German Edition: DOI: 10.1002/ange.201814222Superionic ConductorInternational Edition: DOI: 10.1002/anie.201814222

Boosting Solid-State Diffusivity and Conductivity in LithiumSuperionic Argyrodites by Halide SubstitutionParvin Adeli+, J. David Bazak+, Kern Ho Park, Ivan Kochetkov, Ashfia Huq,Gillian R. Goward,* and Linda F. Nazar*

Abstract: Developing high-performance all-solid-state batter-ies is contingent on finding solid electrolyte materials with highionic conductivity and ductility. Here we report new halide-richsolid solution phases in the argyrodite Li6PS5Cl family,Li6@xPS5@xCl1+x, and combine electrochemical impedancespectroscopy, neutron diffraction, and 7Li NMR MAS andPFG spectroscopy to show that increasing the Cl@/S2@ ratio hasa systematic, and remarkable impact on Li-ion diffusivity inthe lattice. The phase at the limit of the solid solution regime,Li5.5PS4.5Cl1.5, exhibits a cold-pressed conductivity of 9.4:0.1 mS cm@1 at 298 K (and 12.0: 0.2 mScm@1 on sintering)—almost four-fold greater than Li6PS5Cl under identical proc-essing conditions and comparable to metastable superionicLi7P3S11. Weakened interactions between the mobile Li-ionsand surrounding framework anions incurred by substitution ofdivalent S2@ for monovalent Cl@ play a major role in enhancingLi+-ion diffusivity, along with increased site disorder anda higher lithium vacancy population.

All-solid-state Li-ion batteries eliminate the flammableliquid organic electrolyte in Li-ion batteries by implementinga solid-state electrolyte (SSE), hence potentially increasingsafety and also volumetric energy density by allowing moreefficient packaging. Developing high performance cells is—amongst many other factors—dependent on the implementa-tion of SSEs that possess high ionic conductivities at roomtemperature, the ability to form good interfaces and whichpresent scalable synthetic routes to their realisation.[1]

Although a number of sulfide, oxide, and phosphate fastion-conducting solids are now known,[2–4] many exhibit draw-backs such as poor mechanical properties, and difficulty intheir processing. For example, garnets based on Li7La3Zr2O12

are popular owing to their chemical stability and highconductivity,[5, 6] but they have the detriment of very lowductility and high-cost precursors. Sulfide-based materials, onthe other hand, exhibit the highest ductility of all the abovecandidates, which is important to optimise solid-solid inter-faces.

Li-argyrodites, Li6PS5X (X = Cl, Br), fall into this latterfamily and are easily synthesized using inexpensive precur-sors. Their cubic crystal structures F(43m

E Care comprised of

PS43@ tetrahedra, with isolated S2@ and X@ ions disordered

over the 4a and 4 c Wyckoff sites in the lattice, and Li+ ionsites that form cagelike Frank-Kasper polyhedra around theanions.[7] While room-temperature Li-ion conductivities ini-tially reported for the Cl (1.9 mS cm@1)[8] and Br phases(0.7 mS cm@1),[9] were within practical ranges, recent valuesare even higher and comparable to those of liquid electro-lytes. Total ionic conductivities are heavily influenced by thesynthesis method, grain boundary contributions, and methodsused for conductivity measurements, including sinteringcold-pressed pellets.[10] Solution-based Li6PS5X synthesisroutes typically give lower ion conductivities (10@5–10@4 mScm@1)[11–13] owing to phase impurities althoughrecent studies using this method report high values forLi6PS5Cl and mixed anion Li6PS5(Cl,Br) argyrodites of 2.4and 3.9 mS cm@1, respectively;[14] and 3.1 mScm@1 forLi6PS5Br.[15] A conductivity of 3.15 mS cm@1 for Li6PS5Clwas reported prepared by a rapid solid-state route[16] anda value of 5 mS cm@1 was recently achieved upon long-termannealing, albeit using pellets pressed at eight tons.[17]

Li6PS5Cl is consequently a promising candidate for all-solid-state Li-ion batteries.[18–22]

