Leaching improvement of ceramic cores for hollow turbine blades based on additive manufacturing

The precision casting method based on alumina-based ceramic cores is one of the main techniques used to manufacture hollow turbine blades. Additive manufacturing (AM) technology provides an alternate solution to fabricating ceramic cores quickly and precisely. As the complexity of the structure increases and the strength of the material improves, the leaching process of the cores becomes more complicated. This study proposes a compound pore-forming method to increase the porosity of ceramic cores by adding a preformed-pore agent and materials that convert to easy-to-corrode phases. The preformed-pore agents (e.g., carbon fibers) can be burned off during sintering to form pores before the leaching, and the easy-to-corrode phases (e.g., CaCO3, SiO2, β-Al2O3) can be leached firstly to form pores during the leaching process. The pores formed in the aforementioned two stages increase the contact area of the cores and leaching solution, thus improving the leaching rate. In the current study, the additive amount of the preformed-pore agent was optimized, and the effect of the easy-to-corrode phases on the comprehensive properties of the cores was then compared. Based on this, the corresponding model was established.


Introduction
Currently, directional column-and single-crystal hollow turbine blades for aero engines use ceramic cores with high performance (e.g., silica, alumina). Because of their high chemical stability and good mechanical properties at high temperatures, alumina-based ceramic cores are commonly used in specific fields [1][2][3]. Traditionally, fabrication of the ceramic cores involves time-consuming procedures and low processing flexibility. The preparation of integral ceramic molds based on additive manufacturing (AM) technology provides a new method for fabricating ceramic cores with complex cooling channels efficiently [4]. Removing alumina-based ceramic cores, as represented by chemical leaching, is the key process after casting. However, high chemical stability makes the removal of cores particularly difficult. However, as the complexity of the structure increases and the strength of the material improves, the leaching process of the AM-based cores becomes more difficult [5]. One method to improve the chemical reaction is by selecting superior leaching solutions. Some types of effective leaching solutions indeed exist (i.e., hydrofluoric acid), which can concert Al 2 O 3 into soluble Al 3? to remove alumina-based ceramic cores efficiently [6]. The basis reaction equation is Li 3 AlF 6-? 2Al 2 O 3 = 3LiAlO 2 ? 2AlF 3 . However, the Fion has severely harmful effects on the human body and environment. Therefore, this method has been abandoned for many years [7]. Alkali-boiling leaching is now a commonly used method. The main research direction is now divided into two categories. One is to increase the leaching pressure, thereby increasing the temperature of the alkali solution and thus the reaction reactivity. High pressure can result in increased depth of the alkali solution into the cores, thereby increasing the contact area of the alkali solution and the cores [8][9][10]. The other research direction focuses on improving the pore structure by adding pore-forming agents to the cores to prefabricate pores, or by adding materials that will convert to easy-to-corrode phases, which can be dissolved in advance in the process of leaching to form pores, thus increasing the contact area of the alkali solution and the cores.
For the method of improving leaching pressure, significantly improving the equipment is necessary. The existence of high pressure typically requires equipment of a high standard, which often poses safety hazards. On the contrary, the method of introducing the pore structure is more efficient. This method can help to improve the porosity of the cores before and during the process of leaching, thus increase the contact area between the cores and leaching liquid. Still, the ceramic cores must meet the synthesis requirements of high-temperature strength, deflection, sintering shrinkage rate, and so on. Thus, this research represents a comprehensive optimization under a multi-objective constraint. In theory, the higher the porosity, the higher the leaching rate. However, the sintering shrinkage rate is bound to rise, and the high-temperature performance should deteriorate. Therefore, conducting systematic research on the pore-forming methods without considerably affecting other properties is necessary.
Therefore, in this study, different pore-forming methods are tested systematically. In addition, a compound method of pore-forming is introduced, which involves adding an easy combustible material to achieve preformed pores while also adding materials to the cores that convert to easy corrosion phases that can be removed prior to the process of leaching; this is likely to increase the porosity considerably. This method will likely achieve maximum porosity during leaching without affecting other properties, thus achieving a higher leaching rate.
In this study, the leaching rate of alumina-based ceramic cores is improved without affecting other properties. This is the basis for the study of the cores and promotes the technology of precision casting of hollow turbine blades [11,12].

