Synthesis of Porous Submicron Gd_0.1Ce_0.9O_1.95 Particles by Carbon Nanoparticle-Addition Ultrasonic Spray Pyrolysis (CNA-USP) and Nanoparticle Preparation by Ball Milling

Abstract

A new ultrasonic spray pyrolysis method, called carbon nanoparticle-addition ultrasonic spray pyrolysis (CNA-USP), is developed to synthesize nanoparticles of electrolyte material for solid oxide fuel cell applications. In CNA-USP, carbon nanoparticles are added in a precursor solution. First, Gd_0.1Ce_0.9O_1.95 (GDC) particles were synthesized from an aqueous solution of Ce(NO₃)₃ 6H₂O and Gd(NO₃)₃ 6H₂O by using the CNA-USP method. The resulting synthesized GDC particles were agglomerated, porous, primary particles on the order of 10 nm in diameter. EDX images revealed uniform distributions of Ce, Gd, and O in these porous particles. Then, these agglomerated, porous submicron GDC particles were ground into primary nanoparticles by ball milling for 24 h. The average diameter of the ground GDC nanoparticles was about double of their average crystallite size.

Copyright 2014 American Association for Aerosol Research

INTRODUCTION

Solid oxide fuel cells (SOFCs) operated at temperatures of 1173–1273 K have been used primarily for stationary applications such as large-scale stationary power plants, because SOFCs (1) have high energy conversion efficiency, (2) do not require expensive catalysts, such as Pt, and (3)　can use a wide range of fuels, including natural gas, propane, butane, and alcohols. However, high temperatures require the use of expensive ceramic components and preclude the use of less expensive metal components for any of the non-electrochemical components of the fuel cell, thus increasing the likelihood of cracks developing during thermal cycling. To permit the use of less expensive metal components in SOFCs, major research efforts have focused on reducing the operating temperature below 1073 K (Haile 2003). SOFCs consist of an anode, a cathode, and a dense electrolyte sandwiched between these electrodes. To reduce the operating temperature, it is important that the electrolyte material has high ionic conductivity and that thin films about 10–15 μm in thickness are used (Haile 2003). Recently, Ce_1-xGd_xO_2-δ (GDC) and Ce_1-xSm_xO_2-δ have been extensively investigated as potential electrolyte materials because they have high conductivity at intermediate temperatures of 873–973 K (Brandon et al. 2004a,b). The cosintering process of electrolyte films and anode substrates (used as a support for electrolyte films) have been investigated as a fabrication method for intermediate-temperature SOFCs (i.e., IT-SOFC) because they do not use expensive and up-scaling difficult thin-film formation equipment such as CVD and PVD methods (Will et al. 2000; Brandon et al. 2004a,b; Beckel et al. 2007; Liu et al. 2007). In the cosintering process, the anode substrate is coated by using a slurry in which a submicron-meter-sized powder of electrolyte material is suspended to form thin electrolyte films. For IT-SOFC it is optimum for the electrolyte film to be as thin as possible, such as a few hundred nanometers in thickness. However, it is difficult to prepare films thinner than a few microns by using slurry coating. Development of a new thin film formation process is, therefore, necessary to realize ultrathin electrolyte films.

We have proposed a novel process to form ultrathin electrolyte films as follows (Arastoo et al. 2012). First, porous submicron agglomerates of nanoparticles composed of electrolyte material are synthesized and broken up into individual nanoparticles by using liquid ball milling. Next, the electrolyte nanoparticles are suspended in a solvent and the suspension is spread on the anode substrate by using spin coating or spray coating to form a nanoparticle thin film. Finally, the nanoparticle film is annealed in a furnace to form a dense thin film. As a first step in the development of this process, we must develop a new method for synthesizing porous, submicron particles.

