Preparation of zirconia thin films by tape casting ...nopr.niscair.res.in/bitstream/123456789/19724/1/IJEMS 9(1) 58-64.pdf · Indian Journal of Engi neering & Materials Scicnccs Vol.

Indian Journal of Engi neering & Materials Scicnccs Vol. 9, February 2002, pp. 58-64

Preparation of zirconia thin fil ms by tape casting technique as electrolyte material for solid oxide fuel cells

A Samson Nesaraj , I Aru l Raj & R Paltabiraman

Central Electrochemi cal Research Institute, Karaikudi 630006, Ind ia

Receil'ed 30 Janllary 200 I ; accepted 24 Novelllber 200 I

The effec t of solvents e' hyl methyl ketone (MEK), toluene, xy lene and etha nol on the properties of thin fi lms of zirconi a prepared by the tape casting technique is reported here. It has been fou nd thaI preparation of the slurry for tape casting pl ays an important role in controlling the tex tural properti es of the membranes. The rheolog ica l fea tures of the CJst tapes are sati sfactoril y related to the solvent rati o. The toluene + xy lene so lvent system has been fo und to provide fine casting and sintering. The physical characteri stics of the green tapes like thick ness, density and weight losse during sintering have also been measured. The rate of heating is determined by TG A an alysis of the green tapes. The ro le of solvents in controlling the properties of tape cast zirconia membranes is also descri bed.

Thin cerami c membranes with hi gh degree of reli ability in their properties at a low cost find application in the fields of electronics and chemical processing technology . The most widely used techniques for the fabrication of such products are hot pressing, hot rolling and tape casting. Tape casting technique is a well-established process technology employed for the fabrication of very thin , uniform and large area membranes with multiple layered structure l

-10

• Recently , there is a growing interest to adopt this technique to produce the electrolyte" -' 5,

cathode ' 6 and interconnector ' 7- ' 9 components for planar Solid Oxide Fuel Cell (SOFC) syste.m. Tape casting process technology involves the di spersion of the inorgani c ceramic particles in a liqui d medium (typi call y non-aqueous), followed by the addit ion of organ ic binders, pl asticizers and deflocculating or defoaming agents to form a slip which is applied onto a glass plate using a doctor blade. The ease of cas ti ng to obtain the des ired thi ckness, the drying and the mechan ical properties of as cast thin sections of the ceramic are strongly dependent on the slip compos ition. The effect of di fferent organic mi xed solvents on the properties of zirconia cast wares prepared by the tape casting tech nique and their characteristics after subjecting them to sin tering are discussed here.

Experimental Materials

The zirconia powder (Zr02, Loba Chemie, India) was sieved through a standard 200 mesh si eve. The

binder PVB (Wilson , India), plasticizers PEG (Otto Chemie, India), BBP (Amrut, India), defoamer silicone oil (Hind High.Vacuum, India) , di spersing agent GTO (Amrut, India) and homogenizing agent cyclohexanone (Burgoyne, India) were used. The HPLC reagent grade ethanol, MEK, xylene and toluene organic solvents were employed.

Measurements The average particle size di stribution of zirconia

powder was obtained from Malvern Easy Particle Size Analyser (Model 3.0) using sodium pyrophosphate buffer and water as the medium. The tape casting process in volves three steps namely, mi ll ing, casting and dryi ng2o-n. Milling in vol ves preparation of castable slurry of the ceramic materi al with binder, plas ti cizer and other functi onal ingredi en ts in suitable proporti on in a solvent system. Known quant ity of fi ne zirconia powder was mixed with polyvinyl butyral (PVB) binder, benzy l butyl phthalate (BBP)/polyethylene glycol (PEG) pl asticizers, si licone oil defoamer, glyceryl trioleate (GTO) dispers ing agent and cyclohexanone homogeni zing agent. Ethanol + MEK (1 :2 v/v), toluene + MEK (1:2 and I: 1 v/v) and to luene + xylene ( I :2 , I: 1 and 2: I v/v) solvent systems were employed. Milli ng was carried out in a ball mill for IO h.