The discovery of new and advanced solid electrolytes hasbeen hindered by incomplete understanding of the funda-mental descriptors that dictate ionic mobility. RegardingLi6PS5X (X = Cl, Br, I) studies of the role of anion polar-izability[23] and halogen disorder are prominent. Early on, itwas demonstrated that the very low conductivity of Li6PS5I(& 4 X 10@4 mS cm@1), is due to the ordering of the larger I@

ion on the 4 a site, compared to the disorder of Cl@/Br@ ionsover both the 4 a and 4c crystallographic sites which invokeshigh conductivity.[9] Prior studies of Deiseroth et al., alsonoted narrowing behaviour in the 7Li NMR line-shapes forLi6PS5I that suggested a more confined motion.[7] Recentmolecular dynamics (MD) simulations of Li6PS5X definedthree types of Li-ion jumps in the cagelike polyhedra formedby Li+ in the stable 48h Wyckoff site, and 24g transition site:short-range 48h-24g-48h “doublet” jumps; short-range 48h-48h intracage jumps, and long-range intercage jumps.[24] Thesestudies clearly demonstrated that halide/sulfide disorder is

[*] P. Adeli,[+] K. H. Park, I. Kochetkov, L. F. NazarDepartment of Chemistry and the Waterloo Institute for Nano-technology, University of WaterlooWaterloo, Ontario N2L 3G1 (Canada)E-mail: lfnazar@uwaterloo.ca

J. D. Bazak,[+] G. R. GowardDepartment of Chemistry & Chemical BiologyMcMaster UniversityHamilton, Ontario L8S 4L8 (Canada)E-mail: goward@mcmaster.ca

A. HuqNeutron Scattering Division, Oak Ridge National LaboratoryOak Ridge, TN 37830 (USA)

[++] These authors contributed equally to this work.

Supporting information and the ORCID identification number(s) forthe author(s) of this article can be found under:https://doi.org/10.1002/anie.201814222.

AngewandteChemieCommunications

8681Angew. Chem. Int. Ed. 2019, 58, 8681 –8686 T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

advantageous for the intercage jump that is responsible forlong-range Li-ion diffusion. These MD simulations alsoprobed the effect of introducing more halogens (and hencemore Li vacancies) into the argyrodite to form the theoreticalcompositions Li5PS4X2, where the halides occupy all of the 4aand 4c sites. The conclusions were that increasing the halogencontent does not significantly alter the conductivity for X =

Cl, Br.The above motivated us to develop halide-rich composi-

tions Li6@xPS5@xCl1+x (x, 0.5) where halide/sulfide disorder ismaintained on the 4a and 4c sites. Here, we elucidate thestructure of the new single-phase Li5.5PS4.5Cl1.5 argyrodite byneutron diffraction; and examine ion conductivity anddiffusivity in the solid solution series Li6@xPS5@xCl1+x (x< 1)by a combination of electrical impedance spectroscopy (EIS),and pulsed field gradient (PFG) NMR. The latter technique isincreasingly more popular as advances in fast ion conductionhave made it more accessible to the rates of motion probed bythe PFG experiment.[25–28] The most halide-rich composition,Li5.5PS4.5Cl1.5, exhibits the highest Li+ conductivity of 9.4:0.1 mS cm@1 at 25 88C—which is about four-fold above thatLi6PS5Cl prepared under identical processing conditions—and a low activation energy (Ea) of 0.29 eV. Sintering resultsin even higher conductivity of 12: 0.2 mScm@1. 7Li PFGNMR spectroscopy and fast magic-angle spinning (MAS)NMR studies shed light on the underlying chemistry. Thesereveal heightened Li+ diffusivity and weakened interactionsof the Li+ cation with the Li6@xPS5@xCl1+x framework owing tothe substitution of the divalent sulfide for the monovalenthalide, and correlated diffusive motion of the cations.

