A glass bottle annealing test confirms that the lehr relaxed the stress created in forming, and it rests on three pieces of evidence: polariscope screening on the line, a quantitative residual stress reading on bottles drawn from the run, and the lehr temperature curve that batch actually travelled through. A reading only means something when the report also names the spot on the bottle, the glass colour and the scale used. The finding that predicts field breakage is a high value concentrated at the base corner, not the absence of stress, because every glass container carries some.
This page deals with annealing and residual stress only. Burst strength under carbonation depends on wall thickness, base geometry and forming, and belongs with internal pressure testing. Whether a neck, body diameter or height matches the drawing is a question for dimensional tolerance standards. Impact and packing are handled on the drop test page.
What to specify in an annealing clause
A clause is enforceable when two readers reach the same verdict from it, and when the evidence it calls for is something the plant already produces for its own process control. Ask for a document nobody keeps and you will receive one written for the occasion.
The clause needs these parts:
- The container, identified by reference with its capacity, glass colour and mould.
- The measurement: instrument class, viewing or measurement location, and the scale the limit is written in.
- The limit as a maximum value, with a separate figure wherever one is justified. A base corner and a sidewall seldom deserve the same number.
- The sampling ratio and acceptance number.
- Disposition of failures. If the clause is silent on what happens to a failed sample, the party with money at stake decides.
- Retention: reference bottles from each batch kept for a stated period, so a complaint that arrives weeks afterwards can be examined on the real goods.
- A change rule obliging the plant to give advance notice of any change in glass composition, mould, lehr setting or container weight. Each of these can shift the stress pattern while the drawing stays the same.
- A shipment reference, so the record that accompanies the goods carries a batch number you can match to your incoming inspection.
Applying AQL to a stress reading
If you buy to a general sampling convention, say in the clause that it applies to stress. Do not assume an inspection level agreed for dimensions will be carried over. AQL is common in container supply for visible defects and dimensional characteristics, and the same discipline works for stress provided the characteristic is measurable and the sample comes from the same population.
Clauses rarely fail on the AQL level chosen. They fail by attaching an attribute-style acceptance number to a variable measurement, which gives a result no one can defend. A workable arrangement has three parts: a fixed count of containers measured per interval, a maximum value for each measurement location, and a rule that a bottle over the limit triggers a further sample from that interval instead of rejecting the batch outright.
When you must also show conformity to an outside requirement, file the stress records and the compliance documents together, not in separate departments. The material declarations that go with each container are listed alongside the bottle and jar collections.
Terms the clause must use precisely
Loose vocabulary is the usual reason an annealing clause cannot be enforced. These are the terms to recognise in a supplier document.
- Annealing point. The temperature where glass still flows enough to shed strain quickly. The lehr's hold section is normally set somewhat above it, so that relief finishes within the time the belt allows.
- Strain point. A lower temperature beneath which the glass is rigid for practical purposes and stress can no longer relax. The span between the two points is the lehr's working window, and the cooling section's job is to cross it without creating fresh strain.
- Residual stress level. Annealed quality stated as a stress value or as an optical retardation measured through the wall.
- Annealing grade. The older container-industry convention of numbered grades, where a lower number means less stress.
- Thermal shock resistance. A container's ability to survive a sudden temperature change such as hot filling or pasteurisation. It is a separate property with its own method, set out in standards such as ISO 7459.
Stress level and grade are different scales and cannot be swapped. A specification that promises "annealed to grade X" without naming the measurement behind the grade specifies nothing. Name the instrument, the location, the sample condition and the numeric limit.
Good annealing usually improves thermal shock behaviour, since less locked-in stress leaves more margin. One good result still does not prove the other. Seams, folds and checks form another family altogether: they are forming defects, inspected by eye and not optically.
Where residual stress comes from
A bottle leaves the forming machine soft and then cools. The outside of the wall contracts and sets first, while the core is still hot and mobile. When the core cools in turn, the rigid skin resists its contraction, and the two zones stay loaded against each other for good. That permanent load is residual stress.
