J O U R N A L O F M AT E R I A L S S C I E N C E : M AT E R I A L S I N E L E C T RO N I C S 1 4 ( 2 0 0 3 ) 1 6 5 ± 1 6 9

Barium ion leaching from barium titanate powder in water DANG-HYOK YOON, BURTRAND I. LEE, PRERAK BADHEKA, XINYU WANG School of Materials Science and Engineering, Olin Hall, 340971, Clemson University, Clemson, SC 29634-0971, USA E-mail: [email protected] A submicrometer-sized commercial BaTiO3 (BT) powder was aged in water under three different conditions: pH, aging times, and pre-heat treatments of the powder. The amount of Ba2 ‡ ions leaching from the BT particles was determined by the EDTA titration method. As predicted by thermodynamic calculation, the greater extent and the faster rate of Ba2 ‡ leaching were found at the lower solution pH, leveling off at pH 8. The pre-heat treatments of the BT powder increased the amount of Ba2 ‡ leaching when compared to the as-received one. This result was shown by the formation of soluble surface BaCO3, which was detected using FT-IR spectroscopy. It was also shown that organic passivation agents were effective reducing the Ba2 ‡ leaching but at high solution pH. # 2003 Kluwer Academic Publishers

1. Introduction

One of the more important obstacles for using a waterbased barium titanate slip system in the multi-layer chip capacitor (MLCC) industry is the leaching of Ba2 ‡ in water, resulting in an inconsistent Ba/Ti ratio in the slip system. An inconsistent Ba/Ti ratio in the powder can change the sintering density and the relative dielectric constant of BaTiO3 (BT) [1, 2]. In addition, leached Ba2 ‡ from the BT surface can interact in an adverse manner with various polymeric additives [3±5]. The Ba2 ‡ ions present in the solution can cause the folding of the chains in the polymeric additives caused by a saltingout mechanism. This folding will impair not only the homogeneous dispersion of BT particles but also lead to the creation of defective holes in the ®red tapes as a result of the burnout of the salted-out polymers. In order to use the water-based slip system in the MLCC industry with greater reliability, understanding and controlling the Ba2 ‡ leaching are essential, not only to preserve the chemical homogeneity in the powder compact but also to prevent adverse interactions with such organic species as dispersants and binders. Better known methods for quantitative analysis of Ba2 ‡ in a solution are inductively coupled plasma (ICP) spectroscopy [6] and atomic absorption spectroscopy (AAS) [7, 8]. A third way to determine the amount of leached Ba2 ‡ is titration with ethylenediamine tetraacetic acid (EDTA) [5, 9, 10]. Methyl thymol blue is used as a color indicator for the Ba2 ‡ ion to detect the titration end point. Adding a small amount of this color agent into the Ba2 ‡ solution forms a blue complex with Ba2 ‡ . When all the free Ba2 ‡ ions are chelated by the EDTA titrant, the indicator is displaced from the barium, causing the color to change from blue to gray. In this titration, 1 mol of EDTA corresponds to 1 mol of Ba2 ‡ ion [5]. 0957±4522

# 2003 Kluwer Academic Publishers

2. Experimental

Cabot BT-8 powder (Cabot Performance Materials, Boyertown, PA), synthesized by a hydrothermal technique, was used in this study. The BET speci®c surface area measured using N2 adsorption was 8:50 m2 =g. A light scattering particle size measurement indicated the median particle size of 0.24 mm. The Ba/Ti molar ratio provided by the manufacturer was 0.998. As a ®rst test, to determine the amount of Ba2 ‡ ions leaching from the BT particles, 6 g of BT powder were added to 24 ml of deionized water. This mixture was then stirred mildly by hand with a spatula for 1 min to disperse the particles. NH4 OH and CH3 COOH were used to adjust the solution pH by using an Accumet 925 pH/ion meter (Fisher Scienti®c, Pittsburgh, PA). The solution pH was readjusted after an equilibration period of 5 h, if drift from the initial pH was observed, and was con®rmed before titration. After being aged for the desired length of time in an air atmosphere, the slurry of BT was centrifuged at room temperature at 3500 rpm for 30 min. A 12 ml aliquote of the supernatant was diluted to 50 ml with deionized water, and the pH was adjusted to 12 by a 10 M KOH solution. The solution was titrated with a 0.001 M EDTA solution using a Mettler titration meter (DL50 Titrator, Mettler Toledo, Switzerland). As a second test, to check the reliability of this method in measuring the amount of Ba2 ‡ , Ba-acetate was dissolved in water containing a known amount of Ba2 ‡ , followed by the EDTA titration. The effect of the addition of any organic material in this blank test was checked using citric acid, oxalic acid and known dispersants by adding 0.5 wt % per BT powder. The used organic materials are listed in Table I. A third test investigating the effect of BT surface cleaning using heat treatment on the amount of Ba2 ‡ 165

