An internal pressure test fills a glass bottle with water, raises the pressure inside to a nominal figure and holds it for a set time to see whether the bottle leaks or breaks. For a carbonated or fermented drink the result only means something when that figure was derived from your product: its carbonation in volumes of CO2 and the warmest temperature it will reach after filling, not the temperature on the filling line. Routine acceptance uses the hydraulic method described in ISO 7458, and a burst test on further samples shows how much margin sits above the pass mark.
What the drink demands of the bottle
The contents set the requirement, so start there. Four properties of the product decide how hard the headspace pushes on the glass.
- Dissolved CO2. Carbonation is stated in volumes of CO2, meaning the volume of gas at standard conditions dissolved in one volume of liquid. More dissolved gas gives a higher equilibrium pressure.
- Temperature. Warm liquid holds far less carbon dioxide than cold liquid, so a sealed bottle gains pressure quickly as it warms. A pressure figure quoted without a temperature is only half a specification.
- Process. Tunnel pasteurisation, hot filling or a warm filling hall raise temperature and pressure together, and the bottle has to take both at once.
- The gas. Nitrogen dosed for pack hardness or oxygen control behaves differently in solution. A nitrogenated beer or coffee can reach pressures its CO2 content would never suggest, so measure it instead of converting.
Distribution matters as much as formulation. A retail product may sit in a hot warehouse, cross an ocean in an unrefrigerated container and stand under several stacked pallets, so the case to design for is the hottest day of the journey. A drink sold locally or on draught never meets that extreme and can fairly use lighter glass. A returnable bottle is a third case: washing, relabelling and repeated handling wear away some of its margin on every trip.
Converting CO2 volumes into a pressure
Brewers and beverage developers talk in volumes of CO2; test laboratories talk in bar or megapascals. Headspace pressure depends on the dissolved quantity, the temperature and the composition of the liquid, with temperature having the strongest effect. One beer at one carbonation level presses noticeably harder in a warm store than in a cold cellar.
A calculation made at filling temperature therefore flatters the bottle. Run it at the highest temperature you expect the product to see, and write that temperature next to the carbonation figure. For a new product, measure the pressure on the real liquid at the real storage temperature; a measured value is much more dependable than a formula.
As a rough ladder for early discussion only: still water and still wine have no meaningful carbonation; lightly sparkling fermented drinks come next; beer occupies the low to middle range depending on style; bottle conditioned and heavily carbonated styles sit above that; carbonated soft drinks are higher again, since they are formulated with a lot of gas for cold serving; and sparkling wine tops the normal range. That is why sparkling wine bottles are the heaviest, most heavily based containers in everyday production.
Service pressure, test pressure and burst pressure
Most disagreements between a buyer and a glass plant come from mixing up three figures.
| Figure | What it is | How it is used |
|---|---|---|
| Service pressure | What the bottle really experiences: headspace pressure at the product's maximum temperature, plus the short peaks during filling and closing | The starting point for everything else; record it in the specification |
| Test pressure (nominal pressure) | The level a sample must hold in the laboratory with no leak and no fracture, set well above service pressure | The pass or fail criterion for deliveries |
| Burst pressure | The level at which the bottle actually breaks, normally well above the test pressure | Shows the remaining margin and lets you compare candidate bottles |
The gap between these figures is the safety margin. It is there to absorb what nobody controls fully: a bottle that got warm, a filling valve that overfilled, a dropped pallet, a scuff at the heel, and the normal spread of strength within a production run.
Write the test pressure and method into the specification, and note separately the service pressure it came from. If you later reformulate, lengthen shelf life or enter a hotter market, the service pressure moves and the test pressure may have to follow; a document that lists only the test pressure gives you nothing to review against. Recording the burst pressure of delivered goods is also worthwhile, because a batch whose burst values are drifting down is weakening before any bottle fails in the field.
A healthy ratio does not insure against a filling line that is out of control. Field failures are driven by local conditions on individual bottles, not by average strength: a chipped finish, a wrongly set valve that over-pressurises the fill, product filled warm and chilled afterwards, or a closure pressed on so hard that it loads the neck. A laboratory result on sample bottles shows none of these.