Three things shape the pattern, and bottle size alone is not one of them:
- The temperature difference across the wall at each moment of cooling. Strain is generated by the difference, not by the absolute temperature.
- The wall itself. Uneven thickness, or a base much heavier than the body, means sections cool at different speeds and a steep gradient forms between them.
- Time spent in the band where glass can still flow. Below that band the structure is frozen, and whatever strain is present stays for the bottle's life.
Even, moderate compression in the surface is not a defect and should not be reported as one. Glass breaks in tension: a crack advances only where the material ahead is being pulled apart, so a compressed skin has to be overcome before a flaw opens. Harm comes from the reverse arrangement, with tension locked into the surface or wall, and from sharp gradients where thick meets thin. A bottle can hit its drawing weight, pass every dimensional check and still fail early because the pattern is inverted in one place.
Methods for measuring residual stress
All routine methods rely on the same effect. Stressed glass rotates the plane of polarised light, by an amount proportional to the stress and to the path length, and by different amounts in two perpendicular directions. Between crossed polarising filters that difference appears as coloured fringes, and their position, spacing and colour describe the stress pattern. ASTM C148 covers the polariscopic examination of glass containers.
| Method | How it is done | What it gives you | Limits |
|---|---|---|---|
| Hand-held crossed-filter viewer | An inspector looks through base and sidewall against a dark background with a light source | Immediate sorting of heavily strained bottles, which is why it runs continuously as a line screen | The judgement is comparative and varies with observer and light source, so it yields no number for a specification dispute |
| Bench polariscope, polarimeter or strain gauge | Sodium or similar monochromatic source, defined viewing distance, rotating analyser | A fringe count or optical retardation at a defined point; some instruments convert this to a wall stress figure | Valid only for the stated location. One point tells you nothing about the remainder of the bottle |
| Destructive sectioning or scattered-light polariscope on a cut sample | The bottle is cut and the section measured optically in a laboratory | Detailed evidence for development, dispute resolution and validating a production screening limit | Slow and it consumes the sample, so it is not a release test |
Location is the weak point of every quantitative reading. A value taken through the base corner of a heavy punt cannot be set beside one taken through the sidewall. Reserve sectioning for cases that justify the effort, such as a first article from a new mould or a failure that keeps coming back.
Reading a stress or retardation figure
Three statements must accompany any number on an annealing report. Without them it cannot be compared with anything, not even the same supplier's earlier results.
Location on the bottle
Stress gathers in predictable places, and an inspector should check them by default: the base corner where wall meets bottom, the heel just above, the mould seam, the shoulder where body narrows to neck, and the zone around the neck ring and finish. A convenient sidewall point can read low while the base corner reads high, and the base corner is where most field failures start. Quote one limit with no location and the easiest spot on the sample will satisfy it.
Glass colour and sample conditioning
The reading is optical and is taken through a curved, possibly tinted wall. Amber and green absorb part of the spectrum, so they interpret differently from flint. A thick base also reads differently from a thin one at equal stress. If a report applies a single limit to several colours, ask whether it was calibrated for each colour or simply copied across.
The scale
A stress value, a fringe count and an annealing grade are three scales. Low is good on each, yet "five" is meaningless until you know which one it sits on. Ask for a table of measured values showing sample size, location, colour, instrument and scale. A supplier who can hand that over is controlling the process. One who offers only a sentence that the bottles are properly annealed is describing an intention.
There is no universal good number. A limit has to be established for the named bottle, glass composition and lehr; the target is a low reading spread evenly through the wall.
How the lehr curve produces the stress pattern
The lehr is a gas- or electrically heated tunnel. Bottles ride through it on a belt or mesh while the temperature is held and then lowered on a schedule. Forming creates stress and the lehr removes it before the glass drops out of the range where it can flow, which makes the schedule the biggest single lever on annealed quality. Asking a plant for its curve tells you more than asking for its certificate.
The tunnel does three jobs in sequence, and each leaves a recognisable mark.