T A B L E I Organic materials used for Ba2‡ leaching experiment Organic materials KD-6 (acrylic graft copolymer ‡ polyethylene glycole) Tween 80 ( polyoxyethylene(20) sorbitan monooleate) Darvan C (ammonium polymethacrylate) APA (ammonium polyacrylate) PAsA±Na ( polyaspartic acid±sodium salt) Oxalic acid ((COOH)2 ? 2H2 O) Citric acid (HOCCOOH(CH2 COOH)2 ? 2H2 O)

leaching was conducted. After treating the BT powder with the temperatures of 400 , 600 and 1000  C for 2 h in air to remove surface adsorbed species, Fourier transform infrared (FT-IR, Magna FT-IR spectrometer, Model 560, Nicolet Instrument) spectra of each sample were immediately taken to prevent re-adsorption of gaseous contaminants on the BT surface. Spectral scans were made in air using a diffuse re¯ectance stage with a resolution of 4 cm 1 after taking a background calibration to minimize the noise effect. As the fourth test, some known organic materials were added to test the ability of passivation on Ba2 ‡ ion leaching from the BT surface at pH ˆ 8 and 12.

3. Results and discussions 3.1. Blank test

Fig. 1 shows the results of the blank test with the different organic species. The y-axis represents the percent difference between the actual amount of Ba2 ‡ and the titrated value. As the data show, the titrated values of the six samples agree with the actual amount of Ba2 ‡ ; only Tween 80 and oxalic acid show a difference. With these two, precipitates were observed during the titration process. It is well-known that alkaline earth oxalates are quite insoluble and hamper any volumetric determination with EDTA [9]. For the same reason, Tween 80 cannot be used with the EDTA titration method to determine the amount of Ba2 ‡ . As a result, we can conclude that the EDTA titration method can be used to detect the amount of dissolved Ba2 ‡ ions accurately unless a precipitate is formed.

Figure 1 Results of the blank test with Ba-acetate solution and with a different organic material in water.

166

3.2. Effect of solution pH on Ba2 ‡ leaching

Fig. 2 shows the amount of Ba2 ‡ leaching and the resultant Ba/Ti molar ratios of the pure BT powder calculated after being aged for 1 and 2 days in various pH solutions. The amount of Ba2 ‡ leaching shows a signi®cant dependency on the solution pH, i.e., the higher leaching amount and the faster leaching rate at the lower pH. Even though the system is not in equilibrium condition, this pH-dependent Ba2 ‡ leaching behavior shows the similar trend of the equilibrium condition which was reported by Lencka and Riman [11]. As explained by theoretical model [11], it was found that there was a linear log relationship of [Ba2 ‡ ] with pH approximately. However, the situation is complicated in normal environments because of the general presence of carbon dioxide in water. The reaction of CO2 dissolved in water with BT leads BaCO3 formation [11]. This BaCO3 increases the Ba2 ‡ concentration in the solution by dissolving according to the solubility product of BaCO3 . As a result, one can expect an appreciable Ba2 ‡ leaching in water unless the pCO2 is extremely low. In contrast, TiO2 (s) is generally regarded as insoluble in water [12, 13]. Therefore it is likely that Ba2 ‡ will release from the BT surface until the surface becomes a ``TiO2 -like'' character. Fig. 3 shows the Ba2 ‡ leaching behavior at pH ˆ 8 and at room temperature as a function of aging time since the usual industrial practice is in pH of 8 to 10. The amount of Ba2 ‡ leaching for the ®rst 30 min could not be measured due to the time needed for the sample preparation. As shown in Fig. 3, the amount of Ba2 ‡ leaching is quite high for the ®rst measurement at 30 min of aging. After 30 min, the leaching rate gradually decreases reaching a plateau at a time greater than 100 h. The results on the time-dependent behavior of Ba2 ‡ leaching in water reported by others [7, 8, 14] vary markedly. Neubrand et al. [7] observed an instantaneous increase of Ba2 ‡ leaching in water within the ®rst few minutes and then a gradual decrease at a pH of approximately 10. They explained this ®nding as a readsorption of Ba2 ‡ ions onto the hydrated BT surface. On the other hand, Blanco-Lopez et al. [8] observed a gradual increase in the amount of Ba2 ‡ leaching in water at pH of 4 within 5 h of aging, after an instantaneous increase at the beginning which agrees with our results.

Figure 2 The amount of Ba2‡ leaching and the resultant Ba/Ti molar ratio as a function of the solution pH without organic material.

Figure 3 The amount of Ba2‡ leaching versus time at pH ˆ 8 and at room temperature.

