ESARDA BULLETIN, No.

43, December 2009

Fingerprinting of Nuclear Material for Nuclear

Forensics
A. Redermeier
Vienna University of Technology, ATOMINSTITUT, Bahnhofstraße 46, 2231 Strasshof, Austria
E-mail: alice.redermeier@aon.at

Abstract formation is necessary. Great efforts are made to

set up databases containing such reference data.
Unknown nuclear material may originate from sev­
eral sources. Nuclear forensics allows by using fin­ Case development in nuclear forensics follows a
gerprinting and comparison with reference data to deductive way (Fig. 1). Taking the available results
determine the origin, the intended use, the last legal into account a hypothesis or a set of hypotheses
owner and the smuggling route. These information are built. Databases and other experts serve for the
are essential in the cause of theft or diversion as knowledge to build the hypothesis. Further analysis
measures of safeguards can be implemented to has to be made to prove the presence of signatures
prevent future thefts. according to the hypothesis. If the signatures are
absent in the sample a new hypothesis has to be
Certain measurable parameters can point to a spe­ developed.
cific material and provide therefore a ‘fingerprint’ of
the unknown material. Comparing the measured pa­
rameters with reference material give clues to the
origin and the last legal owner.
Characteristic parameters and possible information
they contain are presented.
Keywords: nuclear forensics; uranium; plutonium.

1. Introduction: What is nuclear forensics?

As the former Soviet Union disintegrated in the ear-
ly 1990s a new phenomenon was discovered – ‘nu-
clear smuggling’. The first cases were reported in
1991 in Switzerland and in Italy. New questions had
to be answered: intended use of the unknown nu-
clear material, its origin, the last legal owner and the
smuggling route. Techniques to answer these ques-
tions were known from nuclear safeguards and ma-
terial science. Combining these analytic methods
was the starting point of nuclear forensics.
Figure 1: The deductive way in the nuclear forensic
Information that is obtained by nuclear forensic process [1].
analysis can be divided into two groups: endogenic
and exogenic information.
2. The “nuclear fingerprint” method
Endogenic information is meant as self-explaining
information as age, intended use and mode of pro- To characterize an unknown nuclear material a set
duction, for which only model-calculation might be of items are measured to establish the so called nu-
required for data interpretation. clear fingerprint. Information that form the nuclear
fingerprint are:
Exogenic information is meant as information by
comparison. To interpretate exogenic information • Macroscopic parameters (e.g. pellet diameter,
which include geolocation and production data height)
comparison with reference data and reference in- • Isotopic composition

71
ESARDA BULLETIN, No. 43, December 2009

• Elemental composition (such as elements and

impurities)
In cases of a mixture of components, a powder or if
it may be possible that substances of different
chemicals or isotopic composition have been add-
ed, it is necessary to investigate the microstructu-
rual fingerprint which consists of following items:
• Particle morphology
• Particle size and size distribution
• Grain size and size distribution
• Porosity size distribution and density
Figure 2: Calculated isotopic composition of Pluto-
• Dislocation density and character nium for various reactor types (continuous line) and
• Precipitation of other phases. measured isotopic composition (points) [2].

2.1. Isotopic patterns of U and Pu 2.2. Age determination

The following information can be obtained by inves- The age is defined as the time elapsed since the
tigating the isotopic patterns: last chemical processing took place (e.g. produc-
• The presence of small amounts of U-236 will in- tion, reprocessing or purification). The age of the
dicate a contamination with recycled uranium material may serve as exclusion parameter in the
and hence point at reprocessing activities. search for the production or reprocessing plant.

• The isotopic composition of plutonium is a useful Since radioactive isotopes decay at a rate deter-
indicator of the reactor type in which the material mined by the initial amount and the half-life of the
was produced. isotope, the relative amounts of decay products
(daughters) in comparison to the parent isotope can
• Uranium oxide can be found in different forms, be used as chronometer.
e.g. UO2 or U3O8, which give information on vari-
ous points of origin in the uranium fuel cycle. The age of the material is short in comparison to
the half-life of the observed nuclides. Using the ta-
Plutonium is generated as by-product in nuclear re- ble of nuclides many parent/daughter pairs can be
actors when 238U absorbs a neutron creating 239U, found. The optimal nuclide ratios for Uranium are:
which β-decays into 239Np and finally to 239Pu. Also
heavier isotopes of Plutonium are produced by fur- • 234
U/230Th
ther neutron captures. Therefore the isotopic com- • 235
U/231Pa
position may give answers which reactor type was
used to produce the unknown nuclear material. and for Plutonium:

