A series of tests shows the impact of carbon-based coatings in aluminum diecasting operations.

A MODERN CASTING Staff Report

(Click here to see the story as it appears in February's Modern Casting.)

Die castings produced with as-cast or nearly as-cast surfaces are becoming more common, where near-surface layers of the castings are machined. As a result, various problems related to the boundary between chill layers and general internal structures are more apparent. Controlling heat insulation during molten metal infusion, heat transfer during pressurization and thermal diffusion to the mold are critical to diecasting quality. Proper control of these variables will lead to dense cast structures free of chill boundaries.

In the paper “Key Issues for Quality Stabilization of Aluminum Die Castings,” Yuichi Furukawa, Toyota Motor Corp., Toyota, Japan, and Yoshiki Tsunekawa, Toyota Technical Institute, Nagoya, Japan, investigated aluminum diecasting processes. The report includes the results of a durability test of various carbon-based coatings, the effects on solidification structure and a coating’s ability to stabilize diecasting quality.

Question

How does a carbon-based coating in an aluminum diecasting operation affect surface roughness, molten metal flow, solidification structure and casting quality?

1. Background

In the diecasting process, molten metal fills permanent molds at high speeds and is unlikely to penetrate delicate concave shapes inside the mold surface. Once the mold is filled, high pressure is applied to the molten metal, at which time it fills the concaved areas. The concept of the casting interface having both heat insulation and transfer properties is shown in Fig.1.

In a cold chamber diecasting method, metal is poured into cold sleeves, which create initial solidification layers before cavities are filled entirely. The molten metal becomes solid as soon as it touches a mold, forming a chill layer that has better mechanical properties than the casting’s internal structure. Solidification layers, which can become evident during near-surface machining, degrade mechanical properties and impact fluidity.

Conversely, proper hardness cannot be achieved in a heat-insulated condition (which does not create chill layers) or when castings are removed from a mold that is still at a high temperature. Galling (i.e., wear caused by adhesion between sliding surfaces) also might occur in these situations.

To produce high quality castings consistently, the boundaries between chill layers and the casting’s internal structure must be eliminated. To do this, insulation is necessary during the infusion of molten metal to create a casting interface, which transfers heat during solidification.

The study included a method that allows the metal to remain non-wetting during infusion, but then becomes wetting when pressure is added to create a casting interface that transfers heat. Carbon has poor wettability with molten aluminum but has high lubricity. It long has been used as a parting agent. However, because the application of carbon can negatively affect its working environment, it is used only as a plunger lubricant.

For this study, a carbon-based coating resilient in an aluminum diecasting environment, where casting pressure is added, was tested. The report includes durability tests of carbon nanofiber, carbon-graphite and carbon-fullerene coatings and their effects on solidification structure and casting quality.

2. Procedure

When the unevenness of a mold’s surface is significant, aluminum can flow into concave parts and lead to galling defects. Fig. 2 shows a testing setup to evaluate metal flow by mold surface roughness and carbon materials. It consisted of an electric furnace and a 7.9 × 7.9 × 1.2 in. (200 × 200 × 30 mm) sample mold. The test piece mold was tilted 20 degrees to create a slope for metal flow. All coated surfaces of the molds were recorded with a thermo camera to study the heat transfer condition. The condition of metal flow was classified into four grades and each was indexed. The sum of indexes, obtained by pouring molten metal three times, was defined as the metal flow index.

The metal flow test results are shown in Fig. 3. For all types of carbon coatings, metal flow was not smooth on a mirror surface, that is, Ra<1.0 μm. Any type of carbon coating showed a good result at Ra=1.5 μm or above. Infrared images taken by a thermo camera indicated that immediately after metal was infused, the temperature of the test piece surface was within the range of 167F (75C) and 181F (83C) for the case of a metal flow index score of 3 or less; between 143F (63C) and 158F (70C) for a score between 5 to 7; and 122F (50C) or less for a score of 8 or over. In this way, a relationship between the heat transfer at metal flow and the metal flow property was confirmed. The indicated temperatures were corrected using the infrared emissivity that had been measured in advance. The emissivity of the carbon nanofiber, carbon-graphite and carbon-fullerene coatings were 0.9 or higher, which proved the temperature measurements were accurate.

