Metal Injection Molding: Feedstock Rheology, Debinding Kinetics and the 20% Sintering Shrinkage

Engineering guide to metal injection molding: powder loading and feedstock rheology, catalytic debinding kinetics, sintering shrinkage math, MPIF 35-MIM properties, design rules and per-part cost economics.

·

Metal Injection Molding: Feedstock Rheology, Debinding Kinetics and the 20% Sintering Shrinkage

A 24 g stainless hinge, 120,000 pieces a year, four cavities, one tool. The finished part has a 20.00 mm bore spacing held to ±0.06 mm. The tool that makes it is cut to 24.39 mm at that dimension.

Nothing in the finished part is at its moulded size. The cavity holds 81% more volume than the part it produces, the green part that drops out of it weighs 9% more than the finished part, the brown part that leaves the debinding furnace is chalk-strength and 8% lighter than the green, and the sintered part that leaves the 1,340 °C furnace is metal — 97% dense, 18.0% smaller in every linear dimension than the cavity that formed it, and 45% smaller in volume.

That is the whole engineering problem of metal injection molding in one sentence: the tool does not make the part, the furnace does. The tool makes a powder compact that the furnace converts into metal. Every tolerance, every alloy property, every cost line follows from the fact that the geometry is set twice — once by a steel cavity, once by a diffusion process you only partly control.

This guide works the numbers that decide whether the second shaping step lands where you planned: the powder packing that fixes how much the part will shrink, the feedstock rheology that decides whether it fills at all, the debinding square law that limits wall thickness, the sintering kinetics that turn 60 vol% powder into 97% dense steel, and the economics that decide whether any of it beats a machining centre.

1. What MIM is competing against

MIM sits in a narrow window: small, complex, metallic, and numerous. Outside that window it loses badly, and the loss is usually structural rather than incremental.

Process · Typical tolerance · Min wall · Part mass · Tooling · Economic volume · Material yield

MIM · ±0.3% (±0.05 mm floor) · 0.3–0.5 mm · 0.1–150 g · $5k–40k · >5,000/yr · >95%

CNC machining · ±0.005–0.025 mm · 0.2–0.5 mm · any · $0.1k–2k · 1–5,000/yr · 30–60%

Investment casting · ±0.5% · 0.5–1.5 mm · 10 g–50 kg · $1k–8k · 100–20,000/yr · 60–80%

Die casting · ±0.1–0.3% · 0.8–1.5 mm · 10 g–10 kg · $10k–50k · >20,000/yr · 60–75%

Press-and-sinter PM · ±0.1% (die-pressed) · 1.5 mm · 1 g–5 kg · $5k–30k · >10,000/yr · >95%

Metal AM (LPBF) · ±0.1 mm + distortion · 0.2 mm · 1 g–1 kg · none · 1–500/yr · 40–70%

The two neighbours that matter are press-and-sinter and investment casting. Press-and-sinter is cheaper per part and holds tighter tolerances, but it cannot make a cross-hole, an undercut, or a wall that varies in direction — the die has to withdraw. Investment casting makes larger parts with internal passages, but its as-cast tolerance is five times looser and its minimum wall is three times thicker.

MIM takes the plastic injection moulding playbook — 3D cavity, slides, cores, multi-cavity tools, 20-second cycles — and applies it to a material that is 60% metal by volume. The price of that trick is 15–22% linear shrinkage and a debinding step that limits section thickness.

Globally that trade is worth about 4.0 billion in CY2025, forecast to reach 6.2 billion by CY2030 at a 9.2% CAGR (Frost & Sullivan, cited in Indo-MIM's 2026 draft red herring prospectus; other market houses put 2025 at $4.7 billion on a broader definition, so treat the absolute number as ±20% and the growth rate as the reliable part). India is the third-largest revenue contributor after China and the USA — and the largest single MIM producer in the world is Bangalore-based Indo-MIM, which held a 6.8% global revenue share in CY2025, a position it has held for six years, across roughly one million square feet and 4,000+ employees.

