Effect of sintering aids on the densification and electrical properties of SiO2 –containing Ce0.8Sm0.2O1.9 ceramic

Effect of sintering aids on the densification and electrical properties of SiO2 –containing Ce0.8Sm0.2O1.9 ceramic

Accepted Manuscript Title: Effect of sintering aids on the densification and electrical properties of SiO2 –containing Ce0.8 Sm0.2 O1.9 ceramic Author...

2MB Sizes 1 Downloads 32 Views

Accepted Manuscript Title: Effect of sintering aids on the densification and electrical properties of SiO2 –containing Ce0.8 Sm0.2 O1.9 ceramic Authors: Lin Ge, Yiheng Gu, Yanli Zhang, Xueyan Li, Qing Ni, Lucun Guo PII: DOI: Reference:

S0955-2219(18)30051-7 https://doi.org/10.1016/j.jeurceramsoc.2018.01.036 JECS 11704

To appear in:

Journal of the European Ceramic Society

Received date: Revised date: Accepted date:

8-9-2017 17-1-2018 22-1-2018

Please cite this article as: Ge L, Gu Y, Zhang Y, Li X, Ni Q, Guo L, Effect of sintering aids on the densification and electrical properties of SiO2 –containing Ce0.8 Sm0.2 O1.9 ceramic, Journal of The European Ceramic Society (2010), https://doi.org/10.1016/j.jeurceramsoc.2018.01.036 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of sintering aids on the densification and electrical properties of SiO2 – containing Ce0.8Sm0.2O1.9 ceramic Lin Ge a,*, Yiheng Gu a, Yanli Zhang a, Xueyan Li a, Qing Ni a, Lucun Guo a a

College of Materials Science and Engineering, Nanjing Tech University, No. 5 Xinmofan Road, Nanjing, Jiangsu,

IP T

210009, PR China.

E-mail address: [email protected] (L. Ge)

SC R

*Corresponding author

Abstract:

U

In this work, five different metal-oxide additives (metal=Ba, Co, Fe, Li, and Mn) were

N

examined as sintering aids and SiO2 impurity scavengers for Ce0.8Sm0.2O1.9 (SDC). 2

A

mol% additives were loaded into the SDC with ~150 ppm (moderately impure) and

M

~2000 ppm (highly impure) SiO2. Ba-, Co-, Fe- and Mn-oxides showed comparative

ED

sintering-aid effect on both moderately- and highly-impure SDC specimens, but the sintering-assisting effect of Li-oxide was completely neutralized in highly impure SDC.

PT

Regarding electrical property, the deleterious effect of 2000 ppm SiO2 impurity on the

CC E

grain-boundary conduction of SDC can be effectively alleviated by adding Ba-, Co-, Fe-, or Mn-oxides. Microstructure analysis revealed that Ba-oxide reacted directly with SiO2 and consequently enhanced grain-boundary conduction. By contrast, with the

A

addition of Co-, Fe-, and Mn-oxides, the improved grain-boundary conductions of impure SDC were related to the scavenging reactions between Si, Ca (another original impurity) and Sm components.

1

Keywords: Samaria-doped ceria; Solid electrolyte; Sintering aid; Silica scavenging;

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Grain boundary conduction

2

1. Introduction Solid oxide fuel cells (SOFCs) have tremendous potential due to their various potential advantages [1-4], including high electricity conversion efficiency, superior environmental performance, and fuel flexibility. Thus, SOFCs provide critical energy

IP T

solution. At present, yttria-stabilized zirconia (YSZ) is the most commonly used electrolyte because of its high reliability. However, the necessity to operate YSZ-based

SC R

SOFCs at high temperatures (>800 ºC) results in extreme requirements for other SOFC

components. As a consequence, significant efforts have been exerted in developing

U

alternative electrolytes with low operating temperatures [5-8].

N

Doped ceria has been regarded as one of the most promising electrolytes for low-

A

temperature (LT) and intermediate-temperature (IT) SOFCs because of its excellent

M

oxide ionic conductivity within 400 °C‒700 °C [1,2,5,6]. In the low and intermediate

ED

temperature regions, the grain boundary conduction usually dominates the overall conduction of doped-ceria electrolyte in which the specific grain-boundary resistivity

PT

is ~102–105 times higher than that of grain interior [9-11]. The blocking effect of

CC E

siliceous intergranular phases has been reported in CeO2-based electrolytes, where even a few ppm of SiO2 impurity can significantly decrease the grain-boundary conductivity [11–13]. Considering that SiO2 is the most common background impurity in ceramic

A

processing, avoiding its presence even when using relatively pure starting materials is impossible [11,14,15]. Meanwhile, one of the main drawbacks of doped ceria is that densifying below 1550 ºC is difficult [16,17 ]. As a result, considerable research have focused on either 3

synthesizing highly reactive powder or using some additives to accelerate the sintering process. For the sintering aids, several additives, including Co2O3 [18], TiO2 [19], MnO [20], and ZnO [21], have been reported to effectively reducing the sintering temperature of doped ceria. Although the addition of these sintering aids gives rise to an important

IP T

effect on both the grain interior and the grain boundary properties, the exact role of these aids is still debatable [22,23]. The grain interior conductivity may be closely

SC R

related to the diffusion of the additives in the CeO2 lattice, whereas the grain boundary conduction would strongly depend on the impurity (mainly SiO2) level. During the

U

sintering, these additives accumulate/disperse at grain boundaries and react with the

N

dopants and the impurities to form the intergranular phases. The composition, location,

A

viscosity, and wetting properties of the intergranular phases change with the amounts

M

and types of the additives, sintering condition, and cooling rates. Thus, the additives

ED

play either positive or negative role in the grain-boundary conduction. The discrepancies in literature data for doped-ceria electrolytes with same addition may be

CC E

researchers.