Hydraulic, pneumatic and burst test methods
A supplier report should make clear which of three arrangements was used, because their numbers cannot be swapped for one another.
Hydraulic test
The bottle is filled with water, brought up to the specified pressure and held there while it is watched for leaks or breakage. Water stores very little energy, so a failure stays fairly contained and high pressures can be applied safely. Most routine acceptance testing works this way, and ISO 7458 describes the method for the internal pressure resistance of glass containers. The outcome is usually a pass or fail at a stated nominal pressure. Its value depends on the samples coming from a genuine production run, not from hand-picked bottles.
Pneumatic test
Air or another gas replaces the water. This suits cases where gas is the realistic medium, where the bottle must remain dry, or where gas permeation through a closure liner is part of what is being studied. Compressed gas holds a lot of energy, so a burst throws fragments; the rig needs guarding and the procedure must keep people out of harm's way. You will find it in laboratories and special investigations more often than on routine production checks. If you need it, say so in the specification instead of letting the test house choose the medium.
Burst test
Here there is no fixed pass level. Pressure keeps rising until the bottle breaks, and the report gives the pressure at failure and where the fracture started. Use it to establish margin over the design pressure, to rank candidate bottles, or to investigate a complaint where you need to know how near the limit the bottle was running. One burst value is not a specification, since it varies with the sample, the rate of pressurisation and the origin of the break. A series of results is what carries meaning.
Pressure regime and bottle choice by product category
The table sets each product type against the pressure it generates, what that asks of the wall and base, the fitting verification, and the question to put to the plant before choosing a bottle. The pressure descriptions are indicative and for planning only. The binding number is the one derived from your product at its maximum storage temperature, and the test pressure is whatever the applicable standard or your own specification names.
| Product category | Typical carbonation | Pressure to design for | Wall and base demands | Verification | What to confirm with the plant |
|---|---|---|---|---|---|
| Sparkling wine, sparkling cider | High: around five to six volumes of CO2, sometimes more in bottle conditioned styles | The highest of the ordinary categories; several bar at cellar temperature, climbing as it warms | Heavy wall with controlled distribution, a pronounced dome or punt, generous heel, robust finish | Hydraulic test at the standard's nominal pressure, with burst tests on a sample to establish margin | The test pressure the bottle is rated to, the burst pressure achieved, and the product and storage temperatures assumed |
| Carbonated soft drinks, soda | High: commonly about three and a half to four and a half volumes of CO2 | High at ambient storage; these drinks are made for cold serving yet stored and shipped warm | Uniform wall, concave base with a defined standing ring, heel thickness sized for warm storage | Hydraulic test on production samples; burst tests when lightweighting is under consideration | Whether the rating covers warm storage or only filling, and how it shifts with a change of size |
| Craft beer, standard and bottle conditioned | Moderate and in the low range for standard styles; substantially more for bottle conditioned and strongly hopped styles | Moderate for standard styles; materially higher when secondary fermentation happens in the bottle | Straight wall with even distribution; base sized for the strongest style that will share the bottle | Test at the nominal figure, plus a check of pressure during refermentation, not just at filling | Whether the bottle was chosen for the strongest style it will ever hold, and that style's peak pressure |
| Kombucha and similar fermented drinks | Low and variable; often develops in the pack after filling | Low to moderate but unpredictable, which is the real difficulty | A standard wall often suffices; watch the base, and the finish where closures go on at high torque | Baseline pressure test, plus a shelf life study tracking pressure over time | Whether the rating suits a product whose carbonation is not fixed at filling, and how far the process can raise pressure |
| Still water, still wine, still spirits | None | No carbonation pressure; the test serves as a general strength screen | Standard wall; base and finish chosen for line handling and the closure | Nominal pressure test as a routine screen; impact and drop tests for the packing question | Whether a quoted result is a strength indicator or a carbonation rating, since those are different claims |
| Returnable carbonated bottles | Whatever the product needs; the return loop does not change it | As for the product, with an allowance for strength lost to repeated washing and handling | Heavier wall and base than a comparable one trip bottle, so strength lasts through multiple cycles | Burst tests on new bottles and periodic burst tests on bottles sampled from the return stream | The number of trips expected, and whether the rating applies to the first trip or the whole service life |
Where a bottle fails under pressure
Internal pressure stresses glass differently from an impact, which explains how two bottles of equal weight can perform unequally.
