K. D. Ianakiev, B. D. Boyer, A. Favalli, J. M. Goda, T. R. Hill, D. W. MacArthur,
        C. E. Moss, M. T. Paffett, C. D. Romero, M. K. Smith, M. T. Swinhoe,

                    Los Alamos National Laboratory, Los Alamos, NM, USA

                                      Francis Keel
                                    NNSA, DOE, USA


Improving the quality of safeguards measurements at Gas Centrifuge Enrichment Plants
(GCEPs), whilst reducing the inspection effort, is an important objective given the
number of existing and new plants that need to be safeguarded. A useful tool in many
safeguards approaches is the on-line monitoring of enrichment in process pipes. One
aspect of this measurement is a simple, reliable and precise passive measurement of the
186-keV line from 235U. (The other information required is the amount of gas in the pipe.
This can be obtained by transmission measurements or pressure measurements). In this
paper we describe our research efforts towards such a passive measurement system. The
system includes redundant measurements of the 186-keV line from the gas and separately
from the wall deposits. The design also includes measures to reduce the effect of the
potentially important background. Such an approach would practically eliminate false
alarms and can maintain the operation of the system even with a hardware malfunction in
one of the channels. The work involves Monte Carlo modeling and the construction of a
proof-of-principle prototype. We will carry out experimental tests with UF6 gas in pipes
with and without deposits in order to demonstrate the deposit correction.


The spread of enrichment technology and the coming expansion of gas centrifuge
enrichment plants (GCEP) in the United States, France, and elsewhere demands more
advanced, efficient, and effective GCEP safeguards technology. There is a critical need to
monitor GCEP facilities to ensure that nuclear material is not inappropriately enriched or
diverted. Because of the increasing number and size of enrichment facilities, traditional
manned inspections are becoming too expensive, and therefore, the emphasis in the
modern safeguard approach is trending toward unattended monitoring technologies
[1,2,3]. The past field tests of a transmission-based, Go-No Go Continuous Enrichment
Monitor (CEMO) on a low pressure header pipe in the Capenhurst Plant (URENCO) [4]
and the Ningyo-Toge Enrichment Plant in Japan [5] proved this technology is a viable
component for safeguarding GCEPs. Transmission-based enrichment monitoring
technology was installed beginning in the late nineties as part of the Blend Down
Monitoring System (BDMS) and it is still in use for semi-quantitative enrichment
measurements on high pressure pipes in various plants [6,7]. Similar difficulties with
decaying transmission sources and sensitivity of transmission measurement have been
encountered in both CEMO and BDMS applications. LANL is developing an advanced
enrichment monitoring technology aimed to address the difficulties of its predecessors
and improve the accuracy for quantitative measurements [8]. The decaying isotope source
was replaced with an X-ray tube and notch filter [9], providing flexibility in the choice of
transmission energy. The intensity of the transmission source was stabilized by
monitoring the beam entering the header pipe. We have developed various static UF6
sources allowing R&D work on enrichment monitoring technology in the laboratory and
in an environmental chamber [10]. Despite the fact that a transmission measurement
provides independent correction for UF6 gas density, it adds complexity and cost. The
increased capacity and number of GCEPs invites alternative approaches. The IAEA and
the operators have shown a desire to build a safeguards approach around use of the
operator’s accountancy scales for weight measurements combined with an accurate
online enrichment monitor installed on a unit header pipe. This concept uses
authenticated operator accountancy scales to get the mass of the UF6 in the cylinders in
an unattended mode and an Advanced Enrichment Monitor (AEM) at the headers feeding
or withdrawing from the cascades to measure the enrichment of UF6. Hence, this system
will have the enrichment of the feed, product and tails cylinders available without having
to physically measure each cylinder. To overcome the constraints of difficult
transmission measurements it was proposed to use authenticated gas pressure information
similar to that from accountancy scales and load cells [11-13]. The systematic accuracy
and statistical fluctuation of pressure-based correction for UF6 gas density are much
better than those for transmission and X-Ray Fluorescence (XRF) methods. The presence
of light gases can affect the accuracy, as was noted in [5]. Thus this approach
substantially reduces the complexity and cost per instrument and shifts the focus to
passive measurement of the activity of 235U in the gas.

