The cover depicts the chemical and physical complexity of the various species and

interfaces within a molten salt reactor. To advance new approaches to molten salt
technology development, it is necessary to understand and predict the chemical and
physical properties of molten salts under extreme environments; understand their ability to
coordinate fissile materials, fertile materials, and fission products; and understand their
interfacial reactions with the reactor materials. Modern x-ray and neutron scattering tools
and spectroscopy and electrochemical methods can be coupled with advanced
computational modeling tools using high performance computing to provide new insights
and predictive understanding of the structure, dynamics, and properties of molten salts over
a broad range of length and time scales needed for phenomenological understanding. The
actual image is a snapshot from an ab initio molecular dynamics simulation of graphene-
organic electrolyte interactions. Image courtesy of Bobby G. Sumpter of ORNL.
 Molten Salt Chemistry Workshop

Report for the US Department of Energy, Office

of Nuclear Energy Workshop

Molten Salt Chemistry Workshop

Technology and Applied R&D Needs for Molten Salt Chemistry
April 10–12, 2017
Oak Ridge National Laboratory
Co-chairs:
David F. Williams, Oak Ridge National Laboratory
Phillip F. Britt, Oak Ridge National Laboratory

Working Group Co-chairs

Working Group 1: Physical Chemistry and Salt Properties
Alexa Navrotsky, University of California–Davis
Mark Williamson, Argonne National Laboratory
Working Group 2: Analytical Chemistry
Sam Bryan, Pacific Northwest National Laboratory
Sheng Dai, Oak Ridge National Laboratory
Working Group 3: Molten Salt Fission Product Chemistry and Solid Salt Radiolysis
Bill DelCul, Oak Ridge National Laboratory
Tina Nenoff, Sandia National Laboratories
Working Group 4: Materials Compatibility
James Keiser, Oak Ridge National Laboratory
Preet Singh, Georgia Institute of Technology
Working Group 5: Computational Chemistry and Materials Science
Bobby Sumpter, Oak Ridge National Laboratory
Brian Wirth, University of Tennessee–Knoxville
Charles Henager, Pacific Northwest National Laboratory
Office of Nuclear Energy Leads
John Herczeg, Deputy Assistant Secretary for Nuclear Technology Research and Development,
DOE-NE
Stephen Kung, Program Manager, DOE-NE
ORNL Creative Services and ORNL Research Library
Deborah Counce, LeJean Hardin, Cindy Johnson, Kathy Jones, James Kidder

This publication can be seen here:

https://www.ornl.gov/content/molten-salt-chemistry-workshop

Technology and Applied R&D

 Molten Salt Chemistry Workshop

Table of Contents
Abbreviations, Acronyms, and Initialisms ............................................................................................... v
Executive Summary .................................................................................................................................. vii
1. Introduction ......................................................................................................................................... 1
1.1 Workshop Organization ................................................................................................................ 3
1.2 Future Research Directions ........................................................................................................... 3
References ............................................................................................................................................. 4
2. FRD 1: Understanding, Predicting, and Optimizing the Physical Properties of Molten
Salts ...................................................................................................................................................... 5
2.1 Background and Current Status..................................................................................................... 5
2.2 Technical Challenges and Research Directions ............................................................................ 7
2.2.1 Determining the Physical Properties of Molten Salts ....................................................... 7
2.2.2 Predicting the Chemical Speciation, Structure, and Dynamics of Molten Salts
Solutions .............................................................................................................................. 15
References ........................................................................................................................................... 19
3. FRD 2: Understanding the Structure, Dynamics, and Chemical Properties of Molten
Salts .................................................................................................................................................... 23
3.1 Background and Current Status................................................................................................... 23
3.2 Technical Challenges and Research Directions .......................................................................... 28
3.2.1 Determining Molecular Structure by X-ray and Neutron Scattering and
Spectroscopy ....................................................................................................................... 28
3.2.2 Correlating Electrochemistry and Spectroscopy for Online Monitoring and
Predictive Modeling ............................................................................................................ 30
References ........................................................................................................................................... 33
4. FRD 3: Understanding Fission and Activation Product Chemistry and Radiation
Chemistry .......................................................................................................................................... 37
4.1 Background and Current Status................................................................................................... 37
4.2 Technical Challenges and Research Directions .......................................................................... 38
4.2.1 Understanding the Physical, Chemical, and Radiochemical Properties of the
Fission and Activation Products in Molten Salts................................................................. 38
4.2.2 Correlating Fission and Activation Product Behavior with Surrogates .......................... 40
4.2.3 Understanding the Physical and Chemical Impact of Short-lived Isotopes .................... 42
References ........................................................................................................................................... 43
5. FRD 4: Understanding Materials Compatibility and Interfacial Phenomena ............................ 45
5.1 Background and Current Status................................................................................................... 45
5.2 Technical Challenges and Research Directions .......................................................................... 46
5.2.1 Advancing Spectroscopic and Scattering Investigation of MSR Interfaces.................... 46

Table of Contents iii

Molten Salt Chemistry Workshop

5.2.2 Understanding Materials Compatibility in the Molten Salts Reactor

Environment ........................................................................................................................ 47
5.2.3 Understanding Degradation Processes at the Material–Salt Interface ............................ 50
5.2.4 Understanding the Combined Effect of Chemistry and Radiation at the Interface ......... 51
5.2.5 Predicting Interfacial Interactions between Molten Salts and Structural Materials
and/or Gases ........................................................................................................................ 53
References ........................................................................................................................................... 56
6. FRD 5: Guiding Next-Generation Materials for Molten Salt Reactors ....................................... 59
6.1 Background and Current Status................................................................................................... 59
6.2 Technical Challenges and Research Directions .......................................................................... 59
6.2.1 Enabling Rapid Deployment of MSR Materials: Computational Challenges................. 59
6.2.2 Guiding Experimental Exploration of Advanced “Super” Materials .............................. 63
6.2.3 Advancing Code Qualification of New Materials ........................................................... 64
References ........................................................................................................................................... 66
7. FRD 6: Creating a Virtual Reactor Simulation ............................................................................. 67
7.1 Background and Current Status................................................................................................... 67
7.2 Technical Challenge and Research Directions ............................................................................ 67
7.2.1 Creating a Virtual Reactor Simulation ............................................................................ 67
References ........................................................................................................................................... 70
APPENDIX A: WORKSHOP RESOURCE DOCUMENT ................................................................ A-1
APPENDIX B: A DEVELOPER’S PERSPECTIVE ON THE NECESSITY OF SALT
PROPERTY RESEARCH ............................................................................................................. B-1
APPENDIX C: WORKSHOP AGENDA ............................................................................................. C-1
APPENDIX D: WORKSHOP PARTICIPANTS ................................................................................. D-1

iv Table of Contents
 Molten Salt Chemistry Workshop

Abbreviations, Acronyms, and Initialisms

AIMD ab initio molecular dynamics

ARE Aircraft Reactor Experiment

CALPHAD Computer Coupling of Phase Diagrams and Thermochemistry

CMD classical molecular dynamics

DFT density functional theory

DOE Department of Energy

DSC differential scanning calorimetry or differential scanning calorimeter

EMF electromotive force

EXAFS extended XAFS

7
FLiBe Li2BeF4

FRD Future Research Direction

GAIN Gateway for Acceleration of Innovation in Nuclear

GI-XAFS grazing incidence x-ray absorption fine structure spectroscopy

IR infrared

JICS Joint Institute for Computational Sciences

MAS magic-angle spinning

MD molecular dynamics

MOF metal-organic framework

MSBR Molten Salt Breeder Reactor

MSR molten salt reactor

MSRE Molten Salt Reactor Experiment

NE Office of Nuclear Energy

NMR nuclear magnetic resonance

ORNL Oak Ridge National Laboratory

PDF pair distribution function

Table of Contents v
Molten Salt Chemistry Workshop

QM/MD quantum mechanics molecular dynamics

R&D research and development

SEMO semi-empirical molecular orbital theory

TB-DFT tight binding DFT

TGA thermogravimetric analysis

UV-vis ultraviolet-visible

UV-vis-IR ultraviolet-visible-infrared

VR virtual reactor

VUQ validation and uncertainty quantification

WG working group

XAFS x-ray absorption fine structure spectroscopy

XAS x-ray absorption spectroscopy

XRD x-ray diffraction

vi Abbreviations, Acronyms, and Initialisms

 Molten Salt Chemistry Workshop

Executive Summary
At the dawn of this century, nuclear experts in the Generation IV International Forum recognized the unique
capabilities of molten salts in one of six next-generation reactor designs. Since that time, various national
research and development programs, most notably in China, have focused resources on developing molten
salt reactors (MSRs) of various designs. In the past decade in the United States, new designs and
applications have matured to the point that a number of companies have dedicated resources, enthusiasm,
and new approaches to MSR technology development. The US Department of Energy (DOE) and its
national laboratories have supported this evolution by working with multiple companies and universities;
this relationship matured in 2015 with the creation of the Gateway for Acceleration of Innovation in Nuclear
(GAIN) initiative working in concert with the US Industry MSR Technical Working Group.

In response to the sustained resurgence of the MSR concept, the DOE Office of Nuclear Energy
organized the Molten Salt Chemistry Workshop at Oak Ridge National Laboratory on April 10–12, 2017,
for the purpose of identifying innovative science-based, technology driven approaches to accelerate MSR
development and deployment. In the United States and across the world, the MSR concept is recognized
as one of the few game-changing technologies that can overcome existing barriers to the adoption of
commercial nuclear power. Not only is the generation of electricity within a once-through uranium fuel
cycle foreseen; in addition, proponents of MSR technology envision completely new opportunities to
rationalize fuel cycles by securely using a variety of fissile and fertile resources to supply high-
temperature, low-pressure industrial heat and to create a cost-competitive, safe, and more sustainable
commercial nuclear option.

This workshop was rooted in a multidisciplinary dialogue among a wide range of experts spanning
chemistry, materials science, nuclear science, and engineering from academia, national laboratories, the
US government, and commercial organizations. There was strong participation from the US Industry
Technical Working Group, as well as participation from small and large companies with specific interests
and expertise in molten salt technology. Since MSR technology has a chemical focus and chemical
challenges, the participants were focused toward the chemical sciences—but a very strong contingent of
materials, computational, and nuclear scientists/engineers also participated. These experts were organized
into five working groups: (1) physical chemistry and salt properties, (2) analytical chemistry, (3) molten
salt fission product chemistry and solid-state radiolysis, (4) materials compatibility, and (5) computational
science and material science. The Introduction section of this document and the appendices provide
details on how the workshop was conducted and who participated.

Fourteen R&D topics identified during the workshop were integrated into 6 Future Research Directions
(FRDs) after the workshop. The input for and definition of each FRD contained contributions from
multiple working groups and represent a collective effort from the workshop participants.

The 21st century MSR mission is distinctly different from the original thermal-spectrum molten salt
breeder reactor, and the next generation of MSR salts will be different. Thus a deeper understanding is
needed of molten salt physical and chemical properties to optimize the salt composition for new
applications. Choosing the right salt composition requires a knowledge of the salt’s capacity to dissolve
fissile, fertile, and particular fission products and of the physical properties of the salt. Phase diagrams
and thermophysical properties of molten salts are needed in the early stages of development to justify the
design and to define the safety boundaries. The first FRD (Understanding, Predicting, and Optimizing the
Physical Properties of Molten Salts) highlights the need to apply modern measurement techniques and
modeling and simulation tools to accelerate the design, discovery, and characterization of salts optimized
for various types of MSRs. Purification, characterization, and handling procedures for molten salts—as
well as analytical standards for pedigree molten salts—need to be accessible to the scientific community
to ensure the field can rapidly process and share reliable data.

Executive Summary vii

Molten Salt Chemistry Workshop

The structure and dynamics of molten salts play a key role in understanding phase transitions, viscosity,
thermal conductivity, volatility, solubility of fission products, and corrosion. Identifying and
understanding the structure, properties, and reactivity of these complex salt species is dependent upon
coordination chemistry, the redox state and Lewis-acidity of the system. Doing so requires a combination
of advanced measurement tools to identify the structures, and modeling capabilities to understand and
interpret the structure, dynamics, and properties. The second FRD (Understanding the Structure,
Dynamics, and Chemical Properties of Molten Salts) highlights how modern x-ray and neutron scattering
and spectroscopic tools (which were not available during the studies in the 1960s and 1970s) and
electrochemical methods can be coupled with advanced modeling capabilities to provide new insights into
the structure, dynamics, and properties of salt species on the length and time scales needed for
phenomenological understanding. This capability is needed for advancement in all other FRDs.

The third FRD (Understanding Fission and Activation Product Chemistry and Radiation Chemistry)
highlights the need to understand the rapid decay and chemical transmutation of fission and activation
products and the unique phenomena related to radiation-induced chemistry. It is essential to understand which
fission products are soluble, insoluble, partially soluble, and volatile, and how fission products change the
corrosivity of the melt. Highly specialized approaches are needed to address the fate of insoluble fission
products that evolve through complex decay chains, as well as specific radiation effects on chemical species.

The fourth FRD (Understanding Materials Compatibility and Interfacial Phenomena) addresses the need
to fundamentally understand material degradation (such as corrosion) and interfacial reactions (including
the combined chemical and radiation effect) to develop new, more stable materials for MSR. This
requires specialized experimental approaches (including a flow loop), instrumentation, and analysis to
study interfacial reactions—for both out-of-pile and in-pile measurements—to follow trace levels of
dissolved and dispersed species in solution and concentrations of surface species and surface morphology.
Computational modeling is needed to gain a molecular-level understanding of the chemical reactivity of
molten salts to provide insights into mitigating corrosion, fission product off-gassing, gas entrainment,
and materials precipitation.

The fifth FRD (Guiding Next-Generation Materials for Molten Salt Reactors) addresses the need to
develop the next generation of materials that will enable MSR developers to reach the reactor
performance targets. Structural materials proposed for MSR must endure extreme environments,
including high fuences of neutrons, high operating temperatures, and corrosive environments. New
approaches are needed to develop superior alloys and composites that are optimized for MSR which build
off detailed physical, chemical, and materials studies and combine modern experimental approaches—
including high throughput screening—with state-of-the art computational materials science approaches.
Advances in computational modeling can help guide materials development through an understanding of
chemomechanical degradation mechanisms, radiation damage, and thermomechanical degradation of
microstructure and property evolution. Timelines for materials development, testing, and qualification
must be accelerated to incorporate new materials into MSR designs through innovative approaches using
modern analytical and computational tools with new testing protocols.

The sixth FRD (Creating a Virtual Reactor Simulation) focuses on developing the modeling and simulation
tools necessary to understand the behavior of the reactor throughout its lifetime and provide dynamic reactor
and irradiation capsule emulation for chemical and isotopic source terms. To accurately predict MSR
conditions—including neutron transport, thermal hydraulics, isotopic transmutation, thermochemical
properties of the fuel salt, and corrosion effects—a virtual reactor simulator needs to be developed which
builds off the experimental and computational advances made in the previous FRD and incorporates a new
understanding of the physical, chemical, and radiochemical phenomena found in MSRs. This simulation
tool will not only allow understanding of the operational issues but also allow analysis of the fuel cycle,
prediction of waste forms and quantities, and assessment of new reactor concepts, salts, and chemistries.

viii Executive Summary

 Molten Salt Chemistry Workshop

1. Introduction
Since 1948, scientists have sought the creation of a new type of high-temperature fluid-fueled nuclear
reactor that is not constrained by the limitations of solid fuels and conventional coolants. The molten salt
reactor (MSR) emerged as the only viable concept for this class of reactor. In 1954, scientists at Oak
Ridge National Laboratory (ORNL) demonstrated the short-duration but very-high-temperature
capabilities of a molten-fluoride–fueled reactor in the Aircraft Reactor Experiment (ARE) [1]. The fuel
salt was a mixture of NaF-ZrF4-UF4, and the moderator was BeO. The reactor operated successfully for 9
days at an outlet temperature of 860 °C and power up to 2.5 MWt. At the end of the Aircraft Nuclear
Propulsion program, molten salt technology was repurposed with the objective of developing a thermal
spectrum breeder reactor based upon the thorium fuel cycle. From 1965 through 1969, ORNL operated
the 8 MW Molten Salt Reactor Experiment (MSRE), which demonstrated many of the key technologies
needed to justify the pursuit of a commercial MSR. The MSRE fuel salt consisted of 7LiF-BeF2-ZrF4-UF4
and it flowed through a graphite moderator at atmospheric pressure at approximately 650 °C. (7Li was
used to prevent the undesired production of tritium from 6Li by neutron-induced fission.) The operation
demonstrated the use of both 235U and 233U as the fissile driver in the fuel salt, as well as minor refueling
adjustments with 239Pu. The coolant salt in the secondary loop was 7Li2BeF4 (FLiBe). Experience with
these test reactors and the associated research and development (R&D) showed the MSR concept was
viable; but it also highlighted the challenges that remained in the domains of salt chemistry, fission
product chemistry, materials and corrosion chemistry, radiation chemistry, analytical chemistry, and fuel
processing. Figure 1 charts the evolution of MSR concepts.

(a) High-temperature
thermal propulsion short (b) Extended multi-functional test (c) Thermal breeder reactor design
duration engine (Aircraft reactor (Molten Salt Reactor (Molten Salt Breeder Reactor)
Reactor Experiment)
Experiment)

Figure 1. The 20th century evolution of the molten salt reactor concept. | Images courtesy of Oak Ridge
National Laboratory

With the cessation of the US commitment to thorium-breeder reactor designs in 1976, a different focus
eventually took hold for the MSR concept. It is now envisioned as a new way to rationalize the overall
fuel cycle without conventional aqueous reprocessing, with novel fueling and refueling schemes that do
not entail remote solid-fuel refabrication from recycled actinides; and as a potential pathway to safer,
more economical nuclear electricity generation and supply of high-temperature, low-pressure industrial
heat. This vision was recognized in the international commitment to the MSR concept within the
Generation Four International Forum. Strong efforts in MSR development are sustained in many
European and Asian countries (especially China) (see World Nuclear Association). Most of the interest in

Introduction 1
Molten Salt Chemistry Workshop

the US private sector revolves around the use of low-enrichment, high-actinide-content fuel salts, but a
wide variety of concepts are being considered using fluoride and chloride salts (see Appendix A). MSRs
under consideration today are either solid-fueled salt-cooled designs or liquid-salt–fueled designs, as
shown in Figure 2. Solid-fueled salt-cooled reactors contain solid ceramic fuel in prisms, plates, or
pebbles. Molten salt coolants flow over the solid fuel, and the heat is removed in the primary heat
exchanger by a secondary coolant; it transports the heat to a power generating system, such as a gas or
steam turbine. Molten salts have excellent heat transfer properties that rival those of water and permit
them to operate at much higher temperatures than light water reactors and at near-atmospheric pressures
in the primary loop.

Figure 2. In an MSR, the choice of salt determines the mission of the reactor. It imparts versatility for
multiple applications such as electricity production, generation of process heat, or burnup of actinide
waste (see Appendix A). | Image courtesy of Oak Ridge National Laboratory

Liquid-salt-fueled reactors—as demonstrated during the operation of the ARE and MSRE and envisioned
in the Molten Salt Breeder Reactor (MSBR) design—contain all fissile and fertile constituents (and many
fission products) dissolved within a homogeneous fluid fuel that circulates in the primary loop. The vast
majority of fission energy is deposited uniformly throughout the fluid fuel, which serves as both fuel
matrix and primary coolant. Heat is shuttled to the balance-of-plant by a nonradioactive low-pressure
secondary molten salt via two heat exchangers (Figure 2). Fluid-fueled-reactors have the significant virtue
that they do not require solid fuel fabrication/refabrication/qualification; but they impose the requirement
to control and manage all of the fission products that are released into the primary loop as either
dissolved, insoluble, or gaseous species.

2 Introduction
 Molten Salt Chemistry Workshop

As interest continues to grow in MSRs, new applications involve new materials and chemistry
requirements. Although some R&D requirements for salt chemistry and materials are the same as those
remaining at the end of the MSBR development program in 1976 (e.g., see Engel et al. 1979 [2]), other
requirements are driven by the use of new salts (e.g., chlorides), the realization of new conditions, and the
pursuit of new application domains. To advance the potential viability of MSRs, additional fundamental
and applied R&D is needed for salt chemistry, fission product chemistry, materials and corrosion
chemistry, radiation chemistry, analytical chemistry, and fuel processing.

1.1 Workshop Organization

In recognition of the renewed commercial interest in MSRs, the US Department of Energy (DOE) Office
of Nuclear Energy (NE), in coordination with MSR developers, plans to reestablish core competencies
and R&D capabilities in molten salt chemistry at DOE national laboratories and US universities. This
effort will provide a technical basis for supporting the goal of US industries to commercialize MSR
technologies and will support the MSR research community in training a broad base of next-generation
expertise.

To identify potential innovative science-based, technology-driven approaches to accelerate the

deployment of MSR technologies, and to engage the broader scientific community to advance the
knowledge and technology base of molten salt chemistry, DOE sponsored a 3 day workshop, Technology
and Applied R&D Needs for Molten Salt Chemistry. It was held at ORNL on April 10–12, 2017. In
addition to molten salt experts, the workshop included experts in traditional disciplines of chemistry
(physical, analytical, fission product, and materials chemistry) and nontraditional disciplines
(computational chemistry and materials science) from national laboratories, universities, and industry. A
total of 72 experts attended the workshop—15 from universities, 35 from national laboratories, 15 from
industry, and 7 from government agencies. (Appendix B is list of the workshop participants.)

In preparation for the meeting, a Resource Document (Appendix A) was drafted to orient and educate all
participants on the scientific and technology advances underpinning MSR technologies. Emphasis was
placed on identifying fundamental knowledge gaps. The agenda of the workshop is provided in
Appendix C. The workshop began with a half-day plenary session of invited experts from DOE, industry,
universities, and national laboratories to help define where the science and technology stands currently
and where it needs to go. Following the plenary session, the workshop participants separated into five
working groups: Physical Chemistry and Salt Properties, Analytical Chemistry, Molten Salt Fission
Product Chemistry and Solid Salt Radiolysis, Materials Compatibility, and Computational Chemistry and
Materials Science. The panels were charged with identifying (1) research topics that address both short-
term technology barriers and long-term (5–10 years) grand challenges that might enable revolutionary
changes in MSR technology and (2) a set of Future Research Directions (FRDs) that might accelerate
MSR technology development and deployment. At the midpoint of the workshop, each group reported to
all of the workshop participants on its progress toward identifying FRDs. The FRDs were discussed and
overlaps were identified. At the end of the workshop, each working group again reported its
recommendations for FRDs to all the workshop participants. Each working group prepared a report that
captured the research ideas and opportunities discussed in its sessions and the recommended FRDs. These
results were integrated and are summarized below and described in detail in the full report.

1.2 Future Research Directions

The working groups initially identified 14 FRDs; however, these contained significant overlap and thus
were combined to formulate 6 FRDs. The following FRDs are described in detail in the chapters that
follow.

Introduction 3
Molten Salt Chemistry Workshop

• Understanding, Predicting, and Optimizing the Physical Properties of Molten Salts

• Understanding the Structure, Dynamics, and Chemical Properties of Molten Salts
• Understanding Fission and Activation Product Chemistry and Radiation Chemistry
• Understanding Materials Compatibility and Interfacial Phenomena
• Guiding Next-Generation Materials for Molten Salt Reactors
• Creating a Virtual Reactor Simulation

References

1. E. S. Bettis, W. B. Cottrell, E. R. Mann, J. L. Meen, and G. D. Whitman. “The Aircraft Reactor

Experiment—Operation.” Nucl. Sci. Eng. 2, 841–853, 1957.
2. J. R. Engel, H. F. Bauman, J. F. Dearing, W. R. Grimes, and H. E. McCoy, Jr. Development Status
and Potential Program for Development of Proliferation-Resistant Molten-Salt Reactors. ORNL/TM-
6415, Oak Ridge National Laboratory, 1979

4 Introduction
 Molten Salt Chemistry Workshop

2. FRD 1: Understanding, Predicting, and Optimizing the Physical

Properties of Molten Salts
2.1 Background and Current Status

The concept of a homogeneous, molten salt–cooled reactor was first envisioned in 1948 and its
development is documented in many reviews [1]. The MSRE, the most significant MSR demonstration
activity, operated at ORNL from 1965 until 1969. The results from that experiment and subsequent
research have provided a wealth of material properties data and performance information that are
thoroughly documented in the open literature. The MSRE was originally fueled with 235U (and later 233U)
in a molten salt solution composed of 7LiF-BeF2-ZrF4-UF4 of the approximate composition 65-29-5-1
mol%. To appreciate the derivation of this fuel composition, one must recognize the numerous criteria
and constraints behind it, which were described in detail by Grimes [2]. At the top level, neutronics
(parasitic neutron absorption and moderating properties) limit the choice of nuclides that can be
considered; those limits combine with the limits associated with the mission of the reactor (breeder,
burner, or converter) and the neutron energies (fast or slow) in the system.

Figure 3 highlights the neutronic constraints that dictate the choice of elements for thermal and epithermal
reactors (i.e., low neutron capture cross sections) and the chemical constraints that exert a controlling
influence on all reactor applications. [3] The key criterion of the molten salt in a fluid-fueled reactor is
that it must dissolve more than the critical concentration of fissile material (235U, 233U, or 239Pu), dissolve
a high concentration of fertile material (232Th or 238U), and dissolve all fission products without loss of
useful properties or subsequent chemical reaction. It also must be stable toward reactor radiation and high
operating temperature (500–750 °C), survive fission of uranium (or other fissionable materials), and
possess a low vapor pressure over the entire performance envelope. The fluid must also possess heat
transfer and hydrodynamic properties allowing it to serve as a heat-exchange fluid without excessive
power requirements for circulation, and it must be relatively nonreactive with the metal reactor, heat
exchangers, and moderator material (typically graphite). Finally, to achieve a low fuel cycle cost, the fuel
should be inexpensive and economical to purify via efficient schemes designed to recover unburned
fission and bred fissile materials and to remove fission products and neutron poisons.

The neutronic requirements for coolant salts are distinctly different from those for fuel salts (see
Figure 2). For a primary coolant salt, low-neutron-absorbing-salts are necessary to obtain negative void
and temperature coefficients [4]; whereas secondary coolant salts have no formal neutronic requirements,
as the radiation intensities are significantly less, and the consequences of uranium fission and fission
products are absent. In both cases, the coolant salt must be stable and compatible with the metal
construction of the heat exchangers, have good heat transfer and hydrodynamic properties, have a low
vapor pressure over a wide operating range (400 to over 900 °C), and not undergo a violent reaction with
the fuel or steam in the event of a leak.

The requirement that fuel salts be compatible with the container material places a significant constraint
upon the choice of salt constituents. It means that the cationic constituents of the salt must consist of
relatively active, electropositive elements (e.g., fluoride compound free energies of formation
>95 kcal/mol⋅F), whereas the container material must consist of elements that are relatively noble in
comparison (e.g., fluoride compound free energies of formation <75 kcal/mol⋅F). [5] This requirement
eliminates the middle of the periodic chart for consideration as salt constituents and the periphery of the
periodic chart for consideration as alloy constituents (only a small region in the center is suitable for
metallic alloys; see Figure 3). Nonmetallic materials (e.g., ceramics and composite materials) can also be
compatible with a high-temperature salt—for example, graphite is known to be fully compatible with
molten halides to very high temperatures. Although nonmetallic materials have been used and may find

FRD 1 5
Molten Salt Chemistry Workshop

service as core internals, there is currently no joining and fabrication technology for these materials to
support their deployment as the container material for many of the complex components that will be in
contact with the salt.

Figure 3. Early MSR program approach to screening and justifying candidate MSR salts (where σt(abs) is
total neutron absorption cross section). | Image courtesy of Oak Ridge National Laboratory

From a neutronics point of view, non-halide fluid-fueled systems (containing H, B, C, N, O, and Bi), such
as oxides, carbonates, and nitrates, seem like reasonable choices. But they are either not chemically
compatible with potential container materials or cannot dissolve enough fissile and fertile material at
practical temperatures to serve as fluid fuels. Oxides typically have high melting points and are corrosive;
carbonates also have a fairly high melting point and require a CO2 over-pressure to stabilize them; and
nitrates are problematic because of thermal decomposition and their stability in a high radiation field.
Thus, combinations of 7Li+, Be2+, Al3+, Na+, Mg2+, and Zr4+ cations with F- and 37Cl- stand out for use in
fuel salts and primary coolants. (Note: 37Cl- is preferred because the major chlorine isotope [35Cl]
produces 36Cl as an activation product, which is an unwanted long-lived energetic beta emitter.) For
secondary coolant salts, there are no formal neutronic requirements; therefore, isotopic enrichment is not
needed and so combinations of Li+, Na+, K+, Rb+, Mg2+, Be2+ and Zr4+ cations with F-, Cl- or BF4- are
viable [6]. There are limited thermodynamic data for candidate molten fluoride salt systems and even
fewer data for candidate chloride salts systems (as shown in Figure 4).

Reactor salt options need to be optimized in the context of applications that address current needs, new
capabilities, and material developments that have emerged since the MSRE. Over the past half-century,
experimental, computational, and analytical methodologies have shown enormous advances and can be
used to explore structure, properties, and reactivities of molten salts in considerably greater detail. In this
light, this section focuses on the need for state-of-the-art investigations to understand, predict, and
optimize the physical properties of molten salt systems that could be applied to MSRs and to obtain the
data necessary to support reactor design/development.

6 FRD 1
 Molten Salt Chemistry Workshop

LiF NaF KF RbF CsF BeF2 CaF2 LaF3 ZrF4 ThF4 UF4 PuF3 UF3 LiF NaF KF RbF CsF BeF2 CaF2 LaF3 ThF4 UF4 PuF3
x x x x x x x x x x LiF x x x x x x x LiF-NaF
x x x x x x x x x NaF x x x LiF-KF
x x x x x KF
LiF-RbF
x x x RbF
x x LiF-CsF
x x CsF
x x x LiF-BeF2
x x x x BeF2
x LiF-CaF2
x x CaF2
LiF-LaF3
x LaF3
x x LiF-ThF4
ZrF4
x LiF-UF4
x x ThF4
Binary
x UF4
x x x NaF-BeF2
Ternary
PuF3 x NaF-ThF4

UF3 x BeF2-ThF4

Figure 4. Available thermodynamic data (marked by an x) for binary and ternary fluoride salt mixtures for
MSRs developed at the Institute for Transuranium Elements of the Joint Research Centre of the European
Commission, Karlsruhe, Germany. | Capelli, E. Thermodynamic Characterization of Salt Components for Molten Salt
Reactor Fuel, Fig. 1.2, PhD Thesis, University of Delft, 2016, http://samofar.eu/wp-content/uploads/2016/11/2016_
Capelli_Elisa_PhD-thesis.pdf

2.2 Technical Challenges and Research Directions

2.2.1 Determining the Physical Properties of Molten Salts

Preparation and Handling of High-Purity Salts
Although ORNL established processes and practices for purifying and handling fluoride salts for R&D
and the MSRE, [7] there is no national standard for testing and reporting the purity of the salts. Different
purification processes and procedures are needed for other salts (e.g., carbochlorination for chloride salts).
[8] Purification methods are needed not only to eliminate cationic metallic impurities and corrosive
moisture, oxides, and hydroxides in the hydroscopic halide raw materials, but also to eliminate corrosive
anionic contaminants such as sulfates, sulfides, and carbonates. It is necessary to positively identify salt
impurities and quantify the degree of purity achieved. In some cases, this may require going beyond
previous practices, especially for new salt systems. Such purification protocols and tabulated data sets are
readily available in the literature for organic reagents and solvents, but complementary data and
standardized and validated methods do not exist for halide salts intended for nuclear service. The
development of protocols amenable to the production of high-purity salt will also require scaling to
production levels and transferring proven methods and defined purity requirements to commercial
vendors to facilitate more widespread availability of pure, quality-assured salts to accelerate research.

There are many challenges to the accurate measurement of the high-temperature thermophysical
properties of MSR salts. First, a wide range of salt and fuel compositions must be considered, including
fission products and transuranic elements produced during power production. Each component requires
extensive purification to remove impurities—such as oxides, sulfides, metals and water—which demands
the use and engineering of specialized containers that not only are compatible with the salt but also will
rigorously maintain salt purity. For highly accurate determination of salt physical properties,
contaminants (e.g., atmospheric moisture) must be rigorously excluded following purification and during
property measurement. To these ends, purification procedures need to be developed, externally validated,
and published in the open literature; and a single source of pedigreed salt should be established for the
production of analytical standards for the research community. Meeting those needs would advance our
knowledge and confidence in the physical properties of the molten salts, since the community would be
working with materials of similar quality. A primary contribution to the overall scientific and technical
program would be a supply of pedigreed salt from which institutions could obtain small quantities to
serve as standards to validate the purity of larger salt quantities.

FRD 1 7
Molten Salt Chemistry Workshop

For large-scale experiments, analytical-grade salts available from commercial suppliers are of insufficient
purity to accommodate meaningful investigations of molten salt behavior and the reliable characterization of
physical properties essential for predictive modeling efforts and fuel qualification. Both phase behavior and
corrosion responses are especially sensitive to small amounts of impurities. Previous work on the MSRE at
ORNL revealed that minimal salt loop corrosion was achievable when care was taken to afford a high-purity
salt that was preserved under air-free and moisture-free environments. Modest corrosion rates (0.013–0.044
mm/year) have been obtained for purified salts in a closed loop, [9] whereas salts that were used as received
and remained open to atmosphere corroded containments more than four orders of magnitude more rapidly
(up to 167 mm/year). [10] This vast difference in corrosion response highlights the need to develop a series of
best practices in the production, purification, characterization, and handling of molten salts, as well as a
means by which the pedigree and reliability of the reported data can be assessed by the scientific community.

The current standard for quality assurance of nuclear fuels and material is Nuclear Quality Assurance-1,
established by the American Society of Mechanical Engineers. There is a need to establish a quality
assurance hierarchy for molten salt preparation and characterization with commensurate measurement,
documentation, and reporting criteria explicitly documented. Such an approach will improve confidence
in reported results and allow the research community to meaningfully contribute to the collection of high-
quality data required for fuel validation or development of predictive models. For the purposes of
ensuring instrument calibration and good quality control practices, a single source of salt should be
established for the production of analytical standards for purified, pedigreed salt. Additionally,
experimental documentation and representative metrics of data quality need to be accessible for external
reproducibility, validation, and verification. An example of such a system is the Cambridge
Crystallographic Database, which serves as a free repository for single crystal data sets. Crystal data can
be independently analyzed before they are uploaded, enabling the creation of a “checkcif” file that
articulates any suspicious metrics in the proposed structure. These documents are peer-reviewed before
publication of the crystal structure and preserved for subsequent inspection as needed. A similar review,
documentation, and archival infrastructure is necessary to enable the reliable determination of salt
properties and to facilitate future molten salt characterization and MSR development.

Defining Phase Diagrams of Molten Salts

The primary tool for understanding fuel and coolant salts is the phase diagram—as it defines the
temperature-composition space for single-phase liquids; the order and identity of equilibrium phases that
crystallize as the melt freezes; and the solubility of constituents that can grow, such as transuranics,
fission products, corrosion and activation products, and impurities. A variety of phase transitions occur in
molten salts: in a few instances, the melt and frozen solid have an invariant composition throughout the
freezing process (i.e., the line-compound composition); in most instances, the transition is more complex
and the frozen solid is different from the melt composition in a variety of ways; and in rare instances, two
immiscible liquids can form or a melt may cool to a metastable phase that remains liquid below the
indicated melting temperature of the mixture. [11]

A large number of binary and ternary phase diagrams have been measured for chloride and fluoride salt
systems, [12] and a small number of higher-order systems have been characterized over the years (see
Figure 5). The methods used to determine the phase diagrams include differential scanning calorimetry,
electrochemical methods, and solidification along a temperature gradient [13–16]. Since a number of the
species are volatile, their vapor-liquid and vapor-solid equilibria need to be known, along with additional
thermodynamic data, to provide a complete understanding of the phase behavior.

8 FRD 1
FRD 1

Phase Fields
A. ThF4
B. LiTh4F17
C. LiTh2F9
D. LiThF5
E. Li3ThF7
F. LiF
G. Li2BeF4
H. β-BeF2
I. Liquid
(miscibility gap)

(a) Calculated LiF-BeF2 pseudo-binary phase diagram with (b) The liquidus projection of the LiF-BeF2-ThF4 system.
fixed concentration of UF4 (2.55 mol%) and ThF4 (19.95 mol%).

ideal

Molten Salt Chemistry Workshop

BeF2-UF4, 1000K, 1400K

(c) Calculated entropy of mixing: BeF2-AnF4 (An=Th, U)

1000 K and 1400 K. (d) Optimized BeF2-UF4 phase diagram.
Figure 5. Representations for integrating phase diagram understanding. |Reprinted from J. Nucl. Mater. 449, E. Capelli, O. Beneš, and R.J.M. Koning.
“Thermodynamic assessment of the LiF–NaF–BeF2–ThF4–UF4 system.” 111–121, 2014, with permission from Elsevier.
9
Molten Salt Chemistry Workshop

Previous methods of determining phase diagrams of molten salt mixtures used a relatively large number
of samples and were highly labor intensive. The rigorous control and measurement of impurities—
including dissolved metals, oxygen-containing impurities (e.g., oxides, hydroxides, sulfates, carbonates,
oxyhalides), water, and other types of impurities—were not always reported. Therefore, it is difficult to
ascertain the accuracy of some of the existing phase diagrams in the literature. For some systems, there
are discrepancies among various studies that have yet to be reconciled. These flaws reinforce the need for
a source of high-purity salts and for the publication of complete analytical approaches and experimental
practices used to obtain data sets.

