helium-assisted sand casting abstract

43International Journal of Metalcasting/Fall 2012

HELIUM-ASSISTED SAND CASTING

M. Saleem University of Engineering and Technology, Lahore, Pakistan

M. Makhlouf Worcester Polytechnic Institute, Worcester, MA, USA

Copyright © 2012 American Foundry Society

Abstract

This paper reports a novel approach to enhance the rate of heat extraction from sand castings by flowing helium gas through the sand mold during the course of the process. The effect of helium flow rate, flow direction, and mold design on the thermal profile, grain size, secondary dendrite arm spacing, and tensile properties of the castings is investigated. An optimum set of process parameters is identified and a “performance index” for assessing the efficiency of the helium-assisted sand casting process

is defined. It is found that when the helium-assisted aluminum sand casting process is performed with close to the optimum parameters, it produces castings that exhibit a 22 percent increase in ultimate tensile strength and a 34 percent increase in yield strength with no significant loss of ductility and no degradation in the quality of the as-cast surfaces.

Keywords: helium, sand casting

Introduction

Sand casting is the most widely used casting process for both ferrous and non-ferrous alloys,1 however, the process is marred by large grain size structures and long solidifica-tion times. The coarser microstructure has a negative effect on the mechanical properties of the cast components and the long processing time affects the overall productivity of the process. The research reported herein addresses these prob-lems by enhancing the rate of heat extraction from the cast-ing by replacing air, which is typically present in the pores of the sand mold and has a relatively low thermal conductiv-ity2 by helium gas, which has a thermal conductivity that is at least five times that of air in the temperature range of in-terest.3 Various studies have established an inverse relation-ship between the cooling rate and the coarseness of the mi-crostructure of the cast component.4,5 Dendrite arm spacing and grain size are two microstructure features that become refined with an increased cooling rate;5-7 and the quantita-tive characterization of these two parameters is often used to gauge the coarseness (or fineness) of the microstructure.

The secondary dendrite arm spacing (SDAS) is related to the mean cooling rate, during solidification (R) as shown in Equation (1)8

SDAS = CRm Eqn.1

In Equation 1, C is a constant, R is in °C/sec and SDAS is in μm. Extensive studies have been performed to investigate the effect of SDAS on mechanical properties of aluminum al-

loys, especially their tensile properties.4,5,9,10 It is an accepted conclusion that, with other factors kept constant, a smaller SDAS results in better tensile properties.4,5,9-11 Similarly, a fine grain structure results in enhanced tensile properties, less tendency towards hot tearing, better pressure tightness, consistent properties after heat treatment, as well as a finer distribution of secondary phases and pores.7,12

Typically, a sand mold is a porous medium that consists of an aggregate of sand particles with air occupying the voids between the sand particles. Therefore, the type of sand and binder, as well as the particle size of the sand, the volume fraction of voids, the type of gas in the voids, as well as the temperature of the mold all affect the thermal conductivity of the mold.13 Keeping everything else constant, the extrac-tion of heat from the mold may be enhanced by replacing the air in the voids between the sand particles by another gas that has a higher thermal conductivity than air. Helium and hydrogen are two common gases whose thermal con-ductivity is appreciably higher than that of air.6,14,15 There are serious safety issues associated with hydrogen which prevent its use in sand molds.14,15 Helium, on the other hand, is a non-flammable and non-toxic gas that is readily avail-able.3,14,15 In the temperature range between 25°C (77°F) and 500°C (932°F), the thermal conductivity of helium is at least five times that of air.3 Therefore, if the air present within the voids of sand molds is replaced by helium during sand casting of aluminum alloys, then the rate of heat extraction from the casting would significantly increase resulting in a faster cooling rate with all its associated benefits including microstructure refinement and improved tensile properties.16

44 International Journal of Metalcasting/Fall 2012

Background

The use of helium to enhance the rate of heat extraction from metal castings has been reported by many researchers with encouraging results.3,6,14,15,17 Doutre14,15 injected helium at the interface of metallic molds during solidification of the casting and concentrated on cooling time reduction and productivity aspects. He reported reductions in cooling time ranging between 30% and 50% for a range of aluminum al-loys as well as a 29% increase in the production rate when helium was used in making a complex cored casting.14,15 Wan and Pehlke6 used heat transfer equations to show that using helium with metallic molds can result in an increase of 2.3-4.3 times in the interfacial heat transfer coefficient across air gaps of varying thicknesses.6 Argyropoulos and Carletti3 measured the increased interfacial heat transfer co-efficient in metallic molds resulting from the use of helium at the metal-mold interface. They reported a 48% increase in the average interfacial heat transfer coefficient in the ini-tial phase of solidification (i.e., from the beginning of the casting operation to the onset of the air gap) when helium was used in a metallic mold.3 Griffiths17 made castings in a controlled helium environment within an enclosed chamber and reported that the interfacial heat transfer coefficient in-creased by 70% for permanent mold castings and 20% for sand castings. Griffiths also reported a 50-90% increase in cooling rate along with 20% refine-ment in the SDAS, as well as 10-20% improvement in yield strength and ultimate tensile strength of castings in their as-cast condition.17 However Griffiths observed little improvement in the heat treated condition and reported that the mold gases diluted the effect of helium at the mold/casting interface.17