Raw materials
Alumina-based ceramic core samples were shaped using gelcasting technology [4,13,14]. A ceramic slurry with a u = 60% solid content was prepared according to the gradations listed in Table 1, and the fused alumina was used in a technically pure form. The as-shaped ceramic core's green body was further treated through freeze drying [15][16][17], high-temperature sintering, and a dipping processing [18]. The freeze-drying process was conducted in a DTY-1SL vacuum freeze dryer. Then, the core samples sized (60 mm 9 10 mm 9 4 mm) were subjected to a rebaking treatment at 1 500°C for 0.5 h to model the pouring condition during a real investment casting. Through this treatment, the core samples could be used for the etching experiment.
The material used to preform pores was a carbon fiber that was easy to burn with a diameter of approximately 7-10 lm and an average length of approximately 500 lm. The materials that can convert to easy-to-corrode phases included CaCO 3 , b-Al 2 O 3 , and SiO 2 powders. The mean particle sizes of the CaCO 3 , b-Al 2 O 3 , and SiO 2 powders were 5 lm, 5 lm, and 40 lm, respectively. The carbon fiber was mixed with the ceramic slurry at w = 1%, 2% and 3%, respectively, of the total mass, and the corresponding mass of the alumina powder was deducted from the 40 lm series. The CaCO 3 powder was mixed with the ceramic slurry at w = 2%, 4%, 6% and 8%, respectively, of the total mass, and the corresponding mass of the alumina powder was deducted from the 5 lm series. The b-Al 2 O 3 powder was mixed with the ceramic slurry at w = 1%, 2%, 3% and 4%, respectively, of the total mass, and the corresponding mass of the alumina powder was deducted from the 5 lm series. The SiO 2 powder was mixed with the ceramic slurry at w = 1%, 2%, 3% and 4%, respectively, of the total mass, and the corresponding mass of the alumina powder was deducted from the 40 lm series.

Sampling preparation
Previous studies conducted by our group developed an AM-based method to fabricate ceramic cores [3,4,14], and the procedure for this is shown in Fig. 1. Directly driven by CAD digital data, the mold of the turbine is formed by stereolithography (SL), which is one of the AM processes, and the ceramic slurry is formed by the gelcasting method. After freeze drying, the ceramic core is formed by hightemperature sintering. The hollow blade can be cast using the strengthened ceramic core (see Fig. 1). First, the resin molds for the test samples were prepared using an AM apparatus (SPS600B, Xi'an Jiaotong University, Xi'an, China) with a photosensitive resin (SPR 8981, Zhengbang Technology Co., Ltd., Zhuhai, China). A ceramic slurry with a low viscosity (less than 1 PaÁs) and high solid loading (u = 60%) was prepared by ball-milling for 40 min. After degassing for 5 min, the ceramic slurry was poured into the resin mold and then polymerized in situ to form wet green bodies under the action of the initiator and catalyst. After freeze drying for 24 h, the dried green bodies were placed in a furnace, heated to 1 100°C, and maintained at that temperature for 3 h to pyrolyze the resin prototype and organic monomer polymers. Finally, to promote mullitization, the samples were sintered at 1 400°C for 3 h.

Leaching process
During the removal process, the ceramic core samples were immersed in the w = 70% KOH solution [5]. The etching experiment was conducted in an atmospheric Monel kettle at a constant temperature of 220°C.
A series of experiments were performed to investigate the relationships between the leaching rate and poreforming methods. In addition, the open porosity, sintering shrinkage rate, high-temperature strength, and high-temperature deflection of the cores with pore-forming additive materials were measured and analyzed synthetically.