The phase diagram is also a manifestation of the thermodynamics of the system, as equilibrium is
determined by the Gibb’s free energy of the system. Thermodynamic properties of each of the
components (including their heats and entropies of fusion) and mixing properties within phases of varying
compositions define the Gibb’s free energy. If the thermodynamic properties are known, the phase
diagram can be calculated, using various established formalisms [17]. Conversely, thermodynamic
properties can be extracted from the phase diagram.

The thermodynamic properties of a system are determined by the atomic-level structure and inter-atomic
interactions. Modern spectroscopic techniques (as described in Section 3) can provide insights into the
structure and dynamics of molten salts and can be used to validate computational models. Ab-initio
calculations can be used to guide the prediction of thermodynamic properties and identify the appropriate
forms of equations for calculating phase equilibria. Comprehensive databases, which are just beginning to
be developed from data produced decades ago, can be used for data mining and for compiling and
reconciling data to gain insights into thermodynamics properties of molten salts and develop more
complex thermodynamic models. An improved way to query and visualize phase diagram and
thermodynamic databases is recommended, which would support more strategic thinking about the
development of new chemical systems and help in developing more comprehensive thermodynamic
models.

Computational methods and databases can accelerate the collection and analysis of thermodynamic data
and phase diagrams. As a starting point, we need to assess prior studies and identify missing data,
compositions, and thermodynamic accuracy gaps in the experimental data for binary, ternary, and higher
order multi-component systems. Recommended research directions include the following:

1. Use computational studies to define where additional experimental data are most needed.
Experimental data should be taken in regions where the activity coefficients vary sharply, where there
is evidence for short-order and complex-ion formation, from computational and experimental studies
of salt structure and speciation, and for new systems under consideration.

2. Develop advanced methods of high-throughput data collection, with well-quantified uncertainties,

that are applicable to air- and water-sensitive high-temperature salts containing radionuclides and
other toxic elements. For example, high-throughput phase diagram determination might entail
spectroscopic interrogation of an array of simultaneously heated samples, or of a sample of given
composition exposed to a temperature gradient. Approaches for obtaining meaningful data for salts
containing volatile fission and activation products are also needed; experiments performed under
modest pressures may overcome some of these challenges, but such an approach also requires
validation against known phase diagrams.

3. Develop high-purity calibration standards of known pedigree for use in analytical studies using
multiple experimental methods for thermodynamic data collection. It is critical to validate newly
developed methods on well-characterized simple systems, as well as to include quality control
samples during data collection in compliance with good analytical practices.

10 FRD 1
 Molten Salt Chemistry Workshop

4. Identify and investigate potential metastable phases induced by ionizing radiation and compositional
changes due to fission, activation, and decay.

5. Validate density functional theory (DFT) and ab initio molecular dynamics (AIMD) calculations
performed on known salts with experimental data and spectroscopic characterization, followed by
modeling of thermodynamic properties of systems that are new or difficult to measure
experimentally, (e.g., containing higher actinide activation products).

6. Develop appropriate thermodynamic models for use as a predictive tool, enabling investigation
beyond conditions that can be experimentally measured (e.g., short-lived fission products or highly
activated sample). Database development requires selecting appropriate theoretical models for each
specific system, then computationally optimizing the input thermodynamic parameters to match
available experimental data points. One of the challenges is development and application of the
appropriate thermodynamic model for each specific system. The question of appropriate entropies of
mixing is a particular challenge because of the complex speciation, and the possibility of short- to
mid-range molecular order in the salt.

7. Develop tools to query salt property databases and visualize information in a way that is accessible to
a variety of users and most useful in design and safety analysis. Recent developments in modern
computing, data visualization, and user-interface tools need to be translated to those ends. The ability
to trace the results of a query to the initial data sources, accompanied by experimental details, can
take advantage of modern database and content management tools. Uncertainty quantification, error
propagation, and sensitivity analysis should be integral to these models. Such analysis will also
indicate where more experimental data are needed. The user interface will expand the community,
which will also help with prioritization of data needs.

Thermodynamic and Thermophysical Property Determination

In addition to phase behavior, other important physical properties need to be measured for the individual
salts and mixtures of salts, including melting point, density, viscosity, heat capacity, thermal conductivity,
vapor pressure, heat of fusion, expansivity, compressibility, surface tension, and gas and fission product
solubilities. [18] Some physical properties (such as solubility, vapor pressure, and viscosity in the
neighborhood of glass-forming domains) are sensitive to slight compositional changes; and the properties
change rapidly, further underscoring the need for established purification protocols, high-purity salts, and
rigorous analytical practices (Section 2.2.1). Given the number of measurements that need to be made,
high-throughput experimental techniques that can be miniaturized and are amenable to operation in a
glovebox environment could greatly accelerate property measurements and technology development.
However, the need for high-throughput measurements must be balanced against and coordinated with the
need for high-precision reference measurements. This subsection provides examples of some state-of-the-
art reference measurement techniques that may be miniaturized for high-throughput measurements. This
discussion is not comprehensive but provides some examples of measurement techniques to inspire the
reader. It must be stressed that most of these property measurement capabilities for extreme temperatures
and controlled atmosphere are not off-the-shelf capabilities. They are typically customized systems with
specialized procedures that are not standardized.

Calorimetry
Calorimetric methodology and sensitivity have been improved over the past three decades. A calorimeter
suitable for high-temperature solution calorimetry is now commercially available and is being used in
several laboratories at both universities and national labs. Heats of formation and transformation can
typically be measured with an accuracy of ±1–3 kJ/mol for sample sizes of 10–100 mg. Heats of mixing,
heats of fusion of binary and ternary fluorides and chlorides, heats of formation of crystalline ternary

FRD 1 11
Molten Salt Chemistry Workshop

compounds, and heats of dissolution of fission products and other impurities in molten salts as a function
of both molten salt composition and solute concentration can all be measured by calorimetric methods.
However, the precision and accuracy of such measurements will depend critically on the purity of the
starting materials.

Differential scanning calorimeters (DSCs) can provide heat capacities, melting points and heats of
fusion/crystallization. Their accuracy in heat capacity measurements is of the order of 1–2% at room
temperature to 500 °C and 2–5% at 500–1200 °C. Heats of phase transition and fusion can be determined
at an accuracy of 1–5%, depending critically on appropriate calibration standards. Temperature accuracy
below 1000 °C is generally on the order of ±1 °C. There are opportunities to adapt these methods to
radionuclides and air- and water-free environments. Sample containment—especially because of sample
creep along container walls—and materials compatibility pose challenges. New high-throughput methods
on multiple samples, perhaps combining calorimetric and spectroscopic techniques, should be developed.

Viscosity measurements
A number of different methods (e.g., rotational, capillary) have been used for molten salt viscosity
measurement. [19] The miniature rolling-ball viscometer (Figure 6) uses an induction coil for rapid heat-
up and measurement of the rolling ball velocity (μ) as a function of inclination angle (θ). With this
method, it is possible to determine the viscosity of molten salts in a range of 1–5 mPa⋅s at temperatures of
up to 1400 °C. However, a set-up needs to be designed to use less than 10 mL of flux and use
interchangeable tubes with rolling balls.

Figure 6. Miniature rolling-ball viscometer description. | Image courtesy of Prof. Uday Pal, Boston University]

Thermal diffusivity and thermal conductivity

Optical techniques based on both emission spectroscopy and modulated heating have the potential to
measure radiative and thermophysical properties that govern temperature in a molten salt. To
determine thermophysical properties, such as thermal diffusivity or thermal conductivity, thermal lens
spectroscopy can be implemented. [20] This technique uses an acousto-optical modulator to provide
modulated laser heating of the liquid at a small focal point. A probe laser beam is passed through the focal
region and is affected (broadened or deflected) by the change in refractive index caused by the heating
beam. This change produces in turn a change in the detected laser intensity, and the resulting signal can
be fit to heat transfer models to determine the thermal conductivity and thermal diffusivity of the fluid.

12 FRD 1
 Molten Salt Chemistry Workshop

The thermal lens technique has the added benefit of being able to provide an independent measurement of
the absorption coefficient of a fluid by a rearrangement of the lasers and detectors. [21]

Molar volume and electrical conductivity measurement

Combined density and electrical conductivity (ionic and electronic) can be measured using coaxial
electrodes that measure AC impedance spectra and DC currents as a function of electrode depth in the salt
(Figure 7). Using the cell geometry and the measurement data, salt height at temperature, molar volume,
and ionic and electronic conductivities can be determined. Furthermore, the partial ionic conductivities
and the chemical diffusivities of species can be extracted from separate transference number
measurements. [22]

Figure 7. Electrical method for measuring molar volume, ionic and electrical conductivity. | Reprinted from
S. L. Schiefelbein, N. A. Fried, K. G. Rhoades, and D. R. Sadoway. A high-accuracy, calibration-free technique for
measuring the electrical conductivity of liquids. Review of Scientific Instruments, 69(9), 3308–3313, 1998,
(http://dx.doi.org/10.1063/1.1149095), with the permission of AIP Publishing.

Surface tension and contact angle

A precise bubble pressure technique can be designed to measure surface tension by immersing an
appropriate metal tube in the selected molten flux, blowing gas through the tube to acquire a surface
tension value between the gas and the liquid salt, and examining the peak pressure exerted on the gas by
the liquid (Figure 8). In many cases, contact angle is the key piece of information for understanding
interfacial behavior (foaming, intrusion into pores, and wetting surfaces). Additional techniques that can
yield both contact angle and surface tension are the sessile drop technique and the Langmuir-Wilhelmy
plate method. The sessile drop method uses a very high–shutter speed camera and an optical setup to take
an image of a molten flux piece resting on an appropriate substrate. This allows measurement of a contact
angle at a three-phase interface (Figure 8). Using this contact angle, the solid-liquid surface tension can be

FRD 1 13
Molten Salt Chemistry Workshop

estimated. Each of these methods has advantages and disadvantages that will need to be considered based
upon the system being investigated, the type of information desired, and the required accuracy. [23]

Figure 8. Interfacial measurement methods. (a) Bubble-pressure method for measuring surface tension. (b)
Sessile-drop method for measuring contact angle and surface tension. | Image courtesy of Prof. Uday Pal, Boston
University

Volatility and melting point

The melting point and volatilization rate of the salt as a function of temperature can be measured using a
simultaneous DSC and thermogravimetric analysis (DSC/TGA) balance. To ensure salt purity is
preserved, the instrument needs to be designed for operation in an inert-atmosphere glovebox. Although
the recorded weight change of the TGA can be attributed to salt volatilization, the weight change is
dependent upon salt purity (Section 2.2.1). Volatile species can be determined, both for understanding salt
properties and for validating salt purity, by quadrupole mass spectrometer analysis or by condensing
volatilized species on a water-cooled substrate with subsequent analysis by x-ray fluorescence, x-ray
diffraction (XRD), or other techniques. [24] For measuring low vapor pressures, the traditional
transpiration method is reliable and accurate and can be used as a reference method in this domain. It does
not require significant deployment of sophisticated instruments within an inert envelope. [25]

Solubility
Small, thin capillary tubes containing the salt can be rapidly heated and quenched. The tubes with the salt
can be placed in contact with the compound for which the solubility needs to be measured. Both the
solubility and the diffusivity of the dissolving compound can be estimated by measuring the diffusion
profile of the quenched sample (Figure 9).

EMF technique for chemical activity and phase diagrams

The electromotive force (EMF) generated by electrochemical cells can be used to measure the partial
molar Gibbs energies, activity, and activity coefficients, and to determine the phase diagram. For
example, the phase diagram of Ba-Bi system was determined by the EMF technique using a CaF2-BaF2
binary electrolyte at 450–800 °C. [26] This technique is potentially powerful, as it is possible to define
both the thermodynamic properties of the system and the phase diagram that will delineate liquid-state
solubility, solidus temperature, and phase fraction/composition.

14 FRD 1
 Molten Salt Chemistry Workshop

Figure 9. Rapid-quenched capillary setup for specie solubility and diffusivity. | S. Su, T. Villalon Jr., U. Pal, and
A. Powell. “Techniques for measuring solubility and electrical conductivity,” pp. 465–475 in Advances in Molten Slags,
Fluxes, and Salts: Proceedings of The 10th International Conference on Molten Slags, Fluxes and Salts (MOLTEN16), Eds.
R. G. Reddy, P. Chaubal, P. C. Pistorius, and U. Pal. TMS (The Minerals, Metals & Materials Society), 2016. Used with
permission of Springer.

2.2.2 Predicting the Chemical Speciation, Structure, and Dynamics of Molten Salts
Solutions
Computational salt chemistry is a major scientific challenge because of the complex nature of the salts.
Although there has been prior work showing the value of molecular modeling with interatomic potentials
to predict molten salt properties, [27] only a few studies of molten salts have used a fully first principles
approach capable of explicitly treating polarization effects and chemical reactivity. [28] Looking forward,
optimized approaches are needed which employ first-principles MD simulations of a molten salt system
that are validated against existing experimental data. To develop those approaches, the basic
thermodynamic and kinetic properties of molar volume, thermal expansion, bulk modulus, and
diffusivity, for a molten salt at multiple temperatures, need to be computed. Simulations for pure molten
chloride/fluoride salts at multiple temperatures are needed to obtain temperature-dependent
thermodynamic, kinetic, and transport properties; and speciation information is clearly needed for the
prediction of salt phase diagrams and other physical properties. In addition, knowledge of the constituents
is needed to understand interfacial behavior leading to MSR material degradation, such as corrosion.

Another complexity in computational salt chemistry is that the atomic composition of the salt changes as
a function of time as a result of fission product formation and the emission of α, β, and γ-rays, which can
lead to radiolysis in solid salts. In the molten salt, our understanding of the transient chemistry associated
with fission and decay is limited, but what has been observed over longer time periods is that much of the
energy that is deposited in the salt works to restore the salt to an equilibrium composition; and few
structural changes are observed in the salt. Compositional and redox changes arise as results of
transmutation and fission (loss of fissile species, grow-in of soluble fission products). Chemical evolution
of the insoluble fission products, at all time scales, remains an open question Computational electronic
structure and MD methods can be used to extend the limited experimental data sets to cover a broad range
of parameter space. However, these predictions must have the appropriate accuracy required for upscaling

FRD 1 15
Molten Salt Chemistry Workshop

across multiple time and length scales. For example, nucleation is very sensitive to the rate constants
used. An error of 0.3 kcal/mol in a reaction step at the MSR operating temperature of ~ 600 °C can lead to
an overall error in the equilibrium prediction by a factor of about 850 over 40 steps. [29]

Furthermore, the importance of experimental validation of computationally-predicted results cannot be

overemphasized. In previous research studies, computational methods were applied that incorrectly
predicted molten salt structure and were unable to predict the intermediate structure that was later
demonstrated experimentally. [30] The power of predictive approaches is that they provide the ability to
investigate systems under time scales and conditions that are not experimentally accessible; therefore
demonstration of performance accuracy under experimentally testable conditions is an important
prerequisite for computational development. Techniques suitable for physical characterization are
discussed in detail earlier, and spectroscopic approaches capable of providing structural details at the
atomistic and intermediate length scales are discussed in Section 3.

Successful development and validation of this computational research approach will enable a new
capability that goes beyond the traditional approach to optimizing molten salt composition with respect to
fuel and fission product solubility, thermophysical properties, and in-reactor salt degradation. It will also
minimize corrosion/degradation of materials of construction for the reactor, pump, and heat exchanger
components.

Building Accurate Multi-component Models for Prediction of Phase Diagrams

The phase equilibria and thermochemistry of fluoride and chloride molten salts can be well represented
by existing modeling frameworks (Figure 10). While free energies and solution parameters for a number
of the proposed components are available, numerous critical elements are missing in the datasets. High-
throughput first-principles techniques and experimental methods need to be developed to efficiently
provide phase equilibria and thermochemistry information. The models will then need to be expanded to
include any additional components, and validated experimentally. Solution models for viscosity will also
need to be developed, possibly based on the demonstrated successful correlations in oxide liquids.
Importantly, parameters such as constituent solubility, oxidation state, and coordination environment need
to be experimentally determined to allow theoretical models to accurately reproduce and predict the
physicochemical behavior of the melts.

Applying Molecular Dynamics Simulations to Predict Thermophysical and Transport

Properties
It is important to be able to accurately predict and understand liquid structure and dynamics both in the
bulk and at salt/material or salt/vapor interfaces. MD simulation can help explain transport properties,
including viscosity and the origin of different length-scale features in the x-ray structure function, S(q), as
well as real-space radial distribution functions. It can also be particularly useful in understanding the
effects of structural voids and defects formed under radiolytic conditions, as well as determining their
effects on material transport and thermal conductivity. This is particularly important for understanding
possible structural heterogeneities in the molten salt system. MD simulation can be used to understand
wetting and permeation at degraded interfaces where surface energies are important, and to describe salts
under confinement within graphitic or other structural materials when thermodynamic models are
insufficient. AIMD will be needed for processes in the condensed phase in which reactivity may be
important.

16 FRD 1
 Molten Salt Chemistry Workshop

Figure 10. Illustrative phase diagram of invariant points and liquidus projections for a fluoride fuel salt with a
minor amount of Pu and low concentrations of Cr, Ni, Mo. (Diagrams and data such as this will be needed
for MSR complex salt chemistries.) | Image computed from existing molten salt databases; courtesy of Prof. Ted
Besmann, University of South Carolina

A simplistic view of a MSR salt fuel is a set of cations, which are changing their identity and their
oxidation states (through nuclear transmutation and fission), embedded in a sea of fluoride or chloride
anions. The interactions of the cations with the anions is thus quite complex, especially considering
dynamic oxidation states. The current approach is to fit a classical potential model to a set of results from
first-principles electronic structure calculations; if other liquid state data are available, such as densities or
diffusivities, they can also can be used to improve parameterization of the model. This approach can work
well if enough configurations can be sampled and there are no changes in the oxidation state or the
valency (number of “bonds” to a center). However, modeling more complex systems will require many
different force field interactions, with a concomitant increase in complexity. Thus, to treat the atom-atom
interactions, different approaches are needed that are not necessarily biased by the input force field and
can account for changing oxidation states and the presence of open shells due to unpaired electrons. One
approach would be to employ fast electronic structure methods for the atom-atom interactions, for
example, semi-empirical molecular orbital theory (SEMO) [31] and tight binding DFT (TB-DFT). Such
methods require parameterization, but they can be tuned for the atom and oxidation state in a
straightforward manner by comparison to high-level first-principles electronic structure calculations at the
correlated molecular orbital theory or DFT levels. [32] Even though these approaches have the advantage
of being able to deal with changes in electronic structure, they have some disadvantages: (1) the time
scales that can be simulated are short; and (2) calculations of liquid structural properties, such as the x-ray
structure function S(q), and dynamical properties, such as diffusion constants and viscosity, are not
guaranteed to be superior to those derived from classical simulations. Therefore, experimental validation
is an irreplaceable aspect of atomistic modeling.

FRD 1 17
Molten Salt Chemistry Workshop

Computational approaches also need to be developed to bound uncertainties in the results obtained from
simulations of molten salt systems to provide useful information for system models. It is critical to
provide error bars from the computations either through careful analysis of each term in the modeling, as
has been done for first-principles electronic structure of thermodynamic properties, or through proper
sampling. Current approaches for the latter case are usually based on Monte Carlo sampling, which is
computationally very expensive. New mathematical approaches are needed for more efficient uncertainty
analysis.

Ultimately, quantum mechanics, classical dynamics, coarse-grained models, continuum models, and
thermodynamic models must be integrated to achieve predictive insights ranging from chemical reactivity
at the atomistic level to bulk thermophysical and transport properties. Computational studies will require
large simulation sizes on the orders of tens of millions of atoms to cover the required spatial scale. Such
simulations will require access to DOE’s highest-performance computers alongside optimized exascale
software. For example, new and proposed heterogeneous computer architectures, such as graphics
processing units, may require improvements in the software to attain optimal performance.

Impact of Advanced Modeling

The ability to model physicochemical properties of molten salts and predict optimal compositions,
enabling a “materials by design” approach, would dramatically increase the rate of scientific advance.
Identification of promising melt compositions could be performed predictively, allowing for experimental
validation of promising compositions, rather than through brute-force approaches requiring large
investments of time, money, and material. Furthermore, predictive modeling allows for the interrogation
of conditions beyond what can be investigated by traditional experimental approaches, such as prolonged
exposure (years to decades) to elevated temperatures and extreme radiation fields.

A core capability will be the ability to design molten salts from a combination of simulation and
experimental results with the appropriate chemical and physical properties that will provide optimal
operation in a MSR. This will include the choice of an appropriate combination of anions and cations with
the best set of properties to provide reactors that have long lifetimes and can be safely operated with
maximal efficiency. An advantage of simulations is their ability to model complex mixtures with many
cations in different compositions, covering a much broader range of parameter space than can be done
experimentally in a shorter time.

Addressing these scientific challenges will enable scientists and engineers to improve the modeling and
simulation of MSRs, including the underlying physicochemical phenomena captured in the transport
properties used in the continuum theories of momentum, mass, and energy transfer used to design nuclear
reactors. This will allow predictions of effective diffusivities of fission products in complex molten salts,
thermal conductivity, and viscosity, which are parameters used in the traditional modeling and simulation
approach for reactor design. It will enable prediction of the nature and fate of volatile products, including
release rates. The data generated can be used to extrapolate limited experimental data into broader areas
of parameter space for use in reactor design, including new and improved materials. They will be able to
predict the nucleation and growth of insoluble products within a flowing molten salt and be able to bound
the uncertainty in our knowledge of molten salts.

The ability to predict the chemical and thermophysical properties of molten salts over a broad range of
temperature and composition will have significant impact on fields beyond MSR. A near-term goal
should include the development of a global thermodynamic model of MSR salts that is valid under a
broad range of conditions. This will be of great value to the reactor design community, and these
approaches can be applied to the design of other reactor systems, as well as complex industrial processes.
The tools and approaches developed for the design of molten salts will help enable the prediction of

18 FRD 1
 Molten Salt Chemistry Workshop

chemical properties under ionizing radiation conditions for other materials, in particular because of the
necessary validation of computational data against an extensive experimental data set. The development
of new MD approaches is required for simulations of complex systems over long times; and new
approaches for the development of force fields and fast electronic structure methods will be broadly used
by the simulation community for other systems, especially for materials and ionic liquids. [33] Finally,
the ability to reliably predict the solubility of fission products in salts will lead to new concepts of
solvation, which will be broadly used by the MSR community and others investigating the diverse
industrial applications of molten salts.

References

1. M. W. Rosenthal, P. R. Kasten, and R. B. Briggs. “Molten-salt reactors—History, status, and

potential.” Nucl. Appl. Tech. 8, 107–117, 1970.

2. W. R. Grimes. “Molten-salt reactor chemistry.” Nucl. Appl. Tech. 8, 137–155, 1970.

3. W. R. Grimes and D. R. Cuneo. “Molten salts as reactor fuel,” in Reactor Handbook, vol.1,
Materials, ed. C. R. Tipton Jr., C. R. Interscience Publishers, New York, 1960, 2nd ed., p. 425

4. D. F. Williams, and K. T. Clarno. “Evaluation of salt coolants for reactor applications.” Nucl.
Technol. 163(3), 330–343, 2008. http://www.ans.org/pubs/journals/nt/a_3992]

5. W. R. Grimes. Chemical Research and Development for the Molten-Salt Breeder Reactor,
ORNL/TM-1853, Oak Ridge National Laboratory, 1967.
http://web.ornl.gov/info/reports/1967/3445605995384.pdf]

6. D. F. Williams. Assessment of Candidate Molten Salt Coolants for the NGNP/NHI Heat-Transfer
Loop, ORNL TM-2006/69, Oak Ridge National Laboratory, 1960

7. J. H. Shaffer. Preparation and Handling of Salt Mixtures for the Molten Salt Reactor Experiment,
ORNL-4616, Oak Ridge National Laboratory, 1971; C. E. Bamberger, “Experimental techniques in
molten fluoride chemistry” in Advances in Molten Salt Chemistry, vol. 3, ed. J. Braunstein,
G. Mamantov, and G. P. Smith. Springer US, 1975. DOI: 10.1007/978-1-4615-8270-0

8. V. L. Cherginets. “Acid-base equilibria in ionic solvents,” chapter 10.4 in Handbook of Solvents, ed.
G. Wypych. ChemTec Publishing. Toronto, 2001, pp. 633–635; V. L. Cherginets and T. P. Rebrova.
“Studies of some acid-base equilibria in the molten eutectic mixture KCl-LiCl at 700 degrees C.”
Electrochim. Acta 45(3), 469–476, 1999.

9. H. Susskind, F. B. Hill, L. Green, S. Kalish, L. E. Kukacka, W. E. McNulty, and E. J. Wirsing.

Corrosion Studies for a Fused Salt-Liquid Metal Extraction Process for the Liquid Metal Fuel
Reactor, BNL-585, Brookhaven National Laboratory. US Atomic Energy Commission. 1960.

10. B. Mishra, and D. L. Olson. “Molten salt applications in materials processing.” J. Phys. Chem. Solids
2–4, 396-40, 2005.

11. J. E. Ricci. Guide to the Phase Diagrams of the Fluoride Systems, ORNL-2396, Oak Ridge National
Laboratory,1958; J. E. Ricci. “Phase diagrams of fused salts,” in Molten Salt Chemistry, ed. M.

FRD 1 19
Molten Salt Chemistry Workshop

Blander. 1964. 239–365; R. E. Thoma. “The role of phase equilibria in molten salt research” in
Molten Salts—Characterization and Analysis, ed. G. Mamantov. Marcel Dekker, NY, 1969, 81–119.

12. E. M. Levin, C. R. Robbins, H. F. McMurdie. Phase Diagrams for Ceramists, ed. M. K. Reser.
American Ceramic Society, multiple volumes 1964–1992; R. E. Thoma. “Phase diagrams of binary
and ternary fluoride systems.” Chapter 6 in Advances in Molten Salt Chemistry, vol. 3, ed.
J. Braunstein, G. Mamantov, and G. P. Smith. Springer US, 1975. 275–455.

13. L. Rycerz. “Practical remarks concerning phase diagrams determination on the basis of differential
scanning calorimetry measurements.” JTAC 113(1) July, 231–238, 2013.

14. K. A. Romberger, R. E. Thoma, and J. Braunstein. “New electrochemical measurements of liquidus

in LiF-BeF2 system: Congruency of Li2BeF4.” J. Phys. Chem. 76(8), 1154–1159, 1972.

15. W. Weppner, C. Li-chuan, and W. Piekarczyk. “Electrochemical determination of phase-diagrams

and thermodynamic data of multicomponent systems.” Zeitschrift Fur Naturforschung Section A-A,
J. Phys. Sci. 35(4), 381–388, 1980.

16. H. A. Friedman, G. M. Hebert, and R. E. Thoma. Thermal Analysis And Gradient Quenching
Apparatus And Techniques for The Investigation of Fused Salt Phase Equilibria, ORNL-3373, Oak
Ridge National Laboratory, 1963. https://www.osti.gov/scitech/biblio/4726793

17. J. P. M. van der Meer, R. J. M. Konings, K. Hack, and H. A. J. Oonk. “Modeling and calculation of
the LiF-NaF-MF3 (M = La, Ce, Pu) phase diagrams.” Chem. Mater. 18(2), 510–517, January 24,
2006.

18. S. Cantor, J. W. Cooke, A. S. Dworkin, G. D. Robbins, R. E. Thoma, and G. M. Watson. Physical

Properties of Molten-Salt Reactor Fuel, Coolant, and Flush Salts, ORNL-TM-2316, Oak Ridge
National Laboratory, August 1968.

19. G. J. Janz, G. L. Gardner, U. Krebs, and R. P. T. Tomkins. “Molten salts: Volume 4, Part 1, Fluorides
and mixtures electrical conductance, density, viscosity, and surface tension data.” J. Phys. Chem. Ref.
Data 3, 1–115, 1974.

20. M. Franko and C. D. Tran. “Thermal lens spectroscopy,” in Encyclopedia of Analytical Chemistry,
ed. R. A. Meyers, John Wiley and Sons Ltd., 2010, 1249–1279.

21. E. S. Chaleff, T. Blue, and P. Sabharwall. “Radiation heat transfer in the molten salt FLiNaK.” Nucl.
Technol. 196( 1), 53–60, 2016.

22. S. Su, T. Villalon Jr., U. Pal, and A. Powell. “Techniques for measuring solubility and electrical
conductivity. Advances in molten slags, fluxes, and salts,” pp. 465–475 in Proceedings of The 10th
International Conference on Molten Slags, Fluxes and Salts (MOLTEN16), ed. R. G. Reddy, P.
Chaubal, P. C. Pistorius, and U. Pal. TMS (The Minerals, Metals and Materials Society), 2016.

23. T. Villalón Jr., S. Su, and U. Pal. “Surface properties of molten fluoride-based salts: Advances in
molten slags, fluxes, and salts,” pp. 597–605 in Proceedings of the 10th International Conference on
Molten Slags, Fluxes and Salts (MOLTEN16), ed. R. G. Reddy, P. Chaubal, P. C. Pistorius, and U.
Pal. TMS (The Minerals, Metals & Materials Society), 2016.

20 FRD 1
 Molten Salt Chemistry Workshop

24. U. Pal, S. Su, and T. Villalon. “Molten flux design for solid oxide membrane-based electrolysis of
aluminum from alumina,” in Applications of Process Engineering Principles in Materials Processing,
Energy and Environmental Technologies, ed. M. L. Free, S. Alam, M. Zhang, and P. R. Taylor. The
Minerals, Metals, & Materials Series. ISBN: 978-3-319-51090-3 (print), 978-3-319-51091-0 (online).

25. K. A. Sense, C. A. Alexander, R. E. Bowman, and R. B. Filbert Jr. “Vapor pressure and derived
information of the sodium fluoride-zirconium fluoride system: Description of a method for the
determination of molecular complexes present in the vapor phase.” J. Phys. Chem. 61(3), 337–344,
1957.

26. T. Lichtenstein, N. D. Smith, J. Gesualdi, K. Kumar, and H. Kim. “Thermodynamic properties of

barium-bismuth alloys determined by EMF measurements.” Electrochim. Acta 228, 628–635, 2017.

27. B. Jabes, M. Agarwal, and C. Chakravarty. “Structure and transport properties of LiF-BeF2 mixtures:
Comparison of rigid and polarizable ion potentials.” J. Chem. Sci. 124, 261–269, 2012; L. C. Dewan,
C. Simon, P. A. Madden, L. W. Hobbs, and M. Salanne. ”Molecular dynamics simulation of the
thermodynamic and transport properties of the molten salt fast reactor fuel LiF–ThF4.” J. Nucl.
Mater. 434(1–3), 322–327, 2013; H. Luo, S. Xiao, S. Wang, P. Huai, H. Deng, and W. Hu.
“Molecular dynamics simulation of diffusion and viscosity of liquid lithium fluoride.” Comp. Mater.
Sci. 111, 203–208, 2016; M. Matsumiya and R. A. Takagi. “Molecular dynamics simulation of the
electric properties in molten chloride and fluoride quaternary systems.” Electrochim. Acta, 46(23),
3563–3572, 2001; H. O. Nam, A. Bengtson, K. Vörtler, S. Saha, R. Sakidja, and D. Morgan. “First-
principles molecular dynamics modeling of the molten fluoride salt with Cr solute.” J. Nucl. Mater.
449(1–3), 148–157, 2014; M. Salanne, C. Simon, P. Turq, N. Ohtori, and P. A. Madden. “Modeling
of molten salts,” in ed. F. Lantelme and H. Groult, Molten Salts Chemistry. Elsevier, Oxford, 1–16,
2013; M. Salanne, C. Simon, P. Turq, and P. A. Madden. “Simulation of the liquid–vapor interface of
molten LiBeF3.” Comptes Rendus Chimie 10(10–11), 1131–1136, 2007; M. Salanne, C. Simon,
P. Turq, and P. A. Madden. “Heat-transport properties of molten fluorides: Determination from first-
principles.” J. Fluorine Chem. 130(1), 38–44, 2009; J. Wang, J. Wu, G. Lu, and J. Yu, “Molecular
dynamics study of the transport properties and local structures of molten alkali metal chlorides. Part
III: Four binary systems: LiCl-RbCl, LiCl-CsCl, NaCl-RbCl, and NaCl-CsCl.” J. Mol. Liq. 238, 236–
247, July 2017, doi.org/10.1016/j.molliq.2017.03.103; S. Wang, H. Luo, H. Deng, S. Xiao, and
W. Hu. “A molecular dynamics study of the transport properties of LiF-BeF2-ThF4 molten salt.” J.
Mol. Liq. 234, 220–226, 2017.

28. A. Bengtson, H. O Nam, S. Saha, R. Sakidja, and D. Morgan. “First-principles molecular dynamics
modeling of the LiCl-KCl molten salt system.” Comp. Mat. Sci. 83, 362–370, 2014; J. Song, S. Shi,
X. Li, and L. Yan. “First-principles molecular dynamics modeling of UCl3 in LiCl-KCl eutectic.”
J. Mol. Liq. 234, 279–286, 2017.

29. S. M. Kathmann, G. K. Schenter, B. C. Garrett, B. Chen, and J. I. Siepmann. “Thermodynamics and

kinetics of nanoclusters controlling gas-to-particle nucleation.” J. Phys. Chem. C 113(24), 10354–
10370, 2009; L. D. Crosby, S. M. Kathmann, and T. L. Windus. “Implementation of dynamical
nucleation theory with quantum potentials.” J. Comput. Chem. 30(5), 743–749, 2009; S. M. Kathmann,
G. K. Schenter, and B. C. Garrett. “The critical role of anharmonicity in aqueous ionic clusters relevant
to nucleation.” J. Phys. Chem. C 111(13), 4977–4983, 2007; S. M. Kathmann. “Understanding the
chemical physics of nucleation.” Theor. Chem. Accounts 116(1–3), 169–182, 2006.

30. A. Di Cicco, A. Minicucci, and A. Filipponi. “New advances in the study of local structure of molten
binary salts.” Phys. Rev. Lett. 78, 460–463, 1997; A. Minicucci, and A. Di Cicco. “Short range

FRD 1 21
Molten Salt Chemistry Workshop

structure in solid and liquid CuBr probed by multiple-edge x-ray absorption spectroscopy.” Phys.
Rev. B. 56, 11456–11464, 1997; A. Di Cicco. “Local structure in binary liquids probed by EXAFS.”
J. Phys. Condens. Matter. 8, 9341–9345, 1996; A. Trapanati, A. Di Cicco, and M. Minicucci.
“Structural disorder in liquid and solid CuI at high temperature probed by x-ray absorption
spectroscopy.” Phys. Rev. B. 66, 014202, 2002 ; M. Inui, S. Takeda, K. Maruyama, Y. Shirakawa,
and S. Tamaki. “XAFS measurements on molten silver-halides.” J. Non-Cryst. Solids. 192–193, 351–
354, 1995; S. Takeda, Y. Kawakitaa, M. Inuib, and K. Maruyamac. “Local structure on molten
Ag(Cl1-xIx) mixtures.”J. Non-Cryst. Solids. 250–252, 410–414, 1999.

31. J. J. P. Stewart. “Optimization of parameters for semiempirical methods V: Modification of NDDO

approximations and application to 70 elements.” J. Mol. Model. 13, 1173–1213, 2007.

32. D. Feller, K. A. Peterson, and D. A. Dixon. “The impact of larger basis sets and explicitly correlated
coupled cluster theory on the Feller-Peterson-Dixon composite method,” in Annu. Rep. Comput. Chem.,
vol. 12, ed. D. A. Dixon, Elsevier, Amsterdam, 2016, 47–78; K. S. Thanthiriwatte, M. Vasiliu, S. R.
Battey, Q. Lu, K. A. Peterson, L. Andrews, and D. A. Dixon. “Gas phase properties of MX2 and MX4
(X=F, Cl) for M = group 4, group 14, Ce, and Th.” J. Phys. Chem. A, 119, 5790–5803, 2015; M. Vasiliu,
S. Li, D. Feller, J. L. Gole, and D. A. Dixon. “Structures and heats of formation of simple alkaline earth
metal compounds: Fluorides, chlorides, oxides, and hydroxides for Be, Mg, and Ca.” J. Phys. Chem. A
114, 9349–9358, 2010; K. O. Christe, R. Haiges, M. Vasiliu, and D. A. Dixon. “Formation mechanism of
NF4+ salts and extraordinary enhancement of the oxidizing power of fluorine by strong Lewis acids.”
Angew. Chem., Int., Ed., accepted April 2017; M. Chen, T. P. Straatsma, Z. Fang, and D. A. Dixon. “A
structural and electronic property study of (ZnO)n, n ≤ 168: The transition from zinc oxide molecular
clusters to ultra-small nanoparticles.” J. Phys. Chem. C 120, 20400–20418, 2016.

