English edition

MIM mirror finish: residual porosity, sintering and polishing of metal housings

A mirror finish on a metal injection moulded component is established across the manufacturing chain. Final buffing reduces fine surface irregularities, but material removal can also expose subsurface pores or alter edges and functional dimensions. The condition of the material at the eventual finished surface must therefore be considered together with the polishing allowance.

Watch bezels, wearable frames and metal heads for beauty devices are particularly sensitive to reflected light. A small depression that is inconspicuous on a matte surface may become a visible dark point after polishing. This does not, by itself, establish that polishing created a new pore. Existing porosity, particle pull-out, removal of a smeared surface layer and polishing residue require different corrective actions.

Surface finish is more than average roughness

Overall form, waviness, fine roughness and isolated defects describe different aspects of a surface. Form and longer-wavelength irregularities can distort reflected lines. Fine texture affects scattering, while an isolated pinhole can remain unacceptable despite a low average roughness value. A curved part also changes reflected images through its intended geometry; this must be distinguished from unwanted local waviness.

A drawing that specifies only “mirror polish” or a roughness limit leaves important requirements undefined. Cosmetic zones, viewing conditions, allowable local defects, protected edges and functional mating surfaces should be specified together. Gloss, image clarity and dimensional conformity are related but separate requirements.

How polishing exposes subsurface pores

MIM combines metal powder and binder into a feedstock, moulds the required shape, removes the binder and densifies the metal skeleton by sintering. Residual porosity depends on powder packing, moulding uniformity, debinding and the sintering history. The appearance of the as-sintered surface cannot reveal every pore at a greater depth.

A closed pore becomes surface-connected when the removal plane intersects it. For a spherical pore of radius r, centre depth z and material removal depth t, the opening radius is a = √[r² − (z − t)²] when |z − t| < r. This is a geometric explanation, not a model for predicting production acceptance. Real pores are irregular and may be connected; plastic flow and particle pull-out further affect the observed surface.

The opening can increase as the polishing plane approaches the pore centre and decrease after it passes the centre. Sufficient removal may eliminate that particular pore, but another pore may be exposed at a greater depth. Additional removal also consumes thickness and dimensional allowance. Neither continuously improving appearance with polishing time nor a universally denser surface skin should be assumed.

Why equal pore volume fractions can give different surface populations

Bulk density reflects the total pore volume but does not directly indicate how that volume is distributed among individual pores. Consider a geometric model containing uniformly distributed, non-overlapping spherical pores of one radius r, neglecting boundary effects. If the pore volume fraction is φ, the pore number per unit volume is N_V = 3φ/(4πr³). The expected number intersected per unit area of a plane is N_A = 2rN_V = 3φ/(2πr²), because pores with centres within one radius on either side intersect that plane.

Thus, many small pores and fewer large pores can produce different section appearances at the same pore volume fraction. A reduction in counted pores alone does not establish improved densification. Count, opening size and area fraction should be evaluated together, with the detection limit of the observation method specified. Small pores below that limit will not appear in the recorded population.

Actual MIM parts exhibit pore-size distributions, clusters, connected paths and local density gradients. The model is not a production pinhole-yield equation. It explains why bulk density, section pore count and the largest defect size provide different information.

Sections, connectivity and final surface depth

A round pore section does not prove that the void is isolated from the external surface. Features that appear as separate dots in one section may connect through narrow passages at another depth. An elongated depression also requires differentiation from a section through a pore network or an inclusion pull-out site.

Investigations should record size distribution, elongation, clustering and distance from the surface alongside area fraction. Where connectivity determines the treatment route, additional sections or three-dimensional analysis with suitable resolution provide further evidence.

Cosmetic development benefits from examining the depth range in which the final surface may lie, rather than only the as-sintered surface. Combined mechanical and electrochemical removal, together with local removal variation, can place different regions at different depths. Incremental removal on representative development specimens can track opening size and distribution in the same cosmetic zones. Relating destructive sections to finished-part appearance nevertheless requires a common location reference and a record of removal depth.

Powder and feedstock selection

Finer powder offers greater specific surface area and can increase the driving force for sintering. However, oxygen content, agglomeration, binder demand and feedstock rheology also change. Studies of MIM 316L show that particle-size effects depend on composition and sintering atmosphere. The smallest average particle size is not automatically the best choice for every process.

Particle-size distribution, shape, agglomeration, contamination, oxygen and carbon should be reviewed together. Gas and water atomization describe production routes; neither label alone establishes suitability for a cosmetic finish. Even powders carrying the same alloy designation can respond differently because of powder condition and composition.

