Security & Anti-Counterfeit Labels — Testing & Validation Protocol

What a Batch Release Failure Actually Looks Like #

A brand partner running a pharmaceutical supplement line received 200,000 holographic tamper-evident labels. Incoming QC at their 3PL passed them — visual check, no obvious delamination, holograms looked right. Six weeks later, their Asia-Pacific distributor flagged that roughly one in twelve labels on shelf wasn’t responding correctly to their authentication app. The problem wasn’t counterfeiting. The labels were genuine. The UV-fluorescent authentication ink in zone 2 had been applied at 1.8 g/m² instead of the specified 2.4–2.6 g/m², putting emission intensity below the app’s 380 nm excitation threshold. No test at goods receipt would have caught it without a calibrated UV densitometer.

That scenario captures exactly what this article addresses: the difference between checking that a security feature is present and verifying that it performs to specification. These are not the same test, and most goods-receipt protocols only do the first.

The root cause in that case was an adhesive coat weight drift compounded by a press recalibration that hadn’t been logged against our internal SL-CAL Record form. The correction took four days. The reprint, reshipment, and distributor credit cost the brand owner considerably more.

The Parameters That Predict Authentication Reliability #

Six parameters govern whether a security label will authenticate correctly across its working life. Most QC plans cover two or three of them.

Hologram diffraction efficiency should read ≥28% at 550 nm on a calibrated goniospectrophotometer. Below 22%, the rainbow shift visible to the naked eye degrades enough that field staff start flagging genuine labels as suspect. We test every production reel at the 500-metre mark and the reel end — not just at reel start — because embossing pressure can drift as the foil roll tension changes.

UV-fluorescent ink emission intensity needs to be verified against the specified excitation wavelength, not just checked with a generic UV lamp. The specified window for most covert inks is 365–395 nm excitation with emission peak at 430–450 nm (blue) or 520–545 nm (green). Measuring the wrong wavelength gives a false pass. Our UV emission bench uses a 365 nm LED source calibrated against NIST-traceable standards quarterly.

Peel adhesion force on tamper-evident constructions is tested per ASTM D3330 Method F at 300 mm/min on the specified substrate. For standard BOPP facestocks on corrugated, acceptable range is 12–18 N/25mm. For polyester on glass, we target 16–22 N/25mm. Values below the lower bound risk non-activation on cold-chain substrates; above the upper bound on thin-wall PET bottles, the label tears the substrate rather than voiding cleanly.

Void message legibility after attempted removal is confirmed by applying the label to a standard stainless coupon at 23°C ± 2°C and 50% ± 5% RH (per ISO 2811 conditioning), dwelling for 24 hours minimum, then peeling at 180° in a single motion. The residual VOID text must be readable at 300 mm under 500 lux ambient. We do not accept “partially visible” — the message either reads or it doesn’t.

Sequential serial number integrity is verified 100% inline using our camera-based OCR system at 40 m/min web speed. Tolerance for digit skip or duplication is zero. This isn’t an AQL sampling issue — a single duplicated serial number in a run of 500,000 labels undermines the entire chain-of-custody function.

QR/barcode scan rate under ISO/IEC 15416 is tested at a minimum 90° and 45° presentation angle. We accept ≥99.5% first-pass decode rate. Anything below 98% gets quarantined; below 95%, the batch is rejected and root-caused before reprint.

The most commonly overlooked parameter is UV emission intensity, because it requires equipment most goods-receipt teams don’t own. Brands that rely on their contract manufacturer’s own test reports without periodic third-party verification have no real visibility into whether that parameter drifted between the certified sample and the production run.

Decision Framework — Matching Test Depth to Risk Level #

If your product is destined for a regulated category (pharmaceuticals, agrochemicals, alcohol with excise requirements), the authentication features on your labels fall under traceability obligations that exceed typical packaging QC. In the EU, pharmaceutical serialisation under the Falsified Medicines Directive (FMD) and EU 2023/607 requires tamper-evident features that can be verified at every dispensing point. For these runs, we apply our full QC-SL09 protocol: 100% inline serial verification, destructive peel testing on 32 units per batch per AQL 1.0, UV emission bench test on 10 units per reel, and a documented batch release record signed by a QC supervisor before shipment.

If your product is in a mid-risk category — consumer electronics, premium cosmetics, luxury spirits — where authentication matters for brand protection but not regulatory compliance, a tiered approach is defensible. We run 100% inline serial and barcode checks regardless, but destructive peel testing moves to AQL 2.5 (17 units per batch at normal inspection level II per ISO 2859-1), and UV emission is tested once per production order rather than per reel. This reduces test time by roughly 35% without meaningfully changing the defect escape probability for this risk profile.

For lower-stakes brand differentiation labels (visible hologram only, no covert features, no serialisation), AQL 4.0 on visual attributes is generally sufficient. The calibration investment changes accordingly — we still verify diffraction efficiency on each new embossing die, but not on every production reel.

