technologies charles forsberg, guiqiu (tony) zheng, ronald ...xylenepower.com/nt fission fusion...

Full Terms amp Conditions of access and use can be found athttpswwwtandfonlinecomactionjournalInformationjournalCode=unct20

Nuclear Technology

ISSN 0029-5450 (Print) 1943-7471 (Online) Journal homepage httpswwwtandfonlinecomloiunct20

Fusion Blankets and Fluoride-Salt-Cooled High-Temperature Reactors with Flibe Salt CoolantCommon Challenges Tritium Control andOpportunities for Synergistic DevelopmentStrategies Between Fission Fusion and Solar SaltTechnologies

Charles Forsberg Guiqiu (Tony) Zheng Ronald G Ballinger amp Stephen T Lam

To cite this article Charles Forsberg Guiqiu (Tony) Zheng Ronald G Ballinger amp Stephen TLam (2019) Fusion Blankets and Fluoride-Salt-Cooled High-Temperature Reactors with Flibe SaltCoolant Common Challenges Tritium Control and Opportunities for Synergistic DevelopmentStrategies Between Fission Fusion and Solar Salt Technologies Nuclear Technology DOI1010800029545020191691400

To link to this article httpsdoiorg1010800029545020191691400

copy 2019 The Author(s) Published withlicense by Taylor amp Francis Group LLC

Published online 11 Dec 2019

Submit your article to this journal View related articles

View Crossmark data

Fusion Blankets and Fluoride-Salt-Cooled High-TemperatureReactors with Flibe Salt Coolant Common Challenges TritiumControl and Opportunities for Synergistic DevelopmentStrategies Between Fission Fusion and Solar Salt TechnologiesCharles Forsberg Guiqiu (Tony) Zheng Ronald G Ballinger and Stephen T Lam

Massachusetts Institute of Technology Department of Nuclear Science and Engineering 77 Massachusetts AvenueCambridge Massachusetts 02139

Received September 12 2019Accepted for Publication November 7 2019

Abstract mdash Recent developments in high-magnetic-field fusion systems have created large incentives to developflibe (Li2BeF4) salt fusion blankets that have four functions (1) convert the high energy of fusion neutrons intoheat for the power system (2) convert lithium into tritiummdashthe fusion fuel (3) shield the magnets againstradiation and (4) cool the first wall that separates the plasma from the salt blanket Flibe is the same coolantproposed for fluoride-salt-cooled high-temperature reactors that use clean flibe coolant and graphite-matrixcoated-particle fuel Flibe is also the coolant proposed for some molten salt reactors (MSRs) where the fuel isdissolved in the coolant The multiple applications for flibe as a coolant create large incentives for cooperativefusion-fission programs for development of the underlying science design tools technology (pumps instrumenta-tion salt purification materials tritium removal etc) and supply chains Other high-temperature molten saltsare being developed for alternative MSR systems and for advanced Gen-III concentrated solar power (CSP)systems The overlapping characteristics of flibe salt with these other salt systems create significant incentives forcooperative fusion-fission-solar programs in multiple areas

We describe the fission and fusion flibe-cooled systems what has created this synergism what isdifferent and the same between fission and fusion in terms of using flibe and the common challenges Wereview (1) the characteristics of flibe salts (2) the status of the technology (3) the options for tritiumcapture and control in the salt heat exchangers and secondary heat transfer loops and (4) the coupling topower cycles with heat storage The technology overlap between flibe systems and other high-temperatureMSR and CSP salt systems is described This defines where there are opportunities for cooperativeprograms across fission fusion and CSP salt programs

Keywords mdash Salt-cooled reactors fusion flibe salt fluoride-salt-cooled high-temperature reactor moltensalt reactor

Note mdash Some figures may be in color only in the electronic version

I INTRODUCTION

This paper describes the fission and fusion applicationsfor flibe salt with overlapping areas of science technologychallenges and future directions Figure 1 shows the pro-posed affordable robust and compact (ARC) fusion

E-mail cforsbermiteduThis is an Open Access article distributed under the terms of theCreative Commons Attribution-NonCommercial-NoDerivatives Lice-nse (httpcreativecommonsorglicensesby-nc-nd40) which permitsnon-commercial re-use distribution and reproduction in any mediumprovided the original work is properly cited and is not altered trans-formed or built upon in any way

NUCLEAR TECHNOLOGYcopy 2019 The Author(s) Published with license by Taylor amp Francis Group LLCDOI httpsdoiorg1010800029545020191691400

1

system12 the flibe fusion blanket and the proposedfluoride-salt-cooled high-temperature reactor (FHR) Thestartup company Commonwealth Fusion Systems (CFS) isdeveloping the ARC fusion concept in cooperation with theMassachusetts Institute of Technology (MIT) PlasmaScience and Fusion Center Eni SpA (ENI)a is workingwith MIT to address many of the longer-term challengesassociated with this fusion system The startup companyKairos Power is developing the FHR Significant effortson the FHR are underway at MIT the University ofCalifornia at Berkeley the University of Wisconsin atMadison the Georgia Institute of Technology Oak RidgeNational Laboratory and many other institutions In Chinathe Shanghai Institute of Applied Physics (SINAP) isbuilding a 2-MW(thermal) molten salt reactor (MSR)There are smaller efforts to develop these technologies inEurope and elsewhere Major experimental facilities usingflibe salt are under design in the United States and China

Work on flibe blankets for various magnetic and iner-tial fusion systems goes back to the 1960s (Refs 3 through10) but there has been little work in the fusion communityfor the last 15 years Advances in technology (discussedbelow) indicate many future machines will use flibeblankets Work on flibe for reactor systems goes back tothe 1950s and the early development of MSRs Most of therecent work on flibe has been done for the FHR (UnitedStates) and MSR (China) within the last decade

There are other developments that impact technologydevelopments for flibe salt systems There is significantwork with multiple startup companies and other organiza-tions to develop MSRs with other fluoride salts that havephysical properties similar to flibe The EuropeanCommunity1112 has a project to develop a fast-spectrum

MSR using a lithium heavy-metal fluoride salt with many ofthe same tritium challenges associated with flibe salt Theproposed next-generation concentrated solar power (CSP)systems use a sodium-potassium-magnesium-chloride saltas the heat transfer and heat storage salt13 This salt hasa melting point near 400degC an expected peak operatingtemperature of ~750degC and operates under chemicallyreducing conditions to minimize corrosion The temperaturerange and chemically reducing operating conditions aresimilar to flibe salt resulting in technology overlap in thedevelopment of pumps valves instrumentation powercycles and other areas of salt technology Last there isongoing work to develop molten chloride fast reactors(MCFRs) with fuel dissolved in the salt that operate atsimilar temperatures

Most of these developments have occurred in the last5 years thus only recently at MIT and elsewhere haveresearch and development efforts begun to be integratedacross different salt power system technologies Thispaper has the goal to provide an integrated perspectiveon flibe technology and its coupling to other work onhigh-temperature salt systems to accelerate cooperativedevelopment and commercialization of these technologies

II ARC FUSION POWER PLANT

Recent advances in manufacturing have enabled thelarge-scale production of rare-earth barium copper oxide(REBCO) superconducting tape This in turn has enabledthe building of large magnets with double the magneticfield of previous large superconducting magnets Inmagnetic fusion systems the system size scales as oneover the fourth power of the magnetic field Doubling themagnetic field can reduce the size of the fusion machine byan order of magnitude for the same power output resulting

Fig 1 (a) ARC fusion system (b) flibe fusion blanket for ARC and (c) FHR fission system

a ENI is one of the big-five western oil companies

2 FORSBERG et al middot HIGH-TEMPERATURE REACTORS WITH FLIBE SALT COOLANT

NUCLEAR TECHNOLOGY middot VOLUME 00 middot XXXX 2019

in a fusion machine where the power density increases byan order of magnitude with large reductions in cost per unitof power output This is the basis for the ARC fusionconcept12 It is a one-time improvement With REBCOmagnets the maximum magnetic fields that can be gener-ated are limited by the strength of the materials holding themagnets togethermdashnot by the ability to generate magneticfields Figure 1a shows the conceptual design of the ARCThe ARC is the first fusion machine that proposes to usethese new magnetic materials but the expectation is thatmost future magnetic fusion systems will use REBCO orsome future equivalent magnet material to reduce systemsize and cost

The ARC requires a fusion blanket that (1) convertsenergy from fusion neutrons into heat for the powersystem (2) converts lithium into tritium as fast as thetritium is consumed (3) shields the magnets againstradiation and (4) cools the first wallvacuum vessel(VV) that contains the plasma Many different designsof fusion blankets have been proposed with differentcoolants14 High-magnetic-field fusion relative to othermagnetic fusion machines changes the blanket designrequirements in two ways First the power densitiesincrease by an order of magnitude The much higherpower densities make it difficult to cool solid blanketsthat provide shielding and breed tritium Second thehigher magnetic fields create large incentives to usea coolant in the blanket that has low electrical conduc-tivity relative specifically to lithium-lead-cooled

systems to reduce interactions between the flowingcoolant and the magnetic fields These considerationslead to the use of a liquid salt flibe blanket

The ARC is a tokamak where the machine is a torus(Fig 1) Figure 2 shows the cross section There are massiveheat loads from the fusion process on the first wall thusmost of the cold flibe salt from the heat exchanger flowsthrough multiple channels to provide first-wall coolingbefore entering the large flibe blanket and returning to themain heat exchanger The nominal design characteristics ofthe ARC and the flibe blanket are summarized in Table I

Flibe is proposed because it is the best liquid salt forbreeding sufficient tritium fuel for a self-sufficient fusionmachine15 The beryllium in the salt multiplies thenumber of neutrons via a (n 2n) reaction and neutronadsorption by lithium produces tritium Isotopicallyenriched lithium with 90 6Li is used to maximize tri-tium production The breeding ratio is estimated at 11that is for every tritium atom consumed in the fusionreaction 11 tritium atoms are created A minimumbreeding ratio of 1 is required for the fusion reactor tobe self-sustaining in tritium production assuming no tri-tium losses in the system or radioactive decay of tritium

The proposed first-wall structure is a double-wall(Fig 3) structure with a sacrificial layer facing the plasmabacked by Inconel-718 as the structural material a flibecooling channel a second wall with beryllium backed byInconel and the main flibe tank Flibe from the wall-cooling channel empties into the main tank About 25