A cylindrical body copes well. Hoop stress in the wall rises with pressure and with radius and falls as the wall gets thicker, so a wide body needs more glass than a slim one for the same pressure. An uneven wall gives a bottle a weak side and a strong side, not an average. Pressure-critical specifications therefore address thickness distribution as well as the mean, and it is fair to ask a supplier how distribution is controlled on the forming machine instead of accepting a nominal wall figure.
The base is harder to get right. A flat bottom would bend like a plate under pressure, a poor way to load glass. Carbonated bottles almost always have a concave base on a standing ring, and heavier designs add a punt or dome that turns much of the load into compression. The dome profile, the radius where it joins the standing ring and the heel thickness govern burst pressure far more than the body wall. In a well made carbonated bottle, fractures originate at the base corner or heel, not mid-body. A report that states "pass" and omits the failure origin has thrown away its most useful data.
Two other zones need a look. The neck finish is locally thick, alters the load path at the top and takes the most handling knocks. The shoulder, where diameter changes quickly, concentrates stress and is where internal pressure meets any external load. Review a carbonation bottle as a shape, and when comparing two candidates, compare where each one breaks as well as the pressure at which it does.
Working from the drink to a bottle specification
The sequence has five steps, and a specification that skips one looks finished without being so.
- Set the carbonation target and the maximum product temperature, then calculate or measure the pressure at that temperature.
- Decide how the bottle is filled and closed. A counter pressure filler, a gravity fill with carbonation afterwards in the bottle, and bottle conditioning each give the glass a different pressure history, even where the final equilibrium is identical.
- Fix a design pressure: expected service pressure multiplied by a factor that allows for temperature excursions, filling imbalance and strength lost in handling.
- Choose a bottle whose wall and base meet the design pressure at an acceptable weight. The pressure test result is the tool for this step.
- Prove it on samples of the actual bottle at the stated pressure and temperature, and put the result in the purchase documents so later deliveries are judged against it.
Most of the cost is committed at the fourth step. Glass one wall grade too heavy adds cost to every unit and eats into the freight allowance. Glass one grade too light brings line stoppages, complaints and a breakage rate nobody can budget for. The target is the lightest bottle that holds the agreed design pressure with an agreed margin. When comparing offers, ask for test pressure and results instead of bottle weight: weight contributes to strength but does not measure it. We usually begin that comparison from the glass bottle collections, reviewing the container as a whole before settling the pressure detail.
How filling, crowning and pasteurisation change the load
A filling line can impose conditions no laboratory test reproduces. Specifying a bottle and then altering the process is the commonest route to failures the bottle was never designed to survive.
Filling. Product filled cold peaks later, in the warehouse; product filled warm is at high pressure straight away. A counter pressure filler takes the bottle through a controlled pressurise-and-release sequence that is gentler than gravity filling followed by in-pack carbonation. Compare the bottle's rating with the maximum reached during filling, not only with the equilibrium afterwards.
Closing. Too little crowning force and the pack vents or leaks. Too much and the finish is loaded, which can start a neck crack that opens only once the bottle is pressurised. Crown specification and capping head settings belong to the pressure system, though the closure interface is its own subject, covered in our guide to crown finishes, shells and liners. Logging capping torque or crown application force beside each pressure result settles a large share of pressure complaints.
Pasteurising. This is generally the toughest case. In a tunnel pasteuriser, peak pressure arrives when the bottle is hottest, the glass is slightly softer than at ambient and the closure is under thermal load. Assess the rating at pasteurisation temperature, and test samples with the same weight and height as production bottles, never a heavier laboratory stand-in. A bottle that passes at ambient and breaks in the pasteuriser was not badly made. It was specified against the wrong case.
Writing the pressure clause in a purchase specification
A clause that holds up through deliveries, a supplier change and a complaint has six parts. Each one removes a possible argument.
- Container and product. Identify the bottle by reference and name the drink, with its carbonation figure and maximum temperature.