Targeted argyrodite compositions Li6@xPS5@xCl1+x (x = 0,0.25, 0.375, 0.5, 0.55, 0.6) were synthesized. Figure 1 shows theneutron diffraction pattern and Rietveld refinement of the

most highly ionically conductive material, Li5.5PS4.5Cl1.5 (seeSupporting Information for details). Both the refinement andenergy dispersive X-ray analysis (Figure S1 and Table S1 inthe Supporting Information) reveal a composition in excellentaccord with the targeted value. Only the 48h site is occupiedby Li (see Table 1); S and Cl share the 4 a and 4c sites, with the

additional Cl@ (vis a vis Li6PS5Cl) being distributed evenlyover both sites (Figure 2). Naturally, the higher ratio of Cl@

ions to S2@ ions results in a greater fraction of Li vacanciescompared to Li6PS5Cl, evidenced by the lower Li occupancyon the 48h site. The large atomic displacement parameter forthe Li site (0.075 c2) suggests high mobility of the Li-ion asdiscussed below. The other members of the seriesLi6@xPS5@xCl1+x (0< x< 0.5) were also essentially single-phase based on their XRD patterns (Figure S2). The cubiclattice parameters obtained by full-profile fitting of that datashows a solid solution is adopted over this range (Figure 3a).The lattice volume shrinks monotonically with x, evidencedby a small decrease in a = 9.8598(4) c for Li6PS5Cl to a =

9.8061(1) c for Li5.5PS4.5Cl1.5. This owes in most part to the Li+

vacancies, since S2@ and Cl@ have similar ionic radii (170 and167 pm, respectively). Attempts to introduce additionalchlorine into the structure caused significant exsolvation ofLiCl at x> 0.5 (Figure S2) which is detrimental to ionicconductivity. The solvation limit may be dictated by thermo-dynamic instability of the lattice at high vacancy content.Thus, while Li5PS4Cl2 was proposed to be a stable compositionon the basis of theory,[24] Li5.5PS4.5Cl1.5 is, in fact, the endmember in the argyrodite structure. Analogous attempts toincrease the Br@/S2@ ratio in Li6@xPS5@xBr1+x showed a signifi-cant LiBr fraction even at x = 0.25, indicating solutionbehaviour is not adopted due to the larger radius of Br@

(182 pm). Formation of the theoretically anticipated iodideLi6@xPS5@xI1+x phases[24] would also not likely be possible.

Ionic conductivity for Li6@xPS5@xCl1+x (0, x, 0.6) wasdetermined by EIS at variable temperature. Selected compleximpedance plots at 298 K are shown in Figure 3b for x = 0,0.25 and 0.5 and the resultant data is summarized in Table 2.At 298 K, the CPE/R falls beyond the range of the impedanceanalyzer and the tail of the blocking electrodes was fit toobtain the total conductivity values. EIS was also performedat 195 K for Li5.5PS4.5Cl1.5 (Figure 3c, inset, see Tables S2, S3,Supporting Information for details) but bulk and grain

Figure 1. Rietveld fit of Li5.5PS4.5Cl1.5, refined against time-of-flight(TOF) neutron powder diffraction data collected at 298 K (GOF=3.37,Rwp =4.88%). The black circles denote the observed pattern, the redsolid line indicates the calculated pattern, and the difference map is inblue. Calculated positions of the Bragg reflections are represented bythe vertical tick marks in green, and the ticks for the minority LiClphase (1.7 wt%) are shown in magenta.

Table 1: Atomic coordinates, occupation factor and isotropic displace-ment parameters of Li5.5PS4.5Cl1.5 obtained from Rietveld refinement ofneutron TOF data (space group F(43m), a =9.8061(1) b at 25 88C. Thecomposition from the fit is Li5.47PS4.55Cl1.45, very close to the nominal.

Atom WyckoffSite

x y z SOF Uiso [b2]

Li 48h 0.3173(7) 0.3173 @0.0201(9) 0.456(16) 0.075(4)P 4b 1/2 1/2 1/2 1 0.030(2)Cl1 4a 0 0 0 0.615(17) 0.029(1)S1 4a 0 0 0 0.385(17) 0.029(1)Cl2 4 c 1/4 1/4 1/4 0.834(16) 0.037(1)S2 4 c 1/4 1/4 1/4 0.166(16) 0.037(1)S3 16e 0.1188 @0.1188 0.6188 1 0.050(1)

AngewandteChemieCommunications

8682 www.angewandte.org T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2019, 58, 8681 –8686

boundary contributions could not be deconvoluted,[23] andhence the conductivity represents the total. While Li6PS5Clexhibits a conductivity of si = 2.5 mScm@1 at 298 K, consistentwith the literature (1.1–3.15 mScm@1),[8,14, 16] substitution ofS2@ for Cl@ results in an almost exponential increase withx (Figure 3d). The highest ionic conductivity of 9.4:0.1 mS cm@1 is reached for Li5.5PS4.5Cl1.5, almost a four-foldincrease vis a vis Li6PS5Cl. For comparison, pellets annealed

at 550 88C for 10 min (i.e., sintered)showed an even higher value of12.0: 0.2 mScm@1 as a conse-quence of optimizing grain boun-dary conductivity. The low elec-tronic conductivity of Li5.5PS4.5Cl1.5

(& 3 X 10@9 S cm@1; (Figure S3) ispractically important in determin-ing a transference number close toone.