Ultrasonic spray pyrolysis (USP) has been typically used to prepare complex oxides of submicron particles because this method can form highly crystalline oxides from one solution in which three or four precursors are dissolved in stoichiometric ratios of complex oxides. Various oxides such as zirconia, yttria-stabilized zirconia, alumina/platinum composite, ZnO, BaO, ceria, NiO-ceria composite, and LiM₂O₄ have been prepared by using USP (Messing et al. 1993; Djurado and Meunier 1998; Okuyama and Lenggoro 2003; Gaudon et al., 2004; Matsuda and Taniguchi 2004; Chen et al., 2008). Additionally, complex oxides with high crystallinity can be continuously prepared during one passage of source mist in a flow-type reactor. Therefore, USP is a strong candidate as an inexpensive synthesis process for SOFC materials. Recently, Suda et al. (2006) and Kawano et al. (2011) used USP to synthesize NiO-Sm_0.2Ce_0.8O_1.9 (NiO-SCO) composite particles for use as SOFC anode materials. Mesguich et al. (2010) used USP to synthesize Nd₂NiO_4+δ particles for use as SOFC cathode materials. Our research group also synthesized Gd_0.1Ce_0.9O_1.95, NiO-Gd_0.1Ce_0.9O_1.95 composites, and La_0.8Sr_0.2Co₃ (LSC) particles for use as electrolyte, anode, and cathode materials, respectively (Kinoshita and Adachi 2014). Taken together, these processes represent a new USP method for synthesizing porous submicron GDC particles that we call carbon nanoparticle-addition ultrasonic spray pyrolysis (CNA-USP) (Arastoo et al. 2012).

In this study, we describe a method for first synthesizing porous submicron GDC particles using the CNA-USP method, and then grinding with a liquid ball mill to prepare the nanoparticle suspension for use in the spin-coating method.

CNA-USP METHOD

Figure 1 shows the concept of the CNA-USP method (Arastoo et al. 2012). First, the aqueous solution of dissolved precursors and suspended carbon nanoparticles is atomized by using an ultrasonic nebulizer, which forms micron-sized droplets. Droplets containing carbon nanoparticles are sent to a furnace via a carrier gas. In the furnace, the complex of precursors is deposited in droplets, and then evaporation of the water causes the formation of either solid or amorphous submicron particles containing carbon nanoparticles. Next, the material is produced from the precursor complex by thermal decomposition and crystallization by further heating of the submicron particles. Finally, carbon nanoparticles in crystallized submicron particles are oxidized to CO₂ gas by combustion and then evaporated from the submicron particles. As a result, porous submicron particles are formed.

FIG. 1. Concept of carbon nanoparticle-addition ultrasonic spray pyrolysis (CNA-USP).

EXPERIMENT

Figure 2 shows a schematic of the experimental apparatus for the CNA-USP method, consisting of (1) an ultrasonic nebulizer to convert from the precursor solution to micron-sized droplets; (2) a cylindrical furnace to form complex oxide particles by thermal decomposition; and (3) an electrostatic precipitator to collect product particles. In the nebulizer, a precursor solution was vibrated by six ultrasonic oscillators at a frequency of 2.4 MHz (Honda Electronics, HM-2412) and was converted to mist droplets a few microns in diameter. The droplets were then carried into the cylindrical reactor (a ceramic tube of 50 mm inner diameter and 640 mm total length) with air at a flow rate Q = 5 L min⁻¹. The reactor was heated by two electric furnaces; an upper flow side furnace (Furnace 1; 130 mm in length) kept at T_L = 673 K, and a lower flow side furnace (Furnace 2; 400 mm in length) set at various T_H from 873 to 1473 K. Both furnaces were set at an interval of 110 mm and the rector tube between both furnaces was covered by the insulator. The average residence time in the reactor zone heated by the Furnace 2 (t_r = 9.4 s at Q = 5 L min⁻¹) was used as the reaction time in CNA-USP method because we thought that the oxidization of carbon nanoparticles and the thermal decomposition of precursor occurred in the high temperature zone. We used the preheating temperature of T_L = 673 K because the temperature of the reactor zone heated by the Furnace 2 did not attain to T_H when the Furnace 1 was not used. In the reactor, solid particles were produced by thermal decomposition of precursors after the evaporation of the solvent and precipitation of the solute in droplets. The synthesized particles were collected by using an electrostatic particle precipitator (EPP) (Okuyama et al. 1993), which was kept at 403 K using an electric heater to prevent the condensation of water on its walls. The crystalline phase of the synthesized particles was characterized by powder X-ray diffraction (XRD, RIGAKU RINT1500) and the particle shape by scanning electron microscopy (SEM, Hitachi S-4500). The local distributions of metals and oxygen in the particles were evaluated based on energy dispersive X-ray spectrometry (EDX) images. The size distributions of agglomerate particles were measured by dynamic light scattering spectrophotometry (DLS, Otsuka Electronics, ELSZ-2ON). The thermalgravimetric (TG) analysis was carried by thermo gravimetry analyzer (Bruker, MS9610) to evaluate the mass of carbon remaining in synthesized particles. The data were measured in helium gas at the heating rate of 10 K min⁻¹.

