What do you do when a chemical trial result only partly passes?

This guide covers plant or field trials of a chemical product, such as a coolant, coating or cleaner, judged against criteria agreed before the trial. It does not cover regulated release decisions, such as drug batch release, where written procedures and the quality unit decide, or legal and safety limits, which no trial result can waive.

Which conditions decide what happens to a partly passed chemical trial?

Four conditions decide the outcome of a partly passed chemical trial: the weight of the missed criterion, the size of the miss against the test method's precision, the conditions the trial actually covered, and whether the decision rule was written down before the first result. Each belongs on paper before the trial starts, because once results arrive every party has a reason to read them differently.

Weight comes first. Mark every criterion as critical or minor in the trial plan, with one line on why. Precision comes second. For a property measured by a standard method, the method's repeatability and reproducibility set how far apart two honest results can be, and ISO 4259-2:2026 specifies how to use that precision to settle conformance when supplier and receiver results conflict, for homogeneous petroleum and related products (ISO 4259-2:2026). Scope comes third: list the lines, grades, materials and temperature range the trial ran on. The decision rule comes last. ILAC notes that one measurement can pass under one guard band and fail under a larger one, and concludes: "It is therefore expected that the decision rule is agreed before the measurements are taken." (ILAC-G8:09/2019, section 4). A rule agreed after the results is a negotiation, not a decision.

How do you tell a critical miss from a minor miss in a chemical trial result?

A critical miss in a chemical trial result is one that would change the decision to use the product if it held in service; a minor miss is one the user can accept or manage inside the tested conditions. Make that call per criterion before the trial, not per result after it. A practical test is to ask what happens if the miss is real and permanent.

Treat a criterion as critical when it carries any of these: a safety, health or environmental limit; a regulatory or customer specification limit; compatibility with the metals, seals, coatings or other chemicals in the process; or the main reason the trial was run, such as tool life in a coolant change. Treat it as minor when the miss affects handling, appearance, odor, convenience or a secondary performance target, and the product still does the job it was trialed for.

ASTM D3244-24 makes the same point in statistical form for petroleum products and lubricants. The degree of criticality of a specification, expressed as an agreed probability of acceptance, "shall have been mutually agreed upon between both parties prior to execution of actual product testing" (ASTM D3244-24, 4.3). A criticality label written after the result reads as a defense of the product.

Which decision fits each partly passed chemical trial result?

The decision for a partly passed chemical trial follows from the type of miss, and the table below maps each result to approval, conditional approval, retest or rejection, with the scope that goes with it.

The row that causes most disputes is a miss within the test method's precision. The decision there is the retest rule agreed before the trial, not a new limit. A retest rule can be as simple as a second laboratory testing a retained sample, or a staged plan written into the criterion itself. The European Pharmacopoeia dissolution test is a published staged plan: with Q set at 75%, stage S1 passes when each of 6 units reaches at least 80% dissolved; if not, 6 more units are tested and the 12-unit average must reach at least 75% with no unit below 60% (EDQM, 2019). What a plant trial borrows is the structure, not the numbers: how many extra tests are allowed and what each stage must show are fixed before the first result. A confirmed critical miss leads to rejection, whatever the other criteria show.

When does a partly passed chemical trial earn conditional approval?

A partly passed chemical trial earns conditional approval when every critical criterion passed, the miss is minor and real rather than measurement spread, and the conditions under which the product worked can be named. Conditional approval is a written, limited yes: it states what was tested, what missed, what control applies and when the condition is reviewed.

Write the conditions as the trial ran them, not as the product's technical data sheet describes them. If a coolant was trialed on two machining centers cutting one aluminum alloy at one concentration, the approval covers those machines, that alloy and that concentration. Another line, alloy or water supply is a new question.

Name the control for each minor miss. Illustrative examples: a foam miss handled with a defoamer dose limit, a residue miss handled with an extra rinse, an odor miss handled with a ventilation check. Give each control an owner and a review date written as a calendar date.

Set the exit. A conditional approval needs a point where it becomes full approval, is narrowed or is withdrawn: after a set number of batches, a set run time or a follow-up test. Without that point the condition is likely to be forgotten, and the product ends up in use beyond what the trial showed.

How do you record a missed criterion as a separate finding in a chemical trial?