Activation energies for Li-ionmobility in Li6@xPS5@xCl1+x weredetermined from both EIS andPFG NMR temperature-depen-dent measurements (Table 2).Although comparison of thesetwo values implies considerationof the bulk conductivity from EIS,the ideality of the semicircle andcapacitance in Figure 3c corre-spond to bulk transport,[29] and itis safe to assume that grain boun-dary contributions do not signifi-cantly affect the observed trend inFigure 3c. Li5.5PS4.5Cl1.5 exhibits anactivation energy of 0.29 eV, whichis much lower than that of Li6PS5Cl(0.34 eV; Figure 3d). The activa-tion energies obtained from EIS(Figure S4) follow an inverse cor-relation with conductivity (Fig-ure 3 d). The increase in the sitedisorder (defined as the total frac-tion of Cl@ on the S2@ site) from

61% for Li6PS5Cl to 83 % for Li5.5PS4.5Cl1.5 plays a role indecreasing Ea as discussed below. The effect of this halogendisorder and its interplay with ionic transport is illustrated inFigure 4, which shows the results of 7Li PFG NMR diffusivitymeasurements from 270–340 K. Figure 4a shows the markedincrease on diffusivity with increasing x in Li6@xPS5@xCl1+x.Arrhenius plots of diffusivity and conductivity from both PFGand EIS measurements, respectively for x = 0 and x = 0.5

Figure 2. Crystal structure of Li5.5PS4.5Cl1.5 showing the PS43@ tetrahedra, the cagelike polyhedra formed by Li+ ions, free S2@/Cl@ anions, and

comparison of occupancies (SOF) on the 4a and 4c sites between Li6PS5Cl (from ref. [23]) and Li5.5PS4.5Cl1.5.

Figure 3. a) Lattice parameter of Li6@xPS5@xCl1+x vs. x showing that Vegard’s law is obeyed; b) compleximpedance plots at 298 K of the cold-pressed pellets of Li6PS5Cl, Li5.75PS4.75Cl1.25, Li5.5PS4.5Cl1.5, anda sintered pellet of Li5.5PS4.5Cl1.5 (550 88C for 10 min). Inset= magnified view at high frequencies, wherethe impedance is normalized to the respective pellet thickness (d) for better comparison; c) Arrheniusplot for Li5.5PS4.5Cl1.5 and corresponding Nyquist plot at 195 K (inset). The apex frequency is1.11 W 105 Hz, which corresponds to a capacitance of 1.7 W 10@9 F; the a-value is about 0.9, indicatingthe ideality of the CPE; d) conductivity and Ea for Li6@xPS5@xCl1+x (x =0, 0.25, 0.375 and 0.5).

AngewandteChemieCommunications

8683Angew. Chem. Int. Ed. 2019, 58, 8681 –8686 T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org

phases are compared in Figure 4b (see Figure S4 for x = 0.25and 0.375). Li5.5PS4.5Cl1.5 exhibits a diffusion coefficient of1.01 X 10@11 m2 s@1 at 300 K (see attenuation curve, Figure S5),which is remarkably high compared to that reported fromPFG NMR measurements of benchmark Li10GeP2S12, of 2.2 X10@12 m2 s@1,[25] or 3.5 X 10@12 m2 s@1 for ultrafast Li11Si2PS12.