33. T. P. Straatsma, E. J. Bylaska, H. J. J. van Dam, N. Govind, W. A. de Jong, K. Kowalski, and

M. Valiev. “Advances in scalable computational chemistry: NWChem.” Annu. Rep. Comput. Chem.
7, 151–177, 2011; T. P. Straatsma and D. G. Chavarría-Miranda. “On eliminating synchronous
communication in molecular dynamics simulation to improve scalability. ” Comput. Phys. Commun.
84, 2634–2640, 2013; D. G. Chavarria-Miranda, K. Agarwal, and T. P. Straatsma. “Scalable PGAS
metadata management on extreme scale systems,” pp.103–111 in Proceedings of the 13th IEEE/ACM
International Symposium on Cluster, Cloud and Grid Computing (CCGrid). Delft, the Netherlands,
May 2013.; H. J. J. van Dam, W. A. de Jong, E. J. Bylaska, N. Govind, K. Kowalski, T. P. Straatsma,
and M. Valiev. “NWChem: Scalable parallel computational chemistry.” Comput. Mol. Sci., Wiley
Interdisciplinary Reviews, 1(6), 888–894, 2011.

22 FRD 1
 Molten Salt Chemistry Workshop

3. FRD 2: Understanding the Structure, Dynamics, and Chemical

Properties of Molten Salts
3.1 Background and Current Status

The structural and dynamical properties of molten salt systems have been investigated at varying levels of
detail over the past five decades, resulting in established techniques amenable to the characterization and
interrogation of MSR-relevant salts. Methods that have great utility include ultraviolet-visible-infrared
(UV-vis-IR) spectroscopy, x-ray absorption spectroscopy (XAS), Raman spectroscopy, and nuclear
magnetic resonance (NMR) spectroscopy. These methods provide information on (1) oxidation states of
transition metals, lanthanides, and actinides and (2) local structure and bonding of species in solution.
Complementary methods that apply to crystallized salts and salt mixtures in both the solid and liquid
states include XRD, neutron diffraction, and x-ray absorption fine structure spectroscopy (XAFS), with
current instrumentation and sampling methods permitting studies at temperatures of up to 1000 °C or
even higher in some cases. Application of these methods yields direct structural information, including
inter-atomic distances and coordination numbers. Additional information about cation-anion interactions
can be gained from NMR spectroscopy of ions like 7Li+, 19F-, 11B (in BF4-), 23Na+, 27Al3+, selected
lanthanides (e.g., 139La3+) and actinides (e.g., 235U3+,4+) cations, and other relevant NMR active ions.

An idealized perspective of the structure of a

molten salt is conceptualized as a set of cations
embedded in a sea of halide anions, typified by the
binary salt NaCl. Short-range structural properties
are attributable to competition between coulombic
interactions and Pauli repulsion, inducing strong
ordering effects and determining the closest
interatomic distances, respectively. A corollary
resulting from the influence of these contrasting
phenomena is that the innermost coordination shell
will always be completely made up of oppositely
charged species, affording an intermediate-scale
ordering that extends to several solvation shells. [1]
Such conclusions were validated by neutron
scattering measurements performed on molten
NaCl for three different isotopic compositions of Figure 11. Total structure factor, F(Q), data for liquid
Cl-, shown in Figure 11. (Note the difference in the Na Cl (curve A), NaCl (Curve B), and Na Cl
35 37

(Curve C) at 875 °C. The points are the

structure factor profile as a function of isotope.)
experimental data and the curves are derived from
The key finding here is that no evidence exists for the partial structure factors | Reproduced from F. G.
covalency or effects associated with polarization Edwards, J. E. Enderby, R. A. Howe, and D. A. Page. “The
(consistent with the fact that molten NaCl does not structure of molten sodium chloride.” J. Phys. C: Solid State
exhibit a Raman spectrum). The coordination Phys. 8(21), 3483–3490, 1975, http://iopscience.iop.org/
article/10.1088/0022-3719/8/21/018/pdf
number for unlike ions (first coordination shell)
was determined to be 5.8 ± 0.1 and the coordination number for like atoms (second shell) was found to be
13.0 ± 0.5; both quantities are consistent with values for their solid-state counterparts, i.e., in this simplest
case, the melt is close to isostructural with the solid.

However, introduction of any complexity beyond the aforementioned idealized system, such as
substitution of a divalent cation for a monovalent Na+, results in structural effects that remain difficult to
understand and fully rationalize. While the identity of the inner sphere coordinating atoms remains as
anticipated, variation of the inner-sphere cation-anion bond lengths is observed, and different outer sphere

FRD 2 23
Molten Salt Chemistry Workshop

cationic species arise from the interplay of multiple atomic interactions. Polarization-induced intermediate
structure can also occur, resulting in edge-sharing of the first-neighbor coordination shell. Such
phenomena are further influenced by the energy of the outer shell of the atoms, with formation dependent
upon the concentration of the salt and the number of cations present in the outer sphere. For example, the
structure of a ZnCl2 melt was characterized by neutron diffraction and XRD, revealing a tetrahedral first-
shell coordination environment for Zn2+, where corner sharing of adjacent octahedrals resulted in
intermediate ordering. [2] For atomistic simulations to reasonably represent this system, unrealistically
large dispersion interactions were required among the cations, [3] which were reconciled by the
introduction of a component representing polarization of the ionic electron cloud. [4]

It has also been demonstrated that for composite salts, the melt does not rearrange into a simple
combination of component parts. For example, XAFS spectroscopy was used to investigate the melting of
Rb2ZnCl4 [5] in comparison with the liquid structure of ZnCl2 and RbCl. While molten ZnCl2 displayed
tetrahedrally-coordinated corner-sharing chlorines, forming the extended network discussed above, RbCl
displayed significant distortion of the chloride shell around Rb, indicative of significant movement of Rb+
and Cl- in the molten state. [6] However, in molten Rb2ZnCl4, isolated [ZnCl4]2- units were identified with
a chlorine coordination of 7.6 around Rb+, in contrast to a chlorine coordination of 4.8 in molten RbCl.
Thus, the salt structure formed by their combination does not afford the same metal coordination
environment for Rb+ and Zn2+ as in RbCl and ZnCl2, respectively.

In the above example, the room-temperature solid state structure of Rb2ZnCl4 was similar to that in the
molten form; however, this is not universally true. In other instances, atomic rearrangement occurs, so
that the molten structure differs dramatically from that of the solid precursor. For example, Crozier and
colleagues investigated the liquid structures of [Bu4N][MnBr3] and [Bu4N]2[MnBr4], identifying a
bromine coordination number of 3.1 around manganese in the liquid state, compared with a coordination
number of 6 in the solid-state structure. [7] Molten PbCl2 has also shown a change in local coordination
environment upon melting, with a decrease in chlorine ligation from nine to six. [8]

Studies of lanthanide halides have also revealed unanticipated structural changes. Upon melting, LaCl3
displays a first shell coordination number suggesting an equal mixture of [LaCl7]4- and [LaCl8]5-. A
decrease in the La3+-Cl- bond length has also been observed, contrary to what would be anticipated at
elevated temperatures. [9] Insight into the latter comes from work performed on an AgBr salt. It revealed
a similar first-shell bond length contraction upon melting, while radial distribution functions of atomic
positions revealed greater asymmetry toward longer distances and indicated the average bond length to be
increasing, as is expected from thermal expansion. [10] Such behavior suggests in both instances the
metal is more appropriately conceptualized as forming MClx molecules with the anion than as isolated
Mx+ dissolved in a Cl- solvent. Proceeding down the periodic group revealed significant structural
differences in YCl3 compared with LaCl3. The Y3+ – Y3+ distance was 0.25 Å closer than the analogous
La3+ – La3+, which was attributed to different linkages in the different salt. Whereas LaCl3 displays
corner-sharing six-fold complexes, YCl3 is proposed to have edge-sharing complexes. [11] Thus,
structural properties can vary dramatically based on the species of the cation, which has been previously
demonstrated to significantly influence chemical properties, such as solubility of actinide complexes. [12]

While molten salts are regarded as purely ionic solutions, the formation of MClx species upon melting, as
discussed above, suggests covalent interactions are also a significant, albeit unanticipated, possibility.
Further work was performed by Okamoto and colleagues in which YCl3 and YBr3 were dissolved in a
LiCl-KCl eutectic with LiBr, respectively. Investigation of the extended XAFS spectrum of Y revealed a
sharpening of the first major feature and a small decrease in bond length. These results are consistent with
improved stabilization of a YX3 molecular complex upon further dilution with the alkali salt. [13] Copper
and silver halides have also been shown to display significant covalent behavior. Work by Di Cicco and
colleagues indicates nearly total covalent bonding observed in solid CuBr, which is similar to what is

24 FRD 2
 Molten Salt Chemistry Workshop

observed in the liquid state. A quasi-continuous phase change occurred during melting, which afforded a
significant decrease in Cu-Br bond length in the liquid compared with the solid. [14] A similar structure
for liquid CuI was also determined, supporting the same conclusion. [15] Studies of AgCl-AgI mixtures
reveal silver coordination environments composed of two chloride and two iodide atoms at temperatures
below the melting point; but upon melting, the chloride contribution decreased significantly while the
iodide was invariant. This decrease is thought to have occurred as a result of the removal of chloride
atoms from the first coordination shell as the bonding strength between silver and iodide increases. [16]

Chemical bond lengths are representative of bond strengths, indicative in turn of the chemical reactivity
of the molten salt. This is clearly demonstrated by Cho et al., who investigated the covalency of
uranium(III) and uranium(IV) chlorides and observed that as the U-Cl distance became shorter, the
interactions became stronger and the reactivity decreased. [17] As noted above, the molecular structure of
the melt is known to influence the solubility and chemical reactivity of the constituents. Accordingly,
detailed structural knowledge of molten salts is essential for understanding and predicting the chemical
behavior in a MSR, not only to determine bulk physicochemical properties but also to understand the
solubility and behavior of fission, activation, and corrosion products present in the melt.

In addition to the majority constituents in the starting salt, it is necessary to investigate, identify,
understand, and predict the multi-faceted behaviors of the accumulating fission products. There are
numerous examples in the literature of spectroscopic and diffraction studies that are directly applicable to
this research direction; and the behavior of the “sometimes soluble” niobium, tellurium, and insoluble
noble metal fission products (elements between niobium and tellurium) are particularly important. For
example, Toth et al. measured the UV-vis-IR absorption spectrum of telluride species in molten Li2BeF4.
Their work showed that although Te2- was not detectable in Li2BeF4 by optical absorption spectroscopy,
when the Te2- ion was mixed with metallic tellurium, a soluble poly-telluride species formed that could be
detected in the absorption spectrum of the molten salt solution [18] (Figure 12a). This study provides
insight into the solvation properties that could affect tellurium migration in an MSR coolant salt system.

In light of the preference for using low-enrichment fissile constituents, very high actinide salt
concentrations are needed, affording salts with very different chemical properties and coordination
chemistry from dilute actinide-containing solutions. For example, the Raman spectrum of molten LiF-
NaF-ThF4 at 650 °C shown in Figure 12b sheds light on the coordination behavior of Th4+ in molten
fluorides. Eight-coordinated Th4+ was identified in melts with excess fluoride ions by comparing their
spectra with that of crystalline K5ThF9. The existence of seven-coordinated Th4+, present in fluoride-ion-
deficient melts, was identified by shifts in the frequency of the 474 cm−1 band with changes in melt
composition. This represents a case in which the coordination chemistry of an actinide cation is affected
by solvent composition.

High-temperature solid state and liquid NMR can provide insights into the local structure of species in the
melt and interactions of the cation-anion. For example, in the high-temperature 19F NMR spectra for solid
and molten YF3-LiF (Figure 13), there is a substantial down field shift in the 19F resonance with
increasing LiF content. This large shift indicates that the nature of the interaction between F- anions and
Y3+ cations is influenced by the amount of Y3+ in the melt. The higher the quantity of Y3+, the more
bridging the F- ions become. The differences in line width for the solid and the liquid state are due to
differences in relaxation times. Note also in Figure 13 that the 19F NMR resonance pattern of the solid
shows a significant difference from that of the liquid with increasing YF3 content. One interpretation of
this observation is that there are multiple F- site types in the solid that are randomized by the ionic
mobility in the liquid state.

FRD 2 25
Molten Salt Chemistry Workshop

a b

Ultraviolet-visible absorption
spectrum of polytelluride ion in Raman spectra of (A) molten LiF-NaF-ThF4 (40-46-14 mol%)
molten Li2BeF4. at 650 °C and (B) polycrystalline K5ThF9 at 25 °C.

Figure 12. (a) Ultraviolet-visible and (b) Raman spectroscopy results reveal species identity in molten
salts. | Images courtesy of Oak Ridge National Laboratory.

Figure 13. High temperature 19F NMR spectra of YF3–LiF as a function of composition. Black
lines are for the molten state; dotted lines are for the solid state at temperatures 700–1100 °C
(estimated). | Reproduced from Anne-Laure Rollet et al. ”In situ high temperature NMR and EXAFS
measurements in rare-earth fluoride molten salts,” Comptes Rendus Chimie 2004 7(12), 1135–1140. 2004.

26 FRD 2
 Molten Salt Chemistry Workshop

Elsevier Masson SAS. All rights reserved.

In similar measurements for the LaF3-LiF system, the corresponding 139La NMR spectra showed that the
first coordination shell around the La3+ cations changed very little with composition. [19] The radial
structure functions derived from analysis of the extended XAFS (EXAFS) spectra in Figure 14 support
the 139La NMR results for the first coordination shell, but they give evidence of significant differences in
coordination environment beyond the first shell when the solid and molten states are compared.

The results in Figures 11 through 14 provide excellent examples of how the proposed measurement
methods provide the information needed regarding oxidation states, coordination numbers, inter-ion
spacing, local and long-range order, and solvation that will be used to elucidate the structure and
dynamics of MSR-relevant molten salt systems.

Figure 14. La LIII EXAFS spectra of molten LaF3-LIF (20-80 mole%): (a) solid phase (solid line), (b) melt at 780
°C (open triangles), and (c) melt at 850 °C (open diamonds). | A.-L. Rollet, “Local structure analogy of
lanthanide fluoride molten salts.” International Journal of Thermophysics, Vol. 26(4), 1115–1125, 2005. Used with
permission of Springer.

Dissolved species in molten salts exist as highly mobile metal complexes, whereas insoluble and semi-
soluble elements may exist as suspended fine particles (metallic or salt precipitates) or as surface deposits
on the wetted materials. In both instances, the insoluble species exist in dynamic equilibrium with the
dissolved species and temperature, concentration, speciation, and other factors impact this equilibrium. At
the concentrations required for soluble and semi-soluble elements to appear in the salt, the short-range
structure of the salt (complexation between cations and anions) is sufficiently affected to impact a broad
range of thermophysical properties, such as viscosity, diffusion and transport coefficients (thermal or
concentration gradient induced migration), and chemical activity (the propensity of a particular phase to

FRD 2 27
Molten Salt Chemistry Workshop

be stable or instable). These changes in physicochemical behavior dramatically affect critical engineering
factors, such as corrosion behavior and thermal conductivity.
Design and development of a fully functional MSR that meets all the requirements for sustained
operation, including materials compatibility, manageable radionuclide containment and disposition, and
tenable economics will require detailed knowledge of the structure and chemical speciation of complex
ions and other associated species in the molten salt system, as well as real-time monitoring of the species
in solution. The net oxidation/reduction condition of the salt must be tracked and understood as a function
of burnup, and the time evolution (dynamics) of molten salt structure and elemental speciation as a
function of burnup must be fully elucidated on the local and global scales. The influence of these
structural and dynamical behaviors on other thermophysical and transport properties must be measured
and reconciled.

3.2 Technical Challenges and Research Directions

To understand how the structure and dynamics of molten salts impact their physical and chemical
properties—such as viscosity, solubility, volatility, and thermal conductivity—it is necessary to determine
the structure and speciation of the molten salt using advanced spectroscopic and scattering methods that
provide information at the atomic/molecular scale. Where necessary, innovative experimental methods for
microscopic and macroscopic property measurements will be employed, with emphasis on high-
throughput techniques (i.e., ones applicable to radionuclides) to accommodate the efficient gathering of
data for the large information base expected to be needed to underpin/support MSR design and
development. Real-time analysis methods need to be developed for monitoring key chemical species in
solution, allowing optimization of reactor performance and lifetime. This experimental work must be
closely interfaced with first-principles computational and data analysis approaches to establish validated
predictive models for system performance.

3.2.1 Determining Molecular Structure by X-ray and Neutron Scattering and

Spectroscopy
In the nearly five decades following the decommissioning of ORNL’s MSRE, astounding gains have been
made in the development of advanced x-ray and neutron scattering and spectroscopy techniques, which
will enable precise investigation of the structure of molten salts and the coordinate environment of the
actinides and fission products. Third-generation synchrotron light sources typically have brilliance in
excess of 1018 photons s-1 with exquisite polarization, collimation, energy tunability, and low emittance,
enabling real-time analysis with high-energy techniques that were not possible during the MSRE.
Similarly, access to high-flux neutron scattering facilities now allows the investigation of structures,
coordination environments, and buried structural features and dynamics at high temperatures; and
additional information can be gained by leveraging isotopic contrast.

Development of robust methodologies and high-temperature sample cells (up to 850 °C) is essential for
the application of such techniques to the investigation of molten salt structure at operational temperatures
to enhance the understanding of fundamental salt properties in the bulk. The application of XAFS (see the
sidebar “XAFS of Molten Salts” on p. 29) to the molten salts relevant to a MSR has been limited, yet it
provides an avenue for investigating the local atomic structures of melt constituents. It is also a potential
means of identifying fission and corrosion product oxidation states and geometries, particularly in
instances where dilute elemental concentrations prohibit the application of other techniques.

Application of soft x-rays, suitable for XAFS of light elements (before scandium in the periodic table,
with specific MSR-relevance including Na, Mg, and Al), may afford important structural information for
the salts in the melt phase. A significant challenge is the design of sample holders capable of preserving
an air- and moisture-free environment at MSR operational temperatures (up to and exceeding 850 °C)

28 FRD 2
 Molten Salt Chemistry Workshop

while sufficiently enclosing the low-viscosity melt and permitting sufficient penetration of low-energy x-
rays. Another difficulty is the need for sample holders that are amenable to the analysis of low-
concentration components in the melt. Data for trace elements are customarily collected with fluorescence
detectors, but few instances of such measurements exist in the molten salts literature, none of which
involves elevated temperatures or carefully controlled atmospheres. [20]

XAFS of Molten Salts

X-ray absorption fine structure (XAFS) spectroscopy affords a technique for investigating
coordination geometry and oxidation state for most elements in the periodic table. [21] Relying upon
the absorption of an x-ray with appropriate energy for the excitation of a core electron, it is an
element discriminating technique and is largely unencumbered by requirements of phase, temperature,
or structural ordering. [22] For example, XAFS spectroscopy was applied to investigation of the
uranium binding mode by amidoxime-functionalized polymeric adsorbents. A comparison against
small molecule standards previously used to conceptualize uranium binding modes revealed distinctly
different behaviors, and fits of the data allowed for the proposal of unanticipated coordination modes.
XAFS was uniquely capable of obtaining this information because it provides high sensitivity,
elemental discrimination, and minimal restrictions on sample form. It displays the potential of such
advanced quantum-beam techniques for answering previously unassailable science questions.
| Reproduced from C. W. Abney, R. T. Mayes, M. Piechowicz, Z. Lin, V. S. Bryantsev,G. M. Veith, S. Dai, and W. Lin.
“XAFS investigation of polyamidoxime-bound uranyl contests the paradigm from small molecule studies.” Energy
Environ. Sci. 9(2), 448–453, 2016, with permission of The Royal Society of Chemistry.

High energy x-ray or neutron scattering measurement affording the elemental pair distribution function
(PDF) is another quantum-beam technique capable of affording otherwise unobtainable insights regarding
molecular structure. The PDF is the Fourier transform of the total scattering pattern and affords the
probability of finding an atom at a given distance. Investigation of the PDF allows interrogation of bond
lengths and crystal structures and is a particularly powerful technique following the subtraction of
background scattering (i.e., differential-PDF). [24] Recent work has demonstrated the capability of PDF
studies to investigate the exchange kinetics of the first and second coordination environment of aqueous
metal halides (see Figure 15), [25, 26] yet extension of such efforts to molten salts is non-trivial. Neutron-
PDF investigations afford the opportunity to introduce isotopic contrast (due to different scattering cross
sections of the elements), allowing the elucidation of which molecular moieties contribute at a given
spatial range. While the inherent opportunity for investigating the structure of materials devoid of long-
range ordering is obvious, isotopic contrast nevertheless diminishes with increasing Z-number and is

FRD 2 29
Molten Salt Chemistry Workshop

typically insufficient for elements heavier than phosphorous. Therefore, the development of approaches
by which maximum information can be gained from scattering data possessing minimal contrast remains
an area requiring further development. Finally, it is important to mention that these techniques are just a
subset of potential quantum-beam approaches to investigating the structure and dynamics of molten salts,
and that the technique used is contingent upon the scientific question to be answered.

Solid-state and liquid NMR can also be used to

determine the local structure of species in the melt
and the solid and the cation–anion interactions.
Although the temperature range of magic-angle
spinning (MAS) techniques is limited to about
500 °C by the mechanical strength of the zirconia
rotors, high-temperature MAS probes using a rotor
encapsulated in a CO2 laser-absorbing sleeve are
now commercially available. They can be used at
temperatures above 2000 °C and enable container-
less heating and aerodynamic levitation. [27] The
measured chemical shift provides information on
the local environment around the observed nucleus
and depends on the nature and number of
neighbors. As seen in Figure 13, the evolution of Figure 15. Differential-PDF measurements were used
the high temperature 19F resonances in YF3 in to investigate the first and second coordination
varying amounts of LiF (measured at 20 K above environment of aqueous uranyl as a function of
the melting point of the binary mixture, 700– chloride concentration, revealing the formation of
1100 °C) signifies a progressive modification of the inner sphere chloride complexes and reorientation
fluorine environment from free F- in LiF toward of second sphere solvent molecules upon chloride
YFx3-x. The high-temperature 19F and 7Li NMR coordination. The extension of such techniques to
investigating the speciation and structure of molten
spectra also reveal that the coordination salt components is expected to afford new insights
environment of YF3 in varying amounts of LiF is into physical phenomena such as fission product
different between the melt and the fused state. [28] precipitation, corrosion, and gas entrainment.
Since the fluoride is shielded less in the melt, the | Reprinted with permission from L. Solderholm, S.
molten salt is more ionic and thus exhibits a Skanthakumar, and R. E. Wilson. “Structural
correspondence between uranyl chloride complexes in
different coordination environment. This is not solution and their stability constants.” J. Phys. Chem. A
expected to be the case with chlorides, in which the 115(19), 4959–4967, 2011. Copyright 2011, American
shielding is lower than in fluoride and complex ion Chemical Society.
formation has been observed. Coupling high-
temperature NMR with complementary techniques such as EXAFS and Raman spectroscopy leads to a
better description of the local structure of the salt at high temperatures and provides insights into the
origin of the physical and chemical properties.

3.2.2 Correlating Electrochemistry and Spectroscopy for Online Monitoring and

Predictive Modeling
One of the challenges in a MSR is that fission creates an oxidizing chemical environment that can lead to
corrosion of the metal walls. For example, when fission of 235UF4 occurs, four fluoride ions are released
from U4+ and two fission products are formed—such as 92Kr and 141Ba2+ or 140Xe and 94Sr2+—which
subsequently decay to more stable isotopes. However, the net oxidation state of the products differ from
those of the parent actinide, which leads to free fluoride in the system. To prevent corrosion, online
monitoring of the redox properties of the melt is essential to enable real-time analysis of the state of the
molten salt media. The approaches used should be capable of determining the total uranium as well as the
total fraction of uranium in each oxidation state to maintain proper redox chemistry. A mixed-valance

30 FRD 2
 Molten Salt Chemistry Workshop

uranium fuel in a halide salt is the starting point, but the ratio of the oxidation states of U3+/U4+ will
change over the course of the fuel lifetime. Direct monitoring of the fission product buildup is required to
determine the optimum time for fission product removal/molten salt replenishing. Also, monitoring for
trace impurities or corrosion products will enable feedback on the reactor health.

Spectroscopic monitoring of the molten salts system can simultaneously provide information on the
dissolved concentration and oxidation state of optically active species, and on the identity of the dissolved
specie as a complex ion. The measurement of U [29–31], Np [32, 33], and Pu [34–36] by absorption
spectroscopy in molten salt solutions has been extensively studied since as early as the 1960s. The
absorption spectra of the various oxidation states of U, Pu, and Np in molten salt media are quite distinct
(as they are in aqueous solution) and have been used as the basis for quantitative measurement.
Measurement of U, Np, and Pu species in molten salt solutions has also been achieved by electrochemical
methods (see review article by Zhang [37] and references therein). The strength of the electrochemical
methods includes durability in strong radiation fields and high operating temperatures. The combination
of electrochemical methods with spectroscopic measurements, in conjunction with chemometric modeling
techniques, would prove a powerful analytical tool for real-time monitoring. For example, Figure 16a
presents absorption spectra for the interconversion between U(III) and U(IV) by electrochemical
modulation. [30] The spectroelectrochemical interconversion of Np(IV) to Np(III) is shown in
Figure 16b. [32]

U (III)

U (IV)

Figure 16. (Left, a) UV-vis absorption spectra following the reduction of U(IV) to U(III) within an alkali
chloride molten salt. Reprinted with permission from I. B. Polovov, V. A. Volkovich, J. M. charnock, et al. In situ
spectroscopy and spectroelectrochemistry of uranium in high-temperature alkali chloride molten salts.” Inorganic
Chemistry 47(17), 7474–7482, 2008. Copyright 2008, American Chemical Society. (Right, b) The
spectroelectrochemical UV-vis measurement of the conversion of Np(VI) to Np(II) in molten LiCl-KCl
eutectic media | Reprinted from I. B. Polovov, C. A. Sharrad, L. May, V. A. Volkovich, and B. D. Vasin.
“Spectroelectrochemical study of neptunium in molten LiCl-KCl eutectic.” Z Naturforsch A 62(12), 745–748, 2007.

Spectroscopic measurements, [38] electrochemical measurements, [39] and combined spectro-

electrochemical [39, 40] methods have been demonstrated in alkali-halide eutectic solutions containing a
mixture of lanthanide chlorides. These techniques enable the quantitative analysis of mixtures of metal
salts in molten salt media. [38] As shown in Figure 17, the use of multivariate regression techniques for
the quantitative measurement of multiple species with overlapping/interfering absorption bands has been
demonstrated. It would be advantageous to apply this approach to several key actinides (U, Np, and Pu).
Furthermore, both Raman and UV-vis methods can be incorporated to expand the range of interrogation
capabilities.

FRD 2 31
Molten Salt Chemistry Workshop

Figure 17. UV-vis absorbance spectra of lanthanides within alkali chloride molten salt (left) and example
chemometric modeling results to quantify neodymium within single-component and multicomponent
molten salt solutions (right) where training sets data (blue + symbol) were used to build the models that
were applied to validation sets (red triangles); a 1:1 correlation (black line) is bounded by the 95%
confidence intervals (red dashed lines). |Reproduced from C. A. Schroll, A. M. Lines, W. R. Heineman, and S. A.
Bryan. “Absorption spectroscopy for the quantitative prediction of lanthanide concentrations in the 3LiCl-2CsCl
eutectic at 723 K.” Anal. Methods–UK 8(43), 7731–7738, 2016, with permission of the Royal Society of Chemistry.

In addition to redox state, the Lewis acid-base properties of the melt are important in determining the
solubility of the actinides in the melt and the corrosivity toward structural materials. The solubility of
actinides in a molten salt is fundamentally related to the local basicity of the anions surrounding the metal
cation. Phenomenological approaches have been employed to investigate the effects of molten-salt
compositions on the solubility properties of actinides. [41] No attempts have been made to correlate these
observations with a microscopic rationalization of actinide species solubility in molten salts. In other
words, a bridge between macroscopic solubility experiments and microscopic local structural properties is
lacking. This information gap has prevented the development of predictive models of molten salts with
good solubility properties. Research is needed to uncover the correlation between the solubility of
actinides in molten salts and their fundamental spectral fingerprints.

In the framework of Lewis acid-base theory, a metal ion behaves as a Lewis acid while the anionic
network behaves as a Lewis base. The halide anions of a molten salt surrounding the metal ion donate the
negative charge, and this donation reduces the positive charge on the metal ion. [42, 43] The greater the
negative charge borne by the halide anions, the greater will be the electron donation. If the metal ion can
signal (via optical spectroscopy) the extent to which its positive charge changes, then there is the
possibility of using it as a probe for the quantitative measurement of the basicity of the corresponding
molten salt medium. The basicity scale thus obtained has been called “optical basicity”; it is one of the
most fundamental parameters that characterize a specific molten salt, in analogy to the way pH is used to
describe properties of aqueous systems. Metal ions such as Tl+, Pb2+, and Bi3+ have been employed as
optical probes to measure the basicity of industrially interesting oxide glasses. As was pointed out earlier,
actinide solutes are known to have characteristic UV-vis and near-IR absorption bands resulting from f-f
and charge-transfer electronic transitions. Thus, a new online and accurate optical basicity scale is
needed for actinide-containing molten salts that uses optical spectra of actinide ions as direct probes of
the basicity of the melt.

32 FRD 2
 Molten Salt Chemistry Workshop

References

1. M. Salanne and P. A. Madden. “Polarization effects in ionic solids and melts.” J. Mol. Phys. 109,
2299–2315, 2011.
2. S. Biggin, M. Gay, and J. E. Enderby. “The structures of molten magnesium and manganese (II)
chlorides.” J. Phys. C: Solid State Phys 14, 3577, 1981; T. Triolo and A. H. Narten. “X-ray
diffraction study of molten zinc chloride at 323°C.” J. Chem. Phys. 74, 703, 1981; D. A. Allen, R. A.
Howe, and N. D. Wood. “Tetrahedral coordination of Zn ions in molten zinc halides.” J. Chem. Phys.
94, 5071, 1991; A. Zeidler, P. S. Salmon, R. A. Martin, T. Usuki, P. E. Mason, G. J. Cuello, S.
Kohara, and H. E. Fischer. “Structure of liquid and glassy ZnCl2.” Phys. Rev. B. 82, 104208, 2010.
3. L. V. Woodcock, C. A. Angell, and P. Cheeseman. “Molecular dynamics studies of the vitreous state:
Simple ionic systems and silica.” J. Chem. Phys. 65, 1565, 1976; P. N. Kumta, P. A. Deymier,and
S. H. Risbud. “Improved rigid ion model of molten zinc chloride.” Physica B 153, 85, 1988.
4. M. Wilson and P. A. Madden. “Polarization effects in ionic systems from first principles.” J. Phys.:
Condens. Matter. 5, 6833, 1993; M. Wilson and P. A. Madden. “Short- and intermediate-range order
in MgCl2 melts: The importance of anionic polarization.” J. Phys.: Condens. Matter. 5, 2687, 1993.
5. H. Li, K. Lu, Z. Wu, and J. Dong. “EXAFS studies of molten ZnCl2, RbCl, and Rb2ZnCl4.” J. Phys.
Condens. Matter 6, 3629–3640, 1994.
6. K. Lu, H. Li, Z. Wu, J. Dong, and Z. Cheng. “Structural studies of some chloride melts with
EXAFS.” Physica B: Condens. Matter 208–209, 339–343, 1995.
7. E. D. Crozier, N. Alberding, and B. R. Sundheim. “EXAFS study of bromomanganate ions in molten
salt.” J. Chem. Phys. 79, 939–943, 1983.
8. M. Ablanov, H. Matsuura, and R. Takagi. “Extended x-ray absorption fine structure spectroscopy
study of PbCl2 molten salt.” Electrochemistry (formerly Denki Kagaku) 67(8), 839–842, 1999.
9. M. Inui, K. Maruyama, Y. Shirakawa, S. Takeda, and S. Tamaki. “EXAFS measurements on molten
AgBr-AgI mixtures.” J. Non-Cryst. Solids 205–207, 155–158, 1996; M. Inui, M. S. Takeda, K.
Maruyama, Y. Shirakawa, and S. Tamaki. “XAFS measurement on molten silver halides.” J. Non-
Cryst. Solids 192–193, 351–354, 1995.
10. A. Di Cicco, M. Taglienti, M. Minicucci, and A. Filipponi. “Short-range structure of solid and liquid
AgBr determined by mutiple-edge x-ray absorption spectroscopy.” Phys Rev. B. 62, 12001–12013,
2000; A. Di Cicco, and M. Minicucci. “Solid and liquid short-range structure determined by EXAFS
multiple-scattering data analysis.” J. Synchrotron Rad. 6, 255–257, 1999.
11. Y. Okamoto, H. Shiwaku, T. Yaita, H. Narita, and H. Tanida. “Local structure of molten LaCl3 by K-
absorption edge XAFS.” J. Mol. Struct. 641, 71–76, 2002.
12. S. Dai, L. M. Toth, G. D. Del Cul, and D. H. Metcalf. “Solubilities of uranium(IV) dioxide in
magnesium chloride, calcium chloride, and aluminum chloride melts: A comparative study.” J. Phys.
Chem. 100, 220–223, 1996.
13. Y. Okamoto, T. Yaita, and K. Minato. “High-temperature XAFS study of solid and molten SrCl2.”
J. Non-Cryst. Solids. 333, 182–186, 2004.
14. A. Di Cicco, M. Minicucci, and A. Filipponi. “New advances in the study of local structure of molten
binary salts.” Phys. Rev. Lett. 78, 460–463, 1997; M. Minicucci and A. DiCicco. “Short-range
structure in solid and liquid CuBr probed by multiple-edge x-ray-absorption spectroscopy.” Phys.
Rev. B. 56, 11456–11464, 1997; A. Di Cicco. “Local structure in binary liquids probed by EXAFS.”
J. Phys. Condens. Matter 8, 9341–9345, 1996.

FRD 2 33
Molten Salt Chemistry Workshop

15. A. Trapananti, A. Di Cicco, and M. Minicucci. “Structural disorder in liquid and solid CuI at high
temperature probed by x-ray absorption spectroscopy.ˮ Phys Rev B. 66, 014202, 2002.
16. S. Takeda, Y. Kawakita, M. Inui, and K. Maruyama. “Local structure of molten Ag(Cl1-xIx) mixtures.ˮ
J. Non-Cryst. Solids 250–252, 410–414, 1999.
17. Y. H. Cho, S.-E. Bae, D.-H. Kim, T.-H. Park, J. Y. Kim, K. Song, and Y. W. Yeon. “On the
covalency of U(III)-Cl, U(IV)-Cl bonding in a LiCl-KCl eutectic melt at 450 degrees C:
Spectroscopic evidences from their 5f-6d and 5f-5f electronic transitions.” Microchem. J. 122, 33–38,
2015.
18. L. M. Toth and G. E. Boyd. “Raman spectra of thorium (IV) in fluoride complex ions in fluoride
melts.” J. Phys. Chem. 77(22), 2654–2657, 1973.
19. A.-L. Rollet, S. Godier, and C. Bessada. “High temperature NMR study of the local structure of
molten LaF3-AF (A= Li, Na, K, and Rb) mixtures.” Phys. Chem. Chem. Phys. 10, 3222–3228, 2008.
20. A. J. Dent, K. R. Seddon, and T. Welton. “The structure of halogenometallate complexes dissolved in
both basic and acidic room-temperature halogenoaluminate(III) ionic liquids, as determined by
EXAFS.” J. Chem. Soc. Chem. Commun. 315–316, 1990; N. J. Willams, J. Dehaudt, V. S. Bryantsev,
H. Luo, C. W. Abney, and S. Dai. “Selective separation of americium from europium using 2,9-
bis(triazine)-1,10-phenanthrolines in ionic liquids: A new twist on an old story.” Chem. Commun. 53,
2744–2747, 2017; C. W. Abney, C. Do, H. Luo, J. Wright, L. He, and S. Dai. “Controlling the
intermediate structure of an ionic liquid for f-block element separations.” J. Phys. Chem. Lett. 8(9),
2049–2054, 2017.
21. J. J. Rehr and R. C. Albers. “Theoretical approaches to x-ray absorption fine structure.” Rev. Mod.
Phys. 72(3), 621–654, 2000.
22. S. Calvin. XAFS for Everyone. CRC Press, Boca Raton, Florida, 2013.
23. C. W. Abney, R. T. Mayes, M. Piechowicz, Z. Lin, V. Bryantsev, G. M. Veith, S. Dai, and W. Lin.
“XAFS investigation of polyamidoxime-bound uranyl contests the paradigm from small molecule
studies.” Energy Environ. Sci. 9 (2), 448–453, 2016.
24. K. W. Chapman, P. J. Chupas, and C. J. Kepert. “Selective recovery of dynamic guest structure in a
nanoporous Prussian blue through in situ x-ray diffraction: A differential pair distribution function
analysis.” J. Am. Chem. Soc. 127(32), 11232–11233, 2005.
25. L. Soderholm, S. Skanthakumar, and R. E. Wilson. “Structural correspondence between uranyl
chloride complexes in solution and their stability constants.” J. Phys. Chem. A 115(19), 4959–4967,
2011.
26. A. Kalaji, S. Skanthakumar, M. G. Kanatzidis, J. F. Mitchell, and L. Soderholm. “Changing hafnium
speciation in aqueous sulfate solutions: A high-energy x-ray scattering study.” Inorg. Chem. 53(12),
6321–6328, 2014.
27. B. Coté, D. Massiot, F. Taulelle, and J.-P. Coutures. “27Al NMR spectroscopy of aluminosilicate
melts and glasses.” Chem. Geology 96, 367–370, 1992; R. Winter, A. Jonbes, R. Shaw-West,
M. Wolff, P. Florian, and D. Massiot. “Laser-heated high-temperature NMR: A time-resolution
study.” Appl. Magn. Reson. 32, 635–646, 2007.
28. A.-L. Rollet, C. Bassada, A. Rakhmatoulline, Y. Auger, P. Melin, M. Gailhanou, and D. Thiaudière.
“In situ high temperature NMR and EXAFS experiments in rare-earth fluoride molten salts.” Comptes
Rendus Chimie 7(12), 1135–1140, 2004.

