Same specification, different plant: what can still change

Technical article · Eapearl Chemical · 15 December 2025

Two lots can meet the same specification, carry equally correct certificates, and behave differently in a process. Understanding why is the difference between a smooth source change and an investigation.

A specification is a filter, not a description

A specification lists properties somebody decided to control, with limits chosen to exclude material that would not work. It never describes the material completely. Everything outside the list is unconstrained, and everything inside the list is constrained only to a range, which means two lots at opposite ends of the same range are both compliant and not identical.

For most of a material’s life this is invisible, because successive lots come from the same plant running the same chemistry, and the unspecified properties stay quietly constant. A change of manufacturing site, of synthesis route, of a key raw material or of a catalyst removes that quiet constancy without touching a single number on the certificate.

What actually differs between two origins

• Trace impurity pattern — different by-products, at levels below reporting thresholds, carried forward from different starting materials and different separation steps.
• Residual metals — catalyst traces that can matter enormously in a downstream reaction and not at all in a cleaning duty.
• Isomer and homologue ratios — where a product is a mixture by nature, the proportions can shift while the assay total does not.
• Water and acidity behaviour — different drying and finishing equipment gives different starting points and different tendencies during storage.
• Colour and odour precursors — components invisible in a fresh sample that develop on heating or ageing.
• Physical form — for solids, crystal habit, particle size distribution, caking tendency and bulk handling behaviour, none of which the chemical specification usually mentions.
• Stabilisers and additives — identity or level may differ, and a stabiliser that suits one storage regime may not suit another.
• Packaging and its history — container type, liner, headspace and the transport route all affect what arrives.

Why identical certificates can precede a failed batch

The failure mode is consistent across industries. A process was developed on material with a particular unmeasured characteristic, and the process quietly came to depend on it. Perhaps a trace component acted as a mild inhibitor, or a particular particle habit dissolved at a rate the equipment was tuned to, or a low background acidity suited a pH adjustment step. Nobody recorded the dependency, because nobody knew it existed.

When the material changes origin, the dependency is broken and the symptom appears somewhere far from the incoming inspection: a slower reaction, a colour drift, a filtration that takes longer, a finished article that fails on ageing. Because the certificates conform, the investigation looks everywhere else first. The cost is not the material; it is the weeks spent before anyone re-examines the assumption that conforming means unchanged.

Designing the bridging comparison

1. Decide the sensitivity class first. Bulk duty, sensitive process, or regulated end use. The class sets the depth of everything that follows.
2. Secure retained samples of the outgoing material before the changeover. This is the step that cannot be recovered afterwards, and the one most often skipped.
3. Compare analytically beyond the specification. Run both materials by the most informative technique available for the chemistry, looking at the whole trace profile rather than only the controlled parameters.
4. Run a laboratory trial in the actual system, old and new in parallel, with everything else held constant.
5. Run a monitored plant trial on a defined batch, with additional sampling and a written rollback point.
6. Watch the ageing properties of whatever is produced, because several of the differences described above only reveal themselves over time.
7. Record the conclusion against the material’s internal record, including what was compared and what was found equivalent, so that the next change starts from knowledge rather than from memory.

Sensitivity varies more than people expect

The same substance can occupy several sensitivity classes inside one company. Propylene glycol used to thin a technical blend tolerates a great deal of variation; the same molecule in a formulation intended for human contact is governed by monograph compliance and documentation that are tied to the manufacturing arrangement rather than to the molecule. Glycerol from different feedstock origins meets the same purity monograph while differing in origin declarations that some customers must pass on. A solvent such as acetone used for equipment cleaning raises almost no questions, while the same solvent used as a reaction medium raises several.

The practical instruction that follows is to classify by use, not by substance, and to accept that one material may need three different responses to the same announcement depending on where it goes inside the plant.

What to ask, and what to expect in return

A well-handled announcement states what changed and, equally importantly, what did not. The buyer’s questions should establish whether the change is of location, chemistry, raw material or all three; whether specifications and analytical methods are unchanged; whether extended comparative data exist; whether pre-shipment samples are available; whether the previous material remains available during a transition period; and which documents are reissued. Documents tied to a manufacturing location deserve specific attention, because origin statements, regulatory registrations held for a site, and audit certificates do not automatically travel with a molecule to a new plant.

Habits that reduce the exposure permanently

Three habits turn this from a recurring crisis into routine work. Keep retained samples of every material in every sensitive use, indexed and actually retrievable. Record known process sensitivities as they are discovered, so that the dependency on an unmeasured property is written down somewhere before it is broken. And hold specifications that reflect what the process genuinely needs, revising them when a bridging exercise reveals a parameter that mattered and was not controlled. A specification that has been improved after each change eventually becomes a description of the material that works, which is what everyone assumed it was in the first place.

Questions about manufacturing arrangements, comparative data and pre-shipment samples for a named application can be raised through our contact page.