Contents lists available at ScienceDirect

Journal of Alloys and

Compounds

j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /j a l l c o

m

Preparation and characterization of La 9.33Si 6O 26powders by molten salt method for solid electrolyte application

Buyin Li,Jia Liu,Yunxiang Hu ∗,Zhaoxiang Huang

Department of Electronic Science &Technology,Huazhong University of Science &Technology,Wuhan,430074,PR China

a r t i c l e i n f o Article history:

Received 18August 2010

Received in revised form 12October 2010Accepted 27October 2010

Available online 13 December 2010Keywords:

Lanthanum silicate Molten salt method Electrical conductivity Powder synthesis

a b s t r a c t

Lanthanum silicate La 9.33Si 6O 26(LSO)powders with more uniform particle and less agglomeration were obtained at a much lower synthesis temperature by the molten salt method than by the solid-state method.LSO ceramic electrolytes were prepared with these powders and characterized as well.The optimal molten salt synthesis conditions are mass ratio of reactants to NaCl of 1:3and synthesis temper-ature of 900◦C.XRD results showed that when the mass ratio of reactants to NaCl was no more than 1:3,pure LSO phase powder was obtained at 900◦C.XRD and XRF results showed that when synthesis tem-perature was higher than 900◦C,a solid solution type LSO powder with Na replacement for La formed at a fixed mass ratio of reactants to NaCl of 1:3.The involvement of Na in LSO lattice might lead to the lattice contraction in powders and deteriorate the conductance of ceramic electrolytes.The ceramic electrolytes prepared from the pure LSO powder via molten salt process exhibited better electrical properties than those from the powder via solid-state method.

© 2010 Elsevier B.V. All rights reserved.

1.Introduction

Solid oxide fuel cells (SOFCs)draw much attention in recent years,because of their high efficiency and environmentally friendly nature.A basic SOFC consists of three main ceramic components:an anode and a cathode separated by a solid electrolyte.Currently,yttria-stabilized zirconia (YSZ)with fluorite structure is the most common electrolyte material for SOFCs due to its excellent ionic conductivity and negligible electronic conduction at elevated tem-peratures (850–1000◦C),and good chemical stability.Nevertheless,this high operating temperature causes problems in terms of mate-rials selection and lifetime [1–4].So people search for alternative functional materials with higher ionic conductivity at relatively lower temperature [2,5].The apatite structure materials as a new class of oxide-ion conductors for SOFCs,first found by Nakayama et al.[6,7],have high ionic conductivity at intermediate temper-atures (700–800◦C),suitable thermal expansion matching with the electrode material,and other benefits [3–5,8].The conduction mechanism of apatite structure materials is based on interstitial conduction and completely different from the oxygen vacancy conduction mechanism in the traditional materials with fluorite or perovskite related structures [5,9–11].These apatite structure materials have a common chemical formula Ln 10−x (MO 4)6O 2±y (Ln is a metal such as rare earth or alkaline earth and M is a p-block

∗Corresponding author.Tel.:+862787558482;fax:+862787545167.E-mail address:hyx@mail.hust.edu.cn (Y.Hu).element such as P,As,Si,or Ge),and La 9.33SiO 6O 26(LSO)is a typical compound in this series [5,9,10,12–14].Ln site ions can be partly replaced by other ion dopants with lower valance,and the conduc-tivity of the resultant materials decreases with increasing dopant concentration if the ion radius of the dopants is smaller [15–20].

So far the preparation methods of LSO powders mainly include solid-state,sol–gel,and hydrothermal methods [9,12,13,21–24],while little study refers to the molten salt synthesis of them.In the molten salt process,the salt plays a role as flux.The reactants in the molten salt have a high reactivity and mobility,and react easily and completely,which can greatly reduce the synthesis tem-perature.Moreover the product powders have a better uniformity and fewer agglomerates [25].These advantages are expected to be also embodied if molten salt method is applied to the synthesis of LSO powders.

The aim of the present paper is to explore the feasibility of the molten salt synthesis of LSO powders.The originality of this work is to decrease the synthesis temperature of LSO powders via the molten salt method and to obtain ultra-fine powders with nar-row size distribution and dense ceramics with these powders.The effects of the mass ratio of reactants to the salt and the synthesis temperature on crystalline structure of the synthesized powders are investigated in detail.

