What Are the Properties of Feedstock in Metal Injection Molding?
Last year we helped a medical device client push a 17-4PH housing down to 0.35 mm wall thickness. Took us three tooling iterations and about 200 kg of scrapped feedstock before we nailed the flow balance. The thing most engineers don't realize until they're knee-deep in a MIM project: feedstock isn't just "metal powder mixed with plastic." It's the single variable that will make or break your part, your timeline, and probably your sanity.
So let's skip the textbook definitions. If you're reading this, you probably already know MIM combines powder metallurgy with injection molding. What you actually need to know is which feedstock properties matter in production, what can go wrong, and how to avoid the mistakes we made so you don't burn the same cash we did.
Powder Loading: The Number Everyone Fights About
Powder loading is the volume percentage of metal in your feedstock. Higher loading means less shrinkage after sintering, better density, stronger parts. Sounds simple. Clients love asking for 65% loading because some paper they read said it's possible.
Here's reality: we've had customers insist on 65 vol% for a 316L part. First shot, short fill. Second shot, flash everywhere because we cranked pressure. Third shot, finally filled but the green part cracked during ejection. We ended up at 61.5% after two weeks of back-and-forth. Could've saved everyone a lot of headache if we'd just started there.
The practical range for most stainless steels sits between 58-64 vol%, depending on particle size distribution and what your binder can handle. Go below 58% and you're looking at 18-20% linear shrinkage-good luck holding any tolerance tighter than ±0.3%. Go above 64% and your viscosity shoots through the roof; you'll need injection pressures that'll wear out your screw in six months.
What actually determines how high you can push loading:
- Particle size mix: We run a bimodal distribution on most jobs-coarser powder (15-25 μm) with fines (3-8 μm) filling the gaps. Get the ratio wrong and you're either too viscous or you'll see powder-binder separation during molding.
- Particle shape: Spherical gas-atomized powder flows better than irregular water-atomized stuff. But spherical costs 40% more. Guess which one the customer wants until they see the quote.
- Binder chemistry: Some wax-polymer systems tolerate higher loading than water-soluble formulations. We learned this the hard way switching a customer from Catamold to our in-house blend-had to drop loading 3% to get the same flow.
Any experienced metal injection molding service provider will tell you the same thing: optimal loading is specific to your part geometry, alloy, and binder system. There's no universal answer.
Flowability: Where MIM Gets Weird
MIM feedstock doesn't flow like ABS or nylon. Not even close.
Regular thermoplastics have one melting point. You heat past it, viscosity drops predictably, done. MIM binders are cocktails-wax melts at 60°C, then EVA softens at 90°C, then your backbone polymer finally lets go around 150°C. You get this stair-step viscosity curve that makes simulation software cry. We've had Moldflow predictions off by 30% on fill time because the software assumes a smooth viscosity-temperature relationship that doesn't exist in MIM.
The other thing that'll mess with your head: wall slip. In normal injection molding, the melt sticks to the mold wall-zero velocity at the surface, fastest flow in the center. MIM feedstock actually slides along the wall. Sounds like it'd help filling, and sometimes it does. But it also means your shear rate calculations are wrong, your pressure drop predictions are wrong, and that expensive CAE license isn't telling you the full story.
We've stopped trusting simulation for anything except rough ballpark on new molds. Real answer comes from short shots on the actual tool.
What kills flowability in practice:
- Humidity: Feedstock absorbs moisture if you're not careful with storage. We keep everything in sealed drums with desiccant. One batch sat out over a long weekend once, absorbed enough water that we got splay marks on every part. Dried it out in a dehumidifying hopper for 4 hours, problem solved-but that's half a shift gone.
- Fines content: Too much sub-5μm powder and your feedstock turns into peanut butter. Good luck pushing that through a 0.8 mm gate.
- Binder degradation: Run your barrel too hot or let material sit too long, and the wax starts breaking down. Viscosity drops initially (feels great), then you get inconsistent shots because the binder composition is drifting.
Thermal Properties: Why Cooling Takes Forever
Here's something that surprises people coming from plastics: MIM feedstock density runs 4-6x higher than engineering polymers, but thermal conductivity barely budges. Our 17-4PH feedstock tests around 2.9-3.4 W/m·K. Pure iron is 76 W/m·K. Pure wax is 0.25 W/m·K. You'd think 60% metal by volume would put you somewhere in between. Nope.
The reason is heat has to travel through the binder to get from particle to particle. Metal powder doesn't form continuous chains in the feedstock-each particle is an island surrounded by wax and polymer. So heat conducts fast within each tiny metal grain, then hits a binder roadblock before reaching the next grain. Your overall conductivity ends up closer to the binder than the metal.
Practical consequence: cooling time in MIM is basically still plastics math. Don't assume that metal content will speed up your cycle. We made that mistake once quoting a job, promised 18-second cycles based on "it's mostly steel." Actual cycle time: 34 seconds. Ate the margin on that one.