FIG. 2. Apparatus for synthesizing SOFC electrolyte material particles using ultrasonic spray pyrolysis (CNA-USP).

The precursor solution was prepared by dissolving Ce(NO₃)₃·6H₂O and Gd(NO₃)₃·6H₂O in pure water at a stoichiometric ratio of Gd_0.1Ce_0.9O_1.95. The concentration of total metal ions C_total was kept at 0.02 mol L⁻¹. Carbon nanoparticles with carboxyl group on their surfaces (commercial name is Aqua-Black 162, Tokai Carbon Co.) were selected as the carbon source because they are hydrophilic. The average diameter of carbon nanoparticles measured by DLS (Nikkiso, UPA-EX) was 100 nm. Different concentrations of carbon nanoparticles C_c were evaluated from 0 to 24 g L⁻¹. The maximum production rate of the apparatus and yield calculated from the volume of the sprayed precursor solution and the mass of the collected particles were about 3 g h⁻¹ and 70–80% at the synthesis condition of Q = 5 L min⁻¹ and C_total = 0.02 mol L⁻¹, respectively.

Synthesized particles were then ground by using a tumbling ball mill to qualitatively estimate the toughness of the particles. Particles (100 mg) were suspended in ethanol (70 mL). The suspension (70 mL) was then put in a polypropylene pot (61 mm deep, 57 mm inner diameter) together with zirconia balls (70 g balls of 3.0-mm-diameter). The pot was rotated at 108 rpm and different grinding times t_g were varied from 0.5 to 24 h. The shapes of the particles before and after grinding were observed by SEM and their size distributions by DLS. The specific surface areas of particles before and after grinding were determined by N₂ adsorption using a volumetric adsorption measurement instrument (Bel Japan, Belsorp-18).

Concentrations of Ce³⁺ and Gd³⁺ ions, C_Ce3+ and C_Gd3+, in the precursor solution were measured by using an inductively coupled plasma (ICP) analyzer (Seiko Instruments Inc., SPS7800) to clarify the effects of the adsorption of metal ions onto carbon nanoparticles in the precursor solution. Carbon nanoparticles were suspended in a precursor solution for 24 h to adsorb metal ions. They were then removed from the precursor solution by using centrifugal sedimentation at 26,200 rpm for 3 h, to prepare a solution devoid of carbon nanoparticles. Concentrations of metal ions in this carbon-nanoparticle-free precursor solution were measured by using ICP elemental analysis.

RESULTS AND DISCUSSION

Figure 3 shows SEM images of particles synthesized from aqueous solutions of carbon nanoparticle suspensions with C_c = 0, 8, 16, and 24 g L⁻¹, with T_H = 1273 K. When C_c = 0 g L⁻¹, the particles were dense spheres and did not have pores. When C_c = 8, 16, or 24 g L⁻¹, the particles seem to be porous spheres. We measured BET analysis to make clear whether the morphology on particle surfaces in Figures 3b–d is due to actual pores or not. Table 1 shows the specific surface areas S_w obtained from BET analysis of N₂ adsorption isotherms for Gd_0.1Ce_0.9O_1.95 particles synthesized at T_H = 1073 K, C_c = 0, and 16 g L⁻¹. The addition of carbon nanoparticles increased S_w from 6.31 to 15.1 m² g⁻¹, indicating that the morphology on particle surfaces is actual pores.

