Binder Jetting Sintering: Shrinkage, Distortion & Dimensional Control

Binder jetting prints a green body at room temperature and ambient pressure — then hands off the real work to a furnace. Everything that determines whether the final part is dimensionally accurate, fully dense, and mechanically useful happens during debinding and sintering, not during printing. Engineers who treat the furnace cycle as a commodity step are routinely surprised by the results: 20% smaller in every direction, bowed, cracked, or under-dense. This guide gives you the physics and the numbers to control all of it.

1. The Binder Jetting Workflow

Print

The print head deposits an aqueous binder into a powder bed layer by layer, bonding powder particles together to define the part geometry. The printed state is called the green body. Green bodies have approximately 55–60% theoretical density (TD) — the rest is void space filled with unbonded powder and binder. They are fragile: green strength (binder-bonded, no sintering) is typically 1–5 MPa, comparable to damp sand. Green bodies must be handled carefully to avoid damage before curing.

Most commercial systems print at layer thicknesses of 50–200 µm. Print speed is high — HP Metal Jet S100 builds a full 430 × 309 × 200 mm build volume in a single run, enabling 100s to 1000s of parts per build. The print step is not the cost driver; the furnace step is.

Depowder

After printing, the build box is moved to a depowdering station. Loose unbonded powder is removed — by brushing, compressed air, or vibration — leaving the green body surrounded by a small shell of loosely adhered powder on fine features. At this stage the green body is at print density: ~55–60% TD with the remaining ~40–45% volume as open porosity (plus binder).

Debinding

Binder must be removed before the sintering cycle, or the carbon and volatile organics will cause porosity and contamination in the sintered part.

Two debinding routes are used commercially:

Catalytic debinding (used by Desktop Metal, based on BASF process): the green body is exposed to nitric acid vapour at ~110°C. The acid catalytically depolymerises the polyacetal binder, which evaporates leaving an open pore network. Typical cycle time: 4–12 hours. The resulting part is called the brown body — it retains the powder skeleton but has almost no binder remaining (< 0.5% residual carbon). Brown bodies are slightly stronger than green bodies due to slight particle necking during catalyst exposure but remain very fragile (< 5 MPa).

Thermal debinding (used by ExOne, some HP variants): the green body is heated slowly (typically 1–5°C/min) through a binder burnout profile in flowing gas. Temperatures of 200–500°C drive off organic components. The risk is carbon contamination if burnout is incomplete, and distortion if heating is too rapid before the binder network is removed. Thermal debind cycles are longer (12–24+ hours) than catalytic for equivalent part sizes.

Sinter

The brown body is placed on ceramic setter plates or sintering supports and loaded into a batch or continuous sintering furnace. During sintering, solid-state diffusion or transient liquid phase mechanisms (material-dependent) drive densification: the powder particles sinter neck, pores close, and the part shrinks isotropically. Typical sintered densities for well-controlled processes: > 97% TD for stainless steels, > 96% TD for copper alloys, with HIP achievable to > 99.9% TD.

2. Sinter Shrinkage: Physics and Magnitudes

Sintering is a volume-reduction process. As porosity closes, linear dimensions shrink. For a green body at 55–60% TD sintering to 97–99% TD, the volume ratio is:

Volume (sintered) / Volume (green) = TD_green / TD_sintered ≈ 0.57 / 0.98 ≈ 0.58

Linear shrinkage (isotropic) = 1 − (0.58)^(1/3) ≈ 0.17 → approximately 17–22% linear shrinkage in each dimension.

For 316L stainless steel, the industry-standard shrinkage figure is 18–22% linear in XY and Z — and the industry calls it "isotropic" because the ratio between axes is close to 1:1:1 for equiaxed powder particles (D50 ~22 µm, D90 ~38 µm) in a well-controlled process. In practice, small anisotropy exists:

The anisotropy arises from gravity (Z shrinks slightly more as the part sags downward during sintering at high temperature), powder packing differences between print layers (XY vs. Z packing), and binder jetting-induced density gradients from print head passes. Predictable anisotropy enables pre-scaling: most binder jetting systems apply differential XY/Z scale factors to compensate.

Pre-Scaling the CAD File

Every binder jetting workflow requires upscaling the CAD geometry before printing. The scale factor is:

Scale factor = 1 / (1 − linear shrinkage)

For 316L at 20% shrinkage: scale factor = 1 / 0.80 = 1.25× in each dimension.

