Written and maintained by CASRAI Editorial Board
Last updated
A sealed source that leaks is no longer “sealed” in the regulatory sense — it is a contamination hazard that has to be caught before it spreads to instruments, benches, or people. Under the U.S. Nuclear Regulatory Commission’s byproduct-material rules, radiation safety programs don’t leave leak testing to judgment. It’s a scheduled, documented wipe test with a defined trigger threshold and a defined response, and getting the frequency, method, or action level wrong is one of the more common findings in NRC and Agreement State inspections of sealed-source licensees.
The regulatory basis: 10 CFR 39.35
Leak testing of sealed sources is governed in the United States by 10 CFR 39.35, “Leak testing of sealed sources,” part of 10 CFR Part 39 (byproduct material — well logging and other Part 39-licensed uses of sealed sources, including calibration and check sources and fixed gauges). Most Agreement States — the states the NRC has authorized to regulate radioactive materials within their borders under a compatibility program — adopt requirements that are functionally equivalent to 39.35, so the numbers below hold regardless of whether your license is issued by the NRC directly or by your state radiation control program. Confirm your specific license conditions, since a license can impose a tighter interval than the regulatory default, but it can never impose a looser one.
Default interval: every six months
Under 39.35(c)(1), each sealed source in use must be leak tested at intervals not to exceed 6 months. This is the default that applies to the sealed sources most labs actually hold — calibration sources, check sources, and sources embedded in fixed gauges. There is one named exception: an energy compensation source (ECS), a low-activity sealed source built into some instruments specifically to compensate for energy-dependent response, gets a 3-year interval under 39.35(c)(2) instead of six months, because its use pattern and construction present materially lower leak risk. Don’t generalize the 3-year figure to any other source type — it’s specific to ECS.
There is no exemption for a source that has been sitting in storage rather than in active use. Testing is tied to the interval since the last test, not to hours of use, so a source pulled out of a cabinet after a year still needs a current leak test before it goes back into service if the interval has lapsed.
Method: the wipe test
The test itself is a wipe (smear) test, not a survey-meter scan. A swab or filter paper is wiped over the nearest accessible surface to the source where removable contamination would be expected to accumulate — the source housing, the shutter mechanism, or the immediately surrounding surface, depending on geometry — using a leak test kit or method approved by the NRC or the relevant Agreement State. The wipe sample then goes to analysis, typically liquid scintillation counting or a comparable method, performed by a person approved by the NRC or Agreement State to do that analysis. The analysis method itself has a defined sensitivity floor: it must be capable of detecting 185 Bq (0.005 microcurie) of removable activity on the sample, which is the same number that sets the action level below — the test has to be sensitive enough to actually catch a failure at the regulatory threshold, not just report a number.
The action level: 0.005 microcurie (185 Bq)
If the wipe test analysis shows 185 Bq (0.005 microcurie) or more of removable radioactive material, that is a positive result under 39.35(d)(1), and it triggers an immediate, specific response:
- Remove the sealed source from service immediately — it cannot continue to be used pending further evaluation.
- Have it decontaminated, repaired, or disposed of by a licensee authorized by the NRC or an Agreement State to perform that work — not handled in-house beyond removal from service.
- Check any equipment associated with the source (the gauge housing, instrument internals) for contamination, and decontaminate or dispose of that equipment as needed.
- Report the result to the NRC or Agreement State, per the reporting requirements referenced in 39.35(d) — check your specific license and jurisdiction for the exact reporting deadline, since this detail varies by regulatory authority.
0.005 microcurie is a small number by design: it’s meant to catch a breach in the source’s seal well before the leaked activity is large enough to be a meaningful contamination or dose event on its own. Treat a positive result as a containment failure to be acted on immediately, not a borderline reading to double-check first — double-checking happens on a fresh wipe from an authorized tester, with the source already out of service, not by continuing to use it while you decide.
Exemptions: what doesn’t need routine leak testing
39.35(e) exempts several categories of sealed source from the routine leak-test schedule entirely, because their construction or activity level makes a leak either physically impossible or radiologically insignificant:
- Gaseous sources. Sealed sources containing licensed material in gaseous form are exempt — a wipe test can’t meaningfully detect a gas leak the way it detects removable particulate/liquid contamination, so the leak-test framework doesn’t apply to this source type.
- Short half-life. Sources containing material with a half-life of 30 days or less are exempt — the source decays to insignificance faster than a routine test cycle would catch a problem, so the regulation doesn’t require one.
- Low-activity beta/gamma sources. Beta- or gamma-emitting sources with an activity of 3.7 MBq (100 microcuries) or less.
- Low-activity alpha/neutron sources. Alpha- or neutron-emitting sources with an activity of 0.37 MBq (10 microcuries) or less — a lower threshold than beta/gamma, reflecting the higher radiological significance of alpha and neutron contamination at a given activity.
An exemption from leak testing is not the same as an exemption from your license’s inventory and possession-limit requirements — an exempt source still needs to be accounted for; it just doesn’t need a leak-test entry on the schedule.
Recordkeeping
Leak test results must be recorded in units of activity (microcuries or becquerels, matching the action-level units), and the record must be retained and available for NRC or Agreement State inspection. Keep leak-test records with the same discipline you’d apply to dosimetry or survey-meter calibration records — an inspector who asks for the last six leak-test results on a specific source expects to see a continuous, undated-gap-free history, not a reconstruction from memory.
Frequently asked questions
Does a source that’s never removed from its shielded storage still need leak testing?
Yes, unless it falls under one of the 39.35(e) exemptions above. The six-month clock runs from the date of the last test, not from hours of active use, so long-term storage does not pause the testing requirement on its own.
What counts as a “sealed source” for this rule?
A sealed source is radioactive material permanently bonded or encased in a capsule or matrix designed to prevent leakage or escape of the material under the conditions of use it’s licensed for — the calibration sources, check sources, and gauge sources this rule targets are the common examples in a research or clinical lab setting.
Who is allowed to perform the leak test and the analysis?
The wipe itself is typically performed in-house by trained radiation safety staff using an approved kit or method, but the analysis of the wipe sample must be performed by a person approved by the NRC or the relevant Agreement State to do that analysis — many labs send samples to an outside vendor or their institutional radiation safety office for this step.
Is 0.005 microcurie the same threshold everywhere?
185 Bq (0.005 microcurie) is the federal action level under 10 CFR 39.35(d)(1). Agreement States generally adopt the same figure under their compatible regulations, but always confirm against your specific license and state program rather than assuming.