2.Experimental procedures

Starting reactant powders of silicon dioxide (SiO 2,98%)and lanthanum oxide (La 2O 3,99%)were weighed in the nominal composition of La 9.33Si 6O 26.Sodium chloride (NaCl,99%)was taken as the molten salt and weighed.The mass ratios of

0925-8388/$–see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.jallcom.2010.10.215

3173

Fig.1.XRD patterns of powders calcined at900◦C for4h with different mass ratios of reactants to salt.

reactants to NaCl werefixed to be1:1,1:3,and1:5respectively.Then the mixtures of reactants and the salt were ball milled in ethanol medium for4h,dried,ground and sifted to avoid agglomerated powders.The sifted powders were calcined at 900–1100◦C for4h.Next,the as-synthesized powders were washed repeatedly with de-ionized water andfiltered to remove free NaCl until there was no appearance of white precipitate in thefiltrate by adding AgNO3solution.Finally LSO powders were obtained through drying the products at105◦C for6h.In addition,for comparison a LSO powder was prepared by the solid-state reaction for4h at1300◦C which is the lowest required temperature to synthesize the powder via traditional solid-state method.

The LSO powders were pelleted and pressed at200MPa for25s into disks (15mm in diameter,2mm in thickness).The green disks were then heated to1500◦C with a heating rate of300◦C/h in air,soaked for4h,and then furnace-cooled.

X-ray diffraction(XRD)analyses at room temperature were carried out on a diffractometer(X’Pert PRO,PANalytical B.V.)using a Cu K␣radiation source (K␣1=1.54060˚A and K␣2=1.54443˚A).This method has been performed on each powder sample in order to reveal the crystallographic structures and cell param-eters.

X-rayfluorescence(XRF)probe analyses(EAGLE III)were adopted to analyze qualitatively and semi-quantitatively the elemental composition of ceramic sam-ples.

Scanning electron micrograph(SEM)analyses(Quanta200,Philips)was used to observe the micromorphology of the synthesized powders and ceramics.

A.C.conductivity measurement of the ceramics was performed with an accurate impedance analyzer(Agilent4294A)at frequencies of40Hz–20MHz at tempera-tures of300–800◦C.

3.Results and discussion

3.1.Synthesis and characterization of LSO powders

The effect of mass ratio of reactants to salt on the phase com-position of synthesized powders was investigated.Fig.1shows the XRD patterns of powders calcined at900◦C for4h with dif-ferent mass ratios.It is apparent that no NaCl,La2O3,and SiO2 peaks were detected in all synthesized powders,indicating that the presence of NaCl in the reactant mixture is helpful for the reac-tion between the reactants at relatively lower temperature(900◦C) and free NaCl can be completely removed from the product pow-ders by washing repeatedly.Further,at lower salt concentration (reactants:salt=1:1)apatite structure LSO phase(JCPDS49-0443) formed with the companion of other impurity phases La2SiO5 (JCPDS40-0234)and La4.67(SiO4)3O(JCPDS75-1145).However, only did single LSO apatite phase exist at higher salt concentration (1:3and1:5).The XRD results indicate that there is a threshold salt concentration for the preparation of pure LSO powders which lies between1:1and1:3.It seems that excess salt concentration(1:5) has no further help to the synthesis of pure LSO powders.Therefore the mass ratio of1:3was taken for the succeeding

experiments.Fig.2.(a)XRD patterns and(b)(211)peak magnification patterns of powders obtained at different synthesis temperatures by different methods.

In order to explore the effect of the synthesis temperature on the microstructure of LSO powders,powders were synthesized using molten salt method at900,1000,and1100◦C for4h respectively with afixed reactants to salt ratio of1:3.For comparison,a powder was synthesized by the solid-state method at1300◦C for4h.For simplification these four powders were named MS9,MS10,MS11, and SS13respectively.

XRD patterns of these four powders are displayed in Fig.2(a). They all had the same pure apatite structure characteristic peaks (JCPDS49-0443),and no significant impure peaks appeared,indi-cating that pure apatite structure phase could be formed when reactant mixture was calcined at temperatures of900–1100◦C for molten salt method while at1300◦C for solid-state method.More-over,it is apparent that MS9had larger peak intensity than SS13 though MS9was synthesized at a much lower temperature than SS13.This observation shows that the molten salt method allows the synthesis of LSO powders with better crystal growth at a rela-tively low temperature.

