Bubble leak testing · Immersion and dry
Bubble leak tester: immersion and dry vacuum testing for packaging
A bubble leak tester makes a leak visible. The package is held under water in a transparent chamber, the pressure around it is lowered, and gas escaping through a defect rises as a stream of bubbles from the exact point of the leak. DVACI manufactures the CDV range of acrylic vacuum chambers and control modules for this test, plus a kit for packages that contain little air. This page takes the operator's side of the chamber: what the equipment is, how an immersion test runs from filling to venting, the dry variant, how to qualify the test on your own packages, and where the method stops being reliable.
What a bubble leak tester is
A bubble leak tester, also called an immersion leak tester, has three parts: a transparent vessel for the water and the sample, a way to create a pressure difference between the inside of the package and the water around it, and a control that sets and times that difference. Two arrangements exist.
- Vacuum outside the package. The chamber pressure is lowered, and the gas inside the package expands and leaves through any defect (ASTM D3078, bubble emission). The package needs gas inside.
- Pressure inside the package. Air is injected into a submerged package through a needle (ASTM F2096, internal pressurization). It works when the package holds little gas.
In a DVACI system the acrylic (PMMA) chamber has 30 mm or 50 mm (1.2 or 2 in) walls, so the sample can be watched from every side. The control module beside it includes a Venturi vacuum generator that runs on compressed air; an electric pump is optional. The Internal Pressurization Kit adds the second arrangement.
Bubble tests are probabilistic in the sense of USP <1207.2>: the probability of detecting a given leak is below 1. That is a property to document, not a reason to avoid the test. For the equipment side of the standard, see the ASTM D3078 leak tester page; this page does not repeat it.
Bubble test for packaging: the immersion test step by step
Your SOP fixes vacuum level, hold time and acceptance criterion. The sequence stays the same.
- Close the drain valve before any water goes in.
- Fill from the sides. Run the water in along the inner walls, not over the sample position. A falling stream carries micro-bubbles into the water that can later be mistaken for a leak.
- Set the level. The sample must end up fully under water, with the hold-down grid above it.
- Submerge with no trapped air. Lower the package, tilt it and press gently so that air leaves the folds, the gussets and the space under the grid, and wipe off bubbles clinging to the film. Trapped air releases bubbles during pump-down and looks like a leak.
- Position it. Turn the seals toward the clear walls so each can be followed from outside; note which face is tested.
- Close the lid, press START and start watching. Do not wait for the setpoint. A gross leak can show during pump-down, and a large defect may release most of its gas in the first seconds. Module S draws down to the locked setpoint and runs the hold; Module P runs the stored recipe; with Module M the operator opens the valve and reads the gauge.
- Watch the whole hold. A leak gives a continuous stream from one point. A package that swells and opens a seal has failed under vacuum, a different finding from a leak.
- Locate the leak. The stream marks the leak position. Note where the bubbles leave the package, for example the left corner of the bottom seal, not where they break the surface.
- Release the vacuum. Modules S and P vent automatically at the end of the hold; with Module M the operator opens the valve. Open the lid only at atmospheric pressure.
- Record and handle the sample. Write down result and leak location. The package is wet; whether it returns to stock is for your QA team.
Reading the bubbles
- A leak: a stream that starts at one point and continues bubble after bubble through the hold, and may strengthen as the vacuum rises.
- Not a leak: a few bubbles from trapped air in the first seconds that then stop; bubbles clinging to the film or the grid; a fine haze of tiny bubbles on the walls and package, because dissolved air leaves the water as the pressure falls.
- In doubt: release, re-immerse and repeat; run a known-good package alongside; or check the suspect point by internal pressurization.
The dry variant of the bubble test
When a package cannot get wet, the same chamber runs without water. The package must hold air: with no gas inside, nothing expands and the test shows nothing. The operator watches the deformation. A tight package inflates as the vacuum rises and stays inflated through the hold; a leaking one inflates less, deflates during the hold or, if it holds liquid, leaves product on the chamber floor. Run a known-good package alongside for comparison.
Packages that pass are often left undamaged. The dry test locates a leak less exactly than bubbles do and is less sensitive than immersion, so it needs its own control defects and qualification.