[30]

Figure 5a shows the 7Li MAS NMR data forLi6@xPS5@xCl1+x ; x = 0, 0.25, 0.375, 0.50. As expected, onemain resonance is observed, with the small broad resonanceat @1.18 ppm corresponding to the LiCl impurity used forinternal shift referencing (see Supporting Information).Increasing the halogen substitution shifts the isotropicresonance to much lower frequency. The substitution of Cl@

withdraws electron density from the Li environments relativeto S2@, decreasing the paramagnetic component of the

chemical shift. This is in accord with the principlethat the lower ionic charge of the halogen willdecrease the electrostatic attraction of the mobile Liions to the rigid framework. Alternately, the shift canbe viewed as a gradual progression toward a moreionic, “LiCl-like” environment with increasing x,since as noted above, LiCl resonates at a much lowerfrequency than the parent-phase. However, thiseffect is not linear with the value of x ; in particularthe change from x = 0.375 to x = 0.5 is much lessened.Moreover, the symmetric, Gaussian character of theresonances for x = 0 and x = 0.25 gives way to a lineshape with much more pronounced chemical shiftasymmetry, which is a symptom of the greaterhalogen disorder. The fact that this shift trend isnot linear in x points toward a decreased interactionwith the anion framework, and therefore cannot beattributed to the changing statistical distribution ofthe anions alone. This region also shows the mostpronounced decrease in Ea with increasing halogensubstitution between x = 0.375 and x = 0.5 (Fig-ure 5b).

Here, the increase in Li vacancy content isdemonstrated through the Haven ratio, HR = D*/Ds, in Figure 5b where D* is the self-diffusivity ofthe mobile ions measured via PFG and Ds is definedby rescaling the ionic conductivity measured by EIS,sLi, into diffusivity via the Nernst–Einstein equa-tion[31] (see SI for details):[32]

Ds ¼ sLikBTcq2 ð1Þ

The Haven ratio is relatively insensitive totemperature (Figure S6),[33] but is known to beprimarily influenced by the concentration of mobilecharge carriers.[34] A temperature-independentHaven ratio implies that there is no change in thediffusion mechanism over the investigated temper-ature range.[35] For all phases, HR clusters around 0.3,similar to the Li-ion conductors Li6.5La3Zr1.5Ta0.5O12

and Li10SnP2S12—studies that also employed EISand PFG measurements.[28,36] Molecular dynamicssimulations also point to values of HR of thismagnitude for superionic conductors: 0.3–0.4 for

Li10GeP2S12,[37, 38] and 0.43 for LLZO.[37] While the Haven ratio

is relatively unchanged from x = 0 to x = 0.25, it decreasesslightly thereafter, reaching 0.23 at x = 0.5. A Haven ratioHR = 1 corresponds to purely random ion motion, and is onlyrealizable in extremely dilute systems, while lower values ofHR indicate strong correlation of the ion hops via cooperativemechanisms mediated by the presence of vacancies,[34] aneffect that has also been directly observed in other sys-tems.[39,40] In the more extreme case of the fluoride ionconductor LaF3, only one site is mobile and the transport isentirely vacancy-mediated, yielding HR& 0.1.[41]

As noted above, the onset of the decrease in HR (Fig-ure 5b) is coincident with the largest decrease in the Ea, andalso with the smallest change in the isotropic chemical shift.

Table 2: Summary of the ionic conductivity (si, for pellets cold-pressed at 2 tons) at298 K and Ea values obtained from EIS and PFG Arrhenius plots for the synthesizedLi-argyrodites. Error in the ionic conductivities is determined from the span in themeasurements for multiple samples of the same composition (extracted values ofthe impedance analyses are tabulated in Table S2 of the Supporting Information).

Li6@xPS5@xCl1+x si(tot) [mScm@1] Ea (EIS) [eV] :0.01 Ea (PFG) [eV]

Li6PS5Cl 2.5(1) 0.34 0.35(1)Li5.75PS4.75Cl1.25 4.2(2) 0.33 0.343(9)Li5.625PS4.625Cl1.375 5.6(2) 0.31 0.320(3)Li5.5PS4.5Cl1.5 9.4(1) 0.29 0.29(1)Li5.5PS4.5Cl1.5, sintered 12.0(2) – –Li5.45PS4.45Cl1.55 5.9(2) N/ALi5.4PS4.4Cl1.6 3.3(1) N/A

Figure 4. a) Li+-ion diffusivity in Li6@xPS5@xCl1+x obtained from 7Li PFG NMRmeasurements for x = 0, 0.25 and 0.5; b) Arrhenius plots of the diffusivity andconductivity values for x = 0 and x = 0.5 from PFG and EIS. The value for x =0 isin accord with that reported in the literature which ranges between 0.33–0.37 eV[4]

(where values are extracted from lnsT vs. T@1).