The isotopic composition of Plutonium provides in- • 238

Pu/234U
formation to indicate the type of the reactor: • 239
Pu/235U
• The higher the initial 235U enrichment of the fuel • 240
Pu/236U
is, the higher is the 238Pu abundance due to mul-
• 241
Pu/241Am.
tiple neutron capture on 237Np.
The radioactive decay of each of these nuclides is
• The neutron energy spectrum influences the Pu
unique, therefore measuring the parent/daughter
isotopic composition (the softer the spectrum,
ratio allows to calculate the time elapsed since the
the higher is the 242Pu/240Pu ratio).
last chemical separation. If the material has not
The measured isotopic patterns can be compared been fully separated, the chronometer hasn’t been
with model calculations using computer codes (e.g. set properly to zero and so misleading information
ORIGEN or SCALE for the RBMK and the VVER). It about the age is gained. To avoid this systematic
was demonstrated that the main reactor types can error it is recommended to measure various parent/
be separated clearly from each other (Fig. 2). daughter ratios.

72
 ESARDA BULLETIN, No. 43, December 2009

change takes place and the final U-oxide product

carries the signature of the 18O/16O ratio of the used
water. Measuring the oxygen isotope ratio provides
information about the geographical region where
the material was processed. Figure 4 shows the
correlation between the geographic location of the
production site of uranium oxid examples and the
variation in the (18O)/(16O) ratio.

Figure 3: The parent/daughter mass ratios of aged

plutonium material [3].

2.3. Metallic impurities

Nuclear material contains metallic impuritites at var-
ying concentrations. In the starting materials metal-
lic impurities are accompanying elements, during
processing to the intermediate and final product the
impurities are drastically reduced. A reduction fac-
Figure 4: Comparison of the 18O/16O ratio of U3O8
tor of 103 in the impuritity level is possible after
for different uranium mines. Olympic and Ranger
processing. Impurity patterns within the processing
are Australian mines [4].
remain for the most mines the same.
Otherwise metallic impurities may enter in the nu-
clear material at the different processing stages. The 2.4.2. Lead
systematic behind this is not well understood up to From the four stable lead isotopes one is primordial
now, but a full theory must contain that the concen- (natural) 204Pb and the other three are endproducts
tration of the impurities are a function of exposure
of radioactive decay series: 238U  206Pb, 235U 
time to the container material or storage tank as 207
Pb, 232Th  208Pb. Therefore the stable lead iso-
they are leached from the surface of the walls.
tope composition gives information on the initial
Nowadays in sample analysis the ratio of elements U/Th ratio in the mine and on the age of the ore.
of similar chemical behaviour are examined be- Due to the fact that the variations in the composi-
cause the ratio will vary only within narrow limits.
tion for different mines are significant, investigating
the stable lead isotopes can locate the origin mine.
2.4. Stable Isotopes
Measuring the stable isotopes is a established tech- 2.5. Anionic impurities
nique in geolocation. Two substances which are
chemically identical have different stable isotope Uranium crude ore undergoes different chemical
compositions if either their origin and/or their history processes before it becomes uranium ore concen-
are not the same. In nuclear forensics the oxygen and trate (yellow cake). Since uranium is mined from
the lead isotope ratio measurements are applied. ores with different mineralogical natures (acidic, al-
kaline and phosphatic ores) different chemical leach-
2.4.1. Oxygen ing processes are used. In the subsequent process-
Natural oxygen exists in three stable isotopes: 16O ing also different chemical processes are used for
(99.762%), 17O (0.038%) and 18O (0.200%). Tem- the precipitating and dissolving the uranium.
perature, latitude and distance to the sea cause Different acids are used for the different processes
variations up to 5% in the 18O/16O ratio. and may leave anionic impurities (Cl-, F-, Br-, NO2-,
Water is used as a common solvent in uranium NO3-, SO42- and PO43-). These anions may be an in-
processing. During the processing isotopic ex- dicator for the process used.