- Method and pressure. Hydraulic or pneumatic, the nominal test pressure, and the standard it is taken from.
- Holding time and acceptance rule. Define failure. A leak at the finish and a fracture at the base are different findings, though both fail.
- Sampling plan. Samples from a production run, spread across machines and moulds where more than one is used.
- Record. Batch, date, test pressure and result; for burst tests, the burst value and failure location too.
- Change rule. Notice before any change to weight, mould, wall distribution or base design, as each can shift burst pressure while the drawing stays the same.
For a new product or a lightweighting project, add two things. Ask for a first article test on bottles from a normal production run and keep the retained sample, so a dispute can be resolved by re-measuring the same goods. Also state the assumed service pressure, so a change of recipe, pack size or market can be checked against the original design case. Both cost little to request with the order and a great deal to reconstruct later.
Keep the evidence together. A first shipment of a carbonated drink generally needs pressure results, material compliance documents and packing test results in a single file.
Problems that look like pressure failures but are not
Seal and pressure vessel are two specifications that fail differently and are fixed by different people. A bottle that withstands ten bar of hydraulic pressure and still vents at the crown has a closure fault. Finish dimensions, crown shells, liners, capping head settings and leak testing are outside the scope of a strength test.
Thermal history is the other look-alike. When bottles fracture at the base under pressure with no sign of impact, especially in a batch that had been running normally, check residual stress as well as strength. A well formed but poorly annealed bottle fails the test an identical, properly annealed one passes, and the remedy lies in the lehr, not the pressure rating. Our page on verifying annealing and residual stress explains that check.
Two neighbouring decisions also deserve separate treatment. Packing and loading are an impact question: a bottle that survives ten bar in the laboratory can still arrive broken. Bottle weight is a commercial choice as well as a technical one, because the extra margin of heavier glass is paid for on every unit and every pallet. The beer bottle range looks at the same container family from the filling side.
Frequently asked questions
What does an internal pressure test involve?
A sample bottle is filled, normally with water, and held at a defined pressure for a defined period; it passes if it neither leaks nor fractures. The nominal pressure comes from an applicable standard or from the customer's specification. The burst variant continues to failure and reports the pressure and fracture location, which measures margin instead of giving a pass or fail.
Is hydraulic or pneumatic testing the better choice?
Hydraulic, for routine acceptance, because a water-filled bottle fails with little released energy. Choose pneumatic only when gas is the realistic medium, the bottle has to stay dry, or liner behaviour with gas is under study, and expect guarded equipment and a written safety procedure. Name the medium in the specification whenever it affects the result.
How much pressure can a glass bottle withstand?
No single figure applies. It varies with the bottle, wall distribution, base geometry, glass composition and annealing quality. A sparkling wine bottle holds several bar with substantial margin; a light one trip bottle for a still drink is designed and tested on a different basis. Commercially, the useful number is the ratio of the pressure seen in service to the pressure the bottle has been shown to hold.
Which drinks need the test as a pressure requirement?
Any drink that is carbonated, referments in the pack or is pasteurised: sparkling wine and cider, carbonated soft drinks, bottle conditioned beer, kombucha and comparable ferments, and carbonated ready to drink products. For still water, still wine and spirits the same test is only a general strength screen, and its result should not be quoted later as a carbonation rating.
Why do carbonated bottles break under pressure?
Typical contributors are a wall or base distribution outside specification, a steep stress gradient left by annealing, handling damage at the finish, over-pressurisation during filling, and storage or pasteurisation hotter than the design case assumed. If a batch fails with no evidence of impact, request the annealing record first.
Can a bottle be rated for a drink that carbonates after filling?
Yes, provided the rating is based on peak pressure, which for a refermenting product can come long after the filling date. Run a shelf life study that tracks pressure over time across the expected storage temperatures, then select the bottle against the highest measured value plus a margin. Choosing on filling pressure alone is the most frequent error with these products.
What information is needed to plan pressure verification for a product?
Send the product category, its carbonation level or target pressure, the intended batch size and the market where it will be stored and shipped. Carbonation defines the pressure regime and the market defines the maximum temperature case. Batch size indicates whether routine tests on production samples are enough or a burst test series is worth adding.