Download as pdf or txt
Download as pdf or txt
You are on page 1of 84

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343524505

ACTIVITY CHARACTERIZATION STUDIES IN FIR 1 TRIGA RESEARCH REACTOR


DECOMMISSIONING PROJECT

Thesis · August 2020

CITATIONS READS

3 152

1 author:

Antti Räty
VTT Technical Research Centre of Finland
19 PUBLICATIONS   52 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

TRIGA FiR 1 decommissioning View project

DEMONI - DEcommisioning Material characterizatiON and final dIsposal studies View project

All content following this page was uploaded by Antti Räty on 08 August 2020.

The user has requested enhancement of the downloaded file.


University of Helsinki Doctoral school in natural sciences dissertation series

ACTIVITY CHARACTERIZATION STUDIES


IN FIR 1 TRIGA RESEARCH REACTOR
DECOMMISSIONING PROJECT

Antti Räty
Department of Physics
University of Helsinki
Helsinki, Finland

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Science of


the University of Helsinki, for public examination in auditorium 1 of Metsätalo,
Unioninkatu 40, Helsinki, on 7th August 2020, at 12 o’clock.

Helsinki 2020

1
ISBN 978-951-51-6050-8 (nid.)
ISBN 978-951-51-6051-5 (PDF)
ISSN 2669-882X (printed)
ISSN 2670-2010 (PDF)

Helsinki 2020
Hansaprint Oy

2
ABSTRACT

During normal operation of a nuclear reactor neutrons cause fission reactions


in the uranium fuel producing energy. In addition, some of the neutrons from
reactor core are also absorbed into the construction materials close to the core
causing activating nuclear reactions in these materials. Radioactive
characterization is the process of determining the radiological properties of
reactor fuel and activated structural materials. This information is critical to
enable safe and cost-efficient dismantling and waste management plan.
This study consists of the activity characterization work carried out in the FiR1
TRIGA research reactor decommissioning project during the years 2014-2019.
Both reactor fuel and construction materials were studied. The thesis also
describes their essential connections to other aspects in a decommissioning
project.
The approach for the decommissioning waste was to build a three
dimensional neutron transport calculation model of the reactor structures to
calculate the neutron fluxes during reactor operation and to combine this data
with reactor operation history and material-specific activating impurity
concentrations. Conservative assumptions have been used especially
regarding material compositions and operational history modelling
simplifications to provide slightly overestimated activities. Decommissioning
waste from the FiR 1 reactor mainly consists of concrete, aluminium, steel and
graphite. Computational results give an estimate of the activated region
around the reactor core. Measurements from active samples support the
calculated results and indicate that the calculations have been conservative as
intended.
Burnup calculations of spent fuel elements were performed with two
separate calculation models. The results demonstrate the effect of different
approximations on nuclide inventories calculated with two methods.
Comparison with measured dose rates provided reasonable agreement for
practical fuel removal and transportation planning.
Main sources of uncertainty are related to approximations in detailed
material composition and simplifications in modelling the complicated reactor
operation history. Some uncertainties, especially with respect to spent nuclear
fuel, are also related to nuclear cross section data and methods in different
calculation codes.

3
ACKNOWLEDGEMENTS

The scientific research reported in this thesis has been performed in the FiR1
research reactor decommissioning project and many of the results contribute
to the regulatory licensing and technical decommissioning planning of the
reactor. I would like to thank research team leader Petri Kotiluoto and the
decommissioning project manager Markus Airila and for the opportunity to
extend my work to academic level. Their patience and determination set an
example to many.

This thesis has been completed in MATRENA doctoral thesis programme at


the University of Helsinki department of physics. I thank my supervisor Jyrki
Räisänen for the flexibility to combine my PhD studies with professional
research work outside the university.

In addition to basic research education at the University of Helsinki, I have


received my training to many of the scientific tools used in this project in
several SAFIR and KYT projects throughout the years I have worked at the
VTT. Of this I would like to express my gratitude to Finnish Nuclear Waste
Management Fund and all the collaborating organizations in various
projects.

I would like to thank all my co-authors both in Finland and abroad. Especially
the staff at the FiR1 research reactor: Anumaija Leskinen, Tommi Kekki, Olli
Vilkamo, Merja Tanhua-Tyrkkö, Iiro Auterinen, Jori Helin and Perttu Kivelä
for their knowledge and contribution to this project.
Nuclear decommissioning is a multidisciplinary challenge and my research
also benefitted greatly from the collaboration with subcontractors and
assisting staff involved in sampling and other preparatory measures.

During this thesis work I have had the privilege to interact with people at
various societies, especially Karjalainen Osakunta, Karjalainen Nuorisoliitto,
Seniores Carelienses, and Suomen Atomiteknillinen Seura. Being able to
discuss matters also not related to work has contributed greatly to finding my
place in academic community and keeping my motivation during all these
years.

Last but not least I thank my family, relatives and friends for all the support.

4
CONTENTS

Abstract....................................................................................................... 3

Acknowledgements .................................................................................... 4

Contents...................................................................................................... 5

List of original publications ....................................................................... 7

Author’s contribution in the publications.................................................. 9

Abbreviations ............................................................................................11

1 Introduction ..................................................................................... 13

2 Decommissioning and FiR 1 project ............................................... 15

2.1 Project outline and regulatory framework ............................. 15

2.2 FiR 1 reactor ............................................................................ 16

3 Physical and calculational background ........................................... 18

3.1 Radioactivity and activating nuclear reactions ..................... 18

3.2 Main nuclides in construction materials ................................ 21

3.3 Calculation method .................................................................22

3.4 Characterization by the scaling matrix method .....................24

3.5 Activity measurement methods ..............................................26

3.5.1 Key nuclide measurements ..................................................26

3.5.2 Difficult to measure nuclide measurements ....................... 27

3.5.3 Impurity analysis techniques .............................................. 28

4 Activity inventory calculations in FiR 1 project ............................. 30

4.1 Modelling the operating history ............................................ 30

4.2 Neutron transport calculation ................................................ 31

4.3 Point-depletion calculation ....................................................34

4.4 Validation and uncertainties ..................................................39

5
5 Material-spesific activities in FiR 1 decommissioning waste ......... 41

5.1 Concrete .................................................................................. 41

5.2 Steel ........................................................................................ 42

5.3 Aluminium ............................................................................. 44

5.4 AGOT graphite ....................................................................... 46

5.5 Fluental .................................................................................. 49

5.6 Lead ........................................................................................ 50

5.7 Lithiated plastics .....................................................................52

6 Spent research reactor fuel inventories...........................................53

6.1 Description of TRIGA fuel ......................................................53

6.2 Inventory calculations.............................................................54

6.3 Time evolution of nuclides in spent nuclear fuel ................... 57

6.4 Spent nuclear fuel dose rate calculations .............................. 58

7 Applying characterization results in the decommissioning project 61

7.1 Prelimenary dismantling plan ................................................ 61

7.2 Dose rates and radiation safety planning .............................. 62

7.3 Waste management and classification ...................................65

7.4 Clearance of materials............................................................ 66

7.5 Final disposal ......................................................................... 68

8 Summary and conclusions............................................................... 71

References ................................................................................................ 74

6
LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following four publications:

I A. Räty and P. Kotiluoto, FiR 1 TRIGA Activity Inventories for


Decommissioning Planning, Nuclear Technology, Volume 194, Issue 1, April
2016

II A. Räty, T. Kekki, M. Tanhua-Tyrkkö, T. Lavonen, E. Myllykylä, Preliminary


Waste Characterisation Measurements in FiR 1 TRIGA Research Reactor
Decommissioning Project, Nuclear Technology, Volume 203, Issue 2, August
2018

III A. Räty, S. Häkkinen and P. Kotiluoto, Nuclide Inventory of FiR 1 TRIGA


research reactor fuel, Annals of Nuclear Energy, Volume 141, 2020

IV A. Räty, T. Lavonen, A. Leskinen, J. Likonen, C. Postolache, V. Fugaru, G.


Bubueanu, C. Lungu and A. Bucsa, Characterization measurements of fluental
and graphite in FiR 1 TRIGA research reactor decommissioning waste, Nuclear
Engineering and Design, Volume 353, November 2019

The publications are referred to in the text by their roman numerals.

7
8
AUTHOR’S CONTRIBUTION IN THE
PUBLICATIONS

This thesis is based on the following four publications:

I A. Räty and P. Kotiluoto, FiR 1 TRIGA Activity Inventories for


Decommissioning Planning, Nuclear Technology, Vol. 194, pp.
28-38, April 2016

The article consists of a calculation model used to estimate the activities in the
construction materials of the FiR 1 research reactor.
Author’s contribution has been in creating the point-depletion model to
calculate the activating nuclear reactions throughout the reactor operating
history, compiling the theory and the numerical results and interpretation of
the data.

II A. Räty, T. Kekki, M. Tanhua-Tyrkkö, T. Lavonen, E. Myllykylä,


Preliminary Waste Characterisation Measurements in FiR 1
TRIGA Research Reactor Decommissioning Project, Nuclear
Technology, Vol. 203, Issue 2, August 2018

The article consists of material-specific composition and activity


measurements to validate earlier activity inventory calculations reported in
publication I.
Author’s contribution has been in compiling the theory of the scaling matrix
approach, designing the performed measurement setups based on e.g.
knowledge of material-specific activating impurities, compiling the results and
performing comparison with calculated data.

III A. Räty, S. Häkkinen and P. Kotiluoto, Nuclide Inventory of FiR 1


TRIGA Research Reactor Fuel, Annals of Nuclear Energy, Vol. 141,
2020

The article consists of TRIGA fuel burnup calculations with two separate
calculation codes and a fuel rod dose rate calculation model performed with
Serpent code. The author is responsible for collecting the input data,
performing all ORIGEN-S burnup calculations, and compiling the results from
both of the codes. The author created the Serpent dose calculation model,
performed the calculations, and compiled the data presented in this article.

9
IV A. Räty, T. Lavonen, A. Leskinen, J. Likonen, C. Postolache, V.
Fugaru, G. Bubueanu, C. Lungu and A. Bucsa, Characterization
measurements of fluental and graphite in FiR 1 TRIGA research
reactor decommissioning waste, Nuclear Engineering and Design,
Vol. 353, November 2019

The article consists of material-specific composition and activity


measurements of the two special reactor materials that are especially relevant
in waste final disposal. The article partially completes the measurements
performed in article II.
Author’s contribution has been in compiling the theory and phenomena
relevant to designing the experiments, compiling the results, performing
comparison with calculated data and interpretation of the results.

The publications are referred to in the text by their roman numerals.

10
ABBREVIATIONS

ADC Analogue to digital converter


AGOT Acheson Graphite Ordinary Temperature
BNCT Boron neutron capture therapy
BNG Babcock Noell GmbH
DOE Department of Energy (United States)
DTM Difficult to measure
ENDF Evaluated Nuclear Data File
ETM Easy to measure
FiR 1 Finland Reactor 1
HPGe High purity germanium
IAEA International Atomic Energy Agency
ICP Inductively coupled plasma
IFIN-HH Horia Hulubei National Institute for Physics and Nuclear
Engineering
INL Idaho National Laboratory
ISOCS In situ object counting system
JAEA Japan Atomic Energy Agency
KRR-1 Korean Research Reactor 1
LANL Los Alamos National Laboratory
LSC Liquid scintillation counting
MCA Multichannel analyser
MCNP Monte Carlo N-particle
MHH Medizinische Hochschule Hannover
MS Mass spectroscopy
NIREX Nuclear Industry Radioactive Waste Executive
NIST National Institute of Standards and Technology
OECD-NEA Organisation for Economic Co-operation and Development
Nuclear Energy Agency
OES Optical emission spectroscopy
ORIGEN Oak Ridge Isotope GENeration
ppb Parts per billion (concentration unit)
ppm Part per million (concentration unit)
ppt Parts per trillion (concentration unit)
SAR Safety Assessments Report
SCALE Modular Code System for Performing Standardized Computer
Analyses for Licensing Evaluation
SS Stainless steel
STUK Säteilyturvakeskus (Radiation and Nuclear Safety Authority of
Finland)
TDCR Triple-to-double coincidence ratio

11
TEM Työ ja elinkeinoministeriö (Ministry of Economic Affairs and
Employment of Finland)
TRIGA Training, Research, Isotopes, General Atomics
TVO Teollisuuden Voima Oyj
VTT Teknologian tutkimuskeskus VTT Oy (VTT Technical Research
Centre of Finland)
WAC Waste acceptance criteria
YVL guide Ydinturvallisuusohje (Regulatory Guide on nuclear safety and
security)
γ Fission yield of isotope j
Σ Macroscopic fission cross section
𝛷 Integral neutron flux
𝜆 Decay constants
(→ )
𝜎 Transmutation cross section for the production of isotope i into
isotope j
𝜎 Neutron absorption cross section
𝑇/ Nuclide half life
𝜏 Nuclide mean lifetime

12
1 INTRODUCTION

Nuclear decommissioning is a planned activity at the end of life of facilities


that have a regulatory licence to conduct nuclear activities. It consists of all
activities needed to remove from the facilities from regulatory control and to
release the site for limited (e.g. only industrial) or unrestricted use [Chatzis,
2016]. This includes both technological (decontamination and dismantling
techniques etc.) and organizational (preliminary and detailed planning,
training, cost estimation etc.) activities. Nuclear decommissioning differs from
normal industrial decommissioning mainly because of the radioactive and
contaminated systems, structures and components in the facility.

During the normal operation of a nuclear reactor, some of the neutrons from
the reactor core are absorbed into the construction materials surrounding the
core causing activating nuclear reactions in certain elements of these
materials. Radioactive characterization is the process of determining the
radioactive properties of the facility materials. The objective is to provide data
on the quantity and type of radionuclides, their distribution, and their physical
and chemical states. This process is illustrated in Figure 1 [OECD-NEA, 2014]

Figure 1: Graphical sectioning of a characterization process in a nuclear decommissioning project.


[OECD-NEA, 2014]

An estimate of the amount and type of radioactivity in a nuclear facility is


required to enable safety and cost-efficiency in actual dismantling and waste
management. Practically the data is used e.g. in determining factors such as
the need for decontamination, shielding or remotely operated equipment,

13
Introduction

waste management and disposal, and minimizing radiation exposure of


personnel.
The aim of this thesis is to present the main methods and results of
radiological characterization work conducted at VTT Technical Research
Centre of Finland Ltd in the FiR 1 TRIGA research reactor decommissioning
project during the years 2014-2019.
Chapter 2 describes the background of the FiR 1 TRIGA research and the on-
going decommissioning project. Chapter 3 introduces main concepts of
radioactivity, the basis of the activity calculations and physical background of
the measurement methods used in this study. Chapter 4 describes the activity
inventory calculations performed in Publication I to estimate the amount of
radionuclides produced in the reactor structural materials during the whole
operation history. Chapter 5 focuses on describing the activity measurements
in the characterization work. These are used both in validating the calculated
activity data and in future waste management in the dismantling phase. The
chapters also summarize the main results of Publications II and IV. Spent
nuclear fuel burnup calculations are modelled in Publication III with a slightly
different operating history model using two separate codes. This is presented
in chapter 6 along with a comparison with measured dose rates.
Chapter 7 describes how activity characterization is utilized in other parts of
the FiR 1 decommissioning project, mainly waste management, radiation
safety procedures and dismantling methods. Finally, a brief summary and
conclusions are given.

14
2 DECOMMISSIONING AND FIR 1
PROJECT

2.1 PROJECT OUTLINE AND REGULATORY


FRAMEWORK
The FiR 1 research reactor was shut down in June 2015 and is planned to be
dismantled as soon as is technically and legally possible. The project road map
is presented in Figure 2.

Figure 2: Project outline. Figure by VTT.

The spent nuclear fuel from the FiR 1 is intended to either send back to United
States, the country of its origin [IAEA, 2006a], or to be stored in underground
final disposal repository in Finland. All the other decommissioning waste will
be stored in either one of the underground final disposal sites in Loviisa or
Olkiluoto, Finland.
The FiR 1 research reactor, and other nuclear facilities in Finland, are
regulated by both the Nuclear Energy Act and the Radiation Act. Prior to
starting the actual dismantling, section 20 of the Nuclear Energy Act requires
a decommissioning licence to assess the technical solutions for dismantling
and waste management [Finlex, 1987]. Two years after the final shutdown,
VTT submitted to the Finnish Government an application for
decommissioning the FiR 1 research reactor1. The contents of the application
follow the requirements of the Nuclear Energy Decree [Finlex, 1988] as

1 In the 2018 amendment of the Nuclear Energy Act, Section 20a Licensing - Decommissioning of a
nuclear facility was introduced. However, as VTT submitted the application prior to the amendment, the
applied license is formally an operating license (for decommissioning) as defined in Section 20.

15
Decommissioning and FiR 1 project

described in Figure 3 [Airila et al., 2019]. The research work in this thesis sums
up the main radiological characterization results applied also to the
decommissioning license application.