After milling, the viscosity of the slu rry was modified by slow evaporation of the excess so lvent in ai r with uniform stirring before casting in order to produce a good quality tape. A doctor blade

NESARAJ el al.: PEPARATION OF ZIRCON IA THIN FILMS BY TAPE CASTING TECH IQUE 59

asse mbly~.1 was used to obtain the green tapes wi th controlled thickness. T he slurry was poured into the cavity of the doctor blade assembly. As it was moved over the glass plate at a L1ni form speed, the slurry was cas t uniformly on the g lass plate in the form of a thi n sheet to a controlled thick ness with the knife-edge of the doctor blade. T he tape was dri ed slowly and then the cured tape was flex ible in nature . T he su rface morphology of the green zirconia tapes was observed wi th a S-3000 H, Hi tac.'1i Scanning Electron Microscope at an acceleration vol tage of 10 kY and at a magn ification factor of 1000. The e ffect of the solvent system on the properties of the green tape was investigated as a function of thi ckness, density of the tapes wh ile at the curing stage and at the storage. The TGA and DT A curves were obtai ned in air using PL Therma l Sciences Model ST A 1500 T hermal Analysis System to determ ine the weight loss that occurred during heat treatment and the strategy for removal of the binder and other functi onal ingredients24

.

Results and Discussion Powder characteristics

The particle size distribution of the zirconia powder employed for the present investigation is shown in Fig.! , showing thereby, that the majority of particles (77 %) in the bulk powder are about 6 .0 11 m in di ameter.

Tape casting process

The flow chart depicting the various operations involved in the tape casting process is shown in Fig.2. Various ingredients employed for the tape casting process along with the different experimental

100

90

80

70

~ " c: 60

'" .0

.!: 50 Q) CI

'" 40 C 1l Q; 30 a.

20

10

l 11,·_,·",_n.,

I \ I \

I \ i \ r \

\ • 3 4 5 6 7 8 9 10 11 12 13 14

Particle size (urn)

Fig. I- Particle size di stri bution of zirconi a powder

compositions inve tigated are given in Tab le I. Fro m the experimental results obtained some genera l gu ideli nes to be fo llowed while casti ng thin zi rconi a fil ms are: (1) the ratio between orga l1Ic compounds

Binder + Plasticizer +

Solvent

Milling 4 hou rs ~

Ceramic Powder + Solvent

Milli ng 4 hours ~

Defoamer + Homogeneizer

Milling 2 hours ~r

Degassing

Casting

Drying

Stripping

~

Organic Ingredients

removal

~

Sintering

Fig. 2- -Flow chart fo r the zirconi a tape casting process

60 INDI A J ENG. MATER. SCI.. FEBRUARY 2002

and zirconia powder must be as low as poss ible: (2) lhe amoun t of organic solvent must be as minimum as possible to have a homogrneous slurry; (3) the amount or dispersant must be as minimum as necessary to ensure the stabil ity or the sl urry: (4) th e plasti cizer to bi nder ratio must be adjusted LO make the tape flex ible and easy to release and (5) removal l)r the liq ui d solvents whil e mainr<'lInlng a homogeneous, crack-free and dense ceramic film :: orten the biggest challenge of tape casti ng.

The role of organic ingredients

Six rati os of organic solvents used for the preparation of zirconi a tapes were: ethanol + M EK ( 1:2 v/v); toluene + MEK ( I: 1 and 1:2 v/v); and xy lene + tol uene (I: I , 1:2 and 2: I v/v). The solvents dissolve the organic materials and di st ribute them uniform ly in the slurry. They are the vehicles that carry the zircon ia particles in di spersion un til it evaporates and leaves a dense tape on the carrier. To realize effective parti cle packing, the zirconia powder must have very low particle size.

