CERAMICSINTERNATIONAL

Available online at www.sciencedirect.com

http://dx.doi.org/0272-8842/& 20

nCorrespondinnnCorrespond

fax: þ86 21 524E-mail addre

huangfq@mail.s

(2014) 15497–15501

Ceramics International 40 www.elsevier.com/locate/ceramint

Enhanced ionic conductivity of sulfide-based solid electrolyteby incorporating lanthanum sulfide

Zhanqiang Liu, Yufeng Tang, Xujie Lü, Guohao Renn, Fuqiang Huangnn

Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, PR China

Received 16 May 2014; received in revised form 2 July 2014; accepted 2 July 2014Available online 10 July 2014

Abstract

Solid electrolyte is the key point for developing all-solid-state lithium-ion battery with good recyclability and no safety problem. Here, acrystalline phase of lanthanum sulfide (La2S3) was incorporated into glassy sulfide-based solid electrolytes (Li2S–SiS2 or Li2S–P2S5) to form twocomposite systems Li2S–SiS2–xLa2S3 and Li2S–P2S5–yLa2S3. The results demonstrated that La2S3 has the ability to suppress the crystallizationof Li2S–SiS2 or Li2S–P2S5 during the synthetic process. The lithium-ion conductivities of the designed composites can be improved by more thanone order of magnitude in comparison with the pristine samples. And both of these two composites have a wide electrochemical window of morethan 8 V vs. Liþ /Li.& 2014 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: A. Powders: solid state reaction; B. Composites; C. Ionic conductivity; E. Batteries; Lanthanum sulfide

1. Introduction

Traditional liquid electrolytes used in lithium-ion batterycontain toxic and flammable solvents, which bring seriousproblems such as leakage, inflammability and narrow operat-ing temperature range. For safety concerns and particularapplications, the next-generation lithium-ion batteries havebeen heading for replacing the traditional liquid electrolyteswith solid electrolytes [1]. However, the solid electrolytes havenot been widely used because their ionic conductivities are stillmuch lower than those of the liquid electrolytes. To meet therequirements of high performance solid-state batteries,advanced solid electrolytes having similar ionic conductivitywith the liquid electrolyte are highly required. Many inorganicsolid electrolytes have been investigated among which sulfide-based electrolytes show great promising prospects.

10.1016/j.ceramint.2014.07.01114 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

g author. Tel.: þ86 21 69987741; fax: þ86 21 5241 3903.ing author. Tel.: þ86 21 5241 1620;1 6360.sses: rgh@mail.sic.ac.cn (G. Ren),ic.ac.cn (F. Huang).

Sulfide-based electrolytes have much higher lithium-ion con-ductivity than oxide-based electrolytes owing to their higherpolarization ability of S2� than O2�. Li2S�P2S5 [2–5],Li2S�SiS2 [6–8] and Li2S�GeS2 [9–11] are currently themost studied sulfide-based electrolytes. In these systems,P2S5, SiS2 or GeS2 serve as network former and Li2S isnetwork modifier. It is well known that the lithium-ion conductiv-ities of these sulfide electrolytes can be improved by reducing thecrystallinity to form poorly-crystallized or amorphous phases.Unfortunately, these systems usually have low-melting pointsand are easy to form good crystalline phase. So it is very importantto develop strategies to stabilize the glassy phase of sulfide-basedelectrolytes. Although remarkable progress has been made in thepreparation of the glassy systems by doping oxides, sulfides,lithium halides or other chemical materials to improve ionicconductivities, an effective strategy still represents a big challenge[6,12–19]. In this paper, we propose a facile route to improve theionic conductivities of Li2S�SiS2 and Li2S�P2S5 systems byincorporating lanthanum sulfide (La2S3). The La2S3 in thecomposites can suppress the crystallization of Li2S�SiS2 orLi2S�P2S5, and the ionic conductivities of these compositescan be significantly enhanced.

Fig. 1. X-ray diffraction patterns for the samples 3Li4SiS4�xLa2S3.

