The Design and Optimization of a Neutron Time-of-Flight Spectrometer with Double Scintillators for Neutron Diagnostics on EAST

Transcription

1 The Design and Optimization of a Neutron Time-of-Flight Spectrometer with Double Scintillators for Neutron Diagnostics on EAST ZHANG Xing ( ), YUAN Xi ( ), XIE Xufei ( ), FAN Tieshuan ( ), CHEN Jinxiang ( ), LI Xiangqing ( ) School of Physics and State Key Laboratory of Nuclear Physics and Technology, Peking University, Beijing , China Abstract Neutron energy spectrometry diagnosis plays an important role in magnetic confinement fusion. A new neutron time-of-flight (TOF) spectrometer with double scintillators is designed and optimized for the EAST tokamak. A set of optimal parameters is obtained by Monte Carlo simulation, based on the GEANT4 and ROOT codes. The electronic setup of the measurement system is designed. The count rate capability is increased by introducing a flash ADC. The designed spectrometer with high resolution and efficiency is capable of being applied to fusion neutron diagnostics. Applications in mixed-energy and continuous energy neutron fields can also be considered. Keywords: magnetic confinement fusion, neutron diagnostics, TOF spectrometer, GEANT4, flash ADC PACS: Nc, Hs, Mc DOI: / /14/7/24 1 Introduction Magnetic confinement fusion (MCF) is one of the most hopeful solutions to the global energy crisis MeV and 14 MeV neutrons, generated by fusion reactions of D(d, n) 3 He and T(d, n) 4 He, are the only products possible to be measured directly. Neutron diagnostics are useful and reliable for plasma physics research, tokamak operation, blanket design, tritium breeding and radiation protection [1 3]. The neutron flux monitor and neutron energy spectrometry (NES) diagnosis are two alternatives [1,4]. In particular, numerous key plasma parameters can be deduced from the neutron energy spectrum, such as the ion temperature [5], fuel ion isotope ratio [6,7], fraction of fast ion [8], toroidal rotation [9,10], and impurity level [11]. For plasmas with a maxwellian ion distribution, the ion temperature T i has an analytical relation with the full width at half maximum (FWHM) of the neutron energy spectrum, expressed as ( ) 2 FWHM(keV) T i (kev) =, (1) k where k is 82.5 to D plasmas and 177 to D-T plasmas [1,4]. When T i is about 3 kev, FWHM is about 143 kev for D plasmas. In this work we set the requirement on the energy resolution of NES at the 5% level. High energy resolution and detection efficiency are necessary in NES diagnosis. In DT plasmas, the magnetic proton recoil (MPR) neutron spectrometer has made many achievements at JET [12]. It can diagnose the ion temperature [13], plasma rotation [9,10], fuel ion density [6], α-particle knock-on effect [14,15], fuel ion kinetic information during neutral beam injection (NBI) [16] and ion cyclotron resonance heating (ICRH) [17], and so on. In D plasmas, MPR can still work [18,19]. Due to the lower neutron yield, a 2.5 MeV neutron time-of-flight (TOF) spectrometer with much higher efficiency is a better choice. Presently, the TO- FOR [20] at JET is a representative fusion neutron TOF spectrometer. Also, many valuable achievements have been obtained, including the diagnoses on fast fuel ions and their interactions with MHD activity [21,23], impurity level [24], and evaluation of auxiliary heating. Some related models perform well too [25,26]. Another neutron TOF spectrometer has served at JT-60U [27]. The neutron spectrometer using TOF techniques is also under consideration for ITER [28]. Neutron yields of EAST and HL-2A have increased in recent years (EAST will reach n/s [29,30], and both EAST and HL- 2A have already achieved H-mode plasmas [31,32] ). The neutron TOF spectrometer is a prior choice for NES diagnosis at HL-2A [33] and EAST. supported by the State Key Development Program for Basic Research of China (Nos. 2008CB717803, 2009GB107001, and 2007CB209903), the Research Fund for the Doctoral Program of Higher Education of China (No ) and the National Natural Science Foundation of China (No )