The main factors affecting the intensity of the 186-keV line of 235U gas in the pipe are the
background signal coming from the deposits on the pipe wall, 238U daughter products in
the gas, and background external to the pipe. The instrumentation factors of stability of
counting characteristics, spectral interference with the 186-keV peak, and reliability add
to the overall performance of the measurement.

In this paper we describe our research efforts towards development of such a passive
measurement system: modeling sensitivity to UF6 gas and wall deposits in the unit header
pipe; development of laboratory setup of gaseous sources (plated and unplated with UF6
deposits); and a multi-detector head for measuring and correcting for wall deposits. A
design concept for redundant measurements and 186-keV peak stabilization will be


Detector choice
The spectroscopy of the 186-keV region of 235U presents significant challenges to
instrumentation choice: low intensity of the 186-kev peak, interference of 238U decay
products and external background including radiation from dump tanks and UF6
cylinders; low intensity or absence of peaks in the spectrum for pulse height stabilization.
Despite excellent resolution and advances in electrically cooled HPGe detectors, they
cannot be considered as an option because of cost, complexity and sensitivity to micro-
phonics. The LaBr3 detectors offer about 30% better resolution than NaI(Tl) scintillators
but at five times the price for the same detector size and they have internal background.
For the time being, the NaI(Tl) detectors possess the best combination of efficiency,
reliability, and cost. The predecessor applications CEMO and BDMS, based on 2” and
3” NaI(Tl) detectors, are a good resource for detector reliability data. For instance the 3”
detectors used in the BDMS system have shown good reliability record: three obvious
failures for 120 detector-years of operation.

Detector configuration
The experimental data with a 3” diameter detector and 3.3% enriched UF6 gas at 50 Torr
pressure [8] gave a 186-keV peak intensity of 28 cps. This is expected to increase to 38
cps for 4.5% enrichment. If a single detector is used, the statistical fluctuations for a 2.5
hour measurement time (the measurement time used by CEMO deployment) would be on
the order of 0.3% before the background subtraction. The detection efficiency and cost
scale almost proportionally: a larger area detector with 5” diameter by 0.5” thick crystal
would provide 2.7 times higher detection efficiency but with three times higher cost. The
detection geometry of a 5” detector does not fit well to a 4” diameter pipe but would be
the best choice for larger diameter header pipes for uranium tails. Therefore our
preference is to use two 3” detectors to reduce the statistical uncertainties. The
conceptual design of a passive measurement head is shown in Fig.1.

Fig. 1. Design concept of a two-channel passive enrichment monitor: the shielding from composite
tungsten with 11g/cc density is shown in gray; the 100-mm header pipe is shown in green; the NaI(Tl)
crystal in blue; and the detector housing in beige. The electronics is in the detector case.
The redundancy of two independent detection channels would provide much higher
instrumentation reliability and continuity of data even in the case of catastrophic failure
in one of the channels. The cost and performance of a NaI(Tl) spectrometer are essential
for this application. Because of the physical size of a GCEP, the advanced enrichment
monitors need to be powered over the data communication cables, for example, Power
over Ethernet (PoE). Our concept envisions a low cost, low power NaI(Tl) spectrometer
based on an MCA with a new principle of operation. The whole architecture of MCA
gain stabilization and data communication will be presented at a future technical meeting.

The shielding design is based on hollow wax casting from composite heavy tungsten
powder and polymer binder with average density of 11 g/cm3 for cost effectiveness. The
construction consists of two shells with holes to accommodate two 3” dia. NaI(Tl)
detectors. Cylindrical shielding from the same material mounted to the bottom shell
surrounds the crystal and the PMT. A plug from composite tungsten located between the
PMT and electronics is used for back shielding of the crystal. The electronics is
integrated in the unshielded part of the detector case. Preliminary background
measurements in the laboratory with a single detector configuration of the same design
(see Fig 2.a below) show about 1.2 cps count rate in the 186-keV region of interest.