The powder volume fraction must remain compatible with uniform filling and controlled debinding. Increasing the solids loading without confirming processability is not a reliable route to improved finish. Powder–binder separation or uneven filling can lead to local differences in green density, shrinkage and residual porosity. Gate regions, flow ends, junctions and thickness transitions are possible investigation locations, but their significance must be verified for the actual geometry.

Tracing location and manufacturing history

Defects should be grouped by both their location within the part and their manufacturing history before they are combined into a batch average. Recurrence in the same region of one cavity provides a reason to investigate filling and local powder concentration. Variation with furnace loading position across several cavities calls for examination of thermal history and atmosphere.

Changes associated with a polishing tool or work zone instead direct attention to removal and contact conditions. These patterns narrow the investigation; none independently proves a root cause.

Debinding, sintering and local densification

Interconnected voids during debinding provide routes for binder removal and decomposition products. They should not be evaluated as if they were the residual porosity of a finished sintered component. The objective is to remove the binder while retaining a stable metal skeleton, then achieve the required densification during sintering.

Inappropriate removal conditions can produce cracks, swelling or residue. Binder systems and local wall thickness require corresponding process schedules. A temperature or hold time established for one solvent, catalytic or thermal debinding route cannot simply be transferred to another.

Sintering develops particle bonds and changes pore size, shape and connectivity. Connected pores may become isolated as densification progresses. Higher temperatures or longer holds must nevertheless be balanced against grain growth, distortion and trapped gas. Atmosphere selection is also connected to oxide reduction, moisture, oxygen, residual carbon and alloy chemistry. Conditions reported for one feedstock are not universal production settings.

316L and 17-4PH require separate consideration. Their austenitic and precipitation-hardening metallurgical systems, hardness and heat-treatment routes affect subsequent finishing. Mechanical properties, corrosion requirements and appearance must be considered together when setting the sequence.

Relative density compares a measured density with an appropriate reference density. Composition and material state matter, as does liquid ingress into surface-connected pores during measurement. Equal bulk densities do not guarantee equal pore size distributions or equal conditions immediately below the cosmetic surface. Bulk density is useful for process monitoring, while local surface integrity requires a separate assessment.

Oxides, residual carbon and the near-surface region

Oxygen affects more than pore volume. A comparative MIM 316L study observed oxide particles involving silicon and manganese, with consequences for densification and the distribution of alloying elements. Treating every dark feature in a section as an empty void can therefore confuse porosity with non-metallic inclusions. Distinguishing retained oxide particles, actual depressions and particle pull-out sites changes the basis for deciding whether to modify feedstock cleanliness, the sintering atmosphere or polishing conditions.

Debinding cannot be described fully by shape retention or mass loss alone. Residual carbon also affects phase formation and shrinkage during subsequent sintering. Research on injection-moulded 17-4PH related different residual-carbon levels to high-temperature δ-ferrite formation and density gradients across the section.

If the same peak temperature and hold time give different densification patterns, the preceding debinding history and residual carbon warrant investigation. Phases present during sintering must be distinguished from the final microstructure after cooling and heat treatment.

Carbon and oxygen are not independent variables that should simply be minimized without considering their interaction. In some 316L powder systems, carbon participates in oxygen removal, while excessive residual carbon can impair final composition, microstructure and corrosion performance. Changes in powder oxygen and binder-derived carbon can alter the result of an otherwise unchanged thermal schedule.

Development should therefore relate incoming powder data, the debound condition and final sintered chemistry. Carbon additions reported for one study are not transferable recipes for another feedstock.

For a uniform matrix composition, the overall pore-volume fraction is the volume-weighted average of the local fractions. A thin near-surface region may contribute little to the measured bulk density even when its local porosity increases. If that region becomes the polished cosmetic surface, its effect on appearance can be substantial. Stable bulk density accompanied by increasing pinholes in one cosmetic zone therefore calls for comparison of that zone’s sections and removal history, rather than repeated bulk measurements alone.

Distinguishing pore exposure from polishing defects

A small depression may be an exposed pore, a particle pull-out site or another local defect. Directional lines suggest scratches that may originate from a previous stage, coarse abrasive contamination or detached particles. Broad irregular texture can arise from over-polishing, selective removal and material–tool interactions. Haze and staining may involve residues, cleaning, drying or chemical changes.

Observation after cleaning and at the same location after successive finishing stages helps distinguish these mechanisms. A representative cross-section can provide further evidence. Metallographic preparation must itself be controlled: grinding can smear metal over pores, while unsuitable preparation can pull particles out and create apparent voids. Dark regions should not automatically be treated as true porosity in image analysis.

SEM and compositional analysis of remaining particles may assist an investigation. Neither a single image nor analysis of an empty depression can fully reconstruct its formation history. They are most useful alongside manufacturing records and before-and-after comparisons.