Fig. 2. Complex impedance plots for 3Li4SiS4�xLa2S3 with x¼0.3.(a) Complex impedance plots. (b) Enlarged plots from 0 to 500 Ω.

Z. Liu et al. / Ceramics International 40 (2014) 15497–1550115498

2. Experimental details

La2S3, SiS2, P2S5 and Li2S were used as the startingmaterials. All the raw materials were stoichiometricallysynthesized in our lab, and the purities were confirmed byX-ray diffraction. For the incorporation of La2S3 intoLi2S�SiS2, the molar ratio of Li2S, SiS2 and La2S3 was controlledas 6: 3: x (x=0, 0.1, 0.3, 0.5, 0.75 and 1.0). For the additionof La2S3 into Li2S�P2S5, the molar ratio of Li2S, P2S5 andLa2S3 was controlled as 4: 1: y (y=0, 0.1, 0.2, 0.3 and 0.5).The fully mixed precursors were loaded into evacuated andsealed silica tubes. The tubes were placed into the furnace,heated at 750 1C for 10 h and then quenched in water. Thefinal products were collected and ground into fine powders.

X-ray diffraction (XRD) analysis was carried out with anX-ray diffractometer (Rigaku D/Max 2550V, 40 kV 40 mA)with CuKα radiation in the 2θ range from 201 to 601.

For ionic conductivity measurement, powder sample wascold-pressed into a ϕ10� 1 mm pellet and both sides of thepellet were attached with indium plates as current collectors.Then the measurement was conducted in a dry argon flow bycomplex impedance on an impedance analyzer (Chenhua660B) in the frequency range of 0.1 Hz and 0.1 MHz overthe temperature range from 30 to 210 1C.

DC polarization measurement was carried out to determinethe electronic conductivity. Indium foil was used as blockingelectrode and lithium plate as unblocking electrode. Timedependence of the electrical current was measured under aconstant voltage 1 V for 1600 s.

Cyclic voltammogram (CV) of the asymmetric Li/sample/Ptcell was performed on Chenhua 660B to evaluate the electro-chemical stability of the electrolytes at the scan rate of 10 mV s�1.The Lithium plate and Pt plate in the cell are serve as thereference/counter and working electrodes, respectively.

3. Results and discussion

XRD patterns of 6Li2S�3SiS2�xLa2S3 with various La2S3contents are shown in Fig. 1. It can be seen that Li4SiS4 wasobtained at x¼0 [8]. With the increasing of La2S3, the peaksindexed to La2S3 appear and become stronger gradually. Theunchanged phase of La2S3 before and after the syntheticprocess indicates that there is no reaction between La2S3 andLi2S or SiS2. Since the melting temperature of Li4SiS4 isaround 750 1C, the quenching process should result in the poorcrystallinity of Li4SiS4. And it can be observed that the peaksindexed to Li4SiS4 become relatively weaker with the increas-ing of x and almost disappeared at x¼1.0. The poor-crystalLi4SiS4 transforms into amorphous state step by step, implyingthat the added La2S3 can efficiently suppress the crystallizationof Li4SiS4 during the preparation process.

Usually, the coordination number of La3þ is six to nine,while Si4þ is 4-coordinated in Li4SiS4. Considering otherconditions, such as the reaction temperature, reaction time andespecially the big difference of atomic radius between La3þ

and Si4þ , the possibility of substitution of Si4þ by La3þ isvery low. So it can be concluded that the final samples can be

recognized as a kind of composite, poor-crystallized Li4SiS4/well-crystallized La2S3 (3Li4SiS4�xLa2S3).Ionic conductivities of the obtained samples were examined

through the AC impedance spectroscopy. Fig. 2 shows theimpedance plots of 3Li4SiS4�xLa2S3 with x¼0.3 at varioustemperatures. A semicircle in the high-frequency range with aspike in the low frequency range was observed at 30 1C.As typical ionic conductor, the semicircle is interpreted as a

Figofcon

Z. Liu et al. / Ceramics International 40 (2014) 15497–15501 15499

parallel combination of a resistance and a capacitance, whichare attributed to the bulk and grain boundary contribution,respectively. The spike is caused by the electrode contribution,where Warburg impedance exists. By increasing the measure-ment temperatures over 60 1C, no semicircle can be observedbut only spikes left. The total resistance at each temperaturecan be calculated from the intersection of the semicircle withthe real axis at the lower frequency side. Then the totallithium-ion conductivity at each temperature can be derived.