AngewandteChemieCommunications

8684 www.angewandte.org T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2019, 58, 8681 –8686

This points toward a synergy in the complementary effects ofperforming Cl@/S2@ substitution: namely for relatively lowsubstitution levels, the most pronounced effect on the ionictransport is likely from the reduced electrostatic interactionwith the framework, while at greater substitution levels themuch higher vacancy content (and hence more correlated ionmobility)[42] further drives the increased diffusivity. Statisti-cally, there will be a greater chance of an empty Li site in thecage near the Cl@ ions that will facilitate intercage jumps.Furthermore, the decrease in the intercage jump distancefrom 2.88 c (x = 0) to 2.81 c (x = 0.5) owing to the shrinkageof the unit cell will also favour higher jump rates. However,we expect this effect to be small.[23]

Stability of a solid-state electrolyte at its anodic andcathodic limits is an important consideration. Conflictingreports regarding the stability of Li6PS5Cl with lithium metalsuggest that while excellent stability is claimed[16]—presum-ably due to formation of a nanometre-thin passivatinginterphase[43]—slowly increasing interfacial resistance hasbeen observed on prolonged stripping/plating in symmetriccells.[17] We observe comparable, or slightly better stability ofLi5.5PS4.5Cl1.5 using even higher stripping/plating currents(0.25 mA cm@2) and capacity (1.0 mAh cm@2), Figure S8.Cyclic voltammetry (CV) performed in a SS jLi5.5PS4.5Cl1.5 jLi cell shows the current response of Li5.5PS4.5Cl1.5 issignificantly higher than the x = 0 phase owing to its higherionic conductivity (Figure S9a,b). Furthermore, while a verysmall anodic current on the first CV scan for Li5.5PS4.5Cl1.5 isobserved—likely corresponding to oxidation of lattice sulfideto insulating sulfur—this diminishes to virtually zero on thesecond cycle (Figure S10a). In contrast, the anodic current forLi6PS5Cl is initially higher (Figure S10b), and is still meas-

urable on the second sweep. Thus, Li5.5PS4.5Cl1.5

exhibits better anodic stability due to its lowersulfide content, but studies in cells with activecathode materials need to be conducted to fullyevaluate stability as CV studies can be misleading.

In summary, our studies of a new halide-richsolid solution series Li6@xPS5@xCl1+x that employneutron diffraction and variable temperature EIS,in conjunction with 7Li PFG and MAS NMR, revealthe changes to ionic transport that halide substitu-tion incurs. Increasing Cl and Li vacancy contenttriggers a significant and systematic lowering of theactivation barrier and increase in Li-ion diffusivity.This coincides with increase of the Cl@/S2@ disorder,and marks the influence of the monovalent anion.The limit of the solid solution range, Li5.5PS4.5Cl1.5,exhibits particularly high Li-ion diffusivity andquadrupled ionic conductivity of 9.4: 0.1 mScm@1

at room temperature (up to 12 mScm@1 for sinteredmaterials, approaching the best benchmarks). Itsrelatively good stability to lithium metal owing tothe lack of easily reduced metals such as Ge and Sn,and low-cost elements, suggest this material isa good prospect as a solid-state electrolyte. Ourfindings demonstrate that increasing the halidecontent of thiophosphate-based materials toweaken interactions between the mobile Li-ions

and surrounding framework, while increasing site disorderand lithium cation vacancy population to alter the energylandscape, is an important strategy to increase lithium ionmobility in the accelerated search for new solid-state ionconductors.

Acknowledgements

This research was supported by the BASF InternationalScientific Network for Electrochemistry and Batteries, andthrough NSERC Discovery grants to L.F.N. and G.G., anda Canada Research Chair to L.F.N. A portion of this researchused resources at the Spallation Neutron Source, a DOEOffice of Science User Facility operated by the Oak RidgeNational Laboratory. We thank Dr. Sergey Krachkovskiy forhelpful discussions regarding the PFG measurements.

Conflict of interest

The authors declare no conflict of interest.

