How to Read a PVC Resin COA: What Each Line Decides
A certificate of analysis is a lot record, not a spec. Which lines decide acceptance, which predict line trouble, and three common misreads.
A certificate of analysis arrives attached to a shipment, and it is routinely read as if it were a specification sheet. The two documents answer different questions. A specification describes what a grade is supposed to be across all production; a certificate records what one lot measured on the day it was tested. Confusing them is the source of most arguments about material that turns out to be within specification and still unsuitable for the line it was delivered to.
1. Who issues the document and what it covers
The certificate is issued by the laboratory of the plant that produced the lot, and it travels with the goods from there. A trading company passes it on; it does not generate the numbers. That distinction matters when a figure is questioned, because the answer comes from the producing site, not from the seller.
What the document covers is one lot, identified by a batch number and a production date. It does not cover the grade in general, it does not cover the next lot, and it does not describe how the material will run on a specific extruder. Those three limits are worth keeping in mind before any single value is treated as a promise.
2. The lines that identify the lot
Before any test value is read, four fields should be checked, because a certificate that fails here makes the rest irrelevant.
| Field | Why it decides things |
|---|---|
| Grade designation | Must match the grade on the bag, the invoice and the contract; SG-5 on one document and a model number on another is a traceability break |
| Batch or lot number | The only link between the document and the material in the warehouse |
| Production date | Sets which edition of the standard and which method edition the lot was legitimately tested to |
| Standard and method editions | GB/T 5761-2018 for the grade bands; the viscosity method changed in 2026, so the edition on the certificate dates the testing |
The method question is not pedantry. GB/T 1632.2-2025 replaced GB/T 3401-2007 for the dilute-solution viscosity measurement, and the older standard was withdrawn on 1 May 2026. A lot produced after that date carrying the withdrawn edition is either older stock or a document assembled from a template, and both possibilities are worth clearing up before shipment.
3. The lines that decide acceptance
These are the figures where a value outside the band stops the delivery; it does not merely complicate it.
| Item | SG-5 band under GB/T 5761-2018 | What a failure means commercially |
|---|---|---|
| K value / viscosity number | 66 – 68 / 107 – 118 mL/g | The grade is not the grade that was ordered; end product risk follows |
| VCM residue | ≤ 5 ppm | Regulatory and food-contact exposure; buyers in regulated markets cannot accept it |
| Volatile matter incl. water | ≤ 0.40 % | Moisture carried into the extruder; bubbles, voids, weight loss against the invoice |
| Impurity particles | ≤ 16 per 100 g | Specks visible in pipe and sheet; cosmetic rejections and customer claims |
| Whiteness at 160 °C for 10 min | ≥ 78 % | A thermal stability proxy; a low figure predicts colour drift and burn marks downstream |
The K value is the one that carries the most weight and the one most often read loosely. A measured 68.0 against a band of 66 to 68 is compliant. Whether it runs on a line validated at 67 is a different question, and the contract answers it, not the certificate.
4. The lines that predict trouble without failing
The remaining figures rarely stop a shipment, and they are the ones that explain a bad week on the line.
Apparent density, at ≥ 0.48 g/mL for SG-5, decides how much resin each screw flight carries and how the material behaves in a silo. A lot at the bottom of the band feeds more slowly than the previous one, and the operator compensates with temperature, not with the feeder.
Plasticiser absorption, at ≥ 19 g per 100 g, sets how quickly a dry blend takes up plasticiser. In a rigid pipe compound it barely registers. In a soft formulation it changes the mixing cycle, and a low value shows up as a longer blend time or an uneven batch.
Sieve residue at 0.25 mm, ≤ 1.6 %, and sieve pass at 0.063 mm, ≥ 97 %, describe the particle size distribution at both ends. Coarse residue means slower fusion and a risk of gel particles; a low fine fraction means the resin wets out differently in the mixer.
Fish eyes, ≤ 20 per 400 cm², matter most in film and sheet, where each one is a visible defect. In pipe they are less visible and still present as surface interruptions that a buyer will find.
5. Three ways the document is misread
The first is treating one certificate as a specification. A single lot record says nothing about the spread across lots, and the spread is what the line experiences over a month.
The second is comparing numbers across methods. A viscosity number in millilitres per gram and a K value are not interchangeable readings, and a certificate tested under the withdrawn method edition is not directly comparable with one tested under the current edition without noting the difference.
The third is reading a single measurement as a distribution. Each figure on the certificate comes from one sample of one lot. Where the value sits relative to the band is the useful information; the value alone is not.
6. What the certificate cannot say
It cannot describe gelation on a particular screw, because gelation is a function of the screw, the temperature profile and the residence time as much as of the resin. It cannot predict long-run thermal stability under a given stabiliser package. It cannot say whether the lot will hold output at a specific die. And it says nothing about the next lot, which will carry its own document with its own numbers.
What it can do is close a delivery against agreed numbers and, read in sequence, show a supplier's consistency over time. Three consecutive certificates from the same plant, compared line by line, describe the capability more honestly than any single document. The bands each line is being compared against are set out on the specification table, and the questions buyers raise most often about documents and origin are answered on the FAQ page.
SG-5 for the next production run
The grade, the tonnage and the destination port produce a price, a specification sheet and the full document list, usually within 12 hours (GMT+8).