PT

due to different impurity levels and sintering conditions adopted by different

In the present study, the two SiO2-containing Ce0.8Sm0.2O1.9 (SDC) systems with

~150 and ~2000 ppm SiO2 were prepared, which represent moderately impure and

A

highly impure, respectively. The densification and conduction behavior of SDC doped at the 2 mol% level with five different sintering aids (oxides of Ba, Co, Fe, Li, and Mn) was evaluated. How these sintering aids combined with SiO2 impurity affecting the grain /grain boundary conductions was investigated in detail. 4

2. Experimental 2.1 Sample preparation SDC powders were first prepared from commercial CeO2 (99.99%, Yixing Xinwei

IP T

Leeshing Rare Earth Co., China) and Sm2O3 (99.99%, Beijing Founde Star Science and Technology Co., China) powders. Stoichiometric quantities of CeO2 and Sm2O3 were

SC R

mixed with distilled water for 8 h using zirconia balls as milling media in a polyethylene

jar. After being dried and ground, the powders were calcined in air at 1200 °C for 2 h.

U

The obtained SDC powders were characterized using X-ray fluorescence for impurity

N

content (see Table S1 in supplementary materials). The content of SiO2 was around 150

A

ppm and that of CaO was 160 ppm in the SDC starting materials (moderately impure

M

SDC). ~1850 ppm SiO2 was loaded into the SDC powders using diluted SiO2 sol

ED

(Zhejiang Yuda Chemical Co., China) with distilled water. A mixture of the SDC powders and diluted SiO2 sol were thoroughly mixed to form slurries. Then, the slurries

PT

were dried and annealed at 800 °C for 2 h to obtain highly impure SDC powders

CC E

(contain 2000 ppm SiO2). The same method was used to introduce 2 mol% sintering aids into the two groups of SDC powders using nitrate salt as precursor. The two groups of SDC powders with and without 2 mol% sintering aid loading

A

were pressed into cylindrical pellets of 14 mm in diameter. Subsequently, these pellets were sintered in air at 1200 °C‒1450 °C for 10 h. To simplify, the moderately impure (150 ppm SiO2) group specimens are denoted as “SDC2M,” where “M” is the metal component, and the highly impure (2000 ppm SiO2) group specimens are referred to as 5

“SDC2M+Si.” For example, SDC2Co indicates 2 mol% Co-doped SDC specimen (with ~150 ppm SiO2 concentration), and a specimen labeled “SDC2Co+Si” was 2 mol% Codoped SDC containing 2000 ppm SiO2.

IP T

2.2 Characterization The crystal structures of the specimens were identified by X-ray diffraction (XRD,

SC R

ARL-X’TRA, Thermo, USA) with Cu Kα radiation at room temperature. The densities of the sintered samples were measured using the Archimedes method and expressed as

U

relative densities. Theoretical densities were estimated from the unit cell parameters

N

(determined from XRD patterns). The diffractometer was operated at 40 kV and 35 mA

A

in the 2θ range of 20°–80°. The surface morphology of the specimens was investigated

M

with a SEM (Model JSM-6360, JEOL, Tokyo, Japan) equipped with an Oxford energy

ED

dispersive spectroscopy (EDS) system. For scanning electron microscopy (SEM) observation, the surfaces of the samples were sputter-coated with gold. During SEM

PT

observations, EDS microanalyses were carried out locally at a high voltage of 10 kV.

CC E

Average grain sizes were measured from the micrographs and calculated using JEOL SMileView software. AC impedance spectroscopy was performed to determine the oxide ionic

A

conductivity of the sintered pellets. Both sides of the pellets were coated with Ag paste to form electrodes and fired at 700 °C for 10 min before measurement to ensure good bonding. The AC impedance spectra of the samples were obtained at 300 °C–750 °C in air using an impedance analyzer (PARSTAT 2273) at frequencies from 0.1 Hz to 100 6

KHz. Curve fitting and resistance calculations were performed using ZSimpWin software.

3. Results and discussion

IP T

3.1. Synthesis characterization Fig.1 shows the room-temperature XRD patterns of SDC2M+Si (M= Ba, Co, Fe,

SC R

Li, Mn) specimens. All of the diffraction peaks indicated a pure single fluorite phase

(CeO2, Fm3m, ICDD# 34-0394), and other phases were not detected within the

U

limitation of the XRD instrument. The cell parameters were refined by a least-squares

N

procedure using the MDI JADE 5.0 program. The values obtained are given in Table 1.

A

The results indicate that the addition of SiO2 or sintering aids causes only a minor

M

perturbation in the lattice parameters. This finding appears reasonable because the

ED

concentration of the additives is rather low, and these additives are prone to react with

PT

each other at grain boundary regions during sintering [11].

CC E

3.2. Sinterability

Fig. 2 and Fig. 3 show the linear shrinkage and the calculated relative densities of

specimens as a function of sintering temperature, respectively. The results of SDC are

A

also included in each picture to facilitate comparisons. Given the sintering-aid effect of BaO, Co2O3, Fe2O3, Li2O, and MnO2, the SDC2M specimens can obtain their maximum shrinkage and requisite relative densities (>92 %) at 1400 ºC. SDC should be sintered at 1600 ºC to achieve similar shrinkage and relative density values. The sinterabilities 7

of SDC2M+Si specimens varied in different ways when the SiO2 content increased compared with those of SDC2M series samples, including SDC2Co+Si, SDC2Fe+Si, and SDC2Mn+Si, which showed slightly lower sinterabilities below 1400 ºC but were able to densify at 1400 ºC. SDC2Ba+Si showed comparable sinterability with SDC2Ba;

IP T

whereas SDC2Li+Si presented the worst sinterability (even worse than that of pristine SDC sample). Thus, the addition of Li2O failed to densify the SDCSi sample.

