Big-part metal molding takes off—and pays off

Who says metal injection molding, or MIM, is only good for small parts? While the vast majority of MIM components do weigh less than an ounce, advances in materials and processing technology now allow parts whose weight is best given in pounds. The biggest of these large MIM applications, at least so far, comes from Honeywell International Inc. Working with Polymer Technologies Inc. (PTI), a Clifton, NJ, injection molder of metals and plastics, Honeywell has developed a MIM flowbody for jet engines. Essentially a housing for a butterfly valve that handles hot engine gasses, this roughly cylindrical flowbody weighs in at just under four pounds and has an inside diameter of 3.5 inches. "It's easily the largest commercial MIM part today," says Jerry LaSalle, Polymer Technologies' MIM operations director.

To prove out the concept of MIM jet engine parts, Honeywell engineers decided to revamp a successful investment-cast flowbody design that had long been used in commercial aircraft. Outwardly, the MIM flowbody almost exactly duplicates the original cast flowbody, right down to the nickel-based superalloy, Inconel 718 (IN718), used for both versions. "The only thing we changed was our manufacturing process," says Jimmy Lu, Honeywell's principle engineer and program manager on the flowbody project. He makes it sound easy, but it took Honeywell and Polymer Technologies engineers more than two years to develop the MIM process to the point that it could produce such a large part that could meet the aerospace industry's exacting quality standards.

Why MIM? Since the MIM flowbody represents a drop-in replacement for the investment cast original, nothing about the design of the finished flowbody outwardly changed. Yet Honeywell still had plenty of incentive to give MIM a try. "Investment casting is just too expensive," Lu says. Because of the dimensional tolerance limitations of casting, thin wall aerospace components like the flowbody usually start out with plenty of extra material and reach their final dimensions only after heavy finish machining. In the case of a cast flowbody, for example, machining accounts for a whopping half of the manufacturing cost, Lu reports.

And the use of IN718 only makes matters worse. Honeywell needs the superalloy because the flowbody goes into what Lu describes as a "challenging environment"-in which operating temperatures at the inlet hit 940F and the gas pressure reaches 277 psi. With its strength and retention of mechanical properties at high heat, IN718 excels in these conditions, but the material is costly both to buy and to machine. And since a relatively large portion of the casting lost to finish machining, the raw-material price looms large. What's more, the machining process itself can be arduous. "IN718 work hardens like crazy when you try to machine it," LaSalle notes.

MIM, by contrast, gets close to the net shape of the flowbodies without machining. Lu reports that a few critical sections of the MIM component-those involving flanges or mating surfaces-do feature "minor thickening" to leave some metal stock for light machining to final dimensions. "But it's only a small amount compared to casting," Lu says. For an idea of just how much extra metal the cast version requires, consider that the sintered flowbody weighs, by LaSalle's calculations, at least 50% less than the comparable casting does before machining. So even taking increased tooling costs of MIM and the remaining machining operations into account, Lu says that MIM will significantly reduce overall manufacturing costs. He declined to say by exactly how much, but remember, he noted, machining alone represents half of the cast version's cost.

The external similarities of the cast and MIM flowbodies conceal another advantage of metal molding. According to LaSalle, MIM produces a "finer and more uniform grain" than casting, and that improvement in microstructure translates to mechanical property improvements. Honeywell's mechanical testing of MIM, cast, and wrought properties of IN718 bear out LaSalle's explanation. "Our preliminary data shows tensile properties slightly better than cast," Lu says. "But the fatigue properties are much better." Looking at stress levels to failure at one million cycles, the cast IN718 made it 50 ksi while the MIM part reaches 68 ksi.

The process. Despite the advantages that MIM can impart, it's not surprising that most MIM parts have stayed small. Applying the benefits of MIM to a large, tight tolerance part required a lengthy fine-tuning of the entire manufacturing process-from molding through sintering. More than just a matter of sheer size, the hollow flowbody's thin walls (from 0.180 to 0.200 inch thick) presented mold-filling challenges. PTI and Honeywell engineers avoided these filling difficulties with computer simulations using Moldflow 3D (Moldflow Corp., Wayland, MA). These simulations identified the optimal gate locations and process conditions. And even if the hollow part does mold well, it also has to avoid distortion and cracking as it makes its way through a furnace for debinding and sintering steps.

Several enabling technologies made the job possible, starting with the right MIM feedstock. PTI used the PowderFlow system, which another division of Honeywell invented and which relies on a water-based agar binder rather than the waxes that most other systems use. According to LaSalle, this water-based system not only has the high flow characteristics needed to fill the large thin wall molds but also permits higher solids loading than wax-based systems. Higher solids limit shrinkage and, thus, the potential for distortion, he notes.

With the right binder in place, PTI and Honeywell engineers next turned their attention to sintering development. LaSalle reports that IN718 needs "all sorts of special handling" related to high vacuum levels, gas flow, and furnace temperatures. "We're not talking about an Easy Bake oven here," he jokes. Honeywell and PTI also came up with proprietary fixturing techniques that help keep the parts distortion free as they go from their green "as-molded" state to sintered-during which time they shrink by about 16%. The fixturing, for which Honeywell has applied for patents, proved crucial in combating the distortion caused by friction between the shrinking part and its resting plate during sintering, according to Lu.

Honeywell and PTI finally pushed the envelope when it came to use of IN718 in the first place. Though PTI and other molders have reported work with a variety of metal alloys-including cobalt, chrome, and heavy tungsten alloys-the vast majority of MIM applications still involve mild or stainless steels. "We tried 15 or 20 molders when we first started the project and none of them wanted to work with IN718 or a part this size," Lu recalls.

Bigger and better. Though it's currently in production at PTI, the MIM flowbody still has to make its way through some final testing and customer approvals before it can take to the air. Lu expects a quick acceptance, though, especially since MIM only adds strength and removes cost from a proven design. With this flowbody under their belts, Honeywell engineers have already started to evaluate where else MIM might work. For example, Honeywell already makes some corrosion resistant flowbodies from investment cast 347 stainless steel. "We're evaluating using metal injection molded 316L to replace the 347," Lu says. The company's engineers are looking at MIM for other flowbody components, such as valve plates. "Once you establish a robust process for your most challenging part, in our case the flowbody housing, everything else is straightforward," he says. MIM's enhancements to fatigue performance could also pay dividends in the future as Honeywell engineers come up with lighter, thinner flowbody designs that still offer improved fatigue strength.

LaSalle says that many other aerospace and industrial parts could also make sense in MIM. Like the flowbody, the best candidates will combine challenging end-use requirements, costly materials, and the need for heavy machining. "The biggest potential for the technology is to do things that haven't been done before," LaSalle says.

And now, size has become less of an issue. LaSalle believes MIM can turn out parts that dwarf even Honeywell's flowbody, and he notes that PTI has been working on one such part that weighs about 7 lbs and has a 10-inch diameter. "The sky's the limit," he says.