The binder provides strength to the green tapes after the evaporati on of the solvent by bridging the zirconia particles together. PVS25 was used as binder in this work, the amount of which was determined by experiments. The PVS binder was selected because of its excellent green tape functi onal characteri stics, the mai n reason why it is the most commonly used binder

fo r the tape casting of ceramic powders. When th ere was not enough bi nder. the resulting green tape had mi nute cracks on the surface. When the amount of binder was too high, on the other hand, the tapes contained many voids. However, the mechanical strength of the green tapes increased With increase In lhe binder content. Generally the green dens ity values decreased wi th increase in the binder con tent.

The plast icizers modified the stren",th, flex ibi lity and viscos ity of the slip. These were the additi ves that soften the binder in the drying stage. In thi s work PEG and SSP:'!5 were L1 sed as plast icizers . These plasticizers broke the close alignment and binding of the binder molecules thereby increasing the flexibility and workability of the tape. The dispersant coated the ceramic particles and kept them in a stable suspension in the slurry due to electros tatic repulsion . GT026 was used as di spersant in thi s work. Silicone o il :.!7 was added to zi rconia slurry in order to improve the wett ing characte ri sti cs . Cyclohexanone28 was added to the slurry before casting in order to improve the homogeneity of the slurry .

Solvent effects on physical properties

Initi ally, eth anol + MEK solvent system was employed for the preparation of zirconi a slurry . After the tape casting is done with this slurry, the tape results in sk in fo rmation on the top of the surface. This problem is originated because of the quick

Table I- The experimental co mpositions used for tape casting

Exp ZI0 2 PVB PEG BBP GTO Silicone No powder (g) liquid oil

(g) (mL) (mL) (mL) (drops)

30 2.1 2.2 1.8

2 30 5.0 2.2 1.8 2

3 30 7.0 2.2 1.8 2

4 30 8.0 3.0 2.5 2

5 30 7.0 2.2 1.8 2

6 30 7.0 2.2 1.8 2

7 30 7.0 2.~ 1.8 2

g 30 7.0 2.2 1.8 2

9 30 7.0 2.2 1.8 2

Cyclohe Solvent -xanone stem (mL) (mL)

MEK-16 EtOH-8 MEK- 16 EtOH-8 MEK-22 EtOH-II MEK-20 EtOH-IO MEK-20 Toluene-IO Tol uene-20 Xylene-IO Toluene-14.5 Xylene- 14.5 MEK -15 Toluene-IS Toluene- I 0 Xylene-20

Green Tape (thickness mm)

0.55

1.0 I

0.51

0.64

0.46

0.49

0.45

0.86

0.92

Tape qual ity

very less binder less binder cracks seen sticks on glass tape quality improved good

velY good

pinholes appeared good but. rough surface appearance

NESARAJ el af. : PEPARATION OF ZIRCONIA THIN FILMS BY TAPE CASTING TEC HNIQU E 61

evaporation of the solvent from the cast-ware during drying. A solvent system with high boiling point has been chosen. Instead of ethanol, toluene was taken as the solvent keeping the other ingredients as such. The slurry prepared with toluene + MEK was very thin and the cast ware obtained from thi s exhibited improved surface appearance. However, the skin formation on the top of the surface due to the evaporation of the solvent (MEK) was also observed. Therefore, experiment with xylene + toluene (1:2 and 2: I v/v) solvents mixture keeping the other ingredients as such was carried out. The tapes prepared from these combinations have been found to ex hibit improved functional characteristics with modified surface appearance. From the result, we suggest that the use of a solvent system with hi gh boiling point may yield a good ceramic tape. Removal of the liquid solvents while maintalI11ng a homogeneous, crack free, solid tapes often is the biggest challenge of tape casting process. Also, the solvent vapours must be removed uniformly without excessive skinning of the surface.

Finally , these experiments were done with the same solvent mixture but at a different ratio (I: I v/v). This composition yielded tapes with good quality, good fl ex ibility , green density, less shrinkage and uniform surface microstructure. So, the use of 14.5 + 14.5 ml of toluene + xylene solvent ratio may be preferred for

the tape casting of zirconia powders with the same composition of other ingredients. The appearance of in-homogenity, and cracks on the surface of the tapes were found to be very sensitive to the rati o of the solvent components.

