Paper Number: IAEA-CN-184/262
Isotopic separation by laser based technologies: safeguards related aspects
Authors: Denys Rousseau, John Lepingwell, Division of Safeguards Operations A, International
Atomic Energy Agency
Email Address of Main Author: [email protected]
Abstract: The mass difference between isotopes of the same element generates slight variations of the
properties of their electronic cloud. This physical fact makes it possible to separate one isotope from the
others by bringing an appropriate amount of coherent light energy to its constituent atoms or molecules.
This laser material interaction is carried out in the three major known laser processes: atomic vapour laser
isotope separation (AVLIS), molecular laser isotope separation (MLIS), and selection of isotopes by laser
excitation (SILEX).Part of the required energy must be delivered in a selective manner to bring the selected
isotope to an appropriate level of excitation. Since the laser and optics technologies are currently being
developed in a very fast and efficient manner worldwide larger and larger varieties of laser lights are
therefore available. As a result the range of possible spectroscopic options to be used as efficient photoexcitation sequences becomes wider and wider. The technological options to generate the gas material
stream are mainly constrained by the throughput of an installation which changes a lot depending on the
targeted production: large amounts of low enriched uranium (LEU) production or lower amounts at higher
enrichment levels. From a safeguards point of view, the Additional Protocol provides strong tools to
monitor activities in this domain. In particular tracking R&D activities can be done by complementary
access based on the thorough review of all available information. However, there are serious obstacles to
identifying any undeclared activities in this field such as the obsolescence of the section of Annex II of the
AP addressing laser technologies, the development of more compact laser light sources allowing the
reduction of the size of integrated installations, the possibility of using depleted uranium (DU) for such
activities and increased research on laser isotopic separation of stable isotopes. Last but not least, there is no
physical hurdle to apply these three processes to actinides other than uranium.
1. Introduction
The rapid development of laser technologies poses new challenges in detecting and safeguarding activities
related to laser isotope separation (LIS). The diffusion of knowledge about advanced laser technologies and
optics, the development of new and improved laser systems and nonlinear optics, and the rapidly expanding
market for these technologies make tracking LIS-related developments increasing difficult. At the same
time, detailed knowledge of LIS-related technologies remains relatively limited, and may be decreasing
over time as funding for declared LIS research is reduced. To maintain and develop its ability to detect,
identify, and safeguard LIS-related activities the Agency must acquire and analyze a wide range of
information as well as conduct safeguards activities in a focused and information-driven manner.
2. Basic Principles of Laser Isotopic Separation
The mass difference between isotopes of the same element generates slight variations in the properties of
their electronic clouds. This physical fact makes it possible to separate one isotope from another by bringing
an appropriate amount of coherent light energy to the atoms or molecules. This photo-excitation process,
also named laser material interaction, is the first step in the laser isotope separation process, which is then
followed by the extraction and collection of the isotopically isolated material. This very general description
is common to all three laser processes which are known to have been, or are being, developed beyond the
1
laboratory scale: atomic vapour laser isotope separation (AVLIS), molecular laser isotope separation
(MLIS), and selection of isotopes by laser excitation (SILEX).1
The AVLIS process is carried out with atoms of a single element. The physics is based on the ionisation of
such atoms. The basic function of the light is therefore to remove one peripheral electron of the electronic
cloud. The ionisation energy is about 6.2 eV for uranium and many stable metallic isotopes. The chemical
form of the element does not change during the process
The MLIS process is carried out with molecules of a single element. The physics is based on the excitation
of such molecules with enough energy to “break” (or photo dissociate) the molecule. The basic function of
the light is therefore to increase the vibrational energy up to this rupture point. The required energy for such
a physical transformation is about 4-5 eV. Then the chemical form of the photo-dissociated product changes
during the process. In the MLIS process the product becomes UF5 while the feed and the tails remain as
UF6.
The SILEX process is also carried out with molecules of a single element. Regarding its application to UF6
