Journal of Radioanalytical and Nuclear Chemistry, Articles, Vol. I I 1, No. 1 (I 98 7) 111-116

NEUTRON-ACTIVATION ANALYSIS OF THORIUM IN ACRYLIC SAMPLES G. E. AARDSMA,P. JAGAM,J. J. SIMPSON Department of Physics University of Guelph, Guelph, Ontario NIG 21r (Canada} (Received September 10, 1986) The techniques employed to measure Th concentrations in acrylic (PMMA) samples at a level of a few tens of ppt are described. Measurements on eleven different acrylic samples yielded a mean Th concentration of 40-+20ppt.

Introduction We have recently been faced with the problem of measuring the concentration of thorium in acrylic [poly(methyl methacrylate)] at the level of a few tens of parts per trillion (ppt) by weight. The need to make such a measurement arose in the context of the proposed Sudbury Neutrino ObservatoryJ This proposal is to build a 1000 tonne D 2 0 Cerenkov detector deep underground in a nickel mine near Sudbury primarily to measure the solar neutrino spectrum. The D20 would be contained in a cylindrical acrylic tank. An important feature of the use of heavy water is the ability to detect the neutral current reaction v + d --> p + n + v by detecting the neutron. The expected event rate of this reaction is on the order of only a few per day. The presence of 232Th in the acrylic would give rise to free neutrons by photodisintegration of the deuteron (Q = 2.22 MeV) which would be a significant source of background in this detector. To keep this source of background at acceptable levels the thorium concentration in the acrylic must be below about 10 ppt. This paper describes the techniques which we have successfully employed to measure thorium concentrations around 40 ppt in acrylic samples using instrumental neutron activation analysis (INAA). Our emphasis in this work was to explore the sensitivity limits of INAA for the measurement of thorium in acrylic.

Elsevier Sequoia S. A., Lausanne Akaddmiai Kiad6, Budapest

G. E. AARDSMA et al.: NEUTRON-ACTIVATIONANALYSIS OF THORIUM

Experimental The solid acrylic sample to be measured was generally machined to produce a rod 1.6 cm in diameter and 3.8 cm long; the sample mass was then roughly 9 grams. The sample was encapsulated in a cylindrical aluminum container for insertion into the neutron reactor pool. The container had an inside diameter of 1.9 cm and a height of about 4.3 cm. A small piece of aluminum foil (typically 0.05 g) was also included in the container to serve as the standard. The aluminum foil contained 110+10 parts per billion of thorium. This foil was developed as an in-house standard for thorium determination by INAA. It is standardized against three international rock standards (JG1, BCR1, and AGV1) used in geochemical work. Neutron irradiation of the sample was carried out at the 2.0 MW McMaster Nuclear Reactor. z The neutron flux was 1013 n 9 era -2 9 s -~ at the irradiation site and samples were usually irradiated for 2 hours (limitations on irradiation time are discussed below). The concentration of thorium in the acrylic samples was measured by detecting the 312 keV "r-rays from the decay of z33Pa which is produced from the thorium in the sample during neutron irradiation. This isotope has a 27.0 day half-life. After irradiation the short lived isotopes such as 24Na gave rise to a very large background count rate. This background was allowed to decay for two weeks before "/-counting of the sample was begun. The gamma-ray spectra from the activated acrylic samples were recorded Using a low background 208 cm 3 intrinsic germanium gamma-ray detector. 3 The detector was shielded with a 15 cm passive Pb shield and an active five sided cosmic-ray veto outside the Pb operated with a 20/.ts gate to the ADC. The passive Pb shield reduced the background at 312 keV by a factor of 100 compared to that for the unshielded detector. The five sided veto provided an additional factor of two reduction. This yielded an operating background rate (from non-sample sources) of 2 counts/keV/hour. When an activated acrylic sample was placed inside the shielding the background in the 312 keV region rose to between 15 and 150 counts/keV/hour depending upon the trace' impurities present in the sample. The FWHM of the 7-ray peak at 312 keV was 2.5 keV. The MCA was operated at 1 keV/channel. Gamma-ray spectra were generally recorded for 18 to 70 hours for the activated acrylic sample; the standard aluminum foil was usually counted for 18 hours. The parameters given above yield an interference free counting sensitivity of 150 counts per 40 hours in the thorium peak at the required limit of 10 ppt.

112

G.E. AARDSMAet al.: NEUTRON-ACTIVATIONANALYSISOFTHORIUM k

Results

Figure 1 shows the gamma-ray spectrum recorded for an acrylic sample for 67.8 hours. The region from 200 keV to 400 keV has been expanded in Fig. 2. The dominant peak near 312 keV is due to the Cr line at 320 keV. This intense peak was present in all of the acrylic samples. However, the peak due to Th is still clearly visible at 312 keV. t

I 102~ lo

1

O

,

I

500

j

I

1000

t

I

i

I

~500 2000 Energy, keV

Fig. 1. Recorded3,-rayspectrumfor acrylicsampleNo. 8 The final results from the thorium concentration determination of eleven different acrylic samples and one distilled water sample are presented in Fig. 3. The first six samples (Nos 5, 8, 3, 11, 13, and 16) were commercial grade acrylic. Sample No. 8 was obtained from 6.4 mm rod stock; the other five samples were machined from 15.9 mm rod stock. The remaining five acrylic samples (Nos 14, 19, 21, 23, and 24) were provided by the acrylic tank manufacturer. These samples were taken after various stages of the manufacturing process. Sample No. 19 was taken from the raw PMMA powder stock which comprises approximately 50% of the final solid acrylic. The high Th concentration measured in sample No. 23 may have been the result of contamination from a surface polishing compound.

