Performance of the Neutron Diagnostic System for Alcator C-Mod.


C. L. Fiore and R. L. Boivin


MIT Plasma Fusion Center, 190 Albany St., Cambridge, MA 02139


Abstract


Measurement of the global neutron production from the Alcator C-Mod tokamak has been done for all experimental operation in deuterium, which commenced in August 1993. Up to 3× 1012 neutrons/s have been measured for diverse plasmas, including high current (1 MA), pellet injected, and rf heated discharges. Measurements during the initial run period were obtained using two sets of neutron detectors housed in polyethylene loaded moderator stations. Two more moderator stations containing additional neutron detectors have been obtained from LLNL and have been installed in the Alcator C-Mod experimental cell. The system has been calibrated using Cf252 placed at discrete points inside the Alcator C-Mod vacuum vessel. This paper presents a description of the diagnostic, followed by details of the calibration method and results. A discussion of the derivation of ion temperature from the global measurements is included with first results from the integrated system.

I. Introduction
The neutron diagnostic system for Alcator C-Mod is capable of measuring the total fusion yield over the full range of deuterium operation. The anticipated neutron source rate will vary from 1× 1011 n/s at Ti=1 keV and \hat ne=1× 1020 m-3, to 1× 1016 n/s for Ti= 6.5 keV at \hat ne=1× 1021 m-3. A brief description of the experimental layout will be followed by details of the calibration. The first results of neutron rate measurements will be presented along with the derivation of the ion temperature from these data.
II. Experimental Layout
The global neutron detection system for Alcator C-Mod is comprised of 12 fission chambers of varying sensitivity as well as 4 BF3 counters and 4 He3 detectors. These are distributed in 4 moderator stations. The first two moderator stations, built at MIT [1,2], house 3 fission detectors enriched 93% in U235, containing 1.68g, 0.08g and 0.004 g of fissile material respectively, in addition to 1 BF3 counter. The remaining 2 moderator stations were obtained from LLNL and are described in Ref. [3]. These each contain three fission chambers enriched 93% in U235, 2 He3 detectors and 1 BF3 counter. The fission chambers were selected to cover the entire operational regime of the Alcator C-Mod experiment using the count rate mode, between 1× 1011 up to 1× 1016 n/s. There is sufficient overlap between the operating ranges to allow cross calibration between detectors.
One of the moderator stations, containing detectors 1-4, is located directly under the center of Alcator C-Mod, inside the borated concrete igloo which surrounds the tokamak. This position was chosen to minimize changes in sensitivity which can occur when large structures (such as other diagnostic systems) are moved near the detectors. The second station (detectors 5-8) is located near the toroidal midplane, adjacent to the large opening in the igloo which accommodates the buswork for the experiment. Of the remaining moderator stations, one is attached to the cell wall at the toroidal midplane and the other is on top of the igloo above Alcator C-Mod.
III. Calibration
Three calibrations of the neutron detection system have been done. A 66.5\mug source (1.56× 108 n/s) was obtained from ORNL for the purpose of calibrating this experiment in January, 1992. The source strength has been adjusted to allow for the decay of the at each calibration date. In each case, the source was placed at discrete locations inside the Alcator C-Mod vacuum vessel. The source position for one location has been repeated for all three cases to establish consistency between calibrations.
The calibration positions are shown in Fig. 1. The source was placed inside a thin aluminum tube (0.4 mm wall) 3 m long. This allowed the source to be moved to different positions in the experiment while minimizing the accumulated dose to the authors. Use of two tangential ports allowed the source to be placed between ports as well as in port centers. The limited number of positions used for the calibration is due to the long counting time (~ 2 hours) required to obtain good counting statistics on the two most sensitive detectors and the short time allocated for this work by the operational schedule.
A summary of the measured efficiency for the two most sensitive detectors (detectors 1 and 5, the 1.68 g U235 in the original detector set) is shown in Table I. The results for the last two calibrations on detector 1 (located underneath Alcator C-Mod) are the same for the redundant points (F port, KT1, and KT3). The F port efficiencies are significantly lower than was measured in the first calibration (1/23/92) however. This is attributed to the addition of diagnostic equipment and cables in the F port lower vertical ports between the first and second calibration. Note that the B port measurements obtained on the first and third calibrations are nearly equivalent. The B port measurements show the highest observed efficiency because the vertical port is unobstructed with materials. The differences in the efficiencies measured for detector 5 (toroidal midplane at A port) are due to adjustment of the amplifier and discriminator settings.
Since the position of the moderator containing detector 1 is centered underneath Alcator C-Mod, it is expected that the detector response should be insensitive to source positions on the major radius of the device. Little difference was found between data taken with the source placed in a port and that with source placed between ports in the earlier calibrations, so all of the data was averaged to obtain a calibration value for this centered detector. The exception, noted in the previous paragraph, is the high sensitivity of the detector to the source when it was placed in the center of B port. This port has little material which could obstruct the free streaming of neutrons, as opposed to most of the other ports which are filled with diagnostic equipment and cables. Thus a weighted average was used to determine the efficiency for this detector in the final calibration which accounted for a somewhat higher contribution from the exposed area of the three unobstructed ports. A value of 1.67× 10-9 counts/n was obtained both for the weighted average and by averaging all of the available data points as had been done previously. In addition, the 3-d Monte-Carlo neutron transport code(MCNP) [5] was used to calculate the anticipated efficiency for this detector location to a ring source. The MCNP code was also used to calculate the efficiency of the detectors to placed in actual calibration positions in order to gain a normalization value for the ring source calculation. When this was done, an efficiency of 1.64× 10-9 counts/n was obtained from the MCNP calculation.