it is subject to strong confidentiality arrangements. However it is known that the three main gas streams feed, product and tails - remain as UF6. This means that the physics is based on extra-molecular forces. This
implies and explains several characteristics:
•
The excitation mechanism remains identical to MLIS namely increasing the vibrational energy of
the molecule. Some authors are speaking about a condensation repression mechanism.2
•
The level of the required energy is lower than MLIS. This is clearly stated by the SILEX promoters
when they say that SILEX is less expensive than centrifuges and uses less energy.3
3. Structure of an LIS Research and Development Programme
Since there is no declared uranium LIS enrichment facility in operation today, it is proposed to limit the
approach to the activities which may be necessary to perform before designing and constructing such a
plant. Developing one of these LIS processes requires performing various R&D activities which can be
categorised in four main areas or sub-programmes: 1) Spectroscopy, 2) Laser technology, 3) Material
technology and 4) Integration.
3.1 Spectroscopy
The main purpose of a spectroscopy programme is to acquire the necessary knowledge of the basic physical
parameters governing the interaction between the laser and the isotopes to be separated. This includes the
energy and lifetime of excited levels, transitions and associated cross sections, light beam quality, and
selection of sequences. These activities can be accomplished with a very limited amount of material (for
example, 50-100 grams of uranium). The required process wavelengths have to be used but this can be done
at lower repetition rates (10-100 Hz) than those to be used in the final process. Installations dedicated to
spectroscopy studies are therefore incapable of producing more than a few milligrams of enriched material.
Furthermore, the laser technology used is not necessarily the same as would be used in a pilot installation.
The size of such a spectroscopy research installation could be very small (i.e. a single room) and the
research could be conducted by a small but well-qualified staff.
3.2 Laser Technology
Part of the required energy must be delivered in a selective manner to bring the selected isotope to an
appropriate level of excitation. This implies that part of the laser system has to be finely tuneable around the
selected transition level. Laser technologies and associated optical components are currently being
developed in a very fast and efficient manner worldwide. There are therefore more and more technological
solutions to provide larger and larger varieties of laser lights (wavelengths, energy/pulse, repetition rates,
beam profiles). As a result the range of spectroscopic options which are available to be used as efficient
photo-excitation sequences becomes wider and wider.
In the case of AVLIS the historical solution was based on tuneable dye lasers pumped by copper vapour
lasers (CVL). The purpose was to generate 3 or 4 wavelengths of about 2 eV each (~ 0.6 µm) at high
repetition rate (5 – 20 kHz). One of these colours was required to be selective. CVL lasers were not widely
used for other applications and therefore served as both a technological hurdle for those conducting LIS
research and an important indicator of interest in LIS technologies.
However, over the past decade the huge progress made by the yttrium aluminium garnet (YAG) laser
technology, both in repetition rate and mean power, make the CVL solution obsolete. In addition similar
progress in the upper harmonics generated by these lasers can provide solutions for ionising metallic atoms
in only two steps. Another example of recent developments is the use of Ti-Sapphire lasers (a tuneable solid
state laser) in a continuous mode to carry out the selective part of an ionisation process applied to stable
isotopes. Thus, as technology develops new avenues for LIS emerge, and the watch lists and indicators
developed in the past become outdated.
In the case of MLIS there is no doubt that the 16 µm (~ 0.08 eV) wavelength is specific to the excitation of
one of the vibration modes of the UF6 molecule. The historical solution was based on CO2 lasers and a
Raman cell shifter to generate the selective beam, together with ultra violet lasers, or in some cases nonselective IR lasers providing the additional non-selective energy required for multi photonic dissociation of
the UF6. Limitations on the repetition rates of CO2 lasers and reliability of UV lasers in the desired power
range have often been identified as major weaknesses for extended MLIS applications. Again however,
recent progress in the repetition rate and tuneability of CO2 lasers, free electron lasers in the IR range, non
linear optical technologies, and the development of semiconductor lasers offer new perspectives for MLIS
applications.
The details of the SILEX process have been kept strictly confidential. However, the basic characteristics
that are cited above allow one to assume that progress in MLIS laser technologies can benefit the SILEX
process as well. In such a case the lower energy requirements as well as the steady chemical phase of the
