Rhenium and rhenium alloys offer attractive high-temperature pro-perties for use in aerospace applications. However, traditional methods used to make components are time consuming and have high production costs. Historically, rhenium has been produced in standard rod, bar, plate and sheet forms, and components were made from these shapes using electrical discharge machining (EDM) to produce the required shapes to specified tolerances. The method works well, but is time consuming and has a low yield. Therefore, manufacturers are continually looking for ways to improve productivity and reduce costs. Some approaches to reach these goals were discussed in several papers at PM2TEC2001, and are presented below.
CIP to NNS
A recent developments in rhenium powder metallurgy (PM) is cold isostatic pressing (CIP) to a near-net-shape (NNS) using traditional rhenium powder. Rhenium metal powder is packed into a flexible mold having a rigid container to maintain the desired shape (Fig. 1), and the can/ mold assembly is immersed in water contained within a CIP vessel. In this "wet-bag" process, hydrostatic pressure in the range of 210 to 410 MPa (30 to 59 ksi) transfers isostatic pressure to the mold, consolidating the powder into a green NNS rhenium compact.
In their paper "Advances In Powder Metallurgy Rhenium," Todd, Leonhardt and co-authors (Rhenium Alloys Inc., Elyria, Ohio) discussed the use of CIP to make high-performance bi-propellant liquid-fuel apogee rhenium-iridium thrusters under a NASA Phase II Small Business Innovative Research (SBIR) program. The thrusters were made using a collapsing-bag technique, which requires the CIP tooling to match the part shape; this requires a contour can and mold and a one- or two-part solid pressing mandrel.
Rhenium Alloys made the thrusters using a two-part pressing mandrel to produce a NNS green part. The mandrels were separated at the narrowest point of the convergent-divergent thruster cone design, and a slight taper on the mandrel allowed removal from the part without scratching it. These innovations in mandrel design help in extracting the mandrel from the green part (Fig. 2). The amount of powder fill controls the wall thickness of the part.
After removing the mandrel, the green part is lightly presintered at a low temperature while in a horizontal position to ensure that no sag will occur. The thruster is presintered a second time at a higher temperature in the vertical position to promote shrinkage prior to high temperature sintering at 80% of the melting point (m.p. = 3180?C). Sintering can be performed either in a hydrogen atmosphere or a vacuum for several hours to increase density. Sintered density was 20.4 g/cm3 (97% of theoretical). Assintered thruster dimensions were close to those required, and an additional 2-4% shrinkage during HIP put the thruster within the required dimensional tolerances. EDM of the finished part removed rough surfaces and excess material. Final dimensions were achieved using a two-step diamond grinding process.
Rhenium Alloys notes that the 440 N thruster (Fig. 3) was the first produced using the CIP-to-NNS process. The hot processing methods produced a fine-grained microstructure having microporosity inside the grains (Fig, 4); final density was 99.4% of theoretical.
Similar process parameters were used to produce a 490 N thruster having a large exit cone and smaller barrel section (Fig. 5). The part had a final density of 99.9%, with a fine grain microstructure (Figure 6). A shorter production process for smaller 100 N thrusters consisted of a light presintering cycle followed by a shorter sintering cycle. Two thrusters were HIPed simultaneously, reducing both the production time and cost.
Benefits of using this production method to make these parts include a reduction in manufacturing costs of as much as 35%, a 30-40% reduction in manufacturing time, a 70% reduction in the amount of powder required and a 50% reduction in machining requirements.
CIP to NNS also was used to make 38-mm (1.5-in.) radius hemispherical domes (Fig. 7) using a one-piece mandrel. Final wall thickness of the domes is 4 mm (0.157 in.).
Improved powders for advanced NNS methods
Advanced techniques such as vacuum plasma spraying (VPS), direct-hot isostatic pressing of powder (D-HIP), directed light fabrication (DLF) and metal injection molding (MIM) offer the ability to produce NNS parts in large quantities at lower cost and faster turn-around times using less material. However, these methods require high-density, low-oxygen, spherical rhenium powders in a wide range particle sizes.