Bubble emission leak test equipment: what you are specifying
| Part | What it does | DVACI option |
|---|---|---|
| Chamber | Holds water and sample; clear walls on every side | CDV2, 280 × 200 × 200 mm (11 × 8 × 8 in), to CDV6, 800 × 600 × 500 mm (31 × 24 × 20 in); acrylic walls 30 or 50 mm |
| Vacuum source | Creates the pressure difference | Venturi in every module: −750 mbar (−22.1 inHg), up to −800 mbar (−23.6 inHg) in favourable conditions. Optional pump: two-stage rotary to −950 mbar (−28.1 inHg), oil-free to −900 mbar (−26.6 inHg) |
| Control module | Sets level, times hold, vents | Module M, Module S or Module P |
| Internal Pressurization Kit | Injects air into a submerged package | DV-CONT-PI-01 |
All vacuum figures on this page are given at sea level. A vacuum gauge reads relative to the local atmosphere, so the achievable vacuum falls with altitude.
Two limits should not be confused. The Venturi or the pump sets how deep a vacuum can be generated. The chamber sets its own structural limit: the CDV2, CDV3, CDV4+ and CDV5 reach −950 mbar with a pump, while the CDV4 (30 mm walls) and the CDV6 are deliberately capped at −750 mbar, because customers run repeated test cycles and material fatigue rises sharply beyond that value. Choose the chamber by internal size, because the package inflates under vacuum and every seal must stay visible.
Qualifying a bubble test on your own packages
A bubble test is only as useful as your knowledge of what it finds on your own package. Qualification uses intact packages from the lot, packages with defects of known size, and a few intact packages sacrificed to find the vacuum at which they fail.
Control defects
DVACI supplies a tungsten wire of 50 µm (0.002 in) and calibrated pins of 150 µm (0.006 in) and 250 µm (0.010 in) to make defects of known size in a seal or the film. Prepare several samples per size, plus intact references from the same lot. The diameters describe the tool, not a measured leak rate, and should not be quoted as one. They also train and check operators and re-verify the procedure after a change of film, sealing setting or equipment, at the interval your quality plan defines.
Setting vacuum level and hold time
Raise the vacuum stepwise on intact packages until a seal opens or the package deforms excessively: that is the level your package will not tolerate. Set the test vacuum below it, with a safety margin that protects seals from the test itself and absorbs package-to-package variation. Then run the control-defect packages at that vacuum and note when bubbles first appear for each defect size, lengthening the hold until more time no longer helps. Liquid-filled packages need a longer hold, because a liquid moves through a defect more slowly than a gas, and a viscous product slows flow through a small channel.
If a defect size only appears above your safety margin, the method does not detect that defect on this package at a safe vacuum: write that limit into the procedure instead of raising the vacuum past the margin. The ASTM D3078 leak tester page gives the same five-step method with typical orders of magnitude.
Then lock the result: the setpoint is lockable in Module S and stored in a recipe in Module P, and the SOP states vacuum, hold, water, orientation and acceptance criterion. DVACI supplies IQ/OQ/PQ documentation for the CDV range (scope per configuration to confirm with DVACI).
What a bubble leak tester cannot do
- Prove a package is tight. The method is probabilistic (USP <1207.2>): a package that shows no bubbles was not seen leaking under the test conditions, which is not proof of tightness. Probabilistic does not mean statistical sampling; it means the probability of detecting a given leak is below 1.
- Find small leaks. It targets gross leaks; a narrow channel may not release a visible stream within the hold.
- See through the contents. Product can block a small opening, and the package material can press against it under vacuum.
- Work without gas. With little or no headspace nothing expands.
- Remove the observer. Results depend on attention, lighting and viewing angle, so the procedure fixes them and control defects check the operators.
- Measure a leak rate. The result is pass or fail, with a location.
Where your risk assessment needs smaller leaks detected or a number, a deterministic method such as vacuum decay (ASTM F2338) is used alongside; a CDV chamber is not a vacuum decay instrument. The bubble test keeps a clear role: it is quick, it shows where the leak is, and it suits seal set-up, start-up checks and complaint investigation.