Figure 3: Structure of the licensing documentation for FiR 1, following the Finnish Nuclear Energy
Decree (document names simplified). [Airila et al., 2019]

2.2 FIR 1 REACTOR


FiR 1 is a 250 kW TRIGA Mark II type open pool reactor from General Atomics.
It was in operation between 1962 and 2015. The reactor is entirely above
ground and is surrounded by a concrete shield, as shown in Figure 4. The
thermal reactor power was upgraded from original 100 kW to 250 kW in 1967.
The reactor core and reflector assembly (Figure 4) are located near the
bottom of an aluminium tank 6.4 metres deep and 2.0 metres in diameter.
The FiR 1 reactor has four beam tubes extending from the reflector
assembly through the water and concrete to the outer face of the shield
structure. These beam tubes have been used, for instance, for material
research with neutrons during the first decades of the FiR 1 reactor’s
operational history. In the late 1980’s, the beam tubes were plugged
Originally, the reactor had a 1.2 x 1.2 x 1.7 metre graphite thermal column
extending from the outer surface of the reflector assembly and penetrating the
reactor tank and shield structure (see Figure 4). In 1995-1996 the thermal
column was replaced by an epithermal boron neutron capture therapy (BNCT)
station [Savolainen et al., 2013] [Kankaanranta et al., 2012]. A BNCT
irradiation room was also built from steel tube elements filled with heavy
concrete and a heavy steel-framed lead door. The BNCT beam moderator and
collimator structures consist of several different materials, the most important

16
of which is the FluentalTM neutron moderator [Auterinen & Hiismäki, 1994],
which is used to shape the neutron energy spectrum to be optimal for BNCT.
Bismuth, lead and lithiated polyethylene were also used for gamma and
neutron shielding.

Figure 4: Vertical and horizontal section views of the FiR 1 reactor with thermal column. [General
Atomic, 1962a]

All in all, in the decommissioning planning, 17 different materials whose


activation need to be considered were identified. These are listed in Table 1.
[Publication I]. “Other” consists of e.g. lithium enriched shielding plastics,
wood, bismuth and bitumen. Their contribution to total activity is negligible.

Table 1: Complete inventory of parts to be decommissioned. [Publication I]

Material Density (g/cm3) Volume (cm3) Mass (g)


Concrete in biological shield 2.44 2.50E+07 6.10E+07
Heavy-aggregate concrete in 3.50 3.79E+06 1.33E+07
thermal column
17-4 PH steel 7.78 2.24E+06 7.27E+06
AGOT graphite in thermal 1.70 2.28E+06 3.88E+06
column
Steel in shadows 7.90 4.15E+05 3.28E+06
AlMg3 inside the tank 2.66 1.18E+05 1.46E+06
Fluental 2.99 4.50E+05 1.33E+06
Aluminium in tank and tubes 2.66 4.32E+05 1.15E+06
AGOT graphite 1.70 3.25E+05 5.53E+05
Al in thermal column 2.66 1.36E+05 3.62E+05
Stainless steel 7.90 1.28E+04 9.46E+04
Boral in thermal column 2.53 3.83E+04 8.24E+04
Other 2.02E+06 5.89E+06

17
Physical and calculational background

3 PHYSICAL AND CALCULATIONAL


BACKGROUND

3.1 RADIOACTIVITY AND ACTIVATING NUCLEAR


REACTIONS
An atom consists of a nucleus of protons and neutrons and an electron cloud
around it. The nucleus is held together by the strong force, which overcomes
the force of repulsion between the positively charged protons. The net energy
associated with the balance of the strong force and the force of repulsion is
called the binding energy. If the electrons in cloud around the nucleus are not
in the lowest possible energy levels, an atom is said to be excited.
Radioactive decay is a stochastic process in which an unstable atomic nucleus
releases energy by emitting radiation. The decaying nucleus is called the
parent radionuclide, and the process produces a daughter nuclide. A daughter
nuclide can also decay, which produces a decay series. Stable nuclides are
nuclides that are not radioactive and consequently do not spontaneously
undergo radioactive decay.
The rate at which the decays occur in a sample of certain unstable nuclei is
directly proportional to the number of these nuclei in the sample. This
proportionality constant is called the decay constant 𝜆. If a sample contains N
atoms of a radionuclide, decay constants 𝜆 are defined as a function of time as
follows:

= −𝜆𝑁 (1)

The amount of radionuclides at instant t can be expressed by solving


differential equation (1):

𝑁(𝑡) = 𝑁 𝑒 , (2)

where 𝑁 means the number of radioactive atoms at time point 0.


Time constant 𝜏 is defined as the mean lifetime of decaying atoms. Another
typically used parameter is half-life 𝑇 / , which is the time taken for half of the
radionuclide atoms in a sample to decay. Applying equation (2) one can
calculate a correspondence:

𝑇/ = = 𝜏ln (2), (3)


Where,
𝜏= (4)

18
The main forms of radioactivity in reactor structural materials are beta and
gamma rays, which are caused by absorption reactions of neutron from the
reactor core. Negative beta particles (electrons, β−) are emitted by a nucleus
when a neutron is transformed into a proton during nuclear transformation.
Beta (or alpha) decay may simply proceed directly to the ground (lowest
energy) state of the daughter nucleus without gamma emission, but the decay
may also proceed wholly or partly to higher energy states (excited states) of
the daughter. In the latter case, gamma emission may occur as the excited
states transform to lower energy states of the same nucleus. 60Co, for instance,
decays by undergoing beta radioactivity and forms a stable nucleus of 60Ni.
The transformation, accompanied by the emission of an electron and an
antineutrino, results in an excited 60Ni nucleus 999 times out of 1 000. The
nucleus loses the 2 158.80 keV of its excess energy by emitting two sequential
gamma rays.
The gamma emission is typically instantaneous, though with some nuclides it
can take place with a delay of up to several hours, e.g. around six hours for
99mTc.

Gamma-rays have well-defined energies. Measuring the energy of the gamma


rays emitted allows for positive identification of the emitter nucleus.
Due to the time scales and applications considered in this study, the following
chapters omit the beta decay in case of profilic beta delayed gamma emitters
(e.g. 60Co) and refer to them only as gamma active.

In nature, radioactivity is produced especially by the slow decay of the


primordial uranium and thorium and by nuclear reactions in the high
atmosphere caused by cosmic radiation. Radioactivity can also be produced
artificially by bombardment of stable isotopes with particles generating
unstable isotopes. This work is focused on neutron induced reactions, but
particle bombardment can mean also e.g. photons, electrons, protons and
deuterons.
Cross-sections express the likelihood of particular interaction between an
incident neutron and a target nucleus. These can be further divided into
microscopic and macroscopic cross sections. The former defines the effective
target area of a single nucleus, while the latter represents the effective target
area of all of the nuclei contained in certain volume.

The most typical neutron-induced reactions resulting in activation of a target


nuclide are:

1) Radiative neutron capture


2) Particle emission reactions
3) Fission reaction
4) Inelastic scattering

19
Physical and calculational background

Neutron capture means that a target nucleus absorbs a neutron resulting in an


isotope with higher mass number. The product nucleus is often unstable. Two
examples of radioactivity production via neutron capture in FiR 1 structural
materials are Fe(n, γ) Fe or Co(n, γ) Co
An (n,2n) reaction is a special case of neutron absorption where a high-
energy neutron induces a reaction with a product nucleus of lower mass
number. Some of these reactions also produce the same radioactive isotopes
as thermal neutron capture reactions, e.g. 60Ni(n,2n)59Ni and 58Ni(n, 𝛾)59Ni in
FiR 1 materials.
Particle emission reactions mean that a target nucleus absorbs a neutron,
emitting particles such as an alpha, proton, two neutrons or a deuteron. The
unstable product nucleus generally de-excites through β− emission back to a
stable nucleus. Particle emission reactions can result in different chemical
elements than the target isotope. For example, technetium-99 can be yielded
from molybdenum, which is a typical impurity in stainless steel. Other typical
examples from FiR 1 reactor materials are: 𝑀𝑛(𝑛, 𝑝) 𝐶𝑟 → 𝑀𝑛, and
6Li(n,𝛼)3H.

Fission means that a heavy target absorbs a neutron and divides into two
large segments, simultaneously releasing 2-3 neutrons. Fission occurs mainly
in the reactor fuel, but the biological shield concrete can also contain very
small amounts of uranium or thorium.
Inelastic scattering means that the target nucleus does not absorb a
neutron, but part of the neutron energy is transferred to the target in such a
way that the kinetic energy is not conserved. As a result, the incoming particle
causes the nucleus it strikes to become excited or to break up. In nuclear
reactors, inelastic collisions are important especially in neutron moderation
process.

In a typical research reactor with a total power of a few hundred kW, neutron
fluxes in the core area components are around 1011−1013 n cm-2s-1. The FiR 1
research reactor has been in operation for around 46 000 hours during the
years 1962-2015. Consequently, the neutron fluence in the areas close to the
core has been around 1.7 × 10 − 1.7 × 10 n cm-2s-1. Neutrons are often
divided according to their kinetic energy into thermal (below 0.625 eV),
resonance (0.625 eV to 1 MeV), and fast (> 1 MeV) groups assuming that all
the particles within a group have the same physical properties. The division
originally inherits from reactivity calculations, but it is applied in the
calculations in this thesis as well.
The calculation method in Publication I consists of two stages. Neutron
transport calculation with MCNP code [MNCP, 2003] uses ENDF/B [IAEA,
2019c] formatted cross section libraries with hundreds of groups, but the
point-depletion code ORIGEN-S code [Gauld et al., 2011a] limits the number
of groups to only three. This is still reasonable, because both the activation
cross sections and the neutron fluxes in the construction materials are centred
on the thermal energy region and the final production rates are the weighted

20
sums of fluxes and cross sections. 59Co, for instance is one of the most
important activating impurities of reactor materials. Its cross sections are
presented in Figure 5 [JAEA, 2019]. The three neutron energy groups used in
this work are added to the plot.

Figure 5: Cross sections of 59Co reactions. [JAEA, 2019]

3.2 MAIN NUCLIDES IN CONSTRUCTION


MATERIALS
Radionuclides can be classified e.g. according to their production mechanism,
measuring method, chemical properties, or biological hazardousness. The two
main factors for nuclide classification in nuclear decommissioning are:

1) Potential occupational exposure during decommissioning


2) Potential population exposure due to migration of long-lived nuclides
from a decommissioning waste repository.

In the first case, the (relatively short-lived) gamma-emitting nuclides that


determine the doses to the decommissioning personnel. Population exposure
due to nuclide migration from a repository to the biosphere, on the other hand,
is governed by the long-lived nuclides in the waste. They are typically not
gamma active and do not cause notable doses during dismantling work.
Nevertheless, they should be taken into account in waste management
planning.

Table 2 [IAEA, 1998] lists the most typical nuclides in decommissioned


nuclear facilities. Nuclides in spent nuclear fuel are excluded. Typically, only a
few isotopes dominate the waste inventories.

21
Physical and calculational background

Table 2: Typical neutron activation products [IAEA, 1998]

Active Half-life (years) Target Principal


nuclide isotope production reaction
3H 12.3 6Li n-𝛼
14C 5730 14N n-p
13C n- γ
22Na 2.6 23Na n-2n
23Na γ-n
36Cl 3.01× 105 35Cl n-γ
39Ar 269 39K n-p
41Ca 1.03× 105 40Ca n-γ
59Ni 76 000 58Ni n-γ
60Co 5.272 59Co n-γ
63Ni 100 62Ni n-γ
93Zr 9.5× 105 92Zr n-γ
93mNb 12 93Nb n-n
93Mo 3500 92Mo n-γ
94Nb 20 000 93Nb n-γ
94Mo n-p
99Tc 2.13× 105 98Mo n-γ
108mAg 130 107Ag n- γ
110mAg 0.68 109Ag n- γ
133Ba 10.4 132Ba n-γ
152Eu 13 151Eu n-γ
154Eu 8.6 153Eu n-γ
155Eu 4.76 154Eu n-γ
166mHo 1200 165Ho n-γ
54Mn 0.85 54Fe n-p
55Mn n-2n
65Zn 0.67 64Zn n- γ
55Fe 2.74 54Fe n- γ
125Sb 2.76 124Sn n- γ

3.3 CALCULATION METHOD

The activation mechanisms in both the spent nuclear fuel and the reactor
structures are governed by the Bateman equations describing the
disintegration and neutron absorption reactions (both capture and fission)
[Stacey, 2007]:

= ∑ 𝛾 𝛴 𝛷 + ∑ 𝜆( → ) + 𝜎 ( → ) 𝛷 𝑛 − 𝜆 + 𝜎 𝛷 𝑛 (5)

22
where
𝛾 = fission yield of isotope j from fissile or
fissionable isotope k
𝛴 = macroscopic fission cross section of isotope k
𝛷 = neutron flux
𝜆 = decay constants
𝜎( → )
= transmutation cross section for the production
of isotope i into isotope j
𝜎 = neutron absorption cross section

Here j = 1 … N, where N is the number of calculated isotopes n. The first sum


runs over k fissile or fissionable isotopes and the second sum over i=1…M,
where M represents all the other nuclides in the system. In case the structural
materials do not contain any fissile isotopes, the fission yield is zero.

The specific activities of all the involved radionuclides can be determined by


solving this group of equations with numerical integration. In practice, this
requires two separate stages, i.e. a particle transport code to solve the neutron
fluxes and a point-depletion code, which takes the energy dependent neutron
flux values from the transport calculations together with the material
composition data and operating history to determine the quantity of neutron
activation products. Figure 6 gives an overview of the calculation process
[IAEA, 2019a].

Figure 6: General approach for activity calculations [IAEA, 2019a]

23
Physical and calculational background

The neutron transport code requires a model of the neutron source (reactor
core) and the surrounding structures. For the neutron transport phenomena
modelling, the minor impurities of materials are insignificant. However, in the
activation source term calculations the detailed material composition is very
important, since significantly activating elements can be very minor
impurities, which have little effect on neutron transport phenomena. Since the
activation model assumes that the neutron target is homogeneous, the
calculation has to be repeated for all the individual components or structural
pieces of interest, where this assumption is valid enough.

The calculation method in publication I uses the Monte Carlo based code
MCNP [MCNP, 2003] for neutron transport calculations and the activation
reactions are calculated using the point-depletion code ORIGEN-S [Gauld et
al., 2011a]. The approach is illustrated in Figure 7 [Publication I].

Figure 7: Applied calculation system [Publication I]

3.4 CHARACTERIZATION BY THE SCALING


MATRIX METHOD
Activity inventories in decommissioning waste can be estimated with
calculations, but they must also be validated with material samples and
activity measurements before and during dismantling work. The legal
requirement is that the total activity and nuclide vector for each waste package
must be known [STUK, 2013a]. Gamma activities are relatively easy and fast
to measure but measuring beta activities requires chemical separation and
more time-consuming techniques.
A nuclide vector is defined as a normalized material-specific variable
representing the relative contribution of each nuclide to the total activity.
Because during dismantling dozens (or even thousands for larger nuclear
facilities) of waste packages need to be measured, a method called the scaling
matrix approach is employed [STUK, 2013a]. In this approach, before actual

24
dismantling, material samples of each (plausibly) activated material are
collected and studied in detail. Later, for the actual waste inventory, only the
gamma activity of the waste packages is measured, and the difficult-to-
measure beta activities are scaled assuming the same ratio between nuclide-
specific activities as in the earlier measured samples. The procedure is further
illustrated in Figure 8.

Figure 8: Characterization scheme. [Publication II]

The gamma active nuclides used in scaling are called key nuclides. Identifying
appropriate key nuclides requires considering their half-life, radiation energy,
chemical challenges in final disposal, toxicity, etc., with a conservative
approach, but also enabling the measurements to be later performed
systematically for dozens of waste packages during dismantling. Table 3 lists
the tentatively identified key nuclides and other nuclides to be scaled. Nuclides
are sorted according to their mass numbers, but nuclides with half-lives of less
than half a year are omitted [Ruokola, 2016]:

Table 3: Identified key nuclides and other main nuclides. Modified from table in Reference
[Ruokola, 2016] by removing short-lived nuclides.

Material Key nuclides Other nuclides


Concrete 54Mn,60Co, 65Zn, 133Ba, 3H, 14C, 36Cl, 41Ca, 55Fe, 63Ni, 79Se
134Cs 152Eu, 154Eu

Aluminium 54Mn, 60Co, 65Zn 55Fe, 63Ni

Steel 54Mn, 60Co 14C, 55Fe, 63Ni, 79Se, 94Nb

Graphite 60Co, 152Eu, 154Eu 3H, 14C, 36Cl, 55Fe, 63Ni, 133Ba

Fluental 3H 14C

Lead 54Mn, 60Co 75Se ,110mAg, 119mSn, 125Sb

Bismuth 60Co 36Cl, 55Fe, 63Ni, 65Zn, 108mAg, 110mAg

25
Physical and calculational background

Contamination can be taken into account by estimating the contamination


sources and calculating their nuclide vector separately. If the gamma
spectrometric measurement results include activities from both activation key
nuclides and contamination key nuclides, total activity is calculated by
summing the two nuclide vectors with appropriate weights. If the key nuclide
is the same (e.g. 60Co from active steel dust), a conservative approach has to
be used to determine the difficult-to-measure nuclides. Practically this means
using the higher fraction between the activities of DTM and key nuclides. The
procedure is illustrated in Figure 9.