34 FRD 2
 Molten Salt Chemistry Workshop

29. T. Nagai, A. Uehara, T. Fujii, O. Shirai, N. Sato, and H. Yamana. “Redox equilibrium of U4+/U3+ in
molten NaCl-2CsCl by UV-Vis spectrophotometry and cyclic voltammetry.” J. Nucl. Sci. Technol.
42(12), 1025–1031, 2005.
30. I. B. Polovov, V. A. Volkovich, J. M. Charnock, B. Kralj, R. G. Lewin, H. Kinoshita, I. May, and
C. A. Sharrad. “In situ spectroscopy and spectroelectrochemistry of uranium in high-temperature
alkali chloride molten salts.” Inorg. Chem. 47(17), 7474–7482, 2008.
31. T. Fujii, A. Uehara, T. Nagai, T. J. Kim, N. Sato, Y. Sakamura, and H. Yamana. “Electronic
absorption spectra of U(IV) in molten LiCl-SrCl2 and CsCl-SrCl2.” Electrochemistry 77(8), 667–669,
2009.
32. I. B. Polovov, C. A. Sharrad, L. May, V. A. Volkovich, and B. D. Vasin. “Spectroelectrochemical
study of neptunium in molten LiCl-KCl eutectic.” Z Naturforsch A 62(12), 745–748, 2007.
33. D. H. Kim, T. H. Park, S. E. Bae, N. Lee, J. Y. Kim, Y. H. Cho, J. W. Yeon, and K. Song.
“Electrochemical preparation and spectroelectrochemical study of neptunium chloride complexes in
LiCl-KCl eutectic melts.” J. Radioanal. Nuclear Chem. 308(1), 31–36, 2016.
34. J. L. Swanson. “Plutonyl species in molten chloride salt solutions.” J. Phys. Chem. 68(2), 438–439,
1964.
35. T. Nagai, A. Uehara, T. Fujii, O. Shirai, M. Myochin, and H. Yamana. “Redox equilibria of Pu4+/Pu3+
and PuO22+/Pu4+ couples in molten NaCl-CsCl eutectic as measured by absorption
spectrophotometry.” Radiochim. Acta 97(4–5), 209–212, 2009.
36. O. Shirai, A. Uehara, T. Fujii, H. Yamana, T. Nagai, and M. Myochin. “Absorbance of plutonium in
NaCl-2CsCl molten salt mixtures,” in Proceedings of the 4th Workshop on Molten Salts Technology
and Computer Simulation, Tokai, Japan, December 20, 2004, Japan Atomic Energy Research
Institute, 2005.
37. J. S. Zhang. “Electrochemistry of actinides and fission products in molten salts: Data review.”
J. Nucl. Mater. 447(1–3), 271–284, 2014.
38. C. A. Schroll, A. M. Lines, W. R. Heineman, and S. A. Bryan. “Absorption spectroscopy for the
quantitative prediction of lanthanide concentrations in the 3LiCl-2CsCl eutectic at 723 K.” Anal.
Methods—UK 8(43), 7731–7738, 2016.
39. C. A. Schroll, S. Chatterjee, T. G. Levitskaia, W. R. Heineman, and S. A. Bryan. “Electrochemistry
and spectroelectrochemistry of europium(III) chloride in 3LiCl-2KCl from 643 to 1123 K.” Anal.
Chem. 85(20), 9924–9931, 2013.
40. C. A. Schroll, S. Chatterjee, T. Levitskaia, W. R. Heineman, and S. A. Bryan.
“Spectroelectrochemistry of EuCl3 in four molten salt eutectics 3LiCl-NaCl, 3LiCl-2KCl, LiCl-RbCl,
and 3LiCl-2CsCl at 873 K.” Electroanalysis 28(9), 2158–2165, 2016.
41. S. Dai, L. M. Toth, G. D. DelCul, and D. H. Metcalf. “Solubilities of uranium(IV) dioxide in
magnesium chloride, calcium chloride, and aluminum chloride melts: A comparative study.” J. Phys.
Chem. 100(1), 220–223, 1996.
42. J. A. Duffy and M. D. Ingram. “Optical basicity. 4. Influence of electronegativity on Lewis basicity
and solvent properties of molten oxyanion salts and glasses.” J. Inorg. Nuclear Chem. 37(5), 1203–
1206, 1975.
43. J. A. Duffy, and M. D. Ingram. “Establishment of an optical scale for Lewis basicity in inorganic
oxyacids, molten salts, and glasses.” J. Am. Chem. Soc. 93(24), 6448–6454, 1971.

FRD 2 35
 Molten Salt Chemistry Workshop

4. FRD 3: Understanding Fission and Activation Product Chemistry

and Radiation Chemistry
4.1 Background and Current Status

Understanding the fate and transport of fission and activation products in MSRs is key to monitoring,
predicting, and controlling reactor performance and salt properties (Figure 18). Although the decay
pathways of fission and activation products are well established and independent of the chemical
environment, the chemical and physical properties of these elements and their interactions with their
surroundings are different at each step of the decay chain; they are complicated by the fact these elements
exhibit distinct chemical behaviors in different salt mixtures and behave differently in different short-
decay regimes. The general understanding is far from complete and remains a major scientific and
technology challenge associated with the MSR system. It is essential to understand which fission
products remain dissolved in the salt; which ones are volatile; and which ones precipitate on surfaces
(and where), which ones are sometimes soluble, and how the fission products change the corrosivity of
the melt. Furthermore, determination of fission product interactions and solubility is necessary for each
unique salt system (e.g., fluoride salts vs. chloride salts vs. new salts). To achieve this level of
understanding requires knowledge regarding the chemical nature of the fission and activation products—
i.e., the chemical species that exist under particular conditions of salt composition, temperature, and redox
state—and how they may react or distribute on surfaces in a dynamic environment.

Figure 18. Complex behavior of molten salt and insoluble and volatile fission products. The metal strip
was in the MSRE pump bowl gas space; it shows evidence of mist in the gas space and noble metal
fission products. | Images courtesy of Oak Ridge National Laboratory

Although fresh fuel salt mixtures have been studied, understanding of transport phenomena, effects of
short-term decay, neutron activation, and interfacial interactions are necessary. Individual fuel atoms in
the salt undergo fission (see Figure 19), splitting into a few of 40+ elements representing more than 400
different isotopes, while neutron capture generates various transuranic and other activation products (such
as 36Cl). Fission liberates large amounts of energy, causing various local disruptions such as ionization or
material sputtering. The individual fission product atoms will be highly ionized and will quickly interact
with nearby atoms, reaching one of several possible end states dependent on their chemistry and the
atoms with which they interact. The interactions with the salt, structural materials, or other fission and

FRD 3 37
Molten Salt Chemistry Workshop

activation products, causes complex speciation. The process is repeated with many variant outcomes for
each fission product. The end result can be macroscopically understood based on whether the fission
products are soluble, insoluble, volatile, or corrosive. Each of these have implications for the reactor and
the physical properties of the salt. On the mesoscale, interactions will be highly dependent on interfacial
chemistries between fission products and the materials present. Understanding this and translating it to the
macroscale effects, such as corrosion, is necessary.

Figure 19. Yields of different elements from fission of 235U. | Image courtesy of Peter Lux. In “Composition of Spent
Fuel,” Peter Lux Blog, http://www.plux.co.uk/. Licensed under a Creative Commons license Attribution-NonCommercial-
ShareAlike 3.0 unported license

At some time after the fission event, the isotopes of the fission products reach chemical equilibrium and
behave according to their elemental identities, eventually forming stable species. Soluble species are
incorporated within the salt solution and often behave according to their elemental periodic group;
insoluble species likely deposit on surfaces, such as the heat exchanger, or at the interface between wetted
and non-wetted surfaces; volatile species are driven to the off-gas system; and corrosive species may
degrade structural and container materials (Figure 18). To understand the impact of fission products on
reactor operation, an understanding the basic physics and chemistry at the atomic, molecular, meso-scale,
and continuum length scales, as well as phenomena that occur over a wide range of time scales is
required. With full knowledge of the fission product decay pathways and their chemistry, we should be
able to design systems that mitigate detrimental effects of fission products with regard to reactor
performance and disposition in waste streams.

4.2 Technical Challenges and Research Directions

4.2.1 Understanding the Physical, Chemical, and Radiochemical Properties of the

Fission and Activation Products in Molten Salts
As previously discussed, there is a need to understand the chemical bonding and coordination
environment (and associated energies) between the solutes (actinides, fission products, and activation
products) and the molten salt solvent, since interactions could form out to a length scale that includes

38 FRD 3
 Molten Salt Chemistry Workshop

many atoms (in fact, determining the number of atoms that must be included is one of the underlying
uncertainties.) Spectroscopic studies (described in Section 3) can provide insights into the coordination
environment of the elements, but computational insights would be particularly useful for the more
radioactive and short-lived species that are hard to study experimentally. Unfortunately, little theoretical
work has been reported on the solvation of individual fission products. Although various computer codes
can be used to obtain this information (e.g., the methods described for understanding the solvation of Gd3+
dissolved in LiCl and KCl), [1] the more atoms that are involved, the greater the computational expense.
Therefore, appropriate assumptions for simplifying the models will need to be developed and validated. In
moving from single atoms to complex fission product species, more solute-solvent configurations must be
included, as well as an understanding of the internal bonds within the fission product complex. To solve
the problem satisfactorily will require combining the results from computer codes that handle atomic- and
molecular-scale interactions and properties all the way up to bulk material properties and energetics
associated with changes in the physical states of transmuting isotopes.

To understand the binding, solvation, and properties of fission and activation products, a representative
set of fission products (several from each of the possible end states, e.g., soluble, insoluble, sometimes
soluble, and volatile) that can be both modeled and measured in the laboratory (potentially though the use
of nonradioactive surrogates) needs to be investigated. A comparison between experimental and computer
models is extremely important, since only validated computational models can to trusted to provide
insights into a broader range of fission products and yield meaningful results. This means that the systems
that are studied must also possess unique, reliable experimental signatures that extend from single atoms
to more complex molecular species.

In addition to out-of-pile chemical studies, in the future, it will be necessary to develop experimental
systems with in situ monitoring for the irradiation of salts to probe dynamic fission product chemistry
effects and their impact on the physical properties of the molten salt (and reactor operation). With this
understanding, we can begin to design systems to mitigate any negative effects associated with fission
product transport, localization, or specific radiation/interfacial chemistry. For example, one importance
result from in-pile studies in the MSRE was the discovery that the fission product tellurium contributed to
surface cracking in the alloy Hastelloy N. Tellurium is unusual in that it has a high vapor pressure at the
operating temperature of the MSRE, so it can survive as both a dissolved gas and (possibly) as a soluble,
reactive ion. [2] To determine the cause of the observed surface cracking, a series of temperature-
controlled tests were conducted both to confirm that tellurium was contributing and to subsequently learn
how to mitigate its degradation effects. This study showed how a particular fission product’s behavior
changed with salt redox potential, how it reacted at the interface, and how the negative effect could be
mitigated in future designs.

The tellurium study is an example of what needs to be done in greater depth for a range of “salt-
insoluble” fission products that reside in the middle of the periodic chart. Although the understanding of
tellurium is the most advanced finding for these fission products, it is not yet complete, especially for the
new proposed salt compositions. An experiment design that accomplishes these goals with dynamic
online measurement capability, as distinct from static sample analysis, will greatly accelerate the
understanding of fission product behavior.

Validating the ex situ surrogate testing requires high radiation fields, high-temperature operation, and
realistic production and decay of fission products. In preparation for operation of the MSRE in the 1960s,
a series of experiments were conducted at the Materials Test Reactor in Idaho and the Low Intensity Test
Reactor and the Oak Ridge Research Reactor at ORNL (Figure 20). A similar set of experiments is
needed for next-generation molten salts for the MSR to address the concerns discussed above.

FRD 3 39
Molten Salt Chemistry Workshop

Figure 20. Natural circulation loop used in the Oak Ridge Research Reactor. The experiment tested both
7LiF-BeF2-ZrF4 and 7LiF-BeF2-ZrF4-UF4 salts. | Image courtesy of Oak Ridge National Laboratory

4.2.2 Correlating Fission and Activation Product Behavior with Surrogates

Before establishing the ability to install a molten salt test loop within a reactor with useful emulation of a
range of MSR environments, the path for learning could proceed in a limited way with out-of-pile studies
using both nonradioactive and radioactive surrogates to understand the behavior of fission and activation
products. Whereas in situ studies will be very complex and will require the development of advanced
analytical techniques, ex situ studies can be started immediately. These studies can be carried out with
both nonradiological surrogates and the actual fission products. Ex situ studies can involve capsule testing
of materials prior to thermal convection loop experiments to understand the influence of the fission
products on reactor materials. The anticipated radiological fission products include elements such as I,
Xe, Cs, Cd, In, Sn, Sb, Te, Zr, Nb, Mo, Tc, Ra, and Tl (see Figure 19). The stable isotopic forms for most
of these elements are routinely used as surrogates in adsorbent and separations studies. Therefore,
laboratory studies can be performed in a variety of research settings to gain insights into fission product
behavior without requiring access to hazardous radioactive materials.

There are numerous published examples of successful studies with surrogate/stable isotopes. For example,
Nenoff et. al, have used stable iodine gas (I2 and organic iodides such as CH3I) to study the fundamental
materials chemistry of iodine fission gas capture and storage materials. Structure property relationships
between nanoporous materials such as zeolites and metal-organic frameworks (MOFs) and the iodine
species have been detailed (Figure 21). [3] Subsequent studies indicate there is no degradation of the

40 FRD 3
 Molten Salt Chemistry Workshop

framework structures due to gamma irradiation. Furthermore, stable isotope studies of cesium and
strontium ion capture from aqueous and seawater solutions were performed to confirm the stability and
selectivity of a crystalline silicotitanate framework before commercialization and use at the Fukushima
Daiichi nuclear reactor accident site. [4]

Figure 21. Metal-organic frameworks for capture of fission products (left): A schematic of iodine capture
by silver-containing mordenite (MOR). Whereas pre-reducing the silver MOR yields a mixture of γ/β-AgI
nanoparticles and subnanometer α-AgI, direct iodine uptake by silver-exchanged MOR produces
exclusively subnanometer α-AgI. | Reprinted with permission from K. W. Chapman, P. J. Chupas, and T. M.;
Nenoff. “Radioactive iodine capture in silver-containing mordenites through nanoscale silver iodide formation.”
J. Amer. Chem. Soc. 132 (26), 8897–8899, 2010. Copyright 2010, American Chemical Society. Right: Captured I 2
crystallographically located inside the small pored ZIF-8 MOF | Reprinted with permission from D. F. Sava, M. A.
Rodriguez, K. W. Chapman, P. J. Chupas, J. A. Greathouse, P. S. Crozier, and T. M. Nenoff. “Capture of volatile iodine,
a gaseous fission product, by zeolitic imidazolate framework-8.” J. Amer. Chem. Soc. 133(32), 12398–12401, 2011.
Copyright 2011, American Chemical Society.

Some fission and activation products are commonly substituted with surrogate elements for laboratory
studies, for example, cesium for plutonium and europium for americium. Requirements include
comparable atomic size, redox states, and chemical reactivity. For MSRs, the substitution of rhenium for
technetium is particularly important for investigation of its solution chemistry and volatilization
behaviors. [5] It is important to note that in one set of glass waste form degradation studies, rhenium
transport was reported to behave quite differently from technetium transport under hydrothermal
conditions. [6] Experiments reacting Re, ReO2, and TcO2 with Cl2 and HCl in molten chloride eutectics
also revealed different chemical products, with a volatile TcOxCly proposed. [7] Therefore, control
experiments need to be performed to ensure surrogate elements possess representative chemical and
physical properties compared with the fission product generated under MSR operational conditions.
Radiological studies, either with appropriate radiological isotopes or with the introduction of the
appropriate decay energy, are needed to confirm the experimental results obtained with the stable or
surrogate isotopes. For example, the electron beam from a TEM can be used to irradiate surrogate fission
products [8] and the resulting reactions and redox responses can be studied. [9]
As noted earlier, modeling the effects of radiological degradation of the salt or containment metals will
provide a predictive understanding of the behavior of radiological fission products in the molten salt
operational environment. It will also provide a computational means of near-term investigation of surrogate
behavior compared with behavior of the actual radionuclide. These computational findings will rapidly
provide information, which will necessarily be validated or revised upon completion of the more involved
and complicated studies of radioactive materials following in the course of MSR development. [10]

FRD 3 41
Molten Salt Chemistry Workshop

4.2.3 Understanding the Physical and Chemical Impact of Short-lived Isotopes

Short-lived fission products in the salt are of great concern because of the chemical differences between
the original elements and the daughter products. The effects from short-lived fission products are complex
and special experimental methods are required to probe them as they undergo many iterations and
speciations until becoming stable isotopes. For instance, 137I (t1/2 = 24 sec) will be soluble in the salt;
however, upon decay it becomes the entrained gas 137Xe (t1/2 = 3.81 min), which can further decay to 137Cs
(t1/2 = 30 year) (Figure 22). If the system is flowing rapidly, the xenon will be removed from the salt (if
not entrained) and will decay to cesium. Cesium from entrained 137Xe will redissolve into the halide salt
and will stay soluble based on its free energy as a fluoride or chloride. Alternatively, cesium from the
liberated 137Xe, removed during circulation, will result in cesium plating within the outgassing system.
Another example is the 99Zr fission product (t1/2 = 2 sec), which will decay to 99Nb (t1/2 = 15 sec), then to
99
Mo (t1/2 = 2.75 days) and eventually to 99Tc (t1/2 = 2.1×105 year), which will stay in the metal state and
not be soluble in the molten salt; thus it will plate out on the reactor components or heat exchanger.
Cadmium- 131,-132 (t1/2 = 68 msec, 97 msec) provides a third example; it decays to 131In (t1/2 = 280
msec), 131Sn (t1/2 = 56 sec), 131Sb (t1/2 = 23 min), and 131Te (t1/2 = 25 min), and finally to 131I (t1/2 = 8 days).
Iodine-131 decays to 131Xe gas, which is stable and must be removed from the melt to minimize neutron
poisoning in the reactor. These phenomena are depicted in Figure 22. Although individual elements can
be experimentally studied with surrogates, it is difficult to study the dynamics and changing chemistries
of short-lived fission products. Thus, computational methods need to be developed in tandem with the
best experimental methods to provide insights into the impacts of short-lived isotopes on the reactor
chemistry and corrosion.

Figure 22. Examples of short-lived fission products that change solubility behavior during decay. | Image
courtesy of Oak Ridge National Laboratory

42 FRD 3
 Molten Salt Chemistry Workshop

References

1. S. Hazebroucq, G. S. Picard, and C. Adamo. “A theoretical investigation of gadolinium (III) solvation

in molten salts.” J. Chem. Phys. 122, 224512, 2005; S. Hazebroucq, G. S. Picard, C. Adamo, T.
Heine, S. Gemming, and G. Seifert. “Density-functional-based molecular-dynamics simulations of
molten salts.” J. Chem. Phys. 123, 134510, 2005.
2. E. L. Compere, S. S. Kirslis, E. G. Bohlmann, F. F. Blankenship, and W. R. Grimes. Fission Product
Behavior in the Molten Salt Reactor Experiment, ORNL-4865, Oak Ridge National Laboratory, 1975.
3. T. M. Nenoff, M. A. Rodriguez, N. R. Soelberg, and K. W. Chapman. “Silver-mordenite for
radiologic gas capture from complex streams: Dual catalytic CH3I decomposition and I confinement.”
Micro. Meso. Mater. 200, 297–303, 2014; K. W. Chapman, P. J. Chupas, and T. M. Nenoff.
“Radioactive iodine capture in silver-containing mordenites through nanoscale silver iodide
formation.” J. Am. Chem. Soc. 132(26), 8897–8899, 2010; D. F. Sava, K. W. Chapman, M. A.
Rodriguez, J. A. Greathouse, P. S. Crozier, H. Zhao, P. J.Chupas, and T. M. Nenoff. “Competitive I2
sorption by Cu-BTC from humid gas streams.” Chem. Mater. 25(13), 2591–2596, 2013; D.F.S.
Gallis, I. Ermanoski, J. A. Greathouse, K. W. Chapman, and T. M. Nenoff. “Iodine gas adsportion in
nanoporous materials: A combined experiment–modeling study.” Ind. Eng. Chem. Res. 56(8), 2331–
2338, 2017; J. T. Hughes, D. F. Dava, T. M. Nenoff, and A. Navrotsky. “Thermochemical evidence
for strong iodine chemisorption by ZIF-8.” J. Am. Chem. Soc. 135(44), 16256–16259, 2013.
4. T. M. Nenoff and J. L. Krumhansl. “Cs+ removal from seawater by commercially available molecular
sieves.” Solvent Extraction and Ion Exchange 30, 33–40, 2012; C. A. Abney, K. M. L. Taylor-
Pashow, S. R. Russell, Y. Chen, R. Samantaray, J. V. Lockard, and W. Lin. “Topotactic
transformations of metal-organic frameworks to highly porous and stable inorganic sorbents for
efficient radionuclide sequestration.” Chem. Mater. 26, 5231–5243, 2014.
5. M. H. Langowski, J. G. Darab, and P. A. Smith. Volatility Literature of Chlorine, Iodine, Cesium,
Strontium, Technetium, and Rhenium: Technetium and Rhenium Volatility Testing, PNNL-11052,
Pacific Northwest National Laboratory, 1996.
6. D. A. McKewon, A. C. Buechele, W. W. Likens, D. K. Shuh, and I. L. Pegg. “Tc and Re behavior in
borosilicate waste glass vapor hydration tests.” Environ. Sci Technol. 41(2), 431–436, 2007.
7. V. A. Vokovich, I. May, J. M. Charnock, and B. Lewin. “Reactions and speciation of technetium and
rhenium in chloride melts: A spectroscopy study.” Phys. Chem. Chem. Phys. 4, 5753–5760, 2002.
8. M. Nyman, B. X. Gu, L. M. Wang, R. C. Ewing, and T. M. Tenoff. “Synthesis and characterization of
a new microporous cesium silicotitanate (SNL-B) molecular sieve.” Micro. Meso. Mater. 40, 115–
125, 2000.
9. C. L. Tracy, M. Lang, J. M. Pray, F. Zhang, D. Popov, C. Park, C. Trautmann, M. Bender, D.
Severin, V. A. Skuratov, and R. C. Ewing. “Redox response of actinides materials to highly ionizing
radiation.” Nature Comm. 6, 6133, 2015.
10. A. Dunn, R. Dingreville, and L. Capolungo. “Multi-scale simulating radiation damage accumulation
and subsequent hardening in neutron-irradiated α-Fe.” Model. Simul. Mater. Sci. Eng. 24(1), 015005,
2015; R. Dingreville and S. Berbenni. “On the interaction of solutes with grain boundaries.” Acta
Mater. 104, 237–249, 2016.

FRD 3 43
 Molten Salt Chemistry Workshop

5. FRD 4: Understanding Materials Compatibility and Interfacial

Phenomena
5.1 Background and Current Status

The reemergence of MSR concepts for future energy production has prompted an evaluation of the state
of knowledge and fit-for-use (or lack of such) of materials in contact with the coolant/fuel. Materials
compatibility under such extreme conditions presents unique challenges, many of which have been
defined and addressed, at least for fluoride salts, during the course of the MSRE and subsequent MSBR
program. [1] However, concerns remain regarding the chemical compatibility and application of Ni-Mo
alloys (specifically Hastelloy N) in molten fluoride salt because of the lack of long-term operational
experience and the possibility of embrittlement by fission products. As an expanded performance
envelope has been proposed in some recent MSR designs, including higher temperatures and use of
chloride salts, it is apparent that alternative structural materials need to be evaluated. Chloride salts for
MSR applications have a much smaller compatibility database, require higher temperatures to liquefy,
and appear to display greater corrosion issues, although results show a great degree of variability.
Candidate nickel-based alloys and/or advanced ceramics, such SiC-SiC composites, encompass greater
alloy strength (without significant degradation of fracture toughness), radiation tolerance, and resistance
to corrosion and oxidation than the alloys developed during the MSRE.

Molten salt coolants and fuels present a number of extreme environmental demands on reactor structural
and functional materials. These include

• Resistance to corrosion and embrittlement by the fuel/coolant salt

• Low rates of mass transfer burden due to thermal or dissimilar material gradients in the reactor
system
• High strength and fracture toughness with minimal/acceptable degradation over time due to radiation
damage and high temperatures
• Corrosion/oxidation resistance by the external reactor environment (whether air or a secondary
coolant)
• Fabricability, including joining and cladding (for certain applications)

In general, materials degradation processes underlying these requirements are approached as either
irradiation or chemical compatibility effects. Significantly less attention is paid to their combined
influence or the origins of those effects at the atomistic scale. Similarly, although mechanical and
chemical behaviors tend to be treated separately, their effects can be synergistically compounding;
relevant examples include stress corrosion cracking in aqueous systems and oxidative embrittlement of
nickel-based alloys. To some extent, the problems can be separated by using corrosion-resistant coatings
on materials to provide adequate mechanical strength, although this practice introduces the additional
issue of coating integrity. The manifestations of these effects can be different for alloys, graphite, and
silicon carbide composites.

Materials degradation (corrosion) occurs at an interface between the salt and a metal in a MSR. Thus,
understanding the molecular-level chemistry of the fluid, solid, and gas interfaces is key to controlling the
interfacial reactions that result in materials degradation. Interfacial behavior pertinent to MSRs is an
under-investigated area—in particular at atomistic scales—because the boundary regions between phases
or molecular regions comprise only a small fraction of all species in any sample, making their focused
investigation particularly difficult. For MSRs, a molecular-level understanding is needed of the salt/alloy,

FRD 4 45
Molten Salt Chemistry Workshop

salt/graphite, salt/gas, and salt/salt interfaces; and an understanding of their molecular structure and
chemical reactivity will provide important insights into mitigating phenomena such as corrosion, fission
product off-gassing, gas entrainment, and material precipitation.

5.2 Technical Challenges and Research Directions

5.2.1 Advancing Spectroscopic and Scattering Investigation of MSR Interfaces

The ability to characterize the molecular-level chemistry that takes place at the liquid/solid interface is
key to understanding interfacial reactions such as corrosion. A number of surface-sensitive techniques
have been developed in the 20th century to address the challenges of characterizing an interface,
including infrared (IR) adsorption spectroscopy (attenuated total reflectance, reflection-adsorption
infrared spectroscopy, transmission IR), Raman spectroscopy, sum frequency generation, UV-vis
spectroscopy, and x-ray and neutron-based techniques, such as reflectometry. [2] Second harmonic
generation, which was used to investigate molecular monolayers adsorbed to surfaces, [3] has been
extended to study the air-water interface [4] and more recently the interfaces of spherical particles [5] and
colloidal systems. [6–8] Recent developments in quantum-beam approaches have also enabled the
investigation of interfacial structure and reactivity. Perhaps of greatest note are x-ray and neutron
reflectivity, which provide a molecular density profile of molecular ordering as a function of distance
from the interface. [9] For example, a review of the water/mineral interface reveals perturbations of the
liquid structure near the solid surface boundary. [10] This liquid structure is expected to have a crucial
influence on physical interfacial behaviors—such as diffusion, adsorption, and dielectric properties—that
directly affect phenomena of immediate pertinence, such as metal precipitation or alloy corrosion.

Numerous grazing incidence techniques have also been developed for the investigation of various
interfaces. For example, grazing incidence XAFS (GI-XAFS) has been used to investigate the speciation
of Er3+ at the electrolyte-vapor interface; it revealed simple hydration of Er3+ in the bulk phase to form
[Er(OH2)8]3+, but formation of a neutral [(H2O)6-7ErCl3] at the surface (Figure 23). [11] This work was
complementary to an investigation of the same system using x-ray reflectivity. [12] Grazing incidence x-
ray diffraction provides information regarding in-plane structure and ordering at interfaces, and x-ray
fluorescence near total reflection provides element-specific information regarding the binding and
coverage of target metal ions.

There is a need to characterize the molecular-level chemistry at the molten salt/solid interface. The use of
approaches like those described to investigate interfaces within a molten salt would enable the
identification of metal species and oxidation states in the salt phase. There is also great potential to
answer unresolved questions regarding corrosion mechanisms in MSRs, in particular via the development
of hyphenated techniques in which manipulation of interfacial behavior is achieved electrochemically.
Corrosion studies are commonly performed by applying a voltage to a metal surface or creating a
chemically reactive environment and observing how metal components react. [13–15] Electrochemical
impedance spectroscopy is routinely used to explore corrosion of a metal surface, because it provides
information regarding solution phase resistance, charge transfer properties of the surface, and diffusion
characteristics of the solution phase. Traditional corrosion analysis is performed using post-mortem
methods—such as transmission electron microscopy with electron energy loss spectroscopy or energy
dispersive spectroscopy, x-ray diffraction, and optical spectroscopy—in which a sample is analyzed after
retrieval from a reaction vessel. While informative, this approach is limited by challenges associated with
discerning the chemistry after it has occurred. In contrast, the spectroscopic approaches discussed above
allow the direct interrogation of the interface itself and, in some instances, are even amenable to in situ
reaction monitoring.

46 FRD 4
 Molten Salt Chemistry Workshop

Figure 23. GI-XAFS results reveal unanticipated molecular complex formation at the water/gas interface.
The extension of similar techniques to molten salts could be used to answer fundamental questions
regarding corrosion mechanisms, particularly if successfully paired with electrochemical control. | Reprinted
with permission from M. K. Bera, G. Luo, M. L. Schlossman, S. Lee, and M. R. Antonio. “Erbium(III) coordination at the
surface of an aqueous electrolyte.” J. Phys. Chem. B 119(28), 8734–8745, 2015. Copyright 2015, American Chemical
Society.

Development of multi-faceted in situ and in operando approaches to observe interfacial physical

phenomena is needed to advance MSR development. The advancement of corrosion science is greatly
enhanced by the ability to collect information as the chemistry occurs. For example, a combination of
electrochemical methods with advanced x-ray and neutron scattering techniques would allow the
exploration of corrosion in operando, providing process-relevant information that cannot be observed
during traditional post-mortem analysis.

5.2.2 Understanding Materials Compatibility in the Molten Salts Reactor Environment

Results from the Aircraft Nuclear Propulsion program in the 1950s and the MSRE in the 1960s clearly
showed the benefits of the Ni-Mo alloy system for use with molten fluoride salts. Most existing high-
temperature nickel-based superalloys were designed for operation in oxidizing external atmospheres (e.g.,
air in the MSRE radiator and secondary coolant piping/tanks, and supercritical steam in the MSBR
secondary heat exchanger). Therefore, their corrosion resistance is derived from the formation of a
passivating layer of chromia or alumina scale. The situation with regard to salt-side corrosion is
fundamentally different because of the dissolution of passivating oxide layers in fluoride and chloride
salts. Corrosion in these salts is instead controlled by the thermodynamic stability of the bare alloy
surface in the halide salt environment. Degradation occurs by dissolution of the containment material into
the salt, driven by the difference in free energy of formation between the salt constituents and the most
susceptible transition metal corrosion product (the more negative the free energy of reaction, the greater
extent to which the reaction will occur). In the simplest systems, these reactions are governed by
elemental solubility considerations and the kinetics of the rate-controlling step of the dissolution process.
Those dependencies are reflected in the fundamental rate equation for solvation of solids in liquids [16]:

FRD 4 47
Molten Salt Chemistry Workshop

Ji = ks,i(Co,i – C) , (5.1)

where Ji is the rate of mass loss of element i due to dissolution, Co,i is the solubility of component i in the
salt, C is the actual concentration of i in the bulk of the liquid, and ks,i is the effective solution rate
constant. (Each constituent of the condensed phase is separately governed by Eq. (5.1) but, for
simplification, the subscripts are not used hereinafter.) Because J is proportional to (Co – C), high
solubilities do not necessarily translate into high dissolution rates (and vice-versa), because the driving
force for dissolution depends on the actual concentration of the solute in the salt as well as Co. [17, 18]
Elimination of concentration gradients by convective or mechanical mixing thus forms the basis for the
difference between isothermal and/or stagnant salt systems and nonisothermal dynamic ones (such as in
most reactor schemes) with regard to material stability. This can be particularly important for heat
exchangers and systems with large temperature differentials in which deposition can block flow passages
and lead to thermal excursions or excessive localized radioactivity. Furthermore, preferential dissolution
of one constituent into the salt can destabilize the structure of the solid or its interface with the liquid, in
some cases, leading to porous zones.

Under nonisothermal conditions normally encountered in reactor systems, [19] C should be uniform
within a flowing liquid system. Consequently, J is only dependent on time (t) through ks because (Co – C)
is effectively constant. Common practice then allows J to be described in a generalized kinetic form:

J = m/tn . (5.2)

The time dependence of the dissolution rate process (n) can be determined by fitting measured mass loss
(m) using Eqs. (5.1) and (5.2). Therefore, direct measurement of mass loss from corrosion affords a
straightforward method of determining a key kinetic parameter for dissolution, yielding information
regarding rate-controlling steps (surface reaction, liquid-phase diffusion through boundary layer, and
solid-state diffusion) [16, 17, 20, 21] for corrosion mechanisms under nonisothermal conditions.
Collection of these data at various temperatures affords the representation of different positions in a
molten salt loop. Such kinetic information can guide and validate appropriate modeling and lifetime
prediction approaches for a particular solid-salt system.

The flux description described by Eq. (5.1) applies equally well to deposition when C > Co. If the
dissolution and deposition behaviors are controlled by the respective forms of Eq. (5.1), measurements of
mass change as a function of loop position (x) should yield data resembling the schematic depiction
shown in Figure 24. Using the appropriate analysis methods and simplifying assumptions, measurements
of the mass—or, for alloys and compounds, composition change—as a function of loop position (and
therefore temperature) can provide further data regarding kinetics and the rate controlling mechanism of
mass transport associated with a given material–liquid salt system. Initial results could be obtained in ex
situ studies to guide future experimental investigations.

The free energy driving force for dissolution of the solid can be controlled by manipulating the redox
potential of the salt. While more reducing conditions were found to suppress corrosion, other reactions
between the salt and solid, such as alloying, become more likely. [22] Therefore, corrosion control by
manipulation of the redox potential is bound by the necessity of identifying conditions that minimize both
oxidative and reductive degradation reactions; limiting other deleterious reactions; and, for fuel salts,
maintaining the necessary U3+/U4+ ratio. The ability to effectively control the salt redox properties in real
time requires implementation of online monitoring (discussed in Section 3).

48 FRD 4
 Molten Salt Chemistry Workshop

(b) Idealized profile of mass change per unit

surface area for a liquid circuit. Tmax, Tmin are
maximum, minimum loop temperatures,
(a) Mass transfer schematic respectively; x is loop position.
Figure 24. Steps and balance points in temperature-gradient mass transfer. | Image courtesy of Oak Ridge
National Laboratory

Impurities in the as-produced salts, such as HF, HCl, H2O; residual metal oxides; and/or polyvalent metal
ions can also cause or affect degradation. They can change the driving forces (free energies) for
compound formation/alloying of the solid, provide new reaction pathways for dissolution and deposition,
and affect kinetic factors governing mass transfer of the principal products. In closed systems with
purified salts, these impurity effects should be limited and insignificant (after an initial short time period)
compared with the persistent, continuous corrosion mechanisms.

As discussed in Section 4, fission products that develop in the salt during reactor operation can also
impact materials performance. Tellurium, for example, significantly impacts the mechanical performance
of Hastelloy N in fluoride salts by promoting cracking via leaching of chromium from the alloy.
However, the cracking can be reduced by controlling the redox potential of the salt through variation of
the U3+/U4+ ratio, as shown in Figure 25. [23] This approach will also affect the dissolution rate of the
alloy constituents. A proven alternative approach was using metal additions to form metal alloys, for
example, adding niobium reduced the propensity for this type of embrittlement. [23]

A rigorous predicative capability for corrosion and environmental effects in the molten salt operating
environment is currently lacking. Although this gap is not unique with respect to MSR systems, the
technology would certainly profit from research to advance the fundamental understanding of corrosion
processes at the atomistic level, as well as provide the necessary underpinnings and experimental data for
accurate modeling and simulation of materials behavior. Such developments would facilitate the
identification and remediation of design issues while quantifying lifetime prediction for MSR concepts. A
focus on predictive capabilities would provide the framework to conduct a coordinated experimental
program in corrosion, mass transfer, physical and mechanical properties, and radiation effects. These
would directly impact and inform the modeling effort and provide the means for validation of models and
the simulations built on them. Such efforts need to address

• Reaction pathways and kinetics

• Prediction of failure mode and lifetime

FRD 4 49
Molten Salt Chemistry Workshop

• Materials, component, and structural design rules

• Mechanistic underpinnings for input into modeling and simulation modules
• Coupling of radiation and mechanical effects to corrosion and mass transfer phenomena

Figure 25. Plot showing the effect of the salt oxidation potential (as measured by U4+/U3+) on the
propensity for tellurium embrittlement of Hastelloy N. | Image courtesy of Oak Ridge National Laboratory

5.2.3 Understanding Degradation Processes at the Material–Salt Interface

To understand the degradation of salt-facing structural materials, it is critical that the corrosion
mechanisms at the material–salt interface be well understood. However, to date, studies have mostly
relied on macroscopic ex situ measurements in static conditions, leaving a knowledge gap in our
understanding of the nature of the material–salt interface and the kinetic parameters that determine its
evolution with time. This knowledge gap needs to be addressed by designing, constructing and
establishing techniques and methodologies for in situ characterization of the material–salt interface (see
Section 5.2.1). In situ spectroscopy/diffraction analysis of material surfaces in the molten salt under both
static and ex situ flow loop testing will greatly advance the understanding of relevant corrosion
mechanisms.