SC R

Consequently, the electrochemical property of SDC2Li+Si is not comparable with other densified specimens.

U

Table 1 also lists the average grain sizes of the sintered dense specimens. During

N

the sintering process, the additives and the impurities may temporarily segregate at the

A

grain boundary areas when the temperature is below their solution temperatures.

M

Nevertheless, grain growth can be increased or decreased by these grain boundary

ED

segregations. Meanwhile, the grain-boundary segregated particles inhibit grain growth by pinning the boundaries. If these grain boundary segregations can react with each

PT

other to form low melting point eutectic compounds, the grain growth process can be

CC E

accelerated by rapid diffusion within the liquid phase during sintering. The much smaller grain size of SDCSi (5.138 μm) than that of SDC (8.933 μm) indicated that the addition of 2000 ppm SiO2 suppressed the grain growth. For the SDC2M specimens,

A

the addition of 2000 ppm SiO2 exhibited a slight positive (for SDC2Fe and SDC2Mn) or negative effect (for SDC2Ba and SDC2Co) on the grain growth. The SDC2Li showed relatively high average grain size value (7.573 μm), which indicated that Li can enhance strongly the sinterability of SDC (~150 ppm SiO2 level). When the SiO2 8

impurity content increased to 2000 ppm, the addition of Li did not show any sinteringaid effect. This finding suggested that SiO2 may react with Li, and 2000 ppm SiO2 may be sufficient to consume the remaining Li. Actually, Duan et al. investigated the Li-SiO system and suggested that, with increased percentage of SiO2, several different

IP T

lithium silicate phases can exist above 1000 ºC (e.g., Li8SiO6, Li4SiO4, Li6Si2O7, Li2SiO3, Li2Si2O5, and Li3Si2O7) [24]. This result may explain why 2 mol% Li cannot

SC R

show any sintering-aid effect on SDC2Li+Si specimen (containing ~2000 ppm; ie, 0.6

U

mol% SiO2).

N

3.3. Electrochemical impedance spectroscopy (EIS) analysis

A

Fig. 4 shows the AC impedance (Nyquist) spectra of the dense samples measured

M

in air at 348 °C. The relaxation time for oxygen ion generally transfers through well-

ED

separated grain interior, grain boundary, and electrode. Thus, the AC impedance plots contained three arcs, which correspond to electrode, grain boundary, and grain interior

PT

contributions in increasing frequency. In the present work, specimens exhibit a typical

CC E

spectrum with two arcs. The first arc represents the grain boundary contribution wherein the capacitance lies in the nF range [25,26]. The grain interior arc was not well resolved due to the limited frequency range used during the measurements (0.1 Hz−100

A

kHz). Fortunately, the grain boundary contribution can be clearly distinguished in the impedance spectra below 500 °C. Thus, the grain interior resistance of the specimens can be estimated by subtracting the grain boundary contributions from the total resistance (𝑅𝑔𝑖 = 𝑅𝑡𝑜𝑡 − 𝑅𝑔𝑏 ). 9

Table 2 lists the values of the grain interior resistivity (𝑅𝑔𝑖 ), grain boundary resistivity (𝑅𝑔𝑏 ), and total resistivity (𝑅𝑡𝑜𝑡 ) of the prepared dense specimens. For most specimens, 𝑅𝑔𝑖 values were similar at 16.7‒29.4 Ω m, whereas the SDC2Fe and SDC2Li specimens showed much higher 𝑅𝑔𝑖 values (50.97 and 45.88 Ω m,

IP T

respectively). This finding may be attributed mainly to the relatively lower densities of SDC2Fe (92.57% relative density) and SDC2Li (92.89% relative density), whereas the

SC R

other specimens showed higher relative density values (above 95%). Compared with

𝑅𝑔𝑖 , 𝑅𝑔𝑏 values showed significant variation. The 𝑅𝑔𝑏 of SDC (containing ~150 ppm

U

SiO2) was 62.49 Ω m whereas that of SDC+Si was 264.17 Ω m. The increase in 𝑅𝑔𝑏

N

caused by the higher content of SiO2 clearly demonstrated the negative impact of the

A

siliceous impurity on the grain boundary conductivity of SDC. However, the 𝑅𝑔𝑏

M

value decreased to 91.59 Ω m, 66.96 Ω m, 33.4 Ω m, and 22.45 Ω m by the addition of

ED

2 mol% Co, Fe, Mn, and Ba for the highly impure (2000 ppm SiO2) group specimens “SDC2M+Si,” respectively. Thus, the addition of these four kinds of sintering aids can

PT

effectively alleviate the deleterious effect of 2000 ppm SiO2 impurity on the grain-

CC E

boundary conduction of SDC. Table 2 also shows remarkable decrease in grain-boundary resistivity for

SDC2Fe+Si, SDC2Mn+Si, and SDC2Ba+Si versus SDC2Fe, SDC2Mn, and SDC2Ba,

A

respectively. Furthermore, the grain-boundary resistivity values of SDC2Mn+Si (33.4 Ω m) and SDC2Ba+Si (22.45 Ω m) were even much lower than that of SDC (62.49 Ω m), and SDC2Fe+Si (66.96 Ω m) was very close to SDC.

10

3.4. Microstructure analysis and discussion To understand the changes in sinterability and electrical behavior, the specimens were observed through SEM. For SDC, SDCSi and most of the moderately impure (150 ppm SiO2) group specimens “SDC2M,” a second phase can hardly be found, and the

IP T

surface microstructure appeared homogeneous (as shown in Fig. S1). According to previous reports [11], the resistive siliceous impurity phase may spread along the grain

SC R

boundaries and result in a highly dispersed distribution. For the same reason, the secondary phases in “SDC2M” may simply be highly dispersed and low in content.