Figure 9: Contamination calculation. [Räty, 2019]

3.5 ACTIVITY MEASUREMENT METHODS

3.5.1 KEY NUCLIDE MEASUREMENTS

The gamma spectrometric measurements in Publications II and IV were


performed using an In Situ Object Counting System (ISOCS) setup by
CANBERRA [Canberra, 2019a] based on a liquid nitrogen cooled HPGe
detector and spectroscopy software GENIE2000 [Canberra, 2019b].
The gamma ray detection with a germanium detector is a calorimetric
measurement, where preferably all the energy of a single gamma photon is
absorbed in the detector. The deposited energy moves electrons from the
semiconductor material valence band to the conduction band, causing a
charge pulse that is directly proportional to the deposited energy. The released
charge is collected with a charge-sensitive preamplifier, which integrates and
amplifies the signal to produce a step function pulse, the amplitude of which
is proportional to the total energy deposited. The preamplifier signal is
directed to a spectroscopy amplifier that magnifies, filters and shapes it to
enhance the signal-to-noise ratio. The pulse height of the analogue signal
produced by the detector and shaped by the amplifiers is finally determined
by an analogue-to-digital converter (ADC). The digital pulse height value is
routed to a multi-channel analyzer (MCA) to be displayed as a spectrum of
detected energies. The gamma photons have discrete energies. Therefore, the

26
narrow, easy to identify peaks in the gamma spectrum correspond to the total
energy deposit of a single gamma ray. Events where part of the photon energy
escapes from the detector form the continuous background beneath the peaks
(see Figure 11). A schematic view of the ISOCS components is presented in
Figure 10 [Wall brink et al., 2002].

Figure 10: Electronic components of a gamma spectrometer. [Wallbrink et al., 2002]

Figure 11: Example of a measured gamma spectrum in an aluminium sample. [Publication II]

3.5.2 DIFFICULT TO MEASURE NUCLIDE MEASUREMENTS

Measuring the beta and alpha emitters requires separating them individually
from the sample matrix and other radionuclides before measurement. This
involves decomposing the sample completely with either acids or burning
methods. These nuclides are referred to as difficult-to-measure nuclides
(DTMs).

The methods presented for difficult-to-measure beta active nuclides in


Publications II and IV utilized liquid scintillation counting (LSC) with various
decomposing methods. They have also been supplemented by composition
measurement using mass spectrometers and different emission spectrometric
methods.

27
Physical and calculational background

In liquid scintillation counting, aqueous radionuclides are mixed with a liquid


scintillation cocktail, an organic aromatic compound in an organic solvent.
The scintillator molecules are excited by the transference of the kinetic energy
of charged particles via solvent molecules to scintillator molecules. When the
excited scintillators are relaxed, they emit light photons by fluorescence.
Typically, an LSC measurement also requires separating the studied nuclides
from the sample matrix using, e.g., resins or oxidizing processes. Since this
separation process is never ideal, the final results need to be scaled with the
separation yield.
For the FiR 1 materials, dissolution was carried out using acid digestions.
Depending on the studied nuclides, separation and purification is typically
carried out by means of precipitation, resins, distillation, etc. Volatile
radionuclides (3H, 14C) have also been separated from solid matrices using an
oxidizer or a pyrolyzer. Activity measurements were carried out using a Hidex
300 SL liquid scintillation counter with TDCR technology and yield
measurements using inductively coupled plasma optical emission
spectroscopy (ICP-OES). [Publication II] [Publication IV].

3.5.3 IMPURITY ANALYSIS TECHNIQUES

Since decommissioning waste activity is typically due to very small impurities


in construction materials, material composition analysis is an essential part of
validation measurement. Publications II and IV include composition
measurements using inductively coupled plasma mass-spectrometry (ICP-
MS) and inductively coupled optical mass spectrometry (ICP-OES).
ICP–MS is based on using a high temperature ionisation source (ICP)
coupled to a mass spectrometer. The plasma ionizes the sample and creates
atomic and small molecular ions, which are then separated on the basis of their
mass-to-charge ratio and a detector receives an ion signal proportional to the
concentration. For radioactive waste characterization, it is essential that ICPS-
MS can separate different isotopes of the same element. The theoretical
sensitivity of ICP-MS measurement is as low as the ppt level
[Thermoscientific, 2019]. Indicative detection limits are presented in Figure
12.

28
Figure 12: Indicative detection limits for high resolution ICP-MS at the Centre for Nuclear Safety,
VTT. Figure by VTT.

Prior to the elemental analysis, the solid material needs to be dissolved.


Microwave digestion is the typically used technique. However, for example in
the case of graphite, material solubility was shown to be a limiting factor.
[Publication IV].
ICP-OES uses the inductively coupled plasma to produce excited atoms and
ions that emit electromagnetic radiation at wavelengths characteristic of a
particular element. The intensity of the emissions from various wavelengths of
light are proportional to the concentrations of the elements within the sample.
VTT uses Agilent SVDV 5100 ICP-OES for the analysis of stable elements
at the ppm level [Agilent, 2019]. In the case of FiR 1 characterization, ICP-OES
has been especially used in concrete composition measurement and
determination of yields after radiochemical analysis of DTM radionuclides
[Publication II] [Leskinen et al., 2020].

29
Activity inventory calculations in FiR 1 project

4 ACTIVITY INVENTORY
CALCULATIONS IN FIR 1 PROJECT

4.1 MODELLING THE OPERATING HISTORY


The FiR 1 research reactor was in operation between 1962 and 2015. To get
the time dependence over the whole core lifetime, the 53-year operating
history needs to be divided into several time intervals, within each of which
the neutron flux distribution is assumed to vary negligibly. This means taking
into account all different fuel loading configurations and updates in reactor
structures.

Two major modifications during these years were:

1) Thermal power was increased from 100 kW to 250 kW in 1967. Main


modifications in the core configuration were adding one control rod and
taking into use a new type of stainless steel cladded fuel elements with
higher uranium content.

2) Building of the BNCT station in the mid-1990s.

In total, FiR 1 has been used for around 11 500 MWh in 46 different fuel
loading configurations. Since detailed operating hours from years 1962-1975
were not available, the activation models presented in Publications I and III
calculate the operating hours from the known accumulated produced energy
of 2980 MWh. Operating hours for years 1976-2015 were based on the
operating diary. Produced thermal energies for each year are presented in
Figure 13.
Thermal energy
produced (MWh)
450
400
350
300
250
200
150
100
50
0
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014

Year

30
Figure 13: Assumption of thermal energies produced each year in MWh. Colours represent the
assumptions used in the activation model: orange represents operating hours calculated from
accumulated burnup (thermal power 100 kW), green represents operating hours calculated from
accumulated burnup (thermal power 250 kW), blue represents the real operation diary prior to
building the BNCT station (thermal power 250 kW) and red represents the real operation diary after
building the BNCT station (thermal power 250 kW). [Publication III]

Since the operating history of the FiR 1 reactor is very complicated with
thousands of start-ups and shut-downs, this limits the alternatives for a
practical calculation method. The deterministic point-depletion method in
ORIGEN-S (publications I, II and IV) enables almost arbitrary amount of
operating periods, but modelling thousands of start-ups and shut-downs
would be practically impossible with Monte Carlo methods.
The calculations in Publication I use a model where the neutron fluxes have
been first calculated in stationary operating conditions with the Monte Carlo
code MCNP and this has been combined with a point-depletion code ORIGEN-
S to model the whole operating history.
The operating history in the ORIGEN-S activation model obtained the
operating hours during the early years of 1962-1976 according to an
accumulated total burnup of 124.09 MWd. The reactor history during years
1976-2015 was modelled according to actual irradiation hours. All significant
holidays and maintenance breaks were also taken into account. This was
performed on a yearly basis by modelling a typical irradiation schedule for a
week and chaining these irradiation weeks for each year in succession. Total
neutron flux was considered constant during all irradiations and zero in
between them. In addition, a long decay period following the shutdown of the
reactor was included in the activation model to investigate the expected
activity levels during the disassembly and the final disposal.

Fuel inventory calculations in Publication III were conducted somewhat


differently. In the studies described in Publication III the activation of the fuel
was calculated by both a fuel element-specific Serpent [Leppänen et al., 2015]
Monte Carlo model and a simplified ORIGEN-S activation model based on
pre-calculated fluxes in each element-type. ORIGEN-S calculations were
performed similarly to Publication I, but Serpent calculation was modelled
following the accumulated burnup and only the last few years were modelled
in higher detail. This is justified by means of an iterative process described in
more detail in Publication III.

4.2 NEUTRON TRANSPORT CALCULATION


Neutron fluxes for the ORIGEN-S model in Publications I and III were
calculated using a Monte Carlo based code MCNP5 [MCNP, 2003]. MCNP is a

31
Activity inventory calculations in FiR 1 project

calculation code developed by the Los Alamos National Laboratory (LANL). It


uses the Monte Carlo method to simulate neutron particle tracks through the
constructed 3D geometry

Before 1994, the reactor was equipped with a thermal column that was
replaced by a different moderating structure to deliver epithermal neutrons
for the BNCT facility. In addition, in the late 1980s the horizontal beam ports
were plugged. Since the structure of the reactor and consequently the neutron
flux is vastly dissimilar in each case, three basic neutron transport MNCP
models were constructed to cover the entire operating history:

1. MCNP model with thermal column and open beam ports


2. MCNP model with thermal column and plugged beam ports
3. MCNP model with BNCT facility and plugged beam ports

Two-dimensional cross sections of the MCNP geometry are shown in Figures


14 and 15 for the thermal column model and BNCT facility model, respectively.

Figure 14: Horizontal and vertical cross sections of the MCNP geometry with the thermal column.
[Publication I]

Figure 15: Horizontal and vertical cross sections of the MCNP geometry with the BNCT facility.
[Publication I]

32
As a default, the MNCP calculated neutron fluxes are presented per fission
neutron. The results were scaled to 250 kW nominal reactor power with a
factor of 1.8875×1016, based on average energy released (about 200 MeV) and
average number of 2.43 neutrons produced in each fission. In the final output,
the neutron fluence rates were scored for thermal (below 0.625 eV),
epithermal (between 0.625 eV and 1 MeV) and fast neutron energy range
(above 1 MeV) as specified in the ORIGEN-S manual [Gauld et al., 2011a].

Thermal and total fluence rate maps for each of the three modelled reactor
geometries are given in Figure 16 [Kotiluoto & Räty, 2016]

Figure 16: MCNP simulated total neutron fluence rates a) with thermal column and open beam
ports, b) with thermal column and plugged beam ports, and c) with BNCT beam. Pictures d), e)
and f) demonstrate the thermal fluxes from the same neutron transport models. [Kotiluoto & Räty,
2016]

The MCNP simulations were based on ENDF/B-VI cross sections. The largest
factors of uncertainty were assumptions in material definitions and the
simplification of simply using air in the open beam tubes. In practice, beam
ports have been used for experiments with instrumentation and shielding.
These experiments have not been documented in detail, but the used
approximation is highly conservative with respect to neutron fluence outside
of the tube, since open beam ports result in much higher neutron fluxes to the
surrounding structures. High statistical uncertainty in the calculated neutron
fluence was only encountered in areas where the fluence rate is low and
insignificant with respect to neutron activation.

33
Activity inventory calculations in FiR 1 project

4.3 POINT-DEPLETION CALCULATION

After the neutron source calculations, the obtained fluxes were collapsed to
three-group form and used to model the production of different nuclides in all
structures using the point-depletion code ORIGEN-S. ORIGEN-S is a
computer code system designed by Oak Ridge National Laboratories for
calculating the build-up, decay and processing of radioactive materials. It
incorporates standard ENDF/B-VI based cross-section libraries, fission
product yields, decay data and decay photon data [Gauld et al., 2011b].

The ORIGEN-S calculation method is based on solving the nuclide


concentrations by applying a power series approximation of the matrix
exponential solution to the Bateman equation and separate secular
equilibrium approximations for handling short-lived nuclides.
Firstly, the contribution from short-lived nuclides (𝑇 / < 0.1 × user
defined step length) is determined with the linear chain method [Gauld et al.,
2011b]. Secondly, the coefficient matrix A is constructed by neglecting the
short-lived nuclides. This starts by writing the group of Bateman equations in
matrix form as:

𝑁̇(𝑡) = 𝐴𝑁(𝑡), (6)

where N is the vector of nuclide atom densities and A is the transition matrix
containing the rate coefficients for radioactive decay and neutron absorption,
e.g. 𝐴 = −𝜆 − 𝜙𝜎, where 𝜆 is the decay constant, 𝜎 the reaction cross section
and 𝜙 the neutron flux. At this point, the short-lived nuclides are omitted,
giving the formal solution:

𝑁(𝑡) = exp(𝐴𝑡) 𝑁(0) (7)

Expanding the exponential to series, the solution becomes

( )
𝑁(𝑡) = ∑ !
𝑁(0) (8)

A straightforward solution of equation (7) would require storing the whole


transition matrix. To optimize the calculation procedure and memory usage,
N is further expressed as a recursion relation:

𝐶 ≡ 𝑁 (0), 𝐶 = ∑ 𝑎 𝐶 , (9)

where t is the user-defined time step and aij represents the first-order rate
constants for the formation of nuclide i from j. Thus:

34
𝑁 =∑ 𝐶 (10)

The convergence of series (8) is ensured with certain internal procedures in


ORIGEN-S that are not in the scope of this thesis [Gauld et al., 2011a].

The contribution from long-lived nuclides to short-lived nuclides is solved by


assuming their decay and transmutation chains to be in secular equilibrium
at the end of the time step

=∑ 𝑎 , 𝑥 = 0, (11)

where xi is the final concentration of long-lived nuclides obtained in the matrix


exponential solution. This can be solved by iterating

𝑥 = ∑ 𝑎, 𝑥 , (12)
,

Finally, the contributions to the final concentrations of short-lived nuclides


are saved and the contribution to long-lived nuclides is added to the initial
concentrations of those nuclides.

Neutron fluxes in ORIGEN-S are either determined from user input or, in the
case of nuclear fuel, calculated from input thermal power. In the latter case of
input thermal power, fluxes are calculated using microscopic cross sections,
the recoverable energy released, and the fissionable and absorbing nuclide
concentrations at the current point in time. Data on recoverable energies and
cross sections are based on ENDF/B-VI evaluations.
The calculations in Publications I and III used ORIGEN-S as a standalone
code. Therefore, the neutron spectra were specified by defining the three-
group neutron spectrum weighting factors THERM, RES and FAST as
specified in SCALE code package version 6 [Gauld et al., 2011a] and by
determining the effective cross section in a medium as a weighted sum of cross
sections in these groups. In theory, the THERM weighting factor is used to
adjust the 2200-m/s cross sections in the library for a thermal neutron
spectrum for the system. RES and FAST (defined with user-input fluxes) are
used to weight the resonance and fast group cross sections in forming effective
one-group values. In this case, the flux is not read from any binary library.
Having input values for THERM, RES, and FAST, ORIGEN-S then combines
these weighting terms with the three-group cross sections to form the effective
one-group cross sections, σeff, used in the calculation of reaction rates based
on the total thermal flux as follows [Gauld et al., 2011b]:

𝜎 = 𝑇𝐻𝐸𝑅𝑀 × 𝜎 + 𝑅𝐸𝑆 × 𝜎 + 𝐹𝐴𝑆𝑇 × 𝜎 (13)

35
Activity inventory calculations in FiR 1 project

where the sigmas are cross sections in thermal, resonance and fast groups.
( )
𝜎 = the 2200-m/sec neutron absorption cross section, 𝜎 = ∫ , 𝑑𝐸
and 𝜎 = the fission spectrum averaged cross section for all the reactions
above > 1 MeV. ORIGEN-S assumes that the thermal cross sections have a 1/v
dependence [Gauld et al., 2011b]:

, ( )

𝑇𝐻𝐸𝑅𝑀 = 𝜎(1/𝑣) = √
(14)

The other quantities are calculated from user-input fluxes in the thermal,
resonance and fast energy region 𝜙 , 𝜙 and 𝜙 [Gauld et al., 2011b]:

𝑅𝐸𝑆 = , 𝐹𝐴𝑆𝑇 = . (15)

In Publication I, fluxes throughout the activation model were kept the same as
in the 250 kW model. This is a highly conservative assumption as most of the
active nuclides produced before the power increase from 100 kW to 250 kW
will have already decayed by the year 2015.
Since all of the individual time steps in the ORIGEN-s input were only a few
hours, all nuclides in that are listed in final results six months after shutdown
(see Tables 4 and 5) are considered “long lived” in the ORIGEN-S
methodology.

In the MCNP Monte Carlo model, the reactor was divided into around 200
structural parts and all of these were modelled with ORIGEN-S on a
component-to-component basis. The aim was to divide the neutron transport
model in a similar manner to real dismantling planning. In particular, the
areas in the upper parts of the reactor where there has been practically no
neutron flux were omitted. For instance, most of the concrete is inactive and
can be considered as ordinary waste. Because the initial assumptions were
conservative with respect to amount of activating impurities and neutron
fluxes, the actual amount of radioactive waste is plausibly smaller.

Table 4 lists the material-specific total inventories and nuclides with


highest contribution to the total activity in decommissioning waste six months
after final shutdown [Kotiluoto & Räty, 2016]2. Spent nuclear fuel is excluded
and discussed further in Chapter 6. The six-month time point was initially
chosen because it was estimated that first sampling and measurements could
be topical at that point. Moreover, using an underestimated decay time for

2 Table VII in Publication I present partially different total activites. The input data on material
compositions and neutron fluxes in the BNCT-station structures was updated in late 2016. This had an
effect in calculated activities especially in Fluental, graphite and aluminium. The presented quantities in
Table 4 correspond to the latest data available.

36
later waste management planning is also a conservative approach. However,
results are easy to calculate for any later phase following the law of radioactive
decay. This time point is also referred to in the calculated activities in later
chapters of this thesis, unless it is mentioned otherwise.
The table also includes the old thermal column structures decommissioned
in 1995. The quantities in Table 4 are presented for all structural parts of the
considered material. Since the activity is unevenly distributed, some parts may
have much higher activity per mass unit and the actual packed waste volumes
will be much higher. In practice, the decommissioning waste also contains a
small amount of operating waste, such as ion-exchange resin and used
personal protection equipment, but their contribution to the total activity is
negligible.

Table 4: Calculated material-specific total inventories in FiR 1 structures and main contributing
isotopes six months after shutdown. [Kotiluoto & Räty, 2016]

Material Volume Mass Main isotopes Total activity


(m3) (kg) (TBq)
Biological shield 25 61 000 3H, 60Co, 152Eu, 0.105
normal concrete 41Ca, 39Ar, 14C

Graphite 2.61 4450 3H, 14C, 152Eu, 0.461


60Co, 133Ba, 36Cl

Steel 0.449 3 540 63Ni, 55Fe, 60Co, 1.91


59Ni, 14C

Aluminium 0.840 2230 55Fe, 65Zn, 63Ni, 0.031


60Co, 54Mn, 59Fe

Fluental 0.450 1330 3H, 14C 1.3


Other 3H, 14C, 108MAg,

- Lithiated plastic 1.353 1990 63Ni, 137Cs 0.428


- Bitumen 0.126 132 <1E-4
around the tank
- Wood 0.154 100 <1E-9
- Boral 0.042 105 0.003
- Bismuth 0.079 776 0.002
- Lead 0.242 2700 <1E-5
- Boronated 0.066 152 <1E-6
concrete (beam
port plugs)
- Heavy- concrete 3.790 13300 <1E-6
( thermal column
door)
Sum 37.7 9430 4.24
0

37
Activity inventory calculations in FiR 1 project

Nuclide-specific fractions of total activity (i.e. nuclide vectors) are given in


Table 5 [Kotiluoto & Räty, 2016]. This list shows only the nuclides with highest
contribution to total activity, and the values are taken from highly activated
representative parts. The reference time for these results is also six months
after shutdown. Nuclide-wise fractions in individual parts of the same material
may differ slightly due to different neutron spectra and heterogeneity of the
material. The results do not take into account any diffusion of volatile nuclides.
This is a conservative assumption but may have a large effect on tritium
activities especially in concrete. Both Tables 4 and 5 list only one values for
steel and aluminium, although actually there are several types of them.

Table 5: Calculated nuclide-specific fractions of total activity for main materials six months after
shutdown [Kotiluoto & Räty, 2016]

Material Half life Total activity (TBq),


(years) Fraction of total
activity by isotope (%)
Graphite 0.461
3H 12.32 87.36
152Eu 13.54 5.60
60Co 5.27 4.80
14C 5 730 1.60
154Eu 8.593 0.48
36Cl 301 300 0.13
133Ba 10.74 0.03
134Cs 2.065 0.01
10Be, 39Ar, 37Ar, 131Cs, 59Fe, <1E-5
58Co, 49V, 131Ba, 35S

Steel 1.91
55Fe 2.737 36.51
63Ni 101.0 33.66
60Co 5.271 28.99
59Ni 76 000 0.33
51Cr 0.076 0.26
14C 5 730 0.20
54Mn 0.857 0.03
59Fe 0.122 0.02
58Co, 75Se, 94Nb <0.01
Aluminium 0.0312
55Fe 2.737 88.99
65Zn 0.671 9.61
63Ni 101.0 0.90
54Mn 0.857 0.42
59Fe 0.122 0.06
60Co 5.271 0.01
46Sc, 45Ca, 51Cr <0.01

38
Fluental 1.3
3H 12.32 >99.99
55Fe 2.737 <1E-3
35S 0.239 <1E-4
14C, 54Mn, 22Na, 33P, 32P, 59Fe, <1E-5
60Co, 51Cr

Concrete 0.105
3H 12.32 88.81
152Eu 13.54 9.28
39Ar 269 0.37
41Ca 102 700 0.32
154Eu 8.593 0.29
60Co 5.271 0.26
151Sm 90.0 0.19
14C 5 730 0.19
63Ni 101.0 0.12
55Fe 2.737 0.10
133Ba 10.74 0.03
36Cl, 45Ca, 155Eu, 59Ni < 0.01
Lead <1E-5
108mAg 438.0 50.99
63Ni 101.0 40.40
113mCd 14.1 3.01
121mSn 43.9 2.61
121Sn 0.003 2.03
125Sb 2.759 0.56
59Ni 0.39
Plastic 0.4
3H 12.32 >99.99
14C, 10Be <1E-5

4.4 VALIDATION AND UNCERTAINTIES


MCNP and ORIGEN-S are well-validated and widely used codes for
decommissioning planning purposes. Similar calculations have been
previously performed for several nuclear power plants, e.g. [Love et al., 1995]
[Philippen et al., 2018] [Volmert et al., 2016] [Schlömer et al., 2016] [Pantelias
& Volmert, 2015]. Systematic code validation goes beyond the main focus of
this thesis. Instead, the quality of the calculated results in this thesis is
primarily dependent on uncertainties related to several approximations being
made in different stages of the calculations process.
The results of neutron transport calculation are mainly dependent on the
assumption that all information shown on structural drawings (i.e. geometry,
dimensions, materials etc.) considered in activation calculations is valid and
reliable, including details such as penetrations and beam ports. The MCNP
calculations in Publication I were mainly based on conservative assumptions.

39
Activity inventory calculations in FiR 1 project

For example, the horizontal neutron beam tubes were assumed to be empty
during operation, although in practice they contained research equipment that
absorbed a proportion of the neutrons.

Another point is to divide the structures into proper size components or pieces,
because the flux will be averaged over the total volume of the component. The
single specific activity value calculated for the component will not show a
difference in specific activity between regions located very close to or remote
from the source. MCNP calculations in Publication I have been validated by
studying the reaction rates of earlier activation analyses [Kotiluoto & Räty,
2016]. The results show an uncertainty in the order of about 10-15%. However,
part of the uncertainty may be due to the very small target applied.

The neutron transport model in Publication I used a next-event-estimator


variance reduction, but separate weight windows were not used. This is mainly
because the activity calculations need to study larger areas and an acceptable
statistical precision was achieved just by increasing the calculation time.
However, in case of small material samples used for validation, the neutron
transport model was modified by adding a weight windows mesh to the studied
sample location.

Uncertainties in point depletion calculations are mainly related to the


information regarding the chemical compositions of the structures being
exposed to neutron fluxes. In most cases, the information available on the
chemical composition of metals, concretes and other material comes from
manufacturer specifications. Because the activating impurities may have only
a small effect on the mechanical properties of the material, these impurities
are not necessarily taken fully into account in construction specifications.
Moreover, the point-depletion code ORIGEN-S treats materials as
homogeneous masses while in reality concentrations of minor impurities may
have anomalies in different structural parts. Publications II and IV include
studies on how to take into account these uncertainties by measuring various
samples and calculating their activities similarly as for the whole reactor.
These are discussed more in chapter 5. Moreover, it has been shown that
ORIGEN-S produces reliable activity estimates for activated samples from
other reactors as well [Leskinen et al., 2020].

Another possible source of error in point depletion calculations is the


complicated operating history. The reactor activation model used at FiR 1 was
built to include all maintenance breaks, longer holiday periods, etc. This
makes the input rather long. However, data on produced energy is available
and the calculations were also tested by assuming the energy to be produced
in a single run. The differences in calculated activities of long-lived
radionuclides were minor. Nevertheless, the more detailed activation model
was used because an oversimplified operating history underestimates the
amount of shorter lived nuclides, which is not a conservative approach.

40
5 MATERIAL-SPESIFIC ACTIVITIES IN
FIR 1 DECOMMISSIONING WASTE

Calculated inventory values need to be verified with measurements before and


during dismantling work. Prior to fuel removal, the main restriction to
sampling is that the structural integrity of the reactor core structures must be
maintained. Consequently, the preliminary characterization measurements in
Publications II and IV mainly analyse low-active samples from the edges of the
reactor.

5.1 CONCRETE

The aluminium tank in the FiR 1 reactor is surrounded by a 1.5-metre thick


layer of ordinary concrete, which serves as a biological shield. Concrete makes
up the largest proportion of the waste generated in a nuclear decommissioning
project. The specific activity of concrete is much lower than, e.g., metal
components in core structures, but optimizing characterization and clearance
procedures for active concrete as distinguished from non-active concrete is
essential to minimize the amount of active waste.

To obtain the initial composition data for activation calculations, FiR 1


concrete was studied by drilling three samples from inactive parts of the
reactors. The composition of these samples was studied with an optical
emission spectrometry method [Publication II]. As expected, the concrete
contained at least lithium, calcium, iron, nickel, europium and barium, which
have been identified as the activating elements producing the highest
contribution to the total activity also in other studies [Hou et al., 2005].

The measured composition was used to perform the activation calculations


reported in Publication I. A conservative approach was taken by assuming that
the activating impurities from the three concrete samples were at the
maximum measured value.
Activated volumes were calculated following the clearance regulations of
the Finnish Radiation and Nuclear Safety Authority [STUK, 2013a]. With
conservative approximations discussed in previous chapters, these
calculations estimated that concrete is activated around the core to about a 1.5
m radius and around the beam tubes to about 0.5 m radius (see 16).
The radionuclides producing the highest contribution to total activity are
152Eu, 154Eu, 60Co, 45Ca, 3H, 63Ni, 59Ni, and 55Fe. Since concrete is a porous

material, part of the 3H activity may have diffused from the material. Current

41
Material-spesific activities in FiR 1 decommissioning waste

approach is to estimate conservatively that all the tritium is still in the


concrete.

Publication II included a gamma spectrometric measurement of a concrete


sample drilled during the construction of the BNCT station in the 1990s. The
measurement was performed in April 2016. The computational specific
activity for this sample was around 10 Bq/g. The measurement results and
their comparison with calculations are listed in Table 6.

Table 6: Comparison of measured and calculated gamma activities in concrete. [Publication II]

Nuclide Measured Measurement Calculated Calculation


(Bq/g) uncertainty (%) (Bq/g) uncertainty (%)
60Co 2.5E-02 6.9 2.6E-02 10
152Eu 9.6E-01 2.3 8.1E-01 10

The uncertainties of the calculated specific activities are based on estimated


sample heterogeneity, which has a linear correspondence to activation
calculation uncertainty. Moreover, the neutron spectrum also has uncertainty.
At the interface between different materials, especially, flux variations can be
high. Measurement uncertainties arise mainly from separating the
background from the measurement data and converting the sample activity to
specific activity. Bearing these points in mind, the calculated specific activities
corresponded to the measured values reasonably well, although the measured
sample was quite old and had a very low activity. Nevertheless, the results
demonstrate that activity calculations are capable of producing reliable
estimates of nuclide activities and 60Co and 152Eu nuclides are natural choices
for future concrete waste measurement.

To further validate the concrete activity estimates, samples from active parts
of the concrete were drilled in December 2018. These are also used to verify
the activities of beta active radionuclides. The measurements were on-going at
VTT in 2019. Initial results indicate that the earlier calculated activities in
Publication I have been conservative.

5.2 STEEL

Pressure vessel and core components in nuclear reactors are usually made of
steel. These parts are exposed to very high neutron fluxes and account for a
very high proportion of total activity in decommissioning materials. The
typical nuclides with highest contribution to the total activity a few years after
reactor shutdown are 14C, 55Fe, 60Co, 63Ni and 94Nb [Evans et al., 1984]. 55Fe
has the highest activity during the initial decades, but it causes only small dose

42
in longer term storage. The most important nuclides are 60Co and 63Ni. As
these are produced from activating impurities in steel, their fractions were
conservatively overestimated for parts without known detailed composition.

Although most core structures in the FiR 1 reactor are aluminium, the
decommissioning waste still contains some steel components. Most of the
mass comes from steel shadows around beam tubes, but their calculated
specific activity is relatively low (1.5-12 kBq/g). Other important parts were the
bellows and clamps around the beam tubes (1.5-8.5 MBq/g), the irradiation
ring (6.5 MBq/g), and small bolts and dowel pins inside the tank (0.25-750
MBq/g). These require special waste packages with heavy shielding [BNG,
2017a].

As the core structures cannot be sampled prior to removing the fuel rods, data
on FiR 1 steel component activities currently relies wholly on the
computational estimates presented in Publication I. These were compared in
Publication I with TRIGA reactor results from the Medical University of
Hannover and the Korea Atomic Energy Research Institute [Hampel et al.,
2002], [Park et al., 2003]. The reactors are referred to here as MHH and KRR-
1, respectively.
The main difference between FiR 1, MHH and KRR-1 is that the thermal
column in FiR 1 was disassembled in 1995 and MHH has only one radial beam
tube. Material impurities may also have several differences. KRR-1 has been
used in total for around 3 700 MWh and MHH for around 2 188 MWh. The
concrete of the biological shield in all three reactors was made from ordinary
concrete from local aggregate material.
Table 7 lists the specific activities of the rotary specimen rack and graphite
reflector of both the KRR-1 reactor in Korea and the MHH reactor in Germany
[Publication I].

Table 7: Specific activities of certain components of FiR 1, KRR-1 and MHH reactors. Data is
combined from Tables IX and X in Publication I

Item Specific Specific Specific Mass Mass in


activity in activity in activity in in FiR 1 KRR-1
FiR 1 (Bq/g) KRR-1 (Bq/g) MHH (Bq/g) (kg) (kg)
Rotary 55Fe: 1.4E07 55Fe: 8.9E05 60Co:5.0E06 6.7 3.4
specimen 60Co: 2.8E07 60Co: 7.0E06
rack (steel 59Ni: 4.5E05 59Ni: 3.0E04

parts)
Core 3H: 5.0E05 3H: 4.1E03 60Co: 2.0E03 550 770
reflector 14C:8.8E03 14C:1.9E02
60Co: 2.8E04 60Co: 1.4E02

43
Material-spesific activities in FiR 1 decommissioning waste

All aluminium and steel parts of the rotary specimen rack were homogenised
in the FiR 1 calculations as a single component with a specific activity of 6.56
MBq/g. A major proportion of this comes from 60Co, which produced from
natural cobalt impurity in both steel and aluminium.
When comparing the specimen rack, the specific activity in FiR 1 is much
higher, but the specific activities are logical compared to the total energy
produced. Comparing the reflectors, the specific activities of 60Co in MHH and
FiR1 follow the same pattern, but the specific activities in KRR-1 seem much
lower than expected. However, data on original activating impurities is not
publicly available.

Samples of FiR 1 active steel components will be collected later to measure and
form a complete nuclide vector. This procedure has been preliminarily studied
in a joint project with VTT and the University of Helsinki using reactor
pressure vessel type steel provided by Finnish power companies [Leskinen et
al., 2020].

5.3 ALUMINIUM

The FiR 1 reactor contains aluminium in the reactor tank, in the horizontal
beam tubes, and in most of the equipment inside the tank. Aluminium parts
are illustrated in Figure 17. According to the manufacturer’s specification, the
aluminium should be of AlMg3 type. The tank, thermal column walls and
beam tubes were manufactured in Finland by Ahlström Ltd, whereas smaller
special components (grid plates, tubes, etc.) were manufactured in the US by
General Atomics. As studied in Publication I, the main activating impurities in
aluminium are cobalt, iron, nickel and zinc, although the specific activities are
much lower than in steel parts.

44
Figure 17: FiR 1 MCNP model. Aluminium parts in the reactor shown in green. [Publication II]

Aluminium in FiR 1 structures has been studied in Publication II by collecting


samples from the inactive upper parts of the tank. The approach involved
measuring the composition of inactive samples using the ICP-MS technique
and irradiating a number of the samples to validate the activation calculation
method. The measured composition was used to calculate the nuclide vector
presented earlier in Table 4.

To further validate the results, three of the inactive samples were irradiated in
the FiR 1 reactor before shutdown in June 2015. The samples were irradiated
in the irradiation ring around the core in a neutron flux of 2.32×1012 n cm-2s-1
(thermal flux 1.15×1012 n cm-2s-1). The flux is based on earlier MCNP
calculations for irradiation ring sample positions. Two of the samples were
irradiated for four hours and one for 14 hours.
The two samples with the same irradiation time were studied with an
ISOCS gamma spectrometry system [Publication II] four months after the
irradiation. Specific activities were also calculated using the same method as
for all reactor structures described in Publication I. The calculation contains
some uncertainties due to neglecting the shielding effect of the irradiation
capsule, and the neutron flux may have small variations in the irradiation ring.
The comparison of results is presented in Table 8.

Table 8: Specific activities of the studied samples. [Publication II]

Nuclide Half-life Measured Measurement Calculated Calculation


(days) activity uncertainty activity uncertainty
(Bq/g) (%) (Bq/g) (%)
46Sc 83.79 5.5 2.9 7.4 10
51Cr 27.7 700 4.7 640 10

45
Material-spesific activities in FiR 1 decommissioning waste

54Mn 312.2 1.8 7.9 4.1 10


59Fe 44.5 14 2.9 21 10
60Co 1925 5.2 3.0 18 10
65Zn 243.9 81 3.0 130 10

Both calculated and measured results show the nuclides with highest
contribution to the total activity to be 51Cr, 55Fe, 65Zn, 54Mn, 59Fe, 59Ni, 63Ni,
46Sc and 60Co. Their activity fractions are not the same as in Table 5, since

many of the nuclides are short lived and the decay time is not exactly the same.
This applies especially to short-lived 51Cr, for which the calculated value is not
conservative. Nevertheless, the calculated specific activities correspond to the
measured values reasonably well and provide a basis for further aluminium
waste characterization and classification. Considering the half-lives and
nuclide-specific clearance levels, 60Co is the best choice the key nuclide for
aluminium.

5.4 AGOT GRAPHITE


Nuclear graphite, or reactor graphite, is a synthetic material manufactured
from filler coke and pitch. The graphite in the FiR 1 reactor is the AGOT
(Acheson Graphite Ordinary Temperature) type manufactured by Union
Carbide [Carlsson et al., 2014]. AGOT is a common graphite type in older
nuclear facilities, but it is no longer commercially available [Woodcraft et al.,
2003]. According to the general information supplied by the manufacturer,
AGOT typically has a porosity of 24% [National Carbon Company, 1955].
Porosity is a crucial characteristic of graphite as the nitrogen contained in the
pores is easily activated by the reaction 14N(n,p)14C.

The FiR 1 reactor contains graphite in the ring-shaped reflector around the
core, in a few graphite rods that were used in the core, and in the thermal
column that was disassembled in 1995-1997. Activities in FiR 1 graphite were
calculated in Publication I. The nuclides with highest contribution to total
activity are 3H, 14C, 152Eu, 154Eu, 60Co, 36Cl and 133Ba (see Table 5).

Many of the challenges of active graphite and its characterization are related
to waste final disposal. These were studied in Publication IV and are discussed
further in chapter 7.

Publication II reports 14C activity measurements from samples collected from


the old thermal column graphite. In the measurements performed at the
University of Helsinki radioisotope laboratory in Viikki, dry oxidation of
carbon was used. The activity of the emission gases was determined with liquid
scintillation counting. The oxidizer used pure oxygen to fully combust the
sample to water and carbon dioxide at a temperature of 960 °C. Water and air

46
rinsing through the reaction column collect 14C into a liquid scintillation
counter bottle. The operating principle of the oxidizer is illustrated in Figure
18

Figure 18: Operating principle of the oxidizer system [Publication II]

The final results show that the specific activity of 14C in the samples varied
from 0 to 1.66 kBq/g. The measured activities were compared with the
calculated values from Publication I. 14C specific activities were estimated to
vary from 0 to 20 kBq/g depending on the distance from the core in the
thermal column. The calculated activity profile is presented in Figure 19
[Publication II].

1000
50
Measured dose rate (microSv/h)
C-14 spesific activity (kBq/g)

5
100
0,5 40 90 140 190 240

0,05
10

0,005

0,0005 1
Distance from the core (cm)

Measured spesific activity of C-14 from the samples (kBq/g)


Calculated spesific activity of C-14 (kBq/g)
Measured dose rate (microSv/h)

Figure 19: Calculated specific activities of 14C in thermal column graphite as a function of distance
from the core. The figure also contains the measured specific activities of the collected seven
material samples. [Publication II]

47
Material-spesific activities in FiR 1 decommissioning waste

Figure 19 also contains the measured activities and dose rates of the collected
samples. The exact location of the samples in the old thermal column was not
recorded during dismantling. However, when the samples were collected, dose
rates from the long graphite blocks were measured. Based on these dose rate
measurements, the samples were determined to have been collected from
graphite blocks around the middle of the thermal column. Specific activities of
the samples in Figure 19 have been plotted based on an approximate
assumption of the sample locations. Comparing the measured and calculated
specific activities, a good correspondence is seen. However, without data of the
exact sample locations, this correspondence does not proof that the
calculations are correct.

The activities in FiR 1 graphite were studied also by performing elemental


analysis of inactive graphite samples and radiochemical measurements of the
most important nuclides, as described in Publication IV.
Graphite is known to be hard to dissolve; therefore, different dissolution
and extraction methods were tested for graphite samples. Eventually, full
solution required using HClO4, which enabled measuring all other nuclides
but chlorine. The measured composition was used to re-perform the activation
calculations for the thermal column graphite. The results were compared with
the activities of 14C and 152Eu previously measured and reported in Publication
II. In theory, the activity fractions between the most important nuclides 152Eu
and 14C should remain about the same. Table 9 compares the measured and
calculated values of samples from various distances from the reactor core.

Table 9: Comparison of activity fractions 14C/152Eu in thermal column graphite. [Publication IV]

Measured Calculated
Average 0.44 0.47
Standard deviation 0.12 0.14
Relative standard deviation (%) 28 30

The difference between measured and calculated values was around 7%. This
is acceptable, especially considering the quite high standard deviations of the
measurement series.
The fractions of specific activity between 36Cl and 14C were also compared
with earlier assumptions. The average value was 36Cl/14C=0.0012 with a
relative standard deviation of 4.7%. However, due to uncertainties in the
chlorine impurity measurements, this result still contains very large
uncertainties.
VTT collected more samples from the dismantled thermal column in
summer 2019. The plan is to continue the measurements using a full
combustion method to ensure that all of the chlorine is released.

48
While ICP-MS measurements provided data on the inactive initial
composition of graphite, radiochemical measurements of difficult-to-measure
radionuclides from active samples are also reported in Publication IV. This
included the radiochemical procedure for the analysis of 3H, 14C, 36Cl, 55Fe and
63Ni in seven graphite samples from the graphite thermal column.

The overall procedure consisted of oxidative acid digestion, in which


chlorine was oxidized to Cl2 and carbon was oxidized to CO2. The fractions
were then separated using exchange resins and, finally, measurements were
carried out using a liquid scintillation counter.
The results were compared with both the oxidizer measurements of 14C
performed earlier and calculated results of the nuclide vector. The results also
show good agreement between the radiochemically analysed and earlier
oxidizer-measured 14C results. The results also show that volatile 3H and 14C
are present in the graphite; therefore, in further studies, sampling needs to be
carefully carried out in order not to lose volatile radionuclides, for example
due to heating of the sample during drilling. The fractions of 36Cl/14C activity
in the samples were between 0.04 and 0.06. These are on the same magnitude
as the reference value 36Cl/14C=0.08 used in the current final disposal
estimates. However, as chlorine can be heterogeneously distributed in the
material, further samples are required to verify the activities.

5.5 FLUENTAL
Fluental is neutron moderator material designed to modify the neutron
spectrum optimal for BNCT. It is a mixture of 30% metallic Al (by weight),
69% AlF3, and 1% LiF [Savolainen et al., 2013] [Auterinen & Hiismäki, 1994].
Due to the large neutron absorption cross section of 6Li, Fluental produces a
huge amount of tritium via the reaction 6Li(n,α)3H, which accounts for a very
high proportion of total decommissioning waste activity [Publication I]. This
is almost the only active radionuclide in Fluental. Moreover, tritium emits only
low energy beta particles with an average energy of 5.7 keV, which cannot be
easily measured and whose health effects are relatively small. However, due to
its biological mobility, tritium is often considered as a toxic substance.
Final disposal of Fluental has special chemical aspects, which are discussed
further in Chapter 7.

The activation of lithium in Fluental was studied in Publication IV by reporting


the material composition measurements and a series of measurements
performed to study the tritium content of irradiated Fluental. Firstly,
composition of Fluental samples was measured with the HR ICP-MS
technique. The main objective was to determine the exact 6Li concentration in
the material. The composition of seven redundant samples was measured. The
results showed that the 6Li impurity would be around 150 ppm. Other
activating impurities were minor.

49
Material-spesific activities in FiR 1 decommissioning waste

To further validate the tritium calculation method, Fluental samples were sent
for irradiation in a research reactor in Pitesti, Romania, as reported in
Publication IV. The activities of the irradiated samples were then measured at
various instants after the irradiation.

After irradiation, the samples were stored in hermetically sealed containers


with 4 ml of water. Since tritium is known to favour binding with water
[Jacobs, 1968], the aim was that the leaking tritium would immerge into the
water and longer term leakage of tritium from Fluental could be studied by
measuring the samples and tritium in the immersion water after intermediate
time points.
At each measurement time point, tritium in the water and inside the vial
were measured with an oxidation technique at a temperature of 800 0C. Gases
were collected in purified water and tritium activity in water was determined
using the LSC technique. Tritium release during storage was also measured
from the immersion water and air inside the glass vials.

In addition to the tritium in samples, part of the total tritium is assumed to be


diffused into the water and air during storage and transportation. These were
all measured individually and the final results of these activities were summed.
Analysing various time points showed that there was very little uncontrolled
leakage. This is also supported by earlier studies, which indicate that tritium
is known to remain well in aluminium-lithium alloys [Roth et al., 1985]
[Shiraishi & Nagasaki, 1965] [Louthan Jr, 2000]. Moreover, leaked tritium in
the immersion water and in the air inside the containers was minor compared
to the amount of tritium inside the Fluental sample. This indicates that was
tritium is released from Fluental very slowly. Consequently, at the time of
decommissioning the Fluental in the FiR 1 reactor is still expected to contain
most of the tritium formed during reactor operation. Moreover, the leakage of
tritium during intermediate storage is plausibly small.

5.6 LEAD

Lead has been used as a shielding material in the BNCT station of the FiR 1
reactor. High purity (e.g. almost antimony-free) lead was originally chosen for
the BNCT station shielding. The lead was exposed to a neutron flux of around
108 n cm-2s-1 during the years 1997-2015 and, according to the inventory
calculations presented in Publication I, the lead is expected to have a very low
specific activity of less than 10 Bq/g. The nuclide with highest contribution to
the total activity are 110mAg, 108Ag, 60Co and 51Cr.

50
Publication II describes the collection of lead samples from the BNCT station
shielding materials behind the Fluental moderator. Drilling holes were chosen
such that the distance of the samples from the BNCT station neutron channel
is the same, thus representing redundant parallel samples. The location of the
sample drill is illustrated in Figure 20 [Publication II]

Figure 20: Location of lead drilling holes. The numbered items in the figure are: 1) Boral plate, 2)
Aluminum, 3) Fluental moderator, 4) Lead shielding, 5) Lithium carbonate plastic shielding, 6)
Bismuth plate and cone, 7) NatLi RX215 plastic shielding, 8) Li 3.5% enriched RX215 plastic
shielding, 9) Li 7% enriched plastic collimator, 10) Steel encapsulation for lead shielding.
[Publication II]

Impurities of the inactive parts of the drilled lead samples were measured with
the ICP-MS technique. The main impurities were silver, cobalt, chromium,
bismuth and manganese. Antinomy concentration was measured to be below
0.1 ppb. The activated lead samples were measured using an ISOCS gamma
spectrometer. Comparison of calculated and measured values showed a good
correspondence in the specific activity of 110mAg. Other expected neutron-
induced activities in lead, 108mAg, 55Fe, 63Ni and 65Zn, turned out to be below
the minimum detectable activity of the measurement. All the values were
clearly below Finnish clearance levels. Consequently, lead in the BNCT station
can very likely be cleared from regulatory control.

51
Material-spesific activities in FiR 1 decommissioning waste

5.7 LITHIATED PLASTICS

FiR 1 contains lithium-enriched plastic that has been used as a neutron


shielding material in the BNCT irradiation room. The plastic is a porous
mixture of lithium carbonate (Li2CO3) and polyethylene paraffin. The
shielding plastic is estimated to contain some tritium due to thermal neutron
absorption of 6Li. No other significant activities are expected. Nevertheless, a
large part of the tritium may have released from the material, and total
activities may thus be close to clearance limits.
The neutron transport calculation model in Publication I does not contain
a detailed model of the neutron flux and spectrum inside the BNCT irradiation
room, but current estimates [Räty & Koivuranta, 2018] indicate that the flux
can be at highest around 106 n cm-2s-1. However, the fluxes are expected to vary
considerably inside the room. With the assumed neutron flux, computational
estimates show that specific tritium activity in the shielding plastic can be at
highest around a few hundred Bq/g.

Publication II included a study in which plastic samples were measured using


the full combustion method at the Horia Hulubei National Institute for Physics
and Nuclear Engineering (IFIN-HH) in Pitesti, Romania. The studied samples
were drilled from six locations around the room. The measurement setup was
identical to that used to measure Fluental samples in Publication IV.
Specific activities were obtained by incinerating the samples, retaining the
tritiated water, and determining the activities with a liquid scintillation
counter. The measured specific activities varied from 60 Bq/g to 1000 Bq/g.
Shielding plastics were identified to have possible re-use value as shielding
material in other plants in Finland. Due to their very low activity, they were
sold for re-use following the case-specific clearance procedure of the Finnish
Nuclear Act [STUK, 2013a].

52
6 SPENT RESEARCH REACTOR FUEL
INVENTORIES

6.1 DESCRIPTION OF TRIGA FUEL


TRIGA type reactor fuel consists of 20 w-% enriched uranium is a solid,
homogeneous mixture of hydrided uranium-zirconium alloy. The ZrH matrix
has very fast thermal feedback in neutron moderation and enables operating
the reactor in pulse mode. Each element is clad with a 0. 7 mm thick
aluminium or steel can. The rods also have 10.2 cm long reflector graphite
plugs at their ends. The active fuel part of each of the elements is
approximately 3.56 cm in diameter and 35.6 cm long. Aluminium end fixtures
are attached to both ends of the can, making the overall length of the fuel-
moderator element 72.5 cm. The weight of a fuel element is around 3 kg.
TRIGA Mark II type research reactors contain 79 fuel elements in their cores.
[General Atomic, 1962a]

Three types of TRIGA fuel have been used in the FiR 1 reactor:

1) aluminium-clad fuel rods with 8 w-% uranium content,


2) stainless steel -clad (SS) fuel rods with 8.5 w-% uranium, and
3) SS-clad fuel rods with 12 w-% uranium.

The reactor contains three boron carbide (B4C) control rods and a boronated
graphite pulse rod.
The last core configuration is presented in Figure 21. The configuration is
intentionally asymmetrical to provide optimal neutron flux for the BNCT
station.

53
Spent research reactor fuel inventories

Figure 21: Last core configuration of the FiR 1 TRIGA research reactor [Publication III]

6.2 INVENTORY CALCULATIONS


Radionuclides formed in the FiR 1 fuel during its 53 years of operation are
studied in Publication III. Two techniques were applied, Monte Carlo based
continuous energy reactor physics code Serpent 2 [Leppänen et al., 2015] with
internal burnup and depletion calculation routine, and the point-depletion
code ORIGEN-S [Gauld et al., 2011a] using MCNP5 [MCNP, 2003] pre-
calculated three-group neutron fluxes as spectrum weighting factors.

After the final shutdown in the burnup calculation, a decay calculation was
performed emphasizing two time points: 12 hours after final shutdown and 9
months after final shutdown. The first one was chosen to be able to compare
the activities of short lived nuclides and the latter because at that time dose
rates were measured from all the fuel rods and this data could be used to
validate the calculations.
A list of nuclides was compiled for this study based on their importance in
spent nuclear fuel burnup calculation validation, in long-term waste analyses,
health impacts, or high migration probabilities in accident conditions [OECD-
NEA, 2002], [OECD-NEA, 2011], [Nirex, 2004].
The initial fuel composition was based on the fuel vendor’s original
specifications [General Atomics, 1962b]. The calculation model segmented the
fuel elements axially into five parts (see Figure 22). The three middle segments
were fuel and the end segments consisted of the graphite plugs. Graphite
composition was taken from a measurement reported in reference [Tress &
Leibundgut, 1999], except that the boron concentration was adjusted
according to the composition measurement reported in Publication IV.
Aluminium ends at the top and bottom of the element were omitted.

54
Figure 22: Fuel rod with axial segmentation (left) and neutron flux distribution (neutrons/cm2s)
(logarithmic scale) (right). Grey segments at the ends of the fuel rod (left) represent graphite
reflectors. [Publication III]

As explained in Chapter 3, the development of activity inventories in reactor


fuel is governed by the Bateman equations (5). To take into account variations
in neutron flux distribution, the calculation model would have to take into
account all different fuel loading configurations and updates in reactor
structures. Since it would be practically impossible to model thousands of
start-ups and shut-downs with the Monte Carlo method, ORIGEN-S and
Serpent calculations used different simplifications in modelling the operating
history.
The Serpent calculations included radionuclide inventory data for each
individual fuel rod. The reactor operating history, excluding the last four and
half years, was modelled in yearly burnup steps so that the correct annual
energy production was reached. This was studied with iterative process by
extending the period of shorter burnup steps one year at a time and comparing
the effect on nuclide-wise activities. This iterative process indicated that using
a longer period than 2011-2015 of short burnup steps did not have any
significant effect on nuclide-wise inventories.
ORIGEN-S point-depletion calculations were not conducted for each
individual fuel rod. Instead, more approximate rod-type specific results were
calculated for the three different fuel rod types used in the reactor. However,
since computing time is not a limiting factor with deterministic codes,
ORIGEN-S applied the same operating history model as in Publication I.
Moreover, the neutron spectra and fission thermal power were also based on
MCNP5 calculated averaged weighting factors, similarly as in Publication I.

Total activity of all the studied nuclides for the whole core differed by 16%
between ORIGEN-S and Serpent 2. The results were further divided into
transuranium, short-lived (T1/2< 12 months) and long-lived nuclides. For

55
Spent research reactor fuel inventories

illustration, Figure 22 presents the calculation results of the transuranium


isotopes as ORIGEN-S (Serpent - ORIGEN-S / Serpent). For visual reasons,
the figures show the absolute relative differences. However, the nuclides in
which the calculated relative difference (Serpent - ORIGEN-S / Serpent) is
negative, are highlighted with red colour. Nuclides whose activity is less than
1 kBq are neglected.

1E+14 90%
1E+13 80%
1E+12
70%
1E+11

Relative difference (%)


60%
1E+10
Activity (Bq)

1E+09 50%
1E+08 40%
1E+07
30%
1E+06
20%
1E+05
1E+04 10%
1E+03 0%
Am-242m

U-235
U-233
U-234
U-238
U-236
U-232
Pu-242

Pu-241

Pu-239

Pu-238
Pu-240
Cm-243
Cm-244

Cm-242
Am-241

Am-243
Np-239

Np-237

Activity calculated with Serpent Relative difference to ORIGEN-S

Figure 23: Comparison of calculated transuranium isotope inventories nine months after
shutdown. Nuclide-specific activities calculated with Serpent. Red colour in the bars indicates
negative difference between Serpent and ORIGEN-S results. [Publication III]

Differences of the activities of long-lived nuclides with the highest


contribution to the total activity, 137Cs, 90Sr and 144Ce were around 28-37
percent so that the values calculated with ORIGEN-S were generally higher.
Differences between activities for the nuclides contributing 99 % to the activity
of all examined nuclides were also around 30 %. For 153Gd, 244Am and 126Sn the
differences were several hundred percent. With short-lived and metastable
nuclides there was no systematic pattern which code calculated the higher
values.
The main cause for differences in the activity inventory may be due to
differences in neutron flux arising from the use of one fuel loading in the
calculation of the flux for the ORIGEN-S activation model. During the actual
fuel loadings, some fuel elements were removed from the reactor and some
(not necessarily the same number) were put back. The ORIGEN-S activation
model in this work was designed for fast and approximate estimates of fuel

56
inventories. It is based on average rod-type specific values for neutron
spectrum and total neutron flux. Consequently, variations in individual fuel
rods can cause a difference between the approximate ORIGEN-S model and
the more detailed Serpent model. The effect of averaging over each fuel rod
type was examined by calculating the relative standard deviations of nuclide-
specific activities in different fuel rods calculated by Serpent. The standard
deviations were calculated for the nuclides with the largest differences
between Serpent and ORIGEN-S.
Some differences are also possible due to nuclear data and the different way
the branching ratios to isomeric states are determined in the two codes. Some
nuclides (e.g. 242Cm and 232Te) also have a second metastable state, which is
handled in ORIGEN-S only by scaling the activity from the ground state
[Gauld, 2006].
The effect of cross section data libraries was tested by repeating the Serpent
calculation with two libraries. With few nuclides, differences of several dozen
per cent were noticed. Similar results are known to exist with SCALE
calculations [Wieselquist et al., 2017]. Nevertheless, the largest differences
were found with metastable states with a negligible contribution to the total
activity. The nuclides with higher contribution to the total activity had only
minor differences.
Relative standard deviations of the nuclides in different rods varied
significantly from 40 to 140% and they were mainly higher than the differences
in the calculated inventories. Differences in the calculated long-lived nuclides
were higher, but they also contained large differences with calculations
performed with different cross section libraries, e.g. for 119mSn the difference
was around 92%.

6.3 TIME EVOLUTION OF NUCLIDES IN SPENT


NUCLEAR FUEL
The burnup calculation models used in Publication III also included a longer
decay period after final shutdown of the reactor. This time evolution of total
activities in different nuclides during the first 10 000 years is presented in
Figure 24. For visual reasons, total activities have been divided into short- and
long-lived and transuraniums only from Serpent results. For comparison, the
activity of a core containing only natural uranium would be around 0.5 GBq
[IAEA, 2019b].

57
Spent research reactor fuel inventories

1E+17
1E+16
1E+15
1E+14
Activity (Bq)
1E+13
1E+12
1E+11
1E+10
1E+09
1E+08
1E+07
1E-02 1E+00 1E+02 1E+04 1E+06
Decay time (years)

Total Serpent short lived transuranium


long lived Total ORIGEN-S

Figure 24: Time development of fuel inventories after shutdown. For visual reasons, division to
transuranium, short and long lived nuclides is presented only from Serpent results. [Publication III]

Dominating nuclides in the fuel inventories nine months after shutdown are
137Cs, 90Sr, 144Ce, 147Pm, 85Kr, 95Zr, 106Ru, 241Pu, 151Sm, 89Sr and 134Cs, which

contribute over 99% of the total activity of the studied nuclides. Publication III
represents the activities of the 11 nuclides mentioned above over a 10-year
cooling period. Only 144Ce, 147Pm, 95Zr and 89Sr have a half-life of less than a
year, and their contribution to decreased activity is seen clearly.

In a decommissioning project, the fuel is removed from the core after a cooling
period ranging typically from a few months to a few years. Total activities in
FiR 1 fuel decrease two orders of magnitude during the first year after
shutdown, but after that any major changes would take dozens of years.
Consequently, no significant changes in transportation arrangements or
radiation safety procedures can be assumed if the cooling period is extended
by a few years

6.4 SPENT NUCLEAR FUEL DOSE RATE


CALCULATIONS
A fuel examination project was performed in April 2016 by the Idaho National
Laboratory (INL) to determine and document the overall condition of the used
fuel. During the examination, dose rates were measured for each individual
fuel rod at the FiR 1 reactor.

The measurement setup consisted of a Geiger–Müller tube based area monitor


probe AMP-200 by Rotem Industries Ltd. [Rotem, 2019] with a 4.25 x 0.8 inch
(10.8 x 2.0 cm) detector calibrated with a 137Cs source. The measurement

58
configuration included a stand for the fuel element and fixed clear plastic tube
containing the detector fixed in place to keep the distance (both radially and
vertically) constant. The distance from the fuel element surface was
approximately 4 inches (10 cm). Measurements were performed under water
in the reactor tank. The manufacturer claims that the detector energy response
is ±10% [Rotem, 2019].

The data measured by INL was used to compare the results of a separate
photon transport calculation that used element-specific fuel inventories from
both Serpent and ORIGEN-S burnup calculation such that Serpent results
were calculated for each individual rod and ORIGEN-S calculated the results
only rod-type wise dose rats. The calculation was performed for all the
elements individually by modelling elements in underwater conditions and
placing a detector at a 10 cm distance from the surface of the assembly at its
axial midpoint. Serpent calculated the photon fluence and energies, which
were converted to absorbed dose rates using NIST mass energy-absorption
coefficients [NIST, 2019]. The absorbed doses were converted to roentgens
with a coefficient 1 R= 0.00877 Gy [Automess, 2019]. Figures 25 and 26 shows
comparison of the measured dose rates and relative differences between
measured and calculated gamma dose rates (measured - calculated /
measured) in individual fuel elements. The rod-type wise dose rates that were
calculated using ORIGEN-S activity inventories are illustrated with green and
red colour. Their relative difference is compared to the average of all the dose
rate rates in the corresponding rod-type. [Publication III].

180 65%

160 55%
Measured dose rate (R/h)

45%

Relative differece (%)


140
35%
120
25%
100
15%
80
5%
60 -5%
40 -15%
Each column represents an individual rod
Measured dose rate Relative difference (%)

Figure 25: Rod-specific measured dose rates and relative differences between calculated and
measured dose rates in Al-clad rods. Red column represents an average rod-type wise result that
has been calculated with ORIGEN-S inventories. [Publication III]

59
Spent research reactor fuel inventories

300
55%
250 45%
Measured dose rate (R/h)

Relative differece (%)


35%
200
25%
150 15%
100 5%
-5%
50
-15%
0 -25%
Each column represents an individual rod

Measured dose rate


Figure 26: Rod-wise measured dose rates and relative differences of calculated and measured
dose rates in SS clad rods. Light blue bars represent the measured dose rates in 8.5% U-cont.
rods and dark blue in 12% U-cont., respectively. Red and green (8.5% U-cont. and 12% U-cont.,
respectively) columns represent average rod-type wise results that were calculated with ORIGEN-
S inventories. [Publication III]

The average difference between measured and calculated dose rates was
20.8% and the median difference 22.1% with a standard deviation of 19.6%.
The highest difference was 58.7%. Sources of uncertainty were related
especially to uncertainties in fuel inventory calculations and conversion of
photon flux to absorbed dose. Moreover, uncertainties in the detector
electronics and signal collection were not taken into account.

Nevertheless, the measurement also has uncertainties related to, e.g.,


calibration with a single nuclide, and the obtained results were considered
acceptable for practical fuel removal arrangements.

60
7 APPLYING CHARACTERIZATION
RESULTS IN THE
DECOMMISSIONING PROJECT

7.1 PRELIMENARY DISMANTLING PLAN


In 2016 VTT arranged a procurement process to produce detailed dismantling
work instructions and a safety classification scheme in order to obtain
documentation suitable for the procurement of dismantling work. German
company Babcock Noell GmbH (BNG) was contracted as a result.

The work utilized the preliminary activity characterization results presented


in Publication I. As the majority of the reactor structures are inactive, all active
parts of the reactor will be removed first and subsequent work carried out as
normal industrial dismantling. This approach reduces both the radiation
doses to dismantling personnel and enables easier radiation protection
measures. Most active components inside the tank (irradiation ring, graphite
reflector) can be lifted out in separate pieces, and beam tubes will be removed
by wire cutting and pulling out the active rectangular blocks (Figure 27).

Figure 27: Cutting the beam tubes (Figure by BNG)

Main work phases and their estimated durations are listed in Table 10. In total,
the dismantling is estimated to take around one year.

61
Applying characterization results in the decommissioning project

Table 10: Phases of dismantling work and their estimated duration

Phase Work Duration


(weeks)
1 Preparatory measures 4
2 Dismantling of reactor tank internals (high activity) 8
3 Dismantling of cooling circuits and heat exchangers 4
4 Dismantling of beam tubes and active concrete 13
5 Dismantling of rest of the systems (low activity) 6
6 Concluding measures 4
7 Clearance measurements 13
SUM 52

7.2 DOSE RATES AND RADIATION SAFETY PLANNING


A decommissioning project has two overarching radiation safety tasks: (1)
protecting personnel during the dismantling process and (2) ensuring the
long-term safety of decommissioning waste storage.
Because time scales of these tasks are different, different nuclides are of
main importance during these stages. During dismantling, long-lived beta
active nuclides are less important and the main focus is on direct radiation
from gamma active nuclides. In the final disposal stage, the primary focus is
on the limitation of radiation dose to the public from long-lived nuclides
dissolved from packed waste along with natural groundwater
According to the current planning on dismantling [BNG, 2016], after the
removal of spent nuclear fuel, the dismantling process can be divided into two
separate stages:

1) Removing the active components from inside the tank


2) Dismantling the active concrete structures using wire saw and
pneumatic hammers.

An MSc thesis was commissioned in 2017 to study the radiation dose rates
from the main tasks in each of these stages [Haapamäki, 2018]. In addition, to
estimate the radiation doses to workers, the dose rates model was used to
estimate direct dose rates around the building. These calculations do not
necessarily determine the final dose assessment but provide a conceptual basis
for choosing the final waste packages and planning the main parts of the
dismantling process.
The work was based on the results presented in Publication I. The dose
estimations were limited to the main gamma-active nuclides in the activation
products, since the FiR 1 reactor has not had any major contamination

62
accidents during its operating history, and it is assumed that body internal
dose can be avoided with carefully planned personal protection.
The Monte Carlo based MCNP code was used to calculate the external dose
rates around the packages and dismantled components. It was assumed that
there are no leaks of radioactive material from the packages.

The final results show that the irradiation ring, the reflector, and some of the
most active steel parts inside the tank will require a special shielded waste
package and remote handling tools, but all the other waste can be packed with
relatively light procedures (steel barrels etc.) focusing just on contamination
control.
Figure 28 shows as an example the calculated dose rates of the graphite
reflector being lowered to the ground floor with a crane. This calculation case
included only a thin steel case with wall thickness of 2 cm around the reflector.
The final packaging will include additional shielding, e.g., a concrete box.
However, for logistical reasons, some of the heaviest packages have remain on
the ground floor during the dismantling.

Figure 28: Calculated dose rates inside the reactor building from a graphite reflector in a steel
shielding case being lowered with a crane. [Haapamäki, 2018]

The results of dose rate calculations in [Haapamäki, 2018] also included


conceptual calculations of different package types and reactor hall area
classification into green (below 0.025 mSv/h), orange (between 0.025 mSv/h
and 1 mSv/h) and red (over 1 mSv/h) areas based on direct dose rates. The
main results show that the irradiation ring, the graphite reflector, and some
small steel parts require additional shielding, but other decommissioning
waste can be stored in simple barrels or other packages without any special
shielding for external dose rate.

63
Applying characterization results in the decommissioning project

Final waste packaging plan will be a compromise depending on e.g.


radiation safety, compatibility with the waste acceptance criteria of the storage
site, crane capability, logistics both inside the reactor hall and on highway
transportation, etc.
In the MSc thesis also the dose rate at certain locations inside the reactor
hall were studied. Final results show that dose rates in the studied location
vary from 1 to 800 µSv/h depending on the phase of work and location.
Dose assessments with several final waste package and waste storage
options are still on-going in 2019, and the data from these assessments will be
utilized to plan the final radiation safety procedures. While it is very difficult
to provide precise estimates of the duration of each work phase, current
estimates set the total work time at around 20,000 person hours and the
collective dose to the workers from the whole project at less than 10 mmanSv
[TEM, 2017][BNG, 2017b].

Dismantled low- and intermediate level waste from the FiR 1 reactor is
planned to be disposed of in Finland at either one of Finland’s two nuclear
power plant sites. This will plausibly include an intermediate storage time of a
few years during which some of the waste can be free released as their activities
decay naturally. Practical options for intermediate storage of dismantling
waste include storage at either of the Finnish nuclear power plant sites, storage
in the FiR 1 reactor hall, or, alternatively, direct transport to final disposal at
either of the Loviisa or Olkiluoto power plant sites.

The option of storing waste packages inside the reactor hall was also examined
in the commissioned dose rates study [Haapamäki, 2018]. It was found that
to avoid dose to the public outside the building additional shielding is needed
around the most active waste packages, and that this can be carried out
relatively easily with several different shielding material options. Figure 29
presents an example from that study of calculated dose rates from dismantling
waste. In the figure all the components inside the tank have been removed and
packed into shielded packages. The packages are stored inside the reactor and
additional shielding has been built around them using heavy-concrete bars
from BNCT irradiation room. This type of situation is possible, if there is a long
delay before VTT can transport the waste from the FiR 1 reactor site to the final
disposal site either in Loviisa or Olkiluoto. The results demonstrate that the
dose rates both in the working area around the reactor and outside the reactor
hall are below 10 𝜇Sv/h.

64
Figure 29: Calculated dose rates from the stored dismantling waste. [Haapamäki, 2018]

7.3 WASTE MANAGEMENT AND CLASSIFICATION


The main requirements for waste classification arise from national regulations
and final disposal site waste acceptance criteria (WAC). This has an effect on
dismantling methods, packing, measurements, transportation and both long-
term and intermediate storage of waste. National approaches differ
significantly [IAEA, 1994]. A schematic classification based on specific
activities and half-lives is presented in Figure 30 [IAEA, 2009].

Figure 30: Classification of nuclear waste [IAEA, 2009].

65
Applying characterization results in the decommissioning project

Spent nuclear fuel is the only high-level waste material in the FiR 1 reactor.
The total activity of the other decommissioning waste is estimated to be
around 4.24 TBq. Estimated material-wise activities were listed in Table 4.
According to current estimates, most of the decommissioning waste can be
classified as low-level waste, but some components close to the core will be
intermediate level waste. The ‘other’ category contains e.g. plastic shielding
materials and secondary waste, such as protective clothing, filters,
contaminated dismantling equipment (drills, saws, concrete sludge, etc.) and
decontamination fluids.

In practice, waste sorting will consist of a two-stage process utilizing the


scaling factors (Publications II and IV) determined prior to dismantling [Räty
et al., 2018], as follows:

1) Preliminary survey of each component of dismantled piece of waste.


2) Monitoring for final classification and compliance with clearance levels.

The preliminary survey facilitates segregating the materials into streams that
have the potential to be cleared and streams to be sent to final disposal. The
preliminary survey also provides information on the spatial distribution of
radioactive contamination and on the radionuclides present. This will be
performed with quick hand-held dose rate and contamination measurement
devices.

Monitoring for final classification and compliance with clearance levels will be
performed by measuring the waste packages with a gamma spectrometer. The
effect of self-absorption will be calculated by considering the dimensions and
weight of the batch of material. The activities of other than the measured
nuclides will be obtained through scaling factors. Final bookkeeping will
include the total activity, waste nuclide vector and external dose rate for each
waste package. Samples from each waste package will also be collected to
enable later verification of the results.

7.4 CLEARANCE OF MATERIALS

Clearance levels are defined as the set of values, established by the radiation
safety authority, expressed in activity concentrations at or below which a
source of radiation can be released form regulatory control. Nuclide-wise
clearance levels are defined in STUK YVL guide D.4 [STUK, 2013a].
Nuclear waste may be cleared from regulatory control following a general
or a case-specific procedure. In a general clearance procedure, the destination
of the materials released from the facility need not be designated, or is only
designated in its outline, and the activity levels to be applied are fixed. In a
case-specific clearance procedure, the recipient of the material and the waste

66
management process must be defined. Slightly higher specific activities than
in general clearance procedure are allowed, but a limit on total activity is still
applied [STUK, 2013b].

If the specific activities of key nuclides are below the general clearance levels,
the activity concentration in materials to be cleared may be averaged over a
maximum of 500 kg of material; however the activity in any single item with a
mass of less than 30 kg may not exceed the value obtained by multiplying the
respective activity concentration level by a factor of 30 kg. In the case of items
with masses higher than 30 kg, the measurements may be performed for
blocks of a few hundred kg.

Based on the results of Publication I, Table 11 lists the most important nuclides
that contribute to the total activities of FiR 1 decommissioning waste. The
nuclide-wise general clearance levels are also listed.

Table 11: Key activating nuclides. [Räty, 2017]

Nuclide Half-life General clearance Material


level (Bq/g)
3H 12.3 years 100 Fluental, plastic
14C 5730 years 10 Graphite, concrete
60Co 5.3 years 1 Graphite, concrete, steel,
aluminium, lead
55Fe 2.73 years 1000 Steel, aluminium, concrete
152Eu 13.5 years 0.1 Graphite, concrete
154Eu 8.59 years 0.1 Graphite, concrete
63Ni 101 years 100 Aluminium, steel, lead,
concrete
59Ni 76 000 years 100 Steel, Aluminium
65Zn 243 days 1 Aluminium
108mAg 418 years 1 Lead, bismuth
210Po 138 days 0.1 Bismuth
125Sb 2.76 years 1 Lead
121mSn 43.9 years 10 Lead
210mBi 3 000 000 10 Bismuth
years
110mAg 249 days 1 Bismuth
39Ar 269 years 10 Concrete
133Ba 10.7 years 10 Concrete
151Sm 96 years 1000 Concrete
36Cl 301 000 years 1 Graphite

Based on the activities calculated in Publication I, most of the concrete, lead


and aluminium parts in the FiR 1 reactor can be released from regulatory

67
Applying characterization results in the decommissioning project

control. This applies especially to the structures farther from the reactor core.
Whether case-specific clearance and delayed clearance will be applied,
depends on the final waste final disposal option and finding suitable re-use for
the materials. [Ruokola, 2016] [Räty, 2017]

7.5 FINAL DISPOSAL


The decommissioning strategy assumes that the decommissioning waste from
the FiR 1 reactor will be placed in one of the final repositories for low- and
medium-level waste owned by the Finnish nuclear power companies [TEM,
2017]. The main requirement in the FiR 1 decommissioning project waste
management plan is to characterize and pack the waste such that they fulfil all
the requirements related to final repository waste acceptance criteria. In
practice, this means that the FiR 1 decommissioning waste results in only a
minor addition to the radiation impacts caused by other wastes in the
repository and thus does not exceed the dose constraints set for the final
disposal sites. In addition to radiation dose, safe final disposal also requires
that the waste does not cause such chemical effects on other wastes or on the
barrier system that would enhance radionuclide release or transport.

The engineered barrier system in the final underground repositories,


consisting of waste packages, sealing of the repository, etc., creates conditions
that effectively limit the release of radioactive substances from the
repositories. The large amount of concrete in the repository creates long-
standing alkaline conditions where the corrosion of steel and dissolution of
minerals is slow.
Use of suitable waste packages is of critical importance. The planned waste
packages need to provide sufficient shielding during the dismantling process
and transportation. In addition, they need to be compatible with the power
plant waste acceptance criteria and, ideally, would be directly suitable for final
disposal. Most typical packages are steel barrels and concrete and steel boxes
of different sizes and wall-thicknesses. A number of waste package options
have been examined in the dismantling planning documentation [BNG,
2017a] and in an MSc thesis conducted in 2018 [Haapamäki, 2018].

Finnish final repositories will include much larger quantities of radioactive


concrete and steel from the decommissioning waste of power reactors.
Consequently, the contribution of FiR 1 steel and concrete waste is minor
compared to the total waste volume and nuclide-wise activities of these waste
in the final disposal repository. However, the inclusion of FiR 1
decommissioning waste will yield a small increase in activities of at least 3H,
14C and 36Cl. Of these, 14C and 36Cl are especially important for long-term

safety. Publication IV focuses on materials that are not used in Finnish nuclear

68
power plants, namely Fluental and graphite, and these have consequently not
been studied as thoroughly.

The main material in Fluental moderator is aluminium. The chemical


behaviour of FiR 1 aluminium in active waste final disposal conditions has
been studied at VTT in a reference study [Carlsson et al., 2014]. The main
challenge is corrosion via reaction as presented in [Zhang et al., 2009]:

Al + 3H2O + OH- = 3/2 H2 + Al(OH)4 (16)

The reaction produces heat and generates hydrogen pressure. In principle,


these can accelerate the diffusion of radionuclides in the bedrock [Carlsson et
al., 2014]. However, according to a publicly reported study, due to the very low
corrosion rate the effect is acceptable considering the small amount of waste
in question [Harriague et al., 2007].

The composition of Fluental was measured in Publication IV. The activities of


all the other nuclides except 3H are negligible. 3H results from the reaction
6Li(n,α)3H. As concluded in earlier chapters, the radionuclides in other

aluminium parts (54Mn, 60Co, 65Zn, 55Fe, 63Ni) are the same as in steel parts,
and consequently their incremental effect is minor. Since tritium is also
relatively short-lived from the point of final disposal, even the high total
activity of Fluental does not present any particular challenge to final disposal.

Activities in FiR 1 graphite were calculated in Publication I. In the final


disposal repositories, waste is kept in saturated alkaline conditions. In these
conditions the graphite matrix remains stable, but the leaching rates of long-
lived radionuclides 14C and 36Cl are important considerations [Carlsson et al.,
2014].
36Cl is formed in the graphite, because in the manufacturing process of

reactor grade graphite, bulk graphite is purified from impurities with high
neutron absorption cross sections (especially boron). This is usually
performed with a chemical purification process using halogens, which
volatilize impurity metals into halides. Various gases, such as sulphur
hexafluoride, carbon tetrafluoride, and carbon tetrachloride, have been used
in this process, and for example Freon-12 or chlorine have been most widely
used in the US [Chapman, 2010] [Nightingale, 1962] [Lee et al., 2015]. The
reaction is presented in reference study [Nightingale, 1962] as:

𝐵 𝑂 + 3𝐶 + 3𝐶𝑙 → 2𝐵𝐶𝑙 + 3𝐶𝑂 (17)

Traces of 35Cl can activate via the reaction 35Cl(n,γ)36Cl into 36Cl with a very
long half-life of around 300,000 years [IAEA, 2016]. In the preliminary
analyses the inventory of 36Cl in FiR 1 graphite was estimated based on the
calculated inventories in the Romanian TRIGA 14 MW research reactor. The

69
Applying characterization results in the decommissioning project

calculated inventories show that the ratio of 36Cl/14C is roughly 0.07-0.08.


[Bislinghoff et al., 2000]. A ratio of 36Cl/14C=0.08 was ultimately used in the
preliminary phases of decommissioning planning [Kotiluoto & Räty, 2016].

14C is an especially important nuclide in waste disposal because it spreads


rapidly in the biosphere and can be heterogeneously distributed in the
graphite matrix [Norris, 2015].
In the final disposal conditions 14C can divide into organic form as CH4
(almost non-sorbing) and inorganic form as CO2 (more sorbing). The organic
form of CH4 spreads more rapidly in the final disposal conditions. The
amounts of 14C in each fraction, the release rate of the slow-releasing fraction,
and the speciation of 14C released depend on the disposal conditions (pH and
presence of oxygen) and partially on the characteristics of the graphite itself.
Under high pH conditions, the diffused inorganic 14C is retarded by
precipitation and sorption, while organic 14C and gaseous 14C are diffused out
from the repository system without any significant retardation. It has been
reported that in the R1 research reactor in Stockholm, Sweden, the fractions
of inorganic CO2 and organic CH4 are around 65% and 35%, respectively, and
release to gaseous phase was low (< 0.1%) [Magnusson, 2002]. Current final
disposal analyses at the FiR 1 reactor assume conservatively that all of the 14C
radionuclide inventory will release in organic form as CH4 [Nummi, 2018]
[Pitkäoja, 2019]

36Cl is released by two mechanisms: initial rapid diffusion (labile fraction),


followed much slower release. On the whole, the proportion of labile fraction
varies from a few per cent to 90%, mainly depending on the initial
temperatures [Carlsson et al., 2014]. Given that there is a cooling period of
several years prior to final disposal, in practice only a small portion of 36Cl is
expected to be in labile form. Current estimates of the final disposal of FiR 1
graphite assume (highly) conservatively that all of the 36Cl is released in labile
form. Considering the package plan, geometries and surface areas of the
graphite components, this indicates that almost all 36Cl would be diffused from
the graphite during the first 30 years. This is a highly conservative assumption
in respect to the diffusion of 36Cl in the graphite matrix.
Current estimates show that the disposal of FiR 1 decommissioning wastes
into Finnish final repositories does not exceed the dose constraints set for the
final disposal sites [Nummi, 2018] [Pitkäoja, 2019]. However, these estimates
are based on calculated nuclide inventories, which still have the uncertainty
that diffusion of gaseous nuclides cannot be taken into account. Another
source of uncertainty is that because exact material specifications are not
available, initial compositions and the homogeneity of the impurity
distributions are partially based on assumptions. These have been taken into
account by assuming conservatively high impurities, but the results need to be
verified with systematic measurements before and during the dismantling
work.

70
8 SUMMARY AND CONCLUSIONS

The focus of this thesis is to study the radiological properties of the FiR 1
TRIGA research reactor structure and spent nuclear fuel and their essential
connections to other aspects in a decommissioning project.
Characterization in the FiR 1 decommissioning project is based on a
calculation system that combines neutron flux calculated with the MCNP code
with irradiation history modelled with ORIGEN-S. The methodology is
presented in Publication I. The results have been preliminary validated with
measurements presented in Publications II and IV. The method has been
further validated with Monte Carlo calculations in Publication III by
comparing fuel inventories calculated with two methods and by comparing
calculated dose rates with measured values.
Detailed results have been published in technical reports for VTT’s internal
use and the articles presented in this thesis discuss the main methods,
phenomena and some examples of more detailed data. Several other VTT
technical reports are also referred to.

The decommissioning waste of the FiR 1 reactor consists mainly of concrete,


aluminium, steel and graphite. Computational estimates in Publication I show
that most of the reactor structural parts are inactive and can be released from
regulatory control. Activated parts consist mainly of steel and aluminium parts
that have undergone neutron irradiation near the core region, and concrete
near the core or around the beam tubes. Conservative assumptions have been
used, especially with respect to material compositions and the operational
history of the beam tubes.

The measurements in Publication II and IV support the calculated results and


indicate that the calculations have been conservative as intended, although
some components are inaccessible prior to removal of spent nuclear fuel. Some
activities have also been compared with publicly reported results from other
TRIGA type reactors. Considering their distinct characteristics, differences in
total inventories were reasonable.
Main uncertainties in decommissioning waste characterization are related
to detailed material compositions, as activating nuclides are typically small
impurities in materials and their concentrations are not necessarily taken into
account in construction specifications.

The burnup calculations of spent fuel elements in Publication III demonstrate


the effect of different approximations on calculated estimates of fuel nuclide
inventory. Differences between total activities calculated with the two codes
were small, but for some especially metastable nuclides calculated models
have very high uncertainties. Main sources of uncertainty are related to the

71
Summary and conclusions

complicated research reactor operating history, which requires simplification,


especially with Monte Carlo methods. Nevertheless, some uncertainties are
also related to nuclear cross section data and the model of branching ratios in
the code. Comparison with measured dose rates provided reasonable
correspondence for practical radiation safety and fuel transport planning, but
the calculation still contains uncertainties related to photon flux to dose rate
conversion and in the simplified measurement setup.

Since 1960’s General Atomic has installed 69 TRIGA type research reactors in
the world. 31 of them are currently decommissioned or permanently shut
down. [IAEA, 2020]
Compared to industrial power reactors, the special aspects in research reactors
set boundary conditions to decommissioning planning. These include
especially different material, more complicated applications and operation
history and several structural changes during their lifetime.
Current international experience shows that planning and preparation for
decommissioning before shutdown in most research reactors has been
insufficient. Plans for decommissioning have been usually on a conceptual
level and infrastructure for spent nuclear fuel and decommissioning waste has
been either missing or inadequate. In some cases, there have also been
challenges with inadequate funding, defective decommissioning-orientated
regulations and organizational human factors. [IAEA, 2002b]
From mechanical point of view, executing the dismantling is a relatively
well-established process, but the responsibility on radiation safety cannot be
transferred to a third party de-construction company. Designing radiation
safety and waste management procedures requires good understanding of
activities in the reactor structures. IAEA’s recommendation is that the
radiological characterization and completing nuclide vectors of the facility
should be carried out well before final shutdown. This enables a smooth
transition to from reactor operation to decommissioning. [IAEA, 2006b]

The main lesson learned in the FiR 1 reactor activity characterization is to


maintain a holistic approach focusing on waste end point issues and
regulations. Compared to the international situation, Finland has final
disposal repositories available and funding is secured via state waste
management fund. Nevertheless, current work has indicated that e.g.
collecting material data and samples much earlier would have contributed to
the quality of results and active communication with other stakeholders,
especially power companies and the regulator is important to establish
compatible measurement and data management methods.
The activity inventory estimates presented in this study are currently used
in detailed radiation safety and dismantling planning. Moreover, the work is
gradually shifting towards larger scale implementation in actual dismantling
and waste management. Since they both are heavily dependent by chosen
waste storage and its waste acceptance criteria, communication and co-

72
operation with power companies and the regulator will be even more valuable
in the future.

Future work in characterization requires completing the nuclide vectors with


materials that have been currently inaccessible. An important consideration is
to base all calculations on measured material compositions and ensure that
the measured samples reliably represent a larger amount of decommissioning
waste.
For scientific interests, building a deterministic full-core model of the
reactor could provide more data for more systematically testing the effects of,
e.g., cross section libraries and operating history approximations between
Monte Carlo and deterministic models. Moreover, a more detailed
measurement setup could provide better data for dose rate calculations.

Universally, the reactor-specific key aspects of radiological characterization


are geometry, construction, and material properties. The well-established
decommissioning techniques used in research reactors are also applicable
throughout the nuclear industry. The methods studied here thus provide
valuable knowledge for VTT for applications in other future nuclear
decommissioning projects.

73
References

REFERENCES

[Ackermann, 2017] T. Ackermann - Methods for the radiological


characterisation of the FiR 1 TRIGA research reactor decommissioning waste.
Master’s Thesis, Stellenboch University, South Africa, 2017

[Agilent, 2019] Agilent Technologies Inc - 5100 ICP-OES spectrometer,


technical features. [online] [access 2019, October 23];
https://www.agilent.com/en/products/icp-oes/icp-oes-systems/5100-icp-
oes

[Airila et al., 2019] M. Airila, I. Auterinen, P. Kotiluoto, A. Räty - Lessons


learned during planning and first phases of decommissioning of the Finnish
TRIGA FiR 1. In Proceedings of Research Reactor Fuel Management
Conference, RRFM 2019, Jordan, March 24-28, 2019

[Auterinen et al., 1998] I. Auterinen et al. - Metamorphosis of a 35 Years


Old TRIGA Reactor into a Modern BNCT Facility. In Proceedings of the Eighth
International Symposium on Neutron Capture Therapy for Cancer, La Jolla,
California, USA, September, 1998

[Auterinen & Hiismäki, 1994] I. Auterinen & P. Hiismäki - Design of an


epithermal neutron beam for the TRIGA reactor in Otaniemi. In: I. Auterinen,
and M. Kallio (eds.): Proceeding of the CLINCT BNCT Workshop, Helsinki
1993. TKK-F-A718 Technology. Helsinki University of Technology. Helsinki,
2–4, 1994

[Automess, 2019] Automess GmbH, Radiation Quantities and Units,


[Online] [access 2019, August 6];
https://www.automess.de/Messgroessen_E.htm

[Bislinghoff et al., 2000] B. Bisplinghoff, M. Lochny, J. Fachinger, H.


Bruecher - Radiochemical characterisation of graphite from Juelich
experimental reactor (AVR). Nuclear Energy 39(5):311-315, 2000

[BNG, 2016] Babcock Noel GmbH, Dismantling Waste Management


Planning of Decommissioning FiR1 TRIGA Mk II: Work Instruction 05,
Dismantling of the biological shield with beam tubes and reactor tank, FIR-
INS-BNG-1005, 2016 (confidential)

[BNG, 2017a] Babcock Noel GmbH, Dismantling Waste Management


Planning of Decommissioning FiR1 TRIGA Mk II: List of waste packages,
Technical report FIR_LIST_BNG_1001_Rev4, 2017 (confidential)

74
[BNG, 2017b] Babcock Noel GmbH, Dismantling Waste Management
Planning of Decommissioning FiR1 TRIGA Mk II: Verification of radiation
protection, Technical report FiR-TRE-BNG-1002_Rev3, 2017 (confidential)

[Canberra, 2019a] Canberra Inc. - In Situ Gamma Spectroscopy with


ISOCS, an In Situ Object Counting System. Canberra Application Note
[Online] [access 2019, April 22];
http://www.canberra.com/literature/gamma_spectroscopy/application_not
es/InSitu-ISOCS-M2352.pdf

[Canberra, 2019b] Canberra Inc. - GenieTM 2000 Basic Spectroscopy


Software. Canberra Industries [Online] [access 2019, April 22];
www.canberra.com/products/radiochemistry_lab/.../G2K-BasicSpect-SS-
C40220.pdfG2K-BasicSpect

[Carlsson et al., 2014] T. Carlsson et al. - Chemical aspects on the final


disposal of irradiated graphite and aluminium. VTT Technology 156, ISBN
978-951-38-8095-8, 2014

[Chapman, 2010] J. Chapman - Testing of Small Graphite Samples for


Nuclear Qualification. Master’s Thesis, Idaho State University, USA, 2010

[Charters & Aggarwal, 2006] G. Charters, S. Aggarwal - Material sample


collection with tritium and gamma analyses at the university of Illinois nuclear
research laboratory TRIGA nuclear reactor. In Proceedings of WM’06
conference, Tuscon, Arizona, USA, 26.2.-2.3., 2006

[Chatzis, 2016] I. Chatzis - Decommissioning and environmental


remediation: an overview, IAEA Bulletin, April 2016

[Choppin et al., 2013] G. Choppin, J. Liljenzin, J. Rydberg, C. Ekberg -


Radiochemistry and nuclear chemistry (4th ed.). Amsterdam, Netherlands,
Elsevier, 2013

[Evans et al., 1984] J.C Evans et al. - Long-Lived Activation Products in


Reactor Materials. U.S Nuclear Regulatory Commission, NUREG/CR-3474,
1984.

[Finlex, 1987] Nuclear Energy Act with all amendments implemented


[Online] [access 2019, October 13];
(https://www.finlex.fi/fi/laki/ajantasa/1987/19870990) (in Finnish)

75
References

[Finlex, 1988] Nuclear Energy Decree with all amendments implemented


[Online] [access 2019, October 13];
https://www.finlex.fi/fi/laki/ajantasa/1988/19880161 (in Finnish)

[Gauld, 2006] I. Gauld, Validation of ORIGEN-S Decay Heat Predictions


for LOCA Analysis, In proceeding of PHYSOR-2006, ANS Topical Meeting on
Reactor Physics, Vancouver, Canada, 2006 September 10-14

[Gauld et al., 2011a] I.C. Gauld et al. - ORIGEN-S: A Scale System Module
to Calculate Fuel Depletion, Actinide Transmutation, Fission Product Buildup
and Decay, and Associated Radiation Source Terms. Oak Ridge National
Laboratory, ORNL/TM-2005/39, version 6.1, 2011.

[Gauld et al., 2011b] I. C. Gauld et al. - Isotopic Depletion and Decay


Methods and Analysis Capabilities in SCALE. Nuclear Technology, 174:169-
195, 2011

[General Atomic, 1962a] Gulf General Atomic - The 100-kW TRIGA Mark
II Pulsing Reactor FiR-1, A Technical Description. 1962

[General Atomic, 1962b] Gulf General Atomic - Specifications: Individual


TRIGA Fuel Elements, FiR 1-4th fuel shipment. Technical description, 1962
(confidential)

[Haapamäki, 2018] A. Haapamäki - Säteilysuojelu ydinlaitoksen


käytöstäpoistossa: Case FiR1. Master’s Thesis, Lappeenranta University of
Technology, 2018

[Hampel et al., 2002] G. Hampel et al. - Calculation of the Activity


Inventory for the TRIGA Reactor at the Medical University of Hannover
(MHH) in Preparation for Dismantling the Facility. In proceedings of WM´02
Conference, Tuscon, USA, 2002

[Harriague et al., 2007] S. Harriague, C. Barberis, E. Cinat, C. Grizutti, H.


Scolari - Disposal aspects of RA-1 research reactor decommissioning waste. In
International Atomic Energy Agency IAEA - Disposal aspects of low and
intermediate level decommissioning waste: results of a coordinated research
project 2002–2006, IAEA TECDOC 1572, IAEA, Vienna, 2007

[Hou et al., 2005] X. Hou, L.F. Ostergaard, S.P. Nielsen - Determination of


63Ni and 55Fe in nuclear waste samples using radiochemical separation and
liquid scintillation counting. Analytica Chimica Acta, 535:297-307 , 2005

76
[IAEA, 1994] International Atomic Energy Agency IAEA -
Decommmissioning Techniques for Research Reactors. IAEA Technical
Report Series 373, 1994

[IAEA, 1998] International Atomic Energy Agency IAEA - Radiological


Characterization of Shut Down Nuclear Reactors for Decommissioning
Purposes. IAEA Technical Report Series 389, 1998

[IAEA, 2002] International Atomic Energy Agency IAEA -


Decommissioning techniques for research reactors: Final report of a co-
ordinated research project 1997–2001, IAEA-TECDOC-1273, 2002

[IAEA, 2006a] International Atomic Energy Agency IAEA - Return of


Research Reactor Spent Fuel to the Country of Origin: Requirements for
Technical and Administrative Preparations and National Experiences. Proc. of
a Tecnical Meeting, IAEA-TECDOC-1593, 2006

[IAEA, 2006b] International Atomic Energy Agency IAEA -


Decommissioning of Research Reactors: Evolution, State of the Art, Open
Issues, Technical Report Series No. 446, 2006

[IAEA, 2009] International Atomic Energy Agency IAEA - Classification of


Radioactive Waste. IAEA Safety Guide GSG-1, 2009.

[IAEA, 2018] International Atomic Energy Agency IAEA - Research


Reactor Database. [Online] [access 2019, May 29];
https://nucleus.iaea.org/RRDB/RR/

[IAEA, 2016] International Atomic Energy Agency IAEA - Processing of


Irradiated Graphite to Meet Acceptance Criteria for Waste Disposal. IAEA
Tecdoc 1790, 2016

[IAEA, 2019a] International Atomic Energy Agency IAEA - Methodologies


for Assessing the Induced Activation Source Term for Use in
Decommissioning Applications. IAEA Safety Report 95, 2019

[IAEA, 2019b] International Atomic Energy Agency IAEA - What is


uranium? [Online] [access 2019, September 3];
https://www.iaea.org/topics/spent-fuel-management/depleted-uranium

[IAEA, 2019c] International Atomic Energy Agency IAEA - Evaluated


Nuclear Data File (ENDF) database. [Online] [access 2019, October 14];
https://www-nds.iaea.org/exfor/endf.htm

77
References

[IAEA, 2020] International Atomic Energy Agency IAEA – Research


reactor database. [Online] [Access 2020, January 12];
https://nucleus.iaea.org/RRDB/RR/ReactorSearch.aspx

[Jacobs, 1968] D.G. Jacobs - Sources of tritium and its behavior upon
release to the environment. Oak Ridge National Laboratory report, 1968

[JAEA, 2019] Japan Atomic Energy Agency JAEA - Graphs of Cross


Sections given in JENDL-4.0. [Online, access 2019, April 29]
https://wwwndc.jaea.go.jp/j40fig/j40fig03.html,

[Kankaanranta et al., 2012] L. Kankaanranta et al. - Boron neutron capture


therapy in the treatment of locally recurred head-and-neck cancer: Final
analysis of a phase I/II trial. International Journal of Radiation Oncology
Biology Physics, 82(1):67–e75, 2012.

[Kotiluoto, 1997] P. Kotiluoto - BNCT-hoitohuoneen mallinnetut


säteilyarvot. Säteilysuojeluraportti, VTT Kemiantekniikka,
Ydinreaktoriryhmä, BNCT-hanke, 28.8.1997.

[Kotiluoto & Räty, 2016] P. Kotiluoto, A. Räty - FiR 1 activity inventories


for decommissioning planning. VTT Technical Report VTT-R-03599-16, 2016

[Lauridsen, 2017] K. Lauridsen, Advances and Innovations in Nuclear


Decommissioning, Woodhead Publishing Series in Energy, 2017

[Lee et al., 2015] S-M Lee et al. - Bulk graphite: materials and
manufacturing process. Carbon Letters, 16(3):135-146, 2015

[Leppänen et al., 2015] J. Leppänen, et al. - The Serpent Monte Carlo code:
Status, development and applications in 2013. Annals of Nuclear Energy,
8:142-150, 2015

[Leskinen et al., 2020] A. Leskinen, S. Salminen-Paatero, A. Räty, M.


Tanhua-Tyrkkö, T. Iso-Markku, E. Puukko - Determination of 14C, 55Fe, 63Ni
and gamma emitters in activated RPV steel samples - a comparison between
calculations and experimental analysis. Journal of Radioanalytical and
Nuclear Chemistry, 323:399-413, 2020

[Love et al., 1995] E.F. Love et al. - Use of MCNP for characterization of
reactor pressure vessel internals from decommissioned nuclear reactors.
Idaho National Engineering Laboratory, INEL-95/0419, 1995

[Louthan Jr, 2000] M.R. Louthan Jr., Aluminum-Lithium Technology and


Savannah River’s Contribution to Understanding Hydrogen Effects in Metals,

78
Symp. Proc. 50 of Excellence in Science and Engineering at Savannah River
Side, WSRC-MS-2000-0006, Aiken, 2000

[Magnusson, 2002] Å. Magnusson, Measurement of the distribution of


organic and inorganic 14C in a graphite reflector from a Swedish nuclear
reactor. Department of Physics, Lund University, Sweden, Report 01/02,
LUNFD6/(NFFR-5016)/1-60, 2002

[MCNP, 2003] X-5 Monte Carlo Team, MCNP – A General Monte Carlo N-
Particle Transport Code, Version 5, Los Alamos National Laboratory, LA-UR-
03-1987, 2003

[National Carbon Company, 1955] Typical properties – reactor graphite.


Information sheet from National Carbon Company, December 12, 1955

[Nightingale, 1962] R.E Nightingale (ed.) - Nuclear graphite. Academic


Press, New York, ISBN 978-1-4832-2854-91, 1962,

[Nirex, 2004] United Kingdom Nirex Ltd. - The Identification of


Radionuclides Relevant to Long-Term Waste Management in the United
Kingdom. Nirex Report no. N/105, 2004

[NIST, 2019] NIST Standard Reference Database 126, Mass Attenuation


Coefficients and Mass Energy-Absorption Coefficients. [Online] [access 2019,
May 6]; https://www.nist.gov/pml/x-ray-mass-attenuation-coefficients

[Norris, 2015] S. Norris (ed.) - WP5 Review of Current Understanding of


Inventory and Release of C14 from Irradiated Graphite (D5.5). EU project
CAST (CArbon-14 Source Term) deliverable D5.5. 2015

[Nummi, 2018] O. Nummi - Prelimenary Safety Assessment for the FiR1


Decommissioning Waste. Fortum Power and Heat Technical Report NUCL-
03608, 2018 (confidential)

[OECD-NEA, 2002] OECD-NEA - Chernobyl: Assesment of Radiological


and Health Impacts. OECD-NEA Radiation Protection report, 2002. [Online]
[access 2019, February 14]; http://www.oecd-nea.org/rp/chernobyl/c02.html

[OECD-NEA, 2011] OECD-NEA - Spent Nuclear Fuel Assay Data for


Isotopic Validation - State-of-the-art Report.
NEA/NSC/WPNCS/DOC(2011)5, June 2011. [Online] [access 2019, February
14]; www.oecd-nea.org/science/wpncs/ADSNF/SOAR_final.pdf

79
References

[OECD-NEA, 2014] OECD-NEA - R&D and Innovation Needs for


Decommissioning Nuclear Facilities. NEA Working party report No. 7191,
2014

[Pantelias & Volmert, 2015] M. Pantelias, B. Volmert - Activation


Neutronics for a Swiss Pressurized Water Reactor. Nuclear Technology,
192:278-285, 2015

[Park et al., 2003] S.K. Park et al. - Korea Research Reactor – 1 & 2
Decommissioning Project in Korea. In proceedings of WM´03 Conference,
Tuscon USA, 2003

[Philippen et al., 2017] P-W. Philippen et al. - Numeric Determination and


Validation of Neutron-Induced Radioactive Nuclide Inventories for
Decommissioning and Dismantling of Light Water Reactors. Nuclear
Technology, 201(1):66-79, 2018

[Pitkäoja, 2019] J. Pitkäoja, Long-term safety assessment for disposal of


VTT's decommissioning wastes in Loviisa LILW repository, MSc. thesis,
University of Jyväskylä, 2019

[Pöyry, 2014] Pöyry Finland Oy, VTT Technical Research Centre of Finland
- FiR 1 -tutkimusreaktorin käytöstäpoisto, ympäristövaikutusten
arviointiselostus, 16X156093. October 2014. (in Finnish)

[Rotem, 2019] Rotem Industries Ltd. - AMP-200 Area Monitoring Probe,


Technical spesifications. [Online] [access 2019, August 7];
https://www.rotem-radiation.co.il/product/amp-200-meter/

[Roth el al, 2000] E. Roth et al. - Irradiation of lithium aluminate and


tritium extraction. Journal of Nuclear Materials, 133-134:238-241, 1985

[Ruokola, 2016] E. Ruokola - FiR 1 Decommissioning plan: Clearance of


radioactive materials and structures. VTT Reseach Report VTT-R-03458-16,
2016

[Räty, 2017] A. Räty - Amounts of FiR1 decommissioning waste to be


cleared. VTT Research Report VTT-R-00042-17, 2017

[Räty, 2019] A. Räty - Activation calculations of decommissioning waste.


STUK neighboring area co-operation workshop on reactor graphite
characterization and 14C measurements, Espoo, Finland, 14.-15.5.2019
(confidential)

80
[Räty et al., 2018] A. Räty, A. Leskinen, M. Tanhua-Tyrkkö - FiR1
tutkimusreaktorin purkua valmisteleva karakterisointi ja näytteenotto. FiR1-
B27, Rev. 1.0, 2018 (in Finnish)

[Räty & Koivuranta, 2018] A. Räty, S. Koivuranta. BNCT-tilan Li2CO3-


suojamuovien tritium-pitoisuus. VTT Research Report VTT-R-06829-17,
2018 (in Finnish)

[Savolainen et al., 2013] S. Savolainen et al. - Boron neutron capture


therapy (BNCT) in Finland: Technological and physical prospects after 20
years of experiences. Physica Medica, 29(3):233–248, 2013

[Schlömer et al., 2016] L. Schlömer et al. - Activation calculation for the


dismantling and decommissioning of a light water reactor using MCNP with
ADVANTG and ORIGEN-S. In proceedings of 13th International Conference
on Radiation Shielding & 19th Topical Meeting of the Radiation Protection and
Shielding Division of the American Nuclear Society, Paris, France, 2016

[Shiraishi & Nagasaki, 1965] K Shiraishi, R. Nagasaki - Behaviour of


Tritium in Aluminium-Lithium Alloys. Nuclear Science and Technology,
2:499-505, 1965

[Skoog et al., 2007] D. Skoog, F. Holler, S. Crouch - Principles of


instrumental analysis (6th ed.). Canada: Thomson Brooks/Cole, 2007

[Stacey, 2007] W.M. Stacey - Nuclear Reactor Physics. WILEY-VCH


Verlacg GmbH & Co. KgaA, 2007

[STUK, 2013a] Säteilyturvakeskus STUK - Matala- ja keskiaktiivisten


ydinjätteiden käsittely ja ydinlaitoksen käytöstäpoisto. YVL-ohje D.4, ISBN
978-952-478-902-8, 2013 (in Finnish)

[STUK, 2013b] Finnish Radiation Safety Authority STUK - Exemption of


radiation use from safety licensing. Radiation safety guide ST 1.5. [Online]
[access 2019, May 5]; https://www.stuklex.fi/en/ohje/ST1-5

[Taylor & Holm, 2006] K.E Taylor, R.L. Holm - Multi-phased sampling
method to characterize a university TRIGA research reactor. In proceedings of
WM’06 Conference, Tuscon, Arizona, 2006

[TEM, 2017] Ministry of Economic Affairs and Employment of Finland -


VTT Technical Research Centre of Finland Ltd’s licence application for
decommissioning. [Online] [access 2019, March 11]; https://tem.fi/en/vtt-
technical-research-centre-of-finland-ltd-s-licence-application-for-
decommissioning

81
References

[Thermoscientific, 2019] Thermofished Scientific Ltd. - Product


description. [online] [access 2019 April 22]
https://www.thermofisher.com/order/catalog/product/IQLAAEGAAMFAB
WMAFC

[Tress & Leibundgut, 1999] Paul Scherrer Institut - Chemische Analysen,


Lauf Nr. 7001, 1999 (graphite elemental composition analysis by Gerhard
Tress and Fritz Leibundgut from PSI including attached AGOT graphite
technical specifications by National Carbon Company from years 1955, 1957
and 1958)

[Viitanen & Leppänen, 2014] T. Viitanen, J. Leppänen - Validating the


Serpent Model of FiR1 Triga Mk-II Reactor by Means of Reactor Dosimetry.
In proceedings of 15th International Symposium on Reactor Dosimetry, Aix en
Provence, France, May 2014

[Volmert et al., 2016] B. Volmert et al. - Validation of MCNP NPP


Activation Simulations for Decommissioning Studies by Analysis of NPP
Neutron Activation Foil Measurement Campaigns. In proceedings of ISRD 15
– International Symposium on Reactor Dosimetry, Aix-en-Provence, France,
2016

[Wallbrink et al., 2002] P. J. Wallbrink et al. - Radionuclide Measurement


Using HPGe Gamma Spectrometry. In: Zapata F. (eds) Handbook for the
Assessment of Soil Erosion and Sedimentation Using Environmental
Radionuclides. Springer, Dordrecht, Online ISBN 978-0-306-48054-6, 2002

[Wieselquist et al., 2017] W. Wieselquist et al. - Overview of Nuclear Data


Uncertainty in Scale and Application to Light Water Reactor Uncertainty
Analysis. United States Office of Nuclear Regulatory Research, NUREG/CR-
7249, 2017

[Woodcraft et al., 2003] A.L Woodcraft et al. - A replacement for AGOT


graphite? Physica B: Condensed Matter, 329–333:1662–1663, 2003

[Zagar el al., 2004] T. Zagar et al. - Long-lived activation products in


TRIGA Mark II research reactor concrete shield: calculation and experiment.
Journal of Nuclear Materials, 335:379–386, 2004

[Zhang et al., 2009] J. Zhang, M. Klasky, B.C. Letellier -The aluminium


chemistry and corrosion in alkaline solutions. Journal of Nuclear Materials,
384:175–189, 2009

82
View publication stats

83

You might also like