Reheat and hold
Bottles arrive hot but already cooling. The hold section returns the entire wall to a temperature at which strain relaxes and keeps it there until relaxation is largely complete through the full thickness, not only at the skin. Too short a hold, or a belt running too fast for its load, produces bottles that were warmed but never relaxed. They carry moderate stress spread evenly through an otherwise well-formed wall. In high-output work this is the most frequent shortfall, and visual inspection cannot see it.
Controlled cooling
From the hold, the glass descends through the annealing range at a pace it can follow without new strain building across the wall. The pace is usually written as a maximum temperature drop per minute, and it is the most important figure on the curve because it has to suit the wall weight being run. Thick sections cool slowly and cannot be rushed, so a heavy-based bottle needs a gentler curve than a light one. Run both on one schedule and the heavy one tends to suffer: cooling too quickly for the thickest section leaves a gradient between base and sidewall, and the crack eventually starts there, often well after packing.
Final cooling to handling temperature
Here the concern moves downstream. Bottles packed warm go on shrinking inside a pack already closed around them, and cartons, dividers and shrink film behave differently on hot glass. Bottles that reach a filling line still hot raise a further problem. If you pack for immediate use, agree a maximum temperature at which bottles are presented for shipment; leaving the tunnel is not proof the lehr has finished.
Evidence to check and what an abnormal result looks like
The table turns an annealing clause into the evidence that supports it. Every entry is a starting point for discussion, because the applicable limit depends on the container, the glass composition, the lehr and the process downstream.
| Evidence | Accepted practice | Warning sign | How it shows up as breakage | Confirm with the plant |
|---|---|---|---|---|
| Lehr curve for the production run | Hold temperature and cooling rate recorded against time, with the maximum rate stated and shown to suit the heaviest section | A flat or retyped curve, the same curve for containers of plainly different wall weight, or a hold set below the annealing point of the composition | A marginal curve trims impact and thermal margins across the run, so you see a raised overall breakage rate and no single defect | Is the curve logged per run or per machine, is it retained, and can it be retrieved months later |
| Polariscope screening on the line | Crossed-filter viewing of every container or at a set frequency, with a documented viewing position on base and sidewall and a written diversion limit | A screening step with no limit, position or record, so the outcome depends on the shift | High-stress bottles enter stock and later fail at the base corner in the customer's hands | Viewer location, who does the check, what unit the diversion limit is in, and where diverted bottles go |
| Quantitative stress reading | Bench measurement with controlled light, defined point and stated scale, on bottles pulled from production and not from the sample room | Figures with no location, scale, colour or sample size; one limit reused across colours and wall weights | A high base-corner value forecasts cracks in filling, in transport and in the first weeks on shelf | Default measurement points, instrument used, and whether the limit was set for this container or inherited |
| Sampling ratio and frequency | A stated count per interval taken across the run, with plan and acceptance number in writing | A plan described as "regular" or "random"; all samples pulled at start-up, when the lehr is most stable | Drift over a long run goes unseen, and a bottle that passed first article repeats a defect nobody measured | Interval, count per sample, cavity and machine sampled, and whether the plan changes with belt load |
| Batch record and traceability | One document per batch tying together date, machine, mould, colour, curve reference and stress results, retrievable by batch number | Curves and readings in separate files without a batch number | A repeat complaint gets a repeat sample, not a root cause, and the failure returns with the next order | Can the record ship with the goods, in what format, and does it survive a shift or staff change |
| Thermal shock cross-check | Periodic hot and cold water bath test to a stated temperature difference for hot-filled or pasteurised containers, following methods such as ISO 7459 | Failure at a temperature difference the bottle should meet in normal service, which points to stress and not to the bath | Cracks on the filling line: each one stops the line and loses the contents with the bottle | The real temperature difference on your filling line, and whether the check uses it or an assumed value |
| Handling and storage after the lehr | Cooling to a stated handling temperature before packing, with check bottles retained so a fault is traceable after the pallet is wrapped | Warm bottles packed straight away with nothing retained | Carton distortion and handling damage that later get blamed on annealing, obscuring the true cause | Exit temperature from the lehr, where retained bottles are stored and for how long |
Breakage patterns that point back to the lehr
Annealing faults show themselves in a few recognisable forms, and each implicates a different part of the curve. Filing them all under "glass quality" delays the correction.
