Pressure Decay Leak Testing — When to Use It

Put the package in, pressurize, wait, watch the number fall. That's the whole test — and the entire art is in what you do with the falling number. Here's the physics, how a leak size becomes a pass-fail line, and where the method sits in a packaging program.

The test in thirty seconds

Pressurize the package's interior or headspace with a dry gas — usually nitrogen. Isolate it. Watch the pressure over a fixed interval.

A package with an intact seal holds nearly flat. A package with a leak bleeds pressure faster than the reference allows. The difference — how much the pressure dropped, compared against what a known-good package does over the same interval — is the result. Pass or fail, on the package in front of you, in seconds.

That's the test. What's left in this guide is everything that makes the test a program instead of a number.

What's actually driving the decay

Here's the part the spec sheet doesn't say: the pressure isn't falling because of the leak alone.

The package is a volume that flexes. As it pressurizes, the material stretches, and as it sits, it relaxes back. Temperature moves the gas. In a polymer package, a little gas permeates the wall over the test interval. All of that happens on a perfect package, and it shows up as a small, predictable decay.

So the test isn't "does the pressure drop?" Every package's pressure drops. The test is "does it drop more than a known-good package does, over the same interval, under the same conditions?" The reference — a set of known-good packages run the same way — defines the normal band. A leak pushes the reading outside it.

That's why the method is calibrated on your own package, in your own fixture, at your own line conditions. The normal band is a property of the package and the setup, not of the instrument.

How a leak size becomes a pass-fail line

The acceptance criterion is set by deliberate failure. Take good packages, make holes of a known size — 50 µm, 100 µm, 200 µm, whatever the program requires — and run them through. The hole has to fail, every time, with margin. Then set the pass-fail line between the known-good band and the smallest failing hole, with the margin your risk assessment requires.

That's where the instrument's capability sets the floor. The MedIntegrity MFT-1000, for example, is specified for leak-rate detection as low as 2 microns, depending on package specification — the package and the fixture decide what the instrument can actually see. The leak limit is a property of your package plus your risk assessment, and the instrument has to be able to see it.

The margin matters more than the limit. A limit set with no margin passes on the day the line is tuned and fails on the day it isn't. A limit with margin is a limit that survives the shift change.

Why it's a deterministic, every-unit method

In the framework of USP <1207>, pressure decay is a deterministic method. It gives a direct, measurable result on every single unit. Run the package in front of you, get the number, make the call on that package. The unit you test is the unit that ships — no sampling, no inference, no statistics doing the work the test should do.

Two properties make it the in-process default.

It's non-destructive. The package survives the test and goes into the carton. For a fill-finish line, that's the difference between testing the product and testing a throwaway surrogate. And it's fast — the test interval is seconds, which is the only speed a line running at production rate will accept.

No dye, no consumables, no sample prep. The method runs the same way on the first shift and the last, and the only thing that changes is the package.

Where it sits against the other methods

Pressure decay is one of several pressure-based methods, and the differences are about what the package can do, not about which is better.

Pressure decay — the package is pressurized from the inside or its headspace, and the decay is watched. Right for hermetic and semi-hermetic packages you can pressurize: vials, pre-filled syringes, rigid containers, many blisters and pouches with a defined volume.

Vacuum decay — the environment around the package is pulled to vacuum and the pressure change is measured. Right where the package can't take an internal pressure, or where the geometry works better from the outside. It's the method the ASTM F2338 framework describes for leak detection, and it catches what pressure decay can't see on certain formats.

Dye ingress — the package sits in dye and a leak lets it in. Brutally sensitive, and destructive: the tested unit is a used unit. It earns its place in validation and failure investigation, not in the production line.

The method choice is a function of the package. The instrument has to cover the methods your package set needs, and the MFT series — MFT-600 through MFT-1000 — is built around the pressure-based methods for pharma packaging: blister, pouch, sachet, and rigid containers, with traceable electronic records.

Where it runs on the line

Pressure decay runs in-process, at pack-out, on every unit. That's the primary job — the per-unit call that keeps the failing package out of the carton.

It also runs at release, on samples from the finished lot, as the confirmation that the in-process test is still doing its job. And it runs in validation, against the deliberately failed packages, to prove the method sees the leak it's supposed to see. Three runs, three purposes, one method — and the records from all three are the program.

The MFT-1000 is specified for flexibility across a wide range of products and packaging formats, which matters on a line that runs more than one package: the fixture changes, the reference band is re-established per format, and the method is the same. GMP qualification — the IQ/OQ/PQ package — is what moves the instrument from the receiving dock into the validated environment.

Setting acceptance criteria you can defend

The criteria come from your package, your failure modes, and your risk assessment — in that order, and not from a sister product's procedure.

What does this package actually do when it fails? Does the seal part at the crimp, does the film delaminate, does the port seal blow? Each mode has a size, and the acceptance criterion has to see the smallest one your product can't afford to ship. Work backward from that hole, set the line, validate against it, document it.

And the validation doesn't retire. A new package supplier, a different film, a changed crimp, a line speed that moved the fixture loading — any of them moves where the failure modes live, and the method gets re-run against the new reality. The criteria are a living document, and the record shows when they were last proven.

Records

A per-unit test on a production line generates a lot of data, and a lot of data has to be real data. The MFT series carries electronic, traceable records — mapped to 21 CFR Part 11 for the GMP environment — so every pass and every fail is a record with a timestamp, a package ID, and a method behind it. When the deviation comes, the record is the evidence, and it's already written.