2 Fig.1 (a) The global arrangement diagram of TOF spectrometer in vertical view, (b) The schematic diagram of the primary scintillator in detail, (c) The schematic diagram of the secondary scintillator in detail (color online) In this work, the design of a neutron TOF spectrometer with double scintillators for the EAST tokamak is described in section 2, based on the principle of a constant TOF sphere [34]. A Monte Carlo (MC) routine is developed for optimizing the dimensions of the scintillators in section 3. A setup of the measurement system is presented in section 4. A flash ADC is introduced into the electronics diagram of the TOF spectrometer. Specifications of the spectrometer are shown in section 5. In section 6, conclusions are drawn. 2 Geometrical construction of TOF spectrometer The TOF spectrometer consists of two fast plastic scintillators, which are placed on the surface of a constant TOF sphere (shown in Fig. 1). The incident neutron, with energy E n, is an approximately parallel beam after passing through a collimator. A primary scintillator (S1) plays a combined role of a scattering target, adequately thin to make sure single scattering is dominant, and a recoil proton detector, generating the start timing signal t 1. A secondary scintillator (S2) covers a wide range of scattering angles θ ± θ by a large area. The stop timing signal t 2 is generated when the scattered neutron is detected in S2. Just considering the neutron scattered on hydrogen in S1 with angle θ, its energy is E n(θ) = E n cos2θ and flight path is L(θ) = Dcosθ. In non-relativistic kinematics, the corresponding flight time t can be expressed as m n L t = t 2 t 1 = 2 (θ) m n D 2E n(θ) = 2. (2) 2E n So the incident neutron energy E n can be deduced by E n = 1 ( ) 2 D 2 m n. (3) t From Eq. (2), it is obvious that the flight time t is independent of θ when S2 lies on the surface of the sphere. High efficiency can be obtained by expanding the area of S2 without deterioration of the energy resolution of the spectrometer. The design of the spectrometer requires a compromise between high detection efficiency and good time resolution, due to the finite dimensions of the two scintillators and incident neutron beam, which will cause a deviation of the scattered neutron s energy E n(θ) and the flight path L(θ). Extensive MC simulation is useful in designing and optimizing the detector geometry. Geometrical dimensions in the designed model are listed in Table 1. Table 1. Geometrical dimensions in designed model Symbol Description Value D Diameter of TOF sphere 2 m θ Central scattering angle 30 o Φ Diameter of collimator 6 cm d1 Thickness of S1 2.5 cm ϕ Tilted angle of S1 55 o d2 Thickness of S2 2 cm a Width of S2 32 cm ratio Ratio of b/a in S2 1.5 area Area of bottom surface in S cm 2 3 Optimization by Monte Carlo simulation 3.1 Description of simulation routine A MC routine is developed, based on the GEANT4.9.3 [35] and ROOT5.26 [36] codes. The module of transport and nuclear reaction utilizes the builtin GEANT4 physics list QGSP BERT HP and the standard electromagnetic processes. The cross section data of nuclear reactions are taken from the ENDF/B- VI library. The semi-empirical expressions of light re- 676

3 ZHANG Xing et al.: The Design and Optimization of a Neutron TOF Spectrometer with Double Scintillators sponse functions in Ref. [37] are used for p and α. The light responses of the scintillators for heavy ions are neglected because of their low light yields and low energy. Threshold settings, being consistent with the experimental setup, are included. In the following subsections, the routine firstly scans the effect of the spectrometer s dimensions on the time resolution and detection efficiency, including θ, Φ, d1, ϕ, d2, area (a b) and ratio (b/a) in Fig. 1. Incident neutron energy is set at 2.45 MeV. Firstly, the TOF spectrum is formed by recording the time elapsed t, as seen in Eq. (2). The pulse height spectra of S1 and S2 are also recorded. Secondly, the routine creates an energy-tof matrix in a range of 1 7 MeV. Last, the shape of the collimator is designed and simulated using the MCNP code [38]. 