The crucial parameters controlling the quality of the slurry and the tape were the following ratios: ceramic/binder, ceramic/total liquid, binder/total solid, binder/plasticizer, binder/(binder + plasticizer), binder/total liquid, plasticizer/total solid , plasticizer/total liquid, dispersant/total solid and (binder + plasticizer)/total solid . These ratios were found to influence the microstructure of the cast wares. The typical values employed in thi s work are presented in Table 2.

Characteristics of the green tapes The weight losses when the ceramic tapes were

subjected to annealing for 2 h at different temperatures are shown in Table 3. The green density of the as cast ceramic tapes and the density values as a function of annealing temperature for 2 h duration are presented in Table 4. The weights and the density values were found to decrease as the annealing temperature was increased. These experiments were carried out to understand the gradual escape of all the organic functional ingredients from the tapes leav ing a crack free surface microstructure. The thickness of

Table 2. - Typical ratio values used for the zirconi a tape casting slip formulation

Compositi on Rat io 2 3 4 5 6 7 8 9

Ceramic/ 14.28 6.00 4.285 3.75 4.285 4.285 4.285 4.285 4.285 Binder Cerami c/ 1.00 1.00 0.76 0.80 0.83 0.83 0.80 0.83 0.8:\ Total Liquid Binder/ 0.065 0.14 0.18 0.21 0.1 8 0. 18 0.1 8 0. 18 0. 18 Total Solid Binder/ 0.525 1.25 1.75 1.454 1.75 1.75 1.75 1.75 1.75 Plasticizer Binder/ 0.34 0.55 0.63 0.59 0.63 063 0.63 0.63 0.63 (Binder + Plasti cizer) Binder/ 0.07 0.16 0.17 0.2 1 0.19 0. 19 0.20 0. 19 0.19 Total Liquid Plastici zer/ 0. 12 0. 11 0.10 0. 14 0.10 0. 10 0. 10 0. 10 0.10 Total Solid Plast icizer/ 0.13 0. 13 0.10 0. 14 0.11 0. 11 0. 11 0.11 0.1 1 Total Liquid Dispersant/ 0.03 0.Q28 0.027 0.026 0.027 0.027 0.027 0.027 0.027 Total Solid (l3 inder + 0.19 0.25 0.29 0.35 0.29 0.29 0.29 0.29 0.29 Plasti cize r) / Tutal So lid

62 INDIAN J ENG. MATER. SCI., FEBRUARY 2002

Table 3 - The weight loss observed on zirconia tapes (area: 3 x 3 cm2) at different annealing temperatures

Tape No.

2 3 4 5 6 7 8 9

Initial Weight

(g)

2.13 1.05 1.60 1.01 0.45 0.91 1.83 1.78

Weight at 473 K (g)

1.97 1.00 1.55 0.98 0.44 0.84 1.52 1.52

Weight at 673 K (g)

1.47 0.75 1.17 0.78 0.36 0.67 1.26 1.20

Weight at 873 K (g)

1.45 0.60 1.08 0.71 0.32 0.56 1.22 1.20

(0/0 ) Weight loss

at 873 K

31.92 42.86 32.50 29.70 28.88 35.16 33.33 32.58

Table 4- Densi ty data obtained on tape cast zirconi a tape (area 3 x 3 cm2)

Tape Nos.

2 3 4 5 6 7 8 9

Green Densi ty (g/mL)

2.3432 2.8888 2.3088 2.5505 2.4225 2.4737 2.3643 2.14n

*Densityat 473 K (g/m L)

2.1672 2.3641 2.2366 2.4747 2.2407 2.2222 2.0105 1.8559

*Densityat 673 K (g/mL)

1.6171 1.8115 1.3978 1.9696 1.5555 1.6487 1.6867 1.5151

*Densi tyat 873 K (g/mL)

1.6110 1.4492 1.3043 1.7927 1.6707 1.4590 1.6735 1.5503

Shrinkage (0/0 ) in Density at

873 K

3l.25 49 . .83 43.5 1 29.71 31.03 41.02 29.21 27 .88

*Density of the tapes at different temperatures when the tapes were subjected to annealing for 2 h each

the dried green tape was found to depend upon the slip viscosity, casting flow rate, blade gap and reservoir height behind the doctor's blade).