Vertical cracks from the base
Short cracks that start at the base and climb the heel with no impact mark indicate a steep gradient between a heavy base and a lighter body. They tend to surface during filling or in the days that follow, and to cluster on one mould instead of scattering across the run. A burst that originates at the base corner and shows a smooth fracture face is the same fault further advanced. The remedy is a slower cooling rate for the heavy section, or a slower curve for that bottle if it shares belt speed with lighter ware.
Cracks at the shoulder or neck
Without impact evidence, these implicate the hold, frequently one too brief for the wall thickness. Because the first symptom is a chip or crack close to the neck, the closure or finish often gets blamed first. A related sign is a bottle that fills without trouble but cracks when hot liquid or hot water touches it: a thermal shock failure that began from stress it should never have carried.
Slightly raised breakage across the whole run
The third pattern has no crack type of its own. Breakage sits a little above what the pack and line can absorb, evenly across moulds and across the run. This is an entire run annealed at a marginal setting. It is the hardest to catch, since every individual bottle passes inspection, and the easiest to wave away as normal breakage. Comparing stress readings from the suspect batch with an earlier batch that performed well is the only dependable way to tell it from handling damage.
When bottles break at filling instead of in transit and nothing suggests impact, request the annealing record before any other document.
What to request before and after the production run
Make the request before ordering, because part of this evidence cannot be recreated. A curve requested afterwards usually arrives, but it is seldom the one that ran.
Before the run:
- The lehr schedule planned for your specific container, with hold temperature and stated maximum cooling rate.
- Confirmation of the measurement points on the bottle and the reporting scale.
- How the schedule is set for a container of different weight. The reply shows whether the plant adjusts per bottle or runs a house setting.
None of this involves commercial disclosure, and a plant in control of its process can answer all of it.
After the run:
- The record itself, not a pass statement: batch, date, machine and mould, the curve reference, and stress results with sample size and measurement location.
- A retained sample from the batch, held by both sides where possible. A later dispute then becomes a re-measurement of the same goods and not an exchange of opinions.
- For first article approval, bottles from normal production and not a pilot. A lehr carrying a single container at low load behaves differently from the same lehr with a full belt.
When we draft the clause and evidence list for a project, we need the bottle type, the intended use, and whether the product is hot filled or held under pressure. Those answers decide whether the limit, the thermal shock check and the sampling plan sit at the conservative or the standard end, and whether one acceptance point on the base is enough or several locations must be measured and reported separately. The most helpful input is the drawing or reference of a container that has performed well, together with the failure that prompted the enquiry; the answer usually lies in the gap between them.
Frequently asked questions
Does an annealing test measure how strong the bottle is?
Not directly. It measures the locked-in stress that subtracts from the strength the container would otherwise have. The checks involved are crossed-filter screening, a quantitative reading on a drawn sample and a review of the lehr curve for the batch.
What counts as a good residual stress reading?
No single figure applies to all bottles, and a supplier who quotes one without asking about the container, colour and measurement location is simplifying. A sound limit is tied to a named scale, a named point and a named colour, and was set for that bottle instead of borrowed from another.
What should a lehr temperature curve show?
A hold hot enough, for the time available, to relax strain through the full wall; a controlled descent the heaviest section can follow; and final cooling to a stated handling temperature. Check that the maximum cooling rate was chosen for your bottle's wall weight, since a schedule built for a light container will usually damage a heavy-based one.
Which standards apply to annealing checks?
ASTM C148 describes polariscopic examination of glass containers. ISO 7459 covers thermal shock, which serves as a cross-check on bottles destined for hot filling or pasteurisation but does not replace a stress reading.
Will better packing solve an annealing problem?
No. Packing limits the impact a bottle receives, whereas annealing sets how much stress it carries before any impact. A highly stressed bottle can fail in filling, in storage or under a load that a well-annealed one would shrug off. Extra cushioning hides the symptom at the carton and leaves the filling line and the customer exposed, so verify the two separately: packing by impact testing, annealing by stress measurement.