Scale factors are determined empirically per alloy–machine–powder lot–sintering cycle combination using calibration test coupons. A 50 × 50 × 10 mm rectangular block is printed, sintered, and measured in XY and Z. Deviations from target are fed back into the scale factor. Initial scale factors provided by machine vendors are starting points; lot-to-lot powder variation can shift shrinkage by ±1%, enough to take a part out of tolerance.

For complex parts, uniform scaling is insufficient — different cross-section sizes sinter at slightly different rates (thick sections cool more slowly; thin features sinter faster). Advanced compensation uses section-specific scale factors or topological sintering simulation software.

3. Gravity-Induced Sag and Distortion

At sintering temperatures, metals are well above their creep threshold. At 1360°C — the sintering temperature for 316L — the material has negligible creep strength. The part behaves as a viscoplastic body under gravity. Thin, unsupported overhangs sag. Tall, slender features lean. Flat parts warp at the unsupported edges.

Typical distortion failure modes:

Cantilever sag: A horizontal shelf extending > 20 mm from a support point will sag measurably — often 0.5–2.0 mm for a 5 mm thick shelf at 316L sintering temperatures. The sag follows a beam-deflection curve: maximum at the free end.

Tall feature lean: Features with aspect ratio > 5:1 (height:width) tend to lean in the direction of the temperature gradient in the furnace (which is rarely perfectly uniform). A 100 mm tall, 15 mm wide cylinder may lean 0.5–1.5 mm from vertical.

Edge curl: Flat plates and flanges curl upward at the edges during sintering, driven by differential shrinkage between the hotter top surface (exposed to radiation) and the cooler bottom surface (on the setter plate).

Ceramic Setter Plates and Sintering Supports

Setter plates (typically Al₂O₃ or Y₂O₃-stabilised ZrO₂) provide the base on which parts rest during sintering. Flat, smooth setter plates minimise friction, allowing the part to shrink without restraint — friction-induced tension would cause cracking. Setter plate flatness tolerance should be < 0.2 mm over the part footprint.

For overhanging or cantilevered features, sintering supports — separate ceramic or metal structures placed under the part — prevent sag. These are analogous to machining fixtures: designed for the specific part geometry and destroyed (or reused for identical parts) after the sinter cycle. Design principles:

• Contact area should be minimised (point or edge contact) to avoid adhesion between the support and the sintering part
• Setter material must not react with the part alloy at sintering temperature — alumina is suitable for steels; yttria-stabilised zirconia or boron nitride-coated alumina for more reactive alloys
• Support height must account for shrinkage: the part will descend as it densifies

For high-volume production, HP and Desktop Metal provide sintering simulation software that predicts distortion and recommends support placement before printing.

4. Density Curve: Green to Sintered

Understanding the density evolution through the process enables defect diagnosis:

The green body density is determined by powder particle size distribution and binder jetting print parameters (saturation, layer thickness). Higher green density means less shrinkage required and less distortion. Bimodal powder distributions (fine + coarse particles filling interstices) can increase green density to 65–70% TD, reducing required linear shrinkage to 12–15%.

5. Residual Porosity and Its Effects

Sintered binder jetting parts at 97–99% TD retain 1–3% porosity, distributed as isolated spheroidal pores (0.5–50 µm diameter) and occasional inter-particle cavities from powder packing defects. This porosity affects mechanical properties differently by application:

Tensile strength: Sintered 316L at 97% TD achieves UTS ~500–550 MPa, yield ~180–210 MPa, elongation 40–50%. Machined wrought 316L: UTS ~520 MPa, yield ~205 MPa, elongation 50%. The difference is small — residual porosity at 97% TD has limited effect on static tensile properties for ductile alloys.

Fatigue life: More sensitive to porosity. Pores act as stress concentrations. High-cycle fatigue strength of sintered 316L at 97% TD is typically 30–40% lower than wrought at equivalent surface finish. Post-sinter HIP (99.9% TD) recovers most fatigue performance.

Pressure tightness: Parts with interconnected porosity (if sintering is incomplete or density < 95% TD) leak. For pressure vessels and hydraulic manifolds, HIP or infiltration with a lower-melting-point material (e.g., Cu infiltration into steel) is needed. Well-sintered parts at > 97% TD are generally pressure-tight for low-pressure applications (< 10 bar).