Furthermore,it is observed that the strongest(211)peak moved slightly to the right side in turn from MS9to MS11 (Fig.2(b)).According to Bragg’s diffraction law,the(211)inter-planar spacing in the synthesized powders decreased from MS9 to MS11.In order to prove this inference,the Philips X’pert Highscore software was used to calculate the cell parameters, and the results are shown in Table1.For comparison parame-ters of SS13,and the standard LSO and NaLa9(SiO4)6O2(JCPDS 32-1109)are also listed.It can be seen that cell parameters a, b,and c and cell volume V all decreased from MS9to MS11. For SS13and MS9these parameters approached to those of3174 B.Li et al./Journal of Alloys and Compounds509 (2011) 3172–3176

Table1

Cell parameter comparison between the synthesized powders and standard LSO and NaLa9(SiO4)6O2.

Standard LSO SS13MS9MS10MS11Standard NaLa9(SiO4)6O2

Cell parameters a,b(˚A)9.712.7108±0.00329.7134±0.004.7022±0.002.6943±0.00129.6920

Cell parameter c(˚A)7.18587.1906±0.00417.1842±0.00227.1824±0.00157.1823±0.00097.1820

Cell volume,V(˚A3)587.08587.22587.02585.51584.56584.26

Table2

Conductance data,elemental composition,and average grain size for different ceramic samples.

SS13MS9MS10MS11

@800◦C(S/cm)8.78×10−4 1.09×10−3 2.25×10−4 1.58×10−4

E a(eV)0.810.77 1.03 1.05

Na:La:Si(atom%)0:57.83:42.170:56.23:43.77 6.52:47.68:45.813.83:49.01:37.16 Grain size(␮m) 2.3 5.0 2.8 3.2

the standard LSO,while for MS11to those of the standard NaLa9(SiO4)6O2.

The changes in(211)peak position and cell parameters from MS9to MS11might result from the involvement of Na ions in the reaction.Na+probably occupied in part the sites and the vacancy of La3+and entered the LSO lattice at high temperatures,which led to the formation of a solid solution of Na x La9.33−x/3(SiO4)6O2. Since Na+has a slightly smaller radius(0.97˚A)than La3+(1.06˚A), the replacement of Na+for La3+led to decreased crystal parame-ters(i.e.lattice contraction).The presence of elemental Na in the synthesized powders via molten method was confirmed by the composition analysis on ceramics made from powders MS10and MS11through XRF(See Table2).Notice that XRD results show absence of free NaCl in these powders.It can be concluded that Na existed in the form of solid solution in these powders.Further-more no elemental Na was detected in SS13and MS9powders. Notice that the synthesis temperature increased from900◦C for MS9to1100◦C for MS11.Therefore there was a synthesis tempera-ture threshold for Na+replacing La3+in the molten salt process,and when synthesis temperature exceeded this threshold the replace-ment amount increased with the increasing temperature.

Fig.3shows the SEM morphology of SS13and MS9powders. Both powders consisted of spherical shape particles.The mean par-ticle size for MS9was250±20nm while600±100nm for SS13.In addition MS9had less agglomeration than SS13.The contrastive pictures directly confirm that molten salt method allows the syn-thesis of homogeneous LSO powders with smaller crystallites at a much lower temperature compared with solid-state method.These particles can be used to prepare dense ceramics.

3.2.Preparation and characterization of LSO ceramics

LSO ceramics were sintered at1500◦C in air for4h using SS13, MS9,MS10,and MS11powders respectively.The bulk densities

of Fig.4.SEM images of SS13,MS9,MS10,and MS11ceramics sintered at1500◦C in air for4h.

SS13,MS9,MS10,and MS11ceramic samples were4.91,5.04,4.93, and4.g cm−3,respectively.Fig.4shows the SEM images of the four obtained ceramic samples.It can be seen that all these sam-ples were very dense and could satisfy the application in SOFCs [12].Furthermore,the three ceramic samples from molten salt method powders had a bigger average grain size than the one from solid-state method powder under the same sintering condi-tion.The average grain sizes of these four ceramics were2.3␮m (SS13),5.0␮m(MS9),2.8␮m(MS10),and3.2␮m(MS11).Finally, the grains of the SS13ceramic only developed3.5times than

its Fig.3.SEM images of LSO powders prepared by solid-state method(SS13)and molten salt method(MS9).B.Li et al./Journal of Alloys and Compounds509 (2011) 3172–3176

3175

Fig.5.Arrhenius plots of the overall electrical conductivity of LSO ceramics sintered from different powders.

staring powder(Fig.3),while the growth multiple was about20 times for the MS9ceramic.This result indicated that the powders prepared by molten salt method had a higher growth activity than the powder by solid-state method.

