Fusion Division & Technologies, Neutron & Gamma Diagnostics

Neutron Emission Measurements on FTU Tokamak


1. Introduction
2. Time-resolved neutron yield
3. Time-integrated neutron yield
4. Fast fluctuations of neutron emission
5. Space/time-resolved neutron emission
6. Neutron spectroscopy

1. Introduction

FTU is a high magnetic field (Bt=8 T) compact tokamak: the major and minor radius of the torus are respectively 93.5 cm and 30 cm, the maximum plasma current is 1.6 MA, the electron density is >1020m-3 and the electron temperature is ~1-3 keV. Deuterium gas is used and the typical duration of a discharge is ~1.5 s. Energy confinement times of the order of 60 ms have been achieved.

Neutrons with 2.45 MeV energy are generated in the D(d,n)3He branch of the D-D fusion reactions. The tritons resulting from the other branch, D(d,p)t, may undergo a D(t,n)4He reaction, producing some 14.1 MeV neutrons (in FTU the ratio between the 2.45 MeV and 14.1 MeV neutrons is ~0.2%). Due to the effect of the structures surrounding the plasma (vacuum vessel, poloidal coils, toroidal magnet, etc.), the energy spectrum of the escaping neutrons is very broad, including thermal energies. The spectrum outside the FTU tokamak is not measured, but can be calculated by means of neutron transport codes such as MCNP.

The present maximum neutron emission rate with ohmic heating of the plasma is ~3x1012 ns-1, but higher values (up to 1014 ns-1) are expected with the full auxiliary heating power (~2.5 MW of lower hybrid radio frequency coupled to the plasma).

The neutron detection systems presently operating on FTU are of three types: 1) proportional counters (BF3 and fission chambers) for time-resolved absolute neutron yield measurements; 2) foil activation technique for time-integrated absolute neutron yield measurements; 3) scintillators for time-resolved neutron emission measurements; 4) collimated scintillator for space/time-resolved neutron emission measurements.

2. Time-resolved neutron yield

The time resolved neutron emission is measured by six BF3 chambers and three 235U fission chambers (FC), operated in pulse mode, with time resolution respectively of 5ms and 1ms. The BF3 chambers are located on the top of the tokamak, while the fission chambers lie on the equatorial plane. The BF3 chambers (Alnor Oy mod. 2002B) are of two types: low sensitivity (surrounded by a polyethylene and boron plastic shield) and high sensitivity (with the boron plastic shield removed); together they provide time-resolved neutron yield measurements for a dynamic range of 3 orders of magnitude (109-1012 ns-1). The fission chambers (Centronic mod. FC144/300, with 300 microg/cm2 of 235U) are embedded in a moderator-shielding assembly made of lead (80% in weight) and polyethylene: their dynamic range is 1011-1014 ns-1. A complete description of this system is given in ref. [1].

These detectors have no neutron energy resolution; they need an in-situ calibration, performed periodically, which allows to establish a relationship between the counts and the actual neutrons produced by the plasma. A neutron source (252Cf), with energy spectrum similar to that of the 2.45 MeV fusion neutrons, placed inside the vacuum chamber at different positions in order to simulate the extended plasma, is used for calibration. Several factors contribute to the error on the calibration constant: calibration statistical uncertainty (~3%, BF3; ~6%, FC), energy spectrum effect (~10%), limited number of calibration positions (~10%) and drifts of electronics (~10%, BF3). The total error amounts to ~15% (FC) and ~18% (BF3), to which the statistical error on the counts must be added: typical count rates are 4*104 cps at 1012 ns-1 (low sensitivity BF3 chambers) and 4*103 cps at 1012 ns-1 (FC).

The gamma-ray sensitivity of these detectors is quite low: for the BF3 the manufacturers quote 1.8 cps as a contribution from a gamma radiation background of 200 R/h. Negligible effects on the fission chambers are reported to be produced by a gamma-ray background up to 105 R/h.

The neutron yield measurement represents a diagnostic tool for the determination of the plasma ion temperature; if the plasma is in thermal equilibrium, the reaction rate is described by a maxwellian distribution function f(v) and the neutron yield is proportional to ni2Tialfa, where alfa is in the range 6-4 for 1 keV < Ti <3 keV. From the above relation the ion temperature can be calculated, assuming a parabolic squared radial ion temperature profile and using the measured radial electron density profile from a DCN interferometer. In FIGURE 1 is shown the time evolution of the neutron emission in a 1.1 MA discharge (#5546) obtained as an average of the signals of the six BF3 chambers and the three FC (saturated chambers are excluded); the FC signal shows poor statistics due to low count rate. In FIGURE 2 is plotted the derived central ion temperature, compared with the ion temperature measured by two other independent methods (x-ray crystal spectroscopy and charge exchange).