Figure 5 Speci®c surface area and median particle size changes of BT powder with pre-heat treatment temperature.

Utech [14] explained that Ba2 ‡ leaching in water at pH 8, 10 and 12 occurred instantaneously, reaching a steadystate within 5 min, i.e., no Ba2 ‡ concentration change after 5 min. Regarding the amount of Ba2 ‡ leaching, the average leached Ba2 ‡ concentration in our experiment lies between the results of Neubrand et al. [7] and Utech [14]. The variations observed here might come from the uncontrolled CO2 concentration in the solution.

approximately 3383 cm 1 are attributed to surface adsorbed water, of 1426 and 1469 cm 1 to two different carbonates: the former to lattice carbonate and the latter to surface carbonate, and of 785 and 475 cm 1 to BT. Small peaks at 1750 and 2450 cm 1 are also assigned to carbonate species [16]. Semi-quantitative comparison of the species by band intensity was calculated using the ratios of the band intensity caused by the species divided by the band intensity of BT at 785 cm 1 . The increasing OH band height, peaking at 600  C, can be explained by the diffusion of the hydroxyl ion (OH ) to the BT surface which had been incorporated into the lattice during the hydrothermal synthesis of this powder. According to Hennings et al. [17], there is an appreciable amount of lattice-incorporated protons and hydroxyl ions ( & 0.4 mol/mol BT) present in hydrothermally synthesized BT powder. As they found, this OH diffuses to the BT surface by annealing up to 800  C [17]. For the carbonate peaks shown in Fig. 6, the band height ratios at 1426 and 1469 cm 1 were calculated using the baseline method suggested by Brezinski [18]. These carbonate band height ratios were obtained after normalizing the carbonate band heights by that of BT at 785 cm 1 . Both of the lattice and surface carbonate band height ratios at 1426 and 1469 cm 1 increased, peaking at 600  C, and then decreased at 1000  C as shown in Fig. 7(a). This means that the carbonates at the surface decomposed at temperature above 600  C. Both of the small carbonate peaks at 1750 and 2450 cm 1 also agree with the 1426 and 1469 cm 1 carbonates. These results agree with the Ba2 ‡ leaching behavior shown in Fig. 4. This behavior is further demonstrated by the linear relationship between the carbonate band height ratios at 1469 cm 1 and the amount of Ba2 ‡ leaching shown in Fig. 7(b). This ®nding indicates that Ba2 ‡ leaching depends more on the amount of surface carbonate than on the diffusion of the lattice Ba2 ‡ for the pre-heat treated BT powders. As the temperature increases up to 600  C, the BT surface in contact with the atmospheric CO2 forms carbonates which are fairly soluble in water. Evidently, this form of carbonate has no passivating capability for Ba2 ‡ leaching. The effects of organic species on Ba2 ‡ leaching are presented in Fig. 8 for pH ˆ 8 (a) and pH ˆ 12 (b) after 2 days of leaching. In general, the amount of Ba2 ‡

3.3. Surface cleaning and organic species effects on Ba2 ‡ leaching

Pre-heat treatment of BT in air increased the amount of Ba2 ‡ leaching as shown in Fig. 4 with the highest amount of Ba2 ‡ leaching at 600  C and then a decrease at 1000  C. This result cannot be explained by the speci®c surface area change which is shown in Fig. 5. By increasing pre-heat treatment temperature, speci®c surface area is decreased, and particle size is increased. To check the next possible variable, the effect of surface BaCO3 on Ba2 ‡ leaching, FT-IR spectra of the preheated BT were used. The FT-IR spectra of these powders measured within 10 min of the heat treatment shown in Fig. 6 indicate band height differences at approximately 3400 and 1200±1500 cm 1 . According to previous research [15, 16], the broad IR bands of

Figure 4 Effect of heat treatment of BaTiO3 powder on Ba2‡ leaching rate at room temperature.

167

Figure 6 FT-IR results of BT-8 after calcining at various temperatures and enlarged spectra between 1000 and 1800 cm

Figure 7 (a) Normalized carbonate band height ratios at 1426 and 1469 cm 1 after heat-treatment of BT at different temperature in air for 2 h and (b) the linear relationship between the normalized carbonate band height ratios and the amount of Ba2‡ leaching after 48 h aging.

leaching with organic agent increased with organic increase and exceeded that of the sample without the organic at pH ˆ 8. At pH ˆ 12, on the other hand, the amount of Ba2 ‡ leaching gradually decreased with organic increase and showed 2±4 times less amount than the samples without organic. Even though it is known that the polymeric molecular conformation and ionic exchange of polymeric adsorbate are affected by the 168

1

.