Where the calculus genuinely changes: if your label combines a covert ink with a serialised QR code and the QR links to a live cloud authentication platform, every one of those six parameters affects system performance, not just label performance. A label that scans fine in the factory but fails UV authentication in the field creates a false-negative alert in your system — which is arguably worse than a label that simply doesn’t scan, because it flags genuine product as suspect and erodes consumer trust in the platform. For those constructions, we recommend the full QC-SL09 protocol regardless of product category.

Our sampling timeline for a new security label construction with all six parameters is typically 8–10 working days from confirmed specification receipt to first approved sample. Runs involving a new holographic embossing die add 5 working days for die qualification.

Specification Notes for Brand Partners #

When you brief us on a security label project, the minimum we need to develop an accurate quote and test protocol is: the substrate the label will be applied to (material, surface energy, texture), the authentication method (overt hologram, covert UV ink, serialised QR, or combined), whether the label must meet any regulatory serialisation standard, and the intended operating environment (temperature range, humidity, UV exposure during shelf life).

The gap that causes the most sample iterations is incomplete substrate information. “Applied to a bottle” tells us nothing useful — a frosted HDPE bottle and a clear glass bottle with a UV-coating varnish require different adhesive formulations, different void construct depths, and different peel force targets. A photo and a short description of the actual container, including any surface treatment, saves at minimum one sample cycle.

Our standard sampling timeline is 12–15 working days for a construction using existing embossing dies and qualified inks. First articles include a full QC data sheet covering all six parameters. If you need to supply the sample to a third-party authentication technology provider for system calibration, flag that at brief stage — we can produce a larger first-article run (typically 500 units) at no additional tooling cost.

Frequently Asked Questions

Can we use your test reports in place of our own incoming inspection?
Our batch release reports cover all six parameters documented in the QC-SL09 record, including calibration traceability. For regulatory categories like pharmaceutical serialisation, your incoming QC process will still need to satisfy your own GMP documentation obligations — our report supports that, but doesn’t replace it. For non-regulated categories, many brand partners accept our test data as the primary record.

What’s the minimum order quantity to justify the full QC-SL09 protocol?
The destructive sample requirement under AQL 1.0 consumes 32 units per batch. Below around 5,000 labels per order, the test sample represents more than 0.6% of the run, which starts to affect cost per unit noticeably. Below 2,000 units, we discuss with the brand partner whether a reduced protocol (AQL 2.5, 17 units) is acceptable given their risk profile, or whether consolidating multiple SKUs into a single production run makes more sense.

How do you handle calibration drift between production runs?
Each piece of test equipment — UV bench, goniospectrophotometer, peel tester — is on a calibration schedule logged in our metrology register. The UV bench and goniospectrophotometer are recalibrated quarterly against NIST-traceable references. The peel tester load cell is verified monthly with certified weights. If a calibration event falls within an active production order, we re-baseline against the last known-good reading before continuing. Drift events are logged under our CAPA-CAL procedure.

Does the 24-hour dwell condition for void testing match real-world application?
It’s a controlled baseline — it doesn’t replicate every real-world scenario. Cold-chain products applied at 4°C and then brought to ambient before attempted removal will perform differently from labels applied and removed at 23°C. For clients in cold-chain pharmaceuticals or frozen food packaging, we extend dwell testing to include a 4°C application cycle and 20-minute warm-up before peel, which is a protocol we developed after seeing inconsistent void activation on a chilled beverage project. Standard ISO 2811 conditioning covers the baseline; cold-chain application conditions need a separate qualifier.

What happens if a production batch fails serial number verification mid-run?
The press stops. The web is pulled back to the last verified good unit, the OCR system is re-zeroed, and a supervisor reviews the camera log to identify the first error position. The affected section is physically cut and quarantined. We do not resume and trim the bad section out at inspection — the serial sequence integrity has to be continuous from the confirmed start point. This typically adds 45–90 minutes to a run but is non-negotiable for any serialised security label.

Can your authentication test data integrate with our brand protection platform?
Our batch records are exportable in CSV and PDF. If your platform requires a specific data schema (for example, GS1 Digital Link format for QR traceability), we need to know that at brief stage so we configure our OCR capture accordingly. Serialised data in a non-standard format after production is complete cannot be retroactively restructured without re-inspecting the physical labels.

How do holographic embossing die qualification results carry over between orders?
A qualified die retains its qualification unless the embossing foil supplier changes, the foil lot changes significantly in thickness (more than ±2 µm from the qualified lot), or more than 6 months have elapsed since last production use. After 6 months of idle time, we run a 50-metre qualification strip before committing to a full production reel — a small cost against the risk of a diffraction efficiency failure on a 300,000-label run.

---

Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.

---