Embed Size (px)
Citation preview
R. Scott Kemp
Princeton University
Detection of Clandestine Centrifuge Enrichment
American Physical Society panel on Technical Steps to Support Verifiable Nuclear Arsenal Downsizing - 22 April 2009 - Washington DC
sensitive
insensitive
low
high
low
high
false-alarm rate
detection probabilitydete
ction
sens
itivit
y
sensitive
insensitive
low
high
low
high
false-alarm rate
detection probabilitydete
ction
sens
itivit
y
sensitive
insensitive
low
high
low
high
political goals may permit a relaxation of constraints
false-alarm rate
detection probabilitydete
ction
sens
itivit
y
External Constraints
Ability to verify HEU stockpiles, esp. Navel stocks
Measurement uncertainties in reprocessing plants1% of all Pu throughput ( = 80 kg at Rokkosho ) or 10% of Pu in MOX, whichever is greater
Other uncertainties (e.g., dismantlement)
Metric Tons (MT)
HEU Stocks SignificantSource: International Panel on Fissile Materials (IPFM)
The Means of DetectionA partial list ?
Optical
Effluent
Electromagnetic
photography
thermal Infrared
HF gas
UO2F2 aerosols
line notching
power waveform distortion
ULF on transmission lines
ULF radio
atomic magnetometer
Effluents
UF6 + 2 H20 UO2F2 + 4 HF
UO2F2 Aerosol Characteristics
Release Scenario Size Vdep*
Routine 0.01-1 µm ~2 m/day
Massive Spill 1-2 µm ~50 m/day
Source: R.S. Kemp “Initial analysis of the detectability of UO2F2 aerosols produced by UF6 release from uranium conversion plants,” Science & Global Security, 16:3 (2008).
* all calculations assume effective deposition velocity of 86 m/day (0.1 cm/s)
Source Terms
Type FacilityEmissions Rate
95% CI[gUF6/tHM-Unat]
Emissionsfor 1 MT HEU/yr
[gUF6/day]
Centrifuge Capenhurst 0.0010 - 0.0023 0.0013n=194
Source Terms
Type FacilityEmissions Rate
95% CI[gUF6/tHM-Unat]
Emissionsfor 1 MT HEU/yr
[gUF6/day]
Centrifuge Capenhurst 0.0010 - 0.0023 0.0013n=194
Conversion
Honeywell Metropolis
13.4 - 19.5 11.3
Conversion CamecoPort Hope
3.0 - 4.3 0.13Conversion
ComurhexPierrelatte
0.01 - 0.04 0.0082
n=7
n=20
n=3
-104 -102 -98 -96 -94
28
30
32
34
100 km
200 km
300 km
400 km
NOAA HYSPLIT MODELConcentration (mg/m3 UO2F2) averaged between 0m and 100m
Integrated from 1200 18 Jan to 0000 19 Jan 03 (UTC)Continuous UF6 release started at 1200 02 Jan 03 (UTC)
1.0E-06 1.0E-07 1.0E-08 1.0E-09
Sour
ce
32
.41
N 9
9.68
W
from
8m
EDAS METEOROLOGICAL DATA
µ
1x10-8 1x10-9 1x10-10 µg/m31x10-11
0.01 ppb 0.001 ppb city aerosol sampling
0.3 ppb 0.03 ppb rural aerosol sampling
Port Hope Scenario
0.1 ppt
3 ppt
0.01 ppt
0.3 ppt
ULF RadioUltra Low Frequency = 300 - 3000 Hz
350 - 1500 Hz
power supply driving frequency
ν =pvp
2φπD
λ = 200 - 850 km
ULF RadioUltra Low Frequency = 300 - 3000 Hz
350 - 1500 Hz
power supply driving frequency
ν =pvp
2φπD
near-field dipole magnetic field
λ = 200 - 850 km (r << λ/2π)
Bdipole =µoIob2
4π
1r3
(2 cos θ �̂r + sin θ �̂θ)
Bmax =2µoIob2
4π
1r3
nt nc
ULF Radio
A set of generous assumptions
Bmax =4× 10−3
r3Tesla
Io = 1 Ampb2 = 100 cm2
nt = 1000 turnsnc = 2000 machines
near-field dipole magnetic field
Bmax =2µoIob2
4π
1r3
nt nc
Atomic Magnetometer
Source: I.M. Savukov, et. al., “Tunable atomic magnetometers for detection of radio-frequency magnetic fields,” Phys. Rev. Let., 95:063004 (2005)
.01
0.1
1
10
100
10 10 10 10 10 103 4 5 6 7 8
Frequency (Hz)
Sens
itivi
ty (1
0-15 T
Hz-1
/2)
Litz Wire
Optimal Coil
RF Atomic Magnetometer
V = 20 cm3
ULF Radionear-field dipole magnetic field
Bmax =2µoIob2
4π
1r3
nt nc A set of generous assumptions
Bmax =4× 10−3
r3Tesla
Io = 1 Ampb2 = 100 cm2
nt = 1000 turnsnc = 2000 machines
2× 10−17 =4× 10−3
r3r → 40 km without shielding!
δ ∼ 1 mm
Power Line Effects
Source: D.A. Jarc and R.G. Schieman, “Power Line Considerations for Variable Frequency Drives,” Proc. IEEE-IAS 1985 Annual Meeting
Transmission depends on isolation and line impedance
The Means of Detection
Optical
Effluent
Electromagnetic
photography
thermal Infrared
HF gas
UO2F2 aerosols
line notching
power waveform distortion
ULF on transmission lines
ULF radio
atomic magnetometer
