Steel Cleanliness Reports: What Inclusion Ratings Can Prove

An inclusion rating is a result from examined specimens, not a certificate for every cubic millimetre of steel in a heat. Before a buyer compares two headline ratings, the report must show which product and location were sampled, how the surface was prepared and observed, which method variant and edition applied, and where the purchase requirement sets an acceptance limit.

The trace runs from sample to method to rating to acceptance. Break any link and the conclusion narrows. Morphology observed in an optical image can support a method-specific category, but it does not automatically identify a particle’s chemistry or origin. This report-reading task is also narrower than the process comparison covered in forged and cast undercarriage parts.

Identify the material and sample represented

Start with the supplier, mill, and laboratory, then record the heat or cast, lot, grade, and material specification. Identify the product form as rolled or forged when the selected standard requires that scope. Link the inclusion report to the relevant billet, bar, forging, or part and retain the product revision and test date.

For an ISO-based report, the current reference matters. ISO 4967:2026, Edition 4 was published in May 2026 and replaced the 2013 edition. Its public scope covers micrographic assessment of non-metallic inclusions in rolled or forged steel with a reduction ratio of at least 3, using reference charts or image analysis, and notes limitations for some steels. That scope does not establish applicability to a cast product, an inadequately reduced product, or every steel family without review.

Record the stated reduction ratio and its source when ISO 4967:2026 is invoked. Capture the sampling plan, number of specimens, and location within the parent product: end, mid-length, surface-near, centre, or another defined position. State depth and orientation relative to rolling, forging flow, or product axes. Assign unique specimen IDs and preserve chain-of-custody records.

Representativeness must be stated rather than assumed. One polished field from one billet position describes the examined area of that specimen under the method. It does not automatically represent every part, depth, orientation, or position in a heat. The purchase specification or approved test plan should define sampling frequency and the population to which the result may be applied.

Read the method, edition and specimen preparation

ASTM version control is equally important. ASTM’s current physical and mechanical testing catalog lists E45-25 as the active steel-inclusion standard. The ASTM E45 public scope and significance page explains key boundaries, including method variants, examined-surface and sampling limitations, morphology rather than chemical identity, and the fact that the methods do not set acceptance limits. Use the controlled E45-25 text for current procedure details.

Record the exact standard and edition, then name the method variant and result expression. Identify whether the examination used manual chart comparison, direct measurement, or image analysis. Two reports carrying the E45 label may use different variants, field-selection rules, and outputs, so the headline ratings should not be compared until those fields align.

Document specimen preparation, including polishing and whether etching was used or prohibited by the selected procedure. Record magnification, calibration, field or area-selection basis, number of fields, and total observed area. State whether the result uses a worst-field, average, count, or another method-defined rule. Do not insert procedural values from an older edition into a current report.

For automated work, retain equipment identity, image-analysis software and version, detection or classification configuration as required, calibration or verification evidence, operator, and laboratory. Accreditation should be checked for the relevant method and scope; a logo alone does not prove competence for this task. Preserve deviations and any limitation caused by preparation, image quality, or an unexamined region.

Interpret the rating as morphology and distribution evidence

Read the result using the vocabulary of the selected method. Record the reported inclusion category, any thin, heavy, or other method-specific class, and the indicated size, shape, concentration, or distribution. Keep worst-field, average, and count outputs distinct. Preserve units, rounding, and the precise rule that turned observed features into the report entry.

Link every reported field image to specimen ID, location, orientation, magnification, calibration, and field number. The image demonstrates what was detected on the examined surface under the stated preparation and method. It does not reveal unseen volume. A micrographic result should not be substituted for a macro examination or another test with a different scale and purpose.

Record outliers and clusters rather than allowing an average to hide them. If the method requires a worst-field result, retain the field that controls it. If multiple specimens differ, preserve individual results and the relevant summary. Note measurement uncertainty or repeatability information when the quality plan requires it.

Labels sometimes associated with sulfide, oxide, silicate, or globular morphology should be interpreted within the rating method. The shape-based category is not confirmed elemental composition for every particle. Questions about indigenous or exogenous origin require additional evidence and metallurgical interpretation; the optical label alone cannot answer them.

Separate morphology from chemical identification

When the decision requires particle chemistry, state that question separately. Optical microscopy can show morphology, distribution, and response to the selected preparation, but appearance does not uniquely identify composition. A dark elongated feature, for example, should not be assigned a definitive chemical formula or process origin merely because it resembles a chart category.

The public scope of ASTM E2142-08(2023) illustrates that scanning electron microscopy and X-ray analysis can address particle chemical classification under their own methods. This is an alternate-method pointer, not a requirement for every order, and its applicability to the specific order must be checked before use. It does not turn an E45 or ISO 4967 optical rating into chemistry data.