TABLE 1 Specific surface areas S_w obtained from N₂ adsorption isotherms of Gd_0.1Ce_0.9O_1.95 particles before and after 24-h grinding; particles were synthesized at T_H = 1073 K, C_c = 0, and 16 g L⁻¹

Cc [g L^−1]	Grinding	Sw [m^2/g]
0	Before	6.31
16	Before	15.1
16	After	22.9

FIG. 3. SEM images of particles synthesized using CNA-USP method from aqueous solutions of Ce(NO₃)₃ 6H₂O and Gd(NO₃)₃ 6H₂O dissolved in stoichiometric ratios of Gd_0.1Ce_0.9O_1.95 at C_c = (a) 0, (b) 8, (c) 16, and (d) 24 g L⁻¹, at C_total = 0.02 mol L⁻¹, T_H = 1273 K, and t_r = 9.4 s.

Table 2 shows the geometric mean diameter D_pg and geometric standard deviation σ_g determined from the SEM images in Figure 3. When C_c was increased from 0 to 16 g L⁻¹, D_pg increased from 212 to 263 nm, respectively. This increase occurs because the combustion of carbon nanoparticles created pores in the submicron particles. In contrast, when C_c was increased further from 16 to 24 g L⁻¹, D_pg decreased from 263 to 224 nm. The decrease in D_pg was caused by (1) the decrease in C_c in the droplets and (2) the decrease in C_total in the source solution. For the cause (1), a portion of the carbon nanoparticles coagulated and formed agglomerates by the ultrasonic vibration when the C_c was high. The agglomerates deposited on the bottom of the nebulizer by gravitational sedimentation, thus causing a decrease in C_c in the droplets. As the results, the diameter D_pg of porous submicron particles decreased because their pore spaces decreased. For the cause (2), it is considered that Ce³⁺ and Gd³⁺ ions in the precursor solution are adsorbed onto the surface of the carbon nanoparticles because used carbon nanoparticles (Aqua Black 162) have carboxyl groups on their surface and carboxyl groups bind metal ions. Concentrations of Ce³⁺ and Gd³⁺ ions, C_Ce3+ and C_Gd3+, in the precursor solution were measured by ICP elemental analysis to clarify the effects of the adsorptions onto the carbon nanoparticles. Table 3 shows C_Ce3+ and C_Gd3+ in the precursor solution when carbon nanoparticles with C_c = 0–24 g L⁻¹were added. Both C_Ce3+ and C_Gd3+ decreased with increasing C_c. At C_c = 24 g L⁻¹, C_Ce3+ and C_Gd3+ decreased to about 60% of their initial values (i.e., for C_c = 0 g L⁻¹). These results suggest that Ce³⁺ and Gd³⁺ ions were adsorbed on the surface of the carbon nanoparticles, probably due to the carboxyl groups on the surface of nanoparticles. In the CNA-USP method, the increase in particles size with increasing porosity and the decrease with decreasing C_total in the solution and C_c in mist droplets occur simultaneously. The former is dominant in the synthesis at C_c = 8 and 16 g L⁻¹ and the latter is at C_c = 24 g L⁻¹, resulting that D_pg decreases at C_c = 24 g L⁻¹.

TABLE 2 Geometric mean diameter D_pg and geometric standard deviation σ_g obtained from SEM images for Gd_0.1Ce_0.9O_1.95 particles shown in Figure 3; particles were synthesized at C_c = 0–24 g L⁻¹ and T_H = 1273 K

Cc [g L^−1]	Dpg [nm]	σg [-]
0	212	1.58
8	254	1.43
16	263	1.37
24	224	1.51

TABLE 3 Concentrations of Ce³⁺ and Gd³⁺, C_Ce3+ and C_Gd3+, and atomic ratios of Ce³⁺ : Gd³⁺ in precursor solutions to which carbon nanoparticles were added

Cc [g L^−1]	CCe3+ [mM]	CGd3+ [mM]	Atomic ratio of Ce^3+ : Gd^3+
0	21.9	2.49	0.90 : 0.10
8	18.5	2.09	0.90 : 0.10
16	15.8	1.71	0.90 : 0.10
24	13.6	1.38	0.91 : 0.09