Designing the polishing sequence

Each stage should have a defined removal objective and a defined shape to preserve. Initial machining, grinding or lapping can establish the required form and remove deeper marks. Intermediate stages remove preceding abrasive damage. Final polishing adjusts fine texture and optical appearance. Skipping an effective intermediate stage can lead to excessive pressure and dwell time during buffing.

Where geometry permits, changing polishing direction can help reveal scratches left by a previous stage. Tools and abrasives should be separated, with cleaning between stages to reduce coarse-particle carryover. Pressure, relative speed, dwell time, lubrication and tool compliance must be treated as a combined process rather than independent settings.

Removal is not uniform across flats, convex edges, narrow ribs and recesses. Convex edges may round before deeper marks disappear from a recess. Boundaries between polished and brushed zones also require positional control. Fixtures, masking and process order should protect functional holes, sealing surfaces, mating faces and appearance boundaries.

Polishing time is not a direct measure of material removal. Tool wear, curvature, contact area and pressure distribution change the removal rate. Thickness, step height or surface-height measurements at defined locations provide more useful evidence. If the material required to remove a defect exceeds the available dimensional allowance, the upstream surface or part design must be revised.

Final evaluation follows cleaning and drying. Residual compounds must not obscure defects or create a misleading appearance. Handling and packaging then protect the accepted surface from contact marks and contamination.

Applied force and local removal allowance

Equal applied polishing forces do not necessarily produce equal contact pressures. Contact area and pressure distribution change as a tool moves from a flat region to a convex edge or as its pad deforms. Nominal average pressure can be understood as applied force divided by apparent contact area, but this does not describe the local load carried by individual abrasive particles.

When edge loss increases, tool compliance, contact width, entry and exit paths, and local dwell should therefore be reviewed alongside total force. A compliant tool can follow curved surfaces, but it may also wrap around an edge and remove material from the side. Shape following and edge preservation require separate verification.

Removal allowance should be managed at critical locations rather than only as a part average. For sequential processing of the same surface, cumulative removal at a particular location is the sum of the local removal depths from the individual stages. This total must remain within the geometrical allowance at that location.

If the edge allowance is exhausted while scratches remain on a broader surface, extending the overall process time is inappropriate. Adjusting support and tool paths or obtaining more uniform scratch removal in an earlier stage is a more useful response. Before-and-after height comparisons should use an unaffected reference surface or a reproducible datum system. Re-aligning each measurement solely to the changed surface can absorb part of the geometrical change being evaluated.

Electropolishing, HIP and coatings

Electropolishing removes metal by anodic dissolution. It can improve microtopography and surface cleanliness, but it does not fill pores or densify the bulk material. Substrate pores, inclusions and deep scratches may affect the final result. Combined mechanical and electrochemical removal must remain within the dimensional allowance. Current distribution, rack contact, edge changes and removal of electrolyte residues require control.

Hot isostatic pressing is an optional densification treatment for suitable parts. Conventional capsule-free HIP mainly addresses pores isolated from the external pressure medium. When pores remain connected to the surface, the medium can penetrate them and reduce the effective pressure differential. Complete removal of all surface-connected defects cannot be assumed.

HIP selection requires evaluation of pore connectivity, alloy and heat-treatment state, dimensional response, subsequent machining and cost. It does not generally restore edges already lost through polishing or provide a universal remedy for inclusions and oxides. It is not a mandatory step for every mirror-finished MIM component.

Hard thin-film coatings can modify tribological properties, but they should not be specified as a general method for filling substrate pinholes or levelling deep depressions. Substrate appearance and cleanliness are established first, followed by coating adhesion and appearance verification.

Measurement and acceptance

Ra is the arithmetic mean of absolute profile-height deviations; Sa is the corresponding areal parameter. Neither average, on its own, controls the maximum size, number or location of isolated depressions. Measurement direction and location, evaluation length or area, filtering, form removal and instrument resolution should be specified for meaningful comparisons. A single trace may miss a sparse pinhole.

Bulk density, local pore observations, surface-height measurements, optical appearance and dimensions serve different purposes. Cosmetic requirements should define the inspected zones, allowable defect size and count, clustering and proximity to edges. Viewing illumination, angle, distance and the role of magnification should also be agreed. Approved and boundary samples reduce ambiguity.

Gloss alone cannot characterize haze or reflected-image distortion. Small or curved parts require an instrument footprint and geometry appropriate to their surface. A measurement taken on a flat coupon cannot be transferred uncritically to a narrow curved feature.

X-ray CT can help investigate internal pore location and connectivity where justified. Detectability depends on material, thickness, resolution and segmentation. Voxel size is not a universal minimum detectable pore size, and “not detected under the specified conditions” does not prove absence of porosity. Development analysis, periodic process verification and routine production inspection should have distinct roles.

