Sequential Measurements of Cosmic-Ray Neutron Energy Spectrum and Ambient Dose Equivalent on the Ground

Transcription

1 Sequential Measurements of Cosmic-Ray Neutron Energy Spectrum and Ambient Dose Equivalent on the Ground T. Nakamura 1,2, T. Nunomiya 1,2, S. Abe 1, K. Terunuma 1 1 Cyclotron and Radioisotope Center, Tohoku University Aoba, Aramaki, Aoba-ku, Sendai, , Japan nakamura@cyric.tohoku.ac.jp 2 Radiation Systems Division, Fuji Electric Systems Co. Ltd. 1, Fuji-machi, Hino, Tokyo , Japan Abstract. The cosmic-ray neutron energy spectrum and dose rate were measured sequentially for more than two years from October 2000 up to March 2003 by using three neutron detectors, a 3 He-loaded multi-moderator detector (Bonner ball), 12.7 cm diam by 12.7 cm long NE213 liquid organic scintillator, and high-sensitivity rem (dose equivalent) counter. The Bonner ball is composed of a 5.08-cm diam spherical 3 He counter filled with 5 or 10 atm 3 He gas, which is placed at the center of a set of polyethylene spheres of different diameters of 8.1, 11.0, 15.0, 23.0 cm. The rem counter consists of a 12.9-cm diam 5 atm 3 He gas filled spherical proportional counter covered with polyethylene moderator and inner absorber. The measuring interval is every 12 hours for Bonner ball, every 30 minutes for rem counter, and much longer accumulation period for NE213. This measurement was carried out at the Kawauchi campus of Tohoku University in Japan. The neutron energy spectra and the neutron ambient dose equivalent were obtained on the ground level with an atmospheric pressure correction. The neutron spectrum has three major peaks, thermal energy peak, evaporation peak around 1 MeV and cascade peak around 100 MeV. There can be seen a tendency that a cascade peak increases on the day of higher solar activity having higher neutron flux compared with other two peaks. The ambient neutron dose equivalent measured by the rem counter keeps an almost constant values of 4.1 [nsv/h] throughout this time period, after atmospheric pressure correction, and it often decreased about 30% after a large Solar Flare, that is called as the Forbush decrease. The total neutron flux was also obtained by the Bonner ball measurements to be 7.5 x 10-3 [n cm -2 sec -1 ] in average. 1. Introduction The Sun has an 11-year solar cycle, and the solar activity becomes a maximum in the period of During the solar maximum, big solar flares sometimes happen and give large effects on our earth environment. Dynamic outbreaks of geomagnetic storms and radiation showers to the Earth occur sporadically but with increasing intensity and frequency during the years around maximum. Proton is the main element among the solar event particles that reach the earth. When the proton comes into and interacts with the atmosphere, high-energy secondary neutrons and another electromagnetic waves are generated. About 90% of the secondary particles that can reach the ground are photons and muons, and the rest is neutrons. There are several places in the world where neutron fluxes have been sequentially measured using the moderated-type large BF 3 detectors. These detectors however only give the neutron counts, but not energy spectra. Recently, an accumulation of the semiconductor device greatly increases and the soft-errors of SRAM and DRAM on the ground level caused by high-energy cosmic-ray neutrons become a serious problem in the world. In this study, a sequential measurement has been carried out from October 2000 to March 2003 in order to investigate the long-term changes of cosmic-ray neutron flux and dose rate on the ground using three types of neutron detectors, Bonner ball (multi-moderator spectrometer with 3 He counter), high-sensitivity rem counter and NE213 organic liquid scintillator.

