Solubility of Radio Nuclides in dilute Acid and Liquid Scintillator




Zelimir Djurcic and Andreas Piepke
University of Alabama
(Status 8/30/2002)
(Updated 9/27/2002)




Introduction

KamLAND calibration sources have been tested for leakage and surface contamination in two different ways:
(1) By soaking in liquid scintillator which in its composition is similar to that used in KamLAND.

(2) By soaking in 0.1 molar HNO3, followed by rinsing with DI water.

Method (1) has been used to screen all UA sources before delivery to Japan. Stringent limits have been reported. However, the recent finding of an Am contamination on LSU's AmBe neutron sources after soaking them for an extended period of time in 0.1 molar HNO3 has raised the question whether method (1) is indeed capable of detecting such problem. When talking about leaking or contaminated sources one should keep in mind that the source manufacturer typically guarantees less than 100 Bq on the surface for a sealed source. This is 1000 times more than our test sensitivity.
Clearly method (1), by its design, was intended to probe the activity released into the liquid scintillator and therefore to directly asses the impact of the source deployment on KamLAND's background. However, it has to be said that such conclusion relies on the implicit assumption that KamLAND's source deployment is truthfully mirrored in the tests performed. This can at least be questioned. While in KamLAND the body of liquid is infinite, the amount of liquid present in the relatively small soak vessels is finite. Different chemical concentration of a hypothetical contamination could result in different activity releases. While in the leak tests the source is held perfectly still and experiences no mechanical agitation the source deployment in KamLAND requires relatively rapid movement through the liquid. It is not inconceivable that the liquid flow might more effectively wash off a surface contamination.
All this said we decided to ask a different and more unambiguous question: how much activity is present on the surface of the source. The choice of organic liquid scintillator does not seem to be ideal for this test. Metals typically exhibit a rather poor solubility in organics. They do, on the other hand, dissolve very well in dilute acid. This claim is supported by the observation that open calibration sources are always sold in acidic solution by the source manufacturers, never in organic solution. We decided to measure the difference of the effectiveness of cleaning methods (1) and (2) by externally contaminating a piece of stainless steel (same type as used for KamLAND's gamma sources) which had been machined to the same surface finish as the sources. The very low acidity of the cleaning agent (0.1 molar HNO3 corresponds to only 0.6% acid by volume) was chosen to assure compatibility with almost all metals. The cleaning itself should not compromise the source integrity.

Procedure

In order to deposit radio isotopes on the surface of our source dummies we used two different solutions containing radio isotopes. The first consists of 4 molar HCl containing calibrated activities of 51Cr, 57Co, 60Co, 85Sr, 88Y, 113Sn, 109Cd, 123mTe, and 137Cs (all gamma emitters). The use of this source cocktail allows the simultaneous testing of many different radio nuclides. 60Co and 137Cs are used in KamLAND. This data can therefore be used to gauge the validity of the source tests performed previously for these nuclides. The only gamma source material not included in our tests is 65Zn.
In a second measurement we contaminated two different source dummies with a 1 molar HCl solution containing 241Am. The solubility data collected with these samples allows us to put the leakage rate observed or the respective limits into some perspective.
In both cases 5 micro liters of solution were distributed over the surface of two SS pieces using syringes. The source dummies were then dried under a heat lamp for 4 to 24 hours. I was noted that only a fraction of the deployed activities ended up on the metal.
After drying the source dummies were packed in small 4 ml PE bottles and counted. This data established the activity baseline before cleaning.
The source dummies, called bolt 1 and 2 from here on, were then transfered into virgin 4 ml PE bottles and 3 ml of 0.1 molar HNO3 and 3 ml of liquid scintillator was added to perform the source cleaning. The primary source bottles were counted and it was noted that for bolt 1 5% and for bolt 2 20% of the deployed activities were lost to the vial's surface. This loss is just due to the fact that the bolts were touching the bottle's surfaces. It shows how important mechanical manipulations are, lending more weight to our argument on the details of the source deployment given above.
After exposing bolts 1 and 2 for 24 hours to their respective cleaning agents they were recovered from the liquid and rinsed with 3.5 ml of DI water and fresh liquid scintillator, respectively. Counting before the rinsing of bolt 1 showed that the activity adhering to the surface was reduced by a factor 4. Rinsing is therefore an indispensable step in the cleaning. Both the cleaned bolts and the respective cleaning fluids were then counted to determine the effectiveness of the cleaning procedures.
The procedures applied during the Am testing follow those described before. They will therefore not be described again.

