Titanium alloys have been increasingly and widely used

May 2012 Research & Development Characterization of zirconia-based slurries with different binders for titanium investment casting Zhao Ertuan, Kong Fantao, Chen Yanfei, Chen Ruirun and *Chen Yuyong (National Key Laboratory for Precision Hot Processing of Metals, Harbin Institute of Technology, Harbin 150001, China) Abstract: The materials and physical properties of primary slurry are crucial to the surface quality of the finished castings, especially for high reactivity titanium alloys. The aim of this study is to investigate the influence of different binders on the physical properties of primary slurry for titanium alloy investment casting. The zirconia-based slurries with different binders were evaluated by comparing the parameters: viscosity, bulk density, plate weight, suspensibility, gel velocity and strength. The results indicate that a higher viscosity of binder leads to a higher viscosity and suspensibility of slurry with the same powder/binder ratio. The retention rate and thickness of primary layer increase with an increase in the viscosity of the slurry, and a higher retention rate is associated with a thicker primary layer. The gel velocity of the slurry is correlated with the gel velocity of the binder. The green strength and the baked strength of the primary layer are determined by the properties of the binder after gel and by the production of the binder after fired, respectively. Key words: primary slurry; binder; investment casting; titanium alloy CLC numbers: TG 221 Document code: A Article ID: 1672-6421(2012)02-125-06 Titanium alloys have been increasingly and widely used in aviation, aerospace, shipbuilding and other industrial areas because of their exceptional properties, such as the high specific strength and fracture toughness, excellent corrosion resistance and so on [1-2]. Investment casting technique presents advantages in the production of complex structure parts with the key benefits of the higher flexibility in design and lower cost [3]. However, due to the high reactivity of titanium alloys, the reactions between titanium melt and mold materials result in the formation of α-phase [4-5], which will deteriorate the surface and change the mechanical properties of titanium castings. Therefore, conventional refractory materials with a typical composition of Al 2 O 3 -SiO 2 and siliceous binder are not suitable for titanium investment casting [6]. Stable high temperature resistant oxides can serve as primary coating materials for titanium investment casting, for instance, ZrO 2 [7-8], Y 2 O [9-10] 3, Al 2 O [11] 3, MgO [12], and CaO [12]. Zirconia sol is widely used as primary coating binder for titanium investment casting [13-14]. However, the detailed information about the binder for titanium investment casting is very limited due to many reasons. *Chen Yuyong Male, born in 1956, Ph.D, Professor. His research interests mainly focus on hot precision forming of titanium and TiAl alloys. By now, he has published more than 200 papers. E-mail: yychen@hit.edu.cn Received: 2011-10-24; Accepted: 2012-02-09 The mechanical performance of ceramic shell is crucial to the quality of the castings [15-16]. The properties of the ceramic shell are determined by the materials and parameters of slurry. Bundy et al. [17] investigated the characterization of zircon-based slurries. Ferenc et al. [18] evaluated the influence of alumina powder on the rheological properties of zircon/silica slurries for investment casting. Sidhu et al. [19] studied the effect of slurry composition on plate weight in ceramic shell investment casting process. However, little is known about the characteristics of the primary slurry for titanium investment casting. Hence it is significant and necessary to investigate the properties of the primary slurry for producing high quality titanium castings. This paper presents the work to assess the fundamental physical properties of zirconia-based primary slurries with different binders. The experimental results should be helpful in material selection and parameter control during the process of slurry preparation. 1 Experimental procedure 1.1 Materials In the present research, CaO-stabilized ZrO 2 powders (325 mesh) were selected to prepare primary slurry. Zirconium acetate (ZA), ammonium zirconium carbonate (AZC) and silica sol (S) were respectively used as binder. Physical properties of the three binders are given in Table 1. The ZrO 2 powders were mixed with each of the three binders, respectively and stirred completely. Based on the binders, 125