Figure 4 shows XRD patterns of particles synthesized at different T_H = 873, 1073, 1173, and 1473 K compared to JCPDS data of Gd_0.1Ce_0.9O_1.95 (JCPDS 75-0161). The total metal ion concentration C_total was 0.02 mol L⁻¹, carbon nanoparticle concentration C_c was 16 g L⁻¹, and reaction time t_r was 9.4 s. The XRD patterns for T_H = 1073K, 1173K, and 1473K agree with the JCPDS data of GDC and don't have any peaks except shown in it. Particles synthesized with the CNA-USP method were not affected by the carbon nanoparticles although the C_total in the solution decreased with increasing C_c as shown in Table 3. The molar ratios of Ce³⁺ and Gd³⁺ are shown in Table 3 to make clear the cause. The ratios for C_c = 8–24 g L⁻¹ were similar to the ratio for C_c = 0 g L⁻¹ of 9:1, thus causing that XRD patterns were not affected by the carbon nanoparticles. The XRD pattern for particles synthesized using the CNA-USP method at T_H = 873 K shows peaks at similar locations as for JCPDS data of GDC, although the peaks are broader, indicating that the particles synthesized at T_H = 873 K were amorphous. Table 4 shows that average crystallite size D_cry calculated from the most intense reflection peak by using Scherrer's formula. The crystallite size D_cry of GDC increased from = 31.8 to 48.3 nm with increasing T_H from 1073 to 1473 K, indicating that particles synthesized at T_H ≥ 1073 K were highly crystalline.

TABLE 4 Average crystallite size D_cry obtained from XRD patterns of Gd_0.1Ce_0.9O_1.95 particles shown in Figure 4, geometric mean diameter D_pg and geometric standard deviation σ_g obtained from SEM images shown in Figure 5; particles were synthesized at T_H = 873–1473 K and C_c = 16 g L⁻¹

TH [K]	Dcry [nm]	Dpg [nm]	σg [-]
873	12.4	309	1.39
1073	31.8	271	1.44
1273	41.4	263	1.37
1473	48.3	210	1.53

FIG. 4. XRD patterns of particles synthesized using CNA-USP method from aqueous solutions of Ce(NO₃)₃ 6H₂O and Gd(NO₃)₃ 6H₂O dissolved in stoichiometric ratios of Gd_0.1Ce_0.9O_1.95 at T_H = (a) 873 K, (b) 1073 K, (c) 1273 K, and (d) 1473 K compared to (e) JCPDS data for Gd_0.1Ce_0.9O_1.95, at C_total = 0.02 mol L⁻¹, C_c = 16 g L⁻¹, and t_r = 9.4 s.

Figure 5 shows SEM images of GDC particles synthesized at different T_H = 873, 1073, 1173, and 1473K, and at C_total = 0.02 mol L⁻¹, C_c = 16 g L⁻¹, and t_r = 9.4 s. Synthesized particles were porous spheres a few hundred nanometers in diameter. SEM images indicate that for T_H ≥ 873K, many pores were made on the surface of submicron particles. Table 4 shows D_pg and σ_g determined from SEM images shown in Figure 5, for which 210 < D_pg < 309 nm and 1.37 < σ_g < 1.54. Table 4 indicates that D_pg decreased with increasing T_H, due to the higher temperature sintering between primary particles in the porous submicron particles.

FIG. 5. SEM images of Gd_0.1Ce_0.9O_1.95 particles synthesized using CNA-USP method at T_H = (a) 873 K, (b) 1073 K, (c) 1273 K, and (d) 1473 K, at C_total = 0.02 mol L⁻¹, C_c = 16 g L⁻¹, and t_r = 9.4 s.

