English edition

Metal powder selection and mold compensation for MIM sintering shrinkage

Final dimensions in metal injection molding result from the interaction of mold geometry, feedstock characteristics, molding, debinding and sintering. Powder particle size and shape, metal volume fraction, molding uniformity and sintering conditions must be evaluated together before a mold compensation factor can be used consistently in production.

A housing can meet its overall length requirement while failing to assemble because boss spacing or hole locations are incorrect. Conversely, substantial shrinkage can be accommodated when dimensional changes are repeatable. The objective is to reach the required density and shape while controlling variation in functional dimensions.

1. Green parts, debound parts and densification

MIM feedstock combines metal powder with a binder system. The binder enables flow into the mold and supports the shape after cooling. Debinding removes binder through a controlled sequence while retaining the powder skeleton and maintaining routes for further binder removal and gas transport.

Binder removal and macroscopic densification are different processes. Proper debinding aims to retain the principal geometry; substantial dimensional reduction takes place as densification proceeds during sintering. This reduction is associated with changes in pore volume and is distinct from thermal contraction alone. Dimensions used to calculate shrinkage should therefore be compared at a consistent reference temperature.

Particle bonding alone does not establish that the required density has been achieved. Residual porosity, density and the mechanical properties required by the application need separate verification. Insufficient densification is not an acceptable way to obtain a lower shrinkage value.

2. Linear shrinkage, volumetric shrinkage and oversizing

A statement such as “20–30% shrinkage” requires a defined reference: length or volume, green or mold dimensions, and the process stages included. A percentage from one feedstock and process cannot be adopted as a universal MIM design value.

Let L_g be the green-part length and L_s the cooled sintered-part length. Linear shrinkage is s = (L_g − L_s) / L_g. The corresponding green dimension is L_g = L_s / (1 − s). The oversizing factor relative to the final dimension is K = 1 / (1 − s). Use s as a decimal: 20% corresponds to 0.20.

In a hypothetical example, a final length of 10.00mm and 20% linear shrinkage require a green length of 12.50mm. This is 25% larger than the final length. A green length of 12.00mm would shrink to 9.60mm under the same assumption.

For comparison, assumed linear shrinkages of 15%, 18%, 20%, 25% and 30% correspond respectively to oversizing factors of 1.1765, 1.2195, 1.2500, 1.3333 and 1.4286. These are mathematical examples, not recommended material values or measurements of GPN products.

If shrinkage is equal in all three directions, V_s / V_g = (1 − s)³. A hypothetical linear shrinkage of 20% gives a volume ratio of 0.512, or 48.8% volumetric shrinkage. Linear and volumetric percentages are therefore not interchangeable.

The equations above relate green and sintered dimensions. Actual tooling also requires consideration of the cavity-to-green dimensional change, the supplier’s definition of the oversizing factor, local compensation and subsequent operations. When an OSF already describes mold-cavity-to-final dimensions, molding shrinkage must not be counted again.

3. Selecting the metal powder

Begin with the required alloy chemistry, corrosion performance, strength and heat-treatment condition. Powders with the same alloy designation may differ in particle-size distribution, manufacturing route and surface oxidation, affecting molding and sintering behavior.

D50 alone does not describe the distribution. Review D10, D50, D90 and the full distribution curve, and establish whether the result is volume-based or number-based. Measurement and dispersion conditions must be consistent. Laser diffraction reports an equivalent particle diameter based on an optical model; it does not replace direct morphology inspection.

Fine powders can support sintering and small-feature reproduction, but increased specific surface area can also increase binder demand, viscosity, agglomeration and oxidation sensitivity. Coarser powders may offer advantages in cost and handling, while demanding careful assessment of feature detail, surface finish and the conditions needed to reach the target density.

A mixed particle-size distribution may improve packing because smaller particles occupy spaces between larger particles. Changing the blend also changes flow and segregation behavior. Packing performance and uniform movement through the actual mold should both be tested.