Figure 8 Organic material effects on Ba2 ‡ leaching at (a) pH ˆ 8 and (b) pH ˆ 12 after 2 days leaching.

solution pH [19, 20], it cannot explain this different passivation behavior at different pH. To understand the exact mechanism of this passivation, more detailed study is needed.

4. Conclusions

After using the EDTA titration method to measure the amount and rate of Ba2 ‡ leaching from commercial BT

powder under various conditions, the following trends were observed: (1) EDTA titration is an effective measurement method for the amount of Ba2 ‡ unless a precipitate is formed as the result of the reaction of Ba2 ‡ with an added organic species. (2) The amount and rate of Ba2 ‡ leaching are signi®cantly affected by solution pH; i.e., the higher leaching amount and the faster leaching rate occur at the lower solution pH, a result which can be predicted by thermodynamic calculation. (3) Pre-heat treatment of BT powder in air enhances the Ba2 ‡ leaching rate due to the formation of the easily soluble carbonate which has no passivating capability for Ba2 ‡ leaching. (4) Organic species as a Ba2 ‡ leaching passivation agent were effective at pH ˆ 12 but not at pH ˆ 8.

5. 6. 7. 8. 9. 10. 11. 12.

13.

Acknowledgments

We would like to thank Dr S. Costantino of Cabot Performance Materials for supplying the BT-8 powder. This research is based upon work supported by the National Science Foundation under Grant No. DMR9731769. Any opinions, ®ndings and conclusions, or recommendations expressed here are those of the authors and do not necessarily re¯ect the views of the National Science Foundation.

14. 15. 16. 17. 18. 19.

References 1. 2. 3.

and C . H U , J. Am. Ceram. Soc. 73(3) (1990) 531. and K . H O N G , ibid. 84(9) (2001) 2001. G . L . R O Y, A . L . L A F E R R I E R E and J . O . E D WA R D S , J. Inorg. Nucl. Chem. 4 (1957) 106. 4. R . E . C H O D E L K A , in ``The Aqueous Processing of Barium T. L I N

20.

Titanate: Passivation, Dispersion, and Binder Formulations for Multilayer Capacitors'', Ph.D. Thesis in Materials Science and Engineering, University of Florida (1996) p. 63. X . WA N G , B . I . L E E and L . M A N N , Colloids Surf. 202 (2002) 71. H . W. N E S B I T T, G . M . B A N C R O F T, W. S . F Y F E , S . N . K A R K H A N I S , A . N I S H I J I M A and S . S H I N , Nature 289 (1981) 358. A . N E U B R A N D , R . L I N D N E R and P. H O F F M A N N , J. Am. Ceram. Soc. 83(4) (2000) 860. M . C . B L A N C O - LO P E Z , B . R A N D and F. L . R I L E Y, J. Eur. Ceram. Soc. 17 (1997) 281. G . D . C H R I S T I A N , in ``Analytical Chemistry'' 3rd edn. (John Wiley, New York, 1980) p. 221. J . B A S S E T, R . C . D E N N E Y, G . H . J E F F E RY and J . M E N D H A M , in ``Vogel's Textbook of Quantitative Inorganic Analysis'' 4th edn. (Longman, New York, 1978) p. 261. M . M . L E N C K A and R . E . R I M A N , Chem. Mater. 5 (1993) 61. K . O S S E O - A S A R E , F. J . A R R I A G A D A and J . H . A D A I R , in ``Ceramic Transactions'', Vol. 1, edited by G. L. Messing, E. K. Fuller Jr. and H. Hausner (American Ceramic Society, Westerville, OH, 1998) p. 47. C . H E R A R D , A . FA I V R E and J . L E M A I T R E , J. Eur. Ceram. Soc. 15 (1995) 135. B . L . U T E C H , in ``The Effect of Solution Chemistry of Barium Titanate Ceramics'', M.S. Thesis in Solid State Science, The Pennsylvania State University (1990) p. 28. S . W. L U , B . I . L E E and L . A . M A N N , Mater. Res. Bull. 35 (2000) 1303. G . B U S CA , V. B U S CA G L I A , M . L E O N I and P. N A N N I , Chem. Mater. 6 (1994) 955. D . F. K . H E N N I N G S , C . M E T Z M AC H E R and B . S . S C H R E I I N E M AC H E R , J. Am. Ceram. Soc. 84(1) (2001) 179. D . R . B R E Z I N S K I , in ``An Infrared Spectroscopy Atlas for the Coatings Industry'' Vol. 1 (Federation of Societies for Coating Technology, Blue Bell, PA, 1991) p. 52. J . H . J E A N and H . R . WA N G , J. Am. Ceram. Soc. 81(6) (1998) 1589. J . H . J E A N and H . R . WA N G , ibid. 83(2) (2000) 277.

J. LEE

Received 20 June in revised 25 October 2002

169