process material appear to be considerable advantages. In particular the low level of required energy implies
that no UV laser appears to be necessary.4
The laser technology development sub-programme needs the appropriate requirements and basic physical
data to be provided by the spectroscopic studies. It aims to solve laser and optical engineering problems
related to operating such laser systems at high repetition rates over several hundred hours. Such a
programme can be conducted without any process material being present. Therefore, it might not be
possible to determine whether a laser development programme is intended for LIS purposes even through
the use of environmental sampling at the location. Access to process-specific information such as
wavelengths, power, and repetition rate may be necessary to allow determination of the intended end-use.
3.3 Material Processing Technology
The technological solutions to generate the atomic or molecular gas stream to be illuminated are mainly
constrained by the throughput of the installation. Solutions which are not technically or economically
efficient for large amounts of LEU production can still be very efficient for a smaller quantity of material
that is to be enriched to higher levels.
In the case of AVLIS the known solution is based on vapourising material by an electronic beam (generated
by an electronic cathode/gun) impacting a metal ingot contained in a cooled crucible. While this solution is
unavoidable for producing LEU because of the requirement for very high throughputs, simpler solutions can
be employed for smaller scale applications such as laser ablation, joule furnace, or sputtering.5 Each
specific solution has advantages and inconveniences which should be developed on a case by case basis.
One of the main generic problems to solve is how to contain and to return to the crucible the large fraction
of the vapour generated which is not subject to the laser-material interaction process.
In the MLIS and SILEX cases molecular spectroscopy data implies that the molecule gas stream has to be
cooled to improve the isotopic shift. This is why the molecules are cooled by adiabatic expansion of a
multi-component gas stream. The core device is a slit nozzle into which the photo-interaction zone is
integrated. Molecular systems can also benefit from the use of available UF6 handling technology which can
be applied to this process without major difficulty.
All these solutions depend strongly on the process performances (enrichment factor, depletion factor,
separation factor) which thus determine the number of required stages to reach the desired level of
enrichment. One major goal of such a sub-programme is to operate a feed material source for several
hundred hours. It can be conducted with rather low amounts of material in the case of uranium (about half a
metric ton for AVLIS and some tens kilogrammes of UF6 for both molecular processes) with the
understanding that depleted uranium is fully useable for these activities.
3.4 Integration and Process Development
Until this stage in the analysis all of the steps outlined above could be applied to the separation of any
isotopes by laser processes, not just uranium isotopes.
The purpose of a pilot plant dedicated to uranium is to demonstrate the global process performances. It aims
to integrate the laser system and the nuclear part of the installation in which the core technology, namely the
photo interaction/extraction/collection zone, is validated experimentally. Significant results in the three subprogrammes above are needed in order to design, construct and operate such a plant with a high probability
of success.
An experimental AVLIS programme for stable isotopes is a very efficient tool for testing the overall
process since it can be a full scale demonstrator for an integrated pilot plant dedicated to nuclear material.
In particular many stable isotopes - gadolinium, ytterbium, thallium, others - have the same ionisation
energy as uranium. The performance of a laser system developed to separate these stable isotopes (mean
power, repetition rates, fluence, number of wavelengths, beam quality, operational tuneability and
flexibility) can be easily adapted to working with uranium. Such a stable isotope demonstrator could
therefore be reconfigured to work with uranium with a production rate similar to that for the stable isotope.
Similarly, MLIS research on the isotopic separation of SF6, MoF6 and Si2F6 can allow the development and
testing of technology and techniques for UF6/MLIS.6 For MLIS systems the transition to working with UF6,
however, would require the conversion of uranium to UF6 and developing experience in handling UF6
flows. This experience might be gained through existing conversion facilities or through the construction
and operation of an undeclared conversion facility. Clearly, the existence of such a facility would be a
strong indicator of LIS-related activities.
4. Safeguards Considerations
Given the above considerations, the Agency must further elaborate its strategy for detecting and
safeguarding LIS activities. This strategy must take into account the following considerations:

The Agency’s experience in this field is very limited since there is no declared LIS facility under
safeguards which is currently in operation.

The four sub-programmes described above can be conducted at different locations. There is no
functional link to be preserved between the dedicated installations until the integration work has
started.

The comparatively small scale and relative absence of external indicators makes the remote
detection of LIS-related activities difficult.

For some processes (i.e. AVLIS) the relevant technologies may not be familiar to inspectors and
uranium material handling processes may not be evident.
These considerations mean that the Agency’s strategy depends heavily on the acquisition and analysis of
relevant information, as well as field activities. However, the Agency’s access to information and locations
depends on the safeguards agreement with the State.
4.1 Comprehensive Safeguards Agreement (CSA) Based on INFCIRC/153
According to INFCIRC 153, “safeguards will be applied… on all source or special fissionable material in
all peaceful nuclear activities…” However, the amount of information available for determining that there
are no undeclared activities or nuclear material is limited under INFCIRC/153, so the Agency does not draw
a conclusion regarding the absence of undeclared activities or nuclear material. Instead, it draws a
conclusion on whether the declared material remains in peaceful use. Under an INFCIRC/153 type CSA,
the following considerations are relevant to laser isotope separation activities:

INFCIRC/153 does not have any provisions requiring the declaration of nuclear fuel cycle-related
R&D which does not involve nuclear material. Therefore, safeguards activities must focus on those
locations where nuclear material is present.

Undeclared activities related to LIS on nuclear sites would be difficult to detect given the inspection
and access provisions of INFCIRC/153, as would LIS activities outside of such sites.

The initial activities (spectroscopy, material processing) involving nuclear material could be
conducted in locations outside facilities (LOFs) since the amount of nuclear material used is below
one effective kilogram. Actual isotope separation activity, or the conversion of feed material, would
require that the location be declared as a facility (an isotope separation plant).

In the case of LOFs Articles 49 and 50 of INFCIRC/153 are applicable. Under current safeguards
approaches the Agency does not inspect all LOFs on an annual basis, and the mean time between
two PIVs performed at the same LOF may exceed several years, depending of the amount of
nuclear material present.

In the case of LOFs, the provision of a Design Information Questionnaire (DIQ) providing
information in addition to that required by articles 49 and 50 is not mandatory. As a result the
Agency would have no information about laser systems in the LOF, even though such a system
would be included in the DIQ as essential equipment if the location were instead categorised as a
facility.

In theory a full programme with the production of up to one significant quantity (SQ) of high
enriched uranium (HEU) could be conducted with less than one effective kilogram of depleted
uranium. Clearly, however, such an activity would require declaration as a facility, since it would
qualify as an isotope production plant under Article 106 of INFCIRC/153.

Diversion of less than 1 SQ of depleted uranium for use in an undeclared LIS programme becomes
a credible scenario in states with advanced LIS-related activities.