Near-spherical particles can support flow and higher solids loading. Interparticle friction and skeleton stability after binder removal remain relevant. Review agglomerates, satellite particles, internal voids, irregular particles and chemical cleanliness alongside shape.

Oxygen, carbon, moisture and contamination must be considered in relation to the alloy and sintering atmosphere. Surface oxides and carbon changes can affect densification and the final microstructure. Apparent density and tap density characterize powder-bed packing; neither directly substitutes for the metal volume fraction in a green part or for final sintered density.

4. Feedstock solids loading and shrinkage

Solids loading is the metal powder’s volume fraction in the combined powder-and-binder mixture. It must be distinguished from a weight fraction because powder and binder densities differ substantially. For a two-component calculation, φ = (m_p / ρ_p) / [(m_p / ρ_p) + (m_b / ρ_b)]. For multiple binder constituents, add their individual volumes.

At the same target final density, a higher initial metal volume fraction generally requires less volume reduction. Excessively high loading, however, can cause inadequate binder coverage or excessive viscosity and incomplete filling. Excess binder can promote segregation and loss of shape during debinding. Critical loading and the selected operating formulation are different values; the usable formulation needs validation for filling, packing and ejection in the actual mold.

A simple mass-conservation model illustrates the relationship. Assume unchanged metal mass and theoretical density, no additional green-part voids, and isotropic shrinkage. Let φ represent the green metal volume fraction and R the final relative density. Then V_s / V_g = φ / R, L_s / L_g ≈ (φ / R)^1/3, and s ≈ 1 − (φ / R)^1/3.

Using hypothetical values φ = 0.60 and R = 0.98 gives approximately 15.1% linear shrinkage. Increasing φ to 0.64 at the same R gives approximately 13.2%. This is an explanatory model, not a tooling release value. It excludes oxide reduction, composition changes, additional voids, nonuniform green density and mechanical constraints.

Uniform overall formulation does not guarantee uniform local composition. Binder-rich areas may shrink differently. Mixing quality, molding temperature, flow path, gating and packing conditions should be assessed together. Feedstock changes require review of grade, batch, binder system and OSF conditions. Any use of recovered feedstock also needs defined blend limits, thermal-history control and contamination control.

5. Global and local mold compensation

Start with final-product dimensions that account for machining, heat treatment and surface finishing. Connect the final targets to as-sintered, green and tooling dimensions. Scaling an external outline before defining hole locations and datum relationships can produce an inconsistent compensation scheme.

Choose an initial factor from validated supplier data under clearly defined conditions, then verify it using trial parts made with the intended feedstock and processing route. Record whether the factor already includes the cavity-to-green change.

Evaluate overall length, bore diameter, boss height and hole-center distance as separate functional characteristics. Uniform free shrinkage can support a common factor, but core behavior, local density and support constraints may create departures. Flatness and positional tolerances cannot be controlled through a length shrinkage percentage alone; datum alignment and measurement methods must be defined.

In another hypothetical example, a fixed green length of 12.50mm yields 10.000mm at 20.0% shrinkage and 9.975mm at 20.2%. A difference of 0.2 percentage points shifts the final length by 0.025mm, or 25μm. For a fixed L_g, ΔL_s = −L_gΔs. A precise scale factor cannot compensate for a changing process.

Separate a stable mean offset from excessive part-to-part scatter. For precision bores or contact faces requiring machining, ensure that sufficient machining allowance remains across the expected sintered variation. Define which dimensions are obtained through molding and sintering, which are finished later, and where local tooling adjustment is practical.

6. Wall thickness, bosses and ribs

Large bosses attached to thin walls, abrupt thickness changes and asymmetric material concentrations affect filling, cooling, binder removal and sintering differently from a more uniform section. Consider coring nonfunctional mass and using gradual transitions while preserving required stiffness and fastening capacity.

Hollowing a boss must be reviewed together with thread engagement, wall strength, core-pin support and ejection. Ribs may improve stiffness but can create locally thick intersections and new distortion risks. Their location and thickness need molding and sintering validation.

