Materials Screening: Start With the Matrix and Acceptance Criteria

Screening is not analysis. Screening answers one question quickly — does this sample look inside or outside the expected range — and it is allowed to be approximate. The mistake is to build a screening bench that is precise but slow, or to expect a screening result to survive a customer dispute.

Two inputs decide the whole method: the matrix and the acceptance criteria. Both are written down before any instrument is chosen.

Write the acceptance criteria as a number

“Moisture must be low” is not a criterion. “Moisture below 0.8 percent at release” is. The number, the unit, the sample form and the point in the process where it is measured — those four turn a wish into a method. Without them, every out-of-specification discussion becomes an argument about the method rather than about the material.

The matrix decides the preparation

Powders, granules, pastes and liquids behave differently in the same instrument. A powder that cakes traps moisture and releases it slowly; a paste that skins over gives a false plateau; a liquid that splashes contaminates the cell. Preparation is where most screening error enters, and it is the part least often written down.

Describe the sample in the procedure: particle size, portion size, whether it is mixed, and whether it is measured as received or after conditioning. A result without that description cannot be compared with a result from another site.

First-line screening tools

When screening is not enough

Three situations end the screening stage. The result is disputed. The specification is tight enough that the method’s uncertainty covers the limit. Or the material is new, and no historical range exists to compare against. In all three the answer is a reference method, not a better screening instrument.

Make the first pass repeatable

A screening method earns its value by being repeatable across operators and shifts, not by being accurate in the absolute sense. Run it on a known sample at the start of each shift, record the value, and treat a drift as information about the instrument rather than about the material.

What to record

Sample identity, portion mass, preparation, instrument, programme, result and operator. Five of those six are cheap to record and impossible to reconstruct. A screening record without the portion mass is a number with no meaning, because the result depends on how much material was placed in the pan.