Fig 2 ARC machine and flibe blanket

HIGH-TEMPERATURE REACTORS WITH FLIBE SALT COOLANT middot FORSBERG et al 3


of the heat from the system is deposited within thisstructure This includes heat from (1) the plasma on theplasma-facing wall (2) neutron heating from slowingdown of neutrons by the beryllium flibe and Inconeland (3) various other nuclear reactions The berylliummetal layer acts as a neutron multiplier via a high-energy neutron yielding two neutrons Part of the blanketstructure includes a diverter that acts as a drain to removehigh-energy plasma from the system that containsdeuterium tritium and the helium reaction productThis localized area sees much higher heating loads fromthe plasma but lower neutron fluxes It containsadditional special features to withstand the higher loca-lized heat loads Alternative first-wall designs are beingevaluated with the common features of a channel with

high-velocity flibe flow to provide sufficient first-wallcooling and a solid layer of beryllium or some othermaterial for neutron multiplication

While the main nuclear reactions are the berylliumneutron multiplication reaction and the lithium tritiumproduction there are many secondary nuclear reactions(Fig 4) creating short-lived radionuclides This generatesa significant gamma radiation field from the flibe saltflowing outside the reactor and a small amount of decayheat for several weeks after shutdown There will also beactivated corrosion products from materials in contactwith flibe There is no radiation damage to liquid flibesalt however if flibe salt is allowed to freeze at very lowtemperatures irradiating solid flibe salt or other fluoridesalts will generate free fluorinemdasha corrosive gas Uponfreezing there is only a 2 change in volume

The development pathway for the ARC has threemajor steps The first step is to develop and demonstratethe REBCO superconducting magnets at scale with thehigh magnetic fields that create high stresses within themagnet Smaller simpler magnets have been built withREBCO The CFS goal is to do this within 3 yearsThe second step would be a fusion machine calledSPARC (Refs 16 and 17) to demonstrate net energyproduction (more fusion energy out than energy into theplasma) and the associated reactor physics This machinewould operate for short periods of time with the neutronenergy dumped into solid neutron shields and adsorbedby the structure by raising its temperature It would bedesigned to operate for 1000 to 2000 cycles The ARCwould be a commercial demonstration project producingpower with self-sufficient tritium generation It requiresa flibe blanket for heat removal and tritium generation

III FHRS AND LIQUID-FUELED MSRS

The FHR is a new reactor concept18ndash20 less than20 years old that combines a clean fluoride salt coolant(no dissolved fuel or fission products) the graphite-matrixcoated-particle fuel originally developed for high-temperature gas-cooled reactors (HTGRs) and passivedecay heat removal systems from sodium fast reactors(SFRs) The baseline FHR uses pebble-bed fuel and 7Li2BeF4 (flibe) salt coolant enriched in 7Li to minimizetritium production Fluoride salt coolants are used becausethey are chemically compatible with the carbon-based fueland fluorine has a small neutron adsorption cross sectionFlibe is the best salt coolant in terms of thermal hydraulicsand neutronics performance2122 A schematic of thepebble-bed FHR is shown in Fig 1c

TABLE I

Nominal ARC Blanket Design Parameters

Design Parameter Value

Power level (fusion energyplus electrical input)

628 MW(thermal)

Flibe blanket volume 248 m3

Total flibe flow rate 125 m3sTemperature in 800 KTemperature out 908 KTritium generation rate 3 times 1020 tritium atomss

[~2 million CiGW(thermal)day]

Fig 3 ARC double-wall VV structure



Fluoride-salt-cooled high-temperature reactors area family of reactor concepts characterized by the use ofa high-temperature fuel and a fluoride salt coolant wherethere is an equivalent FHR concept for each major variantof an HTGR A liquid salt replaces the gas coolant of theHTGR resulting in a reactor that operates at near atmo-spheric pressure with power densities 4 to 10 timeshigher than the equivalent gas-cooled reactor because

liquids are better coolants than gases Flibe and othersalt coolants are mixtures of fluoride salts to lower melt-ing points Chloride salts are not viable for thermal-spectrum reactors because of their high neutron crosssections The pebble-bed FHR is the most advanceddesign20 with a nominal design shown in Fig 1 anddesign parameters in Table II The Kairos Power FHRis a variant of this design

Fig 4 Nuclear reactions with flibe salt8

TABLE II

Design Parameters of a Pebble-Bed FHR

Parameter Value

Reactor designThermal power [MW(thermal)] 236Core inlet temperature (degC) 600Core bulk-average outlet temperature (degC) 700Primary coolant mass flow rate (100 power) (kgs) 976Primary coolant volumetric flow rate (100 power) (m3s) 054Pressure AtmosphericCoolant 7Li2BeF4Fuel Graphite-matrix coated-particle fuel

Vessel interior from centerlineCentral graphite zone (m) 035Pebble fuel zone (m) 035 to 105Sacrificial pebble graphite zone (m) 105 to 125Graphite reflector (m) 125 outwardPower cycle NACC Modified General Electric 7FB gas turbineBase-load reactor [MW(electric)] 100Base-load efficiency () 425Peak power [MW(electric)] 242Peaking fuel Natural gasGas-to-peak power efficiency () 664

Comparison of power to vessel volumes (economics)a

PWR [MW(electric)m3] 28FHR [MW(electric)m3] 087SFR [MW(electric)m3] 029Pebble bed modular reactor [MW(electric)m3] 024

aThe combination of smaller vessel volume and lower pressure reduces the plant size



There are several incentives for developing the FHRRevenue per unit of thermal heat output is increased becausethe average temperature of delivered heat to the power cycleis ~650degC higher than the average temperature of deliveredheat from helium- sodium- and water-cooled reactors(Table III) Flibe has a melting point of 459degC resulting inminimum salt temperatures near 550degC to minimize the riskof freezing and peak salt temperatures near 700degC becauseof limits of materials for practical heat exchangers Highertemperatures enable (1) higher heat-to-electricity efficiency(2) higher temperature delivered heat to industry and (3)more efficient coupling to heat storage technologies thatenables variable electricity to the grid from a base-loadreactor to enhance revenue (Sec VI) This advantage appliesto all fission fusion and Gen-III CSP salt systems becauseall of the salts have high melting points The safety case isbased on several factors The FHR is a low-pressure reactorwith HTGR fuel that has a high heat capacity The high heatcapacity results in slow heat up of the reactor core under anyaccident conditions Safety is aided by the very high-temperature capabilities of the HTGR fuel where fuel failuretemperatures are above 1650degC the salt coolant has a boilingpoint above 1400degC and a coolant that dissolves most fis-sion products if there are fuel failures

There are also significant ongoing efforts to developMSRs where the fuel is dissolved in a fluoride saltDifferent designs use different fluoride salts23 TheSINAP (Ref 24) is in the process of designing and buildinga 2-MW graphite-moderated MSR using flibe salt In theUnited States the startup company FLiBe Energy Inc isalso proposing to use flibe salt The graphite moderator inthe reactor core has holes for the flowing fuel salt Thedissolved fissile fertile and fission products change someof the properties of the flibe salt including radiative heattransfer Countries in Europe have been working ona fluoride salt fast reactor1112 that uses a lithium heavy-metal fluoride salt This reactor has no graphite in the coreLike the FHR all of the MSRs with lithium-containing

salts use 7Li to minimize tritium generation Tritium isgenerated and thus systems to control tritium releases tothe environment are required that are similar to the FHR

There is also an effort to develop fast-spectrum MSRsby TerraPower where the fissile fertile and fissionproducts are dissolved in a chloride salt25 containing iso-topically separated 37Cl In a fast-spectrum MSR the fuelsalt flows through a 1-m-wide vessel with no internalcomponents while generating heat internally within thesalt from fission The thermal hydraulics of high powergeneration in flowing salt through a wide channel hassimilarities to the ARC fusion blanket The operating tem-peratures of these chloride salts are similar to the FHR

IV DIFFERENCES BETWEEN FISSION AND FUSION FLIBEUSE

Fission and fusion use of flibe results in many com-mon challenges but there are some important differencesdriven by different constraints and requirements26

IVA Power Cycles and Heat Storage

Many FHR and MSR designs include high-temperature heat storage to enable economic base-loadoperation with variable electricity to the grid to matchchanging market requirements and maximize revenue27

as discussed in Sec VI This change in future plantdesigns has major implications for the ARC and manyother fusion systems Most fusion systems operate ina pulse mode For inertial fusion the pulses may beseconds whereas for magnetic fusion the pulses may bemany hours followed by a short period of time beforeback to full power The addition of heat storage decouplesshort-term fusion heat output from the requirement todeliver electricity to the grid or heat to industry Itreduces power plant constraints on fusion system designconsequently the incentives for heat storage are greaterfor fusion machines

IVB Thermal Hydraulics

Flibe is a eutectic mixture of two salts (LiF and BeF2)to lower its melting point Like other fluoride salts flibehas a high Prandtl number (low thermal conductivity)relative to other coolants However there are some impor-tant differences in the use of flibe in fission versus fusionsystems

1 Magnetic fields Fusion systems have very highmagnetic fields that interact with flowing flibe to change

TABLE III

Heat Delivery Temperatures for Different Reactor Coolants

Coolant

AverageCore InletTemperature

(degC)

AverageCore Exit

Temperature(degC)

AverageTemperatureof DeliveredHeat (degC)

Water 270 290 280Sodium 450 550 500Helium 350 750 550Salt 600 700 650



the thermal hydraulics via mechanisms such as suppres-sion of turbulence2829 These heat transfer effects are notfully understood for flibe

2 Radiative heat transfer Flibe is semitransparentand at 700degC radiative heat transfer begins to becomeimportant30 Radiative heat transfer reduces temperaturedifferences in the system This is significant in an FHRparticularly in accident scenarios However it is moreimportant in the fusion system where the peak heat gen-eration in the liquid blanket is next to the inner wall ofa 1-m-thick blanket In such a system longer-distancetransport of radiative heat can become important

3 Geometric effects In the FHR there is goodmixing of the coolant flowing through a pebble bed Ina flibe fusion blanket tank (Fig 1) one concern is unevenflow and formation of vortexes where flibe remains rela-tively stationary in some locations Because of the highinternal power generation in flibe if vortexes form itcould result in locally rapid heating of flibe to the boilingpoint and large variations in flibe salt temperatures at theoutlet of the blanket This has similarities to the challengein fast-spectrum MSRs where the fuel salt is flowingthrough a channel more than 1 m in diameter with veryhigh rates of heat generation in the salt