FIGURE 1: Time resolved neutron yield measurement performed with BF3 chambers and fission chambers in a 1 MA discharge

FIGURE 2: Peak ion temperature derived by neutron measurements and other diagnostics

3. Time-integrated neutron yield

A further diagnostic, the neutron activation system, is mounted on FTU for the measurement of the time-integrated total neutron yield [2]. Initially installed in 1993, it is now fully operational and provides an integrated measurement of the absolute neutron yield which is independent of the BF3 or fission chamber monitors and therefore is used as an important cross check reference. The system is based on the foil activation technique: indium foils (the ~1 MeV threshold reaction 115In(n,n')115mIn is used) are irradiated in the FTU neutron field, close to the vacuum vessel; after every discharge, they are moved to a gamma-ray counting station, where the induced radioactivity is recorded by means of HPGe detectors. The measured local neutron fluence at the irradiation position is related to the total neutron yield through a calibration factor obtained by means of neutron transport calculations (MCNP). The accuracy of the method is presently of the order of ~17%. During the 1994 campaign, the system has provided measurements in agreement within 30% with the BF3 data: the neutron yield was in the range: 1011-1012 neutrons/shot (seeFIGURE 3 ).

FIGURE 3: Comparison between BF3 and activation results

Two HPGe detectors are currently in operation, a coaxial and a well-type HPGe. The installation of the well-type detector has been recently completed [3]; this new detector is particularly suited to the FTU low level activity measurements due to its high geometrical efficiency. The detector has 200 cm3 active volume, 21 mm diameter and 56 mm well depth; the electronic set-up and data acquisition make use of a fast amplifier and a MCA CAMAC system, allowing high total counting rates up to 104 s-1. A particular effort has been devoted to the calibration of the measuring system. In order to assure the counting geometry reproducibility in the efficiency calibration, a dedicated multi-gamma calibration source (59 -1332 KeV energy range) was prepared with the same shape, volume and filling material of the actual samples used in the activation measurements: it consists of 10 spiked indium foils (1 mm thick, 18 mm diameter) stacked together reproducing the FTU indium samples. The full-energy-peak efficiency of the well-type detector at the 115mIn 336.2 line, which is used for the determination of the neutron fluence, is equal to 0.213 with an overall uncertainty of 2% (1 sigma). A further multi-gamma calibration source with the same geometry has been prepared in a gel acqueous solution. A comparison of the full-energy-peak efficiency for the two HPGe detectors measured with the multi-gamma mixed gel source is reported in FIGURE 4 showing a factor gain of 6 for the well-type with respect to the coaxial germanium detector. Moreover, the self attenuation and volume source effects for the FTU indium samples are not negligible are taken experimentally into account with this technique.

FIGURE 4: Full-energy-peak efficiency for the two HPGe detectors measured with the multi-gamma mixed gel source

4. Fast fluctuations of neutron emission

The fast variations of the neutron emission from the plasma are due either to intrinsic magnetic instabilities of the plasma column, which can produce sawtooth-like periodic signals, or to some external means such as additional heating or pellet fuelling. In FTU the time scales of the fluctuations are of the order of ~100 microseconds or less [4]. The need of such high time resolution requires the use of scintillation detectors; the complete set of detectors, which are located on the equatorial plane about 160 cm far from the plasma center, is the following:

a) a NE213 liquid organic proton recoil scintillator (size: 12.7 cm diameter 12.7 cm thickness) coupled through a light-guide to a 12-stage RCA 8575 photomultiplier: the NE213 has a ~10% efficiency for fast neutrons and gamma-rays;

b) a NE422 plastic ZnS(6Li) scintillator (size: 12.7 cm diameter 0.63 cm thickness) coupled to a 14-stage PHILIPS XP2041 photomultiplier; this scintillator is sensitive to thermal neutrons through the 6Li(n,a)t nuclear reaction (55% efficiency at 0.025 eV neutron energy). For optimum neutron moderation, a polyethylene cylinder, 6 cm thick, is placed in front of the detector; the NE422 has very low [gamma]-ray sensitivity (this detector has been temporarily removed from the machine).

c) a NaI crystal scintillator (size: 5.08 cm diameter and 5.08 cm thickness) coupled to a RCA 8575 photomultiplier: this scintillator is highly sensitive to gamma-rays. Presently, it works as a gamma-ray spectrometer for runaway electron studies.