Results

[link to PDF file] Figure 1: The blue spectrum was taken with the contaminated source dummy, the black spectrum with the 0.1 molar HNO3 and rinse water used to clean it, and the red spectrum with the source dummy after a 24 hour acid soak and rinsing with DI water. This data was taken with the nine-activity source cocktail.

Figure 1 shows the spectra measured with UA's low background Ge detector for the various components, using the nine-activity source cocktail. We see from the figure that the blue and black spectra, measured with the contaminated bolt and the cleaning fluid are virtually identical. The red spectrum, taken with the cleaned bolt 1, exhibits much lower rates. This already demonstrates the high efficiency of the acid cleaning.
[link to PDF file] Figure 2: The blue spectrum was taken with the contaminated source dummy (using our nine-activity source cocktail), the black spectrum with the liquid scintillator used to clean it, and the red spectrum with the source dummy after a 24 hour scintillator soak and rinsing.

In contrast to that figure 2 shows that the spectrum taken with the contaminated (blue) and cleaned (red) bolt 2 are virtually identical. The activity removal observed in the scintillator soak is miniscule. The second disturbing observation is that the counting of the soak scintillator (although with a very low concentration) does show that 57Co, 60Co, 88Y, 113Sn, and 137Cs can go into solution. A few percent of the deployed activities is found in the scintillator. This clearly demonstrates that it is possible to transfer activity from a contaminated source into KamLAND scintillator.
[link to PDF file] Figure 3: The same as shown in the preceding two figures but using an open 241Am source. The energy range covers three prominent Am gamma lines. Spectrum taken with the contaminated source dummy is shown in blue, the spectrum with the acid (left panel)/liquid scintillator (right panel) used to clean it is shown in black, and and the red spectrum with the source dummy after a 24 hour soaking in acid (left panel) and liquid scintillator (right panel), respectively.

The following table summarizes our results of the quantitative analysis of the various activities detected. The tables lists the various measured activities. Values in red (acid) and blue (scintillator) denote the fractional activities in [%]. R denotes the ratio of the fractional activities detected on the acid cleaned bolt 2 divided by that measured for the scintillator cleaned bolt 1. R is therefore equivalent to the ratio of the solubilities of the various elements in acid and liquid scintillator, respectively. All errors are statistical only. The systematic error, mainly due to differences in the solid angle when placing different samples on the Ge detector, is of the order 10%. We have chosen not to fold in the systematic error to allow direct judgment on the statistical significance of the various measurements.