8

113

G. E. AARDSMA et al.: NEUTRON-ACTIVATION ANALYSIS OF THORIUM

-~ 106 E E 0

320 Cr

E 10s O

10 3

10 2 2O0

,

I 300

J

I ~. 400 Energy, keV

Fig. 2. Recorded "y-ray spectra for acrylic sample No. 8 (curve 1) showing peaks due to Th and Cr and a distilled water spectrum (curve 2)

~. 102 E

E r

8 "E O r

I0

I

r #5

J

#8

i

03

f

#11

I

I

1

I

I

#13

#16

#14

#,19

#21

I

I

#23 #24 Somple No.

Fig. 3. Measured thorium concentrations in eleven acrylic samples and one distilled water sample

114

G. E. AARDSMA et al.: NEUTRON-ACTIVATIONANALYSIS OF THORIUM The average concentration of Th in the acrylic samples which we have measured is 40+20 ppt. Discussion

When working with such small concentrations the question of contamination of the sample during the measurement cannot be ignored. In particular the aluminum container was considered as a possible source of Th contamination since it has a Th concentration many orders of magnitude greater than the acrylic samples. Fig. 2 shows, besides the acrylic spectrum already mentioned, the gamma ray spectrum obtained from a distilled water sample. The distilled water was simply poured into the aluminum irradiation container which was then capped by cold welding as usual. The measured Th concentration in this sample was 0.6---1.6 ppt which provides evidence against enhancement of the 312 keV peak due to contamination of the sample from Th in the aluminum of the container. This result for the water is shown in Fig. 3 for comparison to the acrylic results. The water measurement also clearly demonstrates the potential sensitivity of this technique. The main limitations on the sensitivity for Th detection in the acrylics were due to background at 312 keV from the trace impurities and the inability to irradiate the samples for longer than two hours. These two limitations and possible means of reducing their effect are discussed in what follows. The activities from the trace impurities which produce the Compton continuum at 312 keV are identified by their principal 7-rays to be: s 1Cr (320 keV), x98Au (412 keV), lz2Sb (564 keV), 124Sb (603, 1691 keV), x~~ (658,885 keV), S9Fe (1099, 1292 keV), 6SZn (1115 keV), XSZTa (1121, 1221, 1231 keV), 6~ (1173, 1332 keV), and 14~ (328, 1596 keV). However, the relative ratio of the activities is not the same from sample to sample. These impurities are not considered to be surface contaminants introduced during the preparation of the samples for irradiation. In some more recent samples obtained from the tank manufacturer the trace impurity activity levels are found to be factors of two to four lower than previous samples. These more recent samples were specially formulated to exclude several components present in the earlier acrylics which are not required for the D20 tank construction. Thus, by excluding these unnecessary components the risk of introducing thorium to the acrylics is reduced and the sensitivity for thorium determination is enhanced. Preliminary results from two such samples gave a thorium concentration of 7+9 ppt. The acrylic samples were observed to suffer severe radiation damage during neutronactivation. This resulted in fracturing of the solid samples and evolution of decomposition gases. After a two hour neutron irradiation at 1013 n 9 cm -2 9 s - ' the aluminum canister was pressurized by the evolved gases and the acrylic samples were 8*

115

G. E. AARDSMA et al.: NEUTRON-ACTIVATION ANALYSIS OF THORIUM fractured to a coarse powder. Attempts to irradiate the acrylic samples for longer than two hours invariably resulted in the rupture of the aluminum canister. This resuited in loss o f some sample into the reactor pool and contamination of the sample by the reactor water. The use of a special canister designed to allow escape o f evolved gases4 could possibly allow much longer irradiation times and remove this limitation. In conclusion, we have demonstrated a procedure which is applicable for thorium concentration measurements in acrylic above 20 ppt. In specially prepared acrylics the absence of some trace impurities improves the sensitivity to approximately 10 ppt. By devising a container suitable for longer neutron irradiations of the acrylic samples we expect ultimately to be able to achieve a sensitivity of 3 - 4 ppt which will be adequate for our application.

Credit is due to the staff of the University of Guelph machine shop for machining of the acrylic samples; Mr. Start BONILLAS of the Reynolds-Taylor acrylic manufacturing company for providing samples for analysis; and the staff of the MeMaster Nuclear Reactor for their excellent technical assistance and close cooperation throughout this work.

References la. lb. 2. 3.

G. T. EWAN et al., Report SNO-85-3, Queen's University, Canada, 1985. D. SINCLAIR et al., Nuovo Cimento, C9 (1986) 308. McMaster Nuclear Reactor, McMaster University, Hamilton, Ontario, Canada, L8S 4K1. J.J. SIMPSON, P. JAGAM, J. L. CAMPBELL, H. L, MALM, B. C. ROBERTSON, Phys. Rev. Lett., 53 (1984) 141, 4. J.I. KIM, H. STARK, I. FIEDLER, Nuel. Instr. Methods, 177 (1980) 557.

116