3.2 the energy of single scattered neutrons in S1 concentrates around 1.84 MeV. In Fig. 3(b), it is seen that the multiple scattering in S1 results in a rise in the low energy range. In the simulations of subsections , 15 kevee (0.27 MeV to proton) and 200 kevee (1.03 MeV to proton) are set as the lower and upper thresholds for S1, and 15 kevee is the lower threshold for S2. Invalid events are rejected effectively. Typical spectra Fig. 2 shows the simulated TOF spectra in the designed dimensions of the spectrometer, in which the time bin width is 0.1 ns. It is seen that the fraction of single scattering (scattering only once and on H in the S1 scintillator) is dominant in a time window of ns, as illustrated in the inset figure. The fraction of the long time tail (low energy tail in the energy spectrum) mainly results from multiple scattering in S1, while the short time tail (high energy tail) comes from the contributions of double scattering on C and H (about 54%) and small angle double scattering on H (about 41%) in S1. Fig.3 Pulse height spectra of S1 (a) and S2 (b). The black histograms are contributed from all events in the TOF spectrum in Fig. 2. The red shadows represent the part of events in the time window of ns. The blue bold lines are the lower and upper thresholds (color online) Fig.2 Typical TOF spectrum from simulation. The inset distinguishes the fraction of single scattering events on hydrogen (red) from total events in the TOF spectrum (black) (color online) The pulse height spectra, corresponding to the simulated TOF spectra, are presented in Fig. 3. The x-axis represents the equivalent electron energy (Eee, unit: kevee ), proportional to the pulse amplitude and channel address in the multi-channel analyzer (MCA). In Fig. 3(a), the red shadow represents the contribution of valid events (in a time window of ns in Fig. 2), which is distributed in a narrow width, because Finally, based on the flight time spectra after pulse height selection, the detection efficiency is calculated by the counts in the time window of ns over the number of incident neutrons. The time resolution is reflected by the FWHM of the time peak in the flight time spectra. According to Eq. (3), the energy resolution is double the time resolution. 3.3 Response to central scattering angle θ The central scattering angle θ is scanned from 10o to 40o. A part of S2 overlaps the collimator if θ < 11o. Neutrons, passing through S1 directly but detected in S2, will increase accidental coincidence events. In another aspect, the small scattering angle means a lower energy of recoil proton in S1. The nonlinearity of the light response function will exacerbate the decrease in 677

4 the pulse height. Consequently, it causes a rapid decrease in efficiency with a low threshold (15 kev ee ) when θ < 25 o, as shown in Fig. 4. The line of FWHM indicates that a smaller θ gives a higher resolution, while the efficiency behavior is opposite. and the ratio n H /n C is 1.1. So the macroscopic cross section of S1 to 2.45 MeV neutron is given as Σ = n H σ H + n C σ C = 0.21 cm 1, (4) where σ H and σ C are the cross sections of neutron interactions with H and C. The mean free path λ of 2.45 MeV neutron is 4.76 cm. In our design, the tilted angle ϕ is set at 55 o, the thickness of S1 d1 is 2.5 cm, and d1 eff is 4.36 cm, close to λ. Fig.4 Response of resolution and efficiency to central scattering angle θ. Black solid line is the FWHM of the time peak in TOF spectrum, blue dash-dotted line shows the efficiency, red dashed line shows the efficiency with lower threshold, and dash-dotted vertical line indicates the design value Finally, the central scattering angle θ is set at 30 o in the design, and optimal FWHM (<2.3 ns, energy resolution <5%) and efficiency (>0.1%) are obtained. 3.4 Response to S1 s dimensions, d1 and ϕ In the simulations of the following three subsections, the events, cumulated in time spectra, are selected by the lower and upper thresholds in S1 and the lower threshold in S2. Among the dimensions of S1, the thickness and tilted angle are variables. The effective area of S1 s surface is determined by the incident beam. The tilted angle ϕ of S1 is scanned from 0 o to 80 o, as shown in Fig. 5(a). It is obvious that the tilted S1 is better than a non-tilted one. In the direction of the incident neutron beam, the effective thickness of S1 (d1 eff = d1/cosϕ) increases with ϕ, which increases the probability of neutron elastic