Feedstock properties we've actually measured in-house (2023-2025, not textbook averages):
| Alloy | Powder Loading (vol%) | Processing Temp (°C) | Melt Viscosity @180°C, 1000s⁻¹ (Pa·s) | Thermal Conductivity (W/m·K) | Green Strength (MPa) | Sintering Shrinkage (%) | Notes |
|---|---|---|---|---|---|---|---|
| 17-4PH | 60–62 | 165–172 | 380–550 | 2.9–3.3 | 18–22 | 16.8–17.8 | Our bread and butter. Clients change drawings constantly, shrinkage compensation gets redone every time. |
| 316L | 62–64 | 158–165 | 280–420 | 2.7–3.1 | 15–19 | 15.9–16.7 | Best flow of anything we run. Thin-wall dream. Lower strength though, customers complain sometimes. |
| 440C | 58–60 | 172–178 | 650–950 | 3.1–3.5 | 24–28 | 18.2–19.1 | High-hardness martensitic. Viscosity is brutal, nozzles wear out fast. We charge extra. |
| Ti-6Al-4V | 59–61 | 182–188 | 500–780 | 2.4–2.8 | 12–16 | 17.0–18.0 | Titanium picks up oxygen like crazy. Mixing temp above 190°C and the whole batch goes in the scrap bin. |
| Fe-2Ni | 63–65 | 155–162 | 320–480 | 3.0–3.4 | 14–18 | 15.3–16.1 | Cheapest alloy we run. Customers always try to negotiate price down even further. |
| WC-10Co | 52–55 | 195–205 | 1200–1800 | 4.2–4.8 | 30–35 | 20.5–22.0 | Cemented carbide. Viscosity so high we clog nozzles three times before getting one good shot. We mostly turn these jobs down now. |
These numbers bounce around batch to batch. Powder supplier changes their atomization parameters, suddenly your viscosity shifts 15%. Winter vs. summer humidity affects storage. That's just how it is. Anyone giving you clean single-value specs either hasn't run enough volume or is copying from a datasheet.
Binder Systems: Pick Your Poison
There's no perfect binder. Every system trades something.
Wax-polymer (what we use most)
Cheap to compound, forgiving process window, decent green strength. Debinding is slow-thermal cycles run 24-48 hours depending on wall thickness. But it's predictable.
Water-soluble
Faster debinding because you leach the soluble component first in a water bath, then finish thermally. Sounds great until winter hits and your shop humidity drops. We ran a water-soluble system one January, every single part cracked at the gate during the water soak. Moisture differential stress. Switched back to wax-polymer for that customer and never looked back.
Catalytic (polyacetal-based)
The Catamold-type systems from BASF. Debind in nitric acid vapor at 110-120°C, takes 6-10 hours depending on section thickness. Fast, clean, dimensionally stable. Also expensive-licensed feedstock costs 30-40% more than what we can compound ourselves. For medical or aerospace jobs where qualification paperwork matters, sometimes worth it. For cost-driven industrial parts, we make our own.
Gel-based (aqueous)
Mostly for ceramics or specialty applications. We don't run these often.
Binder compatibility with your specific powder matters more than people realize. We had a titanium job where the surfactant package in our standard binder caused powder agglomeration during mixing. Took two weeks to figure out why our viscosity was all over the place. Changed surfactant supplier, problem solved. But that's two weeks of production delay and about 80 kg of scrapped feedstock to learn that lesson.
Debinding: Where Parts Go to Die
You can do everything right in molding and still lose parts in debinding. This is where thin sections crack, thick sections blister, and your delivery schedule goes out the window.
The binder has to come out slowly enough that evolved gases can escape through the pore network without blowing the part apart. Too fast and you get internal pressure buildup-cracks, blisters, sometimes parts that literally pop. Too slow and you're tying up furnace capacity for days.
Thermal debinding ramp rates we actually use:
- Under 3 mm wall thickness: 1-2°C/minute through the critical 200-400°C range
- 3-6 mm walls: 0.5-1°C/minute, and we're babysitting the furnace
- Over 6 mm: we try to talk customers into design changes because debind cycles stretch past 72 hours and defect rates climb
Solvent pre-debinding helps a lot for thick sections. Water bath or heptane soak pulls out the soluble binder fraction first, creates open porosity, then thermal finishes the job without as much pressure buildup. Adds process steps but saves parts.
One thing nobody tells you: debinding atmosphere matters. We run nitrogen for most steels, argon for titanium, and have to purge the furnace properly between changeovers. Mixed up the gas lines once. Titanium parts came out blue-gray instead of silver-oxygen contamination, entire batch scrapped, customer was not happy.
Green Strength: Will It Survive Handling?