73
ESARDA BULLETIN, No. 43, December 2009

In practice anion ratios are used because the leach- • The 232U content does not only depend on the
ing rate in the mines can change and absolute anion operation mode of the production reactor but
concentrations differ more than the ratios. also on non-reactor-related history (i.e. length of
storage periods before and after irradiation).
It has been shown that anionic impurities can distin-
guish between different mines (Fig. 5), but there are • HEU may be enriched in a single cascade, but it
also differences between different sampling cam- is also possible to use several interconnected
paigns in the same mine (Ranger-old vs. Ranger- smaller cascades.
fresh and Beverly-old vs. Beverly-fresh in Fig. 5).
• The identification of the enrichment process
(gaseous diffusion vs. gas centrifuge) due to
small differences in the concentrations of the
trace uranium isotopes (232U, 234U and 236U) is
very challenging.

2.6.3. Blending
A special challenge to determine origin of nuclear
material arises if it is blended. In former times Rus-
sia blended the spent VVER fuel with spent fuel of
propulsion reactors, after reprocessing this nuclear
material is suitable for RBMK reactors [9].
Figure 5: Anion ratios in different uranium ore con- Therefore the nuclear material measured value of
centrate samples [5]. the 238Pu/tot.Pu vs. 242Pu/240Pu value will not fit the
simulated values (Specimen R1 in Fig.2).
2.6. Limitations of these techniques Sometimes however cross-contamination has oc-
curred before collection to disguise the origin.
2.6.1. Cross contamination Transmission electron microscopy (TEM) is used to
determine the grain-size distribution and can there-
There are two problems which may cause cross
fore indicate that a powder consists of different par-
contamination by investigating the stable lead iso-
ticle types.
topes. First, natural lead (204Pb) is omnipresent – so
special care has to be taken performing the chemi-
2.6.4. Computer codes
cal separation – second, lead is often used as
shielding material. Estimating the uncertainty of all quantities used in
the calculations (i.e. cross sections) and determin-
Presently the methods of the analysis are continu-
ing how these uncertainties propagate through the
ously improved. If the methods are sensitive enough
entire simulation is called uncertainty analysis.
to detect even metallic impurities from the spatula
used to collect the evidence, new problems for the Uncertainty analysis is not implemented in the ORI-
analysts are created. GEN2 and SCALE code.
An interesting overview how to determine the un-
2.6.2. Reprocessing and enrichment certainty in such computer codes is given in [10].
Presence of 236U can point to reprocessing activi-
ties. It has also been shown that in natural uranium 3. O
 verview of analytical techniques used in
variations in 236U and in 234U abundances occur. nuclear forensics
Better measurement methods have to be developed
Due to the tremendous fingerprint diversity and the
for 236U abundance levels close to natural abun-
requirement of high accuracies of measurement
dance.
many different analytic techniques have to be ap-
If nuclear material was reprocessed for non-peace- plied in nuclear forensics. Figure 6 gives a short
ful purposes the identification of this material is very overview of information gained from Uranium and
challenging. Additional information is necessary to Plutonium samples and which analytic techniques
calculate predictive signatures for enriched uranium are required to obtain this information. A short over-
produced from reprocessed uranium [8]: view of most widely used techniques is given in [2].

74
 ESARDA BULLETIN, No. 43, December 2009

3.1. Including traditional forensic methods

“Wherever he steps, whatever he touches, whatever
he leaves, even unconsciously, will serve as a silent
witness against him. Not only his fingerprints or his
footprints, but his hair, the fibers from his clothes,
the glass he breaks, the tool mark he leaves, the
paint he scratches, the blood or semen he deposits
or collects. All of these and more, bear mute wit­
ness against him. This is evidence that does not for­
get. It is not confused by the excitement of the mo­
ment. It is not absent because human witnesses
are. It is factual evidence. Physical evidence cannot
be wrong, it cannot perjure itself, it cannot be whol­
ly absent. Only human failure to find it, study and
Figure 7: Glove-box for handling of special nuclear
understand it, can diminish its value.” material [7].
Dr. Edmond Locard