flowchart TD
    A["Gas-atomised powder<br/>D50 5–15 µm, spherical"] --> B["Compounding<br/>60 vol% powder + 40 vol% binder"]
    B --> C["Feedstock<br/>η ≈ 100–1000 Pa·s at 10³ s⁻¹"]
    C --> D["Injection moulding<br/>160–190 °C, 50–120 MPa"]
    D --> E["Green part<br/>binder intact, 100% of cavity size"]
    E --> F["Debinding<br/>4–8 h catalytic / 6–24 h solvent"]
    F --> G["Brown part<br/>powder skeleton, chalk strength"]
    G --> H["Sintering<br/>1,250–1,380 °C, 12–18 h cycle"]
    H --> I["Sintered part<br/>95–98% density, −15 to −22% linear"]
    I --> J["Optional: HIP, heat treat,<br/>coining, machining, passivation"]

2. The powder decides the shrinkage before the tool exists

MIM powder is gas-atomised, spherical, and fine. The specification that matters is not the alloy — that is a table lookup — but the particle size distribution, because it sets three things at once: how much binder you need, how fast the part sinters, and how much it shrinks.

Typical production powder is D50 = 5–15 µm with D90 under 22 µm and less than 1% above 45 µm. Compare that with press-and-sinter powder at 45–150 µm or LPBF powder at 15–45 µm.

Specific surface area scales inversely with diameter for spheres:

For 316L at ρ = 8,000 kg/m³, a 10 µm particle carries 75 m²/kg of surface. At 45 µm it is 16.7 m²/kg; at 100 µm, 7.5 m²/kg. That tenfold difference in surface area is why MIM powder costs three to five times more per kilogram than press-grade powder, why its oxygen content has to be held under roughly 0.3% for stainless grades, and why every particle must be wetted by binder — surface you cannot wet is surface that will not sinter.

The same arithmetic explains the fineness requirement. A 10 µm powder puts about 2.4 × 10⁸ particles in every gram. Those particles have to flow into a 0.5 mm rib and reproduce its surface, which a 100 µm powder simply cannot do: the particle is a fifth of the feature.

Two packing parameters govern the rest.

Apparent and tap density measure how the powder packs dry. Spherical gas-atomised 316L typically taps to 4.3–4.8 g/cm³, i.e. 54–60% of theoretical. That is already close to the loading the feedstock will run at.

Critical solids loading is the volume fraction at which the powder plus binder system stops behaving like a fluid. It is measured, not calculated: a torque rheometer mixes binder and adds powder in 1 vol% increments at fixed temperature and rotor speed, and the torque rises linearly until the particles touch and interlock, at which point it turns upward sharply and the trace becomes erratic. For spherical metal powder that break point lands at 58–64 vol%; broad or bimodal distributions push it higher, angular or agglomerated powder pulls it down.

The operating point is set 2–8 percentage points below the critical loading. Dihoru's torque-rheometry work puts the optimum 7–8 points below critical; most production feedstocks sit 2–5 points below. Catalytic polyacetal systems, which flow better at high loading, reach 63–67 vol%. Everything else in this article follows from where that number lands.

3. Feedstock rheology: why 60 vol% is the edge of the cliff

A feedstock is a concentrated suspension, and concentrated suspensions do not behave like polymers. The relative viscosity — feedstock viscosity divided by binder viscosity — diverges as the solids fraction approaches the maximum packing fraction. The Krieger–Dougherty form is the standard description:

With intrinsic viscosity [η] = 2.5 for spheres and φ_m = 0.64 (random close packing), the numbers are brutal:

Powder loading φ · η_r · Feedstock η at 5 Pa·s binder

0.50 · 11× · 57 Pa·s

0.55 · 23× · 115 Pa·s

0.58 · 44× · 220 Pa·s

0.60 · 84× · 422 Pa·s

0.62 · 256× · 1,280 Pa·s

0.63 · 776× · 3,880 Pa·s

Two percentage points of loading — 60 to 62 vol% — multiplies viscosity by three. That is the cliff. Go too low and the part slumps during debinding because there is too much binder and too few particle contacts; go too high and the machine cannot fill the cavity, or fills it while separating powder from binder.

The practical ceiling is around 1,000 Pa·s at shear rates of 10²–10⁵ s⁻¹, with about 100 Pa·s being comfortable. Above 1,000 Pa·s, injection pressure runs out before the cavity fills, and the shear rate needed to thin the material generates enough viscous heating to scorch the binder.

