ratner and A. Nitzan, Solid State Ionics, 9/10 (1983)

1115

19 M. Ratner, P. Johanson and D.F. Shriver, Materials Research Society

Bulletin, 3 (2000) 31

20 Z. Gadjourova, Y. G. Andreev, D. P. Tunstall and P. G. Bruce, Nature, 412

(2001) 520

21 S. H. Chung, y. Wang, S. G. Greenbaum, D. Golodnitsky and E. Peled,

Electrochem. Solid State Lett., 2(11) (1999) 553 

22 D. Golodnitsky and E. Peled, Electrochim. Acta, 45 (2000) 1431

23 Oliver Durr, Wolfgang Dieterich, Philipp Maas and Abraham Nitzan, J. Phys.

Chem. B, 106 (24) (2002) 6149 

24 F. Croce, G. B. Appetecchi, L. Persi, and B. Scrosati Nature 394, (1998) 456

25 A. Dawar, A. Chandra Physics Letters A 376 (2012) 3604

26 P. P. Chu, M. J. Reddy and H. M. Kao, Solid State Ionics, 156 (2003) 141

27 M. Wagemaker, G. J. Kearley, A. A.Well, H. Mutka, and F. M. Mulder, J.

Am. Chem. Soc., 125 (34−36) (2003) 840

28 W. Wieczorek, Z. Florjanczyk and J.R. Stevens, Electrochim. Acta, 40 (1995)

2251

29 W. Liu, N. Liu, J. Sun, P. C. Hsu, Y. Li, H. W. Lee, and Y. Cui, Nano Lett.,

15 (2015) 2740

30 Y. Wang and A. P. Sokolov, Current Opinion in Chemical Engineering, 7

(2015) 113

31 C. A. Angell, C. Liu and E. Sanchez, Nature, 362 (1993) 137

32 A. Ferry, L. Edman, M. Forsyth and D. R. MacFarlane, J. Sun, J. Appl. Phys.,

86 (1999) 2346

33 A. Zalewska, I. Pruszczyk, E. Sulek and W. Wieczorek, Solid State Ionics,

157 (2003) 233

34 R. Blanga, D. Golodnitsky, G. Ardel, K. Freedman, A. Gladkich, Yu.

Revised-final

Rosenberg, M. Nathan, and E. Peled, Electrochimica Acta, 114 (2013) 325

35 R. Blanga, L. Burstein, M. Berman, S. G. Greenbaum, and D. Golodnitsky

J. Electrochemical Society, 162 (11) (2015) D3084

36 G. Ardel, D.Golodnitsky, Yu. Rosenberg and E. Peled, Y.Wang, G.Wang, S.

Bajue and S. Greenbaum. Solid State Ionics, 115 (1998) 477-485

37 Y. Dai, S. Greenbaum, D. Golodnitsky, G. Ardel, Yu. Rosenberg and E.

Peled, Solid State Ionics, 106 (1998) 252

38 D. Golodnitsky, G.Ardel, E.Strauss , E.Peled, Y. Lareah and Yu. Rosenberg

J. Electrochem.Soc, (1997) 144, 10, 3484

39 F. Tariq, R. Haswell, P.D. Lee, D.W. McComb, Acta Materialia 59 (2011)

2109

40 M. D. Uchic, L. Holzer, B. J. Inkson, E. L. Principe and P. Munroe, MRS

Bull., 32 (2007) 408

41 S. Greenbaum and S. Abbrent, Current Opinion in Colloid and Interface

Science, 18, (2013) 228

42 Brabant, L., Vlassenbroeck, J., De Witte, Y., Cnudde, V., Boone, M.N.,

Dewanckele, J., Van Hoorebeke, L. 7(2), (2011). 252

43 T. F. Fuller, M. Doyle and J. Newman, Journal of The Electrochemical

Society, 141, (1994) 1

44 I. J. Curto Sillamoni and M. I. Idiart, Continuum Mechanics and

Thermodynamics, 27, (2014) 941

Revised-final

Revised-final

Fig.4

Revised-final

Fig.6

Revised-final

Fig.7

Revised-final

Revised-final