CHINA FOUNDRY Table 1: Physical properties of the binders Binder Zirconium acetate (ZA) 22.0-24.0 4.0 1.34 8.4 Ammonium zirconium carbonate (AZC) ZrO 2 (SiO 2 ) the slurries were named ZA, AZC and S slurry, respectively. Victawet 12 was used as surfactant in primary slurries to lower their surface tension. Octanol was selected as defoaming agent. The concentration of surfactant and defoaming agent was about 0.5wt.% and 0.3wt.%, respectively. 1.2 Particle morphology and size distribution of ZrO 2 powder The particle morphology of ZrO 2 powders was observed under scanning electron microscopy (SEM), and the particle size distribution was analyzed by a laser diffraction method. 1.3 Density The slurry density was determined by measuring mass and volume. A graduated cylinder with 100 ml slurry was weighed using a balance accurate to 0.1 g, and the weight of the cylinder was subtracted to get the weight of the slurry. Therefore the slurry density was determined. 1.4 Viscosity The dynamic viscosity of slurry was measured using a SNB-2 digital viscometer (accuracy: ±2.0%). All slurries were stirred by an electric agitator sufficiently in a stainless steel cup with 90 mm diameter and the viscosity values were measured at a constant speed. 1.5 Plate weight Plate weight is a measurement used to determine the coating thickness and the ability of slurry coverage and adhesion. A glass plate with the dimension of 76 mm 25.5 mm 1 mm was used. The plate was dipped into the slurry for 10 s and drained for 1 min, and then weighed to determine its coated mass. Plate weight was measured with electronic balance having a least count of 0.01 g. Sidhu et al. [19] gave the equations of retention rate (R) and the thickness of primary slurry layer (H): (1) (2) where Wp is the undipped plate weight, Wd is the dipped weight of the plate, S is the surface area of the plate, and D is the density of the slurry. 1.6 Suspension percentage The suspension percentage of the slurry was measured by filling a 100 ml graduated cylinder to the 100 ml mark. After 12 h, the supernatant of binder phase was completely separated from the powder. The suspension percentage was determined by the volume of the deposited powder. (%) ph Density Viscosity (g cm -3 ) (mpa s) 18.0 9.8 1.38 4.3 Silica sol (S) 29.0-31.0 9.5 1.20 4.3 1.7 Gel test Vol.9 No.2 The gel test was used to investigate the gel velocity of different slurries. A drop of slurry and the corresponding binder were placed on the wax plate, respectively and the gel velocity was compared qualitatively after 3 h. 1.8 Strength Three-point bend test was used to examine the properties of different slurries. The test samples with the dimension of 40 mm 20 mm 6 mm were prepared upon a wax pattern (Fig. 1). The test samples for baked strength were fired at 1,000 for 1 h. Samples were loaded in an Instron 5569 tensile testing machine at a constant load rate of 1 mm min -1 until failure. All tests were conducted on five samples and the average values were determined as experimental results. Fig. 1: Wax pattern for test bar 1.9 Other analysis experiments The primary coating surface was examined by scanning electronic microscopy (SEM), and the X-ray diffraction analysis (XRD) was performed to identify the phase composition of binder after baked. 2 Results and discussion 2.1 Particle morphology and size distribution Refractory particle morphology affects the characteristics of slurry. Figure 2 shows the particle morphology and the particle size distribution curves of differentiation (volume) and integration (calculative volume). It can be seen from Fig. 2 (top left corner) that the morphologies of ZrO 2 particles are highly angular, sharp-angled and have large agglomerates. The average size of ZrO 2 particles is 12.9 μm. Meanwhile the Fig. 2: Particle morphology and size distribution of ZrO 2 powder 126

May 2012 particle size distribution curve is not symmetrical, indicating there are more fine fillers or particles in the ZrO 2 powder used in this study. This kind of graded powder is recommended for investment casting, which is crucial to produce stable slurry with good properties. In addition, the properties of binder are also important for getting stable slurry. 