To aid in the understanding of interface behavior under realistic conditions, flow loops should be
constructed for the integration of thermal gradients, fretting, and erosion into the mechanism studies (see
Figure 20). In-core management and monitoring of corrosion is an area of research requiring much more
investment, and flow loop testing is the ideal first step in investigating potential in situ monitoring
methods and testing new materials. Flow loops will also allow for the effects of joining different
structural materials together to enable a clearer understanding of the overall integrity of the system. Use
of in situ experimental techniques to study the evolution of the material–molten salt interfacial chemistry
is critical to understanding and predicting the degradation behavior and lifetime of the material of interest.

50 FRD 4
 Molten Salt Chemistry Workshop

Characterizing any surface deposits that are formed in the reactor as a result of exposure to molten salts
and fission products, and identifying and quantifying the elements that are preferentially leached out of
the material, will lead to an understanding of the chemical interactions and reaction kinetics between the
molten salt and the material. Further, in situ material surface analysis can be combined with in situ salt
studies using techniques such as microfluidic sampling and inductively coupled plasma optical emission
spectroscopy.

To accurately model material degradation in molten salt, accurate inputs are required; and without a solid
base of experimental data, a modeling effort would be limited in usefulness. For this purpose, a
coordinated effort is needed to gain fundamental experimental data on simple model systems, in which
parameters can be accurately controlled and questions or hypotheses can be answered, coupled to more
complex experiments in a test loop. Greater emphasis must be placed on the use of controlled, science-
based experiments to find thermodynamic parameters, diffusion rates, and kinetic dissolution parameters
that control corrosion and degradation of materials at the molten salt interface. Targeted experiments
may include capsule testing, flow loops, calorimetry, or in situ techniques designed to measure the
dissolution of species in salt.

Using data specifically acquired to generate thermodynamic databases, thermodynamic models can be
constructed. While thermodynamic modeling is not necessarily predictive, it is an important step toward
more complex predictive models that account for corrosion, materials aging, and defect kinetics.
CALPHAD (Computer Coupling of Phase Diagrams and Thermochemistry) [24] codes have proven to be
a useful tool for making thermodynamic predictions and have already been used for calculating phase
equilibria in salts. However, the calculation of phase equilibria relevant to alloy–salt interactions requires
fundamental thermodynamic parameters, such as Gibbs free energies and solubilities, which can only be
obtained from experimental data (see Section 2).

Beyond thermodynamic modeling, true predictive capability will require the development of multiphysics
models that incorporate corrosion kinetics with radiation and salt chemistry. The purpose of these models
is to use the mechanistic understanding of materials behavior to forecast the behavior of new conceptual
materials in these extreme environments. This will allow for the development of materials with targeted
properties.

Computational methods need to be developed and coupled to experimental data to predict corrosion due to
thermal gradients, fission products, and dynamic chemistries in the MSR. Models need to be tightly coupled
to fluid flow, heat transport, and mass transport; and the production and dispersion of soluble gases, such as
xenon or tritium (from neutron-induced fission of 6Li), need to be understood. The exposure of the salt to an
inert atmosphere in the pump section can be modeled as a boundary condition allowing the entrainment of
gases (helium and/or argon) into the coolant. Understanding the distribution of tritium, when present, is
important for determining the optimal placement of tritium-getters and determining places in the coolant
loop where tritium might naturally deposit or evolve. To capture these effects, models need to be developed
to determine a sink of tritium through off-gas or other chemical reactions.

Understanding the interfacial chemistry will help predict material behavior and lifetime and accelerate the
deployment of MSRs. As an added benefit, the development of these in situ techniques for molten salts
will also help with real-time monitoring of the quality of the salts in the reactor and time-dependent
transfer and disposition of materials for accountability.

5.2.4 Understanding the Combined Effect of Chemistry and Radiation at the Interface
The generation of fission products and their subsequent decay is well documented and understood and
supports accurate predictions of their production in molten salt systems. This enables the prediction of

FRD 4 51
Molten Salt Chemistry Workshop

fission product concentrations with high accuracy. However, it does not explain the evolution of dissolved
fission product concentrations completely because of the dynamic effects associated with short-lived
fission products and the complex solubility behavior of many transition metals. Although previous studies
have never revealed any new chemical species due specifically to radiation-induced-reactions in molten
coolant or fuel salt, the possibility exists for new chemistry and compounds to be produced by the highly
energetic fissioning salt environment. Some of these could have detrimental effects (especially if the
reactor operates for 30 years).

The most dramatic impact of ionizing irradiation occurs during the self-irradiation (due to fission and
activation products) of solid salts at lower temperatures, with the production of metal centers and reactive
halogens (F2), and must be considered for the case of spent salt storage and capsule irradiation (see
Appendix A, Section 2.3.3). Based on the composition of the salt, the radiolysis effects of each solid salt will
be unique. In the case of mixed halides—including mixed halides involving fission products such as iodo-
fluorides and iodo-chlorides—interhalogens are very reactive, with different chemical interactions from those
of pure halogen salts. In particular, boron and iodine are hard to remove, and they each have long-term effects
in molten salts. High-intensity gamma-radiolysis studies of solid salt will need to be conducted as a precursor
activity to in-pile salt irradiations to define the necessary safety envelopes for the experiment.

The interactions at the gas–salt interface in the MSRE pump bowl appeared to be much more complex
than in the salt solution or at the container-salt interface (see Figure 18); and molten salt mist was found
in the gas space, leading to deposition of fission products. No corrosion acceleration—akin to that shown
in Figure 26 for zirconium in water reactors—was seen during intense irradiation of coolant salt to high
fluence and high power density. So far, only the tellurium embrittlement effect was obvious for fluoride
fuel salt in the MSRE, as shown in Figure 27.

Figure 26. Irradiation enhancement of corrosion rates. | B. F. Kammenzind, J. A. Gruber, R. Bajaj, and J. D. Smee.
“Neutron irradiation effects on the corrosion of Zircaloy-4 in a PWR environment.” Paper B-T-3965, 18th International
Symposium on Zirconium in the Nuclear Industry, Hilton Head, South Carolina, May 15–19, 2016. Used by permission.

52 FRD 4
 Molten Salt Chemistry Workshop

(a) (b)

Figure 27. Cross-section analysis of INOR-8 (heat 5065) sample tested at 25 °C (a) after being exposed to
the MSRE core for 19,136 h above 500 °C and irradiated to a thermal fluence of 1.5⋅1021 neutrons/cm2
and (b) after being exposed to static unenriched fuel salt for 19,136 h above 500 °C. | Image courtesy of
Oak Ridge National Laboratory

5.2.5 Predicting Interfacial Interactions between Molten Salts and Structural Materials
and/or Gases
There is a need for predictive computational models to describe the molecular, mesoscale, and macroscale
interfacial processes that occur in an MSR over a wide range of time scales (from nanoseconds to years)
(see Figure 28). There are two primary types of interfaces relevant to MSRs: (1) molten salt fluid–solid
and (2) cover gas–molten salt. Both interface types need to be understood in terms of fundamental
interactions and interfacial transport processes. Relevant phenomena include

• Noble metal (fission products) plating on both the structural materials and the moderator
• Deposition of fissile materials on the structural interface and subsequent migration into the bulk solid
phase
• Nucleation of particles and transport and collection at the gas–liquid interface that adversely affect
the mass transport of volatile fission products into the gas phase
• Gas-phase entrainment into the molten salt (in the form of microbubbles from the cover gas) either by
design or unintentionally
• Entrainment of molten salts in the gas (mist formation)
• Corrosion of structural components

The underlying timescales span from nanoseconds, for gas diffusion, to months to years for corrosion.
Current modeling approaches can treat fluids and solids on timescales that typically are from picoseconds
to a microsecond at the atomistic/quantum level of theory, and microseconds to seconds at the mesoscale
level of theory. Extending the time scale assailable by predictive simulations can provide essential
missing knowledge of interfacial transport phenomena, in many instances available by no other practical
means, thereby aiding the design approaches for commercial MSRs.

FRD 4 53
Molten Salt Chemistry Workshop

Figure 28. Illustration of the length and time scales (and inherent feedback) involved in the multiscale
processes responsible for microstructural changes in irradiated materials. | Reprinted from B. D. Wirth, X. Hu, A.
Kohnert, and D. Xu. “Modeling defect cluster evolution in irradiated structural materials: Focus on comparing to high-
resolution experimental characterization studies.” J. Mater. Res. 30, 1440–1455, 2015. Used by permission.

There are several critical needs that must be addressed for interfacial interactions between molten salts
and structural materials and/or gases that pose significant challenges to the current computational-based
approaches:

• Prediction of long time processes (particularly regarding corrosion) by scaling up insights from
molecular interactions
• Prediction of interfacial transport processes
• Prediction of surface layer formation leading to degradation of reactor performance with fission
processes (e.g., metal plating)
• Prediction of surface layer formation leading to degradation of heat exchangers
• The ability to couple chemical and physical phenomena with microstructure and composition

One of the challenges is that interfacial regions are inherently inhomogeneous, making transport
phenomena near such regions difficult to quantify. The transport of reactive chemical species across
interfaces and/or species reactivity at the interface are particularly challenging. Nevertheless, a concerted
approach coupling information from simulations at the quantum level of theory (first principles MD),

54 FRD 4
 Molten Salt Chemistry Workshop

classical MD with accurate/informed force fields, and mesoscale simulations will provide far more insight
into the physicochemical behavior of interfaces than is currently available. Polarization effects and
chemical changes dominate phenomena at interfaces; therefore, the combination of methods is a necessity
despite the challenges of passing information from one level of theory to another in either direction. This
multiscale simulation approach will provide new understanding of interfacial transport phenomena in
view of the challenges of experimental methods for probing molten salt interfaces under realistic
conditions of reactor operation. Hence the research directions should follow from the following three
platforms.

1. Predict dynamics at molten salt–materials and molten salt–gas interfaces: First principles
quantum mechanical MD (QM/MD) simulations are needed for modeling redox states of salt
mixtures, fission and activation products, structural metals, and other impurities near the interfacial
region to predict transfer of species to and from the salt across the interface. The results of this type of
investigation provide information on the mobility of interfacial species; residence times in the
interfacial region; and changes in interfacial structure—including the formation of complexes and
nucleation of particles—that lead to precipitation and coating. With these data in hand, classical
mechanical simplifications can be constructed to allow for longer time-scale and larger length-scale
coarse-grained simulations. As with all modeling, comparison with experimental data is indispensable
for validating calculated results and for imparting confidence in the computational results.
2. Develop approaches for coupling chemical and physical phenomena with microstructure and
composition: Various simulations of molten salts have been performed to study the structure and
dynamics of bulk molten salt mixtures using classical MD (CMD) simulations. Equipped with new
information from interfacial QM/MD, classical models can be built considering redox states,
interfacial speciation, and polarization to carry out simulations that allow for species to cross the
interfacial region at representative length and time scales. Of interest are species involved in
corrosion of the solid–liquid interfaces, migration of fission products into structural material defects,
and escape of volatiles into the gas phase in contact with the molten salt phase. In this effort, trends of
mesoscopic mass transport can begin to be assessed and proposed for the next level of coarse-graining
models aimed at stimulating rate-based processes
3. Predict diffusion and chemical transformations in confined spaces near interfaces: Mesoscale
models and transport simulations based on classical conservation laws of convective-diffusive-
reactive processes in nonequilibrium, when calibrated by subscale models (typically CMD), allow for
further insight into trends of species transport and reactivity in interfacial regions on time scales of up
to seconds and length scales of tens of microns. For example, phase field methods coupled with
accurate thermodynamic data sets and kinetic data can provide detailed transport simulations with
chemical reactivity, including nucleation and new phase formation. Together with cluster dynamics
methods, mesoscale approaches can also address radiation damage processes within this framework.

Experimental and reactor results can be used to improve the confidence of macroscopic simulations at the
continuum level of theory commonly used to develop nuclear reactor simulation codes. Ultimately, this is
the simulation tool used to design, build, and test prototype reactors. Therefore, tools are needed for the
key steps toward connecting the research needed to bridge the knowledge gap and to connect directly to
practical simulation tools used in the licensing space of nuclear reactors. These tools need to be made
broadly available to the community with the necessary computational resources, software, and data sets
(see Section 7 discussion on the Virtual Reactor Simulation).

Progress in these research directions will lay down a scientific foundation for building models and
performing simulations of interfacial transport phenomena of pure electrolytes at high temperatures
(800 °C and potentially higher) in the presence of solid and gaseous interfaces. This class of systems
applies not only to MSRs but also to metallurgy, nonaqueous materials synthesis, electrochemical

FRD 4 55
Molten Salt Chemistry Workshop

processing, and solar energy capture and conversion to solar-based fuels. Therefore, the simulation tools
developed in this effort will apply broadly to various crosscutting applications of relevance to the energy
and materials sectors. Specifically, these tools will enable an improved understanding of impurity
interactions at interfaces, of interfacial selectivity and binding, of the volatility of dissolved species, and
of the effects of confined phases. Numerous crosscutting applications will benefit from the modeling and
simulation framework developed, in addition to MSR research.

The recommended computational research directions, supported and validated by experimental results,
will pave the way to understanding and predicting a class of interfacial phenomena that are serious
concerns in designing commercial MSRs with long operating lives (tens of years) that do not require
reprocessing of the salt. For example, this work will enable the prediction of the corrosion of structural
components by the salt at the inside surfaces and at the heat exchanger, enabling the design of new
materials and processes to minimize corrosion and the prediction of material lifetimes under reactor
conditions. The advances will enable the design of reactor materials and alloy compositions. This will
accelerate the qualification of novel materials for MSR applications. Such work will improve the
operando removal of fission products, xenon, krypton, iodide, and tritium. In addition, it will lead to an
improved understanding of

• The loss of tritium through diffusion in the vessel walls

• Plating of surfaces by fissile products and noble metals
• The role of confinement effects on interfacial transport of the entrained gas phase
• Permeation of fission products and/or fissile species in structural materials with defects and/or grain
boundaries

References

1. J. H. DeVan and R. B. Evans III. Corrosion Behavior of Reactor Materials in Fluoride Salt Mixtures,
ORNL/TM-328, Oak Ridge National Laboratory, 1962; J. H. DeVan, Effect of Alloying Additions on
Corrosion Behavior of Nickel-Molybdenum Alloys in Fused Fluoride Mixtures, master’s thesis,
University of Tennessee, August 1960 (see also reference 20).
2. F. Zaera. “Probing liquid/solid interfaces at the molecular level.” Chem. Rev. 112, 2920–2986, 2012.
3. T. F. Heinz, C. K. Chen, D. Ricard, and Y. R. Shen. “Spectroscopy of molecular monolayers by
resonant second-harmonic generation.” Phys. Rev. Lett. 48(7), 478–481, 1982.
4. K. B. Eisenthal. “Equilibrium and dynamic processes at interfaces by second harmonic and sum
frequency generation.” Annual Rev. Phys. Chem. 43(1), 627–661, 1992.
5. J. I. Dadap, J. Shan, and T. F. Heinz. “Theory of optical second-harmonic generation from a sphere of
centrosymmetric material: Small-particle limit.” J. Optical Soc. Am. B 21 (7), 1328–1347, 2004.
6. S. W. Chan, R. Barille, J. M. Nunzi, K. H. Tam, Y. H. Leung, W. K. Chan, and A. B. Djurišić.
“Second harmonic generation in zinc oxide nanorods.” App. Phys. B 84(1–2), 351–355, 2006.
7. J. Zeng, H. M. Eckenrode, S. M. Dounce, and H.-L. Dai. “Time-resolved molecular transport across
living cell membranes.” Biophys. J. 104(1), 139–145, 2013.
8. W. Fan, S. Zhang, N. C. Panoiu, A. Abdenour, S. Krishna, R. M. Osgood, K. J. Malloy, and S. R. J.
Brueck. “Second harmonic generation from a nanopatterned isotropic nonlinear material.” Nano Lett.
6(5), 1027–1030, 2006.
9. P. S. Pershan and M. L. Schlossman. Liquid Surfaces and Interfaces: Synchrotron X-ray Methods.
1st ed., Cambridge University Press, 2012, p 334.

56 FRD 4
 Molten Salt Chemistry Workshop

10. P. Fenter and N. C. Sturchio. “Mineral–water interfacial structures revealed by synchrotron x-ray
scattering.” Prog. Surf. Sci. 77(5–8), 171–258, 2004.
11. M. K. Bera, G. Luo, M. L. Schlossman, L. Soderholm, S. Lee, and M. R. Antonio. “Erbium(III)
coordination at the surface of an aqueous electrolyte.” Phys. Chem. B 119(28), 8734–8745, 2015.
12. G. Luo, W. Bu, M. Mihaylov, I. Kuzmenko, M. L. Schlossman, and L. Soderholm. “X-ray reflectivity
reveals a nonmonotonic ion-density profile perpendicular to the surface of ErCl3 aqueous solutions.”
J. Phys. Chem. C 117(37), 19082–19090, 2013.
13. G. Gao, F. H. Stott, J. L. Dawson, and D. M. Farrell. “Electrochemical monitoring of high-
temperature molten-salt corrosion.” Oxid. Met. 33, 79–94, 1990.
14. D. Ludwig, L. Olson, K. Sridharan, M. Anderson, and T. Allen. “High temperature electrochemistry
of molten fluoride salt for measurement of dissolved chromium.” Corros. Eng. Sci. Technol. 46(4),
360–364, 2011.
15. A. G. Fernández, A. Rey, I. Lasanta, S. Mato, M. P. Brady, and F. J. Pérez. “Corrosion of alumina-
forming austenitic steel in molten nitrate salts by gravimetric analysis and impedance spectroscopy.”
Mater. Corros. 65, 267–275, 2014.
16. L. F. Epstein. “Static and dynamic corrosion and mass transfer in liquid metal systems. Liquid metals
technology.” Chem. Eng. Prog. Symp. Ser. 53(20), 67–81, 1957.
17. R. B. Evans, III, and P. Nelson, Jr. Corrosion in Polythermal Systems, I. Mass Transfer Limited by
Surface and Interface Resistances as Compared with Sodium-Inconel Behavior, vol. 1, ORNL-4575,
Oak Ridge National Laboratory, 1971.
18. J. H. DeVan. Effect of Alloying Additions on Corrosion Behavior of Nickel-Molybdenum Alloys in
Fused Fluoride Mixtures, ORNL/TM-2021, Oak Ridge National Laboratory, 1969.
19. W. N. Gill, R. P. Vanek, R. V. Jelink, and C. S. Grove, Jr. “Mass transfer in liquid -lithium systems.”
A.I.Ch.E. J. 6, 139–44, 1960.
20. C. E. Sessions and J. H. DeVan. “Thermal convection loop tests of Nb-1% Zr alloy in lithium at 1200
and 1300 °C.” Nucl. Appl. and Technol. 9, 250–259, 1970.
21. J. J. Keyes, Jr. Some Calculations of Diffusion Controlled Thermal Gradient Mass Transfer, CF
report 57-7-115, Oak Ridge National Laboratory, 1957.
22. D. F. Williams. Assessment of Candidate Molten Salt Coolants for the Advanced High Temperature
Reactor (AHTR), ORNL/TM-2006/12, Oak Ridge National Laboratory, 2006.
http://info.ornl.gov/sites/publications/Files/Pub57476.pdf
23. J. R. Keiser. Status of Tellurium-Hastelloy N Studies in Molten Fluoride Salts, ORNL-TM-6002, Oak
Ridge National Laboratory, 1977.
24. P. J. A. Spencer. “A brief history of CALPHAD.” Calphad 32(1), 1–8, 2008.

FRD 4 57
 Molten Salt Chemistry Workshop

6. FRD 5: Guiding Next-Generation Materials for Molten Salt Reactors

6.1 Background and Current Status

Structural materials (ranging from graphite to advanced metallic alloys) proposed for use in MSRs will be
exposed to extreme environments, including high fluences of neutrons, high operating temperatures, and
corrosive environments. Challenges associated with the high temperatures will primarily be the thermal
stability of such materials, such as creep strength and fracture toughness; and irradiation effects could
lead to structural evolutions such as irradiation-induced swelling, creep, or embrittlement. These effects
could be exacerbated by the potential dissolution of solutes into or out of the structural materials from
molten salt/structural material interactions. The compositions of the melt and structural material are
expected to be the dominating factors influencing degradative processes. However, salt composition is
also anticipated to change as a function of time because of the formation of transuranics, fission products,
and corrosion products and the introduction of contaminants. These will create a highly complex and
dynamic environment that will affect the structural materials. Despite the performance already
demonstrated by candidate materials, the development of advanced materials that can mitigate
degradation remains a key element for the safe, successful operation of structural components in MSRs.

The stability of nickel-based alloys, as well as of steels, graphite, and ceramics, under various reactor
irradiation conditions (real or simulated) has been the subject of study for decades; and the various modes
of materials degradation are fairly well defined for different regimes of dose and dose rate. [1]
Nevertheless, advances in the fundamental understanding of radiation damage and incremental
improvements in the radiation tolerance of materials have been made, [2] and efforts to discover and
design new materials with revolutionary stability under irradiation are being actively pursued. [3]
Although many of these studies are not specific to MSRs, several have focused on concepts that require
higher temperatures; but there have not been as many studies on radiation effects. Furthermore, there has
been relatively little effort invested in designing new nickel-based alloys for radiation resistance because
much of the attention of the community over the past 25 years has been on other materials systems (partly
because of the observed poor phase stability of 1970s-era nickel-based superalloys after moderate
irradiation doses at elevated temperatures).

Neutron irradiation effects on alloys in the temperature range of interest for MSRs are mostly cavity
swelling, phase instabilities associated with radiation-induced segregation, irradiation creep, and helium
embrittlement of grain boundaries. [1, 4] Neutron irradiation effects on graphite involve an initial volume
shrinkage that is followed by volume expansion. [5] Unlike for metals, irradiation-induced creep is
beneficial in the case of graphite because it can partially relieve some of the large stresses associated with
high volumetric swelling.

6.2 Technical Challenges and Research Directions

6.2.1 Enabling Rapid Deployment of MSR Materials: Computational Challenges

MSRs represent a unique materials challenge with limited compatible materials availability and open
questions regarding material lifetimes and durability. New improved methods are needed which combine
experimental characterization data with predictive modeling capabilities to develop new materials,
including alloys and composites with the required radiation and chemical stability that are optimized or
qualified for MSR use. In addition, the time for new materials development needs to be greatly accelerated
to meet required timelines for testing and code qualification purposes. Addressing these needs requires
that an innovative combination of computational materials science and high-throughput combinatorial
material testing be developed. A comprehensive set of thermochemical models and data are needed to
elucidate how elemental content, reduction potential, and temperature interact to control chemical

FRD 5 59
Molten Salt Chemistry Workshop

potential, thermophysical properties, and viscosity and predict such properties at the required level of
accuracy. Additional studies and computational methods are needed to increase knowledge of the
microstructure-property-corrosion correlation to improve the ability to successfully develop needed
material solutions for MSRs. Finally, new efficient synthesis and processing methods are needed to
prepare materials compatible with MSR operation.

Methods are needed to predict material behavior in complex molten salt environments. MSR-
compatible materials can be designed to exhibit optimal behavior in complex molten salt environments if
researchers can understand and predict the interactions at the interface between the material and the salt,
as discussed in Section 5. This represents a significant challenge because the salt chemistry and properties
change with time, temperature, and burnup as fission products and tritium are produced. It is imperative
that these computational models be dynamic and comprehensive in treating these interface reactions.
Achieving this capability will require close coupling between model development, material testing in salt
systems, and fundamental investigations of salt interfaces affording an atomistic understanding of the
resulting physicochemical phenomena. Such modeling capabilities are at the crossroads of traditional
computational material science, computational chemistry, and experimental materials science and
experimental physical chemistry. A challenge will be to understand the MSR salt chemistry evolution in
parallel to material testing so that the right experimental conditions are included.

There is a need to understand and predict the microstructural evolution of materials in dynamic
radiation and chemical environments. MSRs create a unique, dynamic chemical and radiation
environment where any given structural material (e.g., graphite or Hastelloy N) will be undergoing
simultaneous radiation damage and chemical attack while the MSR salts are undergoing radiolysis. This
situation represents a significant challenge in that bulk radiation damage of structural materials is coupled
to interfacial salt–material interactions. The effort to address these issues can draw on multiscale
computational methods that are being developed for the study of materials operating under irradiation
conditions or of plasma surface interactions. Radiation damage evolves microstructures in known ways,
which include solute segregation that will compete with salt–interface reactions to produce a unique
interfacial reaction zone. Ideally, this interface segregation can be optimized to provide a protected or
stable interfacial region during MSR operation. Fission product uptake and diffusion, along with tritium
uptake and helium production in the structural material, represent a further challenge in material design.
MSR materials must be designed to mitigate high-temperature helium embrittlement and fission product
corrosion. It will be necessary to predict many of these phenomena—such as helium-bubble nucleation at
grain boundaries—over multiple time and length scales. For many materials, MSR operational
temperatures provide relief from radiation hardening and embrittlement but push the material into a
regime in which helium embrittlement can become an issue as helium is transported to grain boundaries
and retained. This requires the ability to couple detailed helium trap models with helium transport and
mechanics models to design materials that are resistant to helium production.

To identify and design structural materials with MSR compatibility, it is necessary to develop novel
multi-resolution (in terms of physics, length scales and time scales) modeling capabilities (see Figure
29). This is a complex challenge with no suitable methods yet developed that can address these multiple
time and length scales. Ab initio and classical methods will be required for structural material–salt
interactions and for radiation damage resistance. The latter can be partly addressed using classical
approaches coupled with cluster dynamics, spatial-dependent rate theory, kinetic Monte Carlo, and phase
field or other mesoscale methods. However, as yet there are no suitable methods that can address
radiation damage, helium trapping and migration, and tritium production and uptake for complex alloy
systems. In any event, some starting point is required, such as Hastelloy N, for example, for alloy design.
Chemical reaction models will need to be integrated with damage and deformation models to capture the
range of corrosion, radiation damage, and thermal degradation that are anticipated.

60 FRD 5
 Molten Salt Chemistry Workshop

The study of structural material–salt interactions will require the development of improved ab initio
methods (in terms of speed) or improved classical potentials for alloy/salt systems to describe the
complex corrosion and microstructural evolution that will occur at the material–salt interface. Rapid
deployment approaches will rely on close coupling of computational methods and rapid characterization
of materials exposed to salt test loops. However, the lack of knowledge regarding concurrent radiation
damage cannot be overcome using this approach. Thus, a MSR test reactor system is required.

The paradigm envisioned in Figure 29 will enable the elucidation of the fundamental chemomechanical
degradation mechanisms leading to corrosion at molten salt–structural materials interfaces, in addition to
bulk radiation damage and thermomechanical degradation [6–9]. The ability to adequately simulate the
response of such microstructures, to elucidate the interdependent compositional and microstructural
evolution, and to gauge the associated change in thermomechanical properties in the extreme
environments experienced in MSRs is crucial and necessary. Such a capability is essential not only for
preventive and remedial actions to ensure safety and for performance evaluation over the lifetimes of
MSR components, but also for the development of new generations of high-performance, corrosion-
resistant MSR materials.

Compatibility with salt systems will require thermodynamic data for MSR salt fuels, as discussed in
Section 2. The complex elemental content of a thermochemical solution model, together with those
necessary for the panoply of crystalline phases that can precipitate, is unprecedented. Research is
necessary to develop methods that efficiently provide accurate input data for such models and
accommodate means of statistical validation. Determining the most advantageous salt composition will
require the development of relationships for use in optimization algorithms.

Advanced computational methods are needed to model and predict corrosion as a function of
microstructure evolution. Any framework designed to improve the modeling and prediction of the
corrosion behavior in MSR materials must be based on the concurrent coupling between composition
evolution, microstructure evolution, localized deformation, and damage accumulation. Atomic-
/molecular-scale ab initio–based input will be necessary to calibrate this type of model and capture the
adequate fundamental mechanisms influencing corrosion in MSR salts. These models can then be
upscaled to phase field models, for example, to predict phase evolution and corrosion at the microscale
and macroscale. The complexity envisioned for these models, however, is unprecedented and will require
new methods of coarse-graining materials or upscaling to comprehensive mesoscale approaches.

Models are needed to predict materials mechanical degradation processes under MSR conditions. The
mechanical properties of the materials in an MSR, such as creep strength and fracture toughness, will
gradually degrade over time because of the severe environmental exposure. Although adequate models
exist for such degradation mechanisms, the MSR environment is unique. For example, models for high-
heat-flux materials for advanced fusion reactors are generally inadequate when plasma surface
interactions must be coupled to bulk radiation damage models of hardening. Likewise, MSR corrosion
models must be coupled with similar hardening models using new methods. Reactor transients, such as
startup and shutdown, exacerbate this problem by introducing nonequilibrium states into the degradation
mechanism space. This adds additional complexity by forcing the models to consider time-dependent
stresses instead of steady state stresses. As a starting point, radiation damage models of hardening and
swelling should be investigated.

FRD 5 61
Molten Salt Chemistry Workshop

Figure 29. Example of multi-resolution modeling capabilities for microstructure evolution and corrosion. |
(From left) W. Barrows, R. Dingreville, and D. Spearot. “Traction–separation relationship for hydrogen induced grain
boundary embrittlement in nickel via molecular dynamics simulations.” Mater. Sci. Eng., A 650(5), 354–364, January 2016;
F. Soisson. “Kinetic Monte Carlo simulations of radiation induced segregation and precipitation.” J. Nucl. Mater. 349, 235–
250, 2006; X. Q. Ma , S. Q. Shi, C. H. Woo, and L. Q. Chen. “Simulation of γ-hydride precipitation in bi-crystalline zirconium
under uniformly applied load.” Mater. Sci. Eng., A 334, 6–10, 2002; A. Musienko and G. Cailletaud. “Simulation of inter-
and transgranular crack propogation in polycrystalline aggregates due to stress corrosion cracking.” Acta Mater. 57,
3840–3855, 2009; A. Y. Dunn. Radiation Damage Accumulation and Associated Mechanical Hardening in Thin Films and
Bulk Materials. PhD Thesis, Georgia Institute of Technology. 2017.

Predictive multi-resolution modeling capabilities, addressing the degradation of structural materials

subjected to extreme environments experienced in MSRs, will not only advance our current understanding
of corrosion mechanisms due to molten salts but also create a new paradigm in the computational
materials science and computational chemistry community for studying corrosion phenomena for a wide
range of applications beyond MSRs. Indeed, since corrosion of materials is one of the main life-limiting
processes for materials in a variety of settings, and is even more critical for MSRs, the development of
improved corrosion models will greatly benefit predictive capabilities in a variety of fields. This research
will greatly facilitate alloy development by advancing innovative ways of coupling advanced
characterization methods and data sets with advanced computational models of salt/material interactions
and bulk materials degradation models. The advanced chemistry approaches developed will have wide-
ranging impacts in a variety of material-centric fields.

62 FRD 5
 Molten Salt Chemistry Workshop

The ability to predict material properties under extreme conditions is essential to understanding how
materials degrade and fail. The ability to accurately predict material lifetimes provides needed safety
information as well as a foundation for the design of improved materials. The rational design of materials
using these models will allow new and accelerated approaches for material deployment in extreme
environments, such as MSRs, by increasing materials testing throughput. Tailored and optimized
materials will enable MSRs and other technologies required for future energy production. These methods
could impact materials design efforts for other advanced nuclear systems or other high-temperature
material systems.

6.2.2 Guiding Experimental Exploration of Advanced “Super” Materials

Although much work has focused on the quantification of current materials for use in MSRs, there exists
a need for the development of new materials to meet the increased demands of advanced reactor concepts.
The development of INOR-8 (Hastelloy N) after the ARE and its deployment in the MSRE serves as an
example of a successful development of a new material for use in molten salt environments. Although
reactor vendors are currently pursuing designs that incorporate alloys from the current ASME Boiler
Pressure Vessel Code, there is a tremendous opportunity to take advantage of innovative advanced
materials or “super materials.” Although code-qualifying a new alloy or material can take over a decade
and cost tens of millions of dollars, the benefits in terms of operational life, maintenance costs, and safety
could be significant. Nevertheless, most companies are not willing to investigate non-code materials
because of the time required and the cost.

Current materials, including Hastelloy N, cannot effectively address the higher temperatures and more
aggressive radiation and chemical environments presented by some advanced MSR concepts. Proposed
reactors with core temperatures above 900 °C and radiation damage over 1000 displacements per atom
(dpa) represent pathways to energy independence through the use of tailored thermochemical processes
(e.g., thermochemical H2 cycles), increased power-cycle thermal efficiency, and added utility from
enabling flexible fissile fuel options. To achieve these goals requires developing new classes of super
materials that can withstand the more aggressive conditions. Materials must be sufficiently
thermodynamically stable in the coolant/fuel (i.e., molten salt), accommodate imposed mechanical loads,
and be immune to significant dimensional and property changes from radiation damage over the design
lifetime of the reactor system. Performance requirements must be balanced with practical considerations
such as availability, affordability, fabricability, and manufacturability and acceptance by regulatory
bodies.

High temperatures are typically associated with loss of strength, creep, and higher corrosion rates. High
dpa from exposure to neutron flux is associated with hardening, void swelling, and helium embrittlement.
In MSRs, the challenge is compounded by temperature gradient, mass transfer, corrosion, and reactions
with fission products, i.e., tellurium. The materials design space for MSRs includes the use of metals,
ceramics, coatings, claddings, multilayer and nanostructured materials and, as necessary, composites.
This variety of materials needs to be explored comprehensively to discover the super materials needed for
maximum MSR performance. There is also a need for selective design of materials to ensure lifetime
requirements are met. These requirements include maintaining material strength at temperatures in excess
of 750 °C without suffering cracks or embrittlement, resisting corrosion, and also maintaining size and
shape within the design specifications.

The development cycle for new structural materials can be shortened through the application of modern
materials science computational tools, as discussed earlier. Computational studies can be performed to
down-select favorable chemical compositions from thermodynamic and kinetic stabilities; predict
properties; and simulate component lifetimes through known/verified mechanisms of time-dependent
mechanical, corrosion, and radiation behavior before trial materials are manufactured and properties

FRD 5 63
Molten Salt Chemistry Workshop

screened (see the sidebar “Optimizing Superalloys” on p. 65). This approach is an improvement over the
traditional development path of fabricating trial materials from a large chemistry matrix for screening
tests, and it has been successfully employed in a number of DOE–NE programs. Improvements in the
accuracy of these tools and the development of new computational tools and databases to cover additional
material behavioral responses would further shorten the development cycle. When the justification exists
to actually start materials testing, there is also the possibility of much smaller alloy synthesis routes
compared with the traditional large “heats” of years past.

It also needs to be recognized that some of the components in an MSR will have different requirements at
different locations, leading to more severe restrictions on material development. This issue could be
addressed by developing innovative advanced manufacturing methods, such as additive manufacturing,
that would allow tailoring of properties via the adjustment of compositions throughout the structural
component. In addition, innovative methods of materials development can have a significant scientific
impact on how material design is performed. This impact may lead to a deeper understanding of the
underlying scientific principles of how materials behave in extreme environments and how they are best
tailored for use in high-temperature, molten salts environments that are subjected to high radiation levels.
Furthermore, the current limitations in understanding the joining of dissimilar materials, such as joining
ceramics to metal seals, must be addressed; doing so would be an enabling technology.

6.2.3 Advancing Code Qualification of New Materials

A significant hurdle for the development of MSR structural materials is code qualification. The design of
high-temperature coolant boundary structural components for deployment in nuclear reactor systems is
required to follow the construction rules of the ASME Boiler and Pressure Vessel Code (Section III,
Division 5). The qualification of new metallic materials requires multiple steps, including composition
selection, processing and heat-treating steps, and finding a vendor willing to produce large quantities of
the material.

After development, ASME-based qualification of the metal includes long-time scale testing at operational
temperatures (anywhere from 1/5 to 1/3 of the expected usage lifetime), preferred examples of the
material applicability in other non-nuclear extreme environments, corrosion testing, and irradiation-
induced changes to the physical/thermal/mechanical properties. Many of these steps are not easily
accelerated, since the mechanisms controlling the changes may be not be the same at different conditions.
Testing at operational temperatures could easily require 6–10 years of continuous testing for a component
for which a 30 year usage lifetime is desired. In many cases, such a long period of testing is cost-
prohibitive; therefore, materials used in other industries often are leveraged toward ASME code
validation.

Study of the changes caused by the irradiation environment in the reactor also is very time-intensive. In
the current materials test reactors (High Flux Isotope Reactor and Advanced Test Reactor), lifetime
damage accumulation can be sped up by a factor of 5–10. As a result, a 30 year lifetime component
requires a minimum of 3–6 years of irradiation time alone (while the development and construction of the
irradiation capsules and/or the post-irradiation examination of the properties and microstructure could
expand the testing program to 7–10 years). Corrosion studies in the MSR environment will also require
experiments to study the base metal–salt interaction at high temperature, as well as exposure to the salt in
an irradiation environment, since the chemistry of the salt will be continuously changing as fission occurs.
These experiments are technically challenging and require complex experimental setups in addition to
long times to perform the tests. These issues highlight the need for innovative experimental test protocols
and analytical methods to efficiently gather the required information for code qualification.