Most of the deposits measurement work was done for the low pressure pipe headers
where CEMO and CHEM instruments were designed to operate and used high resolution
germanium detectors [14, 15]. The two-geometries method developed by Don Close [16]
uses a HPGe detector and a slit colimator. Two measurements are made along and
perpendicular to the pipe axis to distinguish between the signal from gas and from wall
deposits. Because these two measuremenst have to provide the corrected enrichment
results, the width and length of collimator slit were optimized as a tradeoff between
sensitivity to the signal from gas and sensitivity to deposits. The geometry optimization
is caried out with emphasis primarily on sensitivity to signal (both the 186-keV peak of
    U and the X-ray fluorescence lines from uranium) and secondarily on sensitivity to
deposits. Because only a fraction of the detector area was used, long measurement times
were encountered for both measurements.

In this work we explore a multidetector approach where both measurements are done
simultaneously with different detectors and geometries optimized for sensing the gas and
deposits. Our approach is based on low cost NaI(Tl) detectors: high length to diameter
lontitudinal detectors for deposit measurements and larger diameter detectors for gas
measurements. The longer measurement time required by the lower efficiency of the
deposits measuring detectors is allowed because the signal from the deposits changes
much more slowly than the signal from gas.

The Monte Carlo simulations were performed by means of the LANL MCNPX 2.7c
code. The detector was modeled in the inputs file of the code, and the Pulse-Height Tally
(F8) coupled with the SCX card was used to model the detector performance. In the
modeling of the gas pipe and the uranium deposits we considered two scenarios: the first
one with 1-µm thickness of the deposits on the pipe wall, and the second with 2-µm
thickness; in both cases the pressure of the UF6 gas was fixed at 50 Torr. The deposit was
modeled as a UF6 solid with density of a 5.0 g/cm3. An enrichment value of 3.3% for
both gas and deposits was used. Two basic geometries have been introduced to the model
of the pipe plated with deposits: a) one to study the sensitivity of gas sensing detectors
with different sizes mounted perpendicular to the pipe and b) one to study the sensitivity
of deposit-sensing longitudinal detectors mounted parallel to the pipe. We are reporting
preliminary modeling results used for our design concept. This is work in progress and
will be reported in detail at the upcoming INMM meeting [17].

Gas sensing detectors
The modeling for the gas-sensing configuration was based on an existing shielding of
composite material mounted on the vertical UF6 source without deposit plated on the pipe
surface (Fig. 2)

Fig. 2a. T-shielding with 3” NaI(Tl) detector        Fig. 2b. Corresponding MCNP model image. The
mounted on a vertical UF6 source. A pipe length of   deposit plated surface is shown in blue. The
500 mm was used for the simulations.                 composite shielding material is shown in
                                                     semitransparent gray.

The composite shielding material of tungsten powder with an organic binder was
simulated as tungsten with an average density of ~11 g/cm3. Three different sizes of NaI
scintillation detectors were introduced in the simulations: 2” x 0.5”, 3” x 0.5” and 5” x
0.5”, respectively. The distance between the front face of the detector and the pipe was
varied from 1 to 4 cm. The gamma-ray source (186-keV) was weighted proportionally to
the mass of uranium in the deposits and in the gas, respectively.
The detection efficiency scaled proportionally to the detector area without substantial
reduction in deposit sensitivity. Note that this study presents only the relative changes
between the detectors. The experimental data with 3” detector and 3.3% enriched UF6
gas at 50 Torr pressure [8] gave 28 cps for the intensity of the 186-keV peak. It is
expected to increase to 38 cps for a 4.5% enrichment used recently.

