Photofission for Active SNM Detection II - IEEE Xplore

3 downloads 0 Views 2MB Size Report
funded through the UK Home Office, Ministry of Defence and ... support also from the Office of Naval Research and the Defence. Nuclear Detection Office.
Nl-9

2012 IEEE Nuclear Science Symposiwn and Medical Imaging Conference Record (NSS/MIC)

Photofission for Active SNM Detection II: Intense Pulsed

19P(p,ay)160

Characteristic

y Source

P. Mistry, C. Hill, Member, IEEE, J .O'Malley, J. Precious, M. Ellis, Member, IEEE, R. Maddock, F. C. Young, S. L. Jackson, Member, IEEE, D. G. Phipps, R. Woolf, B. PhIips

material (SNM) for Homeland Security purposes. In active detection, an external source of radiation is used to stimulate fissile material to fission and the subsequent measurement of the prompt and delayed neutron radiation and delayed gamma radiation associated with the fission process is used to enable detection. There are however a huge variety of options of the energies and types of radiation which could be deployed to do this. This includes the use of various energy photon or neutron radiation and also delivery of this radiation in the form of a single intense pulse, vs. as a continuous or high frequency repetitive pulse. In terms of photon sources for the interrogating radiation, the use of narrow band, characteristic gamma radiation (produced in beam target interactions such as 19F(p,ay)160) has been investigated as well as the use of broad band, Bremsstrahlung radiation. Both of these photofission source options were compared with a variety of neutron sources options which ranged from low energy, beam target sources (e.g. 7Li(p,n)7Be), DD and DT fusion neutron sources, DD beam target sources and photonuclear production from (y,D or y,Be). Three of these discussed techniques have been investigated experimentally by AWE in collaboration with the Naval Research Laboratory (NRL), Washington [1, 2]. Experiments [3] using the intense pulsed 19F(p,ay)160 characteristic y source will be discussed here.

*Abstract- An ongoing programme investigating the active detection of special nuclear material (SNM) is being undertaken by the Atomic Weapons Establishment (AWE) in collaboration with the Naval Research Laboratory (NRL). The programme is funded through the UK Home Office, Ministry of Defence and Cabinet Office and the Naval Research Laboratory supported primarily through the US Defence Threat Reduction Agency with support also from the Office of Naval Research and the Defence Nuclear Detection Office. The process by which the UK are applying active detection techniques to border protection and a review of the current challenges and opportunities for this technology as assessed by the authors is provided. As part of this programme, the NRL Mercury IVA was operated

in

positive polarity mode to produce photons 19 16 F(p,ay) 0 reaction, at energies of 6.13,

characteristic of the

6.92 and 7.12 MeV. Protons produced by Mercury interact with a thick Teflon (PTFE) target to produce characteristic gamma radiation. These in turn were used to induce photofission in a depleted uranium (DU) sample.

Eighteen experiments were

fielded in September 20 11, in which thirty-five detectors were 3 fielded, including He tubes, NaI detectors, liquid scintillators and high purity germanium detectors, capable of detecting both gamma radiation and neutrons. The results from a selection of those detectors are discussed here. A

variety

polyethylene)

of and

high-Z

(lead)

hydrogenous

and

hydrogenous

shielding

(borated

configurations

was

employed and positive detection was made up to the maximum shielding tested, 8.Sg/cm2. Effects of secondary reactions in the photon production are visible in the results and some employed

II.

reduction techniques are discussed. Monte Carlo modelling has 3 been employed for a subset of the He tubes fielded. The results have been found to agree within an order of magnitude, but have also been found to die away more quickly than observed in the experimental data.

I.

INTRODUCTION

Researchers at the Atomic Weapons Establishment (AWE) have been investigating active detection of special nuclear Manuscript received November 16, 2012. This work was supported by the U.K. Home Office and Ministry of Defence P. Mistry, J. O'Malley and

e. Hill are with the Hydrodynamics Division,

Atomic Weapons Establishment, Aldermaston UK (telephone: 0118 9823491, e-mail prina.mistry @awe.co.uk) .