The work presented herein investi-gates the possibility of using helium in a continuous flow mode with the understanding that in addition to its better thermal properties compared to air, the forced flow of gas be-tween the sand particles introduces convection, which further improves the rate of heat extraction from the casting, and helps in expelling the process-generated mold gases. The effect of the flow rate of helium, the flow direction, and the mold design on the average as-cast grain size, the average as-cast SDAS, and the room temperature tensile properties of castings was investigated and com-pared to their counterparts produced in a typical sand casting process. In addition, a cost analysis of the he-

lium-assisted sand casting process was performed16 and an optimum set of parameters are identified.

Apparatus, Materials, and Procedures

Figure 1 is a schematic representation of the apparatus used in this work.

Three different modes of supplying helium to the mold were investigated. These are:

(i) Cross flow in a partially encapsulated mold(ii) Cross flow in an un-encapsulated mold(iii) Parallel flow in an un-encapsulated mold

Figure 2 is a schematic representation that illustrates the dif-ference between cross flow and parallel flow. The following modifications were done to the apparatus shown schemati-cally in Figure 1 in order to accommodate the different he-lium supply modes.

(i) Cross flow in a partially encapsulated mold—A partial encapsulation was designed such that it could accommodate sand molds up to a maximum size of 17.5 inches (444.5 mm) x 11.5 inches (292.1 mm) with a fixed height of 11 inches (279.4 mm). Figure 3 shows the main components of this ap-

Figure 2. Schematic representation illustrating the difference between cross flow and parallel flow of helium.

Figure 1. Schematic representation of the general apparatus.


paratus. The design consists of a base plate that provides a pedestal on which the mold rests and through which helium is supplied to the drag, an encapsulation case that provides par-tial encapsulation of the mold and a top sealing assembly that seals the peripheral gap which forms because of mismatch be-tween the sand mold and the encapsulating case. The helium was supplied at the base of the mold by means of a rectangular recess in the base plate. The recess ensured that helium was not supplied at one point, but rather it covered more than the area of the cast part’s surface. An epoxy sealant was used to seal the peripheral interface between the drag and the base plate in order to ensure that helium supplied at the base of the mold did not leak through this interface (a good mechanical seal should nullify the need for this chemical sealant). The gas was thus forced to pass through the mold in the upward direc-tion under the pressure of the incoming helium supply which was allowed to escape through the only exposed surface (top) of the mold.

(ii) Cross flow in an un-encapsu-lated mold - This mode of helium supply requires only the base of the device that is used with the cross flow in the partially encapsulated mold (Figure 3). Though helium was sup-plied in the same manner as for the partially encapsulated mold, more system losses could be expected due to the absence of mold encapsulation.

(iii) Parallel flow in an un-encapsu-lated mold - Two supply plates were designed for supplying helium in this mode: one for the cope and other for the drag. The two plates are shown schematically in Figure 4. These sup-

ply plates only differed from one another in the helium in-jection site location to ensure that the helium was supplied approximately equidistant and parallel to each surface. The supply plates were attached at opposite ends of the pouring basin. Before attaching the supply plates, the side taper of the mold was removed by means of a saw blade and a rectangular groove was also cut in the side of the mold to ensure that he-lium flows parallel to the surfaces and along the whole width of the plate casting. The supply plates were fixed to the mold so that the point of helium supply was positioned at the middle of the groove. The relative positions of the supply plates and mold parts are shown in Figure 5. Here too the epoxy sealant was used to seal the peripheral interfaces between the mold and the supply plates in order to prevent leakage of helium through these interfaces. The relatively shorter travel distance along the casting, the position of the groove (cut closer to the casting surface), the pressure of the incoming gas and the pe-

Figure 3. Schematic representation of the main components of the cross flow apparatus.

Figure 4. Plates used in the parallel flow apparatus. (a) for the cope, and (b) for the drag.

Figure 5. The relative positions of the supply plate and mold parts (a) for the cope, and (b) for the drag.(a) (b)

(a) (b)


ripheral sealing of the supply plate ensured that helium trav-eled along the surfaces rather than leaking and/or moving to the sides. Figure 6 shows the apparatus when the supply plate is clamped and sealed to the drag. Pictures of the apparatuses used in the three helium supply modes are shown in Figure 7.

The cross flow mode was investigated at a helium flow rate of 4 L/min whereas the parallel flow mode was investigated at flow rates of 1 L/min, 4 L/min and 8 L/min parallel to each surface of the casting. In all cases helium was introduced into the mold only when the mold was completely filled. The chemical composition of the alloy was measured using spark emission spectrometry and is shown in Table 1. The alloy was used in the un-grain refined, un-modified condition and had a chemical composition close to that of standard 319 alloy.

