Can Carbonation Integrity
Illustrative case study. Company names and identifying details are fictional. Technical details reflect real industry practice and current regulation.
Carbonation is the beverage can’s structural system: dissolved CO₂ pressure keeps the lightweight aluminum rigid, and when it’s lost, the can goes soft — denting, buckling, failing. But it’s also a microbial signal: in a properly carbonated acidic beverage, spoilage organisms struggle; in a flat can, conditions shift and survivors get their chance. At FizzWorks Beverages, the under-carbonated cans weren’t just a quality issue — soft cans, flat taste, complaints. They were the visible symptom of a filler problem also compromising microbial stability, and carbonation had been drifting low for two months before anyone connected the soft cans to the filler.
Background: a craft soda cannery with a carbonation spec
FizzWorks Beverages (fictional) canned about 50,000 cases of craft soda a month — 40 employees, flavored sodas and seltzers, running a single canning line with an inline carbonator. The carbonation specification was 3.5 volumes CO₂ (the industry standard for sodas), verified by a Zahm-Nagel tester on the filler — one can per hour, manually. The cans were standard 12-oz aluminum, lightweight, dependent on internal pressure for stacking strength in the warehouse.
Challenge: the soft cans
Warehouse staff noticed it first: pallets of cans that felt soft — the cases compressing under stacking weight, cans denting during normal handling. Customer complaints followed: flat-tasting soda, cans that “felt wrong.” The QA investigation tested carbonation on retained cans: 2.0–2.5 volumes, against the 3.5 specification — nearly a third low. Two months of production was affected, and the warehouse held 30,000 cases of soft cans that couldn’t be stacked, shipped, or sold normally.
The microbial dimension emerged from stability testing. The sodas were acidic (pH 3.2–3.5), formulated for the specified carbonation, with CO₂ as part of the antimicrobial hurdle. At 2.0 volumes, the hurdle weakened: higher headspace oxygen, lower carbonic acid — and the stability samples started showing yeast growth at 60 days in the under-carbonated lots. The properly carbonated product was clean. Carbonation wasn’t just texture and structure; it was part of the preservation system.
Investigation: the carbonator that drifted
The inline carbonator — the unit dissolving CO₂ into the beverage before the filler — had a failing CO₂ regulator. The regulator’s diaphragm was degrading, delivering progressively less CO₂ to the carbonator over two months. The carbonator’s own display showed normal operation — it was dissolving everything it received, faithfully — but it was receiving less and less. The failure was upstream of the monitoring, invisible to the system’s own instrumentation.
The hourly Zahm-Nagel checks should have caught it — the logbook showed six weeks of decline, from 3.5 to 2.5 volumes. But readings were logged, not trended: each check was judged in isolation against the specification rather than plotted as the decline it was. The action limit sat at 3.0 — below spec but above the stop threshold — so the readings lingered in the marginal zone for weeks, each slightly worse, none triggering action. The procedure required checking, not analyzing.
The warehouse stacking compounded the commercial damage. The lightweight cans, designed for 3.5 volumes of internal pressure, lost their rigidity at 2.0–2.5 volumes — the cases couldn’t be double-stacked, the pallets couldn’t be warehoused normally, and the 30,000 affected cases occupied triple the warehouse space while the company figured out what to do with them. The carbonation specification wasn’t just about taste; it was structural engineering, and the structure had failed.
Root cause: the untrended drift
1. CO₂ regulator degrading undetected. The upstream regulator failure reduced CO₂ supply gradually, invisible to the carbonator’s own monitoring. The system’s instrumentation covered the carbonator, not its inputs.
2. Manual checks logged, not trended. Six weeks of declining Zahm-Nagel readings sat in the logbook as isolated data points, each marginally acceptable, collectively a clear drift. The procedure required measurement without analysis, and the drift continued until the cans went soft.
3. Carbonation’s structural and microbial roles unrecognized. The organization treated carbonation as a taste parameter — the sensory attribute — while it was simultaneously the can’s structural system and part of the preservation hurdle. The drift’s consequences extended far beyond flavor, into warehouse logistics and microbial stability.
Corrective actions: trending the gas
Immediate: the CO₂ regulator was replaced (and the replacement put on a preventive maintenance schedule — regulators now replaced annually regardless of condition), the carbonator was serviced and verified, and carbonation was confirmed back at 3.5 volumes before production resumed. The 30,000 cases of soft cans were evaluated: the under-carbonated lots with yeast growth were destroyed; the lots without microbial issues but with structural problems were sold at discount for immediate consumption — a commercial recovery that the food safety team approved only after stability testing cleared them.
Within 30 days, the readings went from logged to trended: every result plotted on a control chart, with trend rules (six consecutive declining points triggers investigation, regardless of spec compliance). The action limit was tightened — 3.2 triggers investigation, 3.0 stops the line — because marginal compliance is failure in slow motion. An inline CO₂ monitor went on the carbonator feed, with alarms on supply pressure.
The preservation review was the deeper fix: the formulation’s microbial stability was revalidated at the lower carbonation bound, establishing the actual minimum CO₂ for the preservation hurdle (2.8 volumes — below which the product is microbiologically compromised regardless of taste). The carbonation specification now carries a food safety justification, not just a sensory one, and the warehouse stacking procedures were updated with minimum carbonation requirements for double-stacking. Within 90 days, the plant extended the SPC trending to all its critical filler parameters — fill height, seamer operation, pasteurization — on the principle that drift is the universal failure mode.
Results: the carbonation that’s controlled
Twelve months later: carbonation consistently at 3.4–3.6 volumes, zero soft-can incidents, zero stability failures, and the control chart catching two minor CO₂ supply variations before they became drifts — both investigated, both resolved, neither reaching the product. The inline CO₂ monitor paid for itself in the first quarter by catching a supplier’s under-filled CO₂ bulk tank before it affected production.
Lessons learned: what you’d do Monday morning
Pull your carbonation logs — or your equivalent critical-parameter records — and plot them as a trend, not a list. Look for drift: six points declining, even within specification, is a process moving toward failure. If your procedure requires checking without trending, you’re collecting data and discarding information. Add the control chart; it’s the cheapest upgrade in process control.
Then examine your carbonation specification for its full justification: taste, structure, preservation. If the spec is written as a sensory target, rewrite it as the engineering parameter it is — with the minimum for can integrity and the minimum for microbial stability defined separately. And monitor your inputs, not just your outputs: the carbonator was fine; its CO₂ supply wasn’t.
Yeah, but actually — the logbook with six weeks of declining readings is the artifact of this case: the data was there, the drift was visible, and the system saw each point but never the line. Measurement without trending is expensive record-keeping. Plot the points, apply trend rules, and act on the drift before it becomes the failure. The control chart would have caught it at “gradually.” And carbonation isn’t just fizz — it’s the can’s skeleton and part of its immune system. Treat the gas like the engineering pa