4 Extreme heat fluxes The first wall of the fusionmachine that separates the plasma from the salt has highheat fluxes (017 MWm2) with much higher local heatfluxes in the diverter These heat fluxes are much higherthan found in an FHR In addition there is high internalheating of the first wall from the gamma and neutronfluxes At the same time the peak temperature limits forthe fusion first wall are below the temperature limits forgraphite-matrix FHR fuel

There are good low-temperature simulants for flibethat can be used for many heat transfer tests31 such asDowtherm A however these materials canrsquot simulateradiative heat transfer or impacts of magnetic fields onheat transfer Computational fluid dynamics is required todesign the fusion blanket Separate effects experimentsare required to understand the individual phenomena thatare then combined using computational fluid dynamicsA small number of expensive highly integrated experi-ments may be required to validate design codes Much ofthis work would also be useful for the FHR and MSR

IVC Tritium

The ARC and FHR have tritium challengesTritium production in the FHR with isotopically

separated 7Li is about two orders of magnitude higherthan releases from a pressurized water reactor (PWR)but three orders of magnitude less than the productionrate in a fusion machine In a PWR the tritium isprimarily from the lithium and boron in the water cool-ant In the FHR the lithium is 9999+ 7Li to minimizetritium production For the ARC fusion machine thenominal lithium isotopic composition is 90 6Li tomaximize tritium production Both systems requireefficient tritium removal from salt and control ofreleases to the environment In addition to the highertritium levels in a fusion machine there are two impor-tant differences with regard to tritium managementrequirements in fusion versus fission machines

1 Tritium waste or fuel In an FHR the tritium isa waste or a secondary product whereas fusion machinesrequire efficient recycling for the fusion reactor to beself-sufficient in tritium Tritium has some value todayparticularly its decay product 3He As a consequence foran FHR there is an incentive to recover the tritium for themarket but no requirement for efficient recovery Certaintritium removal options such as gas sparging may spargewith a hydrogeninert gas mixture to (1) remove tritiumand (2) control salt redox (corrosion) potential Extrahydrogen mixed with the tritium is not a concern for anFHR In a fusion machine any normal hydrogen has to beremoved before the tritium is recycled to the plasma

2 Safety basis In a fusion machine the primaryaccident source term is the tritium There are large incen-tives to minimize the tritium inventory to minimize theaccident source term and required safety systems Thiscreates incentives for fast recovery and recycling of tri-tium from the flibe blanket

There has been significant recent work on tritiumcontrol1932 for the FHR including irradiation of differentmaterials in flibe salt at 700degC There has been much lessrecent work on tritium control for fusion for flibe sys-tems In each case neutrons interact with LiF to generatetritium in the form of 3HF and helium Hydrogen fluorideis highly corrosive so a chemical redox agent33 is addedto the salt to convert the 3HF to 3H2 Section V discussesthe common technologies for tritium control

V FLIBE SYSTEMS

The status and challenges with flibe systems arediscussed herein recognizing that major flibe test facil-ities are expected to come online in the next 5 years in theUnited States (FHR ARC) and in China (MSR)



VA Supply Chain

The supply chain challenges to produce flibe areidentical for both systems The FHR will use isotopicallyseparated 7Li while the ARC uses lithium isotopicallyenriched in 6Li Lithium isotopic separation facilities arerequired The United States (University of Wisconsin atMadison Oak Ridge National Laboratory) and China cantoday produce small quantities of purified flibe at labora-tory scale (tens of kilograms per batch) not industrialscale (tonnes) Industrial-scale flibe supply chains arebeing developed by Kairos Power for the FHR in theUnited States and the SINAP in China for salt reactorsThis includes capabilities to isotopically separate lithiumisotopes

VB Properties

Flibe systems and tritium management options fol-low directly from the physical characteristics of flibe Theproperties of flibe and several other fluoride salts areshown in Table IV Flibe3435 is the primary liquid-saltoption for fusion to obtain the high tritium breeding ratiothat is required The other salts could be potentially usedin some types of FHRs and MSRs Flibe has betterneutronic and thermal-hydraulic properties19212236 thanthe other salts For FHRs the disadvantages of flibeversus other salts are the toxicity of beryllium and thecost of isotopically separated lithium

Recent critical reviews2122 have evaluated the phy-sical properties of flibe and the uncertainties associatedwith those physical properties There are significantuncertainties associated with the thermal conductivitypartly because flibe is semitransparent and at highertemperatures radiative heat transport becomesa significant heat transport mechanism The major area

of missing physical property data is measurement ofradiative heat transfer and other optical properties asa function of temperature frequency and impurity levelsThis is important for multiple reasons

1 Heat transfer Above 700degC radiative heat trans-fer becomes important and becomes a significant factor inminimizing peak coolant temperatures This is relevant inthe context of accident conditions for FHRs and MSRs Itis also relevant to reducing peak temperatures in theARC In a large flibe blanket tank or flow channelradiative heat transfer can change temperature distribu-tions and thus flows

2 Instrumentation The optical properties createnew instrumentation options that may be required forfusion and useful for fission Infrared imaging can detecthot spots associated with the first wall through theflibe Second Cherenkov imaging diagnostics from neu-tron interactions with the flibe salt may enable directmeasurement of fusion power

The salt melting points for fission fusion and Gen-IIIsolar systems are high relative to other coolants such aswater helium and sodium Special engineering featuresare required to avoid salt freezing The compensatingfeature is the very high boiling points so that in mostcircumstances there is little risk of boiling

VC Corrosion

Liquid flibe salt is chemically stable in high neutronand gamma radiation fields It is an ionic solution Frozenflibe salt near room temperature in high radiation fieldswill generate free fluorine Lithium fluoride plusa neutron yields tritium fluoride (3HF) and heliumTritium fluoride is highly corrosive To prevent corrosiona chemical reducing agent is required The leading

TABLE IV

Candidate Reactor Coolant

Coolant Tmelt (degC) Tboil (degC) ρ (kgm3) ρCp (kJm3degC)

Li2BeF4 (flibe) 459 1430 1940 4670595 NaF-405 ZrF4 500 1290 3140 367026 7LiF-37 NaF-37 ZrF4 436 2790 350051 7LiF-49 ZrF4 509 3090 3750Water (75 MPa) 0 290 732 4040Water (155 MPa) 0 345 709 4049

Salt compositions in mole percent Salt properties at 700degC and 1 atm The boiling points of the zirconium fluoride salts are notwell known For comparison water data are shown at 290degC (75 MPa boiling point) and PWR conditions with water properties at309degC



candidate is beryllium metal that is slightly soluble inflibe and rapidly reacts with HF (Ref 37) Beryllium plustwo tritium fluorides yields BeF2 plus

3H2mdashhydrogen Inflibe salt hydrogen remains as diatomic hydrogen Otherreducing agents can be used to convert 3HF to 3H2 Thereference ARC design has beryllium exposed to flibe Ifdissolution rates of beryllium are too high or localizeda more powerful reducing agent may be required tominimize beryllium dissolution into the salt

Fluoride salts are fluxing agents that is they dissolveoxide and other coatings The only way to prevent corro-sion of the metal is to have the chemical redox conditionsin the salt more reducing than the metals of constructionthat is the metals must be noble relative to the salt33

Figure 5 shows the chemical redox potential of variouselements The materials of construction determine theredox conditions required to minimize corrosion Withthe requirements for redox control is the associatedrequirement to develop robust methods to measure andcontrol the redox potential to assure corrosion controlMultiple methods to measure redox control have beendeveloped in a laboratory setting but what is required isdevelopment of a redox control strategy and instrumenta-tion with the reliability requirements for power plantoperations In this context there is a large differencebetween MSRs and clean flibe systems In a fluoridesalt MSR the historical strategy to control redox is tocontrol the ratio of dissolved +3U+4U in the salt With thehigh concentrations of uranium in the salt measurementand thus control of redox potential is straightforward

Material choices are first limited by the requirementsto be compatible with flibe salt as shown in Fig 5 Notethat the HFH2 = 001 potential is above the CrCrF2potential implying that HF will react with chromiummetal to produce chromium fluoride and H2 Chromiumis an alloying agent in almost all high-temperature mate-rials Structural material choices depend on if one wantsa code-qualified material Using code-qualified materialssimplifies licensing and code-qualified materials arecommercially available However it requires nearlya decade to fully qualify a structure material There isthe option of adding nonstructural materials includingtungsten and nickel as coatings for corrosion controlAdding carbon to the system such as the fuel in theFHR can increase corrosion rates during irradiation1936

All materials in contact with electrically conductive flibemust be considered because of the potential of galvaniccorrosion between different materials of construction Insome designs of FHRs this has led to the decision to usea single material of construction 316 stainless steel toavoid such interactions Material decisions are at an

earlier stage of development for fusion This paper doesnot go further into this complex subject

Coupled to materials and tritium control is purificationof the flibe salt to minimize corrosion Some impurities inall materials of construction will leach into the salt overtime Neutron irradiation will generate other species in thesalt such as oxygen and by interactions with the materialsof construction creating corrosion products that will dis-solve in the salt Fuel in FHRs and operations will addimpurities Like other coolants the impurities in the saltcoolant can have a large impact on corrosion rates It is notknown at this time what the online purification requirementswill be

Fig 5 Redox potentials of various redox couples asfunction of temperature in fluoride salts Solid linemetal dissolution at aMnthorn of 10minus6 dotted line reductionof oxidants HFH2 represents the mole ratio of HFH2 at1 atm total pressure



The traditional method to purify flibe involves fluor-ination with HF that strips volatile fluoride impuritiesfrom the salt including oxygen as steam controllingredox by the ratio of hydrogen to HF to precipitatemost remaining impurities then followed by filteringany metal particulates in the salt This method is robustand works well as a production method for flibe salt Ifproducing flibe salt from beryllium and lithium fluorideslow-cost feedstocks may come with high levels of impu-rities and thus the need for such a process For a fission orfusion reactor the question is whether simpler processescan be developed for online or batch purification of saltswith lower concentrations of impurities

