No steel is perfectly clean. Deoxidation products, sulfides and refractory fragments survive the melt as non-metallic inclusions — microscopic foreign phases embedded in the solid steel. In most applications a normal population does no measurable harm; in bearings, springs, gears and anything else that lives under cyclic stress, one oversized inclusion in the wrong place is a fatigue crack's birthplace. Inclusion rating is how steel cleanliness gets a number a purchase order can reference.
The four types of ASTM E45
The classical chart method — often called the JK method after its Jernkontoret origins — sorts inclusions by morphology into four types:
| Type | Chemistry (typical) | Morphology on the section |
|---|---|---|
| A — Sulfide | Manganese sulfide | Gray, ductile stringers elongated in the rolling direction |
| B — Alumina | Aluminum oxide | Broken rows of small dark angular particles ("strings of beads") |
| C — Silicate | Silicate glasses | Dark elongated stringers, often continuous and sharp-ended |
| D — Globular oxide | Complex oxides | Isolated, roughly round particles, randomly scattered |
Each type is rated in two width classes — a thin series and a heavy series — because a wide sulfide stringer is a very different defect from a hair-thin one of the same length. The result of a rating is therefore not one number but a small table: severity per type, per series.
How chart severity works
The method is refreshingly literal. The polished specimen is examined unetched at 100× magnification, field by field. Each field's inclusion content is compared against reference severity pictures — increasing half-step severity levels showing progressively more inclusion content for each type and series. The field gets the severity of the chart picture it matches. In the traditional worst-field approach, the specimen's rating is governed by the most severe fields found for each type; the certificate lists something like "A 2.0 thin, B 1.5 heavy, D 1.0 thin".
What image analysis changes is the drudgery and the consistency, not the logic: inclusions are thresholded automatically against the bright matrix, stringers are measured in calibrated µm, morphology sorts particles into their types, and every examined field gets its severity recorded — so the "worst field" is genuinely the worst of all fields examined, not the worst of the fields a tired operator happened to study closely.
Reading the type table like a steelmaker
The type breakdown is a diagnostic, not just a grade. Heavy Type A points at sulfur control and Mn/S balance. Type B rows are the signature of aluminum deoxidation practice and clogged-nozzle events. Type C tracks silicon deoxidation. A scatter of Type D globulars is often calcium treatment doing its job — converting harmful stringers into rounder, less damaging particles. When a customer complains about a rating, the type that moved tells the melt shop where to look.
Practical notes for the lab
- Polish, don't etch. The rating is read on the as-polished surface; etching destroys the contrast the method depends on. Edge-retention and scratch-free polishing matter — a comet-tailed scratch can masquerade as a Type C stringer.
- Sample where the spec says. Inclusion content varies through a billet's cross-section; longitudinal sections at specified locations are the norm. A rating from the wrong plane is a different measurement.
- Mind the deformation direction. Types A, B and C are defined by elongation along rolling; on a transverse section the same inclusions read as globular and the rating loses its meaning.
- ISO and DIN variants exist — different charts, similar logic. Customers specify them by name; use the chart family the order calls for and say so on the certificate.
The one-sentence summary
An inclusion rating is a standardized comparison of what's in your steel against reference pictures of how bad it could be — done type by type, series by series, field by field. Automate the seeing, keep a metallurgist behind the judgment, and the number on the certificate becomes one you can defend in front of any customer's quality team.
Rate a heat you already certified
Compare the module's field-by-field severities against your bench rating — same sample, same charts.