Is it really necessary to test a lot that arrives with a certificate?

The certificate describes what the supplier measured on what the supplier sampled. It is a serious document and a good supplier stands behind it, but it cannot describe what happened afterwards: a mislabelled container, a transfer into a vessel that held something else, moisture picked up through a poor closure, or a delivery that is simply not the one you ordered. Goods-in testing is not an accusation aimed at the supplier. It is the step that confirms the material in front of you is the material the certificate describes.

What is the minimum test worth running on every lot?

An identity test, on every container or on a statistically defensible selection of them. Identity is the only failure that is catastrophic rather than merely expensive, because a wrong material can pass unnoticed into a batch and out of the plant. Infrared spectroscopy against a reference spectrum is fast and decisive for a solvent, and gas chromatography adds the profile that distinguishes close relatives. Everything else in the panel is a quality question; identity is a safety question.

How should containers be sampled?

Under a written plan that says which containers are opened, where in each container the sample is taken, with what equipment, into what container, and by whom. Composite samples save laboratory time but destroy the information you need if the result is out of specification, because a composite cannot tell you which drum was the problem. A common compromise is identity testing per container and the full panel on a composite, with individual samples retained unopened until the lot is released.

What happens when a result falls outside the specification?

The lot stays in quarantine and the first question is whether the result is real. Confirm the method, the standard, the instrument and the sample before confirming the material, because laboratory error is a common cause of an out of specification result and the investigation is much cheaper than a rejected delivery. If the result holds, record it, inform the supplier with the data, and decide formally whether to reject, return or use under a documented concession. What must not happen is a quiet retest until a passing number appears.

Why keep retained samples if the lot has already been released?

Because the questions that matter most arrive later. If a finished batch behaves strangely months after production, the retained sample of the solvent lot used in it is the only way to distinguish a raw material cause from a process cause. Retention also lets you build a trend for parameters that drift slowly, such as acidity or colour, which no single certificate can reveal. Store retains sealed, labelled with the lot, dated, and under the same conditions as the bulk.

Goods-in testing of a solvent delivery, step by step

Technical article · Eapearl Chemical · 29 April 2026

Goods-in testing is the last point at which a raw material problem is cheap. After the lot is released into production it is no longer a delivery question; it is a batch investigation, and it costs a different order of magnitude.

What the certificate can and cannot tell you

A certificate of analysis is a claim about a specific quantity of material, measured at a specific time, by a laboratory you did not observe, on a sample you did not take. For a reputable supplier it is normally accurate, and treating it as dishonest is neither useful nor fair. The gap it leaves is different: it says nothing about what happened between that measurement and your yard.

The events in that gap are mundane and well documented across the industry. A container gets the wrong label. A grade close in name to the one ordered is picked from the warehouse. A closure is not fully seated and a hygroscopic solvent takes on moisture in transit. A part-load is consolidated and the paperwork follows the wrong pallet. None of these is exotic, and none of them is visible on a certificate.

Sampling decides everything that follows

The laboratory result can only be as good as the sample, and sampling is where most goods-in systems are weakest. Write the plan down and follow it rather than improvising per delivery.

How many containers — decide in advance and record the rule. Identity on every container is the strongest position for a material where a mix-up would be serious.
Where in the container — a sample drawn only from the top of a drum misses stratification and any settled water layer; use equipment that reaches the body of the liquid.
Clean, dedicated equipment — a sampler carrying residue from a previous material invalidates everything downstream, and for a water-sensitive solvent, damp glassware is enough to move a result.
Sample containers — tightly closed, of a material compatible with a glycol ether, filled with minimal headspace and labelled at the point of sampling rather than later.
Traceability — every sample carries the lot, the container number, the date and the sampler, or the investigation that needs it will fail.
Safety — opening containers is the step where exposure actually happens, so sampling of a flammable solvent belongs in a controlled, earthed, ventilated location.

Identity first, quality second

Order the tests by consequence. Identity is checked first because getting it wrong is the only failure that can leave the site inside a product. For a solvent, infrared spectroscopy against a retained reference spectrum is quick and unambiguous for gross identity, and gas chromatography adds the compositional profile that separates a material from its close relatives.

Close relatives are exactly the risk. In the propylene series, propylene glycol ethyl ether sits beside its methyl homologue and beside the corresponding acetates, all of which share a naming pattern and are routinely abbreviated in correspondence. The product also normally exists as an isomer mixture, so an identity test should be judged against a reference of the same grade rather than against a single pure isomer, and the acceptance criterion should say so.

The short panel worth running on a routine lot

Appearance — clear, free of visible water and of particulates. Trivial to run, and it catches container and handling problems nothing else will.
Identity — spectroscopic, with a chromatographic profile where the grade demands it.
Assay and profile — the main component with named impurities reported individually rather than folded into a single total.
Water — normally by Karl Fischer titration, with the method stated on both your specification and the supplier’s.
Acidity — an early indicator of ageing and of contamination, and a parameter worth trending across lots rather than reading in isolation.
Colour — on a platinum-cobalt scale; drift usually points at thermal history somewhere in the supply chain.
Peroxides, where the chemistry calls for it — ethers can form peroxides with air and time, so a lot of uncertain age or history deserves a check before it enters a process with any thermal step.

Match every line to a method and a reference. A specification that states a limit without naming the method is an argument waiting to happen, because the supplier and your own laboratory can each be right against different procedures.

When a lot fails

The lot stays in quarantine and the investigation starts with your own laboratory, not with the supplier. Verify the standard, the calibration, the instrument condition and the sample handling, and repeat under a documented plan rather than informally. A repeat performed because the first number was inconvenient, with no record, is the single worst habit a goods-in laboratory can acquire; it converts the whole system into decoration.

If the result holds, notify the supplier with the data, the method and the lot identity, and ask for their retained sample result on the same parameter. Then decide formally: reject, return, downgrade to a less demanding use, or release under a written concession approved by whoever owns the product. Record the decision with a reason. Over time the collected records are also your supplier performance data, and a supplier who answers investigations quickly and with their own numbers is worth more than one who is a little cheaper.

Retains, trends and the supplier record

Keep a retained sample of every lot, sealed and stored like the bulk, for a period defined against the shelf life of the material and the life of the products made from it. Retains answer the question that no forward-looking test can: whether an anomaly discovered later came in with the drum or arose in the process.

Trend the numbers as well as filing them. A parameter that is inside specification but moving steadily is information, and it appears long before any single certificate fails. Water and acidity are the usual early indicators for an ether, and colour is the usual indicator of handling. A comparison against a stable reference grade such as propylene glycol in a neighbouring service can be useful for separating a supply-side drift from a change in your own laboratory. Specifications, method details and reference samples for setting up a goods-in panel are available through our contact page.