2 2. Materials and methods For the long-term sequential measurements of cosmic ray neutrons, we set up an experimental cabin on the ground at the Kawauchi campus of Tohoku University. The cabin is a size of 564 cm x 230 cm x 262 cm and air-conditioned. The measuring position is situated at latitude north and longitude east and altitude of about 70 m. The geomagnetic latitude in Sendai is about 29 and the cut-off rigidity is GV. The Bonner ball is composed of a 5.08 cm diam spherical 3 He counter filled with 5 atm or 10 atm 3 He gas which is placed at the center of a set of polyethylene moderators of different diameters of 0 (Bare), 8.1, 11.0, 15.0 and 23.0 cm [1]. The high-sensitivity rem counter consists of a 12.9 cm diam 5 atm 3 He gas-filled spherical counter covered with polyethylene moderator and inner absorber [2] for measuring the neutron dose rate. The 12.7 cm diam by 12.7 cm long NE213 organic liquid scintillator is also used to measure the neutron energy spectrum in the MeV region, while the Bonner ball gives the energy spectrum over the wide energy region down to thermal energy, but with poor energy resolution, especially in high-energy region above about 10 MeV. The NE213 detector has a sensitivity not only for neutrons but also for gamma rays, then the neutron events were separated from gamma-ray events with the pulse-shape discrimination technique [3] which uses the total and slow gates, according to the fact that the light output pulses from neutrons are slower than those from gamma rays. The measuring interval is every 30 minutes for the rem counter, every 12 hours for the Bonner ball, and much longer accumulation period of about several days for NE213. From the total counts above the cut-off level measured with the Bonner ball, neutron energy spectrum was obtained by unfolding with the SAND-II code [4] and the response functions of five detectors that are calculated with the MCNPX Monte Carlo code [5] up to 1 GeV neutron energy [6]. In this unfolding we used the initial guess spectrum which is given as a typical cosmic-ray neutron spectrum by Goldhagen et al. [7] The neutron dose rate was then estimated by multiplying the neutron energy spectrum with the fluence-to-ambient dose equivalent conversion factor given by ICRP 74 [8]. The total counts of the rem counter were also converted into ambient dose rate by applying the conversion factor of 18.5 cps/(µsv h -1 ), which was recalibrated at JAERI (Japan Atomic Energy Research Institute) using a 252 Cf neutron source [9] for this measurement. The neutron energy spectrum in the MeV energy region was obtained by unfolding the pulse height distribution of NE213 with the FORIST code [10] using the response functions in the energy range of 5 MeV to 800 MeV given by Sasaki et al. [11] 3. Results and discussions 3.1 Atmospheric pressure effect The effect of atmospheric pressure to the neutron dose rate is exemplified in Fig. 1 during the period of March 9 to October 4, The dose rate clearly indicates the inverse effect to the atmospheric pressure and can be fitted to the exponential attenuation curve as follows, y = 2.2 exp (-6.5 x 10-3 x), (1) where y is ambient dose rate in µsv/h and x is atmospheric pressure in hpa. Eq.(1) is generally written as y = A exp( - µ x ), (2) where y is the measured total counts of the Bonner ball and the rem counter. All measured total counts, y, at any atmospheric pressure are extrapolated to the value, y, at 1013 hpa as

3 y = y exp[ - µ ( x)]. (3) The µ values obtained by the least mean square fitting are tabulated in Table I. 3.2 Neutron ambient dose equivalent The ambient dose equivalent rates measured by the rem counter were compiled during the daytime from 6:00 am to 6:00 pm and the nighttime from 6:00 pm to 6:00 am, for investigating the effect of solar protons directly coming from the sun. If the neutron dose greatly influenced by solar protons, the neutron dose in daytime facing to the sun will be a little higher than that in the nighttime. Figure 2 shows the ratio of neutron dose rates between daytime and nighttime. The ratio is almost equal to 1 within only 3 % difference, which indicates that this effect is negligibly small. Figure 3 shows the time-sequential data of neutron ambient dose equivalent rates obtained by the rem counter. The dose rate keeps almost constant of about 4.1 nsv/h with a variation of about 5% over the entire time period of April 2001 to March 2003, although there are some remarkable variations at A and B points, especially on April 16 as described later. At the A point, the neutron ambient dose was steeply decreased about 30% and this phenomenon was kept in a few days. In these days, it was snowing in Kawauchi campus and about 20-cm-thick snow was piled up on the hutch. Neutrons were attenuated with the hydrogen in the snow to decrease the neutron dose rate. At the B points, large solar flares were observed by the ACE satellite [12] (see Fig. 4) and the neutron dose rates decreased in about 10%. This phenomenon is called as the Forbush decrease. In Fig. 3, the dose rates estimated by multiplying the Bonner ball spectrum with the fluence-to-ambient dose equivalent conversion factor given by ICRP 74 are also given for comparison. These results show higher fluctuation than the rem counter results, especially in April and May 2001 in the beginning stage. The average value of dose rates given by the Bonner ball is 6.5 nsv/h and is about 60 % higher than 4.1 nsv/h given by the rem counter, which is also described later. Our neutron dose rates are compared in Fig. 4 with the neutron count rates measured by neutron monitors at several places in the world [13], and the proton flux of energy above 30 MeV observed by the ACE satellite that locates between the earth and the sun. The proton flux occasionally increases due to a solar flare. The big fluctuation of the rem counter counts can be seen over the period of April to June in This fluctuation is partly due to an instability in the beginning stage of the data taking system, but is partly influenced by big solar flares occurred on April 3 and 16, which can also clearly be seen at high latitude locations, South Pole, Oulu and Newark. It is already known that the secondary cosmic-ray intensity in the atmosphere surrounding the earth decreases in some cases during a couple of days after a big solar flare due to a big geomagnetic disturbance, called Forbush decrease (FD), and increase on the contrary in some cases just after a big solar flare which emits a large amount of high energy protons of energy above several hundreds MeV that can reach the earth, called ground-level enhancement (GLE). A big fluctuation on April 16 is the successive effect of FD and GLE. 3.3 Neutron energy spectra Figures 5 and 6 show the neutron energy spectra on three different days, April 25, May 20, June 15 in 2001, and May 12, July 12, September 6 in 2002, obtained for 12 hours by the Bonner ball, as typical examples. Three components can be seen in these two figures, the first highest peak is a cascade peak component, which appears around 100 MeV, and this component fluctuates from day to day in about 30% and has a clear tendency that increases on the day of relatively higher solar activity having higher neutron flux. The second middle peak is an evaporation peak component, which appears around a few MeV. The third lowest peak is a thermal neutron peak component, which appears below about 1eV. The peak values of second and particularly third components keep almost constant values of about 7.5 x 10-4 [n cm -2 sec -1 lethargy -1 ] and about 1.8 x 10-4 [n cm -2 sec -1 lethargy -1 ], respectively. Figure 7 shows the comparison of neutron energy spectra obtained in daytime of Dec. 14 without snowfall and of Dec.16 with snowfall of 20 to 30 cm