The elemental composition results were listed in Table2for the four ceramic samples.The Na atomic ratio varied from zero in MS9 and SS13,6.59%in MS10,to13.83%in MS11.It is worth to stress that the Na atomic ratio increased from MS9,MS10,to MS11ceram-ics,i.e.it increased with the increasing synthesis temperature of precursor powders by the molten salt method.This observation supports the conclusion that the replacement of Na+to La3+took place in MS10and MS11powders while did not do in MS9powder due to lower synthesis temperature.

In order to investigate the conductivity of these LSO ceramic electrolytes,silver paste as electrodes was applied to opposite faces of sintered pellets and then heated at850◦C for10min for metallization.A.C.impedance measurement was performed on pellets at frequencies of40Hz–20MHz at300–800◦C.The obtained impedance spectrum showed substantial overlapping of the impedance arcs,and it is difficult to separate the grain bulk and grain boundary contribution from the impedance curve.Therefore, the total impedance is taken to calculate the conductivities of sam-ples herein.These result data are parameterized by the Arrhenius equation:

= 0

T exp

−

E a

(1)

where , 0,E a,k,and T are,respectively,the conductivity,pre-exponential factor,activation energy,Boltzmann constant,and absolute temperature.Arrhenius plots of the four LSO ceramic sam-ples are shown in Fig.5.

First of all,for each ceramic sample,the conductivity increased when the temperature increased.Secondly,it can be seen that among the four samples the MS9ceramics showed the most excel-lent conductance in the range of300–800◦C,including the highest conductivity at everyfixed temperature point and the lowest acti-vation energy E a(Table2).The SS13ceramics took the second best place while MS10and MS11ceramics exhibited the worst conduc-tance.Finally,the conductivity discrepancy among samples became smaller with the increasing temperature.These phenomena will be explained in details as follows.

From MS9,MS10,to MS11ceramic,the conductance became worse and worse,although their precursor powders all were pre-pared by the same way—molten salt method.This might originate from the differences in the average grain size and the amount of Na element in the ceramic samples.The density factor which was often considered elsewhere may be neglected here because all ceramic samples in this work were dense.As we know,the sum of grain bulk and grain boundary contributions was considered to determine the conductivity values for apatite materials[12].MS9had larger grains and therefore less grain boundaries in unit volume than MS10and MS11(Fig.4),which led to MS9having a better conductance than other two.

What is more important is that from MS9,MS10,to MS11 ceramic,the amount of Na element increased,and the involvement of Na deteriorated the conductance of electrolytes.It is well known that oxygen ions in the apatite silicates are confirmed to migrate via an interstitial conduction mechanism:excess oxide ions can be introduced in the structure and facilitate oxide-ion migration in the conduction channel by a complex sinusoidal pathway along the c-axis.Moreover,the pathway is strongly dependent on the ability of the silicate substructure to relax towards the La sites that contain the cation vacancies in the apatite structure[5,9–11]. The involvement of Na in LSO lattice could change the micro-conduction mechanism of samples by two ways.Firstly,the lattice contraction(see Table1)and conduction channel narrowing caused by the incorporation of Na+ions might limit the oxide ions diffusion. Secondly,to maintain total charge balance in the composition,the replacement of lower valence ion Na+for La3+certainly resulted in the reduction in the number of cation vacancies,which had a directly harmful influence on oxide ions migration to the interstitial site[15].

The reason for MS9ceramic exhibited a better conductance than SS13ceramic may be explained as follows.Notice that there was no Na element in both MS9and SS13ceramics.Furthermore,MS9 had a larger average grain size and therefore less grain boundaries than SS13ceramic,which led to MS9having a higher conductivity. However,although MS10and MS11ceramics had a slightly big-ger grain size than SS13ceramic,a much worse conductivity was observed for the former.This might be due to the adverse effects of Na in MS10and MS11on their conductance.

4.Conclusions

In summary,LSO powders with pure apatite phase were suc-cessfully synthesized at relatively low temperature via molten salt method.The optimal synthesis conditions are mass ratio of reac-tants to NaCl of1:3and synthesis temperature of900◦C.When the mass ratio of reactants to NaCl was greater than1:3,impurity phases existed in the powder synthesized at900◦C.When synthe-sis temperature was higher than900◦C,a solid solution type LSO powder with Na replacement for La formed at afixed mass ratio of reactants to NaCl of1:3.The involvement of Na in LSO lattice may be responsible for the lattice contraction in MS10and MS11 powders and for the conductance deterioration in MS10and MS11 ceramic electrolytes.Ceramic electrolytes prepared from the pure LSO molten salt powder(MS9)exhibited better electrical properties than those from the solid-state powder(SS13).