Keywords: argyrodite · Li-ion conductor · neutron diffraction ·PFG NMR spectroscopy · solid electrolyte

How to cite: Angew. Chem. Int. Ed. 2019, 58, 8681–8686Angew. Chem. 2019, 131, 8773–8778

[1] J. Janek, W. G. Zeier, Nat. Energy 2016, 1, 16141.

Figure 5. a) 7Li MAS NMR for Li6@xPS5@xCl1+x (x = 0, 0.25, 0.375, 0.5) b) correlationof the activation energies from both techniques with the 7Li isotropic chemicalshift and the Haven ratio for all values of x under study.

AngewandteChemieCommunications

8685Angew. Chem. Int. Ed. 2019, 58, 8681 –8686 T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org

[2] J. C. Bachman, S. Muy, A. Grimaud, H.-H. Chang, N. Pour, S. F.Lux, O. Paschos, F. Maglia, S. Lupart, P. Lamp, L. Giordano, Y.Shao-Horn, Chem. Rev. 2016, 116, 140.

[3] J. Lau, R. H. DeBlock, D. M. Butts, D. S. Ashby, C. S. Choi, B. S.Dunn, Adv. Energy Mater. 2018, 8, 1800933.

[4] W. Zhang, Y. J. Shao, B. Lotsch, Y. S. Hu, H. Li, J. Janek, L. F.Nazar, C. W. Nan, J. Maier, M. Armand, L. Chen, EnergyEnviron. Sci. 2018, 11, 1945.

[5] R. Murugan, V. Thangadurai, W. Weppner, Angew. Chem. Int.Ed. 2007, 46, 7778; Angew. Chem. 2007, 119, 7925.

[6] D. Rettenwander, G. Redhammer, F. Preishuber-Pfugl, L.Cheng, L. Miara, R. Wagner, A. Welzl, E. Suard, M. M. Doeff,M. M. Wilkening, Chem. Mater. 2016, 28, 2384.

[7] H. J. Deiseroth, S. T. Kong, H. Eckert, J. Vannahme, C. Reiner,T. Zaiss, M. Schlosser, Angew. Chem. Int. Ed. 2008, 47, 755;Angew. Chem. 2008, 120, 767.

[8] R. P. Rao, S. Adams, Phys. Status Solidi A 2011, 208, 1804.[9] P. R. Rayavarapu, N. Sharma, V. K. Peterson, S. Adams, J. Solid

State Electrochem. 2012, 16, 1807.[10] M. A. Kraft, S. Ohno, T. Zinkevich, R. Koerver, S. P. Culver, T.

Fuchs, A. Senyshyn, S Indris, B. J. Morgan, W. G. Zeier, J. Am.Chem. Soc. 2018, 140, 16330.

[11] S. Yubuchi, S. Teragawa, K. Aso, K. Tadanaga, A. Hayashi, M.Tatsumisago, J. Power Sources 2015, 293, 941.

[12] N. C. Rosero-Navarro, A. Miura, K. Tadanaga, J. Sol-Gel Sci.Technol. 2019, 89, 303.

[13] S. Chida, A. Miura, N. C. Rosero-Navarro, M. Higuchi, N. H.Phuc, H. Muto, A. Matsuda, K. Tadanaga, Ceram. Int. 2018, 44,742.

[14] L. Zhou, K. H. Park, X. Sun, F. Lalere, T. Adermann, P.Hartmann, L. F. Nazar, ACS Energy Lett. 2019, 4, 265.

[15] S. Yubuchi, M. Uematsu, C. Hotehama, A. Sakuda, A. Hayashi,M. Tatsumisago, J. Mater. Chem. A 2019, 7, 558.

[16] S. Wang, Y. Zhang, X. Zhang, T. Liu, Y. Lin, Y. Shen, L. Li, C.Nan, ACS Appl. Mater. Interfaces 2018, 10, 42279.

[17] C. Yu, S. Ganapathy, J. Hageman, L. van Eijck, E. R. H.van Eck, L. Zhang, T. Schwietert, S. Basak, E. Kelder, M.Wagemaker, ACS Appl. Mater. Interfaces 2018, 10, 33296.

[18] C. Yu, S. Ganapathy, N. J. de Klerk, I. Roslon, E. R. van Eck,A. P. Kentgens, M. Wagemaker, J. Am. Chem. Soc. 2016, 138,11192.