The binder removal was determined by weight loss measurements with time on programmed annealing.

The DT AfTGA curves obtained in air at a heating rate of 10 Klmin . on the zirconia tapes prepared with

different solvents and compositions are shown 10

Fig. 3. The SEM pictures obtained on the surfaces of the green zirconia tapes are presented in Figs 4 & 5 for compositions 4 & 7 (Table 1). The microstructural features of the tape fabricated with composition 4

reveal irregul arity in the ceramic particles and the bi nder distribution owing to the quick evaporation of the ethanollMEK solvent whereas the distribution of

ceram ic particles and binder was uniform and regular in tape fabricated with composition 7 . It becomes ev ident from the DT NTGA curves that below 623 K, the evaporati on of the sol vent system gets completed . The loss in weight has been recorded as 5 % per min . A furt her weight loss is noticed up to 673 K. Decomposition of binder and pl asticizer gets completed below 773 K. The initial shrinkage IS

due to the removal of the orgamc solvent. Further

11 0

1 00 T.~G,-,-A __ _

~ 90 <; "- 80

70 DTA

273 373 473

100 TGA

~ 90

& 80

70 / DTA

273 373 473

100 TGA

C 90

'" i::!

~ 80

70 DTA

573 673 773

Oeg K

573 673 773

Oeg K

40

30

~ 20 ~

10 g :::;:

o

8 73 9 73

b 50

40

30 ~ ~

20 g 10 j'

- 0

873 973

40

30 V>

20 g g

10 j'

o ----'--

273 373 4 73 573 673

Oeg K

773 8 73 973

Fig. 3-TGA/DT A curves of the zirconia tapes: (a) Composit ion 4: (b) Composi tion 5: and (c) Compos ition 7 of Table 1

NESARAJ el al.: PEPARATION OF ZIRCONIA THIN FILMS BY TAPE CASTING TECHNIQUE 63

Fig. 4-The SEM picture of the surface of green zirconia tape for composition 4

Fig . . 5-The SEM picture of the surface of green zirconia tape for composition 7

64 INDIAN J ENG. MATER. SCI., FEBRUARY 2002

shrinkage is dependent both upon the ratio of binder/plasticizer to solid.

Further, these typical DT A/TGA patterns reveal the si ntering conditions to be adopted for the zirconia tapes. The optimum values for pre-sintering and sintering of the zirconia ceramic tapes are alTived at from the DT A traces as: (I) heating from room temperature to 573 K at a rate of I Klmin in air; (2) soak ing at 573 K for 3 h (solvent removal) ; (3) heating from 573 K to 773 K at I Klmin in air; (4) soak ing at 773 K for 3 h (binder removal); (5) heating from 773 K to the maximum sintering temperature 1773 K at J Kimin. in ai r; (6) soaking at 1773 K for an hour fo r complete sintering and (7) cooling to room temperature at a rate of 1 Klmin in air.

Conclusions Fabrication of thin and flexibl e zirconia ceramic

tapes from fine powder by tape casting technique for app li cation in SOFC system is dealt with. Among the three solvent systems in ves tigated, toluene + xylene ( I: I v/v) gives tapes with excellent microstructure. The organic solvent with hi gh boiling point is found to be preferable. Slow evaporation of the so lvents en hances the surface homogeneity of the green tape with desired fun ct ional characteristics. Acknowledgements

Sincere thanks are due to CSIR for the award of Senior Research Fellowship to Mr A S Nesaraj . The authors are thankful to Dr V Sundaram, Dr R Srinivasan and Mr S Ramu, CECRI, for their help in the experimental measurements. The authors are very thankful to Dr M Raghavan, Director, CECRI, Karaik udi, Indi a, for his constant encouragement dur ing the in ves tigations and kind permission to publish thi s work. Referen('es I Sa lam Loey A. Mat hews Richard D & Robertson Hugh.