Conventional rhenium powder is an irregularly shaped flake powder having poor flow characteristics and high oxygen concentration (~1,000 ppm). Rhenium Alloys produces a PM grade of rhenium having a purity of 99.99%; the -200 mesh powder has an average 3.5 um particle size, a 1.84 g/cm3 apparent density and a 3.03 g/cm3 tap density. The company engaged in a Ballistic Missile Defence Organization (BMDO) Phase I SBIR project to produce a spherical rhenium powder (SReP) having low-oxygen content (<50 ppm), an 11 g/cm3 apparent density, a 12.5 g/cm3 or higher tap density and high flow characteristics (4 s/50 g flow rate). The goal was to produce spherical rhenium powder (SReP) using the rotating electrode/plasma rotating electrode gas-assisted process (REP/GA-REP/PREP). The process consists of a rhenium rod rotating at 15,000 rpm in an argon atmosphere, while high-velocity plasma torch melts the rhenium into droplets. Typical REP particle size is 200 to 600 Km having a bimodal distribution.
The majority of SReP produced by Rhenium Alloys using the REP have diameters ranging from 100 to 300 Km with similar tap and apparent densities, typical for most spherical metal powders. To achieve higher quality powders, the company conducted in-house research, developing a plasma-atomization (PA) process for SReP. The process produces fine particles less than 40 Km in diameter and can produce a size range of 5 to 80 Km. Three plasma-atomized SReP types (Types A, B and C) having average particle sizes of 37 Km, 25 Km and 11 Km, respectively, are shown in Fig. 8, 9 and 10. Apparent and tap densities of the three powder types are similar to those of REP powders (around 10-11 and 12-13 g/cm3, respectively). The PA process also was used to produce Mo-Re and W-Re alloy SReP.
The advantages of using PA-SReP was demonstrated in several advanced consolidation methods including vacuum plasma spraying of type A SReP powder to produce a flat coupon and tube to a maximum thickness of 5 mm (0.2 in.). Type B SReP was made into a 19-mm (0.75 in.) rod using the direct HIP process to achieve a 98% density.
Modeling CIP compaction and sintering
Modeling is used as a cost-effective tool to improve and optimize manufacturing pro-cesses and to conduct parametric studies. For example, in PM, compaction models are used to predict stresses, density distribution, loads and other parameters. While studies of die compaction of powder metals have been conducted, few studies deal with CIP, and there reportedly is a lack of information related to the densification of rhenium powder. This lack of data has been addressed in work at Concurrent Technologies Corp. (Johnstown, Pa.)
"Modeling The Compaction and Sintering of Rhenium," co-authored by Murali Pand-heeradi and colleagues at CTC and Todd Leonhardt (Rhenium Alloys Inc.) reviewed the development of a finite element model to simulate and study the densification behavior of Re powder during cold die compaction and CIP. The simulations were carried out using PCS EliteT and ABAQUST (Hibbitt, Karlsson & Sorensen Inc.) with a user-defined material model. The model was validated by comparing predicted stresses and densities with results obtained from die compaction tests on rhenium.
This work also suggested a methodology for a sintering model for rhenium based on the concept of master sintering curves (MSC), which uses the density distribution predicted by the compaction model as input. A master sintering curve enables prediction of densification behavior under arbitrary time-temperature excursions. The master sintering curve is sensitive to factors such as starting microstructure of the powder, green-body forming process, the dominant diffusion mechanism and sintering heating method. This requires a different MSC if any of the factors change before it can be used to predict sintering behavior under the modified conditions. Experiments to generate data for the MSCs are ongoing.
Powder injection molding
Another consolidation technique of high interest to fabricate small, high-performance rhenium components is powder injection molding (PIM). PIM offers the ability to form intricate shapes to isotropic, near-full density without the need for much secondary machining.
Developmental work to determine optimum processing parameters for spherical rhenium powder is not feasible due to the high powder cost, especially in a large commercial machine that requires a large amount of material for mixing with binder and molding. To overcome this obstacle for further rhenium PIM developmental work, Concurrent Technologies Corp. developed specially designed methods to handle small amounts of feedstock, and to use tungsten powder to test mixing and molding processes before performing trials using spherical rhenium powder.
C.M. Wang and co-authors discuss in their paper "Powder Injection Molding To Fabricate Tungsten and Rhenium Components" some computer tools developed to assist in designing the PIM process including a simple flow-network model to correlate capillary rheometry test data with pressure and flow rate during molding of actual parts. Simulation analysis also is used to define molding pressure, gating size and molding temperatures.
Materials used in the tests included 1.5 Km commercially pure (99.99%) tungsten powder and wax-polymer base binder consisting of paraffin wax, polyolefin, carnuba wax and stearic acid. Compared with spherical rhenium powders for PIM, which are about 5 Km and have good flow characteristics, tungsten particles have an irregular, angular shape and tend to agglomerate. However, tungsten has a similar density and melting point to those of rhenium, and is less expensive.