U

However, the increased secondary phases would be prone to accumulate and should be

N

easier to distinguish with the increase of SiO2 impurity content. This phenomenon

A

provides us the possibility to detect the secondary phases and analyze the relative

ED

M

reaction mechanisms.

3.4.1. The effect of Ba

PT

Fig. 5 shows the surface morphologies of the sintered SDC2Ba+Si specimen.

CC E

Unlike the SDC2Ba specimen, the SDC2Ba+Si specimen presents a unique surface “hairy grains” structure, as shown in Fig. 5a (magnification 3000×). Higher magnification of SEM image (Fig. 5b) revealed the nano-sized secondary phases grown

A

from the grain boundaries and coated on the top grains. Besides these nano-sized secondary phases, another type of secondary phase was also found on the surface of SDC2Ba+Si specimen (marked as “d” area in Fig. 5c) according to the EDS analysis. This secondary phase with large particles was mainly composed of Ba, Ce, and O 11

elements (the average Ba/Ce ratio was close to 1:1). The amount of this type of secondary phase with large grains is very limited and prone to segregating randomly. By contrast, nano-sized secondary phases are widespread. As a consequence, identifying them was not conclusive given the thinness of these nano-sized phases. EDS

IP T

analyzed that from the area “e” in Fig. 5c, five major elements, including Ba, Si, Ce, Sm, and O (Fig. 5e) were revealed. Considering the thinness and sparseness of these

SC R

nano-sized phases, the presence of Ce and Sm may arise from the SDC bulk. Therefore, this widespread secondary phase may be mainly composed of Ba, Si, and O.

U

In fact, since the siliceous impurity is known to be acidic as the cationic field

N

strength of SiO2 is 1.57, basic oxides such as MgO [27], CaO [28], SrO [29], and BaO

A

[30] (cationic field strengths are 0.45, 0.33, 0.28, and 0.23, respectively), have been

M

examined and proved to be the effective scavenger materials [31]. These basic oxides

ED

have high chemical affinity toward the acidic intergranular siliceous impurities. With the direct evidence from microstructural analyses, including transmission electron

PT

microscopy, electron energy-loss spectroscopy, and electron back-scattered diffraction,

CC E

the scavenging mechanisms of MgO and BaO were attributed to the formation of Mg2SiO4 [27] and Ba2Si3O8 [30], respectively.

A

3.4.2. Effect of Co, Fe, and Mn Several transition metal oxides (TMOs) have been examined as sintering aids and/or as SiO2 impurity scavengers for doped ceria electrolytes [16,18-23]. For the sintering aids, the TMOs can promote the grain boundary mobility with a viscous flow 12

sintering mechanism and reduce the sintering temperature by a few hundred degrees [16,18-21]. By contrast, the exact effect of TMOs on the ionic conduction (especially grain boundary conduction) of doped ceria electrolytes is still debatable. For example, some researchers [32,33] suggested that the addition of cobalt in highly pure

IP T

Ce0.8Gd0.2O1.9 (GDC) did not produce major changes in conduction but caused decrease of grain boundary conduction in impure GDC samples. Other researchers have obtained

SC R

clear evidence of enhanced grain boundary conduction with cobalt addition [18,22,23]. Furthermore, the detailed scavenging mechanisms were still not fully elucidated in

U

relevant literature even for some reported TMO scavengers (e.g., Fe2O3) [32,34,35].

N

In the present work, we have examined the microstructure of the SDC2Co+Si,

A

SDC2Fe+Si, and SDC2Mn+Si specimens by using SEM and EDS. Unlike SDC2Ba+Si,

M

these TMO-added ceria electrolytes showed regular ceramic surface morphologies, and

ED

no apparent secondary phases can be observed. Fortunately, the use of SEM backscattering mode showed local inhomogeneities visible as small dark particles of

PT

electrolytes, and the EDS analysis results confirmed these dark particles to be the Si-

CC E

contained secondary phases. Fig. 6 shows the SEM back-scattering images and the secondary particle EDS results of the SDC2Co+Si, SDC2Fe+Si, and SDC2Mn+Si specimens. For SDC2Co+Si and SDC2Mn+Si specimens, the secondary phase particles

A

were formed predominantly at grain boundaries but also within the SDC grains (as shown in Figs. 6a and e). By contrast, the secondary phase particles of the SDC2Fe+Si specimen were mainly located at grain boundaries (as shown in Fig. 6c). It should be noted that, the compositions of the secondary phases in SDC2Co+Si, SDC2Fe+Si, and 13

SDC2Mn+Si were nearly identical: high levels of Sm and Si and moderate amounts of Ce and Ca (see Figs. 6b, d, and f). Considering the fact that EDS spectra can also be collected from background regions, the presence of Ce and partial Sm may arise from the SDC bulk in addition on these secondary phases. Thus, the Si, Ca, and most of Sm

IP T

elements found in the EDS spectra basically originated from the secondary phases. In our previous reports [21,36], the addition of 1 mol% ZnO has proven to be

SC R

effective in alleviating the deleterious effect caused by SiO2 impurities on the grain

boundary conduction of ceria-based electrolytes. Meanwhile, identical secondary

U

phases were detected both in the inner and surface of electrolytes, which mainly

N

consisted of Si (~35%), Ca (~5%), and trivalent rare-earth element (~55%) (as shown

A

in Fig. 7a). Figs. 6 and 7 presents the compositions of secondary phases detected in Co-,

M

Fe-, and Mn-added impure SDC, which are consistent with Zn-added counterpart that

ED

has been reported in detail [36]. These siliceous impurity phases were Ca-dependent but TMO-independent. According to our latest report [37], CaO acted as the “scavenger”

