Testing chemicals, oils and polymers: specifications and certificates of analysis

A specification is the agreed list of tests, methods and limits that a material must meet. For chemicals and materials it is usually a commercial agreement between supplier and customer, often built on a national or international product standard (an ISO, ASTM, BIS or Codex standard) or on the supplier's own grade data sheet.

Unlike a pharmacopoeial monograph, there is rarely one official specification for, say, toluene or polypropylene. Grades differ by purity, colour, melt flow, additives and end use. That is why the specification must name the exact test method and condition for every limit: '60° gloss 85 ± 5 GU' or 'MFR (230 °C/2.16 kg) 12 ± 2 g/10 min', not just 'gloss 85' or 'MFI 12'.

• Identity: is it the right substance or grade? (for example iodine value of an oil, GC fingerprint of a solvent)
• Purity and impurities: how much of the main component and of named impurities?
• Physical properties that matter in use: colour, density, melt flow, hardness, gloss
• Fitness for purpose: properties linked to the customer's process, such as acid value before biodiesel production

Each material family has a small set of core tests that appear on almost every certificate. Learning these covers most of the routine work in a chemicals or materials QC lab.

The same chemical can be sold in several grades, and the grade tells you which specification applies. A technical grade may be 98% pure with no limit on trace metals; an analytical reagent (AR) grade carries tight limits on many impurities; a food or pharmaceutical grade adds limits and documentation required by food or drug law.

Never assume one grade can replace another. A technical-grade solvent that passes purity may still contain an impurity that ruins a downstream reaction or fails a food-contact rule.

• Technical / industrial grade: main component and a few key properties
• Laboratory or reagent grade: purity plus selected impurity limits
• Analytical reagent (AR / ACS-type) grade: tight purity and impurity limits for analytical use
• Food, cosmetic or pharmaceutical grade: limits from the relevant regulation or compendium, plus traceability and change control

A result is only as good as the sample. Liquids can stratify, solids can segregate by particle size, and volatile impurities can escape from a half-empty bottle. Use the sampling plan in the specification or the relevant sampling standard, and use clean, inert containers.

1. 1Check the label, lot number and container condition before opening.
2. 2Mix or circulate liquids where possible; take top, middle and bottom samples from large tanks if the plan requires.
3. 3Melt and mix solid fats completely before sampling; take granules from several bags or a moving stream for polymers.
4. 4Fill sample bottles nearly full and close them at once to limit evaporation and oxidation.
5. 5Label with material, lot, date, sampler and sampling point.
6. 6Keep a retained sample in the same type of container under suitable storage, for the period your SOP sets, in case of a dispute.

For oils, sample and test quickly and protect from light and air. Peroxide and acid values rise as the sample oxidises or hydrolyses.

A certificate of analysis (CoA) is the supplier's statement that a specific lot was tested and met the specification. It is not a guarantee that your own tests will give the same numbers, but it should be traceable and complete.

• Supplier name and address, product name and grade, lot or batch number, manufacture and expiry or retest date
• Each test with its method reference (standard and condition), the specification limit and the actual result with units
• Results, not just 'complies', for key quantitative tests
• Date of testing and the name or signature of the authorised person; the lab's accreditation where relevant (ISO/IEC 17025)
• Statements on storage and, for food-contact or regulated grades, compliance declarations

Red flags: identical results on every lot, results printed with more decimal places than the method supports, a missing method or test condition, or a CoA dated before the lot was made.

Many buyers test only a few key parameters on each lot and rely on the supplier CoA for the rest, after qualifying the supplier. Full testing is done on the first lots and then periodically. The choice of identity and key tests should reflect risk: what would happen in your process if the material were wrong?

Small differences are normal. Every standard method has a repeatability (same lab, same operator) and a reproducibility (different labs) limit, often written as r and R in the precision section. If the difference between your result and the supplier's is within R, the two labs agree within the method's capability.

If the difference is larger, investigate before rejecting the lot. Treat it like an out-of-specification investigation: first check your own work, then compare methods, then compare samples.

1. 1Check calculations, units, instrument calibration and reference materials in your lab.
2. 2Confirm both labs used the same method, edition and condition (for example ISO 1133 at the same temperature and load; gloss at the same geometry).
3. 3Retest your retained sample, ideally by a second analyst.
4. 4Exchange samples with the supplier, or send a split sample to an independent accredited lab.
5. 5Record the outcome and, if needed, agree an improved method or specification.

Customers trust a CoA because the lab works under a quality system. In chemicals and materials testing this is usually ISO/IEC 17025 accreditation for the lab, inside an ISO 9001 quality system for the company. Both expect calibrated instruments, validated or verified methods, trained analysts, controlled records and participation in proficiency testing.

• Calibration: for example temperature and load on melt flow indexers, spring force on durometers, gloss tiles and white tiles on colour instruments
• Control samples: a reference polymer, reference oil or check tile measured routinely and plotted on a control chart
• Proficiency testing: inter-laboratory schemes for oils, fats, polymers and coatings show how your lab compares with others
• Data integrity: raw data such as chromatograms, titration volumes and weighings kept with the result