Green strength is how tough your part is right after molding, before any binder comes out. It needs to survive ejection, handling, maybe some trimming or inspection, then make it into the debind furnace without breaking.
Lower powder loading generally gives higher green strength because you've got more binder holding things together. But then you get more shrinkage later. Higher loading means better final density but the green part is fragile-drops or bumps during handling cause chips or cracks.
For parts with thin sections or unsupported features, we bias toward lower loading (58-60%) and accept the extra shrinkage. Compensate in the tool. For chunky parts that can take some abuse, we push loading higher.
Particle shape affects this too. Angular powder interlocks mechanically better than spherical. But angular flows worse. Always a tradeoff.
Our shop rule: if operators are breaking more than 1 in 50 green parts during normal handling, the feedstock needs adjustment. Either more binder, different binder backbone, or we redesign handling fixtures. Broken green parts are pure waste-you've already paid for the feedstock, the machine time, the operator time. Then it hits the floor and it's all scrap.
When working with a precision injection molding parts supplier, ask about their green part reject rate. It tells you a lot about whether they actually understand their feedstock.
Real Problems We've Solved (and Some We Haven't)
Thin-wall filling on a 316L sensor housing
Customer wanted 0.4 mm walls, 35 mm flow length. Initial feedstock at 63% loading couldn't fill without short shots or burn marks from over-packing. Dropped to 60.5%, added a small percentage of finer powder to help fill the tight sections, adjusted gate location to reduce flow length. Parts finally ran stable on shot 847. First 846 were learning experiences.
Shrinkage variation killing tolerances
Client needed ±0.05 mm on a 12 mm diameter bore. Sintering shrinkage on paper was 17.2%, but actual parts came out anywhere from 16.8% to 17.6% depending on where in the batch the powder came from. Traced it to inconsistent particle size distribution from the powder vendor. Switched suppliers, variation tightened to ±0.2% shrinkage, tolerances became achievable.
Ti-6Al-4V oxygen pickup
This one still bites us occasionally. Titanium is reactive. Any oxygen during mixing, molding, or debinding contaminates the part. We've invested in inert atmosphere gloveboxes for handling titanium feedstock, argon-purged mixing equipment, dedicated titanium-only debind furnaces. It's expensive. We pass that cost to customers who want titanium MIM, and some of them go elsewhere. That's fine-we'd rather turn down work than ship contaminated parts.
WC-Co (cemented carbide) clogging everything
Ran a trial batch two years ago. Carbide powder is hard, abrasive, and the feedstock viscosity was through the roof. Wore out a nozzle in 200 shots. Clogged the hot runner twice. Cycle times were 90+ seconds. We completed the trial, delivered acceptable parts, and decided that's not a market we want to chase. Some jobs aren't worth the equipment damage.
What Actually Matters When You're Picking a MIM Partner
Feedstock is where MIM quality starts. If a shop doesn't control their feedstock-or worse, doesn't understand it-everything downstream suffers.
Questions worth asking any potential custom injection molding supplier:
- Do you compound feedstock in-house or buy it? (In-house means they can adjust formulations for your specific part. Buying means they're locked into someone else's recipe.)
- How do you verify incoming powder quality? (If the answer is "we trust the supplier," walk away.)
- What's your batch-to-batch viscosity variation? (Good shops track this. Sloppy ones don't know.)
- Have you run my alloy before, and can you show me data? (Experience with specific alloys matters. 17-4PH knowledge doesn't transfer directly to titanium.)
We compound everything in-house. We test every powder lot for particle size, chemistry, and tap density before it goes into the mixer. We run rheology curves on every feedstock batch before it hits production. It's extra work, but it's why our Cpk numbers don't collapse randomly mid-run.
Two New Formulations Coming in 2025
We've been developing a couple of specialty feedstocks that should be ready for production trials by Q2:
If either of these fits a project you're working on, get in touch now. We have pilot batch capacity for Q1 qualification runs and can produce sample parts on your geometry. Once these go into regular production scheduling, lead times stretch. Working with an experienced injection molding solutions manufacturer early in development saves iteration cycles later.
The Short Version
MIM feedstock isn't magic-it's chemistry, physics, and a lot of trial and error. Powder loading, flowability, thermal behavior, binder system, debinding response, green strength-they all interact, and optimizing one usually costs you somewhere else.
The shops that run successful MIM operations are the ones that understand these tradeoffs at a practical level, not just from reading papers. We've scrapped enough feedstock and burned enough furnace cycles to learn what actually works. That experience is built into every formulation we run.
Got a part that might be a MIM candidate? Send us the drawing. We'll tell you straight whether it makes sense, what feedstock we'd recommend, and what the realistic tolerances are-no point wasting each other's time on projects that don't fit the process. As a metal injection molding OEM supplier with in-house feedstock development, we can tailor the material to your part instead of forcing your part to fit off-the-shelf material.