PT

of SiO2 impurity in the SDC1Ca1Zn+Si system, whereas ZnO simply acted as the

CC E

“promoter” of the grain boundary Si-scavenging process. In addition, the “scavenger + promoter” strategy dramatically enhanced the SiO2 impurity tolerance of SDC. Obviously, the role of Co, Fe, and Mn in the grain boundary Si-scavenging process may

A

also be the “promoter.” As the “scavenger,” the mobility of CaO was assumed to be negligible during the sintering process. Thus, the rate of Si-scavenging reaction may be mainly controlled by the movement of the molten intergranular siliceous phases. The rate of scavenging reaction was accelerated, and more dispersed impurities were 14

transported to the surfaces of CaO particles when TMOs were added. Thus, the intergranular impurities were gathered effectively. In theory, the cationic field strength of TMOs is mainly within the range of 0.7‒ 0.9. Thus, the chemical affinity of TMOs toward the acidic SiO2 is much lower than

IP T

those of alkaline earth metal oxides. Hence, the selectivity of the scavenging reaction is reasonable. It needs to be emphasized that the positive effect of TMOs on the

SC R

conduction of impure SDC based on the premise that CaO is one of the initial impurities

(~160 ppm) should be emphasized. Therefore, the effect of the sintering aids on the

U

conduction of ceria-based electrolytes may be determined by the types and contents of

N

the initial impurities, which may explain the conflicting results from previous reports

A

[18,22,23,32,33] concerning the exact role of sintering aids on the grain boundary

PT

4. Conclusions

ED

M

conductions of ceria-based electrolytes in some ways.

CC E

The effect of sintering aids on the sinterability and electrical properties of SDC electrolytes strongly interacted with the concentration of SiO2 impurity. For the sinterability, both of the moderately- (~150 ppm SiO2) and highly-impure (~2000 ppm

A

SiO2) SDCs can be densified at 1400 °C (reduced by 200 °C) by adding 2 mol% oxide of Ba, Co, Fe, or Mn individually. The oxide of Li has been found to be one of the most effective sintering aids for the moderately impure SDC, but the sintering-aid effect of the oxide of Li was completely neutralized when the content of SiO2 impurity increased

15

to 2000 ppm. For the electrical property, the deleterious effect of the 2000 ppm SiO2 impurity on the grain boundary conduction of SDC can be effectively alleviated by the addition of Ba-, Co-, Fe-, or Mn-oxides. SEM combined with EDS revealed that the scavenging mechanisms of the four kinds of sintering aids were different. Ba was found

IP T

to react directly with SiO2, whereas the reactions between sintering aids and SiO2 were not observed in Co, Fe, and Mn counterparts. Ca (comes from raw materials) and Sm

SC R

were observed to react with Si to form a secondary phase. Co, Fe, and Mn may act as

“promoter” in this type of reaction, which is similar to our previously reported effect of

N

U

Zn.

A

Acknowledgements

M

The authors are grateful for the financial supported by the National Natural

ED

Science Foundation of China (No. 51502136), the Natural Science Foundation of Jiangsu Province (No. BK20140943), and Jiangsu Planned Projects for Postdoctoral

PT

Research Funds (No. 1601026C). We also acknowledge the support of the Priority

CC E

Academic Program Development (PAPD) of Jiangsu Higher Education Institutions and Top-notch Academic Programs Project of Jiangsu Higher Education Institutions

A

(TAPP).

16

References [1] B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies, Nature 414 (2001) 345-352. [2] E.D. Wachsman, K.T. Lee, Lowering the temperature of solid oxide fuel cells, Science 334 (2011) 935-939. [3] D.A. Medvedev, J.G. Lyagaeva, E.V. Gorbova, A.K. Demina, P. Tsiakaras, Advanced materials for SOFC application: Strategies for the development of highly conductive and stable solid oxide proton electrolytes, Prog. Mater. Sci. 75 (2016) 38-79.

IP T

[4] Y. Guo, M. Bessaa, S. Aguado, M.C. Steil, D. Rembelski, M. Rieu, J.-P. Viricelle, N. Benameur, C. Guizard, C. Tardivat, P. Vernoux, D. Farrusseng, An all porous solid oxide fuel cell (SOFC): a bridging technology between dual and single chamber SOFCs, Energy Environ. Sci. 6 (2013) 2119-2123.

SC R

[5] D.J.L. Brett, A. Atkinson, N.P. Brandon, S.J. Skinner, Intermediate temperature solid oxide fuel cells, Chem. Soc. Rev. 37 (2008) 1568-1578.

[6] L. Ge, K. Guo, L. Guo, Sinterability, reducibility, and electrical conductivity of fast oxide-ion conductors La1.8R0.2MoWO9 (R=Pr, Nd, Gd and Y), Ceram. Int. 41 (2015) 10208-10215.

[7] S.P. Jiang, L. Zhang, H.Q. He, R.K. Yap, Y. Xiang, Synthesis and characterization of lanthanum

U

silicate apatite by gel-casting route as electrolytes for solid oxide fuel cells, J. Power Sources 189 (2009) 972-981.

N

[8] L. Zhang, H.Q. He, H.W. Wu, C.Z. Li, S.P. Jiang, Synthesis and characterization of doped La9ASi6O26.5 (A=Ca, Sr, Ba) oxyapatite electrolyte by a water-based gel-casting route, Int. J. Hydrogen

A

Energ. 36 (2011) 6862-6874.

M

[9] J. An, J. Bae, S. Hong, B. Koo, Y.-B. Kim, T.M. Gür, F.B. Prinz, Grain boundary blocking of ionic conductivity in nanocrystalline yttria-doped ceria thin films, Scr. Mater. 104 (2015) 45-48. [10] F. Giannici, G. Gregori, C. Aliotta, A. Longo, J. Maier, A. Martorana, Structure and Oxide Ion

ED

Conductivity: Local Order, Defect Interactions and Grain Boundary Effects in Acceptor-Doped Ceria, Chem. Mater. 26 (2014) 5994-6006.

[11] X. Guo, R. Waser, Electrical properties of the grain boundaries of oxygen ion conductors: Acceptor-

PT

doped zirconia and ceria, Prog. Mater. Sci. 51 (2006) 151-210. [12] J. Tanaka, J.-F. Baumard, P. Abelard, Nonlinear electrical properties of grain boundaries in an oxygen-ion conductor (CeO2·Y2O3), J. Am. Ceram. Soc. 70 (1987) 637-643.

CC E

[13] R. Gerhardt, A.S. Nowick, M.E. Mochel, I. Dumler, Grain-boundary effect in ceria doped with trivalent cations: Ⅱ, microstructure and microanalysis, J. Am. Ceram. Soc. 69 (1986) 647-651. [14] X. Guo, W. Sigle, J. Maier, Blocking grain boundaries in yttria-doped and undoped ceria ceramics of high purity, J. Am. Ceram. Soc. 86 (2003) 77-87. [15] B.C.H. Steele, Appraisal of Ce1-yGdyO2-y/2 electrolytes for IT-SOFC operation at 500 ºC. Solid State

A

Ionics 129 (2000) 95-110. [16] L. Gao, M. Zhou, Y. Zheng, H. Gu, H. Chen, L. Guo, Effect of zinc oxide on yttria doped ceria, J. Power Sources, 195 (2010) 3130-3134. [17] T.S. Zhang, J. Ma, H.T. Huang, P. Hing, Z.T. Xia, S.H. Chan, J.A. Kilner, Effects of dopant concentration and aging on the electrical properties of Y-doped ceria electrolytes, Solid State Sci. 5 (2003) 1505-1511. [18] J. Ayawanna, D. Wattanasiriwech, S. Wattanasiriwech, P. Aungkavattana, Effects of cobalt metal addition on sintering and ionic conductivity of Sm(Y)-doped ceria solid electrolytes for SOFC, Solid 17

State Ionics 180 (2009) 1388-1394. [19] L. Ge, R. Li, S. He, H. Chen, L. Guo, Effect of titania concentration on the grain boundary conductivity of Ce0.8Gd0.2O1.9 electrolyte, Int. J. Hydrogen Energ. 37 (2012) 16123-16129. [20] M. Zhou, L. Ge, H. Chen, L. Guo, Effect of transition metal oxides doping on Ce0.9Sm0.05Nd0.05O1.95 solid electrolyte materials, J. Adv. Ceram. 1 (2012) 150-156. [21] L. Ge, S. Li, Y. Zheng, M. Zhou, H. Chen, L. Guo, Effect of zinc oxide doping on the grain boundary conductivity of Ce0.8Ln0.2O1.9 ceramics (Ln=Y, Sm, Gd), J. Power Sources 196 (2011) 6131-6137. [22] D. Pérez-Coll, D. Marrero-López, P. Núñez, S. Piñol, J.R. Frade, Grain boundary conductivity of Ce0.8Ln0.2O2-δ ceramics (Ln=Y, La, Gd, Sm) with and without Co-doping, Electrochim. Acta 51 (2006)

IP T

6463–6469.

[23] G.S. Lewis, A. Atkinson, B.C.H. Steele, J. Drennan, Effect of Co addition on the lattice parameter, electrical conductivity and sintering of gadolinia-doped ceria, Solid State Ionics 567 (2002) 152–153.

SC R

[24] Y. Duan, H. Pfeiffer, B. Li, I. C. Rome-Ibarra, D.C. Sorescu, D.R. Luebke, J.W. Halley, CO2 capture properties of lithium silicates with different ratios of Li2O/SiO2: an ab initio thermodynamic and experimental approach, Phys. Chem. Chem. Phys. 15 (2013) 13538-13558.

[25] R. Gerhardt, A.S. Nowick, Grain-boundary effect in ceria doped with trivalent cations: I, Electrical measurements, J. Am. Ceram. Soc. 69 (1986) 641-646.

U

[26] J.T.S. Irvine, D.C. Sinclair, A.R. West, Electroceramics: characterization by impedance spectroscopy, Adv. Mater. 2 (1990) 132-138.

N

[27] Y.H. Cho, P.-S. Cho, G. Auchterlonie, D.K. Kim, J.-H. Lee, D.-Y. Kim, H.-M. Park, J. Drennan, Enhancement of grain-boundary conduction in gadolinia-doped ceria by the scavenging of highly

A

resistive siliceous phase, Acta Mater. 55 (2007) 4807-4815.

M

[28] P.-S. Cho, S.B. Leeb, Y.H. Cho, D.-Y. Kim, H.-M. Park, J.-H. Lee, Effect of CaO concentration on enhancement of grain-boundary conduction in gadolinia-doped ceria, J. Power Sources 183 (2008) 518523.

ED

[29] J.A. Lane, J.L. Neff, G.M. Christie, Mitigation of the deleterious effect of silicon species on the conductivity of ceria electrolytes, Solid State Ionics 177 (2006) 1911-1915. [30] S.-Y. Park, P.-S. Cho, S.B. Lee, H.-M. Park, J.-H. Lee, Improvement of grain-boundary conduction

PT

in SiO2-doped GDC by BaO addition, J. Electrochem. Soc. 156 (2009) B891-B896. [31] J.-H. Lee, Highly resistive intergranular phases in solid electrolytes: an overview, Monatsh. Chem. 140 (2009) 1081-1094.