M. Ellis and J. Precious are with the National Nuclear Security Division,

Atomic Weapons Establishment, Aldermaston UK. R. Maddock is with the Radiation Science Division, Atomic Weapons Establishment, Aldermaston UK. S. L. Jackson, D.G. Phipps and

F.e. Young are with the Plasma Physics

Division, Naval Research Laboratory, Washington DC, USA. R. Woolf and B. PhIips are with the Space Science Division, Naval Research Laboratory, Washington DC, USA

978-1-4673-2030-6/12/$31.00 ©2012 Crown

EXPERIMENTAL SETUP

Calculations performed in advance of the experimental setup revealed that the expected gamma yield above the 6MeV energy range required to induce photofission in depleted uranium (DU) was an order of magnitude lower than those seen in a previous experimental series performed using Bremsstrahlung x-rays [2]. In order to provide a suitable comparison between the two series, the conjugate distances were significantly reduced. This was achieved by placing the DU plate at a 60° angle 25cm from the PTFE target. Two stand clusters were then arranged either side of the Mercury rails as close as physically possible. The decision on which detectors to place on which stand was made favouring photon reflection for the gamma detectors and neutron transmission for the neutron discriminators. The reflection stand featured an array of delayed gamma detectors such as NaI and liquid scintillators. The detectors on the transmission stand were the 3He and another set of liquid scintillators, identical to those on the reflection side for continuity. In addition, on the reflection side, a stand housing two HPGe detectors was positioned, and also a fourth stand, roughly on axis with the DU plate

24

supported a Rhodium counter and a prompt gamma detector. These were positioned to maximise detection of the output from the machine as opposed to the fission products from the DU. The 30.S cm x 30.S cm x 2.S cm thick encapsulated DU target was placed approximately 244 cm from the source and at a range of distances to a number of different radiation detector systems. Detectors used to measure the gamma and neutron signatures arising from the induced fission included 3He tubes, NaI detectors, plastic scintillators, liquid scintillators, LaBr detectors, FC detectors, and high purity germanium detectors. The experimental setup can be seen in schematic form in Fig. 1 and photographically in Fig. 2. Both bare DU and DU shielded by different thicknesses of Pb and 2% borated polythene (BPE) shielding were used as well as a number of inert target materials such as aluminium, copper and lead to assess the 'active background' associated with the interrogation beam on the various detector systems.

a-k gap

Pred 1's (xlOl1)

(m il)

Converter/backing

3.81

3.62

5 PTFE/20 Al

DU

3.81

3.38

DU

3.81

Shot#

Target

1100

DU

3.81

1101

DU

1102 1103 1104

DU

3.81

1.23

5 PTFE/20 Al

1105

DU

3.16

20 PTFE/none

1106

DU Pb

3.81 3.18

3.03

20 PTFE/none

3.81 3.18

3.27

20 PTFE/none

3.38

20 PTFE/none

3.18

3.27

20 PTFE/none

DU

3.18

3.19

20 PTFE/none

DU + 0.158cm Pb

3.18

3.32

20 PTFE/none

1107 1108 1109 1110 1111

1112 1113 1114 1115 1116

DU + 0.158cm Pb DU + 2.54cmBPE

+

5 PTFE/20 Al 20 PTFE/none

2.54cmBPE

DU Cu DU + 0.3175cm Pb DU +

5.08cmBPE

+

0.3175cm Pb

DU +

1117

5 PTFE/20AI

3.18

20 PTFE/none

3.18 3.18

3.05

20 PTFE/none

3.38

20 PTFE/none

3.18

2.76

20 PTFE/none

3.18

3.42

20 PTFE/none

3.18

2.72

20 PTFE/none

5.08cm BPE

DU

There was no anode backing for shots 1101 and 1102. The remaining shots had an AI-Mylar anode backing. The DU plate was encased in 0.63Scm AI. The DU and Pb plates had dimensions: 30.S cm x 30.S cm x 2.S cm. The Cu plate had dimensions: 30.S cm x 30.S cm x 10.S cm. The shielding materials had the following densities form which the areal masses were calculated: PPb 11.3g/cm\ PBPE 0.9Sg/cm3 and PAl 2.7g/cm3. Details of the detectors were presented in [4, S]. The details of detectors fielded in this campaign follows.