Fig. 3 shows the temperature dependences of conductivitiesfor the series of 3Li4SiS4�xLa2S3. The plots of log(σT) vs.103/T indicate that the conductivities of these composites in thetemperature range from 30 to 210 1C follow the Arrheniusequation: σT¼σ0exp(�Ea/RT), where σ0 is the pre-exponentialfactor, Ea the activation energy for conduction, R the gasconstant and T the absolute temperature. So it can be inferredthat the system of 3Li4SiS4�xLa2S3 is a good lithium-ionconductor. Then the corresponding activation energies werecalculated based on the Arrhenius equation.

To further demonstrate the effect of the amount of La2S3 onthe Li4SiS4 system. The composition dependences of the con-ductivity and the activation energy for are shown in Fig. 3(the inset figure). The ionic conductivity and activation energy ofthe obtained Li4SiS4 are calculated to be 3.3� 10�6 S cm�1 and0.43 eV, respectively. As reported, however, the ionic conductiv-ity and activation energy of the fine crystalline Li4SiS4 are5.0� 10�8 S cm�1 and 0.60 eV, respectively [20]. The reasonfor the big differences is the poor crystallinity of the obtainedLi4SiS4, which is caused by the quenching process. So the aboveresult demonstrates that poor crystallinity is benefit for the lithiumion diffusion in Li4SiS4. The XRD analysis has disclosed that thepoor crystallinity will turn into amorphous gradually with theincreasing of La2S3, suggesting the ionic conductivity might befurther enhanced. As shown in Fig. 3, the conductivities ofLi4SiS4�xLa2S3 increase initially and then decrease as the La2S3content increasing from x¼0 to 1.0. The corresponding activationenergy almost follows the reverse trend. Apparently, the contentof La2S3 was optimized to be x¼0.3. And the correspondingconductivity and activation energy were calculated to be

. 3. Temperature dependences of the electrical conductivities for the samples3Li4SiS4�xLa2S3. The inset depicts the composition dependence of theductivities and the activation energies for the samples 3Li4SiS4�xLa2S3.

Figaft

3.4� 10�5 S cm�1 and 0.38 eV, respectively. The optimizedconductivity is ten times higher than that of the pristine Li4SiS4after La2S3 added. Since La2S3 is ionic-insulating, heavy additionwill block the lithium ions and subsequently reduce the ionicconductivity of the composite.Fig. 4 shows the time dependence of the DC conductivity

obtained from current after applying a constant 1 V DCvoltage on sample 3Li4SiS4�0.3La2S3 at room temperature.When using lithium as nonblocking electrode, the DC con-ductivity is almost constant with time and the value, around2.8� 10�5 S cm�1, is in good agreement with the numberobtained from AC impedance shown above. When usingindium as blocking electrode, a large decrease of DC con-ductivity, resulted from the accumulation of lithium atoms onthe electrode, can be observed initially and then the conduc-tivity become almost constant. The constant conductivity of1.80� 10�8 S cm�1 is about three orders of magnitude lowerthan that obtained using the nonblocking electrode. Therefore,it can be concluded that the electronic conduction in the totalconductivity is almost negligible.Electrochemical stabilities of the obtained samples were also

studied by cyclic voltammetry with a potential range from�0.50 to þ10.0 V vs. Liþ /Li. Fig. 5 shows the cyclicvoltammogram of the sample 3Li4SiS4�3La2S3. A cathodiccurrent due to lithium deposition (Liþþe�-Li) is observedat around 0 V on the cathodic sweep up to �0.50 V, and thenan anodic current due to lithium dissolution (Li-Liþ þe�) isobserved at around 0.2 V on the an anodic sweep. There is nosignificant anodic current in the potential up to 10 V, whichmeans the sample of 3Li4SiS4�0.3La2S3 has a wide electro-chemical window up to 10 V vs. Liþ /Li.In order to verify our statement on the effects of La2S3