If a LIS research progresses to the point at which isotope separation is conducted and it is declared
as a facility, the design information questionnaire (DIQ) must be completed and submitted by the
facility operator. The laser system must be defined as essential equipment in the DIQ and key laser
performance parameters provided as well. The associated design information verification (DIV)
activities become of paramount importance in the timely detection of any misuse of the facility
and/or any undeclared production.
INFCIRC/153 presents therefore some limitations as far as detecting undeclared LIS activities is concerned.
Under such an agreement the general objective of the Agency’s verifications is to conclude that all declared
nuclear material remain in peaceful activities. The level of confidence in the safeguards conclusion for a
state with extensive LIS research and large nuclear material (including depleted uranium) inventories may
therefore be restricted by concerns over diversion of nuclear material.
4.2 Additional Protocol (AP)
With both a CSA and an AP in force the Agency is able to draw a conclusion on the absence of any
undeclared nuclear material or activities at the State level. The AP provides strong tools to monitor
activities in the nuclear domain. Tracking nuclear fuel cycle-related R&D activities can be done in
particular through thorough analysis of declarations about activities not involving nuclear material, review
of open source information, and complementary accesses (CAs) based on the review of this information.
Nuclear fuel cycle-related R&D activities are a major part of the AP scope. The articles 2.a(i), 2.a(iv),
2.a(x) and 2.b are of particular importance since they entail declaration of nuclear fuel-cycle related
research and development, manufacture or assembly of laser-based systems, and general plans for the
succeeding ten-year period relevant to the development of the nuclear fuel cycle. All the activities described
in section 3 above are therefore subject to declaration since they are fuel cycle-related, with the exception of
activities exclusively involving stable isotopes.
Under the strengthened safeguards system the Agency also gathers a wide range of additional information
in order to detect research in technologies related to LIS. This information includes scientific and technical
literature which provides substantial insights into the State’s technological capabilities, scientific expertise,
and research programmes on relevant laser and optical technologies. Additional information may also be
acquired by reviewing trade activities in key technologies through export/import information, additional
protocol declarations (2.a.(ix)), and analysis of the State’s infrastructure in relevant areas. To effectively
utilize this information, requires both the ability to efficiently collect it and to conduct in-depth technical
analysis. Thus, technical expertise is an essential element in the evaluation process. The Agency does have
this expertise, but given that research in LIS technologies has been dwindling over the past decade, the
availability of this expertise may decline over time. Furthermore, the technical experts must be familiar
with both the past solutions and the new options being opened up by the rapid development of technology.
Taken together, these information sources can provide guidance to focus complementary accesses on
locations where LIS-related activities are most likely. This information-driven focus is critically important
given the small size and difficulty of detecting LIS activities. Complementary access may be carried out
based on the applicability of different articles of the AP. For research and development activities, the
request for clarifications mechanism provided by articles 2.c and 4.d is a strong tool to investigate possible
undeclared activities in locations other than sites and LOFs.
Despite the additional information and access rights provided by the AP, serious obstacles to identifying
any undeclared activities in this field remain:

Annex II of the AP addressing laser technologies requires significant updating. In particular, section
5.7.13 enumerates the lasers systems based on historical processes which are obsolete today. Fifty
years after being discovered laser technologies are currently being developed worldwide in an
impressive manner. In particular laser development activities conducted in the scope of the inertial
confinement fusion research are now relevant since many associated technologies can be directly
transposed to LIS applications. Annex II of INFCIRC/540 is based on the Part 1 list (the so-called
“trigger list” published in Revision 2 of INFCIRC/254, in 1995, and therefore reflects the laser
systems then used in LIS. The current revision of INFCIRC/254 Part I is Revision 9 of 2007, which
broadens this reference to include laser systems specified in INFCIRC/254/Rev.8/Part 2, which
includes a number of different types of lasers.

The progress in laser technologies allows the size of integrated installations dedicated to R&D
and/or capable of producing 1 or 2 SQs of HEU per year to be reduced significantly. In this respect
it is necessary not to forget the optical fibre technologies. They facilitate the transportation of laser
beams across significant distances without affecting their properties. The result is that the LIS
process material building does not need being adjacent to the building dedicated to the laser
systems.

Depleted uranium (DU) is a useful and efficient material for such activities. The current Agency
practice is to exempt such materials and this approach remains fully justified. However special
effort to verify the non-nuclear use of exempted DU should be undertaken if there are indications of
potential use in a LIS-related programme.

Activities on the isotopic separation of stable isotopes, AVLIS based in particular, may not be
subject to declaration under the AP, even though such integrated installations could be rather easily
adapted to uranium.