This is the famous Locards Exchange Principle that 4. Data interpretation

states that every contact leaves a trace. Traditional Measured exogenic information (such as 18O/16O ra-
forensic methods establish relations between loca- tio, Pb isotopic composition, impurities and micro-
tions, events and individuals - entirely other informa- structure) needs comparison with reference data
tion than the nuclear forensic methods. In summary from known samples.
forensic methods provide information adherent to
the material while nuclear forensic methods provide 4.1. Supporting information
information inherent to special nuclear material [7].
It is important to have access to reference data and
A detailed description of collecting both nuclear to keep information, that may vary with time (e.g. as
and traditional forensic evidence is given in the Nu- seen in 2.5. level of impurities), up to date. The ref-
clear Forensics Support [1]. During the collection of erence information can be gained from Nuclear Ma-
evidence and analysing the evidence in a laboratory terials Database, Open Literature, ITDB, IAEA Re-
special attention has to be taken on sufficient pro- search Reactor Database, other databases and
tection against the radiation. comparison samples.
To fulfil good radiological safety practice the Insti- The Nuclear Materials Database was established in
tute for Transuranium Elements developed jointly collaboration between the ITU in Karlsruhe with the
with the German Federal Criminal Police a so called A. A. Bochvar Institute in Moscow. The database
glove-box, where contaminated evidence can be contains information on fuels for commercial reac-
visually inspected, photographed and fingerprints tors as collected from open literature and from bi-
and DNA samples can be taken (Fig. 7). lateral agreements with fuel suppliers.

Figure 6: Information that can be obtained from nuclear (U, Pu) material and used analytic techniques [6].

75
ESARDA BULLETIN, No. 43, December 2009

Certain data have limited accessibility (commercial- Since the beginning of nuclear forensics in the early
ly sensitive data such as chemical impurities) or are 1990s more and more parameters proved to be use-
not shared (detailed information on weapons grade ful and could be applied for the nuclear fingerprint.
material). Hence it is necessary to do furthermore research
and development and to keep close cooperation
4.2. Exclusion Principle with other sciences whether new characteristic pa-
rameters can be investigated.
In order to avoid a full characterization each time
when unknown nuclear material is found, the exclu-
sion principle is applied. The exclusion principle
6. Acknowledgement
works as follows: The author would like to thank the European Safe-
1. The first measured information (e.g. pellet dimen- guards Research and Development Association,
sions and isotopic composition) are compared Working Group on Training & Knowledge Manage-
with the database entries. ment for organising the 5th ESARDA course.

2. The non-matching records are rejected as they References

could not be manufacturer of the unknown nu-
[1] IAEA; Nuclear Security Series No 2, Nuclear Forensics Sup­
clear material. port, Vienna, Austria, 2006, http://www-pub.iaea.org/MTCD/
Exit condition: A single record is left. This is the publications/PDF/Pub1241_web.pdf
best case. [2] Mayer K. et al; Nuclear forensics—a methodology providing
clues on the origin of illicitly trafficked nuclear materials, Ana-
3. The remaining records, which match the meas- lyst, 2005, 130, 433–441
urements, are compared to each other in order
[3] Wallenius M., Mayer K: Age determination of plutonium mate-
to identify parameters which could distinguish rial in nuclear forensics by thermal ionisation mass spectrom-
between these records. etry, Fresenius J Anal Chem (2000) 366 :234–238
Exit condition: No further parameters can be in- [4] Advances in destructive and non-destructive analysis for envi­
vestigated. ronmental monitoring an nuclear forensics, Karlsruhe, Germa-
ny, 2002, http://www-pub.iaea.org/MTCD/publications/PDF/
4. These parameters are measured next. pub1169_web.pdf
5. The measured data are compared with the re- [5] Badaut V. et al; Anion analysis in uranium ore concentrates by
duced database entries. ion chromatography, Journal of Radioanalytical and Nuclear
Chemistry, Vol. 280, No.1 (2009) 57–61
Go back to 2).
[6] Wallenius M. et al; Nuclear forensic investigations: Two case
studies, Forensic Science International 156 (2000) 55-62
5. Conclusion
[7] Mayer K. et al; Nuclear forensic science – From cradle to ma­
Combining various analytic techniques can give in- turity, Journal of Alloys and Compunds 444-445 (2007) 50-56
formation about the origin, the intended use, the last [8] Joint Working Group of the American Physical Society and the
legal owner and the smuggling-route of unknown American Association for the Advancement of Science; Nu­
nuclear material. But this is only possible if enough clear Forensics Role, State of the Art, Program Needs

reference datas are available and accessible. [9] Wallenius M. et al; Origin determination of Plutonium material
in nuclear forensics, Journal of Radioanalytical and Nuclear
Therefore it is necessary to strengthen the interna- Chemistry, Vol. 246, No. 2 (2000) 317-321
tional cooperation and the cooperation with the fuel [10] NUREG/CR-5625, ORNL-6698; Technical support for a pro­
manufactors. posed decay heat guide using SASH2/ORIGEN-S data

76