addition, La2S3 was also introduced into the Li2S�P2S5system. As shown in Fig. 6, the phases of the series of4Li2S�P2S5�yLa2S3 were examined by X-ray powder dif-fraction. At the point of y¼0, the main peaks are indexed toLi3þ0.55P1�0.11S4 [21]. Similar to Li4SiS4, the peaks ofLi3þ0.55P1�0.11S4 also become weaker and weaker graduallywith La2S3 adding, which implies the crystallinity ofLi3þ0.55P1�0.11S4 is becoming weaker. Interestingly, with

. 4. Time dependence of DC conductivity for the sample 3Li4SiS4�0.3La2S3er applying a constant DC voltage of 1 V at room temperature.

Fig. 6. X-ray diffraction patterns for the samples 2.25Li3þ0.55P1�0.11S4�yLa2S3.

Fig. 7. Temperature dependences of the electrical conductivities for thesamples 2.25Li3þ0.55P1�0.11S4�2yLaS2.

Fig. 8. Cyclic voltammogram of the sample 2.25Li3þ0.55P1�0.11S4�0.2LaS2at room temperature with scanning rate of 10 mV s�1.

Fig. 5. Cyclic voltammogram of the sample 3Li4SiS4�0.3La2S3 at roomtemperature with scanning rate of 10 mV s�1.

Z. Liu et al. / Ceramics International 40 (2014) 15497–1550115500

increasing y up to 0.5, new peaks of LaS2 come up graduallyrather than the originally added La2S3. The conversion ofLa2S3 to LaS2 should be resulted from the high vapor pressureof P2S5 [1,22]. Consequently, the final products should be acomposite in the form of poor-crystallized Li3þ0.55P1�0.11S4and well-crystallized LaS2 (2.25Li3þ0.55P1�0.11S4�2yLaS2).

The conductivities and the corresponding activity energiesof 2.25Li3þ0.55P1�0.11S4�2yLaS2 were also measured by ACimpedance spectroscopy. The temperature dependences ofconductivities are shown in Fig. 7. Just as the former studiedLi4SiS4�xLa2S3 system, the plots of log(σT) against 103/Tindicate that the conductivities of these composite samples inthe temperature range from 30 to 210 1C also follow theArrhenius equation. The reported data for Li3þ0.55P1�0.11S4 isalso displayed in Fig. 7 as reference. It can be seen that for thesample with y¼0 at 30 1C, both of the conductivity,1.82� 10�5 S cm�1 and the activity energy, 0.43 eV aresimilar with the reported data. With LaS2 existed, theconductivity of 2.25Li3þ0.55P1�0.11S4�2yLaS2 reached amaximum value of 7.41� 10�5 S cm�1 at y¼0.1. Then theconductivity decreased gradually with the increase of LaS2 dueto its insulating nature.

The DC conductivity was studied on 2.25Li3þ0.55

P1�0.11S4�0.2LaS2. As shown in Fig. 4, the DC conductivitieswith nonblocking or blocking electrodes were measured to be6.07� 10�5 or 3.40� 10�8 S cm�1, respectively. Therefore, theelectronic conduction in the total conductivity is almost negli-gible. And the obtained value of lithium-ion conductivity iscomparative to that measured by AC impedance spectroscopy. Asshown in Fig. 8, cyclic voltammogram test for 2.25Li3þ0.55

P1�0.11S4�0.2LaS2 shows that the cathodic current and anodiccurrent due to lithium deposition and dissolution can be observed,respectively. And it can be seen that 2.25Li3þ0.55P1�0.11S4�0.2LaS2 also has a wide electrochemical window of more than8 V vs. Liþ /Li.Finally, the lithium-ion conductivities of both Li2S�SiS2

and Li2S�P2S5 systems can be improved by the incorporationof lanthanum sulfide. The formation of composite structureshould be the main reason for the enhancement. To explore theenhancement mechanism of ionic conductivity in multiphasecomposite electrolyte systems, Agrawal [23] proposed severalmodels, all of which emphasize the existence of a space-chargeregion at the interface between the host and the dispersoid. Thehost is assigned as the ionic conducting solid and the