Testing
During the experiments, the open porosity of the core samples was measured by the Archimedes method, and the sintering shrinkage rate was measured by Vernier callipers. The high-temperature strength was tested at 1 500°C in an HSST-6003QP high-temperature stress-strain testing machine (Sinosteel Luoyang Institute of Refractories Research Co., Ltd., China). The high-temperature deflection was tested in a TDV-1600PC high-temperature thermal deformation testing machine (Sinosteel Luoyang Institute of Refractories Research Co., Ltd., China) using ceramic samples of a nominal size of 2 mm 9 6 mm 9 120 mm. The samples were mounted on a silicon-nitride ceramic fixture whose support spacing was 100 mm within the chamber of the testing machine that had been heated at  3 Results and discussion 3.1 Effects of preformed pores on the performances of the cores The effect of the carbon fibers on the microstructures of the ceramic molds is illustrated in Fig. 2. The presence of the carbon fibers in the green bodies, as shown in Fig. 2a, led to a significant increase in the number of pores in the ceramic molds. It is worth mentioning that more open pores could be expected due to the high-aspect-ratio morphology of the carbon fibers, which would considerably enhance the leaching performance. However, the increase in porosity could possibly impair other performances of the ceramic core (e.g., the surface quality and mechanical properties). Therefore, the effects of the carbon fibers should be carefully investigated, and some trade-off must be made to balance all performances.
The results of the open porosities of the core samples with the carbon fiber additive are illustrated in Fig. 3. The open porosity of the core samples without the carbon fiber additive was u = 31.3%, and it increased when the carbon fiber content was increased, indicating u = 34.6% for w = 3%. The experimental results of the leaching rate influenced by the carbon fiber additive are illustrated in Fig. 4. The leaching rate of the core samples without the carbon fiber additive was 13.46%/h, and it increased along with the carbon fiber, indicating v = 18.48%/h for w = 3%.  With the carbon fiber additive, the viscosity of the ceramic slurry increased, which decreased the liquidity of the ceramic slurry when it was gelcasted, resulting in an insufficiently fine structure in the mold. Figure 5 shows the filling results of the tip part in the mold with different carbon fiber additives. As illustrated in Figs. 5a and b, the viscosity of the ceramic slurry is too high when the additive is w = 3% and 2%, respectively, and the tip parts could not be filled, leaving obvious defects in the cores. However, when the additive was w = 1%, the viscosity was moderate and the mold could be fully filled, as illustrated in Fig. 5c.
The addition of the carbon fiber can improve the leaching performance to some extent, but the increase in porosity can damage the mechanical properties. In addition, the theoretical solid phase content of the cores is reduced after the loss of the carbon fiber, and the sintering shrinkage rate increases, thus affecting the precision of the mold and blades. The other properties of the samples with w = 1% carbon fiber were tested as presented in Table 2. It can be seen that with the carbon fiber additive, the high-temperature strength was slightly lower, and the high-temperature deflection and sintering shrinkage rate increased slightly. However, these changes can be ignored. Therefore, the current study determined that the optimal additive amount of the carbon fiber was w = 1%. The leaching rate increased from 13.46%/h to 15.05%/h without affecting the moldfilling capacity of slurry and other properties.

Effect of the compound method of pore-forming on the performances of the cores
To increase the leaching rate further, based on w = 1% carbon fiber additive, materials that could convert to easyto-corrode phases were continuously added, which were leached prior to the process of leaching, with pores subsequently left in situ. The porosity increased further when the materials were being leached, thus the leaching rate was further improved. Figure 6 illustrates the influence rule of CaCO 3 on the leaching rate. Figure 6 indicates that with the increase of  CaCO 3 , the leaching rate improved. The leaching rate of the core samples without the CaCO 3 additive was 15.05%/ h, and it increased when the amount of CaCO 3 additive was increased, with 21.42%/h for w = 8%. Figure 7 illustrates the microstructure morphologies of ceramic samples after leaching without CaCO 3 and with w = 8% CaCO 3 . As shown in Fig. 7, the surface break degree of the samples with w = 8% CaCO 3 additive was greater than that of the samples without CaCO 3 , indicating that the cores with CaCO 3 could be leached more easily.