The two constitutive behaviours you actually need

MIM feedstock is shear-thinning and temperature-sensitive. A power law with an Arrhenius temperature term captures both:

The flow behaviour index n sits between 0.2 and 0.6 for good feedstock. Lower means stronger shear thinning, which is desirable — viscosity collapses as the material enters the gate and races through the cavity, then recovers as it stops. But too low means the melt is so shear-sensitive that it flashes at the gate or fills unevenly between cavities.

The activation energy E sets how fast viscosity changes with temperature, typically 30–60 kJ/mol for wax-polymer systems. A high E is a liability: the melt cools against a 40 °C tool wall, viscosity climbs sharply in the last few millimetres of flow, and the part short-shots at the far end while the near end is fully packed.

Both effects combine in the moldability index, a single figure of merit that ranks feedstocks:

Higher is better. Reported values for production 316L feedstock land around 2 × 10⁻⁶ Pa⁻¹K⁻¹. The index is useful because it penalises the three failure modes simultaneously: high viscosity, weak shear thinning, and strong temperature sensitivity.

4. Moulding: where the defects are born

The moulding step looks like plastic injection moulding and is not. Four differences dominate.

Gates go at the thickest section. Plastic moulders often gate at the thinnest wall to balance flow. MIM gates at the thickest because feedstock freezes fast and because shrinkage has to be fed from the thick section outward. A gate at a thin wall produces a frozen gate, a short shot, and a sink on the far side.

Flow length to wall thickness must stay under 100:1. A 2 mm wall can be fed 200 mm from the gate; a 1 mm wall, 100 mm. Beyond that the melt is below its flow temperature before it arrives. This single ratio rules out more designs than any tolerance requirement.

Jetting is fatal, not cosmetic. If the melt enters the cavity as a free jet rather than spreading against a wall or core pin, it folds over itself and creates an internal weld line with a layer of binder-rich material trapped along it. That line survives debinding as a crack and survives sintering as a pore channel. Gate design should direct flow onto a cavity wall or core pin, and the gate should have radii rather than a sharp land.

Ejection needs more force than plastic. Green parts are weaker and have lower shrinkage on cooling, so they grip the core. Ejector pins are typically 4–6× the area used for the same projected area in plastic moulding, and draft is at least 0.5°, with 1° standard on outer surfaces. Zero-draft surfaces that demould fine in polypropylene will tear in feedstock.

Cycle times run 15–45 s, with melt at 160–190 °C and mould at 30–60 °C. Clamp forces are modest — 25–100 t covers most MIM — because cavity pressures are lower than in plastic moulding and parts are small. Machine hourly rates of $50–80/h are typical in Europe and North America, lower in India.

Tooling follows plastic mould convention: pre-hardened 1.2343 (H11 equivalent) is the standard cavity steel, with hardened inserts at high-wear gates. Cavity count scales degressively — a two-cavity tool costs about 1.5× a single, four cavities about 2.3×, eight about 3.5×, sixteen about 6× — and cycle time does not fall proportionally, because cooling and pressure build-up stay roughly constant. A 16-cavity tool runs about nine times faster than a single, not sixteen.

5. Debinding: the step that sets your wall thickness

Before sintering, 40% of the part by volume is polymer. All of it has to leave without moving the powder.

Three routes dominate, and the choice is a plant-level decision as much as a design one:

Route · Mechanism · Binder · Cycle · Wall limit · Share of production

Catalytic · HNO₃ vapour depolymerises polyacetal at 110–140 °C · POM + polyolefin · 4–8 h · 8–12 mm · Dominant at high volume

Solvent · Heated heptane/hexane dissolves the wax fraction, remainder thermal · Paraffin wax + PE/PP + stearic acid · 6–24 h · 6–10 mm · General purpose

Thermal · Staged pyrolysis and evaporation · Any · 12–48 h · 3–6 mm · Small batches, development

Water-soluble · Hot water dissolves PEG backbone · PEG + PMMA · 2–6 h · 4–8 mm · Niche, solvent-free plants

Catalytic debinding is worth understanding because it is the process most high-volume parts run. The binder is polyoxymethylene — polyacetal, the same polymer as engineering acetal — and it has an unusual property: in the presence of a strong acid it depolymerises cleanly back to formaldehyde monomer at temperatures well below its melting range.