scattering. But in the flight direction of scattered neutrons, the residual thickness of S1 decreases much, which decreases the probability of multiple scattering. As a result, a higher efficiency and narrower FWHM are obtained by increasing ϕ in an appropriate range. Besides, when S1 tends to be parallel to S2, the deviation of L(θ) is much diminished, and FWHM shrinks much. The tilted angle ϕ is the only one variable that can improve resolution and efficiency simultaneously. The thickness of S1 d1 is scanned from 5 mm to 40 mm, as shown in Fig. 5(b). The FWHM increases with an increasing d1, because the deviation of L(θ) increases. The efficiency is increasing and approaching saturation as a result of the saturation of the single scattering probability. The density of S1 is 1.05 g/cm 3, Fig.5 Response of resolution and efficiency to S1 s tilted angle and thickness. Black solid line is FWHM in TOF spectrum, red dotted line is efficiency, and dash-dotted vertical line is the designed value (color online) 3.5 Response to S2 s dimensions, d2, ratio and area In Fig. 6(a), the thickness of S2 d2 is scanned from 5 mm to 40 mm. As a result of the increase in the deviation of L(θ) with increasing d2, the efficiency becomes higher, while FWHM becomes broader. The shape of S2 is inspected by scanning the ratio (b/a) and the area (a b) of the bottom surface, shown in Fig. 6(b) and (c). The ratio varies from 1/4 to 4, when the area is kept at 1536 cm 2. It shows that the efficiency is not sensitive to the ratio. FWHM is the best when S2 s surface is a square (ratio=1). However, a square surface of S2 means the longest length of light guides are needed at two sides, comparable to the width of S2. Propagation and attenuation of light will deteriorate the time resolution distinctly. So a ratio exceeding 1 is selected on the condition that FWHM is less than 2.3 ns. Due to the difficulty in manufacturing a curved S2, a flat one with Φ is adopted. The area(a b) 678

5 ZHANG Xing et al.: The Design and Optimization of a Neutron TOF Spectrometer with Double Scintillators is scanned from 600 cm2 (20 cm 30 cm) to 2400 cm2 (40 cm 60 cm). The efficiency is proportional to area while independent of ratio. Since the non-central region of S2 deviates from the TOF sphere, the FWHM is rapidly broadened when area increases. An intermediate area is feasible for high efficiency and narrow FWHM. efficiency (efficiency area) increases rapidly, which is beneficial for a higher count rate. But the bigger Φ is, the worse the parallelism of the incident beam is. The diameter Φ of the collimator is finally chosen as 6 cm. Fig.7 Response of resolution and efficiency to collimator s area (corresponding to its diameter). Black solid line is FWHM in TOF spectrum, red dotted line is efficiency, blue dash-dotted line is efficiency multiplying area, and dashdotted vertical line is the designed value (color online) 3.7 Fig.6 Response of resolution and efficiency to S2 s thickness, ratio and area. Black solid line is FWHM in TOF spectrum, red dotted line is efficiency, and dash-dotted vertical line is the designed value (color online) Energy-TOF matrix in the neutron energy range of 1 7 MeV Scanning of time-spectrum response in a wide energy range of 1 7 MeV is shown in Fig. 8, in which the time bin width is 0.1 ns and the energy bin width is 0.05 MeV. Lower thresholds of S1 and S2 are set as 15 kevee. The diminished efficiency below 1.1 MeV is due to a lower threshold. Good resolution is kept in the whole range. So this TOF spectrometer can work in a wide dynamic range, up to 14 MeV with possible applications to D-T reactions. It can also measure the continuous energy spectrum of an isotope neutron source (such as an Am/Be source and a Pu/C source). In other applications, when focusing on the fine structure in a local energy range, this matrix can be used to fit the measured time spectrum. Precision of analysis can be improved by shortening the energy bin. In our design, the thickness of S2 d2 is 2.0 cm, the ratio is 1.5, and the area is 1536 cm2 (32 cm 48 cm). 3.6 Response to the collimator s diameter, Φ The diameter Φ of the collimator is scanned from 1 cm to 10 cm, corresponding to area from 0.79 cm2 to 79 cm2. In Fig. 7, it is seen that the FWHM increases with Φ, due to the deviation of En0 (θ) and L(θ) when the off-axis incident neutrons are not scattered at the center point of S1. In another aspect, when Φ is increasing, the efficiency decreases slightly, while the area Fig.8 Energy-TOF matrix in the neutron energy range of 1 7 MeV (color online) 679

6 3.8 Design and simulation of collimator Good shielding and collimation for incident neutron in parallel beam are necessary. Following geometrical optics, 4 types of collimator with a 70 cm length and a 6 cm diameter are designed for the EAST tokamak neutron source, as shown in Fig. 9(a). The model of the source is a port 60 cm in diameter, 10 m away from the tokamak. (A) is a cylindrical hole, (B) is a conical hole, both (C) and (D) are double-truncated conical collimators but with different throat locations. Simulations are accomplished by the MCNP code with the ENDF/B-VII library. 4 Experimental setup of measurement system Fig. 10 presents a block diagram of the electronics setup of the designed TOF spectrometer. The setup consists of two HDN-S2 scintillation detectors, a flash ADC and a digital processing module. The scintillators are connected to three R1828 photomultiplier tubes (PMTs) via light guides. The flash ADC (Agilent U1065A, 4 signal channels with a sampling rate up to 2 G/s and 1 GHz bandwidth) is used to record the anode signals from three PMTs directly and send them to a computer. Fig.10 Block diagram of the electronics setup of the designed TOF spectrometer (color online) Fig.9 (a) Four configurations, (A), (B), (C) and (D), used in collimator simulation. Red bold lines on the left represent the neutron source. The dash-dotted lines in collimators represent the symmetry axis. (b) Distributions of the collimated beam flux from simulation. X-axis is the off-axis distance of a point detector (color online) The uniformity and divergence angle of the neutron beam after collimation are inspected. Fig. 9(b) indicates that the conical hole is the best choice, in which the beam is the most uniform within the radius and the edge of the beam is the steepest. In addition, the beam divergence angle of type B is less than 2 degrees. It should be noted that different types of neutron source need different collimator configurations. The neutron source on EAST is far away from the spectrometer and is distributed over a large area. A long transmission path is beneficial for beam collimation and a conical collimator will reduce neutrons scattered on the wall remarkably. It is predictable that a double-truncated conical collimator (type C) will be better for a point source, such as a target on an accelerator [39]. 680 The digital processing module has two branches, of which one is for timing and the other is for energy selection. In the timing branch, a digital constant-fraction discriminator (DCFD) programme generates a timing sequence: t1, t2 and t3. To eliminate the timing variance, induced from the light propagation time in S2, the mean time tm of t1 and t2 is calculated. Consequently, a pair of t3 and tm in a interval of 200 ns induces a coincidence event. Considering the count rate of S1 is much higher than S2, the coincidence is triggered by the mean time tm. Energy selection is implemented by the thresholds described in section 3.2. Anode signals are integrated by digital verison of charge to digital conversion (DQDC) programme, generating a charge sequence: Q1, Q2 and Q3. Qs is the sum of Q1 and Q2, representing the pulse height of S2. If Qs and Q3 satisfy the thresholds, the value of tm minus t3 will generate a count in the TOF spectrum. The main advantage of the new system is its higher count rate capability, compared to the achievable using a standard NIM amplifier and time-amplitude conversion (TAC) units. By the flash ADC, the theoretical count rate capability of the S1 detector can reach 10 MHz. Considering the fact that the efficiency of S1 is about 28% and the global efficiency of the designed system is 0.11%, the global count rate capability can be increased to 40 khz. In other respects, the new system is more flexible, the diagram is very concise, and the signals can be processed offline.

7 ZHANG Xing et al.: The Design and Optimization of a Neutron TOF Spectrometer with Double Scintillators 5 Specifications of the TOF spectrometer Based on the results mentioned above, the specifications of the designed neutron TOF spectrometer are listed in Table 2. The energy resolution of 4.6% and the area efficiency of cm 2 meet the demand presented at the beginning of this work. Table 3 shows a comparison among our neutron TOF spectrometer, TOFOR [20,40,41], and the one at JT-60U [27,42]. Compared to the spectrometer at JT-60U, the count rate capability is the main improvement of this new system. The TOFOR at JET has a very high area efficiency of 0.12 cm 2 and a good energy resolution of 5.8% for 2.5 MeV neutrons, as shown in the third column of Table 3. The TOFOR operates under a very high count rate capability of 500 khz so that advanced neutron emission spectroscopy diagnosis can be performed. Compared to TOFOR, the presented spectrometer has some space for upgrading, such as improving the catching factor [43] by increasing the number of S2 scintillators. As a result of the lower T i and limited auxiliary heating power, the low neutron yields and narrow neutron spectral broadening of EAST make time-resolved spectra measurements unattainable for the time being, while a good energy resolution is more in need for this spectrometer. We have made a compromise of the tilted S1 and the catching factor in the present design. When T i and auxiliary heating power are much higher, the time-resolved NES diagnosis is feasible. We can take the present one as a prototype, and then easily adopt the non-tilted S1 and an annular array of S2 on the sphere surface. The S1 can also be divided into several layers for a higher count rate capability. The flash ADCs will be still employed for a high count rate capability. 6 Conclusions The neutron TOF spectrometer with double scintillators for EAST has been designed and is under construction. A Monte Carlo simulation routine based on the GEANT4 and ROOT codes has been developed. Responses of the time resolution and detection efficiency to 7 dimensions in the designed spectrometer have been scanned and a set of optimal parameters adopted. An energy-tof matrix has been created and will be extended to a higher energy. The design and simulation of a collimator at the EAST tokamak have also been accomplished. On the whole, the energy resolution of this spectrometer reaches 4.6%, the area efficiency is cm 2, and the count rate capability is up to 40 khz. The spectrometer can be applied in NES diagnosis at EAST. Detailed calibrations with gamma rays and quasi-monoenergetic neutrons from an accelerator as needed have been scheduled. Table 2. The specifications of TOF spectrometer in design Symbol Description Value t Flight time of scattered neutron 92.3 ns t FWHM of time peak in spectrum 2.1 ns η Energy resolution 4.6% ε Efficiency 0.11% ε area Area Efficiency* cm 2 C max Maximum of count rate 40 khz F max Maximum of incident neutron flux** /(cm 2 s) Area efficiency = efficiency area. Assuming that the incident neutron and Gamma ray have the same intensity Table 3. Comparison with similar spectrometers presently in use in fusion research JT-60U (DC-TOF) JET (TOFOR) (Present design) Number of start scintillator Number of stop scintillator Area efficiency cm cm cm 2 Energy resolution 4.3% 5.8% 4.6% Maximum count rate 6 khz 500 khz 40 khz 681