CaCO 3 additive
However, after CaCO 3 was added, the sintering shrinkage rate of the samples significantly increased. Figure 8 illustrates the effect of CaCO 3 on the sintering shrinkage rate, which increased from 0.15% to 5.59% when w(CaCO 3 ) = 8% was added.
The sintering shrinkage rate increased significantly when CaCO 3 was added. This was because CaO, which was resulted from the decomposition of CaCO 3 , reacts with Al 2 O 3 to generate calcium aluminate in the particle contact area. Calcium aluminate can be easily leached, but this process involves liquid phase sintering, which causes sintering densification, thus increasing the sintering shrinkage rate. As illustrated in Fig. 9, the microstructure    additive means that the density is very high and the porosity is very small, indicating that CaCO 3 promotes densification sintering. This means that meeting the hollow turbine blade's precision casting requirements is difficult. Therefore, CaCO 3 cannot be used to improve the leaching performance of ceramic cores.

b-Al 2 O 3 additive
The crystal structure of b-Al 2 O 3 is relatively loose, and the additive's density is 3.31 g/cm 3 , which is less than the 3.97 g/cm 3 of a-Al 2 O 3 . Therefore, the porosity will increase as b-Al 2 O 3 converts to a-Al 2 O 3 after sintering at 1 500°C. Therefore, the cores with b-Al 2 O 3 can be easily leached. Figure 10 illustrates the influence rule of b-Al 2 O 3 on the leaching rate, and shows that with the increase of b-Al 2 O 3 , the leaching rate was improved. The leaching rate of the core samples without the b-Al 2 O 3 additive was 15.05%/h, and it increased when the amount of b-Al 2 O 3 additive was increased, indicating 24.01%/h for w = 4%. Figure 11 illustrates the microstructure morphologies of ceramic samples after leaching without b-Al 2 O 3 and with w(b-Al 2 O 3 ) = 1%. As shown in Fig. 11, the surface of the    Figures 12 and 13 illustrate the influence rules of b-Al 2 O 3 on the high-temperature strength, high-temperature deflection, and sintering shrinkage rate. With the addition of b-Al 2 O 3 , the high-temperature strength decreased, and the high-temperature deflection and sintering shrinkage rate increased. When the mass fraction of b-Al 2 O 3 additive increased to 2%, the high-temperature strength decreased considerably, and the deflection increased rapidly. However, when the mass fraction of b-Al 2 O 3 additive was 1%, the leaching rate reached 17.46%/h, and the high-temperature strength, deflection, and sintering shrinkage rate were 20.89 MPa, 0.47 mm and 0.17%, respectively, which met the hollow turbine blade's precision casting requirements. Figure 14 illustrates the influence rule of SiO 2 on the leaching rate. Figure 14 indicates that the leaching rate of the core samples without the SiO 2 additive was 15.05%/h, and it increased when the amount of the SiO 2 additive was increased, indicating 22.04%/h for w = 4%. Figure 15 illustrates the microstructure morphologies of ceramic samples after leaching without SiO 2 and with w(SiO 2 ) = 3%. As shown in Fig. 15, the surface of the samples with w(SiO 2 ) = 3% additive was greater than that of the samples without SiO 2 , indicating that the cores with SiO 2 could be more easily leached because the porosity was improved when SiO 2 was leached in the early stage of the leaching process. Figures 16 and 17 illustrate the influence rules of SiO 2 on the high-temperature strength, high-temperature deflection, and sintering shrinkage rate. With the addition of SiO 2 , the high-temperature strength decreased, and the high-temperature deflection and sintering shrinkage rate increased. When the mass fraction of SiO 2 additive increased to 4%, the high-temperature strength decreased significantly, and the deflection increased rapidly. The  reason for this is that the presence of a few low-meltingpoint phases in SiO 2 can help to reduce the high-temperature strength, accelerate the creep, and promote the sintering, leading to an increased sintering shrinkage rate. Simultaneously, SiO 2 transformed into cristobalite during the sintering process [19], which severely shrank when cooled to 180-270°C. The shrinkage could easily produce microcracks, thus damaging the high-temperature strength and improving the sintering shrinkage rate. However, when the mass fraction of SiO 2 additive was 3%, the leaching rate reached 20.92%/h, and the hightemperature strength, deflection, and sintering shrinkage rate were 20.60 MPa, 0.48 mm and 0.20%, respectively, which met the hollow turbine blade's precision casting requirements.

Mechanism of pore-forming methods to improve leaching performance
In theory, the higher the porosity, the larger is the contact area between the leaching liquid and the cores and the higher is the leaching rate. Based on the w(C fiber) = 1%, adding materials that can convert to easy-to-corrode phases can improve the leaching performance of the cores. This is because the compound pore-forming method can achieve preformed pores prior to the leaching. In addition, the easy corrosion phases are leached in advance, and thus more pores are generated, thus increasing the leaching rate. The principles behind the process are illustrated in Fig. 18. In a typical condition, combining the two mechanisms, such as performed pores and easy-to-corrode phases, improve the leaching performance. For the former, carbon fibers are used as the additives and open pores are produced by burning out carbon fibers during sintering. For the latter, more pores are introduced by eroding the easyto-corrode phases, such as CaCO 3 , b-Al 2 O 3 , and SiO 2 .
The chemical reaction during the process is The removal efficiency is described quantitatively as the leaching rate by measuring the dry weights before and after the removal process, and the leaching rate (v) is calculated by the following equation where m 1 and m 2 are the weights of the core samples before and after the leaching process, respectively, and t is the etching time.   Fig. 19. A comparison of comprehensive performances of the different samples is presented in Table 3.

Conclusions
The effects of different pore-forming methods on the leaching rates of alumina-based ceramic cores was studied, and the effects of different pore-forming methods on the other properties of the cores was discussed. The addition of carbon fiber increased the porosity of cores, thus improving the leaching performance. When the amount of added carbon fiber was w = 1%, the porosity and leaching rate increased to 32.4% and 15.05%/h, respectively, whereas other performances (e.g., filling capacity and mechanical properties) were maintained. CaCO 3 could improve the leaching performance, but it resulted in an extremely high sintering shrinkage rate. b-Al 2 O 3 and SiO 2 increased the porosity, thus improving the leaching performance. When the mass fraction of added b-Al 2 O 3 and SiO 2 were 1% and 3%, respectively, the leaching rate increased to 17.46%/h and 20.92%/h, respectively, without hampering other properties. As previously mentioned, carbon fiber, CaCO 3 , b-Al 2 O 3 , and SiO 2 are all additives that improve the leaching performance of a-Al 2 O 3 -based ceramic cores. However, the mechanisms of improvement are different. Carbon fibers are used to preform pores (open pores in the case), whereas CaCO 3 , b-Al 2 O 3 , and SiO 2 are based on introducing easy-to-corrode phases into the a-Al 2 O 3 matrix. In this study, the effect of every additive was discussed, and combinations of additives were also investigated by mixing carbon fibers with CaCO 3 , b-Al 2 O 3 , and SiO 2 additive. However, combining two additives of the same mechanism (e.g., CaCO 3 with b-Al 2 O 3 , b-Al 2 O 3 with SiO 2 ) was not conducted in this study. With an increase in the number of design variables, further improvement of the leaching performance could be expected by optimizing the formula of the two additives of the same mechanism. Further studies are recommended to investigate the effects of combining different additives on the leaching performances of ceramic cores.