Deposit sensing geometry.
Unlike the gas sensing-detector with a large surface area and only a half inch thickness,
the deposit-sensing detector is a long cylinder parallel to the pipe. The 3D picture and
cross section of the model geometry are shown in Fig.3. The shielding material and pipe
configuration (size, UF6 gas and deposit density) are the same as for the gas-sensing

Fig 3a. 3D modeling geometry for deposit sensing. The     Fig 3b. Cross section of geometry in Fig. 4a.
1” dia., 2” long deposit sensing detector is situated
parallel to the pipe.

The collimator slot is tangential to the internal surface of the pipe in order to maximize
the wall deposit-to-gas signal. Our initial calculations were made for 1” diameter × 2”
long commercially available detectors used for Pu hold-up measurements. The calculated
efficiencies and ratio between gas and deposit signals for two collimator slit opening
sizes are presented in Fig. 4.

Fig 4a. Efficiency versus detector to pipe distance.   Fig 4b. Deposits to gas ratio versus detector to pipe

Note that the loss of total efficiency for a 0.5 cm slit is compensated for by a higher gas
to deposit ratio. Based on these initial results we designed an experimental multidetector
system with two gas-sensing and two deposit-sensing detectors. The 3D conceptual
drawing of this system is shown in Fig. 5. The design envisions two deposit-sensing
detectors with two collimator slits each that will increase the calculated efficiency by a
factor of four. The longtitudinal detectors and collimators are made as a drop-in inserts to
provide exprimental flexibility for optimizing the deposit-sensing configuration.
Additional doubling of detection efficiency could be acheved by installing two detectors,
face-to-face, on each side of the pipe. The 3” detectors could be collimated with slits
parallel to the pipe axis to optimize the gas-to-deposit ratio. This experimental protoytpe
will be calibrated in a laboratory environment using two UF6 gaseous sources and ready
for field trial on real pipe pipe with deposits. The knowlege gained by this research effort
will be used for designing application-specific attended assay methods for the deposit
calibration of the simpler unattended detection configuration presented in Fig. 1.

Fig. 5. Conceptual design of deposit-sensing detector configuration . The deposit-sensing detectors are 1”
diameter × 2” long detectors situated parallel on both sides of the pipe. Two tangential slits per detector are
used to increase the deposit detection efficiency. The whole construction is supported by two side flanges
clamped to the pipe.

Calibration setup
The calibration of the deposit-sensing detector system will be done in a laboratory
environment using two identical UF6 sources. We have built one source using 50 cm
long, 4” diameter pipe with a passivated surface [10,18]. We are building a second
identical one that will be plated with 3.3% LEU with a real density of 4 µgU/cm2. A
control sample introduced during passivation will be measured independently for
accurate determination of the deposit. Temperature-compensated pressure transducers
will be used to control the gas pressure. We will calibrate the deposit measurement head
using a set of measurements at different gas pressures and two known deposit values.
Different collimator geometries will be tested to determine the sensitivity and statistical
variations of the measurements. The set of two sources and the acquired experience will
be used for development of calibration procedures and actual calibration of other
enrichment monitoring instrumentation.

A concept for a passive enrichment monitor with redundant detection systems was
presented. A prototype of a deposit sensing detection system and calibration setup was
discussed. Our plans are to have the multidetector system built and calibrated in the
laboratory by the end of FY10. We are researching the opportunities for field trial tests on
a header pipe with real deposits.


The work presented here is primarily supported by the NA-22 Global Safeguards
Research and Development Program and is part of a larger program funded by NA-22
and the NA-241 Office of Dismantlement and Transparency.

We would like to acknowledge Alain Lebrun and Donald Close for fruitful technical
discussions on enrichment monitoring technology and deposit measurements.