1I1(/ Ier Chelll Phys . 62 (2000) 263. ,., G ree nwood R. Roncari E & Galassi C, J Ellr Cer([lll Soc, 17

( 1997) 1.103 .

3 Nesaraj Samson A. Pores Joseph S. Prabhu G. Anitha K L. Solaiyan C, Dheenadayalan S, Arul Raj I, Chandrasekaran R & Pattabiraman R , Bull Eleclrochelll , 14 ( 1998) 354.

4 Hyatt Theodore p , CerQ/n Eng Sci Proc, 16 ( 1995) 71. 5 Mi stler Richard E, Ceralll Bull. 69 ( 1990) 1022. 6 Keller Joseph M, Ulbrich Russell R & Haber Richard A. Alii

Ceram Soc Bull, 76 (1997 ) 87. 7 Hyatt Edmond P, Ceram Bull, 68 ( 1989) 869. 8 Landham Rowena R, Nahass Paul , Leu ng Dominic K.

Ungureit Michael, Rhine Wendall E, Bowen Kent H & Calvert Paul D. Am Ceram Soc BIIlI, 66 ( 1987 ) 15 13.

9 Young Albert C, Lin Jane-Chy i, Yeh Tsung-Shou & Che ng Chian- Lii, Maler Chelll Phys. 34 ( 1993) 147 .

10 Forte Susan, Morris John R Jr. & Can non Roger W. Alii Ceralll Soc BIIII, 64 ( 1985) 724.

II Montross C S, Hasse l Van B A, Kawada T. Ynkokawa H & Dokiya M, J Maler Sci, 30 (1995) 3285.

12 Plucknett K P. Cacers C H & Wilkin son D. J Alii Cerulli Soc. 77 (1997) 2137 .

13 Philips S V & Dutta A K. Ceralll Ell er Applic(lI, Sche fi c ld. 3B (1990) 47.

14 Tintinelli A, Rizzo C, Loreti S & Selvaggi A. Proc Secolld fill SYIIIP Solid Oxide Fllel Cells. Greece, edited by Grosz F. Zegers P, Singhal S C & Yamamoto 0 , (Commission of the European Com muniti es, Luxemburg). 1991. 747.

15 Chartier T & Rouxe l, J Eur Ceram Soc, 17 ( 1997) 299 . 16 Kumar D, Upadhyay S & Prakash 0. Maler Lell. 32 ( 1997)

151. 17 Murphy M W, Armstroy T R & Smith P A. J Alii Ceralll Soc.

80 ( 1997) 165. 18 Tai L W & Lessing P A, J Alii Ceralll Soc. 74 ( 199 1) 155. 19 Lessing P A. Tai L W & Kle mn K A. Pmc Firsl f ill Snllp

Solid Oxide Fllel Cells. ed ited by S inghal S C (Electrochemica l Soc., Inc. , NJ). 1989. 337.

20 Chartier T & Rouxel T , J Eur Ceralll Soc, 17 ([ 997) 299. 2 1 Corso G. Lo Casto S, Lombardo A & Fre ni S. Mater Chelll

Phys, 56 ( 1998) 125. 22 Ruys Andrew J & Sorre l Charles C , Ceralll BIIII , 69 ( 1990)

828. 23 Will iams J C, in Treatise 0 11 Materials Sciellce alld

Techllology, Vo1.9, ed ited by Wang F Y, (Academi c Press. New York), 1976, 173 .

24 Salam Loey A, Mathews Richarcl D & Robertson Hugh. J Ellr Ceralll Soc. 20 (2000) 1375.

25 Moreno Rodri go. Alii Ceralll Soc BIIII. 71 ( 1992) 1647. 26 Calvert Paul D, Tormey Ellen S & Pober Richard L. Alii

Ceralll Soc BIIII, 65 (1986) 669 . 27 HOlza D & Grei l P, M([{er Sci Ellg . A 202 ( 1995) 206.

28 Brawn Linda, Morri s John R Jr. & Cannon Roger. Alii Cemlll

Soc BIIII, 64 ( 1985) 727.