How to Compare Two EV COAs in Ten Minutes
Reading one COA tells you what a supplier measured. Comparing two tells you which supplier to pilot with. This is a worked example: two fictional COAs (Supplier A and Supplier B, both for a 10 mL topical-grade EV material), read line by line the way a program manager would.
The two COAs, side by side
| Line item | Supplier A | Supplier B | What the difference means |
|---|---|---|---|
| Batch ID | EV-2606-0847, resolvable to batch records | Batch 88, no traceability statement | A: audit trail exists. B: you are trusting a number |
| Size distribution | 80–150 nm, NTA, per national guideline (No. 17), TEM cross-check | "~100 nm", method not stated | A: comparable across batches. B: not comparable to anything |
| Concentration | 1.0 × 10¹⁰ particles/mL, NTA | "1 × 10¹⁰ particles/mL" | Identical-looking; only A's is anchored to a method |
| Protein content | 0.8 mg/mL with ratio note | Absent | B may be hiding aggregates that inflate the count |
| Endotoxin | < 0.25 EU/mL, LAL, limit printed | "Tested, compliant" | B's "compliant" is a claim, A's is a measurement |
| Sterility | Negative, membrane filtration | "Sterile" | B did not say how |
| Payload check | Marker panel + protein ratio, per batch | "Proprietary blend" | B cannot prove what is inside the particles |
| Retain samples | Kept per batch, 24 months | Not stated | A can survive an audit; B cannot |
| Filings reference | DMF Type II 044197 · INCI 41316 | None stated | A's material chain survives EU/US diligence |
Where the real difference hides
The particle counts are identical — and that is the trap. Everything that decides whether these materials behave the same way in your product lives in the methods, limits and traceability lines. Supplier A's COA is a measurement document; Supplier B's is a brochure with numbers on it.
The ten-minute protocol
- Line up both COAs on the nine rows above.
- Strike any row where the method or limit is unstated.
- Count A's surviving rows versus B's.
- Ask the losing supplier one question per struck row. Evasive answers end the comparison.
The COAs are fictional; the protocol is not. Copy it. The COA it describes is real too: request a TONGNIAN sample COA and run the comparison against whatever is on your desk. Start from the seven-item reading guide if the rows feel unfamiliar.
Get the test panel that matches your batch
Tell us your target category and expected first batch size (a sample run, around 100 units, or 1,000 and up). We will send the matching test panel, characterization references, and a sample COA for your tier.