The NE213 scintillator is used to monitor the fast fluctuations of the 2.45 MeV neutron emission: as this scintillator is sensitive both to neutron and gamma-rays, its output signal is an efficiency-weighted sum of the neutron (n) and the gamma-ray emission. Nevertheless, as the gamma-rays are mainly produced by neutrons interacting with the tokamak structures, the NE213 signal can be considered as representing the true neutron emission. In those cases where the gamma-ray contamination in the NE213 n+gamma signal is too high (e.g. discharges with run-away electrons), the comparison with the n signal from the NE422 and the gamma signal from the NaI allows easy data rejection. The NE213 detector has been recently cross calibrated with the BF3 and fission chambers detectors.

The strong neutron and gamma radiation field at the scintillators' position produced very high count rates in pulse mode operation (>106 cps at 1011 ns-1 in the NE213), causing saturation effects due to pile-up. In order to avoid this problem, all the detectors are now operated in current mode: the time resolution is ~100 microseconds and the analog-to-digital coverter (ADC) sampling rate is 20 kHz. A load resistor converts the anode current into a voltage signal, which is fed, through a RC integrating circuit and an impedence adapter, to the ADC. The photomultipliers are shielded against magnetic fields (up to ~1 T) by means of three concentric metallic cylinders of decreasing magnetic permeability from the outer to the inner shield.

Neutron sawteeth oscillations are present in the majority of FTU discharges (an example is shown in FIGURE 5 with an enlargement of a sawtooth crash: discharge 6166, Ip=1.1 MA). At the sawtooth crash, after an initial fast drop (~500 microseconds or less) the neutron emission reaches its minimum in a total time of 2-3 ms.

FIGURE 5: Neutron sawteeth oscillation in a 1.1 MA discharge

The sawtooth repetition time is <10 ms. The drop is caused by an outward shift of the hot reacting ions. Assuming that the contribution of the density to the neutron drop is negligible, the corresponding drop in ion temperature is given by: DeltaTi/Ti~ DeltaIn/In; as DeltaIn/In is typically < 0.20, DeltaTi/Ti results < 0.05. To improve the signal-to-noise ratio, a phase-locked average of several sawteeth for each discharge is performed during a time interval in which the amplitude and the repetition time of the sawteeth are approximately constant. This "average" overall drop in neutron emission (DeltaIn/In) is found to be increasing with the line-averaged central electron density (FIGURE 6 ), as at higher density the ion-electron coupling increases and DeltaTi/Ti tends to be close to the typical DeltaTe/Te (of the order of 0.15).

FIGURE 6: Neutron drop at sawtooth crash versus line-integrated central electron density (each point is an average over many sawteeth)

5. Space/time-resolved neutron emission

The space/time-resolved neutron emission is measured by the neutron multichannel collimator which has been installed in 1995 (FIGURE 7 ). The multicollimator can also work as a gamma-ray camera for LH power deposition profile studies. The aim of the device is to measure the neutron emission as a function of space and time in a poloidal section of FTU: the measurement can provide information on the ion temperature and on the energy and particle transport in the plasma [5]. The design of the multicollimator [6] was optimized by means of MCNP code calculations for neutron and gamma-ray transport. The system is located under port 3 of FTU and views the plasma through six vertical collimated channels.

FIGURE 7: Neutron multicollimator

As the signal to be measured is produced by D-D 2.45 MeV neutrons, a good neutron shielding can be obtained with hydrogenous materials thanks to the high cross section of hydrogen below 2 MeV. A combination of machined polyethylene (65% in weight) and lithium carbonate powder (35%) was chosen as shielding material: lithium (which efficiently absorbs thermal neutrons) is necessary in order to suppress the production of gamma-rays from thermal neutron capture reactions in the shield structure. The thickness of the shield is 120 cm, including ~26 cm of lead for gamma-ray shielding, for a total weight of ~9 tons. The shielding block attenuates the flux of non-collimated neutrons at the detectrs to a fraction ~1/100 of collimated neutrons.