Records: what each module keeps
| Module | Stored by the module | Written by the operator |
|---|---|---|
| Module M | Nothing | Vacuum, hold, result, leak location, sample ID |
| Module S | Nothing: no record is kept | The same, in a manual log |
| Module P | Named recipes and test records, two user levels; options: thermal or dot-matrix printer, barcode reader, CFR21 software | Visual result and leak location as your procedure requires |
21 CFR Part 11. With ORION OS and the CFR21 software option, which runs on a Windows PC alongside the instrument, the system meets 21 CFR Part 11 requirements: audit trail, individual user accounts through LDAP / Active Directory, PDF export with electronic signature, time-stamping and record integrity. The CFR21 option is required for this; without it, ORION OS does not provide nominative electronic signatures. A point-by-point breakdown of the requirements is available on request.
When internal pressurization shows the leak better
A vacuum bubble test needs gas inside the package to push through the defect. Vacuum-packed products, tightly filled pouches and liquid-filled sachets hold little, so a real leak can stay silent.
The Internal Pressurization Kit reverses the principle. A rubber patch (septum) is stuck to the package, a needle passes through it and air at regulated pressure is injected while the package is held under water; bubbles mark the leak, as in the test described in ASTM F2096. The kit has a manual regulator, a pressure sensor, START and STOP buttons, a 4 mm hose, needles and patches, and takes compressed air at up to 2 bar (29 psi). Where no air line is at hand, the DV-ACC-MC-01 mini compressor can feed it: it is oil-free, quiet enough for a laboratory, and delivers up to 4 bar (58 psi), so its regulator must be set within the module's 2 bar (29 psi) inlet limit.
| Vacuum outside | Internal pressurization | |
|---|---|---|
| Gas needed in the package | Yes | No, air is supplied |
| Package pierced | No | Yes, needle through a patch |
| Standard | ASTM D3078 | ASTM F2096 |
| Typical packages | Pouches, bags, trays with headspace | Vacuum packs, tightly filled packs |
Request a quote for a bubble leak tester
DVACI needs the largest filled package, its material, whether it holds air, whether it may get wet, tests per day and whether you need records. Module M suits occasional tests, development and incoming checks; Module S, routine QC with several operators; Module P, several products or formats and records. Services include installation, training, IQ/OQ/PQ documentation, traceable calibration, maintenance plans, spare parts and remote support; which of these are delivered locally depends on your country.
Request a quote for your bubble leak tester. Package type, dimensions and daily test volume are enough to start. First reply within four business hours; quote within three business days of receiving the information.
Frequently asked questions
Is a bubble leak tester the same as a vacuum decay tester? No. A bubble test is visual and probabilistic: the operator sees bubbles and where they start. Vacuum decay (ASTM F2338) measures a pressure change and is a deterministic method on a different instrument. DVACI CDV chambers run bubble emission, dye ingress, altitude simulation and, with the kit, internal pressurization, not vacuum decay.
How do I choose the vacuum level for a bubble test? Raise the vacuum stepwise on intact packages until a seal opens, then set the test vacuum below that level with a safety margin, and confirm it on packages with control defects. The level is your decision, validated on your own package, and locked in Module S or stored in a Module P recipe.
How long should the vacuum be held? Long enough for the smallest defect that matters to show. Note on control-defect packages when the first bubbles appear for each defect size, and lengthen the hold until more time no longer helps. Liquid-filled packages need a longer hold, because liquid moves through a defect more slowly than gas. The final time comes from your own qualification.
Can a bubble leak tester find very small leaks? Not reliably. The method targets gross leaks and is probabilistic per USP <1207.2>: the probability of detecting a given leak is below 1. Contents or film can block a small opening. Use control defects (50 µm wire, 150 and 250 µm pins) to learn what your procedure finds, and add a deterministic method where smaller defects matter.
Can I test a package without getting it wet? Yes, if it holds air. The dry test uses the same chamber without water: the operator watches the package inflate under vacuum and stay inflated. Passing packages are often left undamaged, but the test is less sensitive than immersion and does not pinpoint the leak, so it needs its own qualification.
Does the equipment record the results? It depends on the module. Modules M and S keep no record; the operator logs each test. Module P stores recipes and test records, with an optional printer and barcode reader; the CFR21 software option adds an audit trail, signed PDF export and LDAP / Active Directory user management.
Request a quote for your bubble leak tester
Package type, dimensions and daily test volume are enough to start. DVACI replies within four business hours and sends a quote within three business days of receiving the information.
Request a quote