The furnace is flushed with nitrogen, brought to 110–140 °C, and nitric acid is metered into the hot gas stream at 0.1–5 vol%. The acid attacks the polymer chain ends, the chain unzips, and formaldehyde leaves as a gas. The off-gas is burned. Because the reaction starts at the surface and moves inward, the part debinds from the outside in, and the front position follows a diffusion-like law.

That law is the design constraint:

Double the wall thickness and debinding takes four times as long. Quadruple it and it takes sixteen. This is why a 2 mm wall debinds in 4 hours and a 10 mm wall takes days — and why the practical wall limit for catalytic debinding is 8–12 mm even though the process is chemically capable of going further. It is not a moulding limit. It is a binder-removal limit, and it is the single most common reason a part that "looks mouldable" is not MIM-able.

The furnace parameters have narrow windows. Below 110 °C the reaction is too slow to be economic. Above about 150 °C the binder can melt and slump before it depolymerises, and the part sags on its setter. Acid concentration below 0.5 vol% starves the reaction; above 2 vol% attacks the powder surface and raises oxygen. Gas flow below 0.5 m³/h per zone leaves acid-starved corners; excessive flow creates temperature gradients across the load.

Residual binder is a sintering problem, not a debinding problem. The target is 90–95% of binder removed before the part enters the sinter furnace; the last few percent burn off during the 600–900 °C oxide-reduction stage. Leave more than that and the carbon from the polymer enters the metal — which is harmless in 316L at 0.02% and catastrophic in a low-alloy steel specified at 0.4–0.6% C.

6. Sintering: 60 vol% powder becomes 97% metal

Sintering is the step that creates the material. The driving force is surface energy: 75 m²/kg of fresh metal surface wants to disappear, and the only way it can is for particles to bond and pores to close.

The capillary pressure pulling two particles together is the Laplace pressure at the neck:

With a surface energy of about 1.5 J/m² for steel and a pore radius of 1 µm, that is 3 MPa. At 0.2 µm it is 15 MPa. That pressure, acting over the whole part, is what compacts the powder — no external press is involved.

Densification follows a characteristic time law. Kingery and Berg's classic analysis gives shrinkage proportional to t^(2/5) for lattice diffusion and t^(1/3) for grain-boundary diffusion. In practice, the whole thermal history collapses onto a single curve — the master sintering curve — using the work-of-sintering integral:

Plot density against Θ and every heating rate, hold temperature and hold time fall on one line for a given powder. That is how a sinter furnace recipe is actually developed: run three heating rates, fit Q, and predict what any other profile will produce. It is also why a 20 °C change in setpoint is not a small change — near 1,300 °C the densification rate is exponentially sensitive to temperature, and a 20 °C deviation can move 17-4PH density by around 0.5%, which is roughly 80 MPa of tensile strength.

Typical profiles:

Alloy · Sinter temp · Atmosphere · Hold · Density · Linear shrink

316L · 1,320–1,380 °C · H₂ or 75/25 H₂/N₂ · 90–180 min · 96–98% · 15–18%

17-4PH · 1,300–1,350 °C · H₂ or vacuum · 90–150 min · 96–98% · 14–17%

420 / 440C · 1,250–1,300 °C · Vacuum or H₂ · 60–120 min · 95–97% · 16–19%

Fe-2Ni / 4605 · 1,240–1,300 °C · N₂/H₂ · 60–120 min · 95–97% · 16–19%

Ti-6Al-4V · 1,200–1,350 °C · Vacuum <10⁻⁴ mbar · 120–240 min · 95–98% · 14–18%

Inconel 718 · 1,260–1,300 °C · Vacuum · 120–240 min · 96–98% · 14–17%

The atmosphere is a material variable, not a utility. Hydrogen is reducing: it strips oxides off the powder and decarburises the surface. That is exactly what you want in 316L and soft magnetic iron, and exactly what you do not want in a 0.5% C low-alloy steel that needs its carbon. Nitrogen-bearing atmospheres can raise strength in austenitic stainless by solid-solution nitrogen — and embrittle titanium. A plant change that alters the atmosphere alters the material, which is why MIM parts are qualified per plant and not per drawing alone.

