Application reports

# 19 : How do Leak Detectors respond in test chambers containing steam

How do Leak Detectors respond in test chambers containing steam? (Part 2)

Question: In your last tip you mentioned that humidity is detrimental for the helium leak detector. But this does not have anything to do with our products! We are performing a water pressure burst test. After drying of the object we are applying a helium tracer gas leak test. Recently a customer has claimed a product as leaky, so for what reason are we missing leaks?

Answer: You still have a humidity problem. You have mentioned a water pressure burst test. During this test water penetrates the leak channels. Water molecules are dipoles. So many of these mini-magnets are sticking together and are forming a three-dimensional network. This network can block leak channels which are ten thousand times bigger than the diameter of a single water molecule. For that reason a leak channel might be blocked and you can miss leaks.

However, this does not mean that matter is not transported through these water-filled leak channels. Water will evaporate from the surface on the inner side of the sample. The water molecules then can change the frequency of quartzes, corrode electrical connections, lead to oxidation, etc.

Additional information: Tracer gas methods are among the most sensitive leak detection methods. However, to make full use of the performance of the method the tracer gas needs to find its way through the leak channels. Leak channels can be blocked by process residuals when the product was in use already. But also remaining water after a water pressure burst test can block the leak channels.

Many users of tracer gas methods claim, that only a burst test will open potential leakages by stressing the material. So it seems only logical to perform a leak test only after the burst test. And the leak test should be applied with a test pressure which corresponds to the maximum operational pressure.

Other users believe that water can block leak channels due to intermolecular forces between individual water molecules. So the reliability of tracer leak tests is questionable after a water pressure burst test and the sensitivity of the method cannot be used effectively.

It is almost 30 years since leak detection experts in Germany's chemical industry investigated whether a leak test can be applied after a water pressure burst test and which leaks still can be detected after the sample was exposed to massive humidity loads (1).

For the experiments a tubular heat exchanger was prepared with ten leaks in the order between <10-3 and < 10-2 mbar · l/s. The heat exchanger was tested with several test methods from foaming agents to Helium leak detection. With test pressures of >5 bar relative, all leaks could be detected with all test methods. The leaks could not be detected with optical inspection.

Next a water pressure burst test was performed with a water pressure of 50 bar relative pressure for a period of 1 hour. The sample was dried by dry air purge for a duration of one hour. Then the leak tests were repeated. Not a single leakage could be detected after this treatment. Only with test pressures of 10 bar could some leaks observed. The detected leak rates were orders of magnitude smaller than with the dry object.

So a drying process with an air purge over one hour is not at all sufficient for preparation of a moist sample for a tracer gas leak test. Only after vacuum drying at 20 mbar at a temperature of 200 °C over a period of 15 hours the majority of leakages could be detected again - but not all of them!

So we strongly recommend to perform a tracer gas leak test prior to a burst test with water. It only makes sence to perform a tracer leak test after a burst test if the burst test is likely to induce leak channels. However, the pre-requisite for the leak testin after a burst test is an efficient drying process.

(1) K.-O. Cavalar, Vakuum-Technik, 29. Jahrgang 1980, Heft 7, S 201ff.
(in German language; abstracts in English and French)