2.2 Viscosity of slurry The viscosity of slurry is determined not only by the powder/ binder ratio, but also by the properties of refractory and binder. Viscosity control is crucial to the production of consistent primary and backup coatings [17]. Especially for primary coating, the viscosity is an important measurement of the flow characteristics of the slurry, which determines the surface quality of the mould shell. Figure 3 shows the changes of the viscosity for 3 types of primary slurries with different powder/binder ratios. It is apparent that the viscosity of slurry increases with increasing powder/binder ratio for all three types of slurries. However, with the same powder/binder ratio, the ZA slurry presents the highest viscosity among the evaluated primary slurries. The differences in viscosity resulted from the different densities and viscosities of binders. The density of zirconium acetate (ZA) and ammonium zirconium carbonate (AZC) is 1.34 g cm -3 and 1.38 g cm -3, respectively, which are higher than the 1.20 g cm -3 for silica sol (S). Therefore, the slurry with the same viscosity requires much more volume percent of silica sol than other slurries. The ZA binder has a similar density to the AZC binder. Nevertheless, the ZA slurry shows higher viscosity than AZC slurry. This can be explained by the difference of binder viscosity. It can be seen from Table 1 that ZA binder has the highest viscosity, about two times as that of the other two binders. A higher viscosity of binder leads to a higher shearing force, and represents a higher dynamic viscosity of slurry. The results indicate that the viscosity of the binder has greater influence on the viscosity of the slurry than the density. ρ ZA = 3.08 g cm -3 Research & Development ρ AZC = 2.87 g cm -3 ρ S = 3.13 g cm -3 Fig. 4: Retention rate (R) and thickness of primary layer (H) of different slurries a thicker primary layer. This result suggests the three types of slurries have good paintability. The ZA slurry presents the highest retention rate (0.11 g cm -2 ) and the thickness of the primary layer (0.37 mm) among the three types of slurries, which is attributes to the highest viscosity of slurry. The S slurry has the lowest retention rate (0.03 g cm -2 ) and the thickness of the primary layer (0.09 mm), but has the highest density, indicating the density of the slurry has nothing to do with the viscosity. It can be concluded that the density of slurries with the same powder and different binders is determined by the characteristics of binder, such as the wettability to refractory powder and the molecular size of binder. In order to investigate the influence of the powder/binder ratio on the retention rate and the thickness of primary layer, the ZA slurries with different powder/binder ratios were prepared. It can be seen from Fig. 5 that the increasing retention rate and the thickness of the primary layer are correlated with the increasing powder/binder ratio. From Fig. 3 and Fig. 5, it is obvious that a higher powder/binder ratio can give a higher viscosity, and a higher viscosity forms a thicker primary layer. However, a thin layer or a thick layer tends to lead to defects of ceramic mold during drying process. In order to seek a proper powder/binder ratio for practical production, the ZA slurries with different powder/binder ratios were coated on a plate and then dried for comparison. It can be seen from Fig. 6 that the plate with low powder/binder ratio (2.0:1) shows the peeling defect [Fig. 6(a)] and the plate with high powder/binder ratio (3.0:1) presents big crack [Fig. 6(c)]. While the plate with a powder/binder ratio of Fig. 3: Viscosities of slurries with different refractory powder/binder ratios 2.3 Plate weight The retention rate indicates the ability of slurry coverage and adhesion on the wax pattern. Figure 4 shows the retention rate (R), the thickness of the primary layer (H) and the density for all three types of slurries with the same powder/binder ratio. The density of ZA, AZC and S slurry is 3.08, 2.87 and 3.13 g cm -3, respectively. As shown in Fig. 4, a higher retention rate is associated with Fig. 5: Retention rate (R) and thickness of primary layer (H) of ZA slurry with different powder/binder ratios 127

CHINA FOUNDRY Vol.9 No.2 (a) (b) (c) Fig. 6: ZA slurry plate with different powder/ binder ratios: (a) 2.0:1; (b) 2.5:1; (c) 3.0:1 2.5:1 presents the consistent and unbroken surface [Fig. 6(b)]. The thin primary layer tends to crack or exfoliate because of short drying time. However, the thick primary layer also tends to crack because of the big stress during long drying time. Moreover, the air mixed in the slurry cannot escape completely for high viscosity slurries during the process of stirring. Figure 7 shows the micro-morphology of the primary layer in Fig. 6(b). Obviously, the uniform surface indicates the proper powder/ binder ratio for ZrO 2 powder and ZA binder is about 2.5:1. The proper powder/binder ratio is crucial to the thickness of primary layer. The low powder/binder ratio produces a thin layer. The refractory sand can be pushed into the thin primary Fig. 7: Morphology of primary coating surface with powder/za