Furnace cycles run 12–18 hours wall to wall, with a 60 kg load being a typical batch. At 45 kW and 0.12/kWh, the energy alone is about 86 per cycle — $1.44/kg. Sintering is not the dominant cost, but it is the dominant schedule constraint: a furnace is either full or it is losing money, which is why MIM quotes cluster around furnace-load quantities.

7. Shrinkage is not a fudge factor — it is the cube root of the loading

Here is the identity that makes MIM tractable. The green part contains metal at volume fraction φ. The sintered part contains the same metal, now at relative density ρ_rel. If the green volume is V_g and the sintered volume is V_s:

The volume shrinkage is therefore:

and the linear shrinkage is its cube root:

At φ = 0.60 and 97% density, S = 16.5%. At φ = 0.58, S = 17.4%. At φ = 0.64, S = 14.1%. In the ideal fully-dense case the volume shrinkage equals the binder fraction exactly — 40% by volume at 60 vol% loading — which is why a 20% linear shrink and a "half the volume disappears" statement are the same statement:

Two consequences follow, and both cost money.

The tool is cut oversize by 1/(1 − S). At 18% shrink that factor is 1.2195; at 20.4% it is 1.2563. A 20 mm bore spacing becomes a 24.39 mm cavity dimension. Every cavity dimension, every core, every slide position is scaled by that number before the tool is cut.

Shrinkage varies, and the variation is the tolerance. Feedstock batch-to-batch shift of 0.6 percentage points is normal — 17.7% versus 18.3%. A tool cut for the lower figure that meets a batch shrinking at the higher one produces a part 0.12 mm under size on a 20 mm dimension, which is twice the entire as-sintered tolerance band. It is why MIM houses run a shrinkage study on the first shots, adjust the sintering profile per batch, and why they will not guarantee a tolerance they have not measured on your geometry.

Shrinkage is also anisotropic. Flow aligns particles and binder, so shrinkage parallel to flow differs from shrinkage across it by typically 0.2–0.5 percentage points, and thick sections shrink differently from thin ones because they densify at a different rate. A part with a 6 mm boss on a 2 mm wall will bow toward the boss. The fix is design (equalise sections), not tooling.

The tolerance conversation, in money

As-sintered tolerance is ±0.3% of nominal, minimum ±0.05 mm, with ±0.5% on some geometries. On 20 mm that is ±0.06 mm. Ask for ±0.02 mm and you are not asking for a tighter tolerance — you are asking for a second process:

Requirement · Process · Capability · Cost impact

±0.5% · As-sintered · Standard · 1.0×

±0.3% · As-sintered, controlled · Standard production · 1.0×

±0.1% · Coining / sizing · IT7–IT8 · +15–30%

±0.02 mm · Grinding or hard turning · IT6 · +50–150%

±0.005 mm · Fine grinding, lapping · IT5 · +200%+

The economics are stark because each tight dimension drags fixturing, inspection and its own scrap rate behind it. A worked case from a European MIM supplier: a 24 g 17-4PH part at 120,000/year arrives with 14 dimensions toleranced between ±0.02 and ±0.05 mm, inherited from a machining drawing. None are achievable as-sintered. After a function analysis, three tight characteristics remain — two sized fits and one ground sealing face — and everything else runs at ±0.3%. Part price index falls from 100 to 62: a 38% cost reduction, with the control plan shrinking from 14 inspected dimensions to 3 special characteristics with Cpk evidence.

The lesson generalises. Put the tight tolerance on the one surface that seals, fits, or locates. Let the rest sinter.

8. What comes out of the furnace

MIM properties are specified by MPIF Standard 35-MIM (2025 edition) or ASTM B883, with ISO 22068 as the international equivalent. The standards define both minimum and typical values, and the gap between them is where most purchase-order disputes live.