Gate placement relates to filling uniformity, surface traces and powder–binder segregation. Sintering orientation relates to gravitational deformation and support friction. Review both in relation to functional and visible surfaces. Where thickness differences are unavoidable, prioritize boss height and spacing, floor flatness and wall inclination in trial-part measurements.

7. Debinding, sintering and support constraints

Solvent, catalytic and thermal debinding routes depend on the binder system. Binder-removal rate, transport paths and skeleton stability need to be balanced, especially in thick sections. Damage introduced during debinding may remain after sintering.

Sintering temperature, hold time and atmosphere influence density, chemistry and microstructure together. Higher temperature may improve densification within a suitable range, but grain growth, slumping and composition changes also require consideration. A single temperature or time cannot be specified for all alloys.

Contact friction between a shrinking part and a setter can restrict free motion. Part weight and high-temperature deformation also matter. Review support position, contact area, setter material and surface condition. Added contact can increase constraint, while insufficient support can allow sagging. Functional geometry and contact marks should be checked while establishing a suitable balance.

Measure length, width and height separately. Investigate repeatable directional differences in relation to feedstock, molding, geometry, support and thermal conditions before adopting directional scale factors. Furnace position, loading arrangement and setter history belong in the trial record. Traceability by lot, cavity and furnace position helps distinguish process variation from a tooling error.

8. Trial and compensation validation

1. Define functional dimensions, datums and the final manufacturing state.
2. Record feedstock grade and batch, binder system and the reference used for the initial factor.
3. Obtain corresponding dimensional and mass data from green, debound and sintered parts, using nondamaging methods for fragile debound parts.
4. Separate mean error, scatter and geometric distortion by cavity, direction and furnace position.
5. Stabilize the cause before adjusting the global factor or local tooling geometry.
6. Revalidate dimensions, assembly, required properties and repeat production after compensation.

Tracking the same specimens through process stages is useful where practical. Otherwise define how samples are matched. Measurement resolution alone is insufficient: fixturing, force, alignment, temperature and repeatability must be assessed.

9. Diagnosing dimensional variation

A consistent global size offset suggests reviewing factor definitions, feedstock changes and mean shrinkage. A correct mean with excessive scatter calls for checks on feedstock uniformity, molding and packing, furnace arrangement and measurement variation. Errors concentrated at a particular bore or boss require examination of core behavior, local thickness and constraint.

Warped floors or long features call for review of geometry, weight, support contact and sintering conditions. Low shrinkage accompanied by inadequate density or properties requires investigation of densification, oxidation and agglomeration. A shift after a raw-material lot change should be compared against particle-size, chemistry, loading and feedstock inspection data.

These are diagnostic candidates rather than conclusive causes. Compare process-stage data to locate where the change began before modifying tooling.

10. Maintaining the validated production basis

Compensation values rely on the validated material and process combination. Changes to powder supplier, feedstock grade, recovered-material ratio, molding conditions, debinding route, atmosphere or setters require revalidation appropriate to their potential effects.

Record material chemistry, distribution, lot and storage; feedstock and molding conditions by cavity; debinding and sintering histories, loading and support; and final inspection datums, temperature, fixturing and acceptance state. Monitor mean and scatter separately. Process capability indices require a stable process and sufficiently controlled measurement variation. A few passing first articles do not establish production capability.

Material specifications support chemistry and property verification. Functional dimensions and assembly acceptance require the product drawing and agreed inspection plan. Density-test methods must account for open porosity and liquid penetration where relevant.

11. Information for a MIM manufacturing review

Provide a final-size 3D model and dimensional drawing. If an enlarged model is supplied, distinguish it from the final-product model and identify the applied scale and reference. Include loads and repeated motion, corrosion environment, heat-treatment requirements, datums and functional tolerances, visible surfaces and finishing requirements, development and production quantities, and inspection or test documentation.

Connecting powder specifications, compensation factors, functional dimensions and trial results enables a specific review of material selection, geometry changes, tooling adjustment and secondary machining.