The optimal powder loading in the molding mix must be known before molding parts (optimal loading is the maximum powder loading that can be molded without forming defects in the final parts). The CTC method uses only 100 cm3 of mix in the mixer to measure viscosity as a function of powder loading, shear rate and temperature. Extruded material and unused feedstock are collected after measurements are taken and are used again with a new powder loading. A computer program was developed to help determine the amount of material to add for a given powder loading in the existing feedstock to obtain a new powder loading.
Another consolidation technique of high interest to fabricate small, high-performance rhenium components is powder injection molding (PIM). PIM offers the ability to form intricate shapes to isotropic, near-full density without the need for much secondary machining.
Developmental work to determine optimum processing parameters for spherical rhenium powder is not feasible due to the high powder cost, especially in a large commercial machine that requires a large amount of material for mixing with binder and molding. To overcome this obstacle for further rhenium PIM developmental work, Concurrent Technologies Corp. developed specially designed methods to handle small amounts of feedstock, and to use tungsten powder to test mixing and molding processes before performing trials using spherical rhenium powder.
C.M. Wang and co-authors discuss in their paper "Powder Injection Molding To Fabricate Tungsten and Rhenium Components" some computer tools developed to assist in designing the PIM process including a simple flow-network model to correlate capillary rheometry test data with pressure and flow rate during molding of actual parts. Simulation analysis also is used to define molding pressure, gating size and molding temperatures.
Materials used in the tests included 1.5 Km commercially pure (99.99%) tungsten powder and wax-polymer base binder consisting of paraffin wax, polyolefin, carnuba wax and stearic acid. Compared with spherical rhenium powders for PIM, which are about 5 Km and have good flow characteristics, tungsten particles have an irregular, angular shape and tend to agglomerate. However, tungsten has a similar density and melting point to those of rhenium, and is less expensive.
The optimal powder loading in the molding mix must be known before molding parts (optimal loading is the maximum powder loading that can be molded without forming defects in the final parts). The CTC method uses only 100 cm3 of mix in the mixer to measure viscosity as a function of powder loading, shear rate and temperature. Extruded material and unused feedstock are collected after measurements are taken and are used again with a new powder loading. A computer program was developed to help determine the amount of material to add for a given powder loading in the existing feedstock to obtain a new powder loading.
Correlating data to actual molding machines
Viscosity cannot be directly used to determine required molding pressure for actual molding, so researchers developed a computer simulation model (flow-network model) to predict pressure and flow rate for a given molding system. Individual sections of the molding system including the barrel, runner, ingate and die cavity are connected to form a network. The pressure at the two ends, the flow rate in each section and the solidification in each section are calculated by solving a set of simultaneous equations derived by combining pressure-flow relations with the mass balance requirements at the junction points.
Pressures and flow rates in capillary rheometry tests are calculated using a flow-network model (Fig. 11). Section dimensions are rectangular (H x W x L) and cylindrical (D x L). The objective is to find viscosity constants to estimate the molding pressure required on the actual molding machine, such as a bench-top device used to mold a tensile specimen (Fig. 12). Figure 13 shows the flow-network model of the bench-top machine. Maximum melt pressure in the barrel (P0) is set at 2,500 psi (17 MPa) to simulate the actual limit of molding pressure of the bench-top machine.
Calculated results for molding 55% powder loading feedstock are shown in Fig. 14. Ram velocity and filling rate decrease when the barrel pressure reaches the 2,500 psi maximum. Ram velocity quickly decreases to a value so small as melt pressure reaches maximum that the mold cannot be filled before the ingate is frozen off. In actual molding, the bench-top machine cannot mold feedstock having a powder loading greater than 53%, which is verified by the model prediction.
For a large commercial injection machine (110-ton) having a maximum melt pressure of 16 ksi (110 MPa), the ram velocity drops when flow reaches the ingate (the point of greatest flow resistance). Therefore, critical loading is where flow velocity can be maintained in the thinnest section-the ingate for this example. Calculations show that optimal loading is 56%, which is verified in a uniformly filled, molded green tensile specimen, which is defect free after debinding.
This article is based on papers presented at PM2TEC2001. Proceedings available from Metal Powder Industries Federation, Princeton, N.J.; tel: 609-452-7700.
Report Abusive Comment