Grade · Condition · UTS (MPa) · Yield (MPa) · Elongation · Hardness · Density

MIM-316L · As-sintered, min · 450 · 140 · 40% · — · ≥7.6 g/cm³

MIM-316L · Typical · 520 · 175 · 50% · 67 HRB · 7.90 g/cm³

MIM-17-4PH · H900, min · 1,070 · 970 · 4% · 35 HRC · ≥7.6 g/cm³

MIM-17-4PH · H900, typical · 1,190 · 1,090 · 6% · 40 HRC · 7.75 g/cm³

MIM-17-4PH · H1025, typical · 960–1,000 · 850–900 · 12% · 33 HRC · 7.75 g/cm³

MIM-420 · Hardened + tempered · 1,300–1,500 · — · 2–3% · 48–52 HRC · 7.6 g/cm³

MIM-4605 · Quenched + tempered · 1,200–1,500 · 1,000–1,200 · 4–7% · 40–45 HRC · 7.6 g/cm³

MIM-Cu · As-sintered · 200–250 · 60–80 · 25–35% · 45–55 HRB · 8.7 g/cm³

Typical production values for design guidance. The contractual minimums live in the current MPIF 35-MIM edition — quote those on the drawing, not this table.

The number to stare at is density. 97% density means 3% porosity, and porosity is not evenly distributed — it sits at prior particle boundaries, at weld lines, and in the last regions to densify. Three percent by volume is enough to matter in three ways:

Carbon is the other quiet variable. ASTM B883 caps MIM-316L at 0.03% C, and that limit is met by process control, not by buying low-carbon powder. Binder pyrolysis adds carbon; a hydrogen atmosphere removes it. Get the balance wrong and a heat of 316L comes out at 0.05% C, sensitised at the grain boundaries, with pitting resistance well below the wrought alloy.

9. Twelve design rules, with the numbers attached

These are the rules that decide whether a part quotes at ₹275 or ₹900 — or whether it quotes at all.

  1. Uniform walls, 2–5 mm ideal. Feasible range 0.8–8 mm; below 0.5 mm is micro-MIM territory; above 8 mm needs coring. Thick-to-thin ratio must not exceed 3:1.
  2. Relieve thick sections. The nominal wall sets cycle time, not the maximum wall. A 12 mm boss shrinks differently from a 2 mm wall and will bow the part.
  3. Draft at least 0.5°, 1° standard. Green parts grip harder than plastics and tear instead of flexing.
  4. Undercuts are possible but expensive. Each one needs a slide, lifter or side core, and adds 20–40% to tool cost.
  5. Internal corner radius ≥ 0.3 mm. Sharp corners are stress risers that crack during sintering, where they cannot be corrected.
  6. Fillet rib intersections at R ≥ 0.5 × wall. Three-wall junctions are porosity nests.
  7. Holes up to L/D = 5:1. Deeper needs a stepped core or a drilled secondary operation. Cross-holes need slides.
  8. Do not sinter load-bearing threads. Sinter a pilot and cut M3–M8 after sintering; it is cheaper and stronger than a moulded thread.
  9. Minimum feature and gap 0.5 mm. Below that, laser cut after sintering or redesign.
  10. Tolerance ±0.3%, floor ±0.05 mm. Tighter means coining or machining, priced accordingly.
  11. Flow length / wall thickness < 100:1. A 2 mm wall fills 200 mm from the gate, no further.
  12. Compensate anisotropy in the drawing. Shrinkage differs along and across flow; tight fits need the tool cut to measured, not nominal, shrink factors.

The corollary: a part with none of the following is probably MIM-suitable — walls over 8 mm, walls under 0.5 mm, flow paths over 80 mm, sharp internal corners, threads over M12, tolerances tighter than ±0.05 mm.

10. The economics: where ₹275 comes from

MIM cost splits into five lines. For a 24 g 17-4PH part at 120,000/year:

Line · Basis · Cost/part

Feedstock · 24 g × 1.5 shot ratio × 45/kg · 1.62 (₹154)

Moulding · 60/h, 35 s cycle, 4 cavities · 0.15 (₹14)

Debind + sinter · 60 kg load, 16 h cycle · $0.35 (₹33)

Secondary (deburr, passivate, inspect) · — · $0.55 (₹52)

Tooling amortisation · 27,600 ÷ 120,000 · 0.23 (₹22)

Total · · $2.90 (₹275)

Three things move that number more than anything else.

Shot ratio. The runner and sprue carry 1.3–1.8× the finished part weight, and cold-runner systems lose 8–12% of material as sprue that is recycled at a discount. On a 24 g part, a 1.8 shot ratio instead of 1.5 adds ₹31/part — 11% of the part cost — for a feature the customer never sees.

