Reading a Chinese Chemical COA: Test Methods, Raw Data, and What Each Parameter Actually Proves
· HARMONY TECHNOLOGY (ZHEJIANG) CO., LTD.
Reading a Chinese Chemical COA: Test Methods, Raw Data, and What Each Parameter Actually Proves
The short version:A certificate of analysis is only as strong as the method behind each number. A COA that lists measured values with methods and limits is verifiable evidence; one that says “conforms” is a claim. This guide explains what each common parameter on a Chinese chemical COA is actually measuring, which analytical method produces it, and how to tell a raw-data COA from a restated one.
1. Purity / Assay — GC-FID (or GC-TCD for permanent gases)
What the number is: the mass fraction of the target compound, reported as area-% or mass-% — these are not the same thing, and Chinese COAs frequently omit which basis was used.
Method: gas chromatography with flame-ionization detection (GC-FID) is standard for R134a, R22, R32, DCM, and most chlorinated solvents; thermal conductivity (GC-TCD) is used where FID response is weak.
What to look for on the COA:
| Field | Why it matters |
|---|---|
| Column & carrier gas | Determines whether the separation is even capable of resolving the impurity in question |
| Internal standard | Without one, area-% can overstate purity by several tenths |
| Limit of detection / quantification (LOD/LOQ) | A spec of “≥99.9%” is unverifiable if LOQ is 0.1% |
| Chromatogram reference | A batch number without a linked chromatogram is a restatement, not evidence |
Practical check: for refrigerants, moisture and non-condensable gas (NCG) limit real-world performance far more often than the assay figure does — yet they are frequently the fields left off a two-line COA.
2. Water Content — Karl-Fischer Titration
Method: coulometric KF for low-ppm work (<100 ppm), volumetric KF for higher ranges.
Typical limits: refrigerants ≤10 ppm (often ≤5 ppm); technical-grade DCM ≤0.05 % (500 ppm); solvent / industrial grade ≤0.02 % (200 ppm); pharmaceutical grade ≤0.01 % (100 ppm) — the same three-grade table used in the DCM specification article, and the only grade table we quote.
What the number proves: water is the single most common cause of acid formation and system corrosion in refrigeration circuits. A ppm figure is only meaningful alongside the sampling point (cylinder valve vs. bulk tank) — KF results shift with handling, so an unqualified “8 ppm” tells you little without knowing where the sample came from.
3. Acidity — Titration (HCl equivalent)
Method: acid-base titration, expressed as ppm HCl, or as mg KOH/g (acid value).
Why it is on a refrigerant COA at all: residual acidity predicts lubricant degradation and metal attack in a running system. Limits are typically ≤1 ppm (HCl) for refrigerant grades. What water and acidity actually do inside a POE system — hydrolysis, free fatty acids and copper plating — is worked through in R134a system compatibility.
Common gap: acid value is often reported as a single figure with no statement of indicator or potentiometric endpoint. Potentiometric titration is less operator-dependent — worth asking which was used when results hover near the limit.
4. Water Content vs. Non-Condensable Gas — Distinguish Them
Both degrade performance; both appear as “moisture-related” complaints. They are measured differently:
| Parameter | Method | Typical limit (refrigerant) |
|---|---|---|
| Water | Karl-Fischer | ≤10 ppm |
| Non-condensable gas (NCG) | Gas burette / GC-TCD after liquid-phase sampling | ≤1.5 % vol (varies by grade) |
A COA listing “moisture” without clarifying water vs NCG is ambiguous — the two have different causes (handling vs. filling/vacuum process) and different fixes.
5. Residue / Evaporation Residue — Gravimetric
Method: evaporate a weighed sample, dry to constant mass, weigh the residue.
What it proves: non-volatile contamination (lubricants, particulates, high-boilers). For solvent grades it is stated as ≤10–50 ppm; for refrigerant grades it is usually absent because the product is distilled, but its absence from the COA is not the same as a passing result — if it is not listed, it was not tested, and that should be stated as such.
6. Batch Number ↔ Drum Label ↔ COA: the Consistency Chain
The only way a COA becomes evidence rather than paperwork is a three-way match:
COA batch number == drum/cylinder label == shipment packing list line
What breaks the chain in practice: a COA covering a production batch while the shipment is a repackaged lot. This is not necessarily fraud — but it changes who issued the QC record (producer’s lab vs. filler’s lab), and the COA should say which.
Practical verification: ask for the labels of two drums within the same shipment — batch numbers that differ across one shipment are normal for repacked product and should be reflected as separate COA lines.
7. COA ↔ MSDS ↔ Regulatory Filing: cross-document arithmetic
A COA is a batch document; an MSDS/SDS is a product document; a REACH registration or F-gas quota filing is a company/volume document. The three must be arithmetically consistent:
| Cross-check | Failure looks like |
|---|---|
| CAS number identical across COA / MSDS / label | Different CAS for the same trade name (e.g. DCM: 75-09-2; R134a: 811-97-2) |
| UN number identical across COA / MSDS / transport docs | UN 1593 (DCM, Class 6.1) vs. UN 3159 (R134a, Class 2.2) — wrong UN means wrong packing instruction |
| Composition in MSDS matches COA assay range | MSDS declaring 99.9 % while COA shows 99.5 % |
| Quota eligibility matches the product’s HFC status | R134a/R32/R410A are HFCs (quota-constrained under the Kigali phase-down and, for the EU market, the 2026 quota year of Regulation (EU) 2024/573); R22 is HCFC (Montreal phase-out, different regime); R404A is an HFC blend |
8. What a COA Cannot Tell You
Technical literacy includes knowing the document’s limits:
- It is a sample, not the shipment — it certifies the batch the sample came from, under the condition the sample was taken.
- It does not certify packing integrity — no COA field reports valve torque, cylinder hydrostatic test date, or linings. Those live in UN packaging marks and periodic inspection records.
- It does not certify origin compliance — origin, quota consumption, and destination-market admissibility are separate filings.
Technical reference for importers and QC engineers. Product grades, limits and quota treatment vary by destination market and by year — figures above are stated as commonly published specification ranges, and current quota/phase-down values should be taken from the applicable regulation for the year of shipment.
What we can provide
On request we can return the records behind the certificate rather than a restated one:
- Measured values with methods — assay (GC-FID), water (Karl-Fischer, coulometric or volumetric), acidity (titration to a stated endpoint) and residue on evaporation (gravimetric), each with its limit, its result as measured, and the method used.
- Raw records — the chromatogram and integration parameters behind the assay figure, and the titration record behind the water figure, where the testing laboratory retains them.
- Sampling point and date — where and when the sample was taken (cylinder valve, bulk tank or line), since a Karl-Fischer figure is not comparable without it.
- Batch-to-label traceability — the batch numbers the certificate covers, matched against the drum or cylinder labels in the shipment; repacked lots are listed as separate certificate lines.
Contact us with: the batch numbers, and the two or three parameters that decide acceptance for your use.
Prepared by HARMONY TECHNOLOGY (ZHEJIANG) CO., LTD. — trading company; the group’s operating entity for this product line.
Need current specs, quota status, or a mixed-load quote for Test Methods, Raw Data, and What Each Parameter Actually Proves? Contact sales@hm-chem.com with your spec & destination port.