Figure 4 shows the experimental setup for a durability test. A given carbon material was coated on the mold surface and a ring-shaped die body was placed on it. No parting agent was used. A total of (44 lbs.; 20kg) aluminum was melted, and 1.1 lbs. (480g) was poured into the ring. A weight of 21.6 lb (9.8 kg) was placed above the casting. The weight, the casting and the die body then were horizontally pulled at once. When they started moving, it was defined as the mold release force. If nothing moved, it was defined that they were stuck.

The test results in Fig. 5 show the CF15 coating, which is the combination of C15 and fullerene coatings, displayed significantly superior durability.

3. Results and Conclusions

Differential thermal analysis was conducted for  carbon nanofiber, fullerene and carbon-fullerene coating in air and nitrogen atmospheres. The thermogravimetry scans of coatings are shown in Fig. 6. In an air atmosphere, the carbon nanofiber coating suddenly dropped in the temperature range of 752F (400C) and 1,112F (600C). It kept reducing until 50%wt. This is because cutting powder of the carbon nanofiber coating was used as a differential thermal analysis sample, and it contained the compound layer that was chemically combined with the test piece mold.

The weights of the fullerene and carbon-fullerene coatings do not decrease until 752F (400C), but then show sharp declines. The fullerene coating starts resolving at 752F (400C) and easily vaporizes into CO and CO2. Compared to other types of carbon materials, fullerene oxidizes easily. However, based on the analysis results, at 752F (400C) or lower, the fullerene coating can control the high temperature degradation of the carbon nanofiber coating. In a nitrogen atmosphere, the fullerene coating remains stable and is not dissolved until 1,652F (900C). The thermogravimetry scan of the carbon nanofiber coating in a nitrogen atmosphere shows a gradual decrease in weight as the temperature increases. Conversely, fullerene and carbon-fullerene coatings demonstrate lower weight decline ratios compared to that of the carbon nanofiber coating up to 1,202F (650C). It is possible that the fullerene coating has a protective effect on the carbon nanofiber coating as observed in an air atmosphere.

During the diecasting process, oxygen in the cavities is expected to be consumed by aluminum, which has a fast oxidization rate, and a nonoxidative atmosphere is created in the cavities. Setting the temperature of molten metal at 1,202F (650C) or lower and the temperature of the mold surface during release at 752F (400C) or lower prevents the high temperature degradation of the coating.

A diecasting test was conducted by using molds with various carbon coatings. The temperature of molten metal was 1,184F (640C). The Scanning Electron Microscopy (SEM) micrographs of a core pin before and after 100 shots are shown in Fig. 6. The before and after comparison of the pins shows the surface and cross section after use are denser than before use.

The density of the carbon nanofiber coating was low compared to that of the carbon-fullerene coating. When molten aluminum is poured into a die, it penetrates and removes the carbon nanofiber coating. By repeating the process with a carbon-fullerene-coated mold, consideration was given to the reasons for the carbonizing of the mold surface when the molten metal used for diecasting was 1,184F (640C), which is lower than ordinary carbonizing temperature. In aluminum diecasting, molten metal immediately absorbs oxygen as it seals and heats the cavity. As a result, it leaves no space for sublimed fullerene to escape, and the fullerene penetrates into the ferrous mold and diffuses. This is fullerene’s low temperature carburization effect when it is sealed and sublimated.