Z. Liu et al. / Ceramics International 40 (2014) 15497–15501 15501

dispersoid as the insulating materials. In our study, Li4SiS4 orLi3þ0.55P1�0.11S4 in poor crystallinity (partially in amorphousstate after lanthanum sulfides addition) are the host andlanthanum sulfide is the dispersoid, respectively. The amor-phous character is beneficial for host covering on dispersoidwith the largest interface area. Along the interface between thepoor crystalline Li4SiS4 or Li3þ0.55P1�0.11S4 and the goodcrystalline La2S3 or LaS2, lithium-ion transfers much freelyand speedily. Another reason is the large amount of defectsexisting in the poor crystalline Li4SiS4 or Li3þ0.55P1�0.11S4will decrease the activation energy of the lithium ion. Inaddition, the existing La ion in the interface will coordinatewith sulfur anion and then decrease the attraction betweensulfur and lithium.

4. Conclusions

In this paper, a composite structure of sulfide-based solidelectrolyte with poor-crystallized host and well-crystallizeddispersoid was designed to increase the ionic conductivity byadding lanthanum sulfide into the Li2S�SiS2 or Li2S�P2S5systems. The lanthanum sulfide in the as-prepared3Li4SiS4�xLa2S3 and 2.25Li3þ0.55P1�0.11S4�2yLaS2 com-posite electrolytes can suppress the crystallization of Li4SiS4 orLi3þ0.55P1�0.11S4. The electrochemical impedance measure-ments demonstrate that the ion conductivities of the designedsolid electrolytes were enhanced by one magnitude in compar-ison with the pristine lithium-ion electrolytes (from 3.3� 10�6

to 3.4� 10�5 S cm�1 for 3Li4SiS4�xLa2S3, and from1.82� 10� to 7.41� 10�5 S cm�1 for 2.25Li3þ0.55P1�0.11

S4�2yLaS2). Both of these composites have a very wideelectrochemical window of more than 8 V vs. Liþ /Li, which isimportant for practical applications. The mechanism for theenhanced ionic conductivity was demonstrated to be related tothe formation of multiphase composites, poor crystallinity ofthe lithium-ion conducting phase and the existing of La ion inthe interface.

Acknowledgments

This work was financially supported by National ScienceFoundation of China (Grant nos. 51125006, 21203234) andScience and Technology Commission of Shanghai Grant12XD1406800.

References

[1] K. Takada, Progress and prospective of solid-state lithium batteries, ActaMater. 61 (2013) 759–770.

[2] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami,Preparation of Li2S-P2S5 amorphous solid electrolytes by mechanicalmilling, J. Am. Ceram. Soc. 84 (2001) 477–479.

[3] F. Mizuno, A. Hayashi, K. Tadanaga, M. Tatsumisago, New, highly ion-conductive crystals precipitated from Li2S-P2S5 glasses, Adv. Mater. 17(2005) 918–921.

[4] J.E. Treveya, Y.S. Jungb, S.H. Lee, High lithium ion conducting Li2S–GeS2–P2S5 glass–ceramic solid electrolyte with sulfur additive for allsolid-state lithium secondary batteries, Electrochim. Acta 56 (2011)4243–4247.

[5] T. Ohtomo, A. Hayashi, M. Tatsumisago, Y. Tsuchida, S. Hama,K. Kawamoto, All-solid-state lithium secondary batteries using the75Li2S � 25P2S5 glass and the 70Li2S � 30P2S5 glass–ceramic as solidelectrolytes, J. Power Sour. 233 (2013) 231–235.

[6] S. Sahami, S.W. Shea, J.H. Kennedy, Preparation and conductivitymeasurements of SiS2–Li2S–LiBr li-ion conductive glasses, J. Electro-chem. Soc. 132 (1985) 985–986.