Bolt 1 (acid cleaning) Bolt 2 (scintillator cleaning)
Nuclide Activity on bolt before cleaning [Bq] Activity in acid [Bq]
% total 		Activity on bolt after cleaning [Bq]
% total 		Activity on bolt before cleaning [Bq] Activity in scint. [Bq]
% total 		Activity on bolt after cleaning
% total 		R
(Bolt 1/Bolt2)
51Cr 5.6±0.8 5.2±0.2
93±14		 -0.026±0.034
-0.5±0.7 		7.3±0.8
 0.24±0.13
3.3±1.8 		8.2±1.0
112±19 		> 98
57Co 1.9±0.2 2.0±0.1
105±12 		0.030±0.009
1.6±0.5 		1.3±0.2 0.097±0.027
7.5±2.4 		1.7±0.2
131±26 		82±30
60Co 13.4±0.3 12.9±0.1
96±3 		0.014±0.003
0.10±0.02 		13.3±0.3 0.52±0.04
3.9±0.3 		13.4±0.4
101±4 		1010±206
85Sr
88Y 8.4±0.2 7.6±0.1
55±2 		0.0084±0.0030
0.10±0.04 		11.1±0.3 0.32±0.03
2.9±0.3 		11.8±0.3
106±4 		1060±426
109Cd 87.2±9.0 88.4±2.6
101±11 		-0.34±0.39
-0.4±0.5 		23.5±7 1.3±1.4
6±6 		49.2±9.6
209±75 		> 255
113Sn 7.8±0.3 6.6±0.1
85±4 		0.22±0.01
2.8±0.2 		4.6±0.2 0.17±0.03
3.7±0.7 		4.7±0.3
102±8 		36±4
123mTe 0.89±0.11 0.75±0.03
84±11 		0.16±0.01
18±3 		0.7±0.1 0.04±0.02
6±3 		0.7±0.13
100±24 		5.6±1.6
137Cs 8.6±0.3 8.4±0.1
98±4 		0.0039±0.0042
0.05±0.05 		11.3±0.3 0.35±0.04
3.1±0.4 		12.2±0.4
108±5 		> 820
241Am 12166±102 14709±454
121±4 		73±8
0.60±0.06 		12780±251 543±12
4.3±0.1 		12086±245
95±0.01 		158±16
Nuclide Activity on bolt before cleaning [Bq] Activity in acid [Bq]
% total 		Activity on bolt after cleaning [Bq]
% total 		Activity on bolt before cleaning [Bq] Activity in scint. [Bq]
% total 		Activity on bolt after cleaning
% total 		R
(Bolt 1/Bolt2)
Bolt 1 (acid cleaning) Bolt 2 (scintillator cleaning)


From the table above we see several interesting things.
(1) The solubility of all radio nuclides tested is much better in acid than in liquid scintillator. This does not come as a surprise! The factors for the relevant activities 60Co, 137Cs, and 241Am are 1000, 800, and 160 respectively. The accuracy of the data presented here is limited by the smallness of the activities left on the acid cleaned bolt. Longer counting (the data shown here corresponds to about one day of counting) would allow to quantify the solubility ratio more accurately. It is also interesting to note that Sn and Te are not as effectively removed by the weak acid treatment as the other elements. To address these would need tests with more concentrated acid.
(2) Scintillator soaking is largely ineffective in removing surface activities. Only a few percent of the surface activity is removed in a one day soak. We have, at this point, no data which would support or reject the assumption that longer soaking would make a difference. Testing of source leakage by scintillator soaking is therefore problematic as it can not be expected that any reasonable transfer of activity takes place. Limits obtained after counting soak scintillator are therefore not a direct measure of the activity present on the source's surface. However, if we are ready to make the assumption that the soak tests properly mirror source deployment then they give some measure of a possible contamination of KamLAND's scintillator by such source.
(3) If we are ready to make the assumption mentioned above then we can divide the activity limits obtained after acid cleaning by a factor 25 to obtain a measure for a hypothetical activity transfer into the scintillator. We also know that the acid soak leaves behind 100 times less activity on the source surface than was there before, if there is no leakage. If we are not prepared to make such assumption then the acid measurement gives a hard limit on any possible contamination, as surface activities are fully transferred into the acid. As an added bonus the acid treatment is actually cleaning the source, as stated before. This is not the case for the scintillator soak. At activity transfers of about 4% the scintillator soak seems more like a symbolic measure.


Conclusion

The data presented here clearly demonstrates the superiority of acid cleaning of sources over scintillator cleaning. Measurements with open activities have also demonstrated that it is possible to contaminate the scintillator as the respective solubilities are not zero.
We should therefore immediately switch to regular acid soaking before deploying calibration sources.

Acknowledgment

We want to thank Jerry Busenitz and Kevin McKinny for their help and assistance during the preparation and execution of these measurements.

This page is maintained by A. Piepke
Last update September 28, 2002