The neutron detectors are six NE213 organic liquid scintillators operated in pulse mode with 2 ms time resolution (for gamma-only detection NaI detectors scintillators are substituted into the detector box); the scintillators are coupled to EMI 9815B photomultipliers and to pulse shape discriminator units (PSD), which allow discrimination between neutron and gamma-ray events (see [7] for a detailed description). Two output signals are available: neutron and gamma channels. The detectors are calibrated in neutron energy and an energy threshold is set at ~1.8 MeV in order to reject the backscattered neutrons so that possibly only virgin 2.45 MeV neutrons are detected. The two output data are corrected for dead time.

The main specifications of the system are the following: 1) spatial resolution of 10% of plasma minor radius; 2) time resolution (for present FTU neutron yields) of ~1 s: the maximum counting rate in the central channel is ~100 Hz at 1012 neutrons/s; 3) spatial intensity variation of a factor 100 with respect to the central emissivity.

The debugging of the system has been carried out during DD operation in 1997. The typical measured neutron count rate is 200 c/s at a total neutron rate of 2 x 10 12 n/s in the central channel, and the intefering breakthrough of gamma events in the neutron channels is less than 0.2%. Measurements of the background signal have been performed, showing the necessity of implementing additional shielding, in particular for channels 1 and 2. First neutron emission profiles have been obtained from ohmic (FIGURE 8 ) and pellet fuelled discharges (FIGURE 9 ) shows the neutron brightness profile, obtained averaging data from 5 consecutive ohmic discharges with similar parameters (ne=1.2 x 1020 m-3, Ip=700 kA, Bt=6 T): the solid line is a fit on the experimental data assuming a neutron emission profile of the form: S(r/a)= S0 [1-(r/a)2]an, where an=8. From this measurement together with that of the total neutron yield, as measured by the BF3 monitors, the ion temperature profile has been also derived [8].

Information on recent operation (1997) of the system can be found here.

FIGURE 8: Neutron emission profile in ohmic discharge

FIGURE 9: Neutron emission profiles after pellet injection

6. Neutron spectroscopy

A portable neutron spectrometer has been set up using a 2" x 2" NE213 scintillator coupled to a photomultiplier. The D-D neutron spectrum on FTU tokamak has been measured for the first time using the central channel of the 6-channel FTU neutron collimator. The spectrum shown in (FIGURE 10 ). has been obtained averaging over several FTU ohmic discharges. A test of the system has been performed with 14 MeV D-T neutrons at FNG (Frascati Neutron Generator): the energy resolution was found to be 3.7% at 14 MeV. The energy resolution obtained for FTU (9.7%) is in good agreement with the formula [9]:


where the constant C is ~14.

The aquisition system was developed using LABVIEW software and standard CAMAC modules. The unfolding program is FLYSPEC [10] which has been rewritten and tested in LABVIEW G language.

FIGURE 10: Neutron spectrum in FTU ohmic discharges


[1] L. Bertalot, B. Esposito, S. Podda and S. Rollet, Rev. Sci. Instrum. 63(10), 4554, (1992).

[2] L. Lagamba, M. Angelone, P. Batistoni, B. Esposito, M. Martone, M. Martone, M. Pillon and S. Podda, ENEA Report, RT/ERG/FUS/95/18, (1995)

[3] L. Bertalot, M.Damiani, B. Esposito, L.Lagamba, S. Podda, P.Batistoni, P.De Felice, R.Biagini, Rev. Sci. Instrum, 68, 1, (1996).

[4] P. Batistoni, M. Rapisarda, D. Anderson, Nuclear Fusion, 30(4), 625, (1990).

[5] B. Esposito et al., Plasma Physics and Contr. Fusion, 35, 1433, (1993)

[6] P. Batistoni, B. Esposito, M. Martone and S. Mantovani, Rev. Sci. Instrum., 66(10), 4949, (1995).

[7] J.M. Adams et al., NIM, A329, 277, (1993).

[8] B. Esposito, L. Bertalot, P. Batistoni and M. Damiani, Proc. 24th EPS Conference on Plasma Physics and Controlled Fusion, Berchtesgaden, Germany, (1997).

[9] Y.I. Kolevatov et al., Neutron and Gamma Spectrometry in Radiation Physics, Moscow, Energoatomizdat, (1990), p. 144

[10] D. Slaughter and R. Strout, Nuclear Instruments and Methods, 198, 349, (1982).

Go to Neutron Diagnostics Home Page

Go to FTU Diagnostics Page