CC E

[32] T.S. Zhang, J. Ma, Y.J. Leng, S.H. Chan, P. Hing, J.A. Kilner, Effect of transition metal oxides on densification and electrical properties of Si-containing Ce0.8Gd0.2O2-δ ceramics, Solid State Ionics 168 (2004) 187-195.

[33] T.S. Zhang, J. Ma, L.B. Kong, P. Hing, Y.J. Leng, S.H. Chan, J.A. Kilner, Sinterability and ionic conductivity of coprecipitated Ce0.8Gd0.2O2-δ powders treated via a high-energy ball-milling process, J.

A

Power Sources 124 (2003) 26-33. [34] T.S. Zhang, J. Ma, S.H. Chan, J.A. Kilner, Grain boundary conduction of Ce 0.9Gd0.1O2-δ ceramics derived from oxalate copercipitation: effects of Fe loading and sintering temperature, Solid State Ionics 176 (2005) 377-384. [35] Q. Dong, Z.H. Du, T.S. Zhang, J. Lu, X.C. Song, J. Ma, Sintering and ionic conductivity of 8YSZ and CGO10 electrolytes with small addition of Fe2O3: A comparative study, Int. J. Hydrogen Energ. 34 (2009) 7903-7909. [36] L. Ge, R. Li, S. He, H. Chen, L. Guo, Enhanced grain-boundary conduction in polycrystalline 18

Ce0.8Gd0.2O1.9 by zinc oxide doping: scavenging of resistive impurities, J. Power Sources, 230 (2011) 161-168. [37] L. Ge, Q. Ni, G. Cai, T. Sang, L. Guo, Improving SiO2 impurity tolerance of Ce0.8Sm0.2O1.9: synergy

A

CC E

PT

ED

M

A

N

U

SC R

IP T

of CaO and ZnO in scavenging grain-boundary resistive phases, J. Power Sources, 324 (2016) 582-588.

19

Fig. 1. XRD patterns of SDC2Co+Si, SDC2Fe+Si, SDC2Li+Si, SDC2Mn+Si and SDC2Ba+Si.

Fig. 2. Linear shrinkage of SDC2M and SDC2M+Si specimens with different “M” versus sintering temperature. (a) M = Ba, (b) M= Co, (c) M= Fe, (d) M= Li, (e) M= Mn.

IP T

Fig. 3. Relative density of SDC2M and SDC2M+Si specimens with different “M”

SC R

versus sintering temperature. (a) M= Ba, (b) M= Co, (c) M= Fe, (d) M= Li, (e) M= Mn.

Fig. 4. Complex impedance spectra of (a) SDC (contains 150 ppm SiO2) versus SDC+Si (contains 2000 ppm SiO2); (b) SDC2Ba versus SDC2Ba+Si; (c) SDC2Co versus

U

SDC2Co+Si; (d) SDC2Fe versus SDC2Fe+Si; (e) SDC2Li and (f) SDC2Mn versus

A

N

SDC2Mn+Si specimens measured at 348 °C.

Fig 5. SEM images of the unpolished surface of SDC2Ba+Si under different

M

magnifications (a) 3000×, (b) 10000×, and (c) 5000×; (d) and (e) are EDS spectra

ED

obtained from the specific areas marked in (c).

Fig. 6. SEM back-scattering-mode micrographs and corresponding EDS spectra of the

PT

secondary phases (marked with circles) of SDC2Co+Si (a and b), SDC2Fe+Si (c and

CC E

d), and SDC2Mn+Si (e and f) samples.

Fig. 7 (a) TEM image of siliceous phase and relevant EDS spectrum in Ce0.8Gd0.2O1.9 containing 1 mol% ZnO and 2000 ppm SiO2 [36]. (b) Fractured cross-section of SDC

A

containing 1 mol% ZnO, 1 mol% CaO, and 4000 ppm SiO2 [37].

20

Intensity (Arb. Unit)

SDC2Ba+Si SDC2Co+Si

SC R

SDC2Li+Si

IP T

SDC2Fe+Si

SDC2Mn+Si

ICDD#34-0394 CeO2 40

50o

2 ( )

60

U

30

70

80

M

A

N

20

Fig. 1. XRD patterns of SDC2Co+Si, SDC2Fe+Si, SDC2Li+Si, SDC2Mn+Si and

A

CC E

PT

ED

SDC2Ba+Si.

21

16

(b)

(a)

12 8

SDC SDC2Ba SDC2Ba+Si 1200

1250

1300

1350

1400

1450

(c)

1200

1250

(d)

12 8

1250

1300 16

1350

1400

1450

(e)

12 8 4

ED

0

1200

1200

1400

1450

1250

1300

1350

1400

1450

N

1200

A

0

1350

SDC SDC2Li SDC2Li+Si

U

SDC SDC2Fe SDC2Fe+Si

4

1300

SC R

16

M

Linear Shrinkage (%)

0

SDC SDC2Co SDC2Co+Si

IP T

4

1250

1300

1350

SDC SDC2Mn SDC2Mn+Si 1400

1450

o

PT

Sintering temperature ( C)

CC E

Fig. 2. Linear shrinkage of SDC2M and SDC2M+Si specimens with different “M”

versus sintering temperature. (a) M = Ba, (b) M= Co, (c) M= Fe, (d) M= Li, (e) M=

A

Mn.

22

90

(a)

(b)

IP T

100

80 70

1250

1300

1350

1400

1450

1200

1250

1300 100 90

1350

(e)

80

1400

1450

1200

ED

70

50

PT

1250

1300

1350

1400

1450

SDC SDC2Li SDC2Li+Si 1400

1450

SDC SDC2Mn SDC2Mn+Si

60

1200

1350

U

A

SDC SDC2Fe SDC2Fe+Si

60

1300

N

70

1200

1250

(d)

(c)

80

50

SC R

90

1200

M

Relative density (%)

50 100

SDC SDC2Co SDC2Co+Si

SDC SDC2Ba SDC2Ba+Si

60

1250

1300

1350

1400

1450

o

CC E

Sintering temperature ( C)

Fig. 3. Relative density of SDC2M and SDC2M+Si specimens with different “M”

A

versus sintering temperature. (a) M= Ba, (b) M= Co, (c) M= Fe, (d) M= Li, (e) M= Mn.