binder ratio of 2.5:1 layer during mold making, which causes rough surface of ceramic mold. As shown in Fig. 8(a), the coarse surface of mold resulted from thin primary layer (the powder/binder ratio is 2.0:1) can deteriorate the surface of the finished casting. While for a thick primary layer (the powder/binder ratio is 3.0:1), the weak cohesive force between the primary layer and the second layer would cause the defect of flaking. Figure 8(b) shows the surface morphology of the thick primary layer of ceramic mold after baking. It is obvious that the casting with defects will be produced using this kind of mold. In contrast, the primary layer with a proper powder/binder ratio (2.5:1) presents an internal surface without defects, as shown in Fig. 8(c). (a) (b) (c) Flaking of the primary layer Fig. 8: Surface morphology of primary layer with different thicknesses (powder/binder ratio): (a) thin primary layer (2.0:1); (b) thick primary layer (3.0:1), (c) proper layer (2.5:1) 2.4 Suspension percentage The stable and homogeneous slurry is correlated with good suspensibility. It can be drawn from Fig. 9 that the ZA slurry presents the best suspensibility (94%), while the AZC slurry (84%) and the S slurry (81%) has a lower but similar suspensibility. From Table 1 and Fig. 9, it is clear that the suspension property of the slurry is closely related to the viscosity of the binder. The higher viscosity of the binder, the better the suspension, indicating the slurry is easy to control and stir. 2.5 Gel test Six drops of different slurries and corresponding binders were placed on the wax plate for comparing the gel velocity, which is a good indicator to characterize the drying time of the primary layer. As seen in Fig. 10, the S binder has already Fig. 9: Suspension percentage of different slurries gelatinized completely, while the ZA binder is semisolid and AZC binder still keeps in liquid state. The slurries display similar behaviors to the binders. It can be concluded that the 128

May 2012 Research & Development (a) (b) (c) Fig. 10: Gel time of different slurries and binders: (a) ZA slurry; (b) AZC slurry; (c) S slurry gel velocity of the slurry is determined by the gel velocity of the binder. At the same time, the S binder presents the best wettability among the three binders, while AZC binder shows the biggest surface tension. The result of the present experiment is helpful for foundry to determine the drying time of primary layer according to the binder to be selected. 2.6 Strength The green strength and the baked strength of test samples prepared with different slurries are shown in Fig. 11. It is apparent that S slurry exhibits the highest green and baked strength, while the ZA slurry and the AZC slurry possess inferior and similar strength. The green strength is attributed to the properties of binder after gel and the baked strength is determined by the production of binder after fired. As seen in Fig. 12, the XRD results show that the ZA and AZC binders have the same phase compositions (monoclinic ZrO 2 and cubic ZrO 2 ) after baking at 1,000. Obviously, the green strength of the ZA and AZC slurries is determined by the same gel product of ZrO 2 colloid, and the baked strength is determined by the Fig. 11: Green and baked strengths of different slurries Fig. 12: XRD of ZA and AZC binders after fired at 1,000 same fired product of ZrO 2 crystalloid. Therefore, the ZA and AZC slurries exhibit the similar green and baked strength. 2.7 Casting defects caused by primary layer The materials and parameters of primary slurry are crucial to the surface quality of the finished castings. Figure 13 shows (a) (b) Inclusion Buckle (c) (d) Penetration Good surface Fig. 13: Surface quality of casting prepared with different powder/za binder ratios: (a) 3.0:1; (b) 3.0:1; (c) 2.0:1; (d) 2.5:1 129

CHINA FOUNDRY the casting surface with different thicknesses of primary layer (ZA slurry). The thick primary layer with a high powder/ binder ratio (3.0:1) could spall off the mould easily during casting. The small pieces of ceramic shell material trapped in the molten metal produce the inclusion defect after solidification, as shown in Fig. 13(a). Moreover, the thick primary layer also presents the rough casting surface because the flaking of primary layer would lead to the melt touching with the coarse second layer. The thick primary layer tends to crack during drying or baking process, which produces the buckle defect [Fig. 13(b)]. Figure 13(c) shows a rough surface of casting caused by a thin primary layer with a low powder/ binder ratio (2.0:1). This is consistent with the result of Fig. 8(a). Figure 13(d) presents a good surface of titanium alloy blade produced by a proper thickness of primary layer with a powder/binder ratio (2.5:1). Obviously, the result demonstrates that the powder/binder ratio of 2.5:1 is proper and crucial for a titanium casting with high quality. 