64 FRD 5
 Molten Salt Chemistry Workshop

The qualification of nonmetallic materials requires a similar development process to that for metallic
materials. The primary difference is that qualification for nonstructural materials usually does not include
long-term testing at operational temperatures (not required for ASME code qualification). But long-term
studies of irradiation-induced changes and corrosion behavior are still critical knowledge requirements for
in-reactor use.

Optimizing Superalloys
Nickel-based superalloys have an exceptional combination of high-temperature mechanical
properties, primarily as a result of the formation of coherent nano-precipitates of the Ni3Al L12-
type in the nickel-based face-centered-cubic matrix. However, achieving similar microstructure and
performance in other low-cost materials continues to be a challenge. Today it is feasible to use a
modern accelerated approach to high-temperature structural alloy development based on atomic-
level control of interfaces. Predictive design of favorable properties in materials under extremes
requires that the microstructural interfaces, such as grain and phase boundaries, be stable under
high temperature and stress. While phase diagrams for stability of individual phases are often well
known, an integrated understanding of the interfacial stability in the presence of individual or
multiple atomic species is still needed. Microalloying elements that maximize stress and
temperature resistance by computing the energetics of higher mobility interfaces are an aspect that
needs to be verified and refined through experiments. Large-scale, high-throughput ab initio
calculations can be the basis for simulating interfacial evolution at high stress and homologous
temperature (ratio of temperature and absolute melting point). A suite of experimental
characterization techniques—such as atom probe tomography, transmission electron microscopy,
neutron diffraction, and mechanical characterization—can be used to help verify the computational
predictions. Such models and experimental results can form a scientific framework for
understanding and controlling interfacial stability in superalloys, thereby dramatically accelerating
the design of next-generation, advanced alloys optimized for molten salt reactors.

FRD 5 65
Molten Salt Chemistry Workshop

References

1. S. J. Zinkle and J. T. Busby. “Structural materials for fission and fusion energy.” Mater. Today 12,
12–19, 2009; R. M. Boothby. “Radiation effects in nickel-based alloys,” in ed. R.J.M. Konings
Comprehensive Nuclear Materials, vol. 4. Elsevier, Amsterdam, 2012. pp. 123–150; A. F. Rowcliffe,
L. K. Mansur, D. T. Hoelzer, and R. K. Nanstad. “Perspectives on radiation effects in nickel-base
alloys for applications in advanced reactors.” J. Nucl. Mater. 392, 341–352, 2009.

2. S. J. Zinkle and L. L. Snead. “Designing radiation resistance in materials for fusion energy.” Annu.
Rev. Mater. Res. 44, 241–267, 2014.
3. J. B. Roberto and T. Diaz de la Rubia. Basic Research Needs for Advanced Nuclear Energy Systems.
Office of Basic Energy Sciences, US Department of Energy, 2006; Y. W. Zhang, G. M. Stocks,
K. Jin, C. Y. Lu, H. B. Bei, B. C. Sales, L. M. Wang, L. K. Beland, R. E. Stoller, G. D. Samolyuk, et
al. “Influence of chemical disorder on energy dissipation and defect evolution in concentrated solid
solution alloys.” Nature Comm. 6, 8736–8739, 2015; G. R. Odette, “Recent progress in developing
and qualifying nanostructured ferritic alloys for advanced fission and fusion applications.” JOM 66,
2427–2441, 2014.
4. T. R. Allen, M. Anderson, R. Ballinger, T. D. Burchell, E. D. Blandford, D. Clark, W. R. Corwin,
G. Flanagan, C. W. Forsberg, P. Hosemann, et al. “Fluoride-salt cooled, High-temperature Reactor
(FHR) Materials, Fuels and Components White Paper,” UCBTH-12-003, Department of Nuclear
Engineering, University of California, Berkeley, California, 2013.
5. L. L. Snead, T. D. Burchell, and Y. Katoh. “Swelling of nuclear graphite and high-quality carbon
fiber composite under very high irradiation temperature,” J. Nucl. Mater. 381, 55–61, 2008; L.L.
Snead. “Fusion energy applications,” in ed. T. D. Burchell, Carbon Materials for Advanced
Technologies. Pergamon, New York, 1999, pp. 389–427.
6. W. Barrows, R. Dingreville, and D. Spearot. “Traction-separation relationships for hydrogen induced
grain boundary embrittlement in nickel via molecular dynamics simulations.” Mater. Sci. and Eng., A
560(5), 354–364, 2016.
7. F. Soisson. “Kinetic Monte Carlo simulations of radiation induced segregation and precipitation.”
J. Nucl. Mater. 349(3), 235–250, 2006.
8. A. Musienko and G. Cailletaud. “Simulation of inter-and transgranular crack propagation in
polycrystalline aggregates due to stress corrosion cracking.” Acta Mater. 57(13), 3840–3855, 2009.
9. A. Dunn, R. Dingreville, E. Martínez, and L. Capolungo. “Synchronous parallel spatially resolved
stochastic cluster dynamics.” Comput. Mater. Sci. 120, 43–52, 2016.

66 FRD 5
 Molten Salt Chemistry Workshop

7. FRD 6: Creating a Virtual Reactor Simulation

7.1 Background and Current Status

MSR technology development requires the use of modeling and simulation to understand the behavior of
the reactor throughout its lifetime, including during normal operation, operational transients (such as
startup and shutdown), and accident conditions. To accurately predict reactor conditions, an appropriately
validated reactor system simulator is needed which incorporates the major physical phenomena that drive
design and operational challenges at a suitable level of fidelity and accuracy. A virtual reactor (VR)
simulator needs to include the reactor core and primary heat exchanger, plus related support systems, such
as off-gas and online salt processing and refueling. The VR simulator should describe phenomena
including neutron transport, thermal hydraulics, isotopic transmutation of the fuel, thermochemical
properties of the fuel salt, and resulting corrosion effects, while also possessing sufficient fidelity to
address material transport due to corrosion and thermal gradient redeposition. The VR would also define
the source term for the MSR. This effort represents a significant scientific challenge because of the
coupling of transport codes with thermodynamic models and material–salt chemistries, but it can build off
the foundational studies outlined in the previous sections of this report.

7.2 Technical Challenge and Research Directions

7.2.1 Creating a Virtual Reactor Simulation

The challenges associated with the deployment of MSR technology are summarized by Briant and
Weinberg [1]:

Two very different schools of reactor design have emerged since the first reactors were
built. One approach, exemplified by solid fuel reactors, holds that a reactor is basically a
mechanical plant; the ultimate rationalization is to be sought in simplifying the heat
transfer machinery. The other approach, exemplified by liquid fuel reactors, holds that a
reactor is basically a chemical plant; the ultimate rationalization is to be sought in
simplifying the handling and reprocessing of fuel.

The existing assumptions and simplifications in current modeling and simulation tools are specialized for
solid-fuel systems; adapting these tools for liquid fuel systems requires significant effort, particularly in
the coupling to the molten salt chemical analysis, as previously discussed. The major scientific challenge
is to develop an accurate simulation of a MSR, including flowing fuel, neutronics, transport, chemistry,
and in situ continuous processing and refueling. The model would need to be dynamic, which represents
a grand challenge because of the complex material properties and transport. The simulator needs to
include the primary loop, i.e., the reactor core and the primary heat exchanger, and related support
systems such as off-gas and online salt processing (if needed) and refueling. The phenomena described by
the simulator must include neutron transport, thermal hydraulics, isotopic transmutation of the fuel,
thermochemical properties of the fuel salt, and corrosion effects. The simulator must also be validated
with well-quantified uncertainties, since the nuclear industry and regulating authorities will need such a
tool to ensure the efficiency and safety of this new technology.

The neutronic response of the reactor depends on salt properties such as density and viscosity, which will
evolve with depletion of the fuel. Changes in fuel composition may also change the preferred speciation
of elements, the liquidus temperature of the melt, and surface-salt interactions. Additionally, tritium
produced by ternary fissions or reaction with salt constituents will easily permeate hot metal alloys and
needs to be taken into account. All these phenomena are coupled together in a MSR system and need to
be accurately described by the simulator.

FRD 6 67
Molten Salt Chemistry Workshop

A transformational advancement of a VR is a salt thermochemical transport modeling and simulation

capability. While this is important for solid-fueled MSR designs, a more extensive contribution will be
made for liquid-fueled systems. The salt thermochemical transport will focus on describing the redox
reactions in the flowing salt and reactivity leading to corrosion, in addition to radiolysis and chemical
speciation in the liquid and gas phases. Online monitoring can provide input and validation of these
computational predictions. The monitoring will be augmented by considerations of reactions with
impurities, and of the nucleation and growth of particulates, which have ramifications for performance,
safety, and off-gas operation.

The molten salt VR must be able to determine the fluence on the core structural material, the neutron
reflector, and the vessel wall, which can be used along with the temperature values and salt composition
to determine the lifetimes of these components. Additionally, these data will enable analysis of tritium
production and the migration through the primary loop. It can also be used to optimize alternative salts
that would reduce tritium production while maintaining good nuclear and chemical properties. The VR
must also be able to couple to power plant system models to investigate transient and safety analysis of
these reactors.

The development of a molten salt VR can leverage existing neutronic and thermal hydraulic capabilities,
but it will need to be extended to capture the complexities of MSRs. A specific challenge that will be
addressed is the “delayed neutron precursor drift” phenomenon, which results in the birth of delayed
neutrons (i.e., the decay of precursors) in a different location from the fission that produces the precursor.
In solid-fuel designs, the fission products are stagnant; but with a liquid-fuel design, they are carried away
with the fluid fuel flow. To capture this effect, delayed neutron precursor groups need to be explicitly
modeled even during steady state operation. The standard delayed neutron precursor models can be
adapted by adding in a convection term. Since the change in the core flow speed changes the effective
delayed neutron fraction in the core, this feature is critical for modeling power transients during accidents
such as a pump trip. In addition to the delayed neutron precursors, there are other elements that require
special treatment because of the flowing fuel, such as xenon, which is a neutron poison. Also, the
redistribution of heat coming from nuclear decay must be properly dispersed throughout the system.

Several key advancements are needed to capture the thermal hydraulic performance of MSRs. The natural
first extension is to add the capability for direct fluid heating as the dominant heat source. Another key
feature is the ability to transport nuclides throughout the system. This capability needs to be extended to
capture multiple species at the same time, which will lead to a coupled set of species transport equations
to account for the interaction between species. An additional complication is the thermophysical
properties variation with concentration of species; since fission products will be generated pointwise, and
convected nonuniformly, the implementations of such variations must be considered. In addition, models
or correlations for dynamic phase diagrams for estimating thermophysical properties and melting point
will be required.

The addition of mass transport and thermochemistry requires information from both the heat and the
momentum fields of the bulk fluid. Additionally, the reactor physics component should provide a
radiation environment that will continually create fission events and transmute the components of the salt
as they are irradiated. It is necessary to account for these effects in considering the thermochemical
reactions in the reactor. The mass transport and thermochemistry will also impact the reactor physics and
the thermalhydraulic solution. The reactor physics model will need the detailed distribution of salt
components throughout the core. For short-lived fission products, there may be a non-trivial gradient in
the concentration during standard reactor operation that must be accounted for. Likewise, the properties of
the molten salt—density, viscosity, thermal conductivity, and specific heat—will change as the salt
changes. Both feedback mechanisms must be accounted for. Figure 30 shows the three major components
of a VR and the data that are needed between the various components.

68 FRD 6
 Molten Salt Chemistry Workshop

Figure 30. A schematic showing the three major components of a virtual reactor simulation and the data
that are needed between the various components. | Image courtesy of Oak Ridge National Laboratory

The virtual simulator code will need validation and uncertainty quantification (VUQ). The VUQ will
require sets of separate effects and integral effect tests to quantify uncertainty for each phenomenon and
each of the implemented multiphysics components. Data are available from international molten-salt loop
tests that can be used for validation. Additionally, the MSR VR should be used to simulate the ORNL
MSRE operation and demonstrate the capability developed to measured data from an operating reactor.
The MSRE operation represents the only data available in the world for a well-characterized operating
MSR and will be crucial for system-level validation.

Two sets of critical experiments were performed at MSRE. The first was performed using 235U fuel
loading. [2] The data measured include the critical position as well as the measurement of various
reactivity coefficients, including temperature, enrichment, differential control rod worth, and the effects
of fuel circulation. The second was performed using 233U fuel. [3] Very similar measurements were
performed with the different driver fuel. Additionally, the dynamic behavior of the MSRE was studied in
detail for these different fuels, [4-6] and its fuel flow behavior characterized. [7]

It is recognized that an MSR is a tightly coupled dynamic multiphysics system consisting of neutronics,
thermal hydraulics, fission product generation and transport, and the chemistry of constituents. The
thermophysical properties will be changing as the fuel moves through the system. The solubility of
solutes will change as new fission products are added. Developing a VR simulator will require
specification of figure of merit or safety parameters and identification and ranking of phenomena with
respect to the figure of merit. Then each physics will have its own constitutive relationships that may
come from experiments or first principles. We anticipate that advanced 3-dimensional meshing
algorithms will need to be developed to adequately handle these demands, with perhaps a phase field
model handling the thermochemistry coupled to other transport models.

Overall, the capabilities of a virtual MSR simulator with a tightly coupled model with fuel flow, salt
chemistry, surface corrosion, neutron transport, and thermal hydraulics would enable advanced design
and analysis of complex fluid-fueled systems. This would allow for the understanding of operational
issues including fuel cycle analysis, fuel loading analysis, prediction of waste forms and quantities, and

FRD 6 69
Molten Salt Chemistry Workshop

prediction of chemical processes required to reduce chemistry issues. The virtual model would also define
the source term for the MSR. In addition, this capability will allow for the assessment of new molten salts
and new reactor designs.

Another application of this simulation capability would be in the design of irradiation test stands and the
analysis of results from irradiation experiments that will be needed for the development of MSR
technology. Nearly all of the phenomena that need to be simulated in a reactor can be probed to some
extent in irradiation tests of increasing complexity, and this simulation capability can be of great value in
designing the irradiation experiments.

References

1. R. C. Briant and A. Weinberg. “Molten fluorides as power reactor fuels.” Nucl. Sci. Engr. 2, 797–
803, 1957.
2. B. E. Prince, S. J. Ball, J. R. Engel, P. N. Haubenreich, and T. W. Kerlin. Zero-Power Physics
Experiments on the Molten Salt Reactor Experiment, ORNL-4233, Oak Ridge National Laboratory,
February 1968.
3. J. R. Engel and B. E. Prince. Zero-Power Experiments with 233U in the MSRE, ORNL-TM-3963, Oak
Ridge National Laboratory, December 1972.
4. T. W. Kerlin, S. J. Ball, R. C. Steffy, and M. R. Buckner. “Experiences with dynamic testing methods
at the Molten-Salt Reactor Experiment.” Nucle.Appl. Tech. 10, 2, February 1971.
5. T. W. Kerlin and S. J. Ball. Experimental Dynamic Analysis of the MSRE, ORNL-TM-1647, Oak
Ridge National Laboratory, October 1966.
6. R. C. Steffy. Experimental Dynamic Analysis of the MSRE with 233U Fuel, ORNL-TM-2997, Oak
Ridge National Laboratory, April 1970.
7. R. J. Kedl. Fluid Dynamic Studies of the MSRE Core, ORNL-TM-3229, Oak Ridge National
Laboratory, November 1970.

70 FRD 6
 Molten Salt Chemistry Workshop

APPENDIX A: WORKSHOP RESOURCE DOCUMENT

Publicly available at https://info.ornl.gov/sites/publications/Files/Pub73652.pdf and included in
the following pages.

Appendix A A-1
A-3
 DOCUMENT AVAILABILITY
Reports produced after January 1, 1996, are generally available free via US Department of Energy
(DOE) SciTech Connect.

Website http://www.osti.gov/scitech/

Reports produced before January 1, 1996, may be purchased by members of the public from the
following source:

National Technical Information Service

5285 Port Royal Road
Springfield, VA 22161
Telephone 703-605-6000 (1-800-553-6847)
TDD 703-487-4639
Fax 703-605-6900
E-mail [email protected]
Website http://classic.ntis.gov/

Reports are available to DOE employees, DOE contractors, Energy Technology Data Exchange
representatives, and International Nuclear Information System representatives from the following
source:

Office of Scientific and Technical Information

PO Box 62
Oak Ridge, TN 37831
Telephone 865-576-8401
Fax 865-576-5728
E-mail [email protected]
Website http://www.osti.gov/contact.html

This report was prepared as an account of work sponsored by an

agency of the United States Government. Neither the United States
Government nor any agency thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal liability
or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents
that its use would not infringe privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute
or imply its endorsement, recommendation, or favoring by the United
States Government or any agency thereof. The views and opinions of
authors expressed herein do not necessarily state or reflect those of
the United States Government or any agency thereof.

A-4
 ORNL/LTR-2017/135

TECHNOLOGY AND APPLIED R&D NEEDS

FOR MOLTEN SALT CHEMISTRY

Resource Document for the

Molten Salt Chemistry Workshop:

Innovative Approaches to Accelerate Molten Salt Reactor

Technologies Deployment

Prepared by David F. Williams

Oak Ridge National Laboratory

Date Published: April 2017

Prepared by
OAK RIDGE NATIONAL LABORATORY
Oak Ridge, TN 37831-6283
managed by
UT-BATTELLE, LLC
for the
US DEPARTMENT OF ENERGY
under contract DE-AC05-00OR22725

A-5
 CONTENTS

CONTENTS .......................................................................................................................... iii

FIGURES ................................................................................................................................v
TABLES .................................................................................................................................v
ACRONYMS ....................................................................................................................... vii
1. INTRODUCTION ...........................................................................................................1
2. MOLTEN SALT REACTOR CHEMISTRIES ..............................................................3
2.1 PHYSICAL CHEMISTRY CHALLENGES .........................................................4
2.1.1 Reasoning Framework for Salt Choices .....................................................4
2.1.2 Phase Diagrams and Solubility Behavior ...................................................6
2.1.3 Thermophysical and Transport Properties .................................................7
2.1.4 Modeling and Simulation ...........................................................................9
2.2 ANALYTICAL CHEMISTRY CHALLENGES (including measurement
instruments and online sensors) .............................................................................9
2.3 IRRADIATION AND FISSION PRODUCT CHEMISTRY CHALLENGES ...11
2.3.1 Molten Salt Fission Product Chemistry....................................................11
2.3.2 Barren-Salt Activation Product Chemistry: Tritium and Sulfur ..............14
2.3.3 Solid Salt Irradiation Behavior .................................................................15
2.4 MATERIALS COMPATIBILITY CHALLENGES ............................................18
2.4.1 Corrosion ..................................................................................................18
2.4.2 Irradiation (Neutron) Damage Limits.......................................................21
3. TECHNOLOGY AND APPLIED R&D NEEDS .........................................................23
4. REFERENCES ..............................................................................................................25
APPENDIX A. THUMBNAIL SUMMARIES OF REACTOR CONCEPTS |UNDER
CONSIDERATION BY MSR TECHNICAL WORKING GROUP ......................... A-1
APPENDIX B. 1969 MOLTEN SALT REACTOR EXPERIMENT VIDEO ...................B-1
APPENDIX C. MOLTEN SALT REACTOR EXCERPTS FROM “AN ACCOUNT OF
ORNL’S 13 RESEARCH REACTORS” BY MURRAY ROSENTHAL ..................C-1

iii

A-7
 FIGURES

Figure 1. Functional categories of molten salts for reactor applications. ....................................... 3

Figure 2. Early MSR program approach to screening and justifying candidate MSR salts. .......... 4
Figure 3. Effect of cationic radius on solubility of tri-fluorides in molten fluorides at 550°C
[6]. ........................................................................................................................................... 6
Figure 4. Trivalent solubility effects in plutonium and minor actinides fueled fluoride salts
[6]. ........................................................................................................................................... 6
Figure 5. Three-dimensional model of LiF-BeF2-ZrF4 phase diagram [11]. .................................. 8
Figure 6. Electroanalytical results from measurement on MSRE fuel [22].................................. 10
Figure 7. MSRE-era electroanalytical and spectrophotometric apparatuses [22]......................... 11
Figure 8. Fission product classes identified during MSRE operation highlighted as groups
in the periodic chart (below) and the fission-yield curve (above). ........................................ 13
Figure 9. Exemplar short-lived fission products that change solubility behavior during
decay in a fuel circuit (typically < 1 minute salt recirculation residence time). ................... 14
Figure 10. MSRE salt phase segregation on cooling [52]. ........................................................... 17
Figure 11. Temperature-gradient mass transfer [53]. ................................................................... 20
Figure 12. Persistent salt corrosion as a temperature-gradient-driven redox cycle [54]. ............. 20
Figure 13. Dual challenges of higher temperatures and displacements for MSR in-core
structural materials [61]......................................................................................................... 22

TABLES

Table 1. Qualitative assessment of salt coolant properties database and predictive methods
[12] .......................................................................................................................................... 9
Table C-1. ARE parameters ........................................................................................................ C-7
Table C-2. MSRE parameters ..................................................................................................... C-9

A-9
 ACRONYMS

ANP Aircraft Nuclear Propulsion Program

ARE Aircraft Reactor Experiment (fuel salt composition: 53.1-40.7-6.2 mol% NaF-
ZrF4-UF4)
MSBR Molten Salt Breeder Reactor (fuel salt composition: 71.7-16-12-0.3 mol% 7LiF-
BeF2-ThF4-UF4
MSR Molten Salt Reactor
MSRE Molten Salt Reactor Experiment (fuel salt composition: 65-29.1-5-0.9 mol% 7LiF-
BeF2-ZrF4-UF4)
ORNL Oak Ridge National Laboratory
PMA plutonium and minor actinides
R&D research and development

vii

A-11
 1. INTRODUCTION

The Department of Energy (DOE) Office of Nuclear Energy is sponsoring the Molten Salt
Chemistry Workshop (Innovative Approaches to Accelerate Molten Salt Reactor [MSR]
Technologies Deployment) to be held at Oak Ridge National Laboratory (ORNL) April 10–12,
2017. This workshop is intended to identify science-based, technology-driven research
opportunities to facilitate and accelerate MSR technologies development.

Highlighted areas will include physical chemistry, irradiation and fission product chemistry,
analytical chemistry (including instruments and in-line sensors), and materials compatibility.

The goals of the workshop are to identify (1) research topics that address both short-term
technology showstoppers and long-term grand challenges that may enable revolutionary changes
in molten salt technologies and (2) a set of priority research directions that may accelerate
molten salt technology development and deployment. The workshop discussion will be focused
on high-risk/high-payoff research in molten salt chemistry and salt chemical properties and
reactions that affect the application of molten salt for advanced nuclear systems.

This resource document was prepared by the author in consultation with subject matter experts
from other national laboratories. In particular, the author acknowledges review and suggestions
by Bill Del Cul and Sheng Dai (ORNL), Guy Frederickson (Idaho National Laboratory), and
Mark Williamson and Sam Sham (Argonne National Laboratory). A draft of this report was
circulated on March 20 to those registered for the workshop for comment (comment deadline
was March 29 for revision) and all comments were addressed. It was released in final form on
April 3.

This document provides information from the scientific research community that underpins MSR
technologies. Emphasis is placed on identifying fundamental knowledge gaps and potential
solutions.

A-13
 2. MOLTEN SALT REACTOR CHEMISTRIES

New molten salt nuclear reactor concepts have recently garnered significant attention in the
United States and other countries. For this workshop, the following systems are emphasized:

• Fast-spectrum chloride fuel salts containing NaCl, sometimes other chloride diluents, and
high concentrations of (U/Pu)Cl3
• Fluoride fuel salts (primarily for thermal spectrum applications)
– Similar to Molten Salt Reactor Experiment (MSRE)/Molten Salt Breeder Reactor (MSBR)
salts containing 7LiF and/or NaF and BeF2 combined with low concentrations (< 5 mol%)
of UF4/PuF3 and sometimes containing larger concentrations of ThF4 (~10 mol %)
– Containing 7LiF and/or NaF and high concentrations of UF4/ThF4 (near eutectic: 25–
30 mol% actinide)
• Primary and secondary coolant salts
– Fluoride salt primary and secondary coolant candidates (e.g., LiF/NaF-BeF2, FLiNaK,
NaF-ZrF4) and sometimes other types of low-pressure secondary coolants (chlorides,
nitrates, liquid metals)
Seven of the eight concepts under consideration by the MSR Technical Working Group in
conjunction with the Gateway for Innovation in Nuclear initiative are fluid-fueled reactors, and
one is for a solid-fueled salt-cooled reactor. Figure 1 provides a functional diagram for these
systems and Appendix A provides a thumbnail description of each one.

Figure 1. Functional categories of molten salts for reactor applications.

A-15
2.1 PHYSICAL CHEMISTRY CHALLENGES

2.1.1 Reasoning Framework for Salt Choices

Salt systems selected for investigation must be justified to ensure that they are viable based on
our current understanding of their physics and chemistry, they are good research and
development (R&D) targets, and there are no better choices. Sometimes lost in the wealth of
recent literature is the underpinning of reasoning that motivated and sustained the early MSR
programs. This traditional approach defined requirements for the application at hand to
downselect salt systems based on their nuclear and chemical behavior, as outlined in Figure 2
[1–4].

Figure 2. Early MSR program approach to screening and justifying candidate MSR salts.

At the top level, neutronics (parasitic neutron absorption and moderating properties) limit the
choices of nuclides that can be considered, in combination with limits associated with the
mission of the reactor (breeder, burner, converter) and neutron energies in the system. Figure 2
highlights the neutronic constraints that strongly limit the choice of constituents for thermal and
epithermal reactors and the chemical constraints that exert a controlling influence on all reactor
applications as a result of the following requirements [1]:
The fuel must dissolve more than the critical concentration of fissionable material at
temperatures safely below the heat exchanger exit temperature. It must consist of elements
of low capture cross sections for neutrons of the energy spectrum of the design chosen.
The mixture must be thermally stable and must possess low vapor pressure over the
temperature range proposed for operation. The fuel mixture must possess heat transfer and
hydrodynamic properties adequate for its service as heat exchange fluid without excessive
power required for circulation. It must be relatively non-corrosive to some (otherwise
suitable) material, presumably a metal, of construction. The fuel must be stable to reactor
radiation, must be able to survive fission of uranium or other fissionable material and must
tolerate fission product accumulation without deterioration of its useful properties [1].

A-16
Early references explain why non-halide fluid-fueled systems (systems containing H, B, C, N, O,
Bi) were rejected for large-scale high-temperature, low-pressure applications based on one of the
chemical criteria listed in the previous paragraph [1–4].

The requirement that fuel salt be compatible with a container material places the most wide-
ranging constraint on the choice of salt constituents because it requires the cationic constituents
of the salt to be composed of relatively active, electropositive elements (fluoride compound free
energies of formation > 95 kcal/mol-F), whereas the container material must be composed of
elements that are relatively noble in comparison (fluoride compound free energies of formation <
75kcal/mol-F) [3]. This requirement eliminates the middle of the periodic chart for consideration
as salt constituents and the periphery of the periodic chart for consideration as alloy constituents
(only a small region in the center is suitable for metallic alloys). Nonmetallic materials (e.g.,
ceramics, composite materials) can also be compatible with high-temperature salt—for example,
graphite is known to be fully compatible with molten halides to very high temperatures.
Although nonmetallic materials have been used and may find service as core internals, there is
currently no joining and fabrication technology to support and deploy these as the container
material for many of the complex components that are in contact with salt.

The neutronic requirements for a primary coolant are distinctly different from those for a fuel
salt, and initial studies indicate that low-neutron-absorbing salts are necessary to obtain negative
void and temperature coefficients [5]. Secondary coolant salts have no formal neutronic
requirements, but consideration is typically given to the ability to recover/remediate leakage of
secondary coolant into primary coolant or fuel salt. The chemical requirements for primary and
secondary coolant salts are similar to those for fuel salts.

A-17
2.1.2 Phase Diagrams and Solubility Behavior

Knowledge of phase diagrams is an essential requirement. Maintaining fissile and fertile material in
solution is a fundamental requirement for solution reactors, so knowing the phase diagram behavior of
actinides is of primary importance. For some MSR concepts, very high burnups and accumulation of
fission products are foreseen, and research must account for how this increasing concentration of
fission/activation products affects the solubility of fissile and fertile materials and the melting point of the
salt. Trivalent elements often play a key role in determining melting point increases and solubility limits
as actinides and fission products accumulate in a fuel salt, as shown in Figures 3 and 4 [6].

Figure 3. Effect of cationic radius on solubility of tri-fluorides in molten fluorides

at 550°C [6].

Figure 4. Trivalent solubility effects in plutonium and minor actinides fueled fluoride salts [6].

A-18
Sometimes unrecognized is the interplay of oxidation-reduction (“redox”) equilibria and phase
diagram behavior. Given the recent interest in salts with high actinide concentrations (up to and
exceeding 30 mol%), the distribution of species between trivalent and tetravalent states
according to the redox condition of the salt will need to be accounted for. For uranium-
containing salts to be noncorrosive toward container materials, a small fraction of uranium must
be trivalent for fluoride salts (UF3) and tetravalent (UCl4) in chlorides. For fluorides, there is a
demonstrated benefit to maintaining salts at more reducing conditions to reduce corrosion and
tellurium-attack [7–8]. Yet these very same conditions tend to increase the trivalent uranium
fluoride content of the salt and could reduce the trivalent solubility margins (trivalents often
exhibit an approximate “group solubility”). The complementary perspective is that trivalent
actinide chlorides are the majority species in chlorides, and the tetravalent solubility associated
with UCl4 and tetravalent fission products will need to be accounted for, as well as the native
volatility of UCl4. Increases in the trivalent content of the fuel salts will occur as a result of
accumulation of activation and rare earths fission products and will likely increase the melting
point of the salt because of a group trivalent solubility effect.

The preceding discussion emphasizes minimizing oxidative corrosion by maintaining the salt in a
reduced state. However, salts kept at very reduced conditions can promote other types of
interactions with container materials and reactor internals, such as alloying, carburization, and
wetting interactions that can cause unwanted intrusion of salt into the pores of core internals
(e.g., graphite pores). This type of interaction between the salt and its container material imposes
a limit on how reducing the salt can be maintained.

Bounds on avoiding both oxidative corrosion and unwanted reductive interactions define a
suitable redox window. Bounds on the redox window for the MSRE and MSBR fluoride salts
were defined (~0.5%–15% UF3/UF4) [9], but those for the desirable redox windows for other
fluorides and chlorides have not yet been established [10].

2.1.3 Thermophysical and Transport Properties

A significant number of molten salt properties have been measured and correlated and this
knowledge should be harvested and used for present needs. However, not all properties are
known or can be accurately estimated, especially those of new salt compositions proposed for
reactor application. A few of the salts under consideration (e.g., 2LiF-BeF2) have a deep
properties database (Figure 5); but for many, we are pushing the boundaries of our knowledge.

A-19
Figure 5. Three-dimensional model of LiF-BeF2-ZrF4 phase diagram. Source: Reprinted from J. Nucl.
Mater. 27(2), R. E. Thoma, H. Insley, H. A. Friedman, and G. M. Herbert, “Equilibrium phase diagram of
the litium fluoride-beryllium fluoridezirconium fluoride system,” 166–180, 1968, with permission from
Elsevier.

For some properties (density, heat capacity), simple methods exist for correlating or obtaining an
approximate estimate of the true value, but for others (melting point, vapor pressure, viscosity,
and thermal conductivity), measurements are usually needed to proceed with confidence.

Phase transition thermodynamics govern melting point and species volatility (vapor pressure)
and reflect the behavior of a number of complex solution species. A priori predictions of these
phase transitions are difficult without some data to determine what species and interactions exist.
Interaction parameters based on lower order datasets (e.g., binary mixtures) can sometimes lead
to reasonable predictions for higher order datasets (e.g., ternary mixtures).

Viscosity and thermal conductivity are defined by constitutive equations for momentum and heat
transfer and depend upon the structure of the molten salt in a complex manner. For thermal
conductivity, trustworthy measurements are scarce because it is difficult to measure this property

A-20
at elevated temperatures in a manner that factors out the parallel heat transfer modes of thermal
radiation and convection. The transmissive and absorptive properties of the salt for thermal
radiation will be needed for detailed heat transfer analyses. With the exception of thermal
conductivity, all salt properties have demonstrated measurement methods that can be refined to
obtain high accuracy and reproducibility. A summary of these points is provided in Table 1,
which is from a previous study.

Table 1. Qualitative assessment of salt coolant properties database and predictive methods [12]

2.1.4 Modeling and Simulation

There is a significant opportunity to extend the knowledge base and strategically fill database
gaps by using modeling and simulation to improve our ability to understand and correlate salt
properties of all types. The understanding of property trends with compositional changes can be
of great value in guiding any measurement campaign.

Phase Transitions/Diagrams
It is typically acknowledged that ab initio prediction of phase transitions/equilibria is more
difficult than the prediction of stationary thermodynamic properties. A thermodynamic
framework that makes use of experimental data on simpler systems is the current method in use
[13]. However, recent progress in predicting phase transitions in molecular crystals suggests a
similar approach might be developed for ionic crystals [14–15].

Thermophysical and Transport Properties (density, heat capacity, viscosity, thermal

conductivity)
Atomistic models, and molecular dynamic simulations in particular, have recently been applied
with success to correlating molten salt physical properties [16–21]. The trend of property
changes with composition is sometimes revealed more accurately than the absolute value of the
property itself; but in some cases, remarkably accurate property estimates can be derived from
reliable atomistic interaction parameters.

2.2 ANALYTICAL CHEMISTRY CHALLENGES (INCLUDING MEASUREMENT

INSTRUMENTS AND ONLINE SENSORS)

Traditional destructive chemical and isotopic analysis methods were successfully applied to
irradiated and unirradiated molten salt samples during the Aircraft Nuclear Propulsion (ANP),
MSRE, and MSBR efforts. Although improvements in some archival sample analyses are likely
9

A-21
needed, the primary R&D need is for short-turnaround (“nearly instantaneous”) feedback from
nondestructive online analytical methods. Ongoing surveillance of the salt chemistry is required
to ensure the reactor is operating within safe bounds with regard to corrosion and other unwanted
chemical interactions; in this regard, measurement of the U3+/U4+ ratio and level of dissolved
alloy constituents (e.g., Cr, Fe, Ni) is the first priority. Electroanalytical and spectrophotometric
methods were tested in the lab during the MSRE and MSBR efforts on unirradiated and
irradiated MSRE salts, and their viability and value was firmly established. Refinement of these
techniques for application (and durability) in the complex MSR solution and its accompanying
high-radiation field was recognized as a significant R&D challenge. Examples of the results and
apparatus used at the end of the MSRE/MSBR program are depicted in Figures 6 and 7.

Figure 6. Electroanalytical results from measurement on MSRE fuel [22].

10

A-22
 Figure 7. MSRE-era electroanalytical and spectrophotometric apparatuses [22].

Spectrophotometry provides fundamental and specific information about species in solution and
has the potential to determine absolute values of most actinide constituents. At the end of the
MSBR program, higher spatial- and energy-resolution gamma radiation detection methods were
being tested. With the advent of much more capable radiation detection and imaging methods in
the intervening years, it is likely that both assay and surveillance by radiation detection methods
will find new applications. Since the end of the MSBR program, development of molten salt
electroanalytical and spectrophotometric methods and apparatuses has continued [23–29].

2.3 IRRADIATION AND FISSION PRODUCT CHEMISTRY CHALLENGES

2.3.1 Molten Salt Fission Product Chemistry

Noble gas behavior was a focus of early fission product studies because these neutron poisons
controlled breeding gain and could produce strong reactivity effects in thermal reactors. The
solubility of noble gases in fluoride salts was found to be very low; therefore, the near-complete
removal of these elements was achieved by use of low-porosity moderator graphite, non-wetting
salts, and straightforward methods to strip the noble gas from the salt.

11

A-23
For the other fission products, it is important to understand where they go and how they behave
to provide a radiation map to support dose control, maintenance, and accident-
mitigation/response, as well as to design features into the plant and components to account for
their behavior and manage their disposition. The detailed chemical behavior of many fission
products remained uncertain at the end of the MSBR program (1976), and not much more is
known today. What we know about their behavior in the MSRE salt is summarized in Figure 8
[30], and we believe other fluoride and chloride salts follow similar patterns of behavior. There
is also a body of evidence developed for fission product behavior during pyroprocessing of solid
fuels using chlorides and fluorides, but the redox conditions applied and salt compositions used
are significantly different from those under consideration as fluid fuels. The short-lived isotope
effects described below are largely absent from the pyroprocessing experience.

12

A-24
Cumulative Fission Product Yield (%)

Mass Number of Fission Product Decay Chain

gas soluble insoluble soluble-gas-soluble

Figure 8. Fission product classes identified during MSRE operation highlighted as groups in the
periodic chart (below) and the fission-yield curve (above). (Upper graphic source: “Fission product
yield,” Wikipedia, https://en.wikipedia.org/wiki/Fission_product_yield.)

13

A-25
Four classes of fission products were identified based on the MSRE experience: gaseous species,
soluble species, insoluble species, and sometimes-soluble species. The noble gases comprise the
gaseous species. The IA (alkali), IIA (alkaline earth), IIIA&B (rare earths and Al-group), IVB
(active refractories), and VIIB (halogens) comprise the soluble species. The transition metals
between Nb and Te comprise the insoluble “noble metal group.” The Nb, Te, and Zn periodic
columns comprise the sometimes-soluble species—their behavior changing with the redox state
of the salt.