There are other purification options includingdistillation38 Distillation was experimentally investigatedfor flibe salt in the 1970s for online purification of MSRsalts and looked potentially attractive However with theexisting 1970s materials of construction the only optionwas vacuum distillation which is a low-throughputprocess with limited separation capabilities Higher-temperature multistage distillation was not viable becauseof the limits in the fabrication of high-temperaturematerials such as molybdenum into complex shapesNew methods of additive manufacturing may enable build-ing high-performance high-temperature distillation col-umns of molybdenum or other materials that wouldenable distillation as a high-capacity low-cost no-additivesalt cleanup option There are other options such as carbonbeds that may be able to remove impurities other thantritium (Sec VD1c) but very little work has been doneon such longer-term options Much of the ongoing work inthis area is associated with MSR salt cleanup and theconversion of wastes into acceptable waste forms

VD Tritium Challenge

Tritium capture and control is a major challenge withflibe and other lithium-containing salts Tritium is thefusion fuel with the requirements for (1) efficient recov-ery for a self-sustaining fusion reactor (2) fast recoveryto minimize total tritium inventory in the fusion plant(minimize inventory and thus accident source term) and(3) avoiding significant releases to the environment Infission machines the goal is to minimize releases to theenvironment with tritium generation rates typically threeorders of magnitude lower than in a fusion machine Inaddition to traditional MSRs with flibe the EuropeanCommunity has been developing a fast-spectrum MSRwith a 7Li heavy-metal fluoride salt1112 where lithium isused to reduce the salt melting point Tritium capture andremoval can be done in the (1) salt (2) certain types of

heat exchangers and (3) secondary heat transfer loopsAll options are discussed herein There is no consensuson preferred options for tritium control and none of thetechnologies have been demonstrated at significant scalein salt systems

VD1 Removal of Tritium from Salt

There are many methods to remove tritium (3H) fromthe salt Recent reviews32 summarize some of theseoptions To understand the performance limits of thesesystems one must understand the behavior of hydrogenin flibe Figure 6 shows hydrogen diffusivity in flibe andother materials that may be found in a flibe fission orfusion system Figure 7 shows the solubility of hydrogenin flibe and other materials that may be found in a flibesystem For diatomic molecules such as hydrogen thesolubility follows Sievertrsquos law in metals including liquidsodium (discussed in Sec VD3c) where the solubility isproportional to the square root of the hydrogen pressureIn liquids such as salts the solubility of diatomic hydro-gen follows Henryrsquos law where the solubility is propor-tional to the pressure of the hydrogen These two physicalproperties drive the choice of options to capture tritium orto prevent its escape from flibe

The permeation (transport) rate of hydrogen is pro-portional to the diffusivity times the solubility In termsof permeation rates diffusivity can be considereda measure of the difficulty of a molecule moving througha material whereas the solubility is a measure of posi-tions that can be occupied by tritium and therefore the

Fig 6 Hydrogen diffusivity in flibe Flinak sodium anddifferent metals Values are extrapolated from availableexperimental range Flibe and Flinak from 773 to 873K(Ref 40) sodium from 742 to 853 K (Ref 42) SS316and Ni from 500 to 1200 K (Ref 39) and tungsten from1100 to ndash2400 K (Ref 41)



number of pathways for a molecule to move througha material Liquid salts have low diffusivity relative tomost materials and low hydrogen solubility Tritium (3H2)moves very slowly through flibe As a result the primarylimit of any tritium removal from a flibe system is thetime it takes the tritium molecules to diffuse from thebulk salt to some surface for removal of the tritium Thisresults in tritium removal options where in each case thesurface area and fluid transport via turbulence oftendetermines the rate of tritium removal

Figure 8 shows a typical schematic of a system toremove tritium from the salt The specific example is fortritium removal using permeation membranes The tri-tium removal system is before the main heat exchanger

in hot flibe salt to reduce the concentration of tritium inthe salt entering the heat exchangers and thus the tritiumconcentration driving force for diffusion of tritiumthrough the heat exchangers Most of the external systemsurface area is associated with the heat exchangers Somesystems to remove tritium from salt will remove tritiumand tritium fluoride (gas spargers carbon beds etc) butin all cases are more efficient in removing tritium gas(3H2) than tritium fluoride because tritium gas diffusesfaster through flibe salt to the surface of the tritiumremoval system Permeation membranes and heat exchan-gers only remove tritium gas Tritium removal systemsthus change the ratio of tritium to tritium fluoride in thesalt The redox control agent is required to maintain thedesired redox potential

VD1a Gas Sparging

In gas sparging32 an inert gas or a hydrogeninert gasis bubbled through the salt to remove the tritiumHydrogen in the sparging gas may be used for redoxcontrol For the FHR tritium with added hydrogen impliesa somewhat larger waste stream with tritium For fusionthe use of hydrogen or deuterium in the gas spargingsystem implies the need to later separate the tritium forrecycling back to the fusion plasma The tritium removaleffectiveness depends upon the bubble surface area

This technology is being developed by SINAP wheregas is injected goes through a structure to create turbulentmixing between the gas and liquid and then througha centrifugal system that results in efficient separation ofgas from the flibe salt The SINAP development work is forFHRs and MSRs In a MSR the system also removesxenon which is a strong neutron poison as well as othergaseous fission products and some noble metals There arelarge economic incentives for the fast removal of xenon tominimize parasitic neutron losses in the reactor core Themany short-lived fission products result in an off-gas systemwith very high radiation levels which creates incentives tominimize equipment size and thus radiation shielding Thereare many other variants such as a gas scrubber where theliquid flows downward over some packing forming a thinlayer on the surface of the packing as the purge gas travelsin the upward direction In these systems there is a trade-offbetween equipment size energy consumption and volumeof expensive flibe in the separation system

VD1b Vacuum Disengager

In a vacuum disengager45 the hot flibe salt is sent toa vacuum tower where spray nozzles create small droplets

Fig 7 Hydrogen solubility versus temperature for dif-ferent materials Values are extrapolated from availableexperimental range Flibe and Flinak from 773 to 873 K(Ref 40) sodium from 742 to 853 K (Ref 42) SS316and Ni from 500 to 1200 K (Ref 39) and tungsten from1100 to 2400 K (Ref 41) The experimental values forthe low estimate of graphite are based on high-densityisotropic Isograph-88 extrapolated from 1123 to 1323 Kand the high estimate is estimated based on recent experi-ments on carbon materials at 973 K (Refs 43 and 44)

Fig 8 Schematic of the tritium control and mitigationsystem



(Fig 9) Tritium diffuses out of the droplets as they fall andis removed by vacuum pumps This option was examinedin the 1990s as part of the US inertial fusion program Inthe initial design the optimum droplet size was 02 mm ina 7-m-tall vacuum tower A two-stage system was esti-mated to provide a decontamination factor of 100 000 andthus meet requirements for tritium capture and allowablereleases The vacuum disengager is between the fusionchamber and the first heat exchanger to remove tritiumbefore entering the heat exchanger The major trade-off isbetween the size of the droplets and vacuum pressurelosses Smaller droplets result in higher diffusion rates oftritium from the center of the droplet to the surface buthigher concentrations of droplets result in the buildup oftritium gas backpressure from the droplet surfaces to thevacuum pumps No experimental work was done We arenot aware of any recent work in this area

VD1c Solid Adsorbers

A particle bed of a solid adsorber can be used to removetritium from the flibe Such beds can have very high surfaceareas The tritium is desorbed from the bed by heating Theleading candidate is a carbon bed because carbon is compa-tible with fluoride salts and has high solubility to hydrogenThe primary choices are carbon and the materials that arenoble (chemically inert) to flibe (Fig 5) There is almosta two order of magnitude difference in the solubility ofhydrogen on different forms of graphite (Fig 7) Recent

experimental work46 at MIT has measured hydrogensolubility at 700degC for a variety of carbon forms as shownin Fig 10 Carbon has several different types of sites thatadsorb hydrogen where the concentrations of each type ofsite depend on how the graphite was processed Most hydro-gen is weakly or strongly adsorbed The quantity of weaklyadsorbed hydrogen depends upon the hydrogen (or tritium)pressure If the goal is a hydrogen bed to remove tritiumfrom flibe one wants a carbon in a form with a highsolubility for weakly adsorbed hydrogen to enable easyrecovery of the hydrogen However this also implies thatin a reactor core that has a carbon form with a high solubilityfor hydrogen that is weakly adsorbed hydrogen (or tritium)will be quickly released in an over-temperature event

In the pebble-bed FHR the fuel and graphite moderatorpebbles spend an average of a month traveling through thereactor core They are then inspected with the fuel pebblesthat have reached their burnup limits sent to spent nuclearfuel storage and the remainder of the pebbles recycled backto the core The specific characteristics of this reactor coredesign may allow the fuel and graphite pebbles in the coreto be used as a tritium removal system by heating thepebbles after exiting the core to remove the weakly boundtritium There is the option47 to add smaller pebbles ofgraphite that fit in the spaces between the fuel pebbles thatare designed specifically for tritium removal Such nonfuelpebbles are at lower temperatures than the fuel pebbles andthus have a greater capacity for tritium adsorption

There are other limitations on the use of carbon insome salt systems Carbon can increase corrosion For anFHR this is not a consideration because the FHR usesa carbon-based fuel Adding more carbon does not

Fig 9 Tritium gas diffusion out of flibe spay region(Droplet size is exaggerated)

Fig 10 Quantities (standard temperature and pressure incubic centimetersgrams) of weak and strong adsorptionat 3 kPa



change system behavior Materials and corrosion controlstrategies must be chosen to address this challenge Ina fusion system a carbon tritium removal system may notbe chosen because of its impact on material choices

VD1d Permeation Membranes

A permeation membrane is a wall with high perme-ability to tritium As shown in Figs 6 and 7 nickel is thepreferred permeation filter due to its relatively highdiffusivity and solubility Permeation membranes areused to purify tritium and are well understood Themajor advantage of a permeation barrier is that itproduces a pure hydrogen gas stream diluted by thesweep gas that is typically helium or argon but no otherimpurities The disadvantage is the high surface area thatis required and may approach the area of the main heatexchangers

The most recent work4849 has been done at theUniversity of Michigan for FHRs Figure 11 showsa modular separation unit The permeation barrier is similarto a cross-flow heat exchanger where the salt flowsperpendicular through an array of tubes with a triangularpitch This is to create turbulent mixing to maximize masstransfer from the salt to the tube The sweep gas flowsinside the tubes and carries away the tritium gas