23

Rgb

30 0

0

50

100

150

200

250

300

ZcosA/L [m]

-ZsinA/L [m]

60

90

120

ZcosA/L [m] 40

(e)

150

-ZsinA/L [m]

30

A

Co 0

ED

SDC2Li

20

Li

0

0

30

60

90

120

0

Rgi

Rgb

0

30

60

40

90

Ba

120

(d)

20

Rgi

150

ZcosA/L [m]

0

SDC2Fe SDC2Fe+Si

Fe

0

30

60

90

120

40

SDC2Mn SDC2Mn+Si

(f) 20

0

Rgi

Rgb

Mn 0

30

60

90

120

ZcosA/L [m]

Fig. 4. Complex impedance spectra of (a) SDC (contains 150 ppm SiO2) versus

SDC+Si (contains 2000 ppm SiO2); (b) SDC2Ba versus SDC2Ba+Si; (c) SDC2Co versus SDC2Co+Si; (d) SDC2Fe versus SDC2Fe+Si; (e) SDC2Li and (f) SDC2Mn versus SDC2Mn+Si specimens measured at 348 °C. 24

150

ZcosA/L [m]

CC E A

150

Rgb

N

Rgb

Rgi

M

0

SDC2Co SDC2Co+Si

PT

-ZsinA/L [m]

-ZsinA/L [m]

20

20

SDC2Ba SDC2Ba+Si

ZcosA/L [m]

40

(c)

(b)

IP T

Rgi

60

40

SC R

SDC SDC+Si

U

(a)

-ZsinA/L [m]

-ZsinA/L [m]

90

150

IP T SC R U N A M ED PT CC E A

Fig 5. SEM images of the unpolished surface of SDC2Ba+Si under different

magnifications (a) 3000×, (b) 10000×, and (c) 5000×; (d) and (e) are EDS spectra obtained from the specific areas marked in (c).

25

IP T SC R U N A M ED PT CC E

Fig. 6 SEM back-scattering-mode micrographs and corresponding EDS spectra of the

A

secondary phases (marked with circles) of SDC2Co+Si (a and b), SDC2Fe+Si (c and d), and SDC2Mn+Si (e and f) samples.

26

IP T SC R U N A M ED

Fig. 7 (a) TEM image of siliceous phase and relevant EDS spectrum in Ce0.8Gd0.2O1.9

PT

containing 1 mol% ZnO and 2000 ppm SiO2 [36]. (b) Fractured cross-section of SDC

A

CC E

containing 1 mol% ZnO, 1 mol% CaO, and 4000 ppm SiO2 [37].

27

Table 1. Specifications, sintering temperatures, refined cell parameters, and average grain size of all specimens sintered for 10 h. Stoichiometry

Sintering Temperature (°C)

Cell Parameter (nm) [a]

Average Grain Size (μm) [b]

SDC

Ce0.8Sm0.2O1.9 +150 ppm SiO2

1600

0.54361±0.0002

SDC2Ba

SDC+2 mol% Ba+150 ppm SiO2

1400

0.54392±0.0001

SDC2Co

SDC+2 mol% Co+150 ppm SiO2

1400

0.54335±0.0003

SDC2Fe

SDC+2 mol% Fe+150 ppm SiO2

1400

0.54325±0.0001

1.483

SDC2Li

SDC+2 mol% Li+150 ppm SiO2

1400

0.54302±0.0003

7.573

SDC2Mn

SDC+2 mol% Mn+150 ppm SiO2

1400

0.54326±0.0002

3.496

SDCSi

SDC+2000 ppm SiO2

1600

0.54322±0.0003

5.138

SDC2Ba+Si

SDC+2 mol% Ba+2000 ppm SiO2

1400

SDC2Co+Si

SDC+2 mol% Co+2000 ppm SiO2

1400

SDC2Fe+Si

SDC+2 mol% Fe+2000 ppm SiO2

SDC2Li+Si

SDC+2 mol% Li+2000 ppm SiO2

SDC2Mn+Si

SDC+2 mol% Mn+2000 ppm SiO2

IP T

Sample

8.933 2.329

N

U

SC R

7.480

1.877

0.54341±0.0002

6.477

1400

0.54331±0.0002

2.457

1400

0.54309±0.0003

--

0.54319±0.0002

4.745

ED

M

A

0.54354±0.0001

1400

Cell parameters refined by MDI JADE 5.0 software.

[b]

Average grain size measured by SmileView (Ver. 2.1) software.

A

CC E

PT

[a]

28

IP T

Table 2. Grain interior (Rgi), grain boundary (Rgb), and overall (Rtot) resistivity values of

SC R

specimens measured in air at 348 °C.

Resistivity (Ω m) SDC+Si

SDC2Ba

SDC2Ba+Si SDC2Co SDC2Co+Si

Rgi

19.79

21.62

27.92

25.78

29.38

Rgb

62.49

264.17

33.57

22.45

59.09

91.59

Rtot

82.28

285.79

61.49

48.23

88.47

114.12

SDC2Fe+Si

SDC2Li

50.97

23.56

45.88

85.09

66.96

77.82

99.67

33.4

90.52

123.7

116.39

56.53

136.06

A

CC E

PT

ED

M

A

N

22.53

SDC2Fe

U

SDC

29

SDC2Mn SDC2Mn+Si 16.72

23.13