To study the changes in carbon crystal structure, the Raman spectroscopy analysis for the carbon-fullerene coating was performed before and after casting. The black area of the surface indicated in Figs. 7a and 7d were an amorphous carbon state. Analysis of the gray area in the same images showed a different structure. Before casting, the main components in the area were C60 fullerene and C70 fullerene, but fullerene could not be found after use. The peak wave numbers of G band were 1,339cm-1 and 1,599cm-1, meaning they were in the state of amorphous carbon. The fullerene might have disappeared, but, as shown in Fig. 8, the carbon-fullerene coating after use was denser, which means it became dense amorphous carbon after repeated heating.

To study the heat transfer properties of the coatings, the temperature of the mold and molten metal in a cavity with and without a carbon-fullerene coating was measured. Figure 9 shows the schematic diagram of the model cavity for the heat transfer measurements. The metal temperature and the mold temperature were each measured at two spots. The temperature of molten metal was measured at Tm1 and Tm2. The mold temperatures were taken at Td1 and Td2, at the same height as Tm1 and Tm2 and 2mm from the cavity surface.

The Td1 profiles in Fig. 10a from zero to six seconds after injection show the increase in the mold temperature when the carbon-fullerene coating was applied. The carbon-fullerene coating mold appears to have a high heat transfer property. However, the Tm1 (Fig. 10b) of the carbon-fullerene coating shows the temperature of the molten metal that just reached the cavity is higher and the cooling rate after the injection is slower. After solidification is complete, the carbon-fullerene coating shows a temperature change delayed by about one second. This delay lowers the hardness of the casting surface at the time of mold release and could lead to galling. At the thick part, it is necessary to improve cooling in addition to using a carbon-fullerene coating.

When there is no carbon-fullerene coating, the temperature of the molten metal at the measuring point was lower, and it also demonstrated undercooling. Since heat from the molten metal is absorbed by the mold before it reaches the measuring point, the temperature at the measuring spot becomes lower. The heat also is used for reheating and/or remelting solidified layers during the filling process. Therefore, when the infusion of the molten metal starts, the mold temperature at the lower part of the cavity is higher if there is no carbon-fullerene coating because the heat is transferred during the filling process. Based on the temperature histories at the lower cavity area, Td2 and Tm2, the molten metal temperatures for cases with and without a carbon-fullerene coating were the same.

However, the temperature of the mold without a carbon-fullerene coating increased by up to 202F (68C). When no coating was applied, the metal cooled at the lower part of the cavity. With the carbon-fullerene coating, heat insulation during infusion and heat transfer during pressurization were achieved.

For oil control valves and valve bodies with complicated oil grooves, foreign matter in the oil grooves is not acceptable. Therefore, when chipping and peeling occur during processing, they should be removed by hand. Then, thorough inspection and cleaning are necessary. By applying carbon-fullerene coating to these parts, the microstructure is improved. Figure 11 shows the cross-section micrographs of a cast component. The component without carbon-fullerene coating (Fig. 11a) has a chill layer about 30 μm thick. Beyond the chill layer, the size of α-Al grain becomes larger and a clear chill boundary is confirmed.

The diameter of α-Al grain size, found on the surface of the cross-section of cast parts produced with a carbon-fullerene coated mold (Fig. 11b), is the same as Fig. 11a. There are no boundaries and the grain sizes show gradual microstructural changes, which drastically reduces chipping and peeling. Additionally, the reduction in blow-hole and galling defects was confirmed. Chipping, peeling and galling are eliminated by obtaining this type of cross-section microstructure.

Aluminum diecasters must avoid solidification during molten metal infusion as much as possible to improve the pressure propagation at the completion of filling. By doing so, due to the physical property of a cavity surface that promotes heat transfer to the mold, a microstructure without boundaries between chill layers and the general internal structure can be obtained. When a carbon-fullerene coating was applied on the mold surface to achieve the above requirements, it improved the strength of the surface layer. As a result, various effects such as the prevention of galling at mold release and the reduction of peeling during machining can be obtained.

This article is based on a paper (13-1243) published in AFS Transactions 2013.