For any chemical investigation, retain how the particle was selected, its specimen and field identity, the analytical method, calibration or quality controls, and uncertainty or detection limitations. A selected particle may not represent all particles in the specimen, much less the heat. Report the analyst’s conclusion only for the evidence actually examined.

Questions about indigenous or exogenous inclusions and process origin also need corroborating melt, refining, casting, deformation, heat-treatment, and handling records as relevant. Process knowledge can frame hypotheses, but morphology or one elemental spectrum does not prove root cause. A metallurgist should integrate the observations and state the scope of the conclusion.

Trace acceptance to the purchase requirement

Keep the test result and acceptance requirement in separate fields. Identify the purchase order, drawing, material specification, customer agreement, and applicable revision. Record the standard, edition, method variant, result expression, sampling frequency and locations, acceptance limits, and rounding rule exactly as invoked.

Then place each reported value beside the corresponding limit without changing its units, class, field rule, or rounding. A valid E45-25 report can still lack an applicable acceptance limit because ASTM’s method does not itself establish acceptability. ISO 4967:2026 likewise supplies a method, not a universal cleanliness grade for every product.

Record deviations and any agreement that authorizes them. Identify the customer, supplier, or other decision authority specified by the quality plan. Other required tests, such as chemistry, mechanical properties, ultrasonic examination, or a product-specific inspection, remain separate. A favorable inclusion rating does not replace them.

Release status should quote the actual authorized disposition. “Report complete” and “method fields aligned” do not mean “lot accepted.” If the sampling plan, limit, method, or decision authority is missing, the report remains incomplete for disposition even when the measured rating appears low.

Control retest, reclassification and report revisions

When a result is questioned, preserve the initial report number and version, the nonconformance or query, and the original specimen identity. State whether the next action is a re-read of the same field, a reclassification using the same images, repolishing the same specimen, testing a new specimen from the same location, or additional sampling from another location. These actions add different evidence.

Record any change in method, edition, operator, laboratory, microscope, software, configuration, preparation, observed area, or classification rule. A lower second rating from another location does not erase the first result. It expands the sampled evidence and may reveal location variability that the decision authority must address.

If a report is revised, identify the superseded version and explain why each result changed. Retain the original data, images, calculations, review and approval history. Do not silently replace a disputed image or average, and do not reclassify only selected fields without recording the selection rule.

Additional sampling and deviations should follow the quality plan and purchase requirement. The final disposition must cite the authorized owner and the complete evidence considered. Retest is a controlled investigation step, not an automatic route from a nonconforming result to acceptance.

Use the report review table to expose three states

The table below keeps the full sample-to-method-to-rating-to-acceptance trace. “Normal” means the relevant evidence aligns. “Missing” means a decision-critical field is unavailable. “Conflict” means records use incompatible methods, locations, classifications, or requirements. These states do not replace authorized disposition.

Review field	Normal	Missing	Conflict	Owner and next evidence
Heat, lot and product	Supplier, mill, lab, heat, lot, grade, form, revision, date, and chain of custody align	Heat, product form, grade, or custody link is absent	Report and purchase records identify different material populations	Material-control owner reconciles primary records
Sample plan and location	Count, parent item, location, depth, orientation, reduction ratio where applicable, and specimen IDs are traceable	Count, location, orientation, or representativeness rule is absent	Reports sample different positions but are treated as the same population	Quality owner obtains the approved sampling plan
Method and observed area	E45-25, ISO 4967:2026, or another exact method, variant, preparation, calibration, field rule, area, operator, and deviations align	Edition, variant, area, field selection, or software version is absent	Old and current editions or different variants are compared directly	Laboratory clarifies the controlled method record
Rating and images	Category, class, output, units, rounding, individual fields, images, outliers, and repeatability context align	Result expression, controlling field, or traceable image is absent	Different classification rules or outputs are treated as equivalent	Metallurgical reviewer reconciles method-specific evidence
Chemistry boundary	Morphology statements remain optical; separate analytical results are limited to identified particles and samples	A chemistry claim lacks an analytical record	Shape labels are asserted as definitive composition or origin	Metallurgist defines any additional analytical question
Limit and disposition	PO/specification, limit, sampling, rounding, deviations, authority, and release status are traceable	Acceptance source, limit, or decision owner is absent	Report method or output differs from the invoked requirement	Authorized quality owner decides clarification, retest, deviation, or disposition

A complete report links a defined specimen population to a current method, observed area, rating, and contractual limit. A missing location or sample count blocks a claim about representation. A report using an old ISO edition or different E45 method variant conflicts with a direct current-method comparison until the differences are reconciled.

The table can make a report ready for metallurgical and quality review while final acceptance remains pending. It does not identify chemistry from morphology, create a universal cleanliness limit, release a heat or lot, or establish the steel quality of any specific supplier.