A two-part (cope and drag) recycled silica sand mold de-signed for casting a rectangular plate was used in the ex-periments. Table 2 shows its relevant technical character-istics. The cast part was a rectangular plate 8 inches (203.2 mm) long and 6 inches (152.4 mm) wide. The height of the plate measured at its sides was 1 inch (25.4 mm). The top surface of the plate had a side web that ran along its periph-ery. The web was approximately 0.375 inches (9.53 mm) wide with a height of 0.6 inches (15.24 mm) so that the effective thickness of the plate was 0.4 inches (10.16 mm). Figure 8 shows a schematic representation of the cast plate

Data acquisition systemHelium supply

Flow meter

Apparatus for partial encapsulation of sand mold

Un-encapsulation sand mold Sand mold with apparatus forparallel flow helium supply

(a) (b) (c)

Figure 7. Actual pictures of the apparatuses (a) cross flow in a partially encapsulated mold (b) cross flow in an un-encapsulated mold (c) parallel flow in an un-encapsulated mold.

Table 1. Chemical Composition of the Alloy as Measured by Spark Emission Spectrometry

Figure 6. Supply plate clamped & sealed to drag for parallel flow helium supply.

and Figure 9 shows the complete casting together with ris-ers, runners and gating system.


For each casting, 40 pounds of alloy were melted in an induc-tion furnace in a silicon carbide crucible. The melt was heated to 880±5°C (1616±41°F) and was degassed with argon for half an hour using a rotating impeller degasser. The melt sur-face was then drossed and a reduced pressure test (RPT) gas sample was taken. If porosity was found inside the solidified RPT sample, or more than three bubbles were detected on the surface of the sample, degassing was continued until a gas-free melt was obtained. At this point a sample was taken for chemical analysis by spark emission spectrometry. The aver-age pouring temperature was 850°C (1562°F).

Five K-type thermocouples (represented by TC in Figure 10) were inserted into the mold cavity to monitor the change in temperature of the plate during solidification of the casting. As shown in Figure 10, four of these thermocouples were near the risers and one was at the geometric center of the casting. The placement of the thermocouples was decided with the help of computer simulation of the process. The thermocouples were inserted and held in place by drilling holes through the cope. A small piece of aluminum tubing (approximately 1.5 inches (38.1 mm) in length) was placed in each hole with one end of the tubing aligned with the cope’s parting surface and fixed in place with a packing of refractory blanket. A thermocouple was pushed through each one of these tubes and into the mold cavity. This ar-rangement ensured that the thermocouples remained snug in their place up to the required depth in the casting. Thermo-couples TC1 and TC2 are nearest to the pouring basin and represent the hot end of the casting. This end is the farthest from the helium supply when helium was supplied in the parallel flow mode. Because of symmetry, the average tem-perature of these two thermocouples is taken to represent the

temperature (and cooling rate) at the hot end of the casting. Similarly, thermocouples TC4 and TC5 are farthest from the pouring basin and represent the cold end of the casting. This end is nearest to the helium supply when helium was sup-plied in parallel flow mode. Again, because of symmetry, the average temperature of these two thermocouples is taken to represent the temperature (and cooling rate) at the cold end of the casting. These thermocouples were connected to a data acquisition system. Data was collected at a sampling rate of 20,000 measurements per second and averaged us-ing 2,000 values to produce a time interval of 0.1 seconds between recorded data points.

A contact profilometer was used to characterize the surface roughness of the top surface of the cast plate by measur-ing R

a and R

z. In each case, twenty measurements were per-

formed and the average was recorded. The conditions used in measuring the surface roughness are shown in Table 3. The reason for choosing the top surface of the plate for char-acterization is that it presented the worst case scenario where helium directly impinged on the solidifying metal.

Eight specimens from each casting experiment were used to determine the room temperature tensile properties of the cast plate. The specimens were sub-size with a rectangular cross section and were machined from the plate casting from the locations shown in Figure 11. Tensile property measure-ments were conducted according to ASTM Standard B557-0618 with a Universal Testing machine. Strain was measured with an axial extensometer with a gage length of 1 inch (25.4 mm). The extensometer was used until the specimen was

Table 2. Technical Characteristics of the Sand Mold

Figure 8. A schematic representation of the cast plate. Figure 9. The complete casting showing the risers, runners, and gating system.


fractured; the testing machine ramp rate was 0.05 inches/min (1.27 mm/min). The data was digitally captured and analyzed to obtain tensile strength, yield strength, elonga-tion and modulus of elasticity. Prior to measuring the room temperature tensile strength, the specimens were heat treated according to a standard T6 schedule (see Table 4).

Microstructure characterization was performed on samples in their as-cast condition. The samples were cut from loca-tions close to the five thermocouples as shown in Figure 12. Each sample was mounted in Bakelite such that the metal-lographic surface was along the thickness of the cast plate. The samples were prepared for optical microscopy follow-ing typical metallographic sample preparation procedures. Two optical microscopes equipped with ultrahigh definition digital camera were used to examine and capture images of the microstructure. The captured images were exported to image processing software in order to characterize porosity, SDAS, and grain size. When required, Keller’s reagent19 was used to etch the samples.