How these insoluble and sometimes-soluble species transition from birth at an atomic level to an
insoluble phase or deposit is not completely understood, and several hypotheses still compete to
explain what occurs [31]. Although a major fraction of the fission products end up as salt-soluble
species (including iodine), the decay behavior of the fission product chain is quite complex in
many cases and can result in partition of short-lived species before the longer-lived isotope
emerges. This situation occurs for many of the noble metal fission products and for the key
isotopes of 137Cs, 99Mo, and 131I, as shown in Figure 9.

Figure 9. Exemplar short-lived fission products that change solubility behavior during decay in a
fuel circuit (typically < 1 minute salt recirculation residence time).

2.3.2 Barren-Salt Activation Product Chemistry: Tritium and Sulfur

Tritium is produced in MSRs by both ternary fission and a number of light-element neutron
reactions. With the exception of its production from the irradiation of lithium, the amounts
produced are similar to those produced in high-temperature reactors and light water reactors; and
its mobility through alloys at high temperatures is a problem that is shared with gas-cooled high-
temperature reactors. It is the use of lithium (and even lithium highly enriched in 7Li) in the fuel

14

A-26
salt or primary coolant that is attended by large amounts of tritium production (rivaling or
exceeding the amounts produced in heavy-water reactors) [32–34].

Although tritium is not particularly long lived and has only weak beta-particle emissions,
because of its mobility, it has the potential to create large amounts of low-level radioactive waste
if it reaches water in the environment or in the balance of plant. It must also be controlled to
levels safe for workers and the public. There is a wealth of information on the technology used to
control hydrogen and tritium transport, and many types of solutions can be envisioned, including
various barriers and trapping approaches. Because of the fluxing property of halide salts, it is
very unlikely that any type of film on salt-wetted alloy surfaces could serve as a barrier to
tritium.

During the operation of the MSRE, a significant fraction of the tritium produced permeated the
primary systems and was not controlled or trapped by the off-gas systems [35]. As part of the
post-MSRE studies, a fluoroborate coolant-salt-trapping modality was identified for the MSBR
fluoroborate secondary coolant [35–36]. Modeling and experimental studies on the behavior of
tritium in MSR systems have continued since the MSBR-era studies [37–39]. It was also
proposed that an oxygen-containing secondary salt, such as a nitrate mixture, would be able to
chemically trap tritium, as was demonstrated with the oxygen moiety in fluoroborates. The
ability to manipulate the chemistry in other fluoride or chloride salts (such as the behavior of
trace oxygen levels as dissolved or undissolved components) to provide tritium trapping is not
known.

In chloride fuel salts not heavily enriched in 37Cl, the activation of 35Cl will produce significant
quantities of 36S by (n,p) activation. Taube predicted sulfur levels in the thousands of ppm at
steady state in a molten chloride fast reactor composed of unenriched chlorides, and it is clear
that this production has both a redox effect and potentially a chemical corrosive/reactive effect
(akin to that associated with its periodic chart neighbor tellurium in group 6A) [40–44].

2.3.3 Solid Salt Irradiation Behavior

The behavior of solid salt during interim or long-term storage is an important issue because solid
halide salts can sustain considerable radiation damage and chemical change upon irradiation by
the alpha, beta, and gamma radiation produced by the decay of dissolved fission and activation
products. The production of metal centers and reactive halogens (e.g., F2) within the salt was
observed in discharged irradiation capsules and the stored MSRE fuel salt [45–46]. The
importance of distinguishing solid-salt irradiation effects from molten-salt irradiation effects was
clearly demonstrated during the testing that led up to the design and operation of the MSRE [47].

Testing of fuel and coolant fluoride salts by in-reactor irradiation confirmed that molten salt is
largely immune to irradiation damage—any primary radiolytic event is rapidly reversed by
recombination. No materials damage or chemical effects of beta, gamma, neutron, or alpha
radiation (or fission fragments) were observed in molten fluorides even when very high
(> 1 kw/cc) in-reactor fission power densities were applied. This testing also revealed that if the
MSRE salt was allowed to freeze and cool to temperatures below 200°C, significant radiolysis
and chemical change occurred very quickly, driven by the intense radiation associated with the

15

A-27
short-lived decay of fission products. Thus any study of irradiated salts must account for this
factor, and the domain where radiolysis occurs must be understood for any new salts.

Molten chloride salts have not seen the intense in-pile power densities proposed for molten salt
fast reactor concepts, so their immunity to radiation damage under these types of conditions will
need to be confirmed. Our understanding of radiolytic effects in chloride salts builds from
experience in using them as reprocessing media to dissolve solid nuclear fuels and perform
separations. Molten alkali chloride mixtures used for pyroprocessing exhibited good chemical
and radiation stability at the radiation levels associated with processing fuel with short decay
periods (3–6 months). Experiments are yet to be conducted to confirm the stability of molten
chlorides under the more intense conditions of irradiation as a fuel salt. Out-of-reactor studies
have also been conducted on solid fluoride and chloride salts using alpha, beta, gamma, and x-
ray irradiation [48–51].

There is another dimension to irradiation damage in solid salts that has not received much
attention. In the solid state, a multicomponent halide is a mixture of different crystal sizes and
types. The distribution of crystal sizes and types depends on the cooling history of the melt. A
solid boule of the unirradiated MSRE salt contained four major phases, as shown in Figure 10.

16

A-28
Figure 10. MSRE salt phase segregation on cooling [52].

17

A-29
The purpose of the study [52] depicted in Figure 10 was to bound how much uranium
segregation could occur during freezing. As irradiated salt is cooled to a solid, not only does the
fissile component distribute according to phase equilibrium and crystallization dynamics but the
fission products also distribute themselves according to their phase equilibria. For the MSRE and
MSBR salts, fission-product phase distributions received some study. In general, it was found (as
expected) that soluble fission products (e.g., Rb, Cs, Sr, Ba, Zr) within the same chemical/space
group (e.g., IA, IIA, IVA) already present in the salt (e.g., LiF, BeF2, ZrF4) would be substituted
into the phases of the major salt constituents. The rare earths were an exception. This class of
soluble fission product was not incorporated into a phase associated with the unirradiated fuel
salt.

The impact of this nonuniform distribution of fission products throughout the solid salt is that
different regions (and even micro-domains) will receive different radiation exposures. Since each
crystal also exhibits a distinct radiolysis behavior, taking account of this heterogeneity is
important, as is accurate thermal modeling, because the radiolysis is strongly influenced by
temperature. Some simplification occurs for the effect of long-energy-track radiation, such as
gamma rays; but the fraction of gamma energy deposited in the salt must also be accounted for
from both internal and external sources (e.g., neighboring tanks/sources).

2.4 MATERIALS COMPATIBILITY CHALLENGES

The identification of suitable core and container materials for molten salt service must take into
account a number of requirements and involves a number of trade-offs. The core and container
material must retain adequate strength, toughness, and creep resistance for the neutron exposure
that is accumulated; endure contact with molten salt over a range of temperatures without
excessive corrosion; be stable with regard to the external atmosphere surrounding the exterior of
the container (and tolerate whatever atmospheric in-leakage that forms the basis for safety
evaluation); and be able to be fabricated into the complex shapes of the primary loop (and
sometimes the secondary loop) of the reactor. For service in the secondary loop, the material
must also exhibit compatibility with a separate working fluid (e.g., steam, CO2, helium for a
power conversion system, or another high-temperature heat transfer agent), but the requirements
for neutron exposure durability are almost completely absent. Although durability under neutron
exposure is a very important property, it is not unique to molten salt systems. Because the
corrosion phenomenon in molten salts is distinctive, it is highlighted first.

2.4.1 Corrosion

There are a number of excellent reviews of the extensive corrosion studies on fluoride salts at
ORNL that led to the development of Hastelloy-N, its highly successful deployment in the
MSRE, and its optimization for the MSBR program [53]. For the reader less familiar with this
history, the following section summarizes some key points. There are parallels with the
chemistry in chloride melts to the principles outlined below for fluorides.

Unlike the case for more conventional oxidizing media, the products of oxidation of metals by
fluoride melts tend to be completely soluble in the corroding media [54]; hence, passivation is
precluded, and corrosion depends directly on the thermodynamic driving force of the corrosion
reactions [55]. Design of a chemically stable system using molten fluoride salts therefore

18

A-30
demands the selection of salt constituents that are not appreciably reduced by available structural
metals and the development of containers whose components are in near thermodynamic
equilibrium with the salt medium.

Examination of the free energies of formation for the various alloy components in Inconel or
Hastelloy-N shows that chromium is the most active of the metal components. Therefore, any
oxidizing reactions with nickel-base alloys are be expected to show selective attack on the
chromium constituent. Such oxidation and selective attack follow from reactions such as the
following [56] for the fuel salt.

Impurities dissolved in the melt attack chromium directly producing dissolved CrF2:

Cr + NiF2 → CrF2 + Ni ,

Cr + 2HF → CrF2 + H2 .

Oxide films on the metal are converted to corrosive dissolved species:

2NiO + ZrF4 → 2NiF2 + ZrO2 .

These reactions are followed by reaction of NiF2 with Cr cited above, and reduction of excess
UF4 to UF3:

Cr + 2UF4 ↔ 2UF3 + CrF2 .

Of course, in the case of a coolant salt with no fuel component, this reaction with uranium would
not be a factor.

The phenomenon described above falls into the class of “temperature-gradient mass transfer”
that occurs without a protective film of any sort and has some features in common with liquid
metal corrosion. When dissimilar materials contain molten salt, an activity gradient corrosion can
occur in the absence of any temperature-gradient driving force (or it can act in combination with
a thermal gradient). The generalized concept of temperature-gradient mass transfer is depicted in
Figure 11.

19

A-31
 Figure 11. Temperature-gradient mass transfer [53].

It was the thermally driven UF4/UF3 redox couple that was believed to be the primary source of
persistent corrosion in the fuel salt loop after impurity-driven corrosion had been played out, and
is depicted in Figure 12.

Figure 12. Persistent salt corrosion as a temperature-gradient-driven redox cycle [54].

Redox processes that are responsible for attack by molten fluoride mixtures on high-temperature
alloys result in selective oxidation of the least noble metal—typically chromium. Removal of
chromium from the alloy occurs primarily in the regions of the highest temperature and results in
the formation of discrete voids in the alloy [57]. These voids are not, in general, confined to the
grain boundaries in the metal but are relatively uniformly distributed throughout the alloy surface
that is in contact with the melt. The rate of corrosion has been measured and was found to be
controlled by the rate at which chromium diffuses to the surfaces undergoing attack [58].

20

A-32
With regard to our current understanding of the corrosion process, there are some important
gaps. The following questions can be posed:

• In the absence of a redox-sensitive soluble species (as was largely the case for the MSRE
2LiF-BeF2 coolant), what is the mechanism of corrosion in a polythermal loop?
• Is it better to have a substantial redox buffer (which will have some redox-temperature
coefficient) in a coolant salt and drive it to a reducing condition, or is it better to have no (or
very low levels of) redox-sensitive species in solution and maintain it in a reducing state by
other means?
• If a redox-sensitive species is a major constituent of the salt (as is the case for a uranium-
containing fuel salt), is it more advantageous to have high concentrations and maintain the salt
in a very reducing conduction (and thus have a very large buffer capacity), or is it better to
have a lower concentration of redox-sensitive species?
– Work in the 1950s during the ANP program indicated that higher uranium concentrations
in the salt accelerated the persistent polythermal corrosion mechanism.
• What is the effect of different salt compositions on polythermal redox-driven corrosion?
– Work in the late 1950s and 1960s indicated that very basic 1 uranium fuel salts (e.g., low
concentrations of UF4 dissolved in FLiNaK, a.k.a. “Fuelinak”) had much larger thermal
coefficients of corrosion than other uranium-containing salts that had fewer uncomplexed
fluoride anions (e.g., neutral 2LiF-BeF2 or acidic NaF-ZrF4 salts).
– The very high uranium- and plutonium-content salts proposed for some new MSR
concepts are significantly different from those tested in the past.
• Are iron-based alloys an option for halide salts if the salt is maintained under very reducing
conditions [59]? What are the salt redox window and compositional window for iron-based
alloys? What is the response of iron alloys to design-basis corrosive inputs compared with
nickel-based alloys?
2.4.2 Irradiation (Neutron) Damage Limits

The irradiation damage limits associated with the use of conventional nickel superalloys and
stainless and ferritic/martensitic steels have been studied extensively [60–62], although
incremental improvements in alloy performance and our understanding of it continue [63].

These conventional materials do not exhibit the long lifetimes typically desired for nuclear
components; this is an especially dramatic shortfall for fast-reactor applications. However, the
thermal neutron effects remain an important constraint limiting higher temperatures and
exposures for nickel-based superalloys. Figure 13 depicts the dual materials challenges posed for
MSR systems as higher temperatures and atomic displacements are sought.

1
Basic in the sense of the Lewis acid-base theory—possessing an abundance of electron donors (uncomplexed
fluoride anions in salts that are predominantly alkali halides such as FLiNaK).

21

A-33
 Figure 13. Dual challenges of higher temperatures and displacements
for MSR in-core structural materials [61].

The situation is not as challenging for container alloys or primary loop components outside the
core because they see considerably less intense radiation fields, but existing materials were still
seen as having some shortfalls at the end of the MSBR program. Whether these shorter material
lifetimes are suitable is an economic decision and depends on the particular system design.

22

A-34
 3. TECHNOLOGY AND APPLIED R&D NEEDS

In the most general terms, the technology and applied R&D needs highlighted in the previous
sections can be summarized as follows:

1. Create a systematic computational approach to salt selection that balances and

encompasses all neutronic and chemical factors.

2. Predict phase diagrams, solubilities, vapor pressures, and all relevant thermophysical and
transport properties.

3. Develop rugged and reliable in-line chemical and isotopic sensors in conjunction with the
development of out-of-pile and in-pile test stands.

4. Create a spatially and temporally accurate predictive model of fission/activation product

production coupled with the ability to accurately track the fate and transport of this
activity.

5. Develop a discriminating model of corrosion and materials compatibility for a variety of

salt redox conditions and salt compositions that permits the most favorable operating
envelope to be identified and the most appropriate salts and materials (or groups of
materials) to be discovered, developed, and designed.

6. Discover new materials and ways to use existing materials that can extend the high-
temperature and neutron flux/fluence boundaries that currently exist.

7. Conceive and formulate a fine-grained predictive understanding of radiolytic processes in

solid salts to permit in-pile testing, interim salt storage, and disposition of salt-based
wastes.

23

A-35
 4. REFERENCES

1. W. R. Grimes and D. R. Cuneo, “Molten salts as reactor fuels,” in Reactor Handbook vol.1
(Materials), ed. C. R. Tipton Jr., Interscience Publishers, New York, 1960, 2nd Edition, p. 425.
2. W. R. Grimes et al., “Chemical aspects of molten fluoride salt reactor fuels,” in Fluid-Fueled
Reactors, ed. J. A. Lane et al., Addison-Wesley, New York, 1958, chap. 12.
http://moltensalt.org/references/static/downloads/pdf/FFR_chap12.pdf
3. W. R. Grimes, Chemical Research and Development for the Molten-Salt Breeder Reactor,
ORNL/TM-1853, Oak Ridge National Laboratory, 1967.
http://web.ornl.gov/info/reports/1967/3445605995384.pdf
4. W. R. Grimes, “Molten Salt Reactor Chemistry,” Nuclear Applications and Technology 8(2), 137–
155 (1970). http://moltensalt.org/references/static/downloads/pdf/NAT_MSRchemistry.pdf
5. D. F. Williams and K. T. Clarno, “Evaluation of salt coolants for reactor applications,” Nuclear
Technology 163(3), 330–343 (2008). http://www.ans.org/pubs/journals/nt/a_3992
6. Y. Hirose, and Y. Takashima, “The concept of fuel cycle integrated molten salt reactor for
transmuting Pu+MA from spent LWR fuels,” paper FP163 at the GLOBAL 2001 Conference, Paris
France (2001).
http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/33/034/33034365.pdf?r=1
7. J. R. Keiser, Status of Tellurium-Hastelloy N Studies in Molten Fluoride Salts, ORNL-TM-6002, Oak
Ridge National Laboratory, October 1977. https://www.osti.gov/scitech/servlets/purl/7295251
8. V. V. Ignatiev et al., “Experimental investigation of tellurium corrosion of nickel-molybdenum alloys
in molten lithium-beryllium- and uranium-fluoride salts,” Atomic Energy 120(6), 2016.
https://link.springer.com/article/10.1007/s10512-016-0148-1
9. W. R. Grimes, “Molten fluoride mixtures as possible fission reactor-fuels,” Journal of the
Electrochemical Society 125(3), C161 (1978). https://www.osti.gov/scitech/servlets/purl/6741721
10. B. R. Harder et al., “Compatibility and processing problems in the use of molten uranium chloride—–
Alkali chloride mixtures as reactor fuels,” in Nuclear Metallurgy 15, P. Chiotti, ed. (1969).
https://www.osti.gov/scitech/biblio/4747246
11. R. E. Thoma et al., “Equilibrium phase diagram of lithium fluoride-beryllium fluoride-zirconium
fluoride system,” Journal of Nuclear Materials 27(2), 166 (1968).
12. D. F. Williams, Assessment of Candidate Molten Salt Coolants for the Advanced High Temperature
Reactor (AHTR), ORNL/TM-2006/12, Oak Ridge National Laboratory, 2006.
http://info.ornl.gov/sites/publications/Files/Pub57476.pdf
13. E. Capelli, O. Benes, and R. J. M Konings, “Thermodynamic assessment of the LiF–ThF4–PuF3–UF4
system,” Journal of Nuclear Materials 462, 43–53 (2015).
14. G. J. O. Beran, “Modeling polymorphic molecular crystals with electronic structure theory,”
Chemical Reviews 116, 5567−5613 (2016).
15. S. Hirata et al., “Ab initio molecular crystal structures, spectra, and phase diagrams,” Acc. Chem. Res.
47, 2721−2730 (2014).
16. D. Corradini et al., “Coordination numbers and physical properties in molten salts and their
mixtures,” Faraday Discuss. 190, 471 (2016).
17. Y. Ishii et al., “Thermal conductivity of molten alkali metal fluorides (LiF, NaF, KF) and their
mixtures,” The Journal of Physical Chemistry B 118, 3385−3391 (2014).

25

A-37
18. A. E. Gheribi et al., “Prediction of the thermophysical properties of molten salt fast reactor fuel from
first-principles,” Molecular Physics 112, 1306–1312 (2014).
19. F. Lantelme and H. Groult, Molten Salts Chemistry: From Lab to Applications (2013).
20. P. A. Madden and M. Salanne, “Thorium molten salts: Theory and practice,” Thorium Energy for the
World, pp. 111–116 (2016). http://link.springer.com/chapter/10.1007/978-3-319-26542-1_15
21. L. C. Dewan et al., “Molecular dynamics simulation of the thermodynamic and transport properties of
the molten salt fast reactor fuel LiF–ThF4,” Journal of Nuclear Materials 434 322–327 (2013).
22. M. W. Rosenthal, The Development Status of Molten-Salt Breeder Reactors, ORNL-4812, Oak Ridge
National Laboratory, August 1972. http://web.ornl.gov/info/reports/1972/3445604501750.pdf
23. H. A. Laitinen, “History of electroanalytical chemistry in molten-salts,” ACS Symposium Series 390,
417–428 (1989).
24. D. K. Corrigan et al., “Advances in electroanalysis, sensing and monitoring in molten salts,” Faraday
Discuss. 190, 351–366 (2016).
25. M. M. Tylka et al., “Method development for quantitative analysis of actinides in molten salt,”
Journal of the Electrochemical Society 162(9), H625–633 (2015).
26. B. Kelleher et al., “Observed redox potential range of Li2BeF4 using a dynamic reference electrode,”
Nuclear Technology 195, 239 (2016).
27. G. Duran-Klie et al., “Dynamic reference electrode development for redox potential measurements in
fluoride molten salt at high temperature,” Electrochimica Acta 195, 19–26 (2016).
28. J. Zhang, “Electrochemistry of actinides and fission products in molten salts—Data review,” Journal
of Nuclear Materials 447, 271–284 (2014).
29. K. Song et al., “Spectroscopic studies of lanthanides ion in high-temperature molten salt,” Applied
Spectroscopy Reviews 50(8), 654–669 (2015).
30. E. L. Compere et al., Fission Product Behavior in the MSRE, ORNL-4865, Oak Ridge National
Laboratory, 1975. http://web.ornl.gov/info/reports/1975/3445605698493.pdf
31. R. J. Kedl, The Migration of a Class of Fission Products in the MSRE, ORNL/TM-3884, Oak Ridge
National Laboratory, 1973. https://www.osti.gov/scitech/biblio/4471292-migration-class-fission-
products-noble-metals-molten-salt-reactor-experiment
32. R. B. Briggs, “Tritium in molten salt reactors,” Reactor Technology 14(4), 335 (1972).
33. R. B. Briggs, “Tritium in molten salt reactors—Reply,” Reactor Technology 15(1), 79 (1972).
34. R. B. Briggs, Calculation of the Tritium Distribution in the MSRE, CF-70-7-13, Oak Ridge National
Laboratory, 1970, and Additional Calculations of the Distribution of Tritium in the MSRE, ORNL/CF
71-7-8, Oak Ridge National Laboratory, 1971.
35. G. T. Mays, Distribution and Behavior of Tritium in the Coolant-Salt Technology Facility,
ORNL/TM-5759, Oak Ridge National Laboratory, 1977.
http://moltensalt.org/references/static/downloads/pdf/ORNL-TM-5759.pdf
36. L. Maya, “Identification of polyborate and fluoropolyborate ions in solution by Raman spectroscopy,”
Inorganic Chemistry 15(9), 2180 (1976).
37. R. B. Briggs and C. W. Nestor, A Method for Calculating the Steady-State Distribution of Tritium in
a Molten-Salt Breeder Reactor Plant, ORNL/TM-4804 Oak Ridge National Laboratory, 1975.
http://web.ornl.gov/info/reports/1975/3445605505619.pdf

26

A-38
38. J. D. Stempien et al., “An integrated model of tritium transport and corrosion in fluoride salt-cooled
high-temperature reactors (FHRs)—Part I: Theory and benchmarking,” Nuclear Engineering and
Design 310, 258–272 (2016).
39. D. Zhang, “Permeation of hydrogen in Hastelloy C-276 at high temperature,” Fusion Science and
Technology 67, 681 (2015).
40. E. Ianovici and M. Taube, Chemical State of Sulphur Obtained by the 35Cl (n, p) 35S Reaction during
in-pile Irradiation, Swiss Institute for Reactor Technology (Wurenlingen) Report, EIR-267 (1974).
http://moltensalt.org/references/static/downloads/pdf/EIR-267.pdf
41. M. Taube and J. Ligou, “Molten plutonium chlorides fast breeder reactor cooled by molten uranium
chloride,” Annals of Nuclear Science and Engineering 1, 277 (1974).
42. E. Ianovici and M. Taube, “Chemical behavior of radiosulphur obtained by 35Cl (n, p) 35S during in-
pile irradiation,” Journal of Inorganic and Nuclear Chemistry 37, 2562 (1977).
43. M. Furrer, “Chemical behaviour of radiosulphur obtained by 35Cl (n, p) 35S during in-core irradiation
of alkali-chloride melts,” Journal of Inorganic and Nuclear Chemistry 39, 1086 (1977).
44. M. Taube, Fast Reactors Using Molten Chloride Salts as Fuel: Final Report (1972–1977), Swiss
Institute for Reactor Technology (Wurenlingen) Report, EIR-332 (1978).
http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/13/648/13648304.pdf
45. D. F. Williams et al., A Descriptive Model of the MSRE after Shutdown: Review of FY 1995 Progress,
ORNL/TM-13142, Oak Ridge National Laboratory, 1996.
https://www.osti.gov/scitech/servlets/purl/230260
46. D. F. Williams, A. S. Icenhour, L. D. Trowbridge, G. D. Del Cul, L. M. Toth, “Radiolysis studies in
support of the remediation of the Molten Salt Reactor Experiment,” Transactions of the American
Nuclear Society 81, 89–90 (1999).
47. W. R. Grimes, Radiation Chemistry of MSR System, ORNL/TM-500, Oak Ridge National Laboratory,
1963. http://web.ornl.gov/info/reports/1963/3445605493999.pdf
48. L. M. Toth and L. K. Felker, “Fluorine generation by gamma radiolysis of a fluoride salt mixture,”
Radiat. Eff. Def. Solids 112, 201–210 (1990).
49. A. B. Lidiard, “The radiolysis of alkali halides—The nucleation and growth of aggregates,”
Zeitschrift fur Physikalische Chemie 206, 219–248 (1998).
50. L. Tandon, Radiolysis of Salts and Long-Term Storage: Issues for Both Pure and Impure PuO2
Materials in Plutonium Storage Containers, LA-13725-MS, Los Alamos National Laboratory, 2000.
http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/32/016/32016719.pdf
51. D. K. Veirs et al., Evidence of Corrosive Gas Formed by Radiolysis of Chloride Salts in Plutonium-
Bearing Materials, SRNL-STI-2010-00051, Savannah River National Laboratory, 2010.
http://sti.srs.gov/fulltext/SRNL-STI-2010-00051.pdf
52. H. A. Friedman and R. E. Thoma, “MSRE fuel fractionation experiments,” in Reactor Chemistry
Division Annual Progress Report for Period Ending January 31, 1965, W. R. Grimes et al., eds.,
ORNL-3789, Oak Ridge National Laboratory, April 1965.
http://web.ornl.gov/info/reports/1965/3445605716008.pdf
53. J. W. Koger, “Chromium depletion and void formation in Fe-Ni-Cr alloys during molten salt
corrosion and related processes,” in Advances in Corrosion Science and Technology, vol. 4, M. G.
Fontana and R. W. Staehle, eds. (1974).

27

A-39
54. W. D. Manly et al., “Metallurgical problems in molten fluoride systems,” Progress in Nuclear
Energy, Series IV 2, 164–179 (1960).
55. J. H. DeVan and R. B. Evans III, Corrosion Behavior of Reactor Materials in Fluoride Salt Mixtures,
ORNL/TM-328, Oak Ridge National Laboratory, September 1962.
http://web.ornl.gov/info/reports/1962/3445605700621.pdf
56. W. R. Grimes, Chemical Research and Development for the Molten-Salt Breeder Reactor,
ORNL/TM-1853, Oak Ridge National Laboratory, 1967.
http://web.ornl.gov/info/reports/1967/3445605995384.pdf
57. L. S. Richardson, D. E. Vreeland, and W. D. Manly, Corrosion by Molten Fluorides, ORNL-1491,
Oak Ridge National Laboratory, March 1953.
http://web.ornl.gov/info/reports/1953/3445603532247.pdf see also
http://web.ornl.gov/info/reports/1969/3445605136769.pdf
58. J. H. Devan, Effect of Alloying Additions on Corrosion Behavior of Nickel-Molybdenum Alloys in
Fused Fluoride Mixtures, University of Tennessee Thesis, August 1960, see also ORNL/TM-2021,
Oak Ridge National Laboratory, 1969. https://www.osti.gov/scitech/servlets/purl/4789147
59. J. R. Keiser et al., The Corrosion of Type 316 Stainless Steel to Li2BeF4, ORNL/TM-5782, Oak Ridge
National Laboratory, April 1977. https://www.osti.gov/scitech/servlets/purl/4789147
60. H. E. McCoy, Status of Materials Development for Molten Salt Reactors, ORNL/TM-5920, Oak
Ridge National Laboratory, January 1978. https://www.osti.gov/scitech/servlets/purl/5195742
61. S. J. Zinkle and G. S. Was, “Materials challenges in nuclear energy,” Acta Materialia 61 (2013) 735–
758.
62. S. J. Zinkle, K. A. Terrani, and L. L. Snead, “Motivation for utilizing new high-performance
advanced materials in nuclear energy systems,” Current Opinion in Solid State and Materials Science
20, 401–410 (2016).
63. L. Jiang et al., “M2C andM6C carbide precipitation in Ni-Mo-Cr based superalloys containing
silicon,” Materials and Design 112, 300–308 (2016).

28

A-40
 APPENDICES

A. Molten Salt Reactor Technical Working Group Descriptions of Reactor Systems under
Consideration
B. 1969 Molten Salt Reactor Experiment Video
C. Molten Salt Reactor Excerpts from “An Account of ORNL’s 13 Research Reactors” by
Murray Rosenthal

A-41
 APPENDIX A.
THUMBNAIL SUMMARIES OF REACTOR CONCEPTS
UNDER CONSIDERATION BY MSR
TECHNICAL WORKING GROUP

A-43
 Elysium Molten Salt Reactor Concept FLiBe Energy
• Rated electric output 1,000 MW e • Liquid Fluoride Thorium Reactor (LFTR) – Thorium fuel cycle
• Chloride fuel salt • Graphite moderated, thermal spectrum, LiF-BeF2-UF4 fuel
• Fast spectrum neutron flux • Freeze valve and drain tank based decay heat rejection
• Core outlet temperature of >600°C • Described in EPRI technical report – 3002005460, October 2015
• Core inlet temperature of ~500°C − https://www.epri.com/#/pages/product/000000003002005460/
• Employs drain valve − https://aris.iaea.org/PDF/LFTR.pdf

www.elysiumindustries.com/ http://flibe-energy.com/

Kairos Power Moltex Energy

• Solid-fueled fluoride-salt-cooled concept with demonstration • UK company – Stable Salt Reactor (SSR) concept

A-3
A-45
targeted by 2030 235
• NaCl-(Actinide)Cl3 fuel salt (LEU, <15% U)
• Uses ceramic TRISO coated fuel particles in a pebble fuel
element design. − Redox stabilized with Zr/ZrCl couple
• Proposes nuclear air combined cycle for flexible, responsive • NaF-ZrF4 coolant salt
power generation
− ZrF2/ZrF4 redox buffer
− co-fired with natural gas or hydrogen for peaking • Natural circulation of fuel salt within vented fuel pins
power
• Steel cladding similar to that used in Dounreay SFR
• PB-FHR described at
https://aris.iaea.org/PDF/Mk1%20PB-FHR%20ARIS%20final.pdf
https://aris.iaea.org/PDF/SSR.pdf
http://kairospower.com/
http://www.moltexenergy.com/

Fig. A-1. Thumbnail descriptions of Molten Salt Reactor Technical Working Group reactor concepts.
 TerraPower Terrestrial Energy USA
• Molten Chloride Fast Reactor (MCFR) uses chloride-based fuel • Integral Molten Salt Reactor concept
salt
− Denatured U/Pu fuel cycle and LEU fluoride salt fuel
• Employs reflector containing constituents such as LiF, NaF, BeF2.
− Reduce dose to containment
− Graphite moderator
− Improve neutron economy
− Sealed primary unit with integrated pumps, heat
− Reduce radiation level within shielding exchangers, and shutdown rods
− Integral heat exchangers • Seven year reactor vessel, moderator, and fuel salt lifetime
• Integral heat exchangers
− Twin reactor module configuration – one module
swapped every seven years
http://terrapower.com/news/terrapowers-continuing-innovation https://aris.iaea.org/PDF/IMSR400.pdf
https://aris.iaea.org/Publications/SMR-Book_2016.pdf
Thorcon Power Transatomic

A-4
A-46
• Fuel salt comprised of NaF and BeF2 with 12mol% actinide • 5% enrichment LEU fuel salt, such as LiF-UF4 eutectic

− 80% of actinide is thorium, 20% uranium (~20% 235U) • Zirconium hydride in movable rods

• Fuel salt and moderator replaced every four years in twin reactor
modular configuration http://www.transatomicpower.com/wp-content/uploads/2015/04/TAP-
https://aris.iaea.org/PDF/ARISThorCon9.pdf White-Paper-v2.1.pdf

http://thorconpower.com/ http://www.transatomicpower.com/

Fig. A-2. Thumbnail descriptions of Molten Salt Reactor Technical Working Group reactor concepts (continued)
 APPENDIX B.
1969 MOLTEN SALT REACTOR
EXPERIMENT VIDEO

A-47
Oak Ridge National Laboratory. 1969. “The Molten Salt Reactor Experiment.” Video.
Available: https://www.youtube.com/watch?v=tyDbq5HRs0o.

B-3

A-49
APPENDIX C. MOLTEN SALT REACTOR EXCERPTS FROM “AN
ACCOUNT OF ORNL’S 13 RESEARCH REACTORS” BY
MURRAY ROSENTHAL

(Complete document at http://info.ornl.gov/sites/publications/files/Pub20808.pdf)

A-51
AIRCRAFT NUCLEAR PROPULSION PROGRAM

In 1946, the Air Force initiated a program to develop a nuclear-powered airplane and contracted
with the Fairchild Engine and Airplane Corporation to evaluate its feasibility. Fairchild set up a
project in an unused building near the K-25 gaseous diffusion plant (now the East Tennessee
Technology Park) in Oak Ridge, Tennessee. Oak Ridge National Laboratory (ORNL) provided
assistance.

There was considerable skepticism, and before getting more deeply involved, the Atomic Energy
Commission (AEC) asked staff at the Massachusetts Institute of Technology to perform an
independent evaluation. A summer project at Lexington, Massachusetts, concluded after initial
doubts that developing nuclear-powered aircraft would cost as much as $1 billion and take 15 y
but was feasible.

So under Cold War pressure, the AEC directed ORNL to establish the Aircraft Nuclear
Propulsion (ANP) Program. It grew to be the Laboratory’s largest activity for a number of years.
General Electric (GE) eventually replaced Fairchild and moved its program to Evendale, Ohio.
But about 180 people remained in Oak Ridge and joined ORNL’s program, which by 1951 filled
a big building at the Oak Ridge Y-12 Plant (now the Y-12 National Security Complex).

Two approaches to nuclear propulsion were considered. In the direct cycle, air for the jet engine
would pass through the reactor and be directly heated to 1,700°F. In the indirect cycle, a molten
metal would flow through a reactor and then through a heat exchanger where it would heat the
air. GE undertook to develop the direct cycle and ORNL undertook development of the indirect
cycle.

Aircraft Reactor Experiment Development

ORNL proposed building a small reactor to guide its development program. Funded by the AEC, this
reactor was called the Aircraft Reactor Experiment (ARE). To permit high temperature, BeO was chosen
as the moderator. The core was to be a stack of hexagonal blocks with vertical passages for the fuel
elements and molten sodium coolant. Because of the time needed for production of the BeO blocks, they
were ordered before the design was completed. The blocks stacked up to form the core are shown in
Figure C-1. The fuel elements themselves were to be long, thin stainless-steel tubes containing UO2.
Details are given in Table C-1.

C-3

A-53
 Figure C-1. Stacked BeO moderator blocks for the ARE.
An extensive research and development program was undertaken in chemistry, materials, high-
temperature sensors, and other areas, and components such as pumps and valves were tested in
loops. Protecting the aircraft’s crew and components and the public from radiation was critical.
Consequently, ORNL launched a broad program on the penetration of gamma rays and neutrons,
including measurements of the shielding properties of a variety of materials and the development
of calculation methods. That led to reactors built just for shielding studies, about which more
will be said later.

As the ARE design progressed, however, two serious questions arose. One came from physics
calculations that showed that the cross section of xenon-135 at elevated temperatures could cause
a positive temperature coefficient in a high-power solid-fuel reactor. The other was doubts about
the reliability of “spaghetti-like” fuel elements operating at the temperatures that would be
necessary.

An alternate idea arose in which a uranium compound would be dissolved in a molten salt and
solid fuel elements avoided. The salt would flow through the core, where it would be heated by
fissions, and then pass through a heat exchanger. The xenon gas was expected to evolve from the
salt and be removed, but even if it did not, much of it would be outside the core in the external
system. So the fuel rods were replaced by a molten salt.

After considering other salts, the choice of carrier salt was a “relatively” low melting
temperature (950°F) mixture of sodium and zirconium fluorides into which UF4 was dissolved.
The fluoride salts when solid look somewhat like green table salt, but as seen in Fig. C-2, when
molten they resemble water. Alkali fluorides are among the most stable of compounds, which
means they do not corrode stainless steels and similar alloys. They also have very low vapor
pressures, and when molten, they are not damaged by radiation.

C-4

A-54
Figure C-2. ORNL photograph of molten salt, featured on the cover of the British Science Journal
in June 1969. Used by permission; (Science Journal was incorporated into New Scientist,
www.newscientist.com, in 1971).
The salt would flow through the core in the passages that were designed to hold the solid
elements, but in a single pass the velocity would have been too low for turbulent flow. This was
dealt with by having the fuel make multiple passes at higher velocity. The core with the multiple-
pass tubes is seen in Figure C-3. The salt emerging from the core was pumped through a finned-
tube heat exchanger, where it was cooled by helium, which was in turn cooled by water, as
shown schematically in Figure C-4.

C-5

A-55
 Figure C-3. Elevation view of the ARE core. (ORNL Drawing 16336).

Figure C-4. ARE flow diagram.

Heat generated in the moderator and reflector was removed in a sodium system that was similar
to the salt system. For safety, the sodium system had two independent circuits, and the fuel salt
system contained a standby pump and heat exchanger to ensure continued operation. Sump-type
centrifugal pumps with overhung impellers were used, and the sumps served as expansion tanks.

C-6

A-56
The reactor was designed to operate with the fuel salt at around 1500°F and at a power of 1 to 3
MW of heat.