The efficiency for detector 5 was determined as a function of the toroidal angle of the source location. A polynomial was fit to the data, and it was integrated to determine an overall efficiency. The MCNP calculation described above was also used. These methods gave efficiencies of 1.70× 10-8 counts/n and 1.74× 10-8 counts/n respectively. Ultimately, this and the remaining detectors will be calibrated relative to detector 1 from actual plasma produced thermonuclear neutrons. This is necessary because additional hardware was installed between the moderator containing detector 5 and Alcator C-Mod following the most recent calibration.
The errors in the rate measurement from known sources are 2.6% in the source rate, 14.7% from the averaging technique, 6.7% from counting statistics, and 2.7% from positioning uncertainty. An additional 7.7% error is used to account for the difference between the energy spectrum and that of the D-D fusion neutrons (from an MCNP calculation for Alcator C-Mod detailed in Ref. [2]). This gives an overall error in rate of 18.3%.
IV. Neutron Rate and Ion Temperature Measurements
Deuterium was introduced as a working gas in Alcator C-Mod in August of 1993. The global neutron production rate was measured for all deuterium discharges. The highest neutron production rates (~ 3× 1012 n/s) were obtained during either deuterium pellet injection experiments or ICRF heated discharges. Examples are shown in Fig. 2.b and Fig. 3.b
The neutron production rate is related to the volume integral of the deuterium ion density and velocity as given by
R = 1/2 \int dV nD2\langle\sigma v\rangleDD/2
where \langle \sigma v\rangleDD is the value of the deuterium fusion cross section averaged over the velocity distribution function. When Ti is below 25 keV, the following approximation can be used [5]:
\langle\sigma v\rangleDD = 2.33× 10-14Ti-2/3 exp(-18.76 Ti-1/3)
where Ti is in keV. This expression cannot be trivially inverted to obtain the ion temperature. Reduction of the ion temperature from the neutron rate is done iteratively. A trial central ion temperature is used with an assumed ion temperature profile and integrated with the measured density profiles over the plasma volume to obtain a calculated neutron rate. The electron density is corrected from the measured plasma Zeff to obtain the deuterium ion density. The calculated rate is compared to the measured neutron rate, and the central ion temperature is adjusted up or down accordingly. This is repeated until the difference between the calculated and measured neutron rates is less than 5%.
Obtaining the ion temperature in this manner is computationally intensive, and was beyond the resources of prior tokamak experiments at MIT. Use of Digital Equipment Alpha workstations reduces the computing time to several minutes, which can be easily completed between tokamak discharges.
The neutron rate is a strong function of the ion temperature in the present operational range of the Alcator C-Mod tokamak. The dependence of neutron rate on ion temperature can be approximated as Ti5.6 for Ti near 1 keV. Using this form to compute the error in the calculated ion temperature, it can be seen that the error in the neutron rate of 18.3% results in an error of only 3.3% in the ion temperature at low temperatures. The error resulting from uncertainty in the density measurements is higher because of the dependence of the neutron rate on the square of the deuterium density. Unfortunately, the error in the electron density is unavailable at this time. Trial reduction of the data in which the assumed ion temperature profile width was varied by 30% resulted in an 8% variation in the ion temperature. The overall error from all sources is not expected to exceed 10% in this temperature range. Good agreement has been found between this measurement of ion temperature and that obtained from Doppler broadening of selected x-ray emissions from impurities in the plasma. Examples of the ion temperature determined from the neutron rate is shown in Fig. 2.a and Fig. 3.a.
V. Summary and Future Work
All four planned moderator-detector stations for measuring global neutron production on Alcator C-Mod have been installed on the experiment. Data from the first two was obtained for all deuterium operation to date, and ion temperatures have been derived from the measured rates. Data from all four systems will be available for all subsequent operation.
Acknowledgements
The authors would like to thank Jody Miller, Bob Granetz and Daniel Lo for assistance with the calibration of the experiment. Data from Earl Marmar, Jim Irby, Tom Luke, and magnetics analysis from Steve Wolfe were used in the reduction of the ion temperature. The authors are indebted for the use of the source to the U.S. Department of Energy's Californium Industrial Loan Program as administered by the Office of Nuclear Materials Production through the facilities of the Oak Ridge National Laboratory. This work was supported by US DOE Contract No. DE-AC02-78ET51013.
References
[1]C. L. Fiore, R.L. Boivin, R.S. Granetz, Rev. Sci. Instrum. 63, 4530 (1992).
[2]C.L. Fiore, R.S. Granetz, Rev. Sci. Instrum. 61, 3166 (1990).
[3]T. Ogawa, K. Oasa, K. Hoshino, K. Odajima, and H. Maeda, Rev. Sci. Instrum., 61, 3181 (1990).
[4]David L. Book, NRL Plasma Formulary, (Naval Research Laboratory, Washington, D. C., 1983).
[5]J. Briesmeister, ``MCNP--A General Monte Carlo Code for Neutron and Photon Transport,'' LA-7396-m, Rev.2, Los Alamos National Laboratory, Sep. 1986.


Figure Captions
Figure 1. Schematic cross section of Alcator C-Mod showing source positions during calibrations and the location of the first two neutron moderator stations.
Figure 2. Neutron rate (a.) and ion temperature (b.) resulting from deuterium pellet injection into a deuterium plasma. The pellet was injected at 0.5 s. The plasma current was 0.8 MA. The central electron density was 2.0× 1020/m3 before the pellet, and 5.0× 1020/m3 following the pellet. The electron temperature exceeded the ion temperature by 0.4 kev prior to the pellet injection, and was equivalent to the ion temperature following the injection.
Figure 3. Neutron rate (a.) and ion temperature (b.) resulting from ICRF injection into a deuterium discharge with a small hydrogen minority. The rf injection was from 0.47 s to 0.67 s. A deuterium pellet was injected at 0.4 s, causing the central electron density to increase from 1.8× 1020/m3 to 5.0× 1020/m3. The electron density had decreased to 2.0× 1020/m3 by 0.52 s, when the neutron rate and ion temperature began to reach their peak values. The central ion and electron temperatures were equivalent prior to the rf injection at which point the ion temperature began to exceed the electron temperature. A 0.5 kev difference was observed at the ion temperature peak, and the two became equal again by 0.6 s. The plasma current was 0.8 MA.
Table I. Calibration Data: Counts/Source Neutron


Source LocationDetector 1Detector 5
3/28/94F1.45× 10-93.28× 10-9
F(-5cm)1.45× 10-93.20× 10-9
F(+5cm)1.51× 10-93.29× 10-9
CTB2.67× 10-92.77× 10-8
CTBC1.8× 10-92.77× 10-8
CTAB1.62× 10-94.13× 10-8
KT11.77× 10-96.29× 10-8
KT31.48× 10-94.38× 10-9
8/23/93F1.54× 10-91.12× 10-9
KT11.77× 10-83.06× 10-8
KT21.50× 10-92.54× 10-8
KT31.36× 10-92.12× 10-8
KT41.41× 10-91.72× 10-8
1/24/92F2.48× 10-95.05× 10-10
B2.42× 10-96.925× 10-9
D2.80× 10-91.91× 10-9
H3.23× 10-91.24× 10-8
K2.80× 10-97.52× 10-9

File converted from TeX by tth; beware of translation errors.