Record each missed criterion in a chemical trial as its own finding, with the criterion as written before the trial, the result, the conditions, the judgment and the action, so that nobody can fold it into an overall pass. A finding that could change the conclusion belongs in the main trial report next to the summary, not in an appendix.

A finding record needs these fields:

• Criterion: the wording and limit fixed in the trial plan, with the test method and its edition.

• Result: every individual value, including failing ones, with sample identification and date.

• Conditions: line, product grade, material, temperature and concentration when the sample was taken.

• Precision check: whether the gap to the limit is larger than the method's reproducibility.

• Weight: critical or minor, as classified before the trial.

• Effect on the conclusion: whether the decision changes if the miss is real.

• Action: retest under the agreed rule, a control, a narrowed scope or rejection, with an owner.

Individual values matter because a summary number can hide the miss. FDA guidance for pharmaceutical laboratories puts it directly: "Reliance on averaging has the disadvantage of hiding variability among individual test results." (FDA, 2022). That guidance is written for drug production; in a chemical trial report the failing value is likewise what the reader needs.

What mistakes turn a partly passed chemical trial into a dispute?

Five mistakes turn a partly passed chemical trial into a dispute between supplier and user: retesting until a pass appears, changing a limit after the result, averaging away a failing value, approving beyond the tested conditions, and leaving a miss out of the report. Each breaks the comparison the trial was set up to make.

Retesting until a pass appears has a name in pharmaceutical quality control. FDA describes repeated testing until a pass appears, with failing results disregarded without scientific justification, and states: "This practice of “testing into compliance” is unscientific and objectionable under CGMP." (FDA, 2022). The same guidance says: "The number of retests should not be adjusted depending on the results obtained." (FDA, 2022). That guidance is FDA's nonbinding recommendation for drug manufacturers under US current good manufacturing practice (CGMP), not for chemical trials, but the reasoning carries over.

Changing a limit after the result voids the comparison, because the original limit was the question the trial asked. Averaging a failing value with passing ones hides the spread the trial needed to show. Approving a product plant-wide after it passed on one line claims more than the trial tested. Dropping a minor miss from the report removes the warning the next user needed.

What should you ask the supplier after a partly passed chemical trial?

Ask the supplier after a partly passed chemical trial what explains the miss, whether it could be measurement spread, what the product needs to pass under the failing conditions, and what the supplier will confirm in writing about the approved scope. The answers decide whether the next step is a retest, a narrowed approval or a different product.

• What causes the miss, in the supplier's view, and which data support that view? Ask for the technical data sheet (TDS), the certificate of analysis (CoA) for the trial lot and any comparable field cases.

• Was the trial lot inside its own specification? A CoA that puts the lot at the edge of its range changes how the result reads.

• Does the supplier's laboratory get the same result on a retained sample, with the same method and edition? When two laboratories disagree on a property of a petroleum product or lubricant, ASTM D3244-24 provides a practice for combining their results into an assigned test value and comparing it with an acceptance limit set before testing (ASTM D3244-24).

• What change would the supplier propose for the failing conditions, and would that change need a new trial?

• Will the supplier confirm in writing the conditions the approval covers?

When should a partly passed chemical trial go to a specialist?

A partly passed chemical trial should go to a specialist when the miss sits near a safety, regulatory or compatibility limit, when supplier and user laboratories disagree by more than the method's reproducibility, or when nobody can say whether the trial conditions matched the plan. Those cases need someone who can judge method, process and chemistry together.

Bring in a statistician when results come from several laboratories or many repeats; ASTM D3244-24 recommends that its dispute practice be conducted under the guidance of a qualified statistician (ASTM D3244-24, 4.6). Bring in the user's quality or safety function when a miss touches a legal limit, because a trial result cannot waive it. Bring in a laboratory accredited to ISO/IEC 17025:2017 when a statement of conformity has to hold up with a third party. When the customer requests such a statement, that standard requires the decision rule to be clearly defined and, unless inherent in the specification, "communicated to, and agreed with, the customer" (ISO/IEC 17025:2017, clause 7.1.3, quoted in ILAC-G8:09/2019). Lubinpla, an applied AI company that structures technical reasoning for specialty chemicals, works a trial result from the customer's own documents and site history: it checks each result against the criteria fixed before the trial, states the scope the pass covers, records misses that could change the conclusion as separate findings and returns one cited document for review, with the user approving each step.