The cycle time for conventional debinding and sintering has been dramatically reduced by the creation of laminar gas flow inside a controlled retort. Adding controlled plasma before laminar gas-flow sintering enables this furnace to produce parts in half the time than conventional furnaces offer today.



Plasma Assisted Debinding and Sintering (PADS) was developed at the Materials Laboratory (MatLab), Department of Mechanical Engineering, Federal University of Santa Catarina in Brazil and resulted in patents assigned to Brazil’s Lupatech S/A, which have been described elsewhere.[1,2,3] A prototype furnace, capable of producing about 35 kg per run, is currently being used in Steelinject Ltda., a Lupatech S/A Company, in Brazil. This work discusses briefly how PADS works. Parts made by PADS were compared with those made conventionally. The advantages of the PADS system and the potential to scale up this process to commercial-sized furnaces have been discussed. This paper discusses preliminary efforts to construct a commercial version of this furnace.

Fig. 1. Schematic of abnormal DC glow discharge (Courtesy MatLab)

The PADS System

Plasma can be defined as a partially or wholly ionized gas with a roughly equal number of positively and negatively charged particles. There are two types of plasma: high-temperature plasma and low-temperature plasma. The PADS uses the low-temperature plasma type, generated by abnormal DC glow discharge. Figure 1 shows it schematically. The chamber is at a vacuum level of 1-5 mbar and has gas species that is ionized by the potential difference between anode (grounded) and cathode (biased on negative voltage). So, the electrons generated in the plasma are attracted and move toward the anode, while a part of these electrons move toward the plasma floating potential, whereas the positive-charged ions move toward the cathode. In the PADS system the gas species used was a mixture of argon and hydrogen. This results in the following scenario:
  • The ionized and the fast neutral atoms bombard the cathode, resulting in heat generation and sputtering.
  • Inelastic collision of electrons with gaseous species in the plasma environment results in chemical reactions like dissociation of the gas molecules and other molecular compounds, such as polymers.
  • The electrons generated in the plasma move toward the anode and floating potential.
In the case of debinding in the plasma environment, electrons, via inelastic collisions, transfer enough energy to the polymer molecules to cause their dissociation into small hydrocarbon radicals (CxHy). The free hydrogen ions from the hydrogen introduced into the chamber will readily cap off the open carbon atoms in the broken chains to form small gaseous molecules such as methane, ethane, propane, etc. These gases are easily swept away by the vacuum pump.

The collision of electrons occurs not only with the polymer molecules present in the vapor phase but also on the surface of the MIM parts, causing the binder on the outer layer of the parts to be removed (before and after binder is molten). As a consequence of the binder removal via dissociation on the outer layer of the part, fresh molten binder quantities are driven from the interior toward the surface by capillary action. Because the plasma-assisted process (PADS) does not wait until the binder evaporates, the process is faster than the conventional process. A more detailed explanation is available in the U.S. Patent 6,579,493,[4] and the equipment is described in other patents.[5,6]

Fig. 2. Process trend in the Elnik MIM 3045 furnace

Experiments

The purpose of the experiments performed was to compare the properties by eliminating the variables introduced by the two different furnaces – the prototype PADS furnace and the conventional Elnik MIM 3045 furnace – working in the principles of laminar gas flow at a partial pressure of 400 mbar.

The materials used in this experiment were 17-4 PH feedstock produced at Steelinject Ltda. from which Steelinject molded tensile bars. Steelinject then sent to DSH Technologies: green molded tensile bars; bars solvent debound to remove the primary binders followed by secondary debinding and pre-sintered to 900°C (1652°F) in the PADS system at Steelinject; and bars debound and sintered at 1370°C (2500°F) in the prototype PADS furnace at Steelinject.

The green parts were solvent debound in a DSH/Elnik SD 3045 solvent debinding unit in perchloroethylene. A number of parts were then pre-sintered at 900°C after removing the secondary binders in the Elnik furnace. The remainder of the parts was sintered in the Elnik furnace at 1370°C. About 500 pre-sintered and sintered pieces were returned to Steelinject for further testing. Steelinject then did their own comparison between the PADS and conventional sintering processes (Fig 2).

Carbon testing was done on the pre-sintered samples made at DSH and Steelinject to ensure that the parts being processed had reached a similar baseline. The following samples were then tested at the DSH/Elnik facilities:

1. Debound and sintered at DSH, (Lot 112107)

2. Debound, pre-sintered at DSH, (Lot 112007) and then sintered at DSH (together with the samples in Lot 112107)

3. Debound and pre-sintered at Steelinject, and then sintered at DSH (together with the samples in Lot 112107)

4. Sintered at Steelinject, and then re-sintered at DSH (together with the samples in Lot 112107)

5. Debound and sintered at Steelinject in the as-sintered condition

The carbon content, density, tensile strength, yield strength and percent elongation were measured for each of the above samples.

Results

The parts pre-sintered at DSH averaged a carbon content of 0.069%, while those pre-sintered at Steelinject showed 0.082%. Based on the carbon levels of the sintered parts we concluded that the difference was not significant.

Table 1 summarizes the results obtained after sintering. There is no significant difference between Samples 1, 2 and 3, which shows that the final sintered product is no different whether debinding is conducted by the traditional method or by PADS. Sample 4 is the same as the parts debound and sintered at Steelinject (Sample 5) except that this sample was re-sintered at DSH with the other samples and showed a nominal, if any, difference compared to the parts sintered by the PADS.