The TG data were measured to examine whether carbon nanoparticles remains in porous particles shown in Figure 5 or not. Figure 6 shows the TG curves of (a) GDC particles used in Figure 5 and (b) carbon nanoparticles added in the precursor solutions. The sample particles were heated in the helium gas and the measuring temperatures T_M was increased from 300 to 1400 K at the rate of 10 K min⁻¹. The ordinates of figures are weight ratio normalized by the initial weight. The weight ratio of GDC particles synthesized at T_H = 873 decreases gently from 89.2% to 84.4% at T_M = 400–1200 K after the sharp reduction at T_M = 300–400 K and decreases sharply at T_M ≥ 1200 K. The weights for T_H = 1073 and 1273 K decrease slightly from 100% to 99.5% and 99.6% to 98.5%, respectively, at T_M = 400–1200 K and decrease sharply at T_M ≥ 1200 K. The weight for T_H = 1473 K decreases slightly from 96.3% to 95.6% at T_M = 400–1200 K after the sharp reduction at T_M = 300–400 K and decreases sharply at T_M ≥ 1200 K. The sharp reductions for T_H = 873 and 1473 K at T_M = 300–400 K are due to the evaporation of moisture adsorbed on the particle surface. The TG curve of carbon nanoparticles was measured to make clear the cause of the weight losses at T_M = 400–1200 K. The weight ratio decreases linearly at T_M ≥ 400 K after the sharp reduction at T_M = 300–400 K as shown in Figure 6b, indicating that this weight loss of carbon nanoparticles is the cause for weight losses of the porous GDC particles at T_M = 400–1200 K. This means that the carbon remains in the porous GDC particles. We think that the cause of sharp reductions at T_M ≥ 1200 K are the deoxidization of GDC by the reaction between the remaining carbon and oxygen in the GDC because the reduction ratios increase with weight losses at T_M = 300–400 K.

FIG. 6. TG data of (a) Gd_0.1Ce_0.9O_1.95 particles synthesized using CNA-USP method at T_H = 873 K, 1073 K, 1273 K, and 1473 K, at C_total = 0.02 mol L⁻¹, C_c = 16 g L⁻¹, and t_r = 9.4 s and (b) carbon nanoparticles added in the precursor solutions.

The amount ratios of remaining carbon to carbon added in the solutions can be evaluated from the weight losses at T_M = 400–1200 K. The remaining carbon ratios calculated from the weight losses of 4.8, 0.5, 1.1, and 0.7% are 0.46, 0.03, 0.09, and 0.05% for T_H = 873, 1073, 1273, and 1473 K, respectively. Effects of the remaining carbon on the electrochemical properties of the electrolyte film should be investigated although the remaining carbon in GDC nanoparticles can be removed by the high temperature annealing at around 1473 K in the electrolyte film formation process.

The performance of GDC as an electrolyte depends strongly on the local distribution of components in the GDC particle. The local distribution of Ce, Gd, and O in particles synthesized by using CNA-USP was observed by using EDX. Figure 7 shows (a) SEM images of GDC particles prepared at C_c = 16 g L⁻¹ and EDX images of Gd, Ce, and O in these particles at C_total = 0.02 mol L⁻¹, T_H = 1073 K, and t_r = 9.4 s. These EDX images show that Ce, Gd, and O were distributed uniformly in the particles.

FIG. 7. (a) SEM images of Gd_0.1Ce_0.9O_1.95 particles synthesized using CNA-USP method at C_c = 16 g L⁻¹ and EDX images of (b) Gd, (c) Ce, and (d) O contained in them, at C_total = 0.02 mol L⁻¹, T_H = 1073 K, and t_r = 9.4 s.

In our proposed process for formation of ultrathin electrolyte films, GDC nanoparticles were produced from porous submicron particles by liquid ball milling. Then, porous GDC submicron particles were suspended in pure water and the suspension was ground with the ball mill. Figure 8 shows SEM images of Gd_0.1Ce_0.9O_1.95 particles before and after the ball milling for 30 min, 1 h, 5 h, and 24 h for particles synthesized at C_c = 16 g L⁻¹, C_total = 0.02 mol L⁻¹, T_H = 1073K, and t_r = 9.4 s. These SEM photographs showed that the only an insignificant quantify of porous submicron particles were ground after 30 min of ball milling, whereas a significant quantity were ground after 1 and 5 h, and most were ground after 24 h.