Defect statistics and measurement selection

Quantify depressions using distinct measures: total count, count per inspected area, opening area fraction, size distribution and the largest observed size. Equal area fractions can describe either dispersed small depressions or a concentration of larger ones, with different appearance consequences. Comparisons require consistent inspected areas, detection thresholds, segmentation and cluster rules, and conventions for defects intersecting the observation boundary.

Surface opening area fraction should not be directly treated as internal volumetric porosity. Defect count and the fraction of rejected parts are also different measures. The latter results from applying predetermined size, location and clustering acceptance rules to individual parts.

A low Sa obtained by selecting a small pore-free region may describe the polished matrix, but it does not establish the appearance quality of the entire part. Conversely, an average dominated by deep depressions may obscure improvements in the fine scratches of the surrounding matrix. Where needed, evaluate a representative cosmetic region separately from the matrix with defects excluded, while identifying the excluded regions and processing rules and retaining the original height data.

Values obtained by deleting or interpolating depressions must not be presented as measurements of a defect-free surface. Missing optical data inside a depression should not automatically be treated as a flat surface; the missing-data condition and the instrument’s measurement capability require review.

Stop criteria and final acceptance

A polishing stage should not end solely at the point where roughness reaches a lower value. Track remaining scratches together with depression count, reflected-image waviness, edge loss and remaining dimensional allowance. If scratches decrease while new depressions and edge loss increase, reconsider the tooling or upstream process instead of extending polishing.

Development stop criteria and final product acceptance criteria serve different purposes, but the final process must satisfy appearance, geometry and functional dimensions together. If measurements or visual decisions near an acceptance boundary vary on repetition, verify the measurement and observation conditions first. A difference smaller than that variability should not be claimed as demonstrated process improvement.

Developing a repeatable production window

Define cosmetic zones, edges, functional surfaces, permitted fixture-contact areas and material-removal allowance before finalizing the process. Relate bulk and local sintered conditions to the surface that will remain after finishing. Track mould cavity, sintering lot, surface location and polishing stage so that local defects can be linked to the relevant conditions.

Compare the same reference locations before polishing, after intermediate stages and after final cleaning. Determine whether scratches decrease while pinholes increase, or whether new waviness appears after a tool or parameter change. Lock the validated feedstock, moulding, debinding, sintering, polishing and cleaning conditions together, and reassess appearance when those conditions change.

Comparing equal time with equal removal depth

An increase in relative density is not, by itself, a complete validation of a process change. Compare pore exposure, reflected-image quality, edges and functional dimensions after matching final removal depth in the same cosmetic zones. Improved sintering may still require more corrective removal if distortion changes, exposing a different subsurface region.

Conversely, reducing removal may lower the pinhole count while leaving deeper marks from an earlier operation. Separating these interactions makes it possible to assess the contributions of densification and polishing changes more reliably.

Register inspection regions to the part geometry and retain surface images together with cumulative removal measurements. The same location means the same lateral coordinates or geometrical region; it does not mean that the same material depth remains visible after polishing. An increased depression count may reflect exposure of deeper pores, tool-induced pull-out, or both.

Comparisons between tool conditions should therefore use comparable starting material and, where practical, include observations at a common removal depth. A comparison made only at equal processing times combines differences in removal rate with differences in the depth exposed.

Interactions, blocks and independent replication

Development trials should reveal interactions rather than rely only on a before-and-after statement that a different tool improved the result. Tool compliance and applied force can, for example, be crossed in a set of combinations, with actual removal, remaining scratches, depressions and edge change measured for each combination. If a lower force helps only with one tool, the result should not be generalized as a force effect alone.

An equal-time test of removal efficiency and a test of surface damage at a common removal depth answer different questions. Their results should be distinguished. Experimental ranges must remain within previously established machine and part limits, and the lowest roughness value alone should not determine the selected condition.

Assigning a different sintering lot to each tool condition confounds material variation with the polishing effect. Include the compared conditions within the same lot and, where appropriate, use lots or working days as blocks in the analysis. Randomize treatment order within each block where feasible and record tool wear, cleaning and conditioning.

Multiple fields on one part describe within-part variation; they are not equivalent to repeating the polishing treatment on separate parts. Evaluating part-to-part variation requires independently treated experimental units. Where multiple parts are processed together, the processing batch should also be identified.

The production objective is a repeatable cosmetic surface that retains its required dimensions. High gloss on one sample is insufficient. Stable mirror finishing requires suitable densification, preserved geometry, controlled removal and consistent acceptance conditions across the complete manufacturing route.