Cavity count. Machine cost per part falls as 1/cavities while tool cost rises degressively. At 60/h and a 30 s cycle: one cavity is 0.50/part, four is 0.125, sixteen is 0.031. The right cavity count is where the tool amortisation curve crosses the machine-cost curve at your annual volume — not the maximum the toolmaker can fit.

Material. Feedstock runs ₹1,400–2,100/kg for 316L, ₹4,300–6,600/kg for 17-4PH, and ₹19,000–38,000/kg for Ti-6Al-4V at ₹94.9/US$. A titanium part that weighs 5 g carries ₹95–190 of powder before any processing. This is why titanium MIM exists almost exclusively for implants and aerospace brackets where the alternative is worse.

The breakeven against machining

For the same 24 g part, MIM variable cost is about 2.67/part and the tool is 27,600. Against a 5-axis machined equivalent at $6.50/part:

At 9.00/part machined, breakeven falls to 4,400 parts. At 5.00, it rises to 11,800. That is the entire volume argument in one line, and it explains why the rule of thumb "MIM needs 5,000–20,000 pieces a year" is not a rule about MIM at all — it is a rule about how much machining the geometry would otherwise require.

A simple turned blank stays cheaper machined at 50,000 pieces. A part needing five setups, two fixtures and an EDM operation can be cheaper as MIM at 8,000.

11. Defect atlas

Symptom · Mechanism · Where it originates · Fix

Short shot at far end · Melt below flow temperature; flow length > 100× wall · Moulding · Move gate, add hot runner, thin the section, raise melt temp

Weld line that cracks in sintering · Jetting folds melt; binder-rich seam · Gate design · Direct flow onto core pin or wall; add gate radii

Blister or crack in debinding · Vapour pressure exceeds green strength · Heating rate too fast · Slow the ramp below 300 °C; reduce wall; switch to catalytic

Slumping on the setter · Binder melts before it depolymerises; insufficient particle contacts · Loading too low, temperature too high · Raise solids loading; hold 110–120 °C; use setters

Sink marks · Thick section feeds shrinkage from a thin gate · Design · Core out the thick section; gate at the thick end

Warpage after sintering · Section thickness variation → differential shrinkage · Design · Equalise walls; ribs instead of bosses; support in furnace

Dimensional drift between batches · Feedstock loading shift of 0.5–1 vol% · Compounding · Per-batch shrinkage study; adjust sintering profile

High carbon, low corrosion resistance · Incomplete debinding; carbon from binder · Debinding / atmosphere · Extend debind; raise H₂ fraction; verify C by combustion

Surface porosity, failed leak test · Residual 3% porosity, connected · Sintering · HIP, or resin impregnation, or raise density

Flash at the parting line · Excess binder, insufficient clamp, worn tool · Moulding · Raise loading, increase clamp, maintain tool

Notice how many of these are set at design or compounding and only discovered at sintering. That latency — a defect appearing two process steps and four days after its cause — is the defining operational difficulty of MIM, and the reason MIM suppliers insist on a design review before quoting.

12. What to put on the drawing

The material callout is not the material. Because the sintering profile creates the properties, a drawing that says "17-4PH, H900" specifies a powder, not a part. A production-ready callout has four parts:

Add the debinding route and the sintering atmosphere to the process flow diagram if the part is going into automotive or medical PPAP. A second plant with the same tool and the same feedstock but a different furnace will not produce the same part — that is physics, not supplier lock-in.

13. When not to use it

MIM loses on five conditions, and it is worth knowing them before commissioning a design review:

Everything else — the 24 g hinge, the 3 g surgical jaw, the 12 g lock bolt, the 40 g turbocharger vane — is a geometry problem, a rheology problem, and a furnace problem in that order. Get the powder loading right and the shrinkage becomes predictable. Get the gate right and the defects disappear. Get the wall thickness right and the furnace will do the rest.

The 20 mm dimension on that hinge is not held by the tool. It is held by 60 vol% powder loading, a 0.6 percentage-point batch window, a sintering profile developed on three heating rates, and a coining station that touches one face. The tool only decides where to start.

More FabFlow blog posts