FIG. 8. SEM images of Gd_0.1Ce_0.9O_1.95 particles (a) before and after (b) 0.5 h, (c) 1 h, (d) 5 h, and (e) 24 h of ball-mill grinding of particles synthesized using CNA-USP method at C_c = 16 g L⁻¹, C_total = 0.02 mol L⁻¹, T_H = 1073 K, and t_r = 9.4 s.

Figure 9 shows size distributions of GDC particles measured by DLS before and after ball milling. The size distribution of porous submicron particles before the ball milling showed two peaks, a minor peak at D_p = 180 nm and a major peak at D_p = 290 nm. The particle diameter of the major peak, D_p = 290 nm, is almost the same as that obtained from SEM photographs, D_pg = 271 nm (Table 4). After 30 min of ball milling, a single peak appeared at D_p = 140, although the original peaks at D_p = 180 and 290 nm disappeared. After 1 h and 5 and 24 h of ball milling, the single peak shifted to D_p = 110, 85, and 65 nm, respectively. The D_p = 65 nm after 24 h of ball milling is about double of the D_cry = 31.8 nm determined from XRD pattern (Table 4), which suggests that porous GDC submicron particles synthesized by the CNA-USP method were ground into near crystallite sizes after 24 h of ball milling.

FIG. 9. Size distributions of Gd_0.1Ce_0.9O_1.95 particles measured by DLS before and after 0.5, 1, 5, and 24 h of ball-mill grinding of particles synthesized using CNA-USP method at C_c = 16 g L⁻¹, C_total = 0.02 mol L⁻¹, T_H = 1073 K, and t_r = 9.4 s.

The specific surface areas S_w obtained from BET analysis for the particles synthesized at T_H = 1073 K and C_c = 16 g L⁻¹ before and after 24-h milling are shown in Table 1. By ball milling, S_w slightly increased from 15.1 to 22.9 m² g⁻¹. The increase in S_w is due to grinding of submicron particles to nanometer-sized particles.

Above results suggest that GDC nanoparticles prepared by CNA-USP method are suitable materials for our proposed process in which ultrathin electrolyte film is formed on the anode substrate by spin or spray coating and sintering of electrolyte nanoparticle suspension if carbon or carbon dioxide generated from carbon nanoparticles don't affect the electrochemical properties of the electrolyte film. The characterization of electrochemical properties and the development of the film formation process are needed to use the GDC nanoparticles as the electrolyte film material.

CONCLUSIONS

For IT-SOFC, A new ultrasonic spray pyrolysis method called carbon nanoparticle-addition ultrasonic spray pyrolysis (CNA-USP) is proposed for synthesizing nanoparticles of electrolyte materials used in solid oxide fuel cells (SOFCs). A key step in CNA-USP is the addition of carbon nanoparticles to a precursor solution. The first step in CNA-USP is the synthesis of Gd_0.1Ce_0.9O_1.95 (GDC) particles from an aqueous solution of Ce(NO₃)₃·6H₂O and Gd(NO₃)₃·6H₂O. In our study, GDC particles synthesized at a furnace temperature T_H between 873 and 1473 K and at a carbon nanoparticle concentration C_c between 0 and 24 g L⁻¹ were porous particles between 210 and 309 nm in diameter and between 12 and 48 nm in crystallite size. EDX images of the porous GDC particles showed that Ce, Gd, and O were distributed uniformly in particles. The second step is to grind the porous GDC particles by using a ball mill. In our study, GDC particles prepared at T_H = 1073 K and C_c = 16 g L⁻¹ were ground into nanoparticles by ball milling for about 24 h. The average diameter of the GDC nanoparticles after 24 h grinding was about 65 nm, about the double of the average crystallite size of 32 nm of porous submicron GDC particles.

ACKNOWLEDGMENTS

The authors appreciate Professor T. Nagaoka and Associate Professor H. Shiigi in the Department of Applied Chemistry at Osaka Prefecture University for the use of their TG analyzer and especially thank their student, Mr. M. Takai, for his efforts at measuring TG.

Additional information

Funding

This work was financially supported in part by The Iwatani Naoji Foundation's Research Grant and a grant-in-aid for Development Science Research (No. 22360331) from the Ministry of Education, Culture and Science of Japan.