If Annex II is revised in the future, consideration should be given to an extensive modernization of
the section on LIS to incorporate new technologies. One possible approach would be to specify
laser parameters, rather than the lasing media, so as to incorporate future technologies as well.
INFCIRC/254/Rev.8/Part 2 already partially incorporates this output-based approach. Any such
revision should also incorporate a clear mechanism for declaring the LIS of stable isotopes.
It must be noted that the Agency’s interest in investigating potential LIS-related activities should not be
taken as evidence of the existence of such a programme. However, in States with active laser research
programmes the ability to distinguish between LIS-related and other activities requires extensive
information evaluation and in many cases access to locations as well. This access, in turn, is essential to
being able to draw, and to re-affirm, the conclusion that all nuclear material remained in peaceful activities
in the State.
5. Conclusions
1/ Under a comprehensive safeguards agreement the Agency does not have all the necessary verification
tools to conclude with high level of confidence that all nuclear material remains in peaceful activities.
While LIS-related activities and the possible use of undeclared nuclear material may be investigated by the
Agency, both information and access are limited. In some cases this may reduce the confidence in the
safeguards conclusion for a State.
With an AP in force the situation changes since it provides a number of appropriate tools. The additional
information and access allow a more complete and thorough investigation of this part of the nuclear fuel
cycle at the State level. However, obstacles to identifying any undeclared activities remain and might
potentially affect in a significant manner the ability to draw the broader safeguards conclusion that all
nuclear material remained in peaceful activities.
2/ Last but not least, there is no physical hurdle to apply these three processes to actinides other than
uranium. In particular LIS can be applied to plutonium.7 From a process point of view the required
enrichment and stripping factors are not as high as those required for enriching uranium by laser. This
relaxation of the process requirements is, however, balanced by plutonium handling difficulties. Very often
plutonium handling devices are subject to experimental qualification with mock up devices before being
operated with plutonium. That could be an additional purpose for a LIS sub-programme applied to stable
isotopes.
In addition, such considerations about LIS applied to plutonium are relevant not only in the framework of
safeguards but also in any possible future international treaty limiting fissile materials, such as the proposed
Fissile Material Cut off Treaty (FMCT). The potential use of LIS to enrich “reactor grade” plutonium in
the 239Pu isotope, thereby making it more effective for weapons, would pose a significant technical
challenge to an FMCT verification system.8
References
[1] C. Schwab, A.J. Damiao, C.A.B. Silveira, J.W. Neri, M.G. Destro, N.A.S. Rodrigues, R. Riva, “Laser Techniques
Applied to Isotope Separation of Uranium,” Progress in Nuclear Energy, Vol. 33, No. 1-2, pp. 217-264, 1998 provides
a good overview of laser-based isotope separation. On Silex see John L. Lyman, “Enrichment Separative Capacity for
SILEX,” Los Alamos National Laboratory, LA-UR-05-3786, 2005.
[2] J.W. Eerkens and J. Kim, "Isotope Separation by Selective Laser-Assisted Repression of Condensation in
Supersonic Free Jets," American Institute of Chemical Engineers (AiChE) Journal, September 2010, vol. 56, no. 9, pp.
2331-2337.
[3] Elaine M. Grossman, "New Technology Offers Detectable 'Signatures,' Advocate Says,"Global Security Newswire,
2 August 2010, http://gsn.nti.org.
[4 ] See Eerkens and Kim, and Lyman.
[5] For example, J.B. De Matos, Nicolau A.S. Rodrigues, "Neutral Atomic Jet Produced by Laser Ablation for Isotopic
Separation Studies," International Workshop on Separation Phenomena in Liquids and Gases, 13-18 June, 2010, Saint
Petersburg, Russia.
[6] V. Yu. Baranov, Yu. A. Kolesnikov, A.A. Kotov, “Laser Photolysis of UF6 Molecules,” Quantum Electronics,
1999, 29 (8) 653-666.
[7] E. Ng, B. Pedrotti, M. DeMicco, “Close Out of the SIS Program,”APT Semiannual Report, March-December 1991,
pp. 1-6, http://www.osti.gov/.
[8] A ban on the enrichment of plutonium was included in the draft FMCT proposed by the United States in May
2006. “Draft Treaty on the Cessation of Production of Fissile Material for use in Nuclear Weapons or Other Nuclear
Explosive Devices,” Conference on Disarmament Working Paper, CD/1777, 19 May 2006.