Fig. 1. Characteristic Gammas experimental setup schematic

=

=

=

A. Liquid Scintillators Eight liquid scintillators were fielded in two similar Pb housings on the floor either side of the Mercury firing point. Each housing contained one lS.2cm square EJ309, two 12.7cm diameter BCSOIA, one 1O.lcm square EJ309. In addition, two small S.lcm x lS.2cm BCSOIA scintillators were situated on the Space Science NaI stand in front of the PMTs. In the second week of experiments one of the small BCSOIAs was swapped for a larger 20.3cm diameter BCSOIA detector. Fig. 2. Characteristic Gammas experimental setup photo

B.

3He Tubes

Eight 3He tubes were fielded and divided into four pairs. All tubes were moderated with a combination of cadmium foil and polyethylene. The charge amplifiers attached to the tubes were not shielded as they have been fielded at a similar distance relative to Mercury without shielding in previous experiments.

In total, 18 shots were fired covering nine different experimental configurations which included bare DU, six shielded DU configurations and two surrogate targets. A summary of the trial configurations is listed in table I. Further descriptions of the radiation detectors fielded as indicated in Fig. 1 are provided in the proceeding sections.

C. HPGe Detectors Four HPGe detectors were available to field. Two were used to record the activation of the AI pucks and were located

TABLE I. SHOT SUMMARY

2S

III. MACHINE PERFORMANCE

some distance from the test cell and control room to avoid recording the shot. A further two detectors were setup in the Mercury cell to record the gamma signal from the DU plate. The HPGe detectors were fielded with additional lead shielding. For shots 1101 - 1109, the HPGe detectors had 5.08cm of lead in front of them, and for the remaining shots, 1110 - 1117, with 2.5cm of lead in front of them.

The machines performance [7] was altered from shot to shot by changing the anode-cathode distance (a-k gap) within the diode, in this way the maximum ion energy was varied from �3.l2MeV to �4.08MeV. The 19F(p,uy)160 reaction is energy dependant, therefore it is important to know the energy of the incident protons if any calculation of the generated characteristic gamma yield is to be carried out. Initial estimates of the diode voltage and ion current are shown in Fig. 3 and 4.

D. Sodium Iodide (NaI) Scintillators Ten 5.lcm x 1O.2cm x 40.6cm NaI scintillators were fielded in two sets of five on the Space Science stand. The detectors were hung downwards, supported by an AI collar, to present the maximum area to the DU plate. The crystals were shielded by 5.1cm BPE (2% borated polythene) and 0.32cm Pb on the front face and 5.lcm BPE and 5.08cm Pb on the back face. The PMTs were shielded by at least 1O.16cm Pb in all directions. In addition, a 10.2 x 10.2 x 40.6cm NaI detector was located above the liquid scintillator PMTs on the reflection stand. This was shielded with �5.1cm Pb. E.

1100"'P1 1101'\'PI 1102""1>1 1103'."",

110�'.ypI 11050"1>1

1108\'PI 1101\"1>1 1101

-111(1"'P1 11111\1»1

Aluminium Activation Pucks

In the second week of testing, two Al activation pucks were deployed to obtain an estimate of the machine produced neutron yield. These were located as close as possible to the converter target. Activation of each puck following each shot was measured using HPGe detectors outside of the cell and control room. A separate HPGe detector was used for each puck. F.

'010

'

T,me

0

�I\S)

me>

"00

","0

Fig. 3. Estimated diode voltage traces from Stella data of shots performed.

" OO·.llonpfo' , 'O,\1lonpfO

:

, , 02l11onpfo

1103',II",.,fO',

1104lfi",-..,1

1 105·,1...., 1.

Rhodium Activation Counter

: : 1109\llonpf.1 ," O',lionpf.:

- 1106'.1.,.., 1 0 - I 107\1...., 1 0

A 5.08cm rhodium activation counter was fielded approximately on the DU plate axis (see Fig. 1) also to assess the neutron output of the machine. The counter is enclosed in a 20cm x 20cm x 61 cm housing. G.

'010

- 108'.1.,.., 1 0

j

,�

-

� '.

1 1 1 1 11_,.,

Prompt Gamma Detector (Plastic Scintillators)

The prompt gamma detector, built by NRL [6], was deployed to assess the flux of characteristic photons produced from the diode. As such it was located directly below the rhodium counter, on axis with the DU plate (again, see Fig. 1) to minimise the fission signal. The detector itself is a plastic scintillator with a PMT and a selection of filters. The shielding was initially varied between 20.3cm and 30.5cm Pb to fmd the optimum. 25.4cm Pb was used for the majority of the shots.

1�

1J..'

tiKJ

t;»J

omo{",)

,.

11Ja

U2J

Fig. 4. Estimated ion current traces from Stella data of shots performed.

Note the inverse relationship between the current and voltage traces due to the fixed impedance of the machine. The machine voltage cannot be measured directly and are instead inferred from measurements of the current using a relationship previously determined by NRL and described in [7]. The estimated neutron yield of each shot is shown in Table 2 as obtained from the rhodium counter.

This paper focuses on the analysis of three of the diagnostics for two different purposes, photon and neutron detection. Both NaI detectors and liquid scintillators were fielded for the purpose of photon detection and the results from these will be compared. In addition, liquid scintillators are also capable of measuring neutrons via pulse shape discrimination and these data will also be compared to the 3He tube results. The 3He detector results discussed in this report will focus on the AWE tubes which were 40cm length and moderated by a combination of 2cm thick polyethylene, 0.95cm flexi-boron and O.013cm cadmium foil.

TABLE II. NEUTRON YIELD FROM RHODIUM COUNTER RECORDS

26

Shot#

Peak Voltage

Measured Neutron Peak Current Yield (xl04)

1100

3.93

31

1101

4.08

31

17.60

1102

3.96

28

10.20

1103

3.97

32

9.89

1104

2.73

22

2.62

1105

3.85

33

5.28

14.40

1106

3.3

32

6.50

1107

3.89

33

6.11

1108

3.44

36

6.48

1109

3.33

31

6.98

1110

3.34

33

16.00

1111

3.35

33

7.56

1112

3.35

27

3.66

1113

3.32

30

2.05

1114

3.37

34

6.50

1115

3.23

26

5.21

1116

3.43

31

5.23

1117

3.12

32

2.88

IV.

55

Die-Away Comparison

.. 4

§

0 u 3

ANALYSIS OF RESULTS Tine(s)

While an extensive number of detectors were fielded, this paper will only discuss the AWE NaI, 3He and liquid scintillator detectors and their analyses. For each detector, the number of counts in a specified time period from the time of flash to lOs has been obtained. For photons these have been further cut for energy above the nonnal background of 2.5MeV. For neutron detectors, no energy cuts were made. In addition, comparison of the die away following irradiation has been plotted for neutron and photon signals to allow comparison of the detector types.

Fig. 6. Die-away counts in helium-3 detectors - comparison between DU and inert material targets

The shot summary for neutron signal is shown below in Fig. 7. The shots are colour coded to indicate if the shot was a bare DU target (green), a shielded DU target (orange), an inert object (pink), or a non-PFTE diode (red). Immediately obvious from this colour-coding is that all shots had a fairly consistent neutron signal, including those with non-PFTE diodes or inert targets. The exception to this is shot 1106 in which the first 5 seconds of the liquid scintillator data was lost. This suggests that the bulk of the neutron signal seen is independent of the diode or target and is in fact machine produced and very early in the time record. This is consistent with efforts made to reduce the machine produced neutron fluence as the series progressed by including crinkled Mylar and reducing the voltage. Shot 1104 suffered from low voltage, and also showed a low neutron signal in the liquid scintillators.

A. Neutron Analysis Helium 3 Counts

as

a Function of Ateal Mass (3HEC)

:t

180 ,------, ----160 t---.�.� . . �.-=. :-: . � ..:o.,-----r 1 '=.= ' P=OI Y =th=en� 140 t--cc-"-'---- -------------1 c : _ . I : :: �:bn i de � 120 1:... � ... ����·;-:·-=-. � ... . ----i.:.:!��� � 100 t--c------'-''1_a------'--.cc:-------------j , � 80 t--c----��--�---_____j � 60t--c-----�----��---_____j 40t--c------�------� -j , �---... 20 t--c-------� '. � �------------j � � " _

Areal Mass (gem"')

Summed Neutron Signal

10

-

Shot Only

-

Fig. 5. Neutron attenuation with shielding in helium-3 detectors

E 0.5

-

0; 3 "'e .,:;I.I!>

Fig. 5 illustrates how the counts observed vary with increasing shielding on the DU plate. Lead shielding is the only material which does not follow the trend of the polythene and combined shielding, This may be due to (y, n) reactions taking place in the lead; causing an increased number of neutron counts, Fig. 6 illustrates a comparison of neutron counts seen in the bare DU and that seen in the inert materials such as lead or copper. There is clear die-away from the fission products in the bare DU shot compared to the inert material shots, where there are hardly any die-away counts.

� 2 ., §o 1$ (..')

-

r-

-

r-

-

r-

r-

-

r- r-

-

r-

-

r- r-

I-

1

-

1-

0.5

-

I-

;-

-

r- f-

r-

ir

rI-

-

r- f-

-

r- f-

-

I-

-

-

r- rI- f-

-

-

1I- f-

1'01 1102 11� 1104 1'06 1106 1107 1'08 1100 1110 11'1 .112 1113 1114 1116 .,1S 1111 Shot Number

Fig. 7. Total number of neutron counts per shot - liquid scintillators

The corrected neutron counts are then compared to the mass thickness of the shielding configuration for those shots with any target shielding. The results are plotted as function of mass thickness in Fig. 8 below. As can be seen, there is no obvious attenuation trend in the data with only a small amount of variation between the different shielding configurations. This is fairly consistent with the Bremsstrahlung data obtained

27

Graph of neutron die-away following photofission of bare

in March 2011 [2] where a clear dependence was not obvious, especially over the low mass thickness range (less than 20g/cm2). In terms of overall signal, the approximate number of neutrons per cm2 at 1m from the DU plate (for a bare shot) was approximately 2. However the signal from the Bremsstrahlung experiments in March 2011 was an order of magnitude higher, with approximately 23 (averaged for bare shots) neutrons per cm2 at 1m. This is despite the DU objects position being much closer to the Mercury front end in the characteristic gamma experiments. This suggests the number of photofissions induced using characteristic gammas is significantly less than those induced using Bremsstrahlung radiation.

DU via characteristic gammas: Average background

corrected 3He for all shieldingconditions

100.0

-0 C ::>

e J2 � � .oc

,



o ::> - 0

-0

V



• •

• •

.. .. .� �

�;

o::I: 7 �

10

-0 ..

S





... ..... . • ••• • • ''''

10

0.1

8g

.. �.o � ::> Q.I VI

Neutron attenuation with shielding

.....

0.1

0.



Time following irradiation in s



NRl3He - bare DU



NRL 3He -1" BPE



NRL

3He -l/16"Pb

• NRL

3He -1/8"Pb

NRL 3He 2" BPE

Fig. 10. Neutron die-away counts - comparison between 3He tubes and 1;

3.5

increasing DU shielding

1-------__

& , .---�-----------1 B.

� c

� 2.5 '-----"-----------__1

The shots in Fig. 11 and 12 are again colour coded in terms of configuration; bare DU target, a shielded DU target (orange), and inert target (pink), or a non-PTFE diode (red). In this case it is easy to see the contribution of the inert objects on the gamma return from photofission. However, a concerning point is the lack of effect of replacing the PFTE target with polyethylene. This suggests that, while we are seeing products of fission in the DU plate, they may not be caused by gamma photons characteristic of the 19F(p,uy)160 reaction as 19F is not present in the polyethylene diode shots. This suggests the fission is being induced by other means. This may be photons from another reaction or possibly neutrons. It is highly unlikely that any other reaction will produce an abundance of photons with energies higher than the photofission threshold in 235U, therefore neutrons are the more sensible assumption. In this experiment, neutrons produced via deuteron interactions with carbon and fluorine are likely to be fairly abundant. These results suggest that a proportion of the induced fission may be attributed to these reactions.

Z

f--_-__-_-_-_--

t .5 ·

Mass thfckneSS8(g/cm2,O

o

Attenuation

shielding

12

relationship

14

for neutrons

Fig. 8. Neutron attenuation with shielding in liquid scintillator detectors.

A comparison of neutron die-away counts detected by the 3He and the liquid scintillators from bare DU following the radiation pulse can be seen in Fig. 9. The effect on die-away neutron counts with increasing amounts of shielding is illustrated in Fig. 10, with no obvious difference between the bare DU shots and the shielded shots. Graph of neutron die-away following photofission of bare DU via characteristic gammas: Averagebackground corrected 3He vs normalised liquid scintillator results II •

I

: :





. .

•• ••



� .



J..

. !:.

Summed Gamma Signal

.."'!

• NR1.3He

- bare au

Time follQWinglrradiaUon in s

-

+lsclnt bare DU

-

Shot Counts Only

1000

"'" E � 800

J..

10

Gamma-ray Analysis

10

100

NE 700

o �800

... AWE 3He -bare DU

� J!l500

C ::J4()() o o ",300 E E200 '" °' 00

Fig. 9. Neutron die-away counts - comparison between the 3He tubes and the liquid scintillators

t-:.

l-

I--

t----

I--

-

-

I-

f-

t-

-

I-- rI--

t-

-

f-

-

r-

r- tI--

t- l1101 11041 110J

t-I- t--

l-

t- t- t- l- I- I"04

"05

1106 1107 1108

IIOS 1110 1111

Shot Number

1112 111:3

'1'4 111� '116 1'17

Fig. II. Total number of gamma counts per shot - liquid scintillators

28

Averaged Gamma Signal for Nal wall - Shot Counts Only

3000

illustrated in Fig. 16, again with no obvious difference between the bare DU shots and the shielded shots, as in Fig. 10.

2500

Graph of gamma die-away following photofission of bare DU via Characteristic Gamma Radiation: Average backgrounil oorrected Nal vs normalised Liquid scintillator results

2000 1500

f-

-

f-

1000

f-

-

f- f-

500

f-

-

f- f-

n

HlOO_O -r------,

. ...... .

18.

n

,.

1101 1102 11 0311041105 11 06 1107 11 08 1109 1110 1111 1 112 111311141115 11161117 Shot Number

100.0

. .......

.

1-------0-"--------'-- '----1 .

.



Fig. 12. Total number of gamma counts per shot - sodium iodides 10.0 -1------_---_---1 0.1 10 100 nma foIlowtng Irradiation in s

The corrected photon counts (with background subtraction) are then compared to the mass thickness of the shielding configuration for those shots with any target shielding. The results are plotted as function of mass thickness in Fig. 13 and 14 below. gOO

E



...

NE u

800 750

!!l 650 c: S 600

E �

(!)

Graph of gamma die-away following photofission of bare



DU via Characteristic Gamma Radiation: Average

550



� •

� .



background corrected Nat for all shielding conditions 0

'" c:



0 ..... v

.

� IY�-22017X'801.511 �

:;:



:> 0 U



� " !!l c:

.

500

I

.. Lseinl - bare OU

shots

Gamma attenutation with shielding �

Nal -bate OU

Fig. 15. Die-away gamma counts in NaI and liquid scintillators for bare DU

850

8. 700 u





. t

• • •

.. .

450 400

12

.00

14

.,

'.

Time following irradiation in s

• Nal-bare DU

Fig. 13. Gamma attenuation with shielding in liquid scintillator detectors.



Nal -1" 6PE



Nal-l/16"Pb

Nal-l/8"Pb

Nal-2" 6PE

Fig. 16. Die-away gamma counts in shielded shots Gamma Attenuation vJth Shielding (Nail

C. Detector Dead-Time �"'" .. � 500



Problems with detector response to the initial radiation flash were encountered. 3He, EJ30119 LS, and NaI show periods of dead time immediately following the radiation pulse. This could be caused by one, all or a combination of the following: Detector saturation observed following the photon/neutron flash; Radiation interference with on-board electronics or; electromagnetic interference (EMI) from Mercury. The liquid scintillators showed additional drop outs at random locations which could be caused by digital data acquisition buffering. It is uncertain as to whether the data acquired following dead­ time arises from fission signatures or environmental effects. An example of dead-time in the 3He detectors is illustrated in Fig 17.

ly:-29.52Sx'848061

� 400 •

U"'"

• � 200



Mass Thickness (gfcm")

Fig. 14. Gamma attenuation with shielding in sodium iodide detectors

The photons clearly display more of dependence on the shielding mass thickness than the neutrons did in Fig. 8 and the relationship appears to be approximately linear; although it must be noted that the scales may not be appropriate to display an exponential relationship that would be expected. A comparison of gamma die-away counts detected by the NaI and the liquid scintillators from bare DU following the radiation pulse can be seen in Fig. 15. The effect on die-away neutron counts with increasing amounts of shielding is

29

Typical