In each case, the amount of porosity in the casting was quan-tified in terms of percent area of pores observed at a magnifi-cation of 50X in the polished samples. A total of eighty four micrographs were taken for each casting and the average of these measurements is reported. Grain size was measured by the Hilliard circular intercept method20 on etched samples observed at a magnification of 10X under polarized light. A

Figure 10. Location of the thermocouples in the plate casting.

Table 3. Conditions Used in Measuring the As-Cast Surface

Roughness of the Plate CastingsTable 4. Details of the Heat Treatment Schedule

Figure 11. Locations within the plate casting where tensile test specimens were machined.

Figure 12. Locations within the plate casting where samples were extracted for microstructure analysis.

total of five micrographs were taken for each measurement. Similarly, the SDAS was quantified by the linear intercept method from micrographs of etched samples taken at a mag-nification of 100X.

Results and Analysis

The cooling curve in the range 650°C (1202°F) to 400°C (752°F) at the hot end (that takes longest time to cool as shown by baseline experiment) of the solidifying plate for the helium-assisted (cross flow mode) and for the helium-assisted (parallel flow mode) sand casting processes together with that for the traditional sand casting process (baseline) are shown in Figures 13 and 14, respectively.


The average time-to-cool the casting from 650°C (1202°F) to 400°C (752°F) for the helium-assisted process in compar-ison to the baseline process is shown in Table 5. The advan-tage of using helium in the sand casting process can be seen from Table 6 and Figure 15, which summarize the benefit in terms of percent reduction in time to cool for the helium-assisted processes compared to traditional sand casting.

Careful examination of Table 5 reveals that there is only a small variation in the time-to-cool between the hot end and the cold end of the plate casting (about 11%) in the case of the baseline casting. This is due to the directional solidifica-

tion that occurs towards the source of the feed metal.7 Table 5 shows that the cooling pattern of the plate is not significant-ly affected by the helium-assisted process in the cross flow mode. However, more variation is observed in the magnitude of the time-to-cool along the length of the plate in the heli-um-assisted process in the parallel flow mode. This variation in time-to-cool when the helium-assisted process is used is caused by the way helium is introduced into the mold. He-lium is supplied from only one end of the mold where it has its maximum cooling effect and it loses its effectiveness as it takes up heat and moves along the length of the hot casting. This observation is important as it suggests that when helium is used in the parallel flow mode, then depending on the size

Table 5. Average time-to-cool the plate castings from 650°C (1202°F) to 400°C (752°F) in minutes

Figure 13. Cooling curves during helium-assisted (cross flow mode) casting and baseline casting obtained at the hot end of the plate.

Figure 14. Cooling curves during helium-assisted (parallel flow mode) casting and baseline casting at the hot end of the plate.


and the design of the part to be cast, more than one input point for helium may be required for correct thermal management. Equally important, helium-assisted sand casting in a parallel supply mode may be used to induce localized cooling of the cast part by carefully designing the locations of the helium input points.

Table 7 shows the Ra and R

z values for parts made with the

helium-assisted sand casting process compared to the Ra and

Rz values for parts made with

the baseline process. Table 7 shows that partial encap-sulation of the mold results in the highest surface rough-ness. This is attributed to the more direct impingement of helium on the solidifying surface. For all the other flow modes, the surface condition of the castings is not much different from that obtained with the baseline process.

Figure 16 shows micrographs representative of the micro-structure and distribution of pores in the baseline cast-ings. In this case, the aver-age volume percent porosity is 0.52%. Figure 17 shows micrographs representative of the microstructure and distribution of pores in the helium-assisted (cross flow) sand castings. More pores are observed in the helium-assist-ed (cross flow mode) castings than in the baseline castings. It is clear that not only the average area of porosity is in-creased in the helium-assisted (cross flow) sand casting pro-cess compared to the baseline process, but also the size of the average pore is increased – more so when the partially encapsulated mold is used than when the un-encapsulat-ed mold is used. On the other hand, less pores are observed in the helium-assisted (paral-lel flow mode) castings than in the helium-assisted (cross flow mode) castings and they are even less than in castings made with the baseline pro-cess. Figure 18 shows mi-

crographs representative of the microstructure typical of the helium-assisted (parallel flow) sand castings and Figure 19 summarizes the results for all the castings. It can be deduced from Figure 19 that helium, when flowing in the cross flow mode, passes through the molten alloy and thus causes more pores as compared to the parallel flow mode where it is not flowing through the alloy but rather parallel to its surface. Ac-cordingly, the cross flow mode may not be viable for making castings with thin section.

Table 6. Percent Reduction in Time-to-Cool in the Helium-Assisted Process Relative to the Baseline

Figure 15. Percent reduction in the time-to-cool from 650°C (1202°F) to 400°C (752°F) relative to the baseline.


Table 7. Surface Roughness of the Plate Castings in the As-Cast Condition

Figure 16. Micrographs representative of the microstructure of the baseline castings.