4 thickness. The evaporation peak and the cascade peak decrease about 25 % due to snowfall, while on the other hand the thermal peak does not change at all. Figure 8 shows the three energy spectra obtained for long accumulation period of the pulse counting with the NE213 detector. These three data are for three periods of high count rate, low count rate, and average (medium) count rate from November to December The NE213 detector could not obtain the neutrons of energy below about 7 MeV because of the difficulty of gamma-ray discrimination as described before, but the unfolded spectra also give two peaks below 10 MeV corresponding to the evaporation neutrons and around 80 MeV to the cascade neutrons. Similarly in Figs. 5 and 6, the high-energy peak increases with increasing the neutron flux, although the increasing rate is smaller than that in Figs. 5 and 6 due to the longer accumulation of pulse counting with NE Summary The long-term sequential measurements of cosmic ray neutrons have been performed since October 2000 up to March 2003 with three different neutron detectors in an air-conditioned experimental cabin on the ground at the Kawauchi campus of Tohoku University. The average values of neutron flux and dose rate throughout this time period are summarized in Table II. The neutron ambient dose equivalent rate estimated from the Bonner ball is about 6.5 nsv/h and this value is about 60% larger than 4.1 nsv/h given by the rem counter. This big difference mainly comes from the fact that the rem counter has a low sensitivity to neutrons above 20 MeV, which is largely deviated from the fluence-to-ambient dose equivalent conversion factor, while on the other hand, the Bonner Ball can give the neutron spectra up to 1 GeV by using the initial guess spectrum extending to 1 GeV, although the response function decrease for high-energy neutrons. The ambient dose obtained by the Bonner ball becomes 3.5 nsv/h in the neutron energy range below 20 MeV, which is close to the rem counter value. The neutron dose obtained by the rem counter, therefore, gives a large underestimation to that by the Bonner ball. The average values of total neutron flux and neutron flux above 20 MeV obtained by the Bonner ball are about 7.5 x 10-3 [n cm -2 sec -1 ] and about 2.1 x 10-3 [n cm -2 sec -1 ], respectively. The latter value is very close to the average value of 1.8 x 10-3 [n cm -2 sec -1 ] obtained by NE213. This work is in parts financially supported by a Grant-in-Aid for Scientific Research of Ministry of Education, Culture, Science and Sports in Japan. References 1. Uwamino, Y., Nakamura T., Hara, A., Two types of multi-moderator neutron spectrometers: gamma-ray insensitive type and high-efficiency type. Nucl. Instrum. Methods, A 239: , (1985). 2. Nakamura, T., Hara, A., Suzuki, T., Realization of a high sensitivity neutron rem counter. Nucl. Instrum. Methods, A 241: , (1985). 3. Kurosawa, T., Nakao, N., Nakamura, T., Uwamino, Y., Shibata, T., Nakanishi, N., Fukumura, A., Murakami, K., Measurements of secondary neutrons produced from thick targets bombarded by high-energy helium and carbon ions. Nucl. Sci. Eng., 132: 30-57, (1999). 4. McElroy, W.N., Berg, S., Crockett T., Hawkins, R.G., A Computer-Automated Iterative Method for Neutron Flux Spectra Determination by Foil Activation. AFWL-TR-67-41, Air Force Weapons Laboratory (1967). 5. Waters, L. S. (Ed.), MCNPX User's Manual verion 2.4.0, LA-CP , Los Alamos National Laboratory, Los Alamos, New Mexico (2002). 6. Nunomiya, T., Study on the deep penetration of neutrons produced by high-energy accelerator and cosmic rays. Doctor Thesis of Tohoku University, (2003) (in Japanese). 7. Goldhagen, P., Reginatto, M., Kniss, T., Wilson, J.W., Singleterry, R.C., Jones I.W., Van Steveninck, W., Measurement of the energy spectrum of cosmic-ray induced neutrons aboard an ER-2 high-altitude

5 airplane. Nucl. Instrum. Methods, A 476: 42-51, (2002). 8. International Commission on Radiological Protection, Conversion Coefficients for use in Radiological Protection against External Radiation. Publication 74. Annals of the ICRP, 26, No.3-4, Elsevier Science, Oxford (1995). 9. Yonai, S., Quantum Science and Energy Engineering, Tohoku University, personal communication, May Johnson R.H., Wehring, B.W., FORIST: Neutron Spectrum Unfolding Code - Iterative Smoothing Technique. ORNL/RSIC-40, Oak Ridge National Laboratory (1976). 11. Sasaki, M., Nakao, N., Nakamura, T., Shibata T., Fukumura, A., Measurements of the response functions of an NE213 organic liquid scintillator to neutrons up to 800 MeV. Nucl. Instrum. Methods, A 480: , (2002). 12. ACE Science Center, California Institute of Technology, Bartol Research Institute, NSF grant ATM , ~ neutronm. Table I. Atmospheric pressure correction factor, µ. year Rem counter x 10-4 Bonner ball x 10-4 diam µ 2001, , , , Table II. Average values of neutron flux and ambient dose equivalent Neutron flux ( x 10-3 n cm -2 sec -1 ) Neutron ambient dose equivalent ( nsv/h ) Total > 20 MeV Total < 20 MeV Bonner ball Bonner ball NE Rem counter 4.1

6 FIG. 1. Correlation of ambient dose equivalent rates measured with the rem counter and the atmospheric pressure during the period from March 9, 2001 to October 4, FIG. 2. Monthly variation of the ratio of neutron ambient dose rates averaged during daytime (6:00 am to 6:00 pm) and nighttime (6:00 pm to 6:00 am) by the rem counter.

7 A B B B A FIG. 3. Time-sequential data of neutron ambient dose equivalent rates measured by the rem counter and the Bonner ball from April 2001 to March 2003.

8 South Pole ~ neutronm/ Oulu University (Finland) Hermanus (South Africa) NEWWARK(USA) ~ neutronm/welcome.html Roma (Italy) ~ svirco/index.html ACE satellite FIG. 4. Comparison of neutron dose rates and neutron count rates measured on the ground at various positions [13] and high-energy proton flux above 30 MeV measured by the ACE satellite [12].

9 FIG. 5. Comparison of neutron energy spectra on three different days, April 25 (high flux), May 20 (medium flux) and June 15 (low flux) in 2001, measured by the Bonner ball. x 10-4 Neutron flux [n cm -2 sec -1 lethargy -1 ] (> 20MeV) [nsv/h] [n cm -2 sec -1 ] 2002/ 7/12 6:00-18: E / 9/06 6:00-18: E / 5/12 6:00-18: E x Neutron energy [MeV] FIG. 6. Comparison of neutron energy spectra on three different days, July 12 (high flux), September 6 (medium flux) and May 12 (low flux) in 2002, measured by the Bonner ball

10 FIG. 7. Comparison of neutron energy spectra measured on the ground with (Dec. 16, 2001) and without (Dec. 14, 2001) snowfall by the Bonner ball. FIG. 8. Comparison of neutron energy spectra measured with NE213 on the days of high, medium(average) and high count rates in 2001.