TL;DR: Switching ink systems mid-project without requalifying substrate adhesion is the most common cause of colour shift and delamination failures in flexible packaging — and it’s avoidable.
TL;DR: In a 2023 project with a personal care brand, we cut ink-related waste from 8.4% to 1.9% across a 12-month production run by switching from solvent-based to water-based gravure and requalifying three substrate-ink pairs.
What the Brand Was Seeing Before We Diagnosed the Problem #
The brief came in mid-2022. A mid-size personal care brand based in the Netherlands was sourcing laminated flexible pouches from a previous supplier and experiencing three recurring quality complaints from their 3PL and retail partners:
- Visible colour shift between production batches, described as a “warm shift” on their hero pink (Pantone 1895 C) — noticeable side-by-side under D65 store lighting
- Occasional delamination at the seal zone on BOPP/PE structures, concentrated at corners
- An ink odour complaint on one SKU, flagged by a retail chain with a restricted substances policy
Each of these looked like separate problems. They were not.
Diagnostic mapping of initial symptoms:
| Symptom | Probable Causes | Confirmed Cause |
|---|---|---|
| Colour shift batch-to-batch | Ink viscosity drift, pigment dispersion inconsistency, substrate surface energy variation | Viscosity not controlled to ±3 sec (Zahn #3); running ±8 sec in practice |
| Delamination at seal zone | Insufficient lamination adhesive, solvent retention under heat seal, ink layer incompatibility | Solvent retention from fast-evaporating ketone carrier — measured at 14 mg/m² vs. 5 mg/m² limit |
| Ink odour | Solvent retention, incorrect curing, contaminated ink batch | Solvent retention confirmed; residual MEK at 11 mg/m² on inner ply |
The incoming QC data from their previous supplier showed no AQL failures — which is exactly why it took eight months before someone connected the dots. AQL 2.5 visual inspection doesn’t catch 4 ΔE colour drift or 12 mg/m² solvent retention. Those need instrument measurement.
The Root Cause Most Teams Miss: Viscosity Drift During Long Press Runs #
The colour shift failure is the one I’d focus on first, because it’s the failure mode that gets misattributed most often. The brand’s previous supplier blamed pigment batch variation. Their ink supplier blamed substrate lot variation. Neither was the primary driver.
On gravure printing lines running solvent-based inks, viscosity is controlled by solvent addition — the press operator adds solvent during the run to compensate for evaporation. On a 6-hour continuous run, if the viscosity target is 18 seconds (Zahn #3) and the operator is checking manually every 45 minutes, the ink can drift from 16 to 22 seconds between checks. That 6-second window sounds narrow. On a gravure cylinder engraved to 55 L/cm with a 28-micron cell depth, that viscosity swing changes ink transfer by roughly 15–20%, which is more than enough to shift a process pink by 3–5 ΔE under D65. The Pantone 1895 C target was specified at ΔE ≤ 2.0 against the approved standard (measured per ASTM D2244), but production was regularly delivering 3.8–4.6 ΔE on the affected runs.
The confirmation method is straightforward: pull timed viscosity logs from the press control system and overlay them against the spectrophotometer ΔE readings for the same press run. In this project, the correlation coefficient between viscosity deviation and ΔE exceedance was 0.81 across 14 sampled production runs — enough to be definitive without needing a regression model.
The non-obvious part: the supplier had inline viscosity control on two of their four gravure presses but not the other two. The affected SKUs had been assigned to the uncontrolled presses. No one had flagged this during job scheduling because the press capability data wasn’t part of the supplier’s standard qualification package.
We log this type of press assignment risk under what we call a P-Class flag in our internal QC-F12 press scheduling review — any job with a ΔE tolerance tighter than ±2.0 gets restricted to presses with closed-loop viscosity control.
Corrective Actions: What We Implemented and What It Cost #
The remediation ran from Q1 to Q4 2023 across four phases:
-
Transition to water-based gravure ink system — the solvent retention problem and the odour complaint both pointed to the same underlying issue: the solvent carrier was wrong for this substrate-structure combination. Switching to a water-based acrylic system on the BOPP/PE structure eliminated the MEK retention issue entirely. Post-switch retention measured at 1.2 mg/m², well within the 5 mg/m² internal limit and compliant with EU 10/2011 migration thresholds for the outer ply. This change required full ink-substrate requalification across three pairs (BOPP/BOPP, BOPP/PE, and matte OPP/PE), which added 6 weeks to the project timeline but resolved two of the three symptoms permanently.
-
Closed-loop viscosity control on all active gravure presses — installed on the two remaining uncontrolled presses, holding viscosity to ±2 sec (Zahn #3) throughout the run. This is the highest-impact fix for colour consistency. After installation, ΔE variance on the Pantone 1895 C target dropped from a range of 1.8–4.6 to 0.7–1.9 across 22 production runs.
-
Substrate incoming inspection protocol — surface energy gate — we added dyne pen testing (38 dyne/cm minimum for BOPP) to incoming BOPP lot acceptance. This catches corona treatment variation before it reaches the press, which had been contributing to adhesion inconsistency at the lamination step. Based on our incoming inspection records from Q1–Q3 2023, roughly one in eight BOPP lots from the previous substrate source failed the 38 dyne/cm threshold. We sourced an alternate BOPP supplier, and the delamination complaints stopped within two production cycles.
-
G7 press calibration on colour-critical jobs — the brand’s retail packaging required consistent colour across three print sites. We implemented G7 grey balance calibration across all gravure cylinders used for the hero pink, bringing inter-run ΔECMC within 1.5. G7 certification isn’t standard practice on all our gravure work, but for packaging going into retail environments with fluorescent D65 lighting, the investment is warranted.
-
Waste tracking baseline and reduction — pre-intervention ink and substrate waste from make-ready and rejects was 8.4% of total material cost on these SKUs. Post-intervention, across the full 12-month 2023 production run, that figure settled at 1.9%. At the production volumes involved (approximately 2.4 million pouches annually), the material cost saving alone offset the press upgrade investment within 11 months.
Prevention: What to Specify Upfront #
If you’re briefing a flexible packaging supplier on a colour-critical laminated pouch, these are the items to include in the technical spec sheet — not to leave for the supplier to assume:
- ΔE tolerance against approved colour standard, measured instrument (specify D65/10°, per ASTM D2244)
- Ink system type required (water-based, solvent, UV) — with rationale if food-contact or retail odour compliance is relevant
- Residual solvent limit in mg/m² (specify inner ply vs. outer ply separately)
- Substrate surface energy minimum (dyne/cm) with test method
- Whether FSC chain-of-custody is required for the paper component, if any
Request the supplier’s press capability sheet showing which presses have closed-loop viscosity control and what their documented ΔE capability is per press.
Specification Notes for Brand Partners #
When you brief us on a flexible packaging project involving process colour or Pantone-matched brand colours, the two pieces of information we need before we can commit to a colour specification are: the approved colour standard (physical sample or digital reference with measurement condition) and the substrate structure you’ve approved or are considering. Ink-substrate interaction determines everything downstream, and a colour target that’s achievable on matte OPP may require a different ink loading on BOPP — sometimes a full ink reformulation.
The brief gap that causes the most unnecessary sample iterations is an unspecified ΔE tolerance. When a brief says “match the approved sample,” we interpret that as ΔE ≤ 2.0 under D65/10° as a default. If your brand standard is tighter (some cosmetic and pharma brands specify ΔE ≤ 1.5), tell us upfront — it changes press and cylinder selection.
Our standard sampling timeline for water-based gravure on BOPP structures is 18–22 working days from confirmed substrate and ink approval. Requalification after a substrate switch adds 8–12 working days. Colour approval rounds beyond the first extend this; most projects reach final approval in two rounds.
What is an acceptable ΔE tolerance for retail flexible packaging?
For most brand colours under D65 store lighting, ΔE ≤ 2.0 (ASTM D2244) is the standard threshold — at ΔE 2.0–3.0, trained observers can detect the shift; above 3.0, untrained consumers can. If your product sits on shelf next to itself in multi-unit facings, the tolerance should be tighter, because side-by-side comparison amplifies visible difference even at ΔE 1.8.
Can water-based gravure ink match the print quality of solvent-based on BOPP?
On properly corona-treated BOPP (38–42 dyne/cm), yes — with the caveat that drying tunnel configuration matters more for water-based systems. Slower water evaporation rate relative to ketone solvents means the drying sections need to be tuned per ink loading. On our gravure lines, we run water-based at 15–20% lower press speed than solvent-based on the same substrate to maintain equivalent dry ink film weight. Whether that speed reduction is acceptable depends on your volume and lead time requirements.
If a substrate lot fails the dyne pen test, does the whole lot get rejected?
Not automatically. Lots that fail at incoming (below 38 dyne/cm on BOPP) can be re-corona-treated in-line on our press, and then re-tested before the press run starts. We re-treat at 40 dyne/cm target and re-test three points across the web width. If the result is consistent, we proceed with a press operator flag to monitor adhesion on the first 200 metres. Lots that read below 32 dyne/cm at incoming are returned — re-treatment can recover 4–6 dyne/cm realistically, not 10+.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