Other options are discussed below for tritiumremoval using heat exchangers or capturing tritium inthe secondary loop However there are large incentivesto remove as much tritium as possible in the primarysystem to minimize the tritium concentration in the

primary system The higher the concentration in theprimary system the larger the driving forces for tritiumreleases through the high-surface-area heat exchangersand the larger inventories and accident source term

VD2 Tritium Barriers

Tritium permeation barriers32 can minimize tritiumleakage In any fission or fusion system more than 99of the external surface area of the flibe system is asso-ciated with the primary heat exchangers Thus the pri-mary tritium challenge is how to prevent migrationthrough the primary heat exchanger On the salt sidethe leading candidate is tungsten with its low solubilityand diffusion coefficient for hydrogen Salt-side optionsare limited because of the requirements to be chemicallycompatible with the flibe salt as defined by the chemicalredox potential (Fig 5) and compatibility with the struc-tural material of the heat exchanger The alternative isa permeation barrier on the nonflibe side of the heatexchangers Oxides and many other layers in the labora-tory have extremely low permeability to hydrogenHowever the real-world experience is not as goodCracks in these surface treatments control real-worldperformance as has been demonstrated in many environ-ments The practical implication is that tritium barriersmust be tested at a reasonable scale (many square meters)under realistic operating conditions The best-preforminghydrogen permeation barrier is not necessarily the mate-rial with the lowest permeability itrsquos the one with thefewest cracks in the barrier or where the environment willseal the cracks

VD3 Removal of Tritium with Heat Exchangers

Tritium can be removed with some types of heatexchangers This can be the primary or a secondary tri-tium removal system In each of these options there arevarious trade-offs between requirements to move heat toa secondary fluid and requirements for tritium captureSome of these options are capable of greatly reducing therisk of salt freezing in the heat exchanger if the fissionfusion machine is shut down

VD3a No Intermediate Loop

No intermediate loop is proposed for FHRs (Ref 20)coupled to nuclear air-Brayton combined cycles(NACCs) In a NACC power system air is compressedis heated by a salt-air heat exchanger and goes toa turbine with the exhaust sent to a heat recovery steam

Fig 11 Schematic of a proposed cross-flow tritiumremoval facility



generator The added steam can be used to produce addedelectricity The important characteristic of this powercycle is that the front-end compressor results in com-pressed air at temperatures between 350degC and 450degCThis high-temperature oxidizing environment maintainsa self-healing oxide layer on the outside heat exchangersurfaces that assures very low hydrogen permeabilityThe system heals any imperfections in the permeationbarrier This option is early in the development processand has not has been tested at scale

VD3b Double-Wall Heat Exchangers

Double-wall heat exchangers have been used inindustry for many decades where heat transfer is requiredbetween highly reactive liquids This type of heat exchan-ger has been considered for decades for flibe service infission and fusion reactors Heat and mass transfer followsimilar laws thus a heat exchanger designed for trans-ferring heat from salt to a second fluid is also good fortransferring hydrogen to the surface of the heat exchangerand diffusion through the double-wall heat exchangerThe properties of flibe salt favor the use of fluted heattransfer tubes to improve heat transfer This is becausethese salts have a low thermal conductivity relative tomost other heat transfer fluids

In the United States most recent work on double-wall heat exchangers for flibe systems has been done atthe University of Michigan5051 and the University ofCaliforniaKairos Power A schematic of a double-wallheat exchanger tube is in Fig 12 Recent studies indicatethat the sweep gas can capture 995 of the tritium goingthrough the heat exchanger Typically the sweep gas ishelium with its high thermal conductivity There are alsoproposals to use a salt as the sweep fluid for tritiumcapture Such a fluid would have a higher thermal con-ductivity and thus a more thermally efficient heat exchan-ger There are plate-type heat exchangers that areequivalent to a double-tube heat exchanger This may bea sufficient level of tritium removal for an FHR but

would not be acceptable by itself for a fusion machinein terms of allowable tritium releases to the environment

There are a large number of design options for tri-tium removal with these systems most of which have notbeen investigated in any detail For example the annularsweep gas could be helium with a small quantity ofoxygen Any tritium entering the annular zone would beoxidized to steam that does not diffuse through metalwalls However oxide layers act as barriers to tritiumdiffusion The oxygen would create a barrier to the diffu-sion of tritium from the flibe salt to the annular zoneIdeally one would want to coat the tube facing the flibesalt with a metal that did not oxidize to allow efficienttransport of tritium into the annular zone to remove thetritium from the system

VD3c Heat Pipes

Sodium heat pipes are proposed to transfer heat from theflibe salt to the power cycle5253 with three additional objec-tives (1) isolate low-pressure flibe reactor system from high-pressure power cycle (2) prevent accidental freezing of flibesalt and (3) capture tritium Each heat pipe (Fig 13) wouldhave its bottom half in the flibe salt and its top half deliveringheat to the power cycle either (1) directly heating steamsupercritical carbon dioxide or compressed air or (2) heatinga secondary heat transfer fluid such as a nitrate or chloridesalt if heat storage is in the power cycle The heat vaporizesthe sodium that then flows as a gas to the colder condensersection of the heat pipe where it condenses to a liquid andflows back to the hot salt zone The liquid salt and steam orair heat exchanger surfaces are separately designed for effi-cient heat transfer The heat pipe is lined with a wick wherecapillary forces wet the entire heat pipe surface and moveliquid sodium from the cold condenser section to the hot saltsection where it is evaporated

Heat pipe performance depends upon the vapor pres-sure of the sodium which depends upon temperatureSodium is chosen because at the melting point of flibe(459degC) the vapor pressure is very low resulting in almostno heat transfer At 600degC the system is efficient at heattransfer Heat pipes can be designed to turn on and shutoff at defined temperatures This characteristic can pro-vide protection against freezing of the flibe salt

The heat pipe can be designed to capture tritium Notritium diffusion barriers are incorporated into the eva-porator zone of the heat pipe to enable efficient tritiumpermeation from salt into the heat pipe Inside the heatpipe the evaporation and movement of the sodium vaporsweeps noncondensable tritium to the condenser sectionThe condenser section includes barriers to tritium

Fig 12 Three-fluid design concept for a double-wallheat exchanger



transfer such as tungsten electroplate on the inside of theheat pipe Tritium is removed by a permeation mem-brane made of nickel or other highly permeable materialto hydrogen The permeation membrane may be eithera tube inside the heat pipe that leads to a vacuum zone tocollect the tritium or a section of the heat pipe wall thatacts as a permeation membrane Unlike a permeationmembrane in salt the required surface area fora permeation membrane in this system is very smallThe sodium vapor sweeps the tritium into the condensersection Diffusion of tritium through the sodium vapor isvery fast relative to a liquid salt because the density ofthe vapor is three orders of magnitude less than thedensity of the salt If a permeation membrane is part ofthe heat pipe wall the area is small because the highsolubility of hydrogen in sodium and the high diffusioncoefficient of hydrogen in sodium relative to salt (Figs 6and 7) implies high rates of hydrogen diffusion throughthe sodium to the permeation membrane

Sodium heat pipes with permeation membranes toremove hydrogen have been designed for industrial andother applications but not for tritium removal Heat pipesare being developed to remove heat from hydrogen-richstreams in chemical plants where a quarter or more of thegas stream may be hydrogen In this environment suffi-cient hydrogen diffuses into the heat pipe to shut it downwithin days The hydrogen is a noncondensable gas thatis swept to the condenser section of the heat pipe andblocks sodium vapor transport In these systemsa permeation tube is added to remove hydrogen from

the heat pipe For these applications the hydrogen con-centrations are many orders of magnitude greater thanwould be found in an FHR or ARC and there is norequirement to avoid some loss of hydrogen Heat pipeshave been built for decay heat removal from salt reactorsbut have not been designed for tritium blockage andrecovery

VD4 Removal of Tritium from Secondary Loop

VD4a Nitrate Salts (Chemically Oxidizing)

Kairos Power and most MSR startup companies pro-pose a nitrate intermediate loop between the primarysystem containing flibe salt and the power cycle27 to (1)isolate the low-pressure flibe system from the high-pressure power cycle (2) act as a backup system fortritium capture and to prevent tritium release and (3)operate as a heat storage system (Sec VI) The nitratesalt is the solar salt used in existing CSP systems (notadvanced Gen-III CSP systems) as a heat transfer andheat storage system It can be used to transfer heat toindustrial customers Nitrate salt systems operate between290degC and 600degC With careful control or chemistry itmay be possible to increase the temperature to 650degCMultiple CSP systems use these nitrate salts thus there islarge-scale industrial experience in these salt systemsincluding large-scale pumps and nitrate heat storagetanks that store gigawatt hours of heat

Fig 13 Heat transfer regions within a heat pipe



Hot nitrate salts are highly oxidizing and thuscreate oxide layers with low permeability to tritiumdiffusion These oxidizing salts will quickly oxidizeany tritium to steam that can be collected in the off-gas system Steam in an oxidizing environment will notdiffuse through steel An intermediate nitrate loopprovides a trap for any tritium This is the near-commercial option to assure no tritium escape Therehas been only limited work on the steam recoverysystems from the nitrate salt

There are differences in the use of nitrate salts inintermediate loops with different salt reactors In theFHR the nitrate salt loop may operate at a lowerpressure than the flibe primary loop to avoid anoxidizing salt entering the reactor coolant and reactingwith the carbon moderator in the event of a heat exchan-ger failure A fusion flibe blanket may not have anycarbon in the primary loop and thus this is nota concern In a MSR the nitrate salt loop may be ina tertiary loop with a secondary loop operating ata higher pressure than the primary loop to avoid trans-port of fuel salt out of the primary loop in the event ofa heat exchanger failure and any interactions of anoxidizing nitrate salt with the dissolved fuel and fissionproducts In a MSR changes in salt chemistry have thepotential to precipitate fuel or fission products creatingother safety concerns

VD4b Halide Salts (Chemically Reducing)

Multiple halide salts are proposed for the intermediateloops of FHRs and MSRs In terms of tritium the commoncharacteristic of all these salts is that tritium will be in thehydrogen form (3H2) This creates the potential for thetritium to enter the secondary salt system and then diffuseout of the salt system through piping or heat exchangersThese systems can use the same tritium removal systemsas used in the primary loop (both chemically reducingchemistry) with some important differences Permeationfrom the flibe salt into the secondary salt implies purehydrogen in the secondary loop There are no other radio-nuclides or HF It is a clean system Second the secondarysalts have much lower costs

The combination of a lower-cost salt and no gammadose reduces the design and cost constraints for anytritium removal systems including carbon beds and gassparging systems in secondary systems Options such asa large gas stripping tower with a fill material (secondarysalt going downward with stripping gas going upward)may become attractive There are several candidate saltseach with different characteristics

1 Flinak Flinak (LiFNaFKF) has been proposedfor the secondary loop because it is compatible with flibeand is relatively nontoxic For an FHR or MSR the majorquestion about Flinak is whether to use isotopically sepa-rated 7Li If natural lithium is used a small leak of Flinakinto the primary system would shut down the reactorThis would be a much smaller concern in a fusion systemwhere one burns out lithium at a significant rate toproduce tritium and thus must include some system toreadjust the lithium isotopics over time

2 Potassium-zirconium fluoride The nonlithiumalternative to Flinak is a mixture of potassium and zirco-nium fluorides with relatively low cross sections to para-sitic neutron capture

3 Sodium-potassium-magnesium-chloride salts Thenext-generation CSP system designs1327 propose usinga sodium-potassium-chloride salt with a melting point of~420degC It is extremely cheap and proposed for large-scalemultigigawatt-hour heat storage systems It is currently pro-posed to be used in the intermediate heat transfer loop of theTerraPower MCFR and would be a leading candidate forthe intermediate loop of any other chloride salt reactor Tominimize corrosion small amounts of magnesium metal aredissolved in the salt creating a chemically reducing environ-ment We are not aware of any studies of hydrogen transportand behavior in these systems

VD4c Steam or Supercritical Carbon-Dioxide PowerCycle

The last tritium trapping system is the power cycle Ifthe cycle operates under oxidizing conditions the powercycle can be a tritium trap If itrsquos a steam cycle then thereis the challenge of tritium separation from large quantitiesof normal water because of rapid isotopic exchange ofnormal hydrogen and tritium in a hot water or steamenvironment These options are reasonably well under-stood because this can happen in CANDU reactors Weare not aware of any assessment of trapping tritium ina supercritical carbon dioxide cycle where one onlyrequires separating water from carbon dioxide

VE Systems Design Analysis

The development of tritium removal systems requiresa system analysis code that includes salt chemistry cor-rosion redox control tritium capture systems and tritiumbarriers It is a systems question (1) where and howtritium should be captured (from salt in heat exchangers



in secondary loop) (2) the performance of permeationbarriers and (3) the redox control strategies that deter-mine allowable materials of construction For fusionmachines the total tritium inventory determines the acci-dent source term and thus creates large incentives tominimize total tritium inventory The question is whatset of options meets the requirements at the lowest costwith the required assurances for tritium recovery and lowreleases to the environment TRIDENT5455 and severalother systems codes exist to address these questions

TRIDENT (Fig 14) models tritium generation bylithium transmutation speciation based on chemical condi-tions (redox potential) absorption in reactor and structuralmaterials including all components in the primary andsecondary system corrosion of structural materials andconvection and diffusion through the molten salt and theheat exchangers It canrsquot be overemphasized that tritiumbehavior is controlled by chemical redox that controls theratio of H2 to HF Currently TRIDENT is equipped with thecapability to simulate various tritium mitigation strategiesincluding the use of an adsorption column a gas spargingtower and a permeation window as well as using chemicalcontrol agents to limit the diffusion due to permeation ofgaseous H2 (Refs 54 and 55) If a new technology isintroduced it must be added to these systems codes

Previous analyses in TRIDENT using an adsorptionbed found that the radioactive release of tritium from thereactor could be reduced by more than 90 usinga graphite bed5657 With high-performance Calgon

Carbon granular carbon OVC 4 times 8 from Sec VD1cthe solubility of carbon adsorbents can be increased by anorder of magnitude significantly reducing the total radio-active release4346 Using TRIDENT a countercurrent car-bon adsorption bed with carbon OVC 4 times 8 is simulated atdifferent 7Li concentrations in the 236-MW(thermal)prototypical FHR (Fig 15) Columns of different sizeswere examined resulting in four different average tritiumgeneration rates 2670 Ciday for 99995 7Li (design casefor FHR) 24 140 Ciday for 9995 7Li 238 430 Cidayfor 995 7Li and 2 450 070 Ciday for 95 7Li enrich-ment In an FHR the lithium enrichments are above 99997Li but lower lithium enrichments are used in this exampleto demonstrate how system performance changes as tritiumproduction increases as in a fusion system The maximumrelease rate is found to be 40 Ciday for 99995 7Li(design case for FHR) 1440 Ciday for 9995 7Li25 240 Ciday for 995 7Li and 300 860 Ciday for95 7Li enrichment An order of magnitude increase inproduction is accompanied by an order of magnitudeincrease in the tritium release rates as the limits of tritiumsolubility are reached with a fixed system size and rate ofcarbon regeneration The time-dependent behavior reflectsthe burnout of 6Li over time in an FHR

To illustrate the mitigation challenges of high tritiumgeneration the size of the adsorption column required foran FHR at 995 7Li enrichment is examined At 9957Li the system has two orders of magnitude greatertritium generation than the base-case FHR The surface

Fig 14 Tritium mechanisms in TRIDENT



area of the bed previously modeled is increased (byincreasing bed length) and the corresponding releaserates are shown in Fig 16

An inverse (1x) relationship is found between thesurface area and release rate as shown in Fig 16When the surface area is doubled from the originalbed the release rate decreased from 25 243 to11 490 Ciday At the extreme case of 10x bed sizethe peak release was found to be 1301 Ciday whichis still an order of magnitude greater than any existingfission reactor in operation today Thus the costs ofmitigating tritium releases in fusion devices will bemuch greater than for highly 7Lindashenriched fissionsystems unless there is a step change improvementin mitigation technology

What such system models show is that there aremultiple existing options to control tritium releasesfrom an FHR Existing technologies can efficientlyrecycle tritium in a fusion machine The challenge isminimizing tritium releases from a fusion machine to

the environment where the allowable releases are aboutone part in a hundred thousand with a system of rea-sonable size and cost The loss of a small fraction of1 of the tritium is not important in terms of fusiontritium breeding but is important in terms of releases tothe environment If one does not efficiently remove thetritium from the salt before the hot salt goes throughthe high-surface-area heat exchanger small amounts oftritium relative to the inventory in the salt leak throughthe heat exchangers but above allowable releasesFurther reductions may need a combination of high-efficiency capture technology (possibly vacuumdisengagers) chemical species control of the tritiumto limit H2 escape (such as a nitrate intermediate loopor oxygen getter in double-wall heat exchangers thatoxidizes tritium to prevent diffusion through metals)or additional barriers to hinder diffusion (such as a heatpipe with permeation membrane or better permeationbarriers) Improvements in any of these domains isbeneficial for both fusion and fission systems

Fig 15 (a) Cumulative tritium generation in 236-MW(thermal) FHR at different 7Li enrichments of 95 995 9995 and99995 and (b) tritium capture on carbon bed at different concentrations

Fig 16 (a) Release rate during operation at different bed surface areas (1x 2x 4x 8x surface area compared to a bed sized 15 mR times 45 m H filled with 09-cm pebbles) and (b) peak release versus the bed surface area



VF Equipment

There is equipment experience with flibe salt The8-MW(thermal) Molten Salt Reactor Experiment(MSRE) in the late 1960s operated successfully usingan overhung centrifugal pump to circulate the primarysalt and clean flibe in the secondary loop The motor wasin the air with the centrifugal pump below the motorThere were no major difficulties with these pumpsbeyond some leakage of oil from seals above into thesalt SINAP plans to operate a small MSR within the next2 years and has a pump development program

Today CSP towers use centrifugal pumps to pumpnitrate salts at temperatures that approach 600degC (Ref 27)The experience with these pumps has been excellentHowever hot nitrate salts are compatible with air andwater whereas flibe salt is not The next generation CSPsystems (Gen-III systems1327) plan to use sodium-potassium-magnesium-chloride salts that require chemicallyreducing conditions and will operate in the temperaturerange of 450degC to 750degC The base-case redox controlagent for these chloride salts is magnesium metal dissolvedin the chloride salt versus beryllium metal dissolved in flibesalt These conditions are similar to those required fora flibe salt system A pilot plant is expected to be builtwithin the next 5 years Work is required to develop largeflibe pumps with massive overlap in requirements to pumpsrequired for Gen-III CSP systems

The MSRE program began to develop valves usinggraphite gaskets for flibe salt operations These werenever deployed Freeze valves were highly successfuland thus the valve program was a low priority We havenot identified other industries that have developed valvesthat would be fully compatible with flibe service but wehave not undertaken any extensive effort to determine ifsuch valves exist in other industrial sectors

VI MARKET CHALLENGES AND IMPLICATIONS FOR FLIBESYSTEMS

VIA Heat Storage

The electricity market is changing because of (1) thegoal of a low-carbon electricity grid and (2) the additionof wind and solar The large-scale addition of wind andsolar results in highly volatile prices with times of low ornegative prices and other times of high prices dependingon wind or solar conditions The goal of a low-carbongrid requires new energy systems that can economicallyprovide the three services of fossil fuels (1) energy

production (2) energy storage and (3) dispatchableelectricity

Adding heat storage to nuclear power plants enablesnuclear power plants to (1) boost revenue in markets withlarge-scale wind and solar and (2) replace fossil fuels inproviding dispatchable electricity including assured peakgenerating capacity while operating the reactor at baseload A recent workshop27 addressed gigawatt-hour heatstorage coupled to Gen-IV nuclear reactors with the par-ticipation of developers of Gen-III CSP systems Thechanges in the market may drive power cycle changesThis has additional implications for fusion First fusionreactors operate in a pulse mode In an inertial fusionmachine this pulse mode may be a fraction of a secondwhereas in the magnetic fusion machines it may be manyhours Heat storage allows the fusion machine to operatein its most efficient mode without considering the impli-cations for the grid Second as discussed above manyheat exchanger and heat storage systems will act to stoptritium escape to the environment There is no require-ment for the fusion community to separately developtechnologies to address these challenges Figure 17shows the system design for heat storage coupled to anylarge-scale heat sourcemdashfission or fusion

To minimize the cost of electricity capital-intensivegenerating assets (fission wind solar and fusion) shouldoperate near their maximum capacity Using this systemwhen electricity prices are high all reactor heat is sent tothe turbine to produce electricity When electricity pricesare low most heat is diverted to heat storage At times ofpeak electricity prices heat from the reactor and heatstorage is sent to the turbine for peak electricity produc-tion that is significantly above base-load reactor electri-city output Peak electricity production can be achievedby (1) oversizing the turbine generator or (2) building

Fig 17 System design for heat storage coupled toa nuclear reactor



a separate peaking steam or gas turbine for peak poweroutput At times of very low electricity prices electricityfrom the grid and from the main turbine operating atminimum load is converted into stored heat withresistance heaters coupled to the heat storage systemThe power plant becomes a seller and buyer of electricityIf heat storage is depleted natural gas or low-carbonbiofuels and hydrogen are used to enable assured peakelectricity production by providing the extra heat thatwould have come from the heat storage system

This has large economic implications for theeconomics of salt power systems The cost of heat storageis a strong function of the average delivered peak tem-perature First the cost of heat storage depends on thetemperature range in the heat storage system If thattemperature range is doubled heat storage costs are cutin half Second the heat-to-electricity efficiency isa strong function of peak temperatures The highertemperatures of salt systems imply much smaller heatstorage per unit of electricity production Table IIIshows the temperature of delivered heat from differentnuclear systems The implication is that a salt-cooledsystem will have an economic advantage relative topower systems that use lower-temperature coolants ifheat storage is required to meet market needs

The primary heat storage materials used today inCSP systems are nitrate salts with solar salt (solar salt60 wt NaNO3ndash40 wt KNO3) the most common saltThese salts are chemically stable in air and waterSensible heat of storage is obtained by typically varyingtemperatures from 290degC to 565degC CSP salts needreasonable margins from decomposition temperatures toavoid hot spots in solar collectors that can degrade thesalt With control of gas compositions over the salt sto-rage tanks and salt chemistry salt storage temperatures inthe 600degC to 650degC range may be possible If the saltsystem is used for backup tritium recovery the off-gassystem must recover tritiated steam Heat storage systemcapital costs in CSP systems are near $20kWh(thermal)or $50kWh(electric) about an order of magnitude lessthan battery storage systems The largest storage systemsizes are measured in gigawatt hours of capacity Nitratesalts can be used to move heat to industrial customers

Nitrate salt storage systems are proposed for SFRs(TerraPower) FHRs (Kairos Power) with solid fuel andclean salt coolants thermal-spectrum MSRs with fueldissolved in the salt HTGRs and fusion machines

Work is underway to develop the next generation ofsalt systems that will allow CSP systems (Gen-III) tooperate at peak temperatures of ~750degC with higher-temperature stored heat The proposed salt is a

sodium-potassium-magnesium eutectic with a meltingpoint near 400degC This salt was chosen because of itsextremely low cost combined with reasonable physicalproperties primarily the melting point Allowable peaksalt operating temperatures could exceed 1000degC Thesesalts are similar to the salts that are used to producemagnesium in electrochemical cells thus there is over-lap in technology with the magnesium industry Thechloride storage salts are proposed to be used withMCFRs (TerraPower) with reactor peak temperaturesnear 750degC The chloride storage salts would couple toflibe fission and fusion systems There is also work oncarbonate salts but at a lower level of effort

VIB Plant Architecture

Heat storage may change reactor power plant systemdesign with the reactor facility inside a security zone thatincludes all vital areas for reactor safety and the storageand power blocks outside this security zone Figure 18shows the plant layout for a CSP or nuclear plant usingsalt storage with this configuration In CSP systems heatstorage simplifies operation On partly cloudy days thepower output of a CSP system varies depending onwhether the clouds are blocking the sun With storage

Fig 18 System design for CSP and nuclear withstorage



the power block does not see such transients and thusstorage simplifies operations Cold salt to the CSP systemalways comes from storage whereas hot salt to the powerblock can come from storage or the CSP system Thesame design can be used for a nuclear power plant asproposed by TerraPower for sodium and salt reactorsTerraPower proposes using nitrate salts to couple withits SFRs and chloride salts to couple to its MCFR

Current reactors put the power block (turbine gen-erator) next to the reactormdasha design that followed thedesign of earlier coal-fired power stations and that wasdeveloped before tight security requirements for nuclearpower plants The separation of the reactor and vital areasfrom the power block creates a clear division betweenareas with (1) requirements for nuclear security mainte-nance licensing safety and construction versus (2) nor-mal industrial requirements This has the potential toreduce costs Second gigawatt-hour heat storage systemsmay become the largest set of structures on site Theywill be in the protected area that has industrial safety andsecurity requirements but in some cases may need to besome distance (100 m) from reactor vital areas Heatstorage systems such as hot salt storage tanks may needto be some distance away because their failure wouldcreate a thermally hot area that could damage buildingsand equipment next to such tanks Last storage isolatesthe reactor from the electricity grid and reduces transientsfrom the grid to reactor and reactor to grid The reactorbecomes a heat generation system

VIC Power Cycles

Salt reactors can be coupled to different powercycles Steam cycles are the near-term option Thelonger-term option is coupling salt reactors to NACCswith heat storage and a thermodynamic topping cycleThis power cycle enables a base-load reactor to providevariable electricity to the grid and can potentially replacenatural gas combined cycle plants Recent papers2058

describe this technologyMolten salt reactors were originally developed in the

1950s as part of the Nuclear Aircraft Propulsion programwhere the goal was to build a jet aircraft with unlimitedrange That is MSRs were chosen because of their abilityto efficiently couple to Brayton power cycles In the1950s gas turbines were inefficient but 60 years ofdevelopment have resulted in dramatic improvements ingas turbine efficiency Salt reactors have two character-istics that enable efficient coupling to Brayton powercycles First in modern gas turbines air is compressedto between 350degC and 450degC Any heat source to couple

to NACCs must provide heat above this temperature Saltreactors typically have minimum salt temperatures near600degC (Table III) Second salt reactors deliver heat ata higher average temperature than other reactor technol-ogies Significant development is required

VII CONCLUSIONS

Research development demonstration and fundingfor fission fusion and CSP have historically been doneby different organizations with results reported in sepa-rate journals and different conferences Technical devel-opments in fission (FHRs MCFRs) fusion (ARC) andCSP (Gen-III) combined with changes in electricity mar-kets (heat storage) in the last 5 years have resulted inmassive technology overlap of high-temperature salt-cooled power systems Going forward this creates largeincentives for cooperative programs across energysources

The ARC and FHR use clean flibe as a coolantThis results in an overlap in technology needs for bothsystems and creates large incentives for cooperativeprograms between fission and fusion to accelerate thedevelopment of these two technologies Flibe blanketsare being considered for other fusion machines andflibe is proposed for several types of MSRs with fueldissolved in the salt In all of these machines there isa strong coupling of salt properties heat transfer tri-tium control and material choices The salts in fluorideand chloride salt systems all have high melting pointsimplying common challenges to prevent salt freezingand similar or identical heat storage and power cyclesThe high temperatures of operation and the need forchemically reducing conditions to limit corrosion implymany similarities in materials selection The salt fissionfusion and CSP programs are not in direct competitionbut rather strategic partners in advanced energysystems

Acknowledgments

We would like to thank ENI the US Department ofEnergy and the SINAP for their support We would like tothank the following individuals for their input and reviewcomments T J Dolan (University of Illinois) R Moir(Lawrence Livermore National Laboratory Retired)P Peterson (Kairos PowerUniversity of California atBerkeley) X Sun (University of Michigan) and M Driscoll(MIT)



ORCID

Stephen T Lam httporcidorg0000-0002-7683-1201

References

1 B N SORBOM et al ldquoARC A Compact High-FieldFusion Nuclear Science Facility and Demonstration PowerPlant with Demountable Magnetsrdquo Fusion Eng Des 100378 (Nov 2015) httpswwwsciencedirectcomsciencearticlepiiS0920379615302337 (current as of Sep 12 2019)

2 A Q KUANG et al ldquoConceptual Design Study for HeatExhaust Management in the ARC Fusion Pilot PlantrdquoFusion Eng Des 137 221 (Dec 2018) httpsdoiorg101016jfusengdes201809007

3 C J BARTON and R A STRELOW ldquoBlankets forThermonuclear Reactorsrdquo ORNL-3258 Oak RidgeNational Laboratory (1962)

4 W R GRIMES and S CANTOR ldquoMolten Salts as BlanketFluids in Controlled Fusion Reactorsrdquo ORNL-TM-4047Oak Ridge National Laboratory (1972)

5 J A BLINK et al ldquoThe High-Yield Lithium-InjectionFusion-Energy (HYLIFE) Reactorrdquo Lawrence LivermoreNational Laboratory (1985) httpswwwostigovservletspurl6124368 (current as of Sep 12 2019)

6 R W MOIR et al ldquoHYLIFE-II AMolten-Salt Inertial FusionEnergy Power Plant DesignmdashFinal Reportrdquo Fusion Technol25 1 5 (1994) httpsdoiorg1013182FST94-A30234

7 A SAGARA et al ldquoBlanket and Divertor Design for ForceFree Helical Reactor (FFHR)rdquo Fusion Eng Des 29 51(1995) httpsdoiorg1010160920-3796(95)80005-I

8 L C CADWALLADER and G R LONGHURST ldquoFlibeUse in Fusion Reactors An Initial Safety AssessmentrdquoINEELEXT-99-00331 Idaho National Engineering andEnvironmental Laboratory (Mar 1999) httpsdoiorg101046j1469-180919996320101x

9 R W MOIR ldquoFlibe Coolant Cleanup and Processing inthe HYLIFE-II Inertial Fusion Energy Power PlantrdquoUCRL-ID-143228 Lawrence Livermore NationalLaboratory (2001)

10 A SAGARA et al ldquoConceptual Design Activities and KeyIssues on LHD-Type Reactor FFHRrdquo Fusion Eng Des81 23ndash24 2703 (Nov 2006) httpsdoiorg101016jfusengdes200607057

11 Molten Salt Reactors and ThoriumEnergy T J DOLAN EdWoodhead Publishing Series in Energy Elsevier Ltd (2017)

12 E MERLE ldquoConcept of European Molten Salt Fast Reactor(MSFR)rdquo CNRS-IN2P3-LPSCGrenoble Institute ofTechnologyUGAmdashFrance (May 23 2017) httpswww

gen-4orggifuploaddocsapplicationpdf2017-05emerle_gif-final23may2017pdf (current as of Sep 12 2019)

13 M MEHOS et al ldquoConcentrated Solar Power Gen3Demonstration Roadmaprdquo NRELTP-5500-67464 NationalRenewable Energy Laboratory (Jan 2017) httpswwwnrelgovdocsfy17osti67464pdf (current as of Sep 12 2019)

14 M ABDOU et al ldquoBlanketFirst Wall Challenges andRequired RampD on the Pathway to DEMOrdquo Fusion Eng Des100 2 (Nov 2015) httpswwwsciencedirectcomsciencearticlepiiS0920379615302465 (current as of Sep 12 2019)

15 M E SAWAN et al ldquoNeutronics Assessment of MoltenSalt Breeding Blanket Design Optionsrdquo presented at 16thTopl Mtg Fusion Energy Madison WisconsinSeptember 14ndash16 2004 UWFDM-1234 University ofWisconsin (Sep 2004) httpftineepwiscedupdffdm1234pdf (current as of Sep 12 2019)

16 M GREENWALD et al ldquoParameter Sensitivities andPhysics Optimization for SPARC-V1C APS-DPPrdquo (Oct21 2019) httpswwwapsorgunitsdppmeetingsannual(current as of Sep 12 2019)

17 D BRUNNER ldquoOverview of SPARC on the High-Field Pathto Fusion Energy APS-DPPrdquo (Oct 21 2019) httpswwwapsorgunitsdppmeetingsannual (current as of Sep 12 2019)

18 C W FORSBERG and P F PETERSON ldquoBasis forFluoride Salt-Cooled High-Temperature Reactors withNuclear Air-Brayton Combined Cycles and FirebrickResistance Heated Energy Storagerdquo Nucl Technol 19613 (Oct 2016) httpsdoiorg1013182NT16-28

19 C W FORSBERG et al ldquoIntegrated FHR TechnologyDevelopment Tritium Management Materials TestingSalt Chemistry Control Thermal Hydraulics andNeutronics Associated Benchmarking and CommercialBasisrdquo MIT-ANP-TR-180 Massachusetts Institute ofTechnology (Oct 2018) httpswwwostigovsearchsemantic1485415 (current as of Sep 12 2019)

20 C ANDREADES et al ldquoDesign Summary of the Mark-IPebble-Bed Fluoride Salt-Cooled High-TemperatureReactor Commercial Power Plantrdquo Nucl Technol 195223 (Sep 2016) httpsdoiorg1013182NT16-2

21 R R ROMATOSKI and L W HU ldquoFluoride Salt CoolantProperties for Nuclear Reactor Applications A ReviewrdquoAnn Nucl Energy 109 635 (2017) httpsdoiorg101016janucene201705036

22 R R ROMATOSKI and L W HU ldquoFluoride-Salt-CooledHigh-Temperature Test Reactor Thermal-HydraulicLicensing and Uncertainty Propagation Analysisrdquo NuclTechnol 205 1495 (2019) httpsdoiorg1010800029545020191610686

23 ldquoOak Ridge Molten Salt Workshoprdquo Oak Ridge NationalLaboratory (2019) httpsmsrworkshopornlgov (currentas of Sep 12 2019)



24 Y ZENG et al ldquoTritium Transport Analysis in a 2-MWLiquid-Fueled Molten Salt Experimental Reactor with theCode TMSR-TTACrdquo Nucl Technol 205 4 582 (2019)httpsdoiorg1010800029545020181507200

25 B FENG et al ldquoCore and Fuel Cycle Performance ofa Molten Salt Fast Reactorrdquo Int Congress on Advancesin Nuclear Power Plants (ICAPP 2019) Juan les PinsFrance May 12ndash15 2019

26 C FORSBERG et al ldquoFlibe Fusion Blankets and FlibeFluoride-Salt-Cooled High-Temperature ReactorsOpportunities for Synergistic Development StrategiesrdquoProc American Nuclear Society Winter MeetingWashington District of Columbia November 17ndash21 2019

27 C W FORSBERG H GOUGAR and P SABHARWALLldquoHeat Storage Coupled to Generation IV Reactors for VariableElectricity from Base-Load Reactors Workshop ProceedingsrdquoProc ANP-TR-185 Center for Advanced Nuclear EnergySystems Massachusetts Institute of Technology July 23ndash242019 INLEXT-19-54909 Idaho National Laboratory (2019)

28 S SMOLENTSEV et al ldquoMHD ThermohydraulicsAnalysis and Supporting RampD for DCLL Blanket in theFNSFrdquo Fusion Eng Des 135 314 (2018) httpsdoiorg101016jfusengdes201706017

29 S SMOLENTSEV et al ldquoAn Approach to Verification andValidation of MHD Codes for Fusion Applicationsrdquo FusionEng Des 100 65 (2015) httpsdoiorg101016jfusengdes201404049

30 C COYLE E BAGLIETTO and C FORSBERGldquoAdvancing Radiative Heat Transfer Modelling inHigh-Temperature Liquid Saltsrdquo 18th Int Topl MtgNuclear Reactor Thermal Hydraulics (NURETH-18)Portland Oregon August 18ndash23 2019

31 N ZWEIBAUM ldquoExperimental Validation of PassiveSafety System Models Application to Design andOptimization of Fluoride-Salt-Cooled High-TemperatureReactorsrdquo PhD Dissertation University of CaliforniaBerkeley (2015) httpsescholarshiporgucitem744882bt(current as of Sep 12 2019)

32 C W FORSBERG et al ldquoTritium Control and Capture inSalt-Cooled Fission and Fusion Reactors Status Challengesand Path Forward (Critical Review)rdquo Nucl Technol 197119 (Feb 2017) httpsdoiorg1013182NT16-101

33 J ZHANG et al ldquoRedox Potential Control in Molten SaltSystems for Corrosion Mitigationrdquo Corros Sci 144 44(2018) httpsdoiorg101016jcorsci201808035

34 E T CHENG H J MERRIL and S Z E DAI-KAIldquoNuclear Aspects of Molten Salt Blanketsrdquo Fusion EngDes 69 205 (2003) httpsdoiorg101016S0920-3796(03)00339-9

35 H MORIYAMA et al ldquoMolten Salts in Fusion NuclearTechnologyrdquo Fusion Eng Des 39-40 627 (1998) httpsdoiorg101016S0920-3796(98)00202-6

36 K R BRITSCH ldquoReview of Fluoride Salt Heat TransferrdquoNucl Technol (submitted for publication)

37 S FUKADA et al ldquoReaction Rates of Beryllium withFluorine Ion for Flibe Redox Controlrdquo J Nucl Mater367-370 Part B 1190 (Aug 1 2007) httpsdoiorg101016jjnucmat200703218

38 B J RILEY et al ldquoMolten Salt Reactor Waste and EffluentManagement Strategies A Reviewrdquo Nucl Eng Des 345 94(2019) httpsdoiorg101016jnucengdes201902002

39 T TANABE et al ldquoHydrogen Transport in StainlessSteelsrdquo J Nucl Mater 123 1568 (1984) httpsdoiorg1010160022-3115(84)90304-0

40 A NAKAMURA S FUKADA and R NISHIUMIldquoHydrogen Isotopes Permeation in a Fluoride Molten Saltfor Nuclear Fusion Blanketrdquo J Plasma Fusion Res 11 25(2015) h t tps pd f s seman t i c scho la ro rg 4d8c d2273bd3e3a99129105ea7e1bc5d6edcaecbpdf (current asof Sep 12 2019)

41 R A CAUSEY R A KARNESKY and C SAN MARCHIldquo416mdashTritium Barriers and Tritium Diffusion in FusionReactorsrdquo Comprehensive Nuclear MaterialsR JMKONINGS Ed pp 511ndash549 Elsevier Oxford (2012)

42 J TROUVE and G LAFLANCHE ldquoHydrogen Diffusionand Permeation in Sodiumrdquo presented at Int Conf LiquidMetal Engineering and Technology in Energy Production(CEA-CONF-7645) Oxford United Kingdom April 9ndash131984

43 S T LAM R BALLINGER and C FORSBERGldquoModeling and Predicting Total Hydrogen Adsorption onPorous Carbon Materialsrdquo J Nucl Mater 511 328 (Dec1 2018) httpsdoiorg101016jjnucmat201809009

44 T ATSUMI and M MIYAKE ldquoAbsorption and Desorptionof Deuterium on Graphite at Elevated TemperaturesrdquoJ Nucl Mater 155ndash157 Part 1 241 (1988) httpsdoiorg1010160022-3115(88)90247-4

45 T DOLAN G R LONGHURST and E G OTEROldquoA Vacuum Disengager for Tritium Removal forHYLIFE-II Reactor Fliberdquo Fusion Technol 21 3P2B1949 (May 1992) httpsdoiorg1013182FST21-1949

46 S LAM et al ldquoWeak and Strong Hydrogen Interactions onPorous Carbons Materials in High-Temperature SystemsrdquoJ Nucl Mater 519 173 (2019) httpsdoiorg101016jjnucmat201903036

47 C W FORSBERG and P F PETERSON ldquoFHR HTGRand MSR Pebble-Bed Reactors with Multiple Pebble Sizesfor Fuel Management and Coolant Cleanuprdquo NuclTechnol 205 5 748 (2019) httpsdoiorg1010800029545020191573619

48 X WU et al ldquoMass Transport Analysis for TritiumRemoval in FHRsrdquo Ann Nucl Energy 121 250 (2018)httpsdoiorg101016janucene201807026



49 X WU et al ldquoPrimary Design of a Tritium Control System forFluoride-Salt-Cooled High-Temperature Reactor SystemsrdquoTritium Charleston South Carolina April 17ndash22 2016

50 S ZHANG et al ldquoA Coupled Heat Transfer and TritiumMass Transport Model for a Double-Wall Heat ExchangerDesign for FHRsrdquo Ann Nