Thin-Rib and High Aspect Ratio Non-Stochastic Sca ff olds by Vacuum Assisted Investment Casting

Cellular structures are a classic route to obtain high values of specific mechanical properties. This characteristic is advantageous in many fields, from diverse areas such as packaging, transportation industry, and/or medical implants. Recent studies have employed additive manufacturing and casting techniques to obtain non-stochastic cellular materials, thus, generating an in situ control on the overall mechanical properties. Both techniques display issues, such as lack of control at a microstructural level in the additive manufacturing of metallic alloys and the difficulty in casting thin-rib cellular materials (e.g., metallic scaffolds). To mitigate these problems, this study shows a combination of additive manufacturing and investment casting, in which vacuum is used to assist the filling of thin-rib and high aspect-ratio scaffolds. The process uses 3D printing to produce the investment model. Even though, vacuum is fundamental to allow a complete filling of the models, the temperatures of both mold and casting are important to the success of this route. Minimum temperatures of 250 ◦C for the mold and 700 ◦C for the casting must be used to guarantee a successful casting. Cast samples shown small deviations relatively to the initial CAD model, mainly small expansions in rib length and contraction in rib thickness may be observed. However, these changes may be advantageous to obtain higher values of aspect ratio in the final samples.

Recent studies have explored the use of additive manufacturing (e.g., selective metal laser or electron beam melting) [35][36][37][38][39] and metal casting [40,41] to produce non-stochastic cellular materials.This has the advantage of allowing in situ geometric and dimensional control, thus, they are able to display tailored properties [42].However, the referred processes are known to have inherent problems being; the difficult microstructural control in additively manufactured metals [43][44][45] and the casting of thin-rib cellular materials.Even though there have been reported casting products with extremely thin sections (~0.1 mm) and high aspect ratios [46], their application in a metallic scaffold shape is still a challenge.Vacuum assisted high pressure die [47][48][49] and low pressure casting [50,51] are able to produce thin-walled and/or high detailed structures, however, they are not suited for the manufacturing of intricate three-dimensional complex geometries.
The present study shows a combination of additive manufacturing, in which fused filament fabrication (PLA) is used to produce the intermediate investment model to be cast in ceramic block by investment casting.Although there are already studies that use similar techniques (e.g., 3D printing of sand molds [52][53][54][55][56], wax, and PLA/ABS [52][53][54][55][56]), the proposed study recurs to vacuum as an auxiliary technique for the filling of thin-rib and high aspect ratio complex three-dimensional metallic scaffolds.Additionally, vacuum is also used to promote the sanity of the resultant samples, as it is known to reduce oxide inclusion [57,58] and porosity [59,60].

Scaffold Design and Selection
Samples were designed using a basic three-dimensional honeycomb unitary cell configuration, in which the cells are assembled in a matrix (9 by 9 by 8) configuration, according to Figure 1.To generate different casting conditions while keeping the casting volume approximately constant (i.e., amount of material), the linear dimensions of the cells [1] were kept constant changing only the rib angle (Table 1).Figure 1 shown the influence of rib angle in the overall scaffold configuration, where the negative angles (Figure 1a,b) are gradually widened to null angles (cubic cell-Figure 1c) and positive angles (Figure 1d,e) samples.An overall theoretical aspect ratio of 6.7 is maintained across the models due to the square cross section of the ribs.problems being; the difficult microstructural control in additively manufactured metals [43][44][45] and the casting of thin-rib cellular materials.Even though there have been reported casting products with extremely thin sections (~0.1 mm) and high aspect ratios [46], their application in a metallic scaffold shape is still a challenge.Vacuum assisted high pressure die [47][48][49] and low pressure casting [50,51] are able to produce thin-walled and/or high detailed structures, however, they are not suited for the manufacturing of intricate three-dimensional complex geometries.The present study shows a combination of additive manufacturing, in which fused filament fabrication (PLA) is used to produce the intermediate investment model to be cast in ceramic block by investment casting.Although there are already studies that use similar techniques (e.g., 3D printing of sand molds [52][53][54][55][56], wax, and PLA/ABS [52][53][54][55][56]), the proposed study recurs to vacuum as an auxiliary technique for the filling of thin-rib and high aspect ratio complex three-dimensional metallic scaffolds.Additionally, vacuum is also used to promote the sanity of the resultant samples, as it is known to reduce oxide inclusion [57,58] and porosity [59,60].

Scaffold Design and Selection
Samples were designed using a basic three-dimensional honeycomb unitary cell configuration, in which the cells are assembled in a matrix (9 by 9 by 8) configuration, according to Figure 1.To generate different casting conditions while keeping the casting volume approximately constant (i.e., amount of material), the linear dimensions of the cells [1] were kept constant changing only the rib angle (Table 1).Figure 1 shown the influence of rib angle in the overall scaffold configuration, where the negative angles (Figure 1a,b) are gradually widened to null angles (cubic cell-Figure 1c) and positive angles (Figure 1d,e) samples.An overall theoretical aspect ratio of 6.7 is maintained across the models due to the square cross section of the ribs.Considering the analysis of mold and casting temperatures in the successful filling of the scaffolds obtained by vacuum assisted investment casting, the CAD models were compared to determine the most difficult sample to manufacture.It is known that in casting procedures, variables Considering the analysis of mold and casting temperatures in the successful filling of the scaffolds obtained by vacuum assisted investment casting, the CAD models were compared to determine the most difficult sample to manufacture.It is known that in casting procedures, variables such as filling length and distance to the mold wall have a fundamental importance in the overall success of the casting procedure.Thus, the samples were characterized in terms of their diagonal cross-sectional area (Figure 2a).According to Figure 2, it may be observed that the sample with higher cross section area (consequently higher values of filling length and lower mold wall distance) is the one with a 20 • rib angle (Figure 2b).Thus, this was the sample that was selected for this study (Figure 1d).such as filling length and distance to the mold wall have a fundamental importance in the overall success of the casting procedure.Thus, the samples were characterized in terms of their diagonal cross-sectional area (Figure 2a).According to Figure 2, it may be observed that the sample with higher cross section area (consequently higher values of filling length and lower mold wall distance) is the one with a 20° rib angle (Figure 2b).Thus, this was the sample that was selected for this study (Figure 1d).

Ceramic Block Manufacturing
Molds were manufactured using plaster (GoldStar Omega+, Staffordshire, UK) to produce a ceramic block.According to the manufacturer guidelines, the plaster was subjected to the thermal cycle showed in Figure 3a.Given that 3D printing is used to produce a PLA investment model it may be observed that, according to the thermogravimetric analysis displayed in Figure 3b, the maximum temperature of 730 °C is able to completely eliminate the PLA model and generate the cavity to be filled in the casting process.

Vacuum Assisted Investment Casting
Molds were introduced into an Indutherm MC15+ casting furnace, while 24 (±0.02)g of A356 alloy, 0.05 g (±0.01) of Al5Ti1B grain refiner and 0.07 g (±0.01) of Al10Sr eutectic Si modifier are placed in a graphite crucible.After the furnace is closed, the metal composition is heated recurring to induction current to the desired casting temperature.Given the low amount of material to be melted, generally, a complete melting can take up to 1.5 min.After such period the melt is subjected to the induction current to allow a homogenous spreading of the grain refiner and eutectic Si modifier for

Ceramic Block Manufacturing
Molds were manufactured using plaster (GoldStar Omega+, Staffordshire, UK) to produce a ceramic block.According to the manufacturer guidelines, the plaster was subjected to the thermal cycle showed in Figure 3a.Given that 3D printing is used to produce a PLA investment model it may be observed that, according to the thermogravimetric analysis displayed in Figure 3b, the maximum temperature of 730 • C is able to completely eliminate the PLA model and generate the cavity to be filled in the casting process.such as filling length and distance to the mold wall have a fundamental importance in the overall success of the casting procedure.Thus, the samples were characterized in terms of their diagonal cross-sectional area (Figure 2a).According to Figure 2, it may be observed that the sample with higher cross section area (consequently higher values of filling length and lower mold wall distance) is the one with a 20° rib angle (Figure 2b).Thus, this was the sample that was selected for this study (Figure 1d).

Ceramic Block Manufacturing
Molds were manufactured using plaster (GoldStar Omega+, Staffordshire, UK) to produce a ceramic block.According to the manufacturer guidelines, the plaster was subjected to the thermal cycle showed in Figure 3a.Given that 3D printing is used to produce a PLA investment model it may be observed that, according to the thermogravimetric analysis displayed in Figure 3b, the maximum temperature of 730 °C is able to completely eliminate the PLA model and generate the cavity to be filled in the casting process.

Vacuum Assisted Investment Casting
Molds were introduced into an Indutherm MC15+ casting furnace, while 24 (±0.02)g of A356 alloy, 0.05 g (±0.01) of Al5Ti1B grain refiner and 0.07 g (±0.01) of Al10Sr eutectic Si modifier are placed in a graphite crucible.After the furnace is closed, the metal composition is heated recurring to induction current to the desired casting temperature.Given the low amount of material to be melted, generally, a complete melting can take up to 1.5 min.After such period the melt is subjected to the induction current to allow a homogenous spreading of the grain refiner and eutectic Si modifier for

Vacuum Assisted Investment Casting
Molds were introduced into an Indutherm MC15+ casting furnace, while 24 (±0.02)g of A356 alloy, 0.05 g (±0.01) of Al 5 Ti 1 B grain refiner and 0.07 g (±0.01) of Al 10 Sr eutectic Si modifier are placed in a graphite crucible.After the furnace is closed, the metal composition is heated recurring to induction current to the desired casting temperature.Given the low amount of material to be melted, generally, a complete melting can take up to 1.5 min.After such period the melt is subjected to the induction current to allow a homogenous spreading of the grain refiner and eutectic Si modifier for 3 min.Finally, the melt is cast into the mold where the sample is allowed to solidify for 10 min before tossing the mold in mild water to remove the plaster.Additionally, to promote a successful filling, vacuum was applied to some sample during the melting and casting procedures.
Table 2 displays the array of mold (T M ) and casting (T C ) temperatures that were studied.For each temperature combination a total of eight samples were produced, in which four where manufactured without vacuum and the remaining samples used vacuum assisted casting.

Influence of Mold and Casting Temperatures
Distinct mold and casting temperatures were used to determine their influence in a successful casting of thin-rib aluminum scaffolds in the presence or absence of auxiliary vacuum.A combination of the referred variables was used, generating the conclusion that their control is fundamental to guarantee a successful casting of the proposed cellular structure.Figure 4 shows a successful attempt to cast the proposed scaffold (20 • rib angle, Figure 1d) using vacuum assistance and mold/casting temperatures of, respectively, 250 • C and 700 • C.
3 min.Finally, the melt is cast into the mold where the sample is allowed to solidify for 10 min before tossing the mold in mild water to remove the plaster.Additionally, to promote a successful filling, vacuum was applied to some sample during the melting and casting procedures.
Table 2 displays the array of mold (TM) and casting (TC) temperatures that were studied.For each temperature combination a total of eight samples were produced, in which four where manufactured without vacuum and the remaining samples used vacuum assisted casting.

Influence of Mold and Casting Temperatures
Distinct mold and casting temperatures were used to determine their influence in a successful casting of thin-rib aluminum scaffolds in the presence or absence of auxiliary vacuum.A combination of the referred variables was used, generating the conclusion that their control is fundamental to guarantee a successful casting of the proposed cellular structure.Figure 4 shows a successful attempt to cast the proposed scaffold (20° rib angle, Figure 1d) using vacuum assistance and mold/casting temperatures of, respectively, 250 °C and 700 °C. Figure 4 shows that the adopted method is able to display a good shape and detail correlation to the CAD model that is displayed in Figure 1d.The unitary cells assume a full three-dimensional honeycomb configuration and are fully connected between themselves, presenting a complete assembly of the scaffold.
It is shown that the use of the described technique, it is possible to reproduce the intended design, however as shown in Figure 5, the use of different temperature has a fundamental role in the obtaining of a successful casting.According to the results presented in Figure 5a, the use of TM = 200 °C and TC = 680 °C with vacuum assistance, is not able to guarantee the successful filling of the model.Further increasing the casting temperature by 20 °C (TM = 200 °C and TC = 700 °C-Figure 5b), it may be stated that the overall filling is increased, however, it has still shown an absence of material in the sample corners.By the increase of mold temperature (TM = 250 °C and TC = 700 °C-Figure 5c) it is possible to guarantee a successful filling of the adopted model, considering that the overall casting process is correctly performed.Figure 4 shows that the adopted method is able to display a good shape and detail correlation to the CAD model that is displayed in Figure 1d.The unitary cells assume a full three-dimensional honeycomb configuration and are fully connected between themselves, presenting a complete assembly of the scaffold.
It is shown that the use of the described technique, it is possible to reproduce the intended design, however as shown in Figure 5, the use of different temperature has a fundamental role in the obtaining of a successful casting.According to the results presented in Figure 5a 5b), it may be stated that the overall filling is increased, however, it has still shown an absence of material in the sample corners.By the increase of mold temperature (T M = 250 • C and T C = 700 • C-Figure 5c) it is possible to guarantee a successful filling of the adopted model, considering that the overall casting process is correctly performed.Overall, it may be determined by the results in Figure 5 that in near complete filling conditions, the increase of mold and casting temperatures is able to enhance the filling by, respectively, 0.16% °C−1 and 0.75% °C−1 .Thus, it is suggested that the influence of casting temperature is more prominent for the filling of complex aluminum thin-ribs in near complete casting conditions.
These suggestions are further supported by the overall comparison of average sample filling showed in Figure 6.As expected, high temperatures of both mold and casting are extremely useful to allow a complete filling of the model (Figure 5a).However, it is also shown that in the absence of vacuum (Figure 5b), it is not possible to cast the scaffolding structures.Even though, the use of high mold and casting temperatures is able to enhance the filling of gravity casting samples, overall the experimental results suggest a maximum filling of approximately 20%.It could be argued that further increasing the mold and casting temperatures could increase the filling and eventually allow a complete casting in non-vacuum conditions.However, it is also known that the superheating of aluminum alloy melts is extremely hazardous for the final casting products by the decrease in microstructural sanity [61,62].Thus, minimal mold and casting temperatures should be used, as long as the casting process is not compromised.Overall, it may be determined by the results in Figure 5 that in near complete filling conditions, the increase of mold and casting temperatures is able to enhance the filling by, respectively, 0.16% • C −1 and 0.75% • C −1 .Thus, it is suggested that the influence of casting temperature is more prominent for the filling of complex aluminum thin-ribs in near complete casting conditions.
These suggestions are further supported by the overall comparison of average sample filling showed in Figure 6.As expected, high temperatures of both mold and casting are extremely useful to allow a complete filling of the model (Figure 5a).However, it is also shown that in the absence of vacuum (Figure 5b), it is not possible to cast the scaffolding structures.Even though, the use of high mold and casting temperatures is able to enhance the filling of gravity casting samples, overall the experimental results suggest a maximum filling of approximately 20%.It could be argued that further increasing the mold and casting temperatures could increase the filling and eventually allow a complete casting in non-vacuum conditions.However, it is also known that the superheating of aluminum alloy melts is extremely hazardous for the final casting products by the decrease in microstructural sanity [61,62].Thus, minimal mold and casting temperatures should be used, as long as the casting process is not compromised.Overall, it may be determined by the results in Figure 5 that in near complete filling conditions, the increase of mold and casting temperatures is able to enhance the filling by, respectively, 0.16% °C−1 and 0.75% °C−1 .Thus, it is suggested that the influence of casting temperature is more prominent for the filling of complex aluminum thin-ribs in near complete casting conditions.
These suggestions are further supported by the overall comparison of average sample filling showed in Figure 6.As expected, high temperatures of both mold and casting are extremely useful to allow a complete filling of the model (Figure 5a).However, it is also shown that in the absence of vacuum (Figure 5b), it is not possible to cast the scaffolding structures.Even though, the use of high mold and casting temperatures is able to enhance the filling of gravity casting samples, overall the experimental results suggest a maximum filling of approximately 20%.It could be argued that further increasing the mold and casting temperatures could increase the filling and eventually allow a complete casting in non-vacuum conditions.However, it is also known that the superheating of aluminum alloy melts is extremely hazardous for the final casting products by the decrease in microstructural sanity [61,62].Thus, minimal mold and casting temperatures should be used, as long as the casting process is not compromised.Considering these facts and observing Figure 5a, it may be stated that the most beneficial combination of mold and casting temperatures for the adopted vacuum assisted investment casting process are, respectively, T M = 250 • C and T C = 700 • C. Additionally, it is known that the use of vacuum is beneficial for the overall sanity of aluminum cast products.Reducing the referred temperatures is not recommended given that the thin-rib configuration is easily clogged by the premature alloy solidification due to the high cooling rates in low temperature molds or near solidus configuration in low temperature castings.

Dimensional Characterization of the Manufacturing Process
The use of investment casting is generally associated with the intent to manufacture samples with a low dimensional deviations, high tolerance, and an elevated replication rate of the investment model [63].Thus, the adopted process is intended to have the capability of fully reproducing the proposed CAD model (Figure 1d). Figure 7 shows a dimensional analysis of the manufactured samples when compared with the proposed CAD model.It may be observed that, on average the rib length tends to be increased, while it is decreased in terms of thickness.Considering these facts and observing Figure 5a, it may be stated that the most beneficial combination of mold and casting temperatures for the adopted vacuum assisted investment casting process are, respectively, TM = 250 °C and TC = 700 °C.Additionally, it is known that the use of vacuum is beneficial for the overall sanity of aluminum cast products.Reducing the referred temperatures is not recommended given that the thin-rib configuration is easily clogged by the premature alloy solidification due to the high cooling rates in low temperature molds or near solidus configuration in low temperature castings.

Dimensional Characterization of the Manufacturing Process
The use of investment casting is generally associated with the intent to manufacture samples with a low dimensional deviations, high tolerance, and an elevated replication rate of the investment model [63].Thus, the adopted process is intended to have the capability of fully reproducing the proposed CAD model (Figure 1d). Figure 7 shows a dimensional analysis of the manufactured samples when compared with the proposed CAD model.It may be observed that, on average the rib length tends to be increased, while it is decreased in terms of thickness.It is known that during its thermal cycle, the plaster releases its water content losing mass, however, the overall mold can suffer an expansion due to the α-to β-transitions in the cristobalite (5.7% in volume) and quartz (4.3% in volume) components of the plaster [64].Additionally, aluminum alloys suffer from considerable shrinkage and linear contraction during solidification [65,66].These effects are able to influence the final cast sample dimensions, in which the higher dimensions of rib length show a similar increase in both horizontal and vertical configurations (2.5% expansion-according to Figure 7).The rib thickness shows a more prominent relative variation possessing a near 10% linear contraction.However, when adopting an absolute dimensional variation, these changes may be rendered as negligible: on average, rib length maximum increase is 0.12 mm, while it is reduced by 0.06 mm in thickness.It is also shown that, according to the values of standard deviation, the overall aspect ratio may assume values between 6.1 and 9.8.This implies that the overall dimension change may be beneficial and further increase the aspect ratio.

Conclusions
A manufacturing route is presented for the production of thin-rib and high aspect ratio aluminum scaffolds that are modeled as an assembly of three-dimensional honeycomb unitary cells.It is shown that a combination different mold and casting temperatures must be employed to ensure a complete filling of the samples.
A minimum temperature combination of 250 °C for the mold and 700 °C for the casting must be used to guarantee the complete filling of the samples.It is suggested that in near complete casting It is known that during its thermal cycle, the plaster releases its water content losing mass, however, the overall mold can suffer an expansion due to the αto βtransitions in the cristobalite (5.7% in volume) and quartz (4.3% in volume) components of the plaster [64].Additionally, aluminum alloys suffer from considerable shrinkage and linear contraction during solidification [65,66].These effects are able to influence the final cast sample dimensions, in which the higher dimensions of rib length show a similar increase in both horizontal and vertical configurations (2.5% expansion-according to Figure 7).The rib thickness shows a more prominent relative variation possessing a near 10% linear contraction.However, when adopting an absolute dimensional variation, these changes may be rendered as negligible: on average, rib length maximum increase is 0.12 mm, while it is reduced by 0.06 mm in thickness.It is also shown that, according to the values of standard deviation, the overall aspect ratio may assume values between 6.1 and 9.8.This implies that the overall dimension change may be beneficial and further increase the aspect ratio.

Conclusions
A manufacturing route is presented for the production of thin-rib and high aspect ratio aluminum scaffolds that are modeled as an assembly of three-dimensional honeycomb unitary cells.It is shown that a combination different mold and casting temperatures must be employed to ensure a complete filling of the samples.
A minimum temperature combination of 250 • C for the mold and 700 • C for the casting must be used to guarantee the complete filling of the samples.It is suggested that in near complete casting conditions, the most important factor is the casting temperature.Additionally, it is shown that the vacuum assistance is needed to ensure the success of the procedure.
In terms of dimensional characterization, due to the volumetric changes by the expansion of plaster during its thermal cycle and alloy shrinkage/linear contraction during solidification, the cast specimens tend to increase their rib length and reduce the rib thickness when compared to the CAD model.However, this combination is able to contribute to enhance the overall high aspect ratio and thin-rib configuration that is proposed for the presented manufacturing method.

Figure 2 .
Figure 2. (a) Model selection by diagonal area analysis and (b) gating and selected model detail.

Figure 2 .
Figure 2. (a) Model selection by diagonal area analysis and (b) gating and selected model detail.

Figure 2 .
Figure 2. (a) Model selection by diagonal area analysis and (b) gating and selected model detail.

Figure 4 .
Figure 4. Developed manufacturing process as-cast sample (T M = 250 • C and T C = 700 • C, using vacuum assistance): (a) front and (b) corner views.
Figure4shows that the adopted method is able to display a good shape and detail correlation to the CAD model that is displayed in Figure1d.The unitary cells assume a full three-dimensional honeycomb configuration and are fully connected between themselves, presenting a complete assembly of the scaffold.It is shown that the use of the described technique, it is possible to reproduce the intended design, however as shown in Figure5, the use of different temperature has a fundamental role in the obtaining of a successful casting.According to the results presented in Figure5a, the use of T M = 200 • C and T C = 680 • C with vacuum assistance, is not able to guarantee the successful filling of the model.Further increasing the casting temperature by 20 • C (T M = 200 • C and T C = 700 • C-Figure5b), it may be stated that the overall filling is increased, however, it has still shown an absence of material in the sample corners.By the increase of mold temperature (T M = 250 • C and T C = 700 • C-Figure5c) it is possible to guarantee a successful filling of the adopted model, considering that the overall casting process is correctly performed.

Figure 6 .
Figure 6.Average sample filling using different combinations of mold and casting temperatures: (a) with vacuum; and (b) without vacuum.

Figure 6 .
Figure 6.Average sample filling using different combinations of mold and casting temperatures: (a) with vacuum; and (b) without vacuum.

Figure 6 .
Figure 6.Average sample filling using different combinations of mold and casting temperatures: (a) with vacuum; and (b) without vacuum.

Figure 7 .
Figure 7. Sample (a) absolute and (b) relative dimensional variations after casting.

Figure 7 .
Figure 7. Sample (a) absolute and (b) relative dimensional variations after casting.