[19] R. Prasada Rao, N. Sharma, V. K. Peterson, S. Adams, SolidState Ionics 2013, 230, 72.

[20] M. Chen, X. Yin, M. V. Reddy, S. Adams, J. Mater. Chem. A2015, 3, 10698.

[21] D. H. Kim, D. Y. Oh, K. H. Park, Y. E. Choi, Y. J. Nam, H. A.Lee, S. M. Lee, Y. S. Jung, Nano Lett. 2017, 17, 3013.

[22] J. Auvergniot, A. Cassel, J. Ledeuil, V. Viallet, V. Seznec, R.DedryvHre, Chem. Mater. 2017, 29, 3883.

[23] M. A. Kraft, S. P. Culver, M. Calderon, F. Bçcher, T. Krauskopf,A. Senyshyn, C. Dietrich, A. Zevalkink, J. Janek, W. G. Zeier, J.Am. Chem. Soc. 2017, 139, 10909.

[24] N. J. de Klerk, I. Rosłon, M. Wagemaker, Chem. Mater. 2016, 28,7955.

[25] A. Kuhn, V. Duppel, B. V. Lotsch, Energy Environ. Sci. 2013, 6,3548.

[26] M. Kaus, H. Stçffler, M. Yavuz, T. Zinkevich, M. Knapp, H.Ehrenberg, S. Indris, J. Phys. Chem. C 2017, 121, 23370.

[27] K. J. Griffith, K. M. Wiaderek, G. Cibin, L. E. Marbella, C. P.Grey, Nature 2018, 559, 556.

[28] A. Dorai, N. Kuwata, R. Takekawa, J. Kawamura, K. Kataoka, J.Akimoto, Solid State Ionics 2018, 327, 18.

[29] G. J. Brug, A. L. G. van den Eeden, M. Sluyters-Rehback, J. H.Luyters, J. Electroanal. Chem. Interfacial Electrochem. 1984, 176,275.

[30] A. Kuhn, O. Gerbig, C. Zhu, F. Falkenberg, J. Maier, B. Lotsch,Phys. Chem. Chem. Phys. 2014, 16, 14669.

[31] P. Heitjans, S. Indris, J. Phys. Condens. Matter 2003, 15, R1257.[32] H. Stoeffler, T. Zinkevich, M. Yavuz, A. Senyshyn, J. Kulisch, P.

Hartmann, T. Adermann, S. Randau, F. H. Richter, J. Janek, S.Indris, H. Ehrenberg, J. Phys. Chem. C 2018, 122, 15954.

[33] B. Roling, C. Martiny, S. Brgckner, Phys. Rev. B 2001, 63, 214203.[34] G. E. Murch, Solid State Ionics 1982, 7, 177.[35] _. Imre, H. Staesche, S. Voss, M. D. Ingram, K. Funke, H.

Mehrer, J. Phys. Chem. B 2007, 111, 5301.[36] P. Bron, S. Johansson, K. Zick, J. Schmedt auf der Ggnne, S.

Dehnen, B. Roling, J. Am. Chem. Soc. 2013, 135, 15694.[37] X. He, Y. Zhu, Y. Mo, Nat. Commun. 2017, 8, 15893.[38] A. Marcolongo, N. Marzari, Phys. Rev. Mater. 2017, 1, 025402.[39] E. Bychkov, Solid State Ionics 2009, 180, 510.[40] M. P. Thomas, N. L. Peterson, Solid State Ionics 1984, 14, 297.[41] V. V. Sinitsyn, O. Lips, A. F. Privalov, F. Fujara, I. V. Murin, J.

Phys. Chem. Solids 2003, 64, 1201.[42] Y. Wang, W. D. Richards, S. P. Ong, L. J. Miara, J. C. Kim, Y. Mo,

G. Ceder, Nat. Mater. 2015, 14, 1026.[43] S. Wenzel, S. J. Sedlmaier, C. Dietrich, W. G. Zeier, J. Janek,

Solid State Ionics 2018, 318, 102.

Manuscript received: December 14, 2018Revised manuscript received: February 26, 2019Accepted manuscript online: April 30, 2019Version of record online: May 23, 2019

AngewandteChemieCommunications

8686 www.angewandte.org T 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2019, 58, 8681 –8686