8 Applications of the spectrometer in other energy ranges can also be considered. In a non-2.45 MeV neutron field or mixed energy neutron field, such as from an accelerator, a D-T neutron generator, a D-T fusion reactor and so on, it can work with different thresholds. In a continuous energy field, e.g., from an isotope neutron source, application is also possible. References 1 Wolle B. 1999, Physics Reports, 312: 1 2 Zheng S, Chen M, Li J, et al. 2007, Nucl. Fusion, 47: Fischer U, Pereslavtsev P, Hermsmeyer S. 2005, Fusion Eng. Des., 75-79: Jarvis O N. 1994, Plasma Phys. Control. Fusion, 36: Jarvis O N. 2002, Nucl. Instr. Meth. A, 476: Kallne J, Gorini G, Ballabio L. 1997, Rev. Sci. Instrum., 68: Okada K, Kondo K, Sato S, et al. 2006, Rev. Sci. Instrum., 77: 10E726 8 Elevant T, Aronsson D, Belle P V, et al. 1991, Nucl. Instr. Meth. A, 306: Gorini G, Kallne J, Ballabio L. 1997, Rev. Sci. Instrum., 68: Giacomelli L, Hjalmarsson A, Sjostrand H, et al. 2005, Nucl. Fusion, 45: Elevant T, Belle P V, Jarvis O N, et al. 1995, Nucl. Instr. Meth. A, 364: Ericsson G, Ballabio L, Conroy S, et al. 2001, Rev. Sci. Instrum., 72: Tardocchi M, Gorini G. 2004, Rev. Sci. Instrum., 75: Kallne J, Ballabio L, Conroy S, et al. 2000, Phys. Rev. Lett., 85: Ballabio L, Gorini G, Kallne J. 1997, Phys. Rev. E, 55: Henriksson H, Conroy S, Ericsson G, et al. 2005, Plasma Phys. Control. Fusion, 47: Tardocchi M, Conroy S, Ericsson G, et al. 2002, Nucl. Fusion, 42: Sjostrand H, Gorini G, Conroy S, et al. 2008, J. Phys. D: Appl. Phys., 41: 1 19 Sjostrand H, Giacomelli L, Andersson Sunden E, et al. 2006, Rev. Sci. Instrum., 77: 10E Gatu Johnson M, Giacomelli L, Hjalmarsson A, et al. 2008, Nucl. Instr. Meth. A, 591: Hellesen C, Gatu Johnson M, Andersson Sunden E, et al. 2010, Nucl. Fusion, 50: Hellesen C, Gatu Johnson M, Andersson Sunden E, et al. 2010, Nucl. Fusion, 50: Giacomelli L, Hjalmarsson A, Kallne J, et al. 2008, Rev. Sci. Instrum., 79: 10E Gatu Johnson M, Hellesen C, Andersson Sunden E, et al. 2010, Nucl. Fusion, 50: Hellesen C, Albergante M, Andersson Sunden E, et al. 2010, Plasma Phys. Control. Fusion, 52: Gatu Johnson M, Conroy S, Cecconello M, et al. 2010, Plasma Phys. Control. Fusion, 52: Shibata Y, Iguchi T, Hoek M, et al. 2001, Rev. Sci. Instrum., 72: Krasilnikov A V, Sasao M, Kaschuck Yu A, et al. 2005, Nucl. Fusion, 45: Wu Y, FDS Team. 2007, Nucl. Fusion, 47: Chen Yixue, Wu Yican. 2005, Plasma Sci. Technol., 7: Duan X R, Dong J Q, Yan L W, et al. 2010, Nucl. Fusion 50: Xu G S, Wan B N, Li J G, et al. 2011, Nucl. Fusion 51: Yang J W, Li W Z, Zhang W. 2003, Sci. Tech. Eng., 3: 360 (in Chinese) 34 Elevant T, Olsseon M, 1991, Neutron time-of-flight counters and spectrometers for diagnostics of burning fuxion plasmas, Report TRITA-PFU-91-05, Royal Institute of Technology, Stockholm, Sweden 35 Agostinelli S, Allison J, Amako K, et al. 2003, Nucl. Instr. Meth. A, 506: Brun R, Rademakers F, 1997, Nucl. Instr. Meth. A, 389: Cecil R A, Anderson B D, Madey R. 1979, Nucl. Instr. Meth., 161: X-5 Monte Carlo Team, 2003, MCNP-A general Monte Carlo N-particle transport code, Version 5, Report LA- UR , Los Alamos National Laboratory, New Mexico, United States of America 39 Glasgow D W, Velkley D E, Brandenberger J D, et al. 1974, Nucl. Instr. Meth., 114: Hjalmarsson A, Conroy S, Ericsson G, et al. 2003, Rev. Sci. Instrum., 74: Gatu Johnson M, Giacomelli L, Hjalmarsson A, et al. 2006, Rev. Sci. Instrum., 77: 10E Hoek M, Nishitani T, Takahashi H, et al. 1999, Fusion Eng. Des., 45: Gorini G, Kallne J. 1992, Rev. Sci. Instrum., 63: 4548 (Manuscript received17 May 2011) (Manuscript accepted 23 December 2011) address of corresponding author FAN Tieshuan: tsfan@pku.edu.cn 682