6. Material-Specific Sintering Cycles

Atmosphere control is critical: H₂ or forming gas (H₂/N₂ mix) reduces surface oxides on the powder particles, improving necking kinetics and final density. For stainless steels, either H₂-rich reducing atmosphere or argon (for oxidation prevention without reduction) is suitable. Pure tungsten and refractory metals require H₂ at very high temperatures to prevent oxide contamination. Never use air or N₂ alone for metals that form stable nitrides (e.g., pure Ti).

Ramp rates to peak temperature are typically 3–10°C/min. Too-fast heating causes binder burnout gas pressure build-up in closed porosity, creating blistering defects. A hold at 400–600°C ("binder burnout hold") is included in most sintering profiles for catalytically debound parts to ensure residual organics evaporate before densification closes the pore network.

7. Dimensional Accuracy and Control Strategies

Sintered binder jetting parts achieve dimensional tolerances of ±0.3–0.5% of nominal dimension (±0.3 mm per 100 mm) for well-controlled processes on calibrated machines. This is better than powder-bed SLS (typically ±0.3–0.5 mm absolute) but worse than LPBF (typically ±0.05–0.1 mm) for absolute tolerance.

Calibration Coupons

Print a set of calibration coupons (L-brackets, cylinders, cubes across the full range of dimensions planned for production parts) in every build. Measure after sintering with a CMM or calibrated callipers. Use deviations to update the scale factors for each axis. For critical features, print witness bars (simple rectangular bars) alongside every production build to verify that the sinter cycle matched the calibration baseline.

Fixture Sintering

For high-precision flat features (sealing surfaces, mating flanges), fixture sintering uses rigid ceramic or refractory metal fixtures to constrain the part geometry during sintering. The fixture prevents warp and lean. The trade-off is increased setup time and fixture cost, but for production parts requiring flatness < 0.2 mm, fixture sintering is often the only way to achieve the tolerance without post-sintering machining of every part.

HIP Post-Sinter

HIP at 1150°C / 150 MPa / 3 h (for 316L) after sintering closes residual porosity and provides a mild stress-relief/anneal. HIP does not significantly change outer dimensions (< 0.1% dimensional change) but reduces variability in part-to-part density, which is an indirect source of dimensional scatter. For medical implants and pressure-critical parts, HIP is standard.

Post-Sinter Machining

For tolerance-critical features — bearing bores, thread forms, sealing faces, precision locating holes — sintered binder jetting parts are left with 0.3–0.8 mm stock and machined after sintering. Sintered 316L and 17-4PH machine similarly to wrought annealed material. 17-4PH can be heat-treated to H900 condition (482°C/1h) before final machining if high hardness is required.

8. Machine Landscape

Digital Metal DM P2500 claims ±0.1 mm tolerance — the tightest in the category — through a combination of fine powder (D50 ~8 µm), fine binder drop placement, and carefully calibrated sinter fixtures.

9. When to Choose Binder Jetting

High-volume metal parts (hundreds to thousands per run): BJ's print speed and large build volumes enable production economics that LPBF cannot match for high-count runs. At volumes above ~500 identical parts, BJ per-part cost drops below LPBF in most steel and stainless alloys.

Complex internal channels and fine features: BJ requires no support structures during printing (the powder bed supports all overhangs). Complex geometries — internal lattices, conformal cooling channels, undercut features — print freely. Only the sintering step imposes any geometry limitation (sag, distortion in very thin/tall features).

Hard-to-weld and crack-sensitive alloys: Alloys like 17-4PH, tungsten, and some tool steels are difficult or impossible in LPBF due to hot cracking. BJ's room-temperature printing with solid-state sintering avoids the weld-solidification cracking mechanism entirely.

Biocompatible implants and scaffolds: 316L and 17-4PH sintered parts meet ISO 10993 biocompatibility requirements. CoCrMo is under development on several platforms. BJ porous scaffolds for bone ingrowth are an active clinical application.

When to avoid BJ: When tolerances tighter than ±0.3% are needed on non-machined surfaces; when the part volume is small (< 10 parts) and the sintering cycle setup cost is not amortised; when the alloy is not yet sintering-qualified on available equipment.

Further Reading

• Binder jetting metals overview — process fundamentals and material landscape
• Shrinkage compensation tool — compute pre-scale factors for your alloy and sintering cycle
• Cost per part calculator — model BJ part economics including furnace cost
• Dimensional accuracy reference — tolerance comparison across AM processes
• Post-processing time estimator — model debind, sinter, and HIP schedule