Counting Results for Acid used to soak the UA 137Cs Source




Zelimir Djurcic and Andreas Piepke
University of Alabama
(Status 9/2/2002)




Procedure

After the identification of statistically significant activities on the surface of LSU's AmBe neutron sources we decided to also perform an acid soak with UA's 137Cs calibration source. In July 2002 the source was soaked for 20 hours in 0.1 molar HNO3. Counting of the soak acid showed evidence for the presence of 123mTe and 137Cs activities on the surface of the source.
The first observation was that the source encapsulation reacted very strongly with the very dilute acid (0.6% acid by volume). This unexpected behavior, not seen for test pieces made from stainless steel, will be discussed by Jerry Busenitz in a separate memo.
The approximately 100 ml soak acid were subsequently counted using our low background Ge detector. The counting efficiency has been experimentally determined using a calibrated mixed standard solution. For a 137Cs surface contamination or leakage we would expect a gamma peak at 661.7 keV.

Results

[link to PDF file] Figure 1: Gamma ray spectrum recorded during 4.844 days of counting with the acid used to soak the source.

Figure 1 shows the full energy range recorded during 4.844 days of counting with the acid used to soak the source. In this representation the only net peak which clearly exceeds the background is seen at 159 keV. [link to PDF file] Figure 2: Peak at 159 keV.

Figure 2 shows the energy range around that peak. The accumulated significance of this peak is with 35 sigma beyond statistical doubt. The fact that no other associated gamma lines are observed makes an unambiguous nuclide identification difficult. In order to identify the decay we performed a time differential measurement. [link to PDF file] Figure 3: Time development of the 159 keV peak counting rate.

The temporal development of the 159 keV peak counting rate is depicted in figure 3. Although this data does not yet allow a conclusive determination of the decay time a short half life is clearly excluded. At this time, 123mTe (T1/2=119.7 d) is the most likely source of contamination. All other decays within ± 2 keV of the peak energy and within the 90% cl. interval of the observed half life can be excluded as other gamma lines associated with those decays are not observed. This is a surprising and unexpected finding. The corresponding 123mTe activity of the acid is with 192±19 mBq (error is dominated by 10% systematics) relatively large. If we take into account that the solubility of 123mTe in 0.1 molar HNO3 is with 18% relatively poor we can infer a total surface activity of 1.1±0.2 Bq for this source.
We can, at this point, offer no explanation how a 123mTe contamination can happen on a 137Cs source. A contamination at the source manufacturer seems the most likely explanation. There open activities are handled all the time and sources typically come certified to have less then 100 Bq on their surface. Clearly our source would have passed such test. Our tests of the solubility of radio nuclides showed that 123mTe is difficult to remove from stainless steel. While a 24 hour soak in 0.1 molar HNO3 would typically only leave less than 0.1% for most other activities on the surface, 18% were left in case of Te. This is how the Te might have resisted source cleaning. Clearly the Cs source needs a more vigorous cleaning before deployment. However, it should be noted that 123mTe constitutes no background for KamLAND as (1) its decay energy is very low, and (2) no charge particles are emitted in this decay. It is on the other hand a warning sign to find foreign activities on KamLAND sources. It seems imperative to also test KamLAND's other gamma sources for surface contaminations. As we have shown earlier exposure to liquid scintillator only removes a small fraction of surface contaminations. Repeated source deployment might hence result in a slow removal of such surface impurities.
[link to PDF file] Figure 4: Energy range around the Cs decay energy of 661.7 keV.

Figure 4 shows the energy range around 661.7 keV, were we would expect the Cs peak. Although, certainly still of limited significance a Cs peak is identified. The peak fit indicates a peak significance of 4.6 sigma. The peak energy was fixed to the known value and the peak width to that determined for the neighboring strong background line. A free fit returned actually the same width and peak energy. Figure 5 below shows the background as measured in the same energy range. [link to PDF file] Figure 5: Background spectrum around 662 keV.

Clearly no such peak is observed in the background spectrum, depicted in figure 5. We therefore conclude that a 137Cs activity of 15±4 mBq (1100 gammas per day) was present on the surface of the source. At this point we don't know whether this activity is due to source leakage or an external contamination. In case of leakage the source has to be either discarded or an additional encapsulation is required. In case of an external contamination an appropriate acid cleaning would suffice to make the source fit for deployment. Only a re-soak can resolve this ambiguity. However, before this can happen it needs to be understood why the source encapsulation showed such a strong reaction with the weak acid.
It is interesting to note that in a 24 hour scintillator soak (use 3.5 ml liquid scintillator; no liquid motion) only 0.5 mBq would have been transferred into the liquid. If we further assume that our soak tests are a truthful simulation of a KamLAND source deployment then this is what we could expect for the activity transfer into the liquid scintillator. The resulting contribution to KamLAND's solar neutrino background would, in this model, be 40 events per day and per source deployment. But any conclusion on the contribution of a source contamination to KamLAND's background has to rely on a set of un-tested assumptions. It is e.g. not clear whether the activity transfer is proportional to the exposure time or the number of deployments. It seems therefore to be the safest strategy to require that no activities larger than say 10% of KamLAND's background be present on the surface of a source. Clearly the 137Cs source would not pass that test!

Conclusion

A 20 hour acid soak of KamLAND's 137Cs showed evidence for the presence of 123mTe and 137Cs activities on the surface of the source. At this point it is unclear whether these are sustained by leakage or just due a cleanable surface contamination. In its current state this source is unfit for deployment into KamLAND. It seems prudent to subject KamLAND's other gamma sources to a similar test at the earliest convenience.

This page is maintained by A. Piepke
Last update September 4, 2002