Table C-1. ARE parameters

Power (heat) 2.5 MW
Fuel salt 53% NaF, 41% ZrF4, 6% UF4
(93% enriched)
Fuel melting point 990°F
Fuel inlet temperature 1225°F
Fuel outlet temperature 1580°F (1620° at peak)
Fuel flow rate 45 gal/min
Moderator BeO hexagonal blocks, 3.6 in.
across × 6 in. high with 1.25 in.
coolant passages
Coolant Helium to water
Container Inconel
Critical November 3, 1954
Shut down November 12, 1954

The salt and sodium pipes and all major components had double walls to provide a space in
which helium was circulated to maintain even temperatures and to provide leak monitoring.
Control and safety rods in the core and fission chambers in the reflector were also cooled by the
circulating helium.

Electrical heaters around the core and the piping raised the temperature before molten salt was
added. All salt and sodium-containing systems were made from Inconel, a high-temperature
nickel- base alloy.

A mill-type building was constructed in Melton Valley near the homogeneous reactors to house
the ARE. All of the reactor systems were located below grade in concrete-lined pits.

ARE Operation

Operation of the ARE began in early November 1954. First the system was heated with the
electrical heaters. Then it was filled with uranium-free carrier salt. And finally, fuel salt as
2NaF·UF4 was added in increments from a heated pot that emptied into the pump bowl.

The fuel concentrate had a higher melting temperature than the carrier salt and tended to freeze
in the fuel addition system. Heating the transfer line and getting the fuel concentrate to flow was
the biggest operating problem of the reactor, and much of three days was spent in clearing plugs.

Overheating of the line with a welding torch caused a small hole which, however, did not prevent
the loading from being completed.

Physics experiments were run at low power for several days, and then the power was raised in
steps to full power over 30 hours. The ARE ran very smoothly. However, small amounts of
radioactive gas leaked into the reactor compartment from the hole in the loading line and set off
C-7

A-57
alarms. That was handled by hooking up a construction compressor that exhausted air from the
compartment and blew it into an uninhabited area about 1000 ft away.

The temperature coefficient of reactivity was negative as expected, and the power level could be
raised or lowered by changing the speed of the blower that cooled the heat exchanger. The total
power extracted in the fuel salt and sodium reached 2.5 MW with a steady-state fuel temperature
of 1580°F, but the temperature exceeded 1620°F in transients. At steady conditions, the
temperature rise across the core was 355°F.

The last experiment conducted was a measurement of xenon retention during a 25-hour run at
full power. It indicated that no xenon remained in the salt. After all scheduled tests were
completed, the reactor was cycled between full and low power 21 times within 12 hours. The
operating plan was to run the ARE for 100 hours; however, when that target was reached, only 9
days of operation had been at power. Running longer was considered, but the experiment was
shut down on schedule because of concern about weakness in the overheated fill line. That
turned out to be a wise decision, because the line opened completely 5 days later and released
radioactive gas into the reactor compartment.

The Air Force was pleased with the performance of the ARE and brought Pratt and Whitney
Aircraft Company aboard to develop the indirect cycle power plant. ORNL began the design of a
compact 60 MW reactor. And in spite of growing skepticism about success and the recognition
that missiles might substitute for bombers, industrial and political support kept the national
program going. But it was killed in March 1961 soon after John Kennedy took office.

Thus ORNL’s ANP program came to an end, but in its 12-year run, it greatly expanded
knowledge of the chemistry and technology of molten salts and made advances in materials,
shield design, and other areas that enlarged the Laboratory’s ability to undertake new projects.

MOLTEN-SALT REACTOR EXPERIMENT

With the ARE having shown the feasibility of molten-salt fuel, ORNL persuaded the AEC to
fund a study of molten-salt power reactors. Two concepts were evaluated, both graphite
moderated and based on the U-233/thorium fuel cycle. In one, the uranium and thorium were in
the same salt. In the other, a thorium salt that formed a fertile blanket was kept separate from the
fuel salt by a graphite barrier. The single-fluid concept was simpler, but a net breeding gain
appeared to be possible in a two-fluid reactor.

The results of the study were encouraging. However, the AEC was still funding the
homogeneous reactor program as well as a liquid-bismuth reactor program at Brookhaven
National Laboratory, and it had more fluid-fuel reactors than it wanted. So a task force was
created to evaluate all three programs. It concluded that the molten-salt reactor “had the highest
probability of achieving technical feasibility.” This led to approval of a program at ORNL to
develop molten-salt power reactors and to the demise of Brookhaven’s program and, as
mentioned earlier, the homogeneous reactor program.

The molten-salt development program proceeded well, and by the end of 1959, the Laboratory
felt justified in proposing a small reactor to investigate the technologies needed for civilian
power. To keep the reactor simple and inexpensive, it had a single region like a converter, but the
C-8

A-58
salt did not contain thorium and in that sense was similar to the core of a breeder. (Later a
concept was developed in which a virtual two-region reactor was created with a single salt by
reducing the moderation in the outer region.) The proposed reactor, called the Molten-salt
Reactor Experiment, was approved by the AEC. Design began in 1960, and construction started
early in 1962. Details are given in Table C-2, and a sketch of the reactor is shown in Figure C-5.

Table C-2. MSRE parameters

Power (heat) 7.4 MW
Fuel salt 65% Li7F, 29.1% BeF4,
5% ZrF4, 0.9 % UF4
Fuel melting point 813°F
Fuel inlet temperature 1,175°F
Fuel outlet temperature 1,225°F
Fuel flow rate 400 gal/min
Coolant salt 66% Li7F, 34% BeF4
Moderator Graphite
Container Hastelloy-N (68% Ni,
17% Mo, 7% Cr, 5% Fe)
Critical on 235U June 1, 1965
Critical on 233U October 2, 1968
Shut down December 1969

C-9

A-59
 Figure C-5. Sketch of the MSRE. (ORNL-LR-DWG 61097R).
The fuel salt chosen for the MSRE was a mixture of the fluorides of lithium-7, beryllium, and
zirconium selected to have good physical and nuclear properties. The melting point was 840°F.
Fluoride salts do not wet graphite, so the MSRE was designed with the bare-graphite core shown
in Figure C-6. It was 54 in. in diameter and made up of 2 in. square bars of specially produced
low-porosity graphite. Flat passages for the salt were machined into the surfaces of the bars.

C-10

A-60
 Figure C-6. Graphite core of the MSRE.
The molten salt flowed from the core to a heat exchanger where it gave up heat to another
fluoride salt that carried it to an air-cooled radiator. As shown in Figure C-7, the salts were
stored below the reactor in heated vessels, from which they could be forced up into the various
systems by gas pressure. “Freeze valves” formed by air-cooled flattened sections of pipe retained
the salts in the core. Figure C-8 is a photograph of the core, fuel pump, and heat exchanger. The
radiator with its doors open is seen in operation in Figure C-9.

C-11

A-61
Figure C-7. Schematic of the MSRE. (ORNL-LR-DWG 63-1209R).

C-12

A-62
 Figure C-8. MSRE core, fuel pump, and heat exchanger in the primary system containment tank.
(ORNL Photo 67051-64).

Figure C-9. MSRE air-cooled radiator with its door open.

The salt pumps were at the high points in the system and like those on the ARE, were sump type
and performed several functions. The sumps served as expansion tanks and were the free
surfaces for the salt systems. A spray of salt into the gas space above the sump allowed noble
gases to escape and be conducted to charcoal beds. In addition, a “sampler-enricher” tube leading
from above provided access to the salt in the sump. With it, capsules of frozen fuel-salt
concentrate could be lowered into the sump on a cable for small fuel additions, or thimbles could

C-13

A-63
be lowered to collect salt samples.

A simple processing facility was installed in a cell near the drain tanks. It contained a tank in
which salt could be sparged with hydrogen fluoride to remove water and oxygen or with fluorine
to remove uranium.

The reactor and all of the components that contacted salt were made of INOR-8, a nickel-base
alloy developed by ORNL and the International Nickel Company for use in molten-salt reactors.
It was later commercialized as Hastelloy-N and is still in use today. All salt piping and vessels
were electrically heated.

Careful provisions drawing on HRT experience were made for maintenance, in particular the
design and layout of components so that they could be removed from above with long-handled
tools. A 16 in. thick shield surrounding the reactor reduced activation of components by
neutrons. Effective procedures were devised to control the spread of radioactivity.

Construction of the MSRE was completed in 1964, and critical experiments were run in mid-
1965. To bring the reactor critical, part of the enriched uranium was added to carrier salt in the
drain tank in the form of a UF4-LiF eutectic. Then the final additions were made 88 grams at a
time using the sampler-enricher.

Approach to full power began early in 1966, but within a few hours, plugs developed at several
points in the fuel off-gas system. Apparently oil vapors that leaked past a gasket in the pump
were polymerized by heat and radiation and deposited in small passages. Only a few grams of oil
were involved, but three months was required to investigate and remedy the problem.

After full-power operation was resumed, the reactor ran well with only a few small problems
until the blades in one of the main blowers that cooled the radiator broke up. Redesigned blowers
were procured and installed, but other problems, including recurrences of the off-gas plugs, kept
interfering with operation until mid-December when an uninterrupted run of 30 days at full
power was achieved. During it the reactor was seen to be stable and load-following as a result of
the strong negative temperature coefficient.

The 30-day operation was followed by progressively longer runs; one lasted for 6 months, during
which the behavior of fission products was investigated. Xenon-135 poisoning was found to be
well below what it would have been if all of the xenon had remained in the core. Between runs,
surveillance specimens were removed from the core for examination, tests made of containment
and instrumentation, and maintenance operations performed.

Even though numerous problems were encountered during the early months of power operations,
they were always overcome in spite of high radiation fields, and the capability for sustained
operation was demonstrated. So at the end of the 6-month run, the fuel salt was transferred to the
processing tank and the uranium removed from it by treatment with fluorine. The emerging UF6
was passed though a bed of hot sodium fluoride pellets that trapped volatile impurities, then was
routed to a series of canisters in the high bay and absorbed on NaF pellets. The fuel-salt
fluorination took only 47 hours, and the UF6 that it produced was clean enough that the canisters
could be handled without shielding.

C-14

A-64
Next, some U-233 that had been made in a production reactor was added to the carrier salt,
making the MSRE the first reactor ever to operate on U-233. Since Glenn Seaborg, chairman of
the AEC, had led the team that first created U-233, he was invited to come to Oak Ridge and
start up the reactor. He is seen in Figure C-10 raising the control rod. Although U-233 has a
smaller delayed-neutron fraction than U-235, the reactor was quite stable and the dynamic
behavior close to predictions. Later a small amount of plutonium was added to show the fissile-
fuel flexibility of molten-salt systems.

Figure C-10. Glenn Seaborg at the MSRE control panel.

The MSRE experience had increased confidence in the performance and practicality of molten-
salt systems, and the Laboratory was anxious to move ahead with further development that
would lead to demonstration of a breeder reactor. Since the budget was quite limited, the reactor
was shut down at the end of 1969 and its funds devoted to development, particularly of a
processing system that would be needed for breeding in a single-fluid reactor. Two problems that
turned up during operation of the MSRE were also investigated. One was the confirmation that
tritium produced in the fuel salt by neutron reactions with lithium, while small in amount, had
diffused through the heat- exchanger tubes and would need to be intercepted in future reactors.
The second was the discovery that although general corrosion was less than expected, fission-
product tellurium had caused fine cracks on Hastelloy-N surfaces.

In spite of the technical success of the MSRE, the AEC was strongly committed to sodium-
cooled fast breeder reactors and in 1973 gave orders to shut down the molten-salt program. A
low-budget extension was soon granted in which solutions of the tellurium and tritium problems
were demonstrated, but the program was stopped completely 2 years later. So ORNL’s molten-
salt adventure came to an end after 20 years.

C-15

A-65
MOLTEN-SALT REACTORS GENERAL REFERENCES

E. S. Bettis, R. W. Schroeder, G. A. Cristy, H. W. Savage, R. G. Affel, and L. F. Hemphill. 1957. “The

Aircraft Reactor Experiment—Design and Construction.” Nuclear Science and Engineering 2:
804–25.
E. S. Bettis, W. B. Cottrell, E. R. Mann, J. L. Meem, and G. D. Whitman. 1957. “The Aircraft Reactor
Experiment—Operation.” Nuclear Science and Engineering 2: 841–43.
D. W. Cardwell, and P. N. Haubenreich. 1971. Indexed Abstracts of Selected References on Molten-Salt
Reactor Technology. Rep. No. ORNL-TM-3595, Oak Ridge National Laboratory, Oak Ridge,
TN.
W. K. Ergen, A. D. Callihan, C. B. Mills, and D. Scott. 1957. “The Aircraft Reactor Experiment—
Physics.” Nuclear Science and Engineering 2: 826–40.
H. G. MacPherson, ed. 1958. “Part II: Molten-Salt Reactors.” In Fluid Fuel Reactors, 565–697. Reading,
MA: Addison-Wesley, 1958.
H. G. MacPherson. 1985. “The Molten Salt Reactor Adventure.” Nuclear Science and Engineering 90:
374–80.
“Molten-Salt Reactor Experiment.” 1966. Oak Ridge National Laboratory, Oak Ridge, TN. Pamphlet.
“A Review of Molten Salt Reactor Technology.” 1970. Nuclear Applications and Technology 8 (2): 102–
219. Special issue comprising a set of ten papers including M. W. Rosenthal, P. R. Kasten, and R.
B. Briggs, “Molten-Salt Reactors: History, Status, and Potential”; Paul N. Haubenreich and J. R.
Engel, “Experience with the Molten-Salt Reactor Experiment”; and W. R. Grimes, “Molten-Salt
Reactor Chemistry.”
M. W. Rosenthal, P. N. Haubenreich, and R. B. Briggs. 1972. The Development Status of Molten-Salt
Breeder Reactors. Rep. No. ORNL-4812, Oak Ridge National Laboratory, Oak Ridge, TN.
Molten-Salt Reactor Program Semiannual Progress Report for Period Ending [August 1958–August
1976]. Rep. Nos. ORNL-2474–ORNL-5132. Oak Ridge National Laboratory, Oak Ridge, TN.
http://web.ornl.gov/info/reports/1976/3445604452553.pdf., (last in a series of semiannual
progress reports that began in 1957; they cover the MSRE and all aspects of the related R&D
program.

C-16

A-66
A-68
 Molten Salt Chemistry Workshop

APPENDIX B: A DEVELOPER’S PERSPECTIVE ON THE NECESSITY OF SALT

PROPERTY RESEARCH
During the initial stages of the design process, it is vital to establish the capability to rapidly evaluate the
base performance of different salt systems in order to select the optimal composition for a given
application. This requires both extensive experimental data for the salt system in question and correlations
and predictive tools that can give insight into the fundamental thermophysical salt properties for systems
that have yet to be explored or have been explored only briefly.

Figure 1 outlines the salt properties data required for input to the framework of a MSR design project.
Although this discussion centers on the design phase of an MSR, the same information will be required
for the safety analysis. The following thermophysical properties must be known as a function of
temperature, both in the normal operating reactor temperature range and at higher temperatures to
account for criticality excursions:

• Melting point (phase diagram, and associated sensitivity with composition)

• Heat capacity
• Thermal conductivity
• Viscosity
• Density
• Vapor pressure
• Gibbs free energy of formation (and the associated redox potential)

Salt Cleanup
Design Team

Composition
Melting Point Melting Point (sets Cold Leg Temperature)
Vapor Pressure Density (Pumps)
Vapor Pressure (Pumps)
Melting Point Viscosity
Density Heat Capacity
Composition Thermal Conductivity
Physics Enrichment Chemistry Emissivity Thermo-
hydraulics
Team Team Team

Burnup Composition Salt Potential

Composition
Temperature

Materials
Team

Figure 1. Organization schematic in a molten salt reactor design project.

These properties must be known for the initial salt composition (at the first startup for the reactor), as well
as during reactor operation, when the salt composition will contain increasing concentrations of fission

Appendix B B-1
Molten Salt Chemistry Workshop

products as a function of burnup (time and power at which the reactor operates), burnout of the fissile
species (for example, 235U → 2 fission products), and an increase in bred species (for example, 238U →
239
Pu).

It will be necessary for national laboratories, universities, and industry to measure these properties where
gaps exist in historically published data. It is likely that each reactor design will require its own unique
data set, based upon its starting chemistry, temperature range, and burnup profile.

The following summarizes the rationale for the properties requirements for each of the reactor design
groups shown in Figure 1.

Physics
This evaluation begins with neutronics, shown as “physics” in Figure 1. If a fuel salt cannot achieve and
maintain criticality, it should not be explored as an option. The multiplication factor (reactivity) of any
given MSR, whether using a thermal or a fast neutron spectrum, depends upon the volumetric density of
fissile isotopes. This is dependent upon the composition of the salt (e.g., percent uranium halide in a
halide salt mix), the enrichment of the fissile isotope (e.g., 235U% within the total uranium), and the
density (molar volume) of the salt mixture (where the salt is a mixture of an actinide salt and other
nonfissile “carrier” salt species, e.g., NaCl, NaF, LiF, or others). The density of the salt mixture is
temperature-dependent, i.e., the salt becomes less dense with increasing temperature. The resulting
negative temperature-coefficient of reactivity is the basis of an MSR’s reactivity control, as well as an
important passive safety feature. Reactor designers need accurate density data to accurately design the
reactor size and to ensure that reactivity is adequately controlled.

The salt density of a mixture is not always an additive combination of the sum of the pure components.
For some salts, the total volume of the salt mixture exceeds the sum of the individual component
volumes. This molar volume anomaly is not completely understood but is thought to be due to complex
ion formation and a resulting open structuring of the salt structure. This stresses the need for experimental
measurements of density, since theoretical calculations are not necessarily predictive of the density of salt
mixtures. Determinations of the density of the salt in and out of a reactor (with and without fission
products) will be necessary to understand reactor control requirements during startup and shutdown.

Thermohydraulics
Once it has been shown that criticality can be achieved and maintained, it must then be determined that
the heat transfer properties of the given system allow for heat to be effectively transported so that power
may be produced. Heat exchanger design requires knowledge of salt density, heat capacity, viscosity, and
thermal conductivity as a function of salt temperature and composition. Thermal radiation transport
coefficients (e.g., absorption, transmission, and fluorescence) and other radiation transport effects may
need to be measured and understood to predict radiative heat transport. Fluid dynamics and heat transfer
properties will determine the size and geometry of the heat exchanger given a designed power output. The
heat exchanger must be designed so as to prevent salt freeze-up at the outlet (cold leg) of the heat
exchanger. Optimal heat exchanger designs will minimize fuel salt volume and therefore reduce the cost
of the reactor.

Materials
Corrosion of reactor structural alloys is strongly affected by the electrochemical potential of the salt
mixture. It is necessary to keep the reactor structural materials at sufficiently cathodic (reductive)
potentials to provide thermodynamic immunity toward corrosion of structural alloys. Specific chemical
effects of fission products that grow in during reactor operation and potentially attack structural alloys

B-2 Appendix B
 Molten Salt Chemistry Workshop

need to be investigated. For example, tellurium, a fission product, was seen in the MSRE to diffuse into
structural grain boundaries and produce cracks by leaching chromium from the Hastelloy N alloy The
cracking could be reduced by controlling the redox potential of the salt. Materials performance as a
function of salt composition, electrochemical potential, and temperature must therefore be studied.

Auxiliary systems (cleanup)

Fission products will have to be managed at some level in all MSR designs. Off-gas hold-up and disposal
systems will need to be designed for gaseous fission products. Some mechanism to remove or manage
insoluble fission products will need to be in place. Some MSR designs may require a soluble fission
product removal system. The composition of the fuel/burnup salt, electrochemical potentials, and vapor
pressures must be known for the successful deployment of these systems.

Down-selection of candidate salts will be accomplished by each MSR design team according to its own
selection criteria for salt properties. Salts with high melting points, or high neutron capture cross sections,
will likely be eliminated from further consideration; and testing of other properties of these salt systems
may not be useful. Fluid-fueled systems require a uranium or plutonium driver as a constituent of the
primary salt—thorium can be included as a fertile constituent in some designs

A quality control program must be instituted to guarantee the quality of the data to the level required.

Access to pure salts for laboratory studies

There is currently no supply chain for many actinide salts. Most researchers will find it necessary to
synthesize their own salts using a variety of synthesis pathways. Although ORNL established processes
and practices for purifying and handling fluoride salts for research and for and the MSRE (J. H. Shaffer in
ORNL-4616; Bamberger in Advances in Molten Salt Chemistry 3 (1975): “Experimental techniques in
molten fluoride chemistry”), there is no national standard for testing and reporting the purity of the salt.
Different purification processes and procedures will be needed for other (e.g., chloride) salts. Although
much, but not all, work can be done with surrogate salt systems, the properties and chemistry of the
actinides will also need to be studied directly.

To enable progress in MSR salt properties research, a central repository is strongly encouraged for both
actinide and surrogate salts. This would provide universities, national labs, and industry with consistent
salts that would allow better cross-comparison between research groups. Site licensing controls exist for
natural uranium salts, limiting the amount of uranium salt that can be stockpiled at a given research
location. A central repository would allow researchers access to adequate quantities of salt, while keeping
their inventories below regulatory limits. Additionally, handing plutonium salts requires extremely
rigorous and specialized controls for any isotopic composition. Special consideration could be given to
doing work with non-fissile plutonium as a way to avoid the nuclear safety and security burden associated
with the effort. Disposal pathways must be identified as a part of any site license, and central repository
disposal services would be most helpful.

Salt purification
Most salts are hydroscopic and contain significant amounts of water. Thus, best practices for dewatering
salts should be studied, published, and standardized. Best practice procedures for establishing clean salt
may involve some or all of the following: handing the salt in an inert atmosphere, heating the salt, and/or
purging the salt with a purifying gas (e.g., HF, HCl or CCl4), perhaps in the presence of hydrogen gas.
Once a salt is purified, standard methods of handling and assaying the purity of the salt economically and
routinely should be developed and standardized.

Appendix B B-3
Molten Salt Chemistry Workshop

Glovebox-compatible advanced instrumentation

All handling and analysis of salts outside of closed or inert-atmosphere-blanketed containers needs to be
avoided, and inert-atmosphere gloveboxes should be used. The dimensions of the analysis equipment
must be such that it can be located within the volume of the glovebox. Electronics must be located outside
the glovebox, since those experienced in the field are aware that standard electronics tend to fail (arcing)
in the inert atmosphere. Easy operation and maintenance by the operator in gloves must be designed into
the instruments.

The following characterization methods are not meant to be a complete list of candidate methods. These
serve as known examples of methods currently in use or being developed for future use:

Melting point determinations

Melting points can be determined in a variety of ways, most commonly with differential scanning
calorimetry (DSC). Standardization of operating parameters for DSC operation needs to be established.
For example, recommendations should be made for temperature ramp rate, number of freeze/thaw cycles,
purity of cover gas, purge time of the cover gas, and sample crucible material. Other melting point
measurement methods may be used, and the operating parameters for these methods should be established
as well.

Density and viscosity

There are a variety of methods for measuring the density and viscosity of a molten salt. The best
method(s) and procedures need to be determined by the molten salt community. The Archimedes-bob Pt-
method is the reference method most commonly employed for density, and various capillary and
rotational methods have been used for viscosity measurement.

Electrochemical potential
Salt electrochemical potential can be calculated with quality Gibbs free energy data, which are
determined from experimental thermochemical data. Electrochemical potential can also be measured
directly in an electrochemical cell that uses a stable reference electrode.

If a given salt system has been shown to effectively produce and transport heat, the feasibility of the
design will then rest on the safety and operational performance of the reactor system, as stated earlier.

The following are some additional questions that can inform this safety and operations evaluation:

• Do the properties of the salt change as a result of the presence of fission products, transuranics, and
radiation, to the point where heat production and transport are significantly affected?

• What are the solubility limits of species that will be present over the course of operations? Is there a
risk that some of these species could precipitate out in ways that would be unfavorable?

• What is the free energy of formation of the species contained in the salt in relation to the reactor’s
structural materials, and can it inform the extent of corrosion that will occur within the system?

• What is the vapor pressure of the species that will be present over the course of operations, to
sufficiently inform considerations for liquid-gas interfaces of all systems containing a fuel salt and its
byproducts?

B-4 Appendix B
 Molten Salt Chemistry Workshop

APPENDIX C: WORKSHOP AGENDA

MOLTEN SALT CHEMISTRY WORKSHOP: Defining Technology and Applied R&D Needs for
Molten Salt Chemistry
Building 5100, Joint Institute for Computational Sciences
Oak Ridge National Laboratory
Chairs: David F. Williams and Phillip F. Britt (Oak Ridge National Laboratory)
Event contact: David F. Williams, 865-574-8853 (office), 865-255-4737 (mobile);
[email protected]
Monday, August 10, 2017
Time Event Lead
7:15–8:00 a.m. ORNL Visitor Check-in and Workshop Registration at Bldg. 5100 Lobby

Alan Icenhour, Associate Laboratory

Director, Nuclear Science and Engineering
Welcome and Workshop Charge (Joint Directorate, ORNL
8:00–8:30 a.m. Institute for Computational Sciences [JICS]
Lecture Hall) – working breakfast John Herczeg, Deputy Assistant Secretary,
Nuclear Technology Research and
Development, DOE-NE

Technical Working Group Perspective on Nick Smith, MSR Technical Working Group
8:30–9:00 a.m.
Workshop Representative

Vic Maroni, Argonne National Laboratory

9:00–9:40 a.m. Challenges in MSR Chemistry
Distinguished Fellow

9:40–10:20 a.m. Group Photo and Coffee Break All

Steve Zinkle, Nuclear Engineering, University

10:20–11:00 a.m. Irradiation Effects in MSR Systems
of Tennessee–Governor’s Chair
James Keiser, Materials Science and
11:00–11:40 a.m. Materials Compatibility in MSR Systems
Technology Division, ORNL
11:40 a.m.–12:00 p.m. Separation Options and Waste Forms Dave Williams (ORNL)
Working Lunch: (JICS Lecture Hall)
• 12:00–12:20—serving line in Lobby
Dave Williams and Phil Britt, Workshop Co-
12:00–1:00 p.m. o Bring lunch to Lecture Hall
chairs (ORNL)
• 12:30–1:00 p.m.
o Working Group (WG) Charge

Parallel Working Group Sessions WG Co-Chairs Locations

• WG#1: Physical Chemistry and Salt WG#1: Alex
1. Rm 130
Properties Navrotsky, Mark
Williamson 2. Rm 262
• WG#2: Analytical Chemistry (includes WG#2: Sam Bryan, (Boardroom)
measurement instruments and online 3. Rm 128 (Lecture
Sheng Dai
1:00–5:30 p.m. sensors) WG#3: Bill DelCul, Hall)
4. Bldg 5700, Rm
• WG#3: Molten Salt Fission Product Tina Nenoff
E104*(Monday
Chemistry and Solid Salt Radiolysis WG#4: James Keiser,
only)
Preet Singh
• WG#4: Materials Compatibility
WG#5: Bobby 5. Roams among all
• WG#5: Computational Chemistry and Sumpter, Brian groups
Materials Science Wirth

Appendix C C-1
Molten Salt Chemistry Workshop

Monday, August 10, 2017 (continued)

Time Event Lead

Dinner for All, WG status from chairs (JICS

5:45–7:00 p.m. Onsite Working Dinner — WG Status
Lobby)

WG Members led by chairs / designated

7:00–9:00 p.m. Parallel Working Group Sessions (cont.)
location above
9:00 p.m. Adjourn Day 1 All

Tuesday, August 11, 2017

Time Event Lead

Working Group Sessions – Working Breakfast WG Co-Chairs Locations

Breakfast served in JICS Lobby WG#1: Alex
1. Rm 130
• WG#1: Physical Chemistry and Salt Navrotsky, Mark
2. Rm 262
Properties Williamson
(Boardroom)
WG#2: Sam Bryan,
• WG#2: Analytical Chemistry Sheng Dai 3. Rm 128 (Lecture
8:00–10:15 a.m. Hall)
• WG#3: Molten Salt Fission Product WG#3: Bill DelCul,
4. Rm 134 (Learning
Chemistry and Solid Salt Radiolysis Tina Nenoff
room)
WG#4: James Keiser,
• WG#4: Materials Compatibility
Preet Singh 5. Bldg. 4100,
• WG#5: Computational Chemistry and WG#5: Bobby Rm 203
Materials Science Sumpter, Brian Wirth

10:15–10:40 a.m. Break and Reconvene in JICS Auditorium All

Report
10:40–11:00 a.m. • WG#1: Physical Chemistry, Salt WG#1: Alex Navrotsky, Mark Williamson
Properties

Report
11:00–11:20 a.m. WG#2: Sam Bryan, Sheng Dai
• WG#2: Analytical Chemistry

Report
11:20–11:40 a.m. • WG#3: Molten Salt Fission Product WG#3: Bill DelCul, Tina Nenoff
Chemistry and Solid Salt Radiolysis
Report
11:40 a.m.–12:00 p.m. WG#4: James Keiser, Preet Singh
• WG#4: Materials Compatibility
Working Lunch – Report
(Serving line in JICS Lobby, return to
Lecture Hall) WG#5: Bobby Sumpter, Brian Wirth
12:00–1:00 p.m.
• WG#5: Computational Chemistry and
Materials Science
1:00–5:30 p.m. WG Sessions (cont.) WG members/designated location as above

5:30–6:30 p.m. Onsite Working Dinner – WG Status Dinner for all, WG chair status (JICS Lobby)

6:30–9:00 p.m. WG Preparation for Final Workshop Plenary All

9:00 p.m. Adjourn Day 2 All

C-2 Appendix C
 Molten Salt Chemistry Workshop

Wednesday, August 12, 2017

Time Event Lead

Report
8:00–8:30 a.m. • WG#1: Physical Chemistry, Salt WG#1: Alex Navrotsky, Mark Williamson
Properties

Report
8:30–9:00 a.m. WG#2: Sam Bryan, Sheng Dai
• WG#2: Analytical Chemistry

Report
9:00–9:30 a.m. • WG#3: Molten Salt Fission Product WG#3: Bill DelCul, Tina Nenoff
Chemistry and Solid Salt Radiolysis
Report
9:30–10:00 a.m. WG#4: James Keiser, Preet Singh
• WG#4: Materials Compatibility
10:00–10:30 a.m. Break All

Report and Discussion

10:30–11:00 a.m. • WG#5: Computational Chemistry and WG#5: Bobby Sumpter, Brian Wirth
Materials Science
John Herczeg, DOE-NE
11:00–11:30 a.m. Final Discussion and Closing Remarks
Dave Williams, ORNL

11:30 a.m. Adjourn Day 3* All

Report Team Agenda (same as previous breakout rooms)

1:30–5:00 p.m. Writing Team Tuesday Breakout Rooms

*You are welcome to lunch at the ORNL Cafeteria on Wednesday. A workshop lunch is not provided on Wednesday.

Appendix C C-3
 Molten Salt Chemistry Workshop

APPENDIX D: WORKSHOP PARTICIPANTS

Registered participants for April 10–12, 2017, Molten Salt Chemistry Workshop at ORNL

Name Organization Role

Abney, Carter Oak Ridge National Laboratory Analytical Chemistry (scribe)
Besmann, Ted University of South Carolina Computational Chemistry/Materials
Betzler, Ben Oak Ridge National Laboratory Computational Chemistry/Materials (scribe)
Bresee, James Department of Energy, Office of Nuclear Observer
Energy
Britt, Phil Oak Ridge National Laboratory Meeting co-chair
Bryan, Sam Pacific Northwest National Laboratory Analytical Chemistry
Calderoni, Pattrick Idaho National Laboratory Analytical Chemistry
Chidambaram, University of Nevada, Reno Materials Compatibility
Dev
Chvala, Ondrej University of Tennessee Computational Chemistry/Materials
Cooper, Thurman Terrestrial Energy USA Ltd Physical Chemistry and Salt Properties

Appendix D D-1
 Molten Salt Chemistry Workshop

Name Organization Role

Czerwinski, Ken TerraPower Physical Chemistry and Salt Properties
Dai, Sheng Oak Ridge National Laboratory Analytical Chemistry (WG co-chair)
DelCul, Bill Oak Ridge National Laboratory Molten Salt Fission Product Chemistry (WG
co-chair)
Dingreville, Remi Sandia National Laboratories Computational Chemistry/Materials
Dixon, David The University of Alabama Computational Chemistry/Materials
Dunstan, Don Energy Storage Solutions Materials Compatibility
Fredrickson, Guy Idaho National Laboratory Molten Salt Fission Product Chemistry
Hammelman, US Nuclear Regulatory Commission Observer
James
Hanson, Michael Elysium Industries Physical Chemistry and Salt Properties
Henager, Charles Pacific Northwest National Laboratory Computational Chemistry/Materials (WG
co-chair)
Herczeg, John Department of Energy, Office of Nuclear Speaker/Observer
Energy
Horton, Linda Department of Energy, Basic Energy Observer
Sciences
Icenhour, Alan Oak Ridge National Laboratory Speaker/Observer
Keiser, James Oak Ridge National Laboratory Speaker/ Materials Compatibility (WG
co-chair)
Kim, Hojong The Pennsylvania State University Molten Salt Fission Product Chemistry
Kruizenga, Alan Sandia National Laboratories Materials Compatibility
Kung, Stephen Department of Energy, Office of Nuclear Observer
Energy
Longford, Des PHDS Co. Analytical Chemistry
Margulis, Claudio University of Iowa Computational Chemistry/Materials
Maroni, Vic Argonne National Laboratory Speaker/Physical Chemistry and Salt Properties
(scribe)
Marsden, Kenneth Idaho National Laboratory Molten Salt Fission Product Chemistry
Massie, Mark Transatomic Power Corporation Observer/ Industry Representative
Mayes, Richard Oak Ridge National Laboratory Molten Salt Fission Product Chemistry (scribe)
Mays, Gary Oak Ridge National Laboratory Physical Chemistry and Salt Properties
McDuffee, Joel Oak Ridge National Laboratory Molten Salt Fission Product Chemistry
McLaughlin, David Westinghouse Electric Co. Ltd Physical Chemistry and Salt Properties
McNeese, James Lawrence Livermore National Laboratory Molten Salt Fission Product Chemistry
Meeks, Noah Southern Company Services Observer/ Industry Representative
Memmott, Brigham Young University Physical Chemistry and Salt Properties
Matthew
Motsegood, Perry Argonne National Laboratory Analytical Chemistry
Muralidharan, Oak Ridge National Laboratory Materials Compatibility
Govindarajan
Navrotsky, University of California–Davis Physical Chemistry and Salt Properties (WG
Alexandra co-chair)

D-2 Appendix D
 Molten Salt Chemistry Workshop

Name Organization Role

Nenoff, Tina Sandia National Laboratories Molten Salt Fission Product Chemistry (WG
co-chair)
Pal, Uday Boston University Physical Chemistry and Salt Properties
Pheil, Ed Elysium Industries Physical Chemistry and Salt Properties
Qualls, Lou Oak Ridge National Laboratory Observer
Raiman, Stephen Oak Ridge National Laboratory Materials Compatibility (scribe)
Rappleye, Devin Lawrence Livermore National Laboratory Analytical Chemistry
Robertson, Sean Transatomic Power Corporation Observer/ Industry Representative
Rohatgi, Upendra Brookhaven National Laboratory Observer
Salay, Michael US Nuclear Regulatory Commission Observer
Scarlat, Raluca University of Wisconsin Physical Chemistry and Salt Properties
Shaka, A. J. University of California–Irvine Analytical Chemistry
Sham, Sam Argonne National Laboratory Materials Compatibility
Singh, Preet Georgia Institute of Technology Materials Compatibility (WG co-chair)
Smith, Nicholas Southern Company Services Speaker/ Materials Compatibility
Sorensen, Kirk Flibe Energy Physical Chemistry and Salt Properties
Stoddard, Michael Alpha Tech Research Corp Molten Salt Fission Product Chemistry
Computational Chemistry/Materials (WG
Sumpter, Bobby Oak Ridge National Laboratory
co-chair)
Sun, Xaidong University of Michigan Materials Compatibility
Terrani, Kurt Oak Ridge National Laboratory Materials Compatibility
Todd, Terry Idaho National Laboratory Molten Salt Fission Product Chemistry
Tortorelli, Peter Oak Ridge National Laboratory (retired) Materials Compatibility
Wells, Daniel Electric Power Research Institute Physical Chemistry and Salt Properties
Williams Dave Oak Ridge National Laboratory Meeting co-chair
Williams, Joseph US Nuclear Regulatory Commission Observer
Physical Chemistry and Salt Properties (WG
Williamson, Mark Argonne National Laboratory
co-chair)
Willit, James Argonne National Laboratory Molten Salt Fission Product Chemistry
Wilson, Dane ThorCon Power Materials Compatibility
Computational Chemistry/Materials (WG
Wirth, Brian University of Tennessee
co-chair)
Zinkle, Steve University of Tennessee Speaker/Materials Compatibility
Ziock, Klaus Oak Ridge National Laboratory Molten Salt Fission Product Chemistry

Appendix D D-3
This report was prepared as an account of work sponsored by an agency of the United States
Government. Neither the United States Government nor any agency thereof, not any of their
employees, makes any warranty, express or implied, or assumes any legal liability or
responsibility for the accuracy, completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use would not infringe privately owned
rights. Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its
endorsement, recommendation, or favoring by the United States government or any agency
thereof. The views and opinions of authors expressed herein do not necessarily state or reflect
those of the United States Government or any agency thereof.