[7] H Takahara, M. Tabuchi, T. Takeuchi, H. Kageyama, J. Ide, K. Handa,Y. Kobayashi, Y. Kurisu, S. Kondo, R. Kanno, Application of lithiummetal electrodes to all-Solid-State lithium secondary batteries usingLi3PO4–Li2S–SiS2 glass, J. Electrochem. Soc. 151 (2004) A1309–A1313.

[8] M. Murayama, R. Kanno, M. Irie, S. Ito, T. Hata, N. Sonoyama,Y. Kawamoto, Synthesis of new li-ionic conductor thio-LISICON lithiumsilicon sulfides system, J. Solid State Chem. 168 (2002) 140–148.

[9] K. Itoh, M. Sonobe, K. Mori, M. Sugiyama, T. Fukunaga, Structuralobservation of Li2S–GeS2 superionic glasses, Physica B 385–386 (2006)520–522.

[10] J. Saienga, S.W. Martin, The comparative structure, properties, and ionicconductivity of LiIþLi2SþGeS2 glasses doped with Ga2S3and La2S, J.Non-Cryst. Solids 354 (2008) 1475–1486.

[11] Y. Ito, A. Sakuda, T. Ohtomo, A. Hayashi, M. Tatsumisago, Preparationof Li2S–GeS2 solid electrolyte thin films using pulsed laser deposition,Solid State Ionics 236 (2013) 1–4.

[12] Y. Ooura, N. Machida, M. Naito, T. Shigematsu, Electrochemicalproperties of the amorphous solid electrolytes in the system Li2S–Al2S3–P2S5, Solid State Ionics 225 (2012) 350–353.

[13] Z. Nagamedianova, A. Hernandez, E. Sanchez, Conductivity studies onLiX–Li2S–Sb2S3 P2S5, Ionics 12 (2006) 315–322.

[14] J.E. Trevey, J.R. Gilsdorf, S.W. Miller, S.H. Lee, Li2S–Li2O–P2S5 solidelectrolyte for all-solid-state lithium batteries, Solid State Ionics 214(2012) 25–30.

[15] A. Hayashi, T. Fukuda, H. Morimoto, M. Tatsumisago, Amorphous solidelectrolytes in the system Li2S–Al2S3–SiS2 prepared by mechanicalmilling, J. Mater. Sci. 39 (2004) 5125–5127.

[16] Z. Nagamedianova, E. Sanchez, Preparation and characterization of novelLi2S–Sb2S3 and Li2S–P2S5Sb2S3 glassy systems, J. Non-Cryst. Solids329 (2003) 13–16.

[17] Z. Nagamedianova, E. Sanchez, Synthesis and characterization of newglasses based on Li2S–P2S5–Sb2S3 system, J. Mater. Sci-Mate. El 18(2007) 547–552.

[18] M. Yamashita, H. Yamanaka, Formation and ionic conductivity of Li2S–GeS2–Ga2S3 glasses and thin films, Solid State Ionics 158 (2003)151–156.

[19] J. Saienga, Y. Kim, B. Campbell, S.W. Martin, Preparation andcharacterization of glasses in the LiIþLi2SþGeS2þGa2S3 system, SolidState Ionics 176 (2005) 1229–1236.

[20] B.T. Ahn, R.A. Huggins, Synthesis and lithium conductivities of Li2SiS3and Li4SiS4, Mater. Res. Bull. 24 (1989) 889–897.

[21] M. Murayama, N. Sonoyama, A. Yamada, R. Kanno, Material design ofnew Li-ionic conductor, thio-LISICON, in the Li2S–P2S5 system, SolidState Ionics 170 (2004) 173–180.

[22] I.G. Vasilyeva, R.E. Nikolaev, The La2S3–LaS2 system: Thermodynamicand kinetic study, J. Solid State Chem. 183 (2010) 1747–1751.

[23] R.C. Agrawal, R.K. Gupta, Superionic solid: composite electrolytephase–an overview, J. Mater. Sci. 34 (1999) 1131–1162.