References

[1]G.J.Offer,J.Mermelstein,E.Brightman,N.P.Brandon,J.Am.Ceram.Soc.92

(2009)763–780.

[2]A.Orera,P.R.Slater,Chem.Mater.22(2010)675–690.

[3]K.Huang,J.Wan,J.B.Goodenough,J.Mater.Sci.36(2001)1093–1098.

[4]Y.Higuchi,M.Sugawara,K.Onishi,M.Sakamoto,S.Nakayama,Ceram.Int.36

(2010)955–959.

[5]T.Kharlamova,S.Pavlova,V.Sadykov,T.Krieger,L.Batuev,V.Muzykantov,N.

Uvarov,C.Argirusis,Solid State Ionics180(2009)796–799.

[6]S.Nakayama,H.Aono,Y.Sadaoka,Chem.Lett.6(1995)431–432.

[7]S.Nakayama,M.Sakamoto,J.Eur.Ceram.Soc.18(1998)1413–1418.

[8]D.Marrero-López,M.C.Martín-Sedeno,J.Pena-Martínez,J.C.Ruiz-Morales,

P.Núnez,M.A.G.Aranda,J.R.Ramos-Barrado,J.Power Sources195(2010) 2496–2506.3176 B.Li et al./Journal of Alloys and Compounds509 (2011) 3172–3176

[9]L.León-Reina,J.M.Porras-Vázquez,E.R.Losilla,M.A.G.Aranda,Solid State Ionics

177(2006)1307–1315.

[10]T.Iwata,E.Bechade,K.Fukuda,O.Masson,I.Julien,E.Champion,P.Thomas,J.

Am.Ceram.Soc.91(2008)3714–3720.

[11]L.Leon-Reina, E.R.Losilla,M.Martınez-Lara,S.Bruque, A.Llobet,

D.V.Sheptyakovc,M.A.G.Aranda,J.Mater.Chem.15(2005)24–

2498.

[12]P.J.Panteix,I.Julien,D.Bernache-Assollant,P.Abelard,Mater.Chem.Phys.95

(2006)313–320.

[13]T.Nakajima,K.Nishio,T.Ishigaki,T.Tsuchiya,J.Sol–Gel Sci.Technol.33(2005)

107–111.

[14]P.J.Panteix,I.Julien,P.Abelard,D.Bernache-Assollant,Ceram.Int.34(2008)

1579–1586.

[15]P.J.Panteix,E.Bechade,I.Julien,P.Abelard,D.Bernache-Assollant,Mater.Res.

Bull.43(2008)1223–1231.[16]P.J.Panteix,I.Julien,P.Abelard,D.Bernache-Assollant,J.Eur.Ceram.Soc.28

(2008)821–828.

[17]E.Kendrick,K.S.Knight,P.R.Slater,Mater.Res.Bull.44(2009)1806–1809.

[18]E.Kendrick,P.R.Slater,Solid State Ionics179(2008)981–984.

[19]E.Kendrick,A.Orerac,P.R.Slater,J.Mater.Chem.19(2009)7955–7958.

[20]J.R.Tolchard,P.R.Slater,M.S.Islam,Adv.Funct.Mater.17(2007)25–2571.

[21]S.Kato,T.Yoshizawa,N.Kakuta,S.Akiyama,M.Ogasawara,T.Wakabayashi,Y.

Nakahara,S.Nakata,Res.Chem.Intermed.34(2008)703–708.

[22]A.Chesnaud,G.Dezanneau,C.Estournès,C.Bogicevic,F.Karolak,S.Geiger,G.

Geneste,Solid State Ionics179(2008)1929–1939.

[23]S.Zec,J.Dukic,M.Puˇs evac,S.Boˇs kovic,R.Petrovic,Mater.Manuf.Process.24

(2009)1104–1108.

[24]S.H.Jo,P.Muralidharan,D.K.Kim,Electrochim.Acta54(2009)7495–7501.

[25]P.Afanasiev,C.Geantet,Coord,Chem.Rev.178(1998)1725–1752.下载本文