3 Conclusions Primary slurry is considered as one of the major components for making ceramic mould. The choice of binder and refractory material for primary slurry plays an important role in titanium investment casting. This paper contains the fundamental evaluation of the effect of different binders on the primary slurry parameters. The following conclusions have been drawn: (1) With the same powder/binder ratio, the slurry with zirconium acetate binder presents the highest viscosity, suspensibility and plate weight, which is attributed to the highest viscosity of the binder. (2)The gel velocity of slurry is determined by the gel velocity of binder. The green and baked strengths of the primary layer are determined by the production of binder after gel and after baking, respectively. (3) The proper powder/binder ratio determines the surface quality of the finished castings. The proper powder/binder ratio for ZrO 2 powder and zirconium acetate binder is about 2.5:1 for practical production. References [1] Nan Hai, Liu Changkui, Huang Dong, et al. Precision casting of Ti-15V-3Cr-3Al-3Sn alloy setting. China Foundry, 2008, 5(1): 12-15. [2] Peters M, Kumpfert J, Ward C H, Leyens C. Titanium alloys for aerospace applications. Advanced Engineering Materials, 2003, 5(6): 419-427. Vol.9 No.2 [3] Nastac L, Gungor M N, Ucok I, Klug K L, Tack W T. Advances in investment casting of Ti-6Al-4V alloy: a review. International Journal of Cast Metals Research, 2006, 19(2): 73-93. [4] Liu Aihui, Li Bangsheng, Nan Hai, et al. Interaction between gamma-tial alloy and zirconia. China Foundry, 2008, 5(1): 44-46. [5] Liu Aihui, Li Bangsheng, Sui Yanwei, Guo Jingjie. Numerical simulation of interfacial reaction between titanium and zirconia. China Foundry, 2010, 7(4): 373-376. [6] Leyens C, Peters M. Titanium and titanium alloys: fundamentals and applications. Wiley-VCH: Cologne, 2003. [7] Jia Q, Cui Y Y, Yang R. Intensified interfacial reactions between gamma titanium aluminide and CaO stabilised ZrO 2. International Journal of Cast Metals Research, 2004, 17(1): 23-27. [8] Lin K F, Lin C C. Interfacial reactions between Ti-6Al-4V alloy and zirconia mold during casting. Journal of Materials Science, 1999, 34(23): 5899-5906. [9] Cui R J, Gao M, Zhang H, Gong S K. Interactions between TiAl alloys and yttria refractory material in casting process. Journal of Materials Processing Technology, 2010, 210(9): 1190-1196. [10] Saha R L, Nandy T K, Misra R D K, Jacob K T. Evaluation of the reactivity of titanium with mold materials during casting. Bulletin of Materials Science, 1989, 12(5): 481-493. [11] Sung S Y, Kim Y J. Economic net-shape forming of TiAl alloys for automotive parts. Intermetallics, 2006, 14(10-11): 1163-1167. [12] Lai P L, Chen W C, Wang J C, Huang T K, Hung C C. A newly developed calcia/titanium modified magnesia-based investment mold for titanium casting. Materials Science & Engineering C - Materials for Biological Applications, 2011, 31(2): 144-150. [13] Jia Q, Cui Y Y, Yang R. A study of two refractories as mould materials for investment casting TiAl based alloys. Journal of Materials Science, 2006, 41(10): 3045-3049. [14] Frueh C, Poirier D R, Maguire M C. The effect of silicacontaining binders on the titanium/face coat reaction. Metallurgical and Materials Transactions B - Process Metallurgy and Materials Processing Science, 1997, 28(5): 919-926. [15] Chen Yanfei, Xiao Shulong, Tian Jing, et al. Improvement in collapsibility of ZrO 2 ceramic mould for investment casting of TiAl alloys. China Foundry, 2011, 8(1): 9-13. [16] Kim M G, Kim S K, Kim Y J. Effect of mold material and binder on metal-mold interfacial reaction for investment castings of titanium alloys. Materials Transactions, 2002, 43(4), 745-750. [17] Bundy J, Viswanathan S. Characterization of zircon-based slurries for investment casting. International Journal of Metalcasting, 2009, 3(1): 27-35. [18] Ferenc J, Michalski J, Matysiak H, Sikorski K, Kurzydlowski K J. The influence of alumina powder on the rheological properties of zircon/silica slurries for investment casting of moulds. In: Proceedings of the Institution of Mechanical Engineers Part B - Journal of Engineering Manufacture, 2009, 223(11): 1417-1421. [19] Sidhu B S, Kumar P, Mishra K. Effect of slurry composition on plate weight in ceramic shell investment casting. Journal of Materials Engineering and Performance, 2008, 17(4): 489-498. 130