   1. “Technical meeting on Techniques for IAEA Verification of Enrichment
      Activities”, STR-349 IAEA, Vienna, Austria, 2005.
   2. W. Bush et al., “Model Safeguards Approach for Gas Centrifuge Enrichment
      Plants,” IAEA-CN-148/98, Symposium on International Safeguards: Addressing
      Verification Challenges, International Atomic Energy Agency, Vienna, Austria,
      October 2006.
   3. A. Lebrun et al., “Improved Verification Methods for Safeguards Verifications at
      Enrichment Plants,” presented at the International Conference on Advancements
      in Nuclear Instrumentation Measurement Methods and their Applications
      (ANIMMA), Marseille, France, June 7-10, 2009.
   4. T. W. Packer and M. R. Wormald, “Continuous Monitoring of Variations in the
          U Enrichment of Uranium in the Header Pipework of a Centrifuge Enrichment
      Plant,” report, SRDR-R221 UK A00623, Sept. 1994.
   5. M. Hori et al., “Development of enrichment measurement technologies at the
      Ningyo-Toge uranium enrichment pilot plant”, IAEA-SM-293/36.
   6. D. A. Close, R. E. Anderson, W. S. Johnson Jr., R. M. Kandarian, P. L. Kerr, C.
      E. Moss, C. D. Romero, L. A. Trujillo, G. W. Webb, and C. R. Whitley,
      Calibration of the Enrichment Monitor for HEU Transparency, in Proceedings of
      the Sixth International Meeting on Facilities Operation—Safeguards Interface, p.
      111, Jackson Hole, WY, September 20-24, 1999.
   7. C. E. Moss, D. A. Close, K. D. Ianakiev, and T. Marks, “Improvements in
      Monitoring Enrichment of HEU Blend Down Under the U.S./Russian HEU
      Transparency Program,” proceedings of the 49th INMM Annual Meeting,
      Nashville, TN, 2008.
   8. K. Ianakiev et al., “Progress in Development of an Advanced Enrichment Monitor
      Based on Transmission Measurements with an X-ray Source and NaI(Tl)
      Spectrometer,” proceedings of the 50th INMM Annual Meeting, Tucson, AZ,
9. J. Goda et al., “Development of a Model of an X-ray Tube Transmission Source,”
    Conference Record of IEEE Nuclear Science Symposium, Orlando, FL, 2009.
10. M. T. Paffett, B. Nolen, T. Hill, C. Moss, and K. Ianakiev, “Vacuum and Gas
    Transfer Technology for the Non-destructive, In-line Determination of Isotopic
    Content in UF6 Process Gas,” proceedings of the 50th INMM Annual Meeting,
    Tucson, AZ, July 12-16, 2009.
11. P. Friend, Executive Summary, URENCO International Safeguard Conference,
    Chester, UK, December 2009.
12. A. Lebrun, “The Role of Instrumentation in Enhancing Safeguards Verifications
    at Enrichment Plants,” Proceedings of the URENCO Safeguards Conference
    2009, Chester, UK, December 2009.
13. B. Boyer, “Defining the Needs for Gas Centrifuge Enrichment Plants Advanced
    Safeguards”, to be reported at this meeting.
14. T. W. Parker et al., “ Monitoring the 235U Enrichment of Uranium deposited on
    pipes in centrifuge plants operated at low pressure using a high resolution gamma-
    ray spectrometer. AERA-Harwell report SRDP-R111, January 1984.
15. D. A. Close, J. C. Pratt, and H. F. Atwater, “Development of an enrichment
    measurement technique and its application to enrichment verification of gaseous
    UF6,” Nucl. Instr. and Meth. A vol. 240, pp. 398-405 (1985).
16. D. A. Close and J. C. Pratt, “Improvement of collimator design for verification of
    uranium enrichment in gaseous centrifuge header pipes of diameter 4.45 cm and
    10.16 cm,” Nucl. Instr. and Meth. A vol. 257, pp.406-411 (1987).
17. A. Favalli et al., “Monte Carlo modeling of a detector system for gas phase
    uranium enrichment measurements,” to be presented at the INMM 51st Annual
    Meeting, Baltimore, MD, July 11-15, 2010.
18. B. Nolen, M. Paffett, K. Ianakiev, T. Hill, and R. Schultz, “Memory Effects of
    Isotopic UF6 Adsorption and Reaction at Interior Metallic Surfaces of Interest to
    the Non-proliferation Community.” to be reported at the 51st INMM Annual
    Meeting, Baltimore, MD, July 11-15, 2010.

To top