Fig. 3. The unique gas-management system creating laminar gas flow

Discussions

The PADS prototype furnace and the Elnik MIM 3045 work in different ways. The intent of this experiment was to eliminate the differences caused by the different processing and evaluate the effect of the plasma debinding on the parts. The difference in how the two furnaces work – the PADS at 1 mbar in a 95% hydrogen and 5% argon mixture and the Elnik at 400 mbar using 100% hydrogen – could result in different sintered properties. It is not the intent to show which furnace results in better sintering, but whether the PADS is an effective debinding process.

Figure 2 shows the furnace trend for debinding and sintering in the Elnik furnace. It has a number of holds to eliminate the secondary binders and a sintering hold. The total process time, which was not optimized and included cooling time, was about 18 hours.

Figure 3 shows the unique laminar gas flow used in the Elnik furnace systems. During debinding, this laminar flow causes the binders to be removed from the furnace chamber and allows these to be collected in the binder trap. Figure 4 shows the collected binders inside the trap.

Fig. 4. Binders collected in the binder trap when debinding is done under laminar flow

Figure 5 shows the profile for the PADS furnace. Typical processing time inside the PADS prototype furnace is about 10 hours plus the cooling time.

It is thus seen from the experiments that while there is no significant difference between parts debound conventionally or debound by the PADS after sintering, PADS has the potential to reduce the cycle time significantly, because in the PADS process debinding takes place while the parts are ramping toward the sintering temperature.

Also, during debinding the binder chains are cut by the impinging electrons, and the open carbon atoms easily attach themselves to the free hydrogen ions in the plasma forming small-molecule hydrocarbon gases such as methane, ethane, propane, etc. (as explained earlier), which are burned at the exhaust. Hence, the PADS furnace is capable of producing a cleaner process where no time needs to be spent to clean out the binder residues. There are no solid binder residues to dispose. Therefore, the process is also environmentally friendly.

Figure 6 shows the effect of debinding using the plasma. Note that there are no binder residues visible in the furnace. Processing with plasma also has the potential of saving about half a day’s worth of processing time for each binder clean-up cycle.

Fig. 5. PADS prototype furnace profile (Courtesy MatLab)

The PADS prototype furnace has the following characteristics:
  • Processing is done at 1-5 mbar for all parts for the debinding step and 100 mbar from the end of the debinding until the end of the sintering step. This may not result in the optimum sintering conditions for all materials.
  • The present prototype furnace has a capacity to produce only 35 kg per run.
  • While the prototype pilot furnace is an excellent research tool, it is not a commercial furnace. It requires many features to be added for worker safety and is difficult to scale up to 200-400 kg per run.
Elnik Systems has taken up the challenge and is working with Lupatech S.A. to commercialize a furnace that incorporates the plasma-assisted debinding with a conventional MIM furnace called PlasMIM. The new furnaces reduce the processing time to about 12 hours per run.

The PlasMIM furnaces which combine plasma-assisted debinding with sintering the parts by the standard laminar gas-flow process under partial pressure are being produced to manufacture 200-400 kg parts per run.

Fig. 6. The PADS prototype furnace after sintering when the plasma discharge was used for debinding. Note the absence of binder residues. (Courtesy Steelinject)

Figure 7 shows the racking system (where the parts will be placed) with the anode, the cathode, the heating elements and the hot zone of the PlasMIM furnace. The parts will be placed on ceramic trays on the racks, which have a total of 1.86 square meters of loading area in the present configuration.

Experimentation using the PlasMIM furnace is proceeding at this time to work out the final details to make production runs with the capacities and speedy debinding as mentioned earlier. Figure 8 shows plasma generated inside the PlasMIM furnace.

Fig. 7. The racking system, the anode, the cathode and the hot zone of the PlasMIM furnace

Summary

  • Plasma-assisted debinding results in a clean furnace with no messy binders to be trapped.
  • Plasma-assisted debinding could result in reducing the total processing time to less than 12 hours per run.
  • A new furnace, PlasMIM, with plasma-assisted debinding that will sinter the parts by laminar gas flow under partial pressure, is being produced to manufacture 200-400 kg parts per run. IH


Fig. 8. Plasma inside the PlasMIM furnace

Acknowledgements

We would like to thank Aloisio N. Klein and Paulo A. P. Wendhausen of the Materials Laboratory (LabMat), Department of Mechanical Engineering, Federal University of Santa Catarina, Florianopolis, SC, Brazil, and Ricardo Machado and Waldyr Ristow of Steelinject (a Lupatech S/A Company) and Nestor Perini of Lupatech S/A., Caxias do Sul, RS, Brazil for their support for this work.

For more information: Claus J. Joens is president of Elnik Systems, Div. of PVA MIMtech, LLC, 107 Commerce Road, Cedar Grove, NJ 07009, USA. Tel: 973-239-6066 ext 12; fax: 973-239-3272; e-mail: cjoens@elnik.com; web: www.elnik.com

Additional related information may be found by searching for these (and other) key words/terms via BNP Media SEARCH at www.industrialheating.com: plasma, debinding, ionized, sinter, laminar gas flow, anode, cathode