(a)

(b)Figure 17. Micrographs representative of the microstructure of the helium-assisted (cross flow) castings, helium flow rate 4L/min. (a) Partially encapsulated mold (b) un-encapsulated mold.


Figures 20, 21 and 22 show that the grain size of castings made by the helium-assisted process is refined compared to the grain size of the baseline castings. Similarly, Figure

23 shows that the SDAS of castings made by the helium-assisted process is refined compared to the grain size of the baseline castings.

Figure 19. Comparison of porosity present in plate castings made by the various helium-assisted casting processes and baseline.

Figure 20. Grain structure of the baseline casting (viewed with polarized light).

Figure 18. Micrographs representative of the microstructure of the helium-assisted (parallel flow) castings. (a) Helium supply: 1 L/min parallel to each face (b) Helium supply: 4 L/min parallel to each face (c) Helium supply: 8 L/min parallel to each face.

(a) (b) (c)


Figure 21. Grain structure of castings made by the helium-assisted process in cross flow mode, helium flow rate 4 L/min. (a) Partially encapsulated sand mold, and (b) un-encapsulated sand mold.

Figure 22. Grain structure of castings made by the helium-assisted process in parallel flow mode. (a) Helium supply: 1 L/min parallel to each face (b) Helium supply: 4 L/min parallel to each face (c) Helium supply: 8 L/min parallel to each face.

(a) (b)

(a)

(b) (c)

Figure 24 shows the average room temperature ultimate and yield strengths and Figure 25 shows the average modulus of elasticity and elongation of all the castings. The room

temperature tensile properties of castings made by the he-lium-assisted sand casting process in the cross flow mode are inferior to their counterparts obtained by traditional


Figure 24. Comparison of the average ultimate and yield strengths of plate castings (T6 temper) made by the various helium-assisted casting processes and baseline.

sand casting despite the refinement in microstructure. This is attributed to the higher level of porosity obtained in the castings made by the helium-assisted sand casting process in the cross flow mode. On the other hand, the room tem-perature tensile properties of castings made by the helium-assisted sand casting process in the parallel flow mode are superior to those of the baseline castings: At a helium flow rate of 4 L/min parallel to each face of the plate, the yield strength increases from 179 MPa to 241 MPa, the ultimate tensile strength increases from 218 MPa to 267 MPa and there is no significant decrease in elongation. Since the al-loy employed in this work has a composition similar to a standard 319 alloy (with a slightly higher Mg content) it had a lower elongation than the standard 319 alloy. For the baseline casting, approximately 1% elongation is in rea-sonable agreement with the handbook value of 2% for sand cast 319-T6 alloys.21

Figure 23. Comparison of the SDAS in plate castings made by the various helium-assisted casting processes and baseline.

Discussion

It can be seen from the results that beyond a certain flow rate, the benefits (i.e., the increase in cooling rate, refine-ment of the microstructure, improvement of the tensile properties, etc.) begin to show diminishing returns with increasing the helium flow rate. Therefore, it is important to determine the optimum helium flow rate. Since an im-proved yield strength of the casting is the targeted outcome of the helium-assisted process (and since this will occur at a cost penalty), then the optimum flow rate would be the one that gives maximum increase in yield strength with respect to the baseline at the minimum flow rate of helium. Based on this logic, a performance index (PI) for the pro-cess is defined herein as a ratio of percent increase in yield strength (relative to the baseline) to the helium flow rate; it is given by Equation 2.

Eqn. 2


Table 8 shows the percent increase in yield strength rela-tive to the baseline and the corresponding performance in-dex for the flow rates employed in this work and Figure 26 shows a plot of the performance index vs. helium flow rate. Since only the parallel flow mode resulted in improved yield strength, so only this mode is considered.

The maxima of the curve shown in Figure 26 occurs at a helium flow rate of 2.63 L/min parallel to each face of the casting and is thus the optimum flow rate beyond which the relative advantage of improved yield strength compared to the baseline casting would begin to diminish.

Figure 25. Comparison of the average modulus of elasticity and elongation of plate castings made by the various helium-assisted casting processes and baseline.

Table 8. Increase in Yield Strength Relative to the Baseline and the Corresponding Performance Index for the Flow Rates Employed in This Work

Figure 26. Performance Index as a function of helium flow rate.


The results indicate that the partially encapsulated mold produces better results than the un-encapsulated mold. The reduction in the average time-to-cool at a flow rate of 4 L/min (compared to the baseline) is about 30% for the par-tially encapsulated mold compared to only 17% for the un-encapsulated mold. Despite the obvious advantage that mold encapsulation provides in reducing the overall time-to-cool the casting, mold encapsulation does not yield a significant advantage in the range of temperature where dendrites grow. This may be attributed to the fact that encapsulation hin-ders the escape of mold gases that evolve during burn-off of the binder. When mixed with helium, these gases seem to adversely affect the efficiency of helium in extracting heat from the solidifying melt. This phenomenon has also been reported by Griffiths,17 who reported the dilution of the ef-fect of helium by binder gases when casting is performed in a box.17 This hypothesis is strengthened by the fact that more pores are observed in the castings when the process is car-ried out in an encapsulated mold than in an un-encapsulated mold.

The results also indicate that introducing helium into the mold by the parallel flow mode is more beneficial than by the cross flow mode. This may be attributed to the follow-ing reasons: (1) In the parallel flow mode, helium is made to flow in a horizontal direction and so it drives a portion of the heat energy out through the sides of the mold thus leaving only a fraction of the total heat energy to be driven through the height of the mold. On the other hand, in the cross flow mode, helium flows from the bottom plate, through the drag mold half to the casting and so it has to travel through the entire height of the mold where it picks up thermal energy, heats up, and loses some of its effectiveness in cooling the casting. (2) In the parallel flow mode helium makes better contact with both surfaces of the casting than in the cross flow mode where it strikes one side and then it flows around the casting. (3) In the parallel flow mode there is a stream of helium moving parallel to each one of the two surfaces of the plate thus making the volume flow rate of helium double that which is used in the cross flow mode.

Conclusions

• Helium-assisted sand casting with a cross flow of helium in either an un-encapsulated mold or a par-tially encapsulated mold is not viable for making castings with thin sections as it increases porosity in the cast component. Partial encapsulation also seems to cause entrapment of the process gases in the cast component.

• Helium-assisted sand casting with a parallel flow of helium is viable for making castings irrespective of part thickness. The only additional required equip-ment for the process is a supply plate. If this plate is made to be part of the mold during the molding process, then clamping and sealing the plate to the sand mold can be eliminated.

• At a helium flow rate of about 4 L/min in the par-allel flow mode, the cooling rate during dendrite formation increases by about 100% at locations near the point of supply of helium (compared to regular sand casting). Under similar conditions, in the cross flow mode, the increase is only 43%.

• In the parallel flow mode, the role of heat convec-tion is limiting and a point of diminishing returns is reached. Accordingly, a “Performance Index” is defined as a ratio of one of the main benefits of the helium-assisted sand casting process (i.e., the increase in yield strength compared to traditional sand casting) to the cost penalty of using helium (as represented by the helium flow rate). It is found that 2.6 L/min (parallel to each face) is the opti-mum helium flow rate that maximizes this perfor-mance index.

• Helium-assisted sand casting in the parallel flow or cross flow mode results in refinement of the sec-ondary dendrite arm spacing and the grain size; however the room temperature tensile properties for T6 heat treated specimens are improved only in castings that are made with the parallel flow mode. This is attributed to the higher porosity obtained when the cross flow mode is used.

• The average SDAS in plates that are sand cast with the helium-assisted process at a helium flow rate of 4 L/min parallel to each face is 46 μm. The average SDAS in the baseline castings is 60 μm. This cor-responds to a 23% reduction in SDAS due to the use of helium.

• The average yield strength of specimens machined from plates that are sand cast with the helium-assisted process at a helium flow rate of 4 L/min parallel to each face is 241 MPa. The average yield strength of specimens machined from the baseline castings is 179 MPa. This corresponds to a 34% increase in yield strength due to the use of helium.

• The average ultimate tensile strength of specimens machined from plates that are sand cast with the helium-assisted process at a helium flow rate of 4 L/min parallel to each face is 267 MPa. The av-erage ultimate tensile strength of specimens ma-chined from the baseline castings is 218 MPa. This corresponds to a 22% increase in ultimate tensile strength due to the use of helium.

• Helium-assisted sand casting with parallel flow mode produces a thermal gradient in the cast part. For this reason, there is a small gradient in SDAS and also in properties. Although the gradient ob-served in the plates is insignificant (due to the small length of the plate), it can be significant in larger parts. In order to mitigate this gradient, he-lium should be introduced into the mold at more than one location. The location of the helium in-put points may be determined be computer simu-lation of the casting process. On the other hand,


when needed, the flow of helium may be used to cause localized cooling or to induce directional so-lidification. In essence, helium would replace chills which may simplify the design of the mold.

• The surface of the cast plate is not affected by he-lium when it is supplied in the parallel flow mode.

• Cost analysis (not presented here) shows that heli-um-assisted sand casting in the parallel flow mode compares favorably to traditional sand casting up to a flow rate of 4 L/min (parallel to each face). Al-though the analysis performed as part of this work is relative and based on assumptions and relations obtained from the open literature, it can serve as basis for company-specific analyses.

Acknowledgments

The authors acknowledge Palmer Foundry, Palmer MA 01069, USA, whose support made this work possible.

REfERENCES

1. Ravi, B., Metal Casting: Computer-Aided Design and Analysis, Prentice Hall of India (Pvt.) Ltd. (2005).

2. Poirier, D.R., Geiger, G.H., Transport Phenomena in Materials Processing, The Minerals, Metals & Materials Society (TMS) (1994).

3. Argyropoulos, S.A, Carletti, H., “Comparisons of the Effects of Air and Helium on Heat transfer at the Metal-Mold Interface,” Metallurgical and Materials Transactions B, vol. 39(3), pp. 457-468 (2008).

4. Zhang, L.Y., Jiang, Y.H., Ma, Z., et al., “Effect of Cooling Rate on Solidified Microstructure and Mechanical Properties of Aluminium-A356 Alloy,” Journal of Materials Processing Technology, vol. 207(1-3), pp.107-111 (2008).

5. Dobrzanski, L.A., Maniara, R., Sokolowski, J.H., “The Effect of Cast Al-Si-Cu Alloy Solidification Rate on Alloy Thermal Characteristics,” Journal of Achievements in Materials and Manufacturing Engineering, vol. 17(1-2), pp. 217-220 (2006).

6. Wan, X., Pehlke, R.D., “Using Helium to Increase Heat Transfer at the Metal/Mold Interface in Permanent Mold Casting,” Transactions of the American Foundry Society, vol. 112, pp. 193-207 (2004).

7. Zalensas, D.L. ed., Aluminum Casting Technology, The American Foundrymen’s Society; (2001).

8. Caceres, C.H., Davidson, C.J., Griffiths, J.R., Wang, Q.G., “The Effect of Mg on the Microstructure and Mechanical Behavior of Al-Si-Mg Casting Alloys,” Metallurgical and Materials Transactions A, vol. 30A,

pp. 2611-2618 (1999).9. Ceschini, L., Morri, A., Morri, A., Gamberini, A.,

Messieri, S., “Correlation Between Ultimate Tensile Strength and Solidification Microstructre for the Sand Cast A357 Aluminum Alloy,” Materials and Design, vol. 30(10), pp. 4525-4531 (2009).

10. Shabestari, S.G., Shahri, F., “Influence of Modification, Solidification Conditions and Heat Treatment on the Microstructure and Mechanical Properties of A356 Aluminum Alloy,” Journal of Material Science, vol. 39(6), pp. 2023-2032 (2004).

11. Jeong, C.Y., Kang, C.-S., Cho, J.-I., Oh, I.H., Kim, Y.-C., “Effect of Microstructure on Mechanical Properties of A356 Casting Alloy,” International Journal of Cast Metals Research, vol. 21(1-4), pp. 193-197 (2008).

12. Davis, J.R. ed. ASM Specialty Handbook for Aluminum and Aluminum Alloys, ASM International (1993).

13. Poirier, D.R., Poirier, E.J., Heat Transfer Fundamentals for Metal Castings, Second Edition with SI Units, The Minerals, Metals and Materials Society (TMS) (1994).

14. Doutre, D., “Increasing the Production Rate of Permanent Mold Casting Through the Use of Helium Injection,” Advances in Industrial Materials, Proceedings of the International Symposium on Advances in Industrial Materials, Calgary, Alberta, pp. 289-301 (1998).

15. Doutre, D., “The Influence of Helium Injection on the Cooling Rate and Productivity of Permanent Mold Casting Process,” AFS International Conference on Permanent Mold Casting of Aluminum, 5th, Milwaukee, WI, pp. 86-96 (2000).

16. Saleem, M.Q., Helium Assisted Sand Casting of Aluminum Alloys, PhD Dissertation, Worcester Polytechnic Institute, Worcester, MA, USA (2011).

17. Griffiths, W.D., Enhanced Mechanical Properties in Al Castings by the Application of Helium, Engineering and Physical Sciences Research Council, (Grant reference: EP/C514718/1).

18. ASTM Standard B557-06, Standard Test Methods for Tension Testing Wrought & Cast Aluminum- and Magnesium-Alloy Products, Annual Book of ASTM Standards (2010).

19. ASM Handbook: Metallography and Microstructure, Volume 9, ASM International (1992).

20. ASTM Standard E112-96 (Re-approved). Standard Test Methods for Determining Average Grain Size, Annual Book of ASTM Standards (2010).

21. ASM Handbook: Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Volume 2, ASM International (1990)


Technical Review and Discussion

Helium-Assisted Sand CastingM. Saleem, University of Engineering and Technology, Lahore, PakistanM. Makhlouf, Worcester Polytechnic Institute, Worcester, MA, USA

Reviewer: Unlike permanent mold operation where the molds are tied up as a casting is freezing inside them, the speed of the sand casting process is not related to the solidi-fication time since filled molds can be allowed to collect and take as long as necessary to freeze - essentially an “accumu-lator” for in-process castings. The slowest speed of 1) ram-ming up molds, 2) pouring them, or 3) shaking them out and 4) post processing the castings could well be the limiting fac-tor on productivity. Having to run the castings through a He-lium injection station could negate the productivity benefit of a consumable mold since the time the casting is tied up at the injection station could become a potential limiting factor analogous to a PM mold, especially if the other factors are less than the enhanced solidification time, which could well be the case. A bit more in-depth discussion of these points would benefit the paper tremendously as far as practicality of application is concerned, i.e. measuring the time to cool over a specific range is not a direct measure of the produc-tivity of a plant with our without He injection. The big ad-vantage that should be stressed is the property enhancement - can a sand caster compete with a PM foundry using He?

Authors: We agree that freezing may not be the bottleneck in the process; however the remark on “productivity” is in-tended to be a general one and not aimed at the time aspect alone but considers the definition of productivity, i.e.,

;

within this definition, the benefit of enhanced properties is inherent. Typically “input” is made of a several factors (time, labor, material, space etc.), whereas “output” is the sellable product produced by the foundry. Undoubtedly, in the context of the helium-assisted process, there are some “input” entities that would be introduced, like the cost of helium, the cost of additional equipment, etc.; however these costs should be compared with the benefits that are obtained. To carry out such an analysis it was necessary that first the extent of benefits, data on the rate of helium consumption and the design of the system that would achieve those benefits be determined; this was one of the objectives of the work presented herein. Granted the helium-assisted sand casting process may not able to compete with the per-manent mold process in as much as sand casting is not able to compete with permanent mold casting; however it has its own usefulness (e.g., the possibility to introduce localized cooling of the mold from outside of the mold that could help with casting/mold design issues).

Reviewer: An explanation for the extremely high melt temperature: 880±5°C (1616±41°F) chosen should be given. Standard pouring temperatures are in the range

of 720-760C - sometimes even lower. At >800C reac-tion between Al and carbon - or SiC becomes an is-sue. Were aluminum carbides seen in the microstruc-tures? Was any loss in Mg noticed from the melts?

Authors: This high melting temperature was necessitated by our laboratory arrangement and the location of the pouring and melting stations relative to each other. We did not observe any harmful effects from the high melting tem-perature; but we certainly do not advocate melting at these temperatures.

Reviewer: The authors should include more discussion as to why the yield strength is better with parallel flow He. The mechanism is not clear. Usually yield is relative-ly insensitive to SDAS, for example. That property usu-ally depends solely on chemistry and heat treatment.

Authors: The reason for the lower yield strength in the case of cross flow helium supply is the comparatively higher level of porosity, which deteriorated the quality of the cast part to the extent of overshadowing benefits obtained from the enhanced cooling rate. The parallel flow helium supply on the other hand, resulted in a faster cooling rate, a smaller grain size, a finer SDAS, and less porosity, which explain the higher yield strength.

Reviewer: Elongation and UTS are sensitive to SDAS. The higher ductility allows the metal to deform further after yield and it work hardens more, giving a higher fi-nal UTS. In this case - in Figure 23 - the ductility is uni-formly poor though. Was porosity significant enough in the baseline samples that the reduction in cross sectional area threw off the apparent measured stress levels? Was the Mg level in the test castings consistent across all batches?

Authors: The alloy used in this work has a chemical compo-sition close to that of 319 alloy, except for a higher Mg con-tent. For such chemical composition, the values we report for the baseline castings are in reasonable agreement with the values reported in handbooks for sand cast aluminum 319-T6 alloy.

In order to ensure uniformity of the melt, a single batch of the alloy was used in all the experiments and spark emission spectrometry measurements were made prior to each pour-ing in order to guaranty that the chemistry was consistent.

Reviewer: The author references a flow rate of 1,4 or 8L of helium. I think it would be better to estimate the veloc-ity through the porous media near the casting surface. How would 1,4,or 8L work in a mold twice as large (or 1/2) the size?

Authors: We have reported the “superficial” fluid flow rate through the mold instead of the “actual” velocity near the surface of the casting. As per D’Arcy’s law2, which governs flow through porous packed beds under low pressure gra-


dients, this “superficial” fluid flow rate can be converted into “superficial velocity”; for a certain mold characteris-tics, the change in mold size would affect this “superficial” velocity only if the area of helium flow is changed.

Reviewer: Some reference is given a possible neutral-izing effect of combustion gasses of the resin in the sand, what would the authors speculate would happen if this were ever tried in green sand casting which has mois-ture present in the sand/clay mix. The authors note that sand casting is widely used, but green sand is likely the most common process of the sand casting techniques.

Authors: Griffiths17 used helium with clay bonded silica sand and an encapsulated mold and reported that the ther-mal conductivity of the mold increased by more than double when He was used. The reader is referred to reference 17 in the manuscript for more details.

Reviewers: A reference is made to recycled sand. Was this thermal or mechanically reclaimed or is the final resin con-tent known? Would a higher content of resin suppress the benefits even more?

Authors: This detail is not available. Nevertheless, we know that mold parameters, such as grain size, permeability, com-position, etc. influence the results. For this reason, mold and molding parameters were all kept constant throughout the work.

Reviewer: Higher porosity is noted in the cross flow experi-ment. How did the authors determine this was actually poros-ity from the helium and not shrinkage porosity. In conven-tional sand casting it is possible to get shrinkage between to chilled surfaces, how can we be sure this did not happen here.

Authors: The shape of the pores observed in castings made by the cross flow process was not typical of shrinkage poros-ity (which is usually much more irregular and inter-dendrit-ic). The observed pores were more or less round, and since we performed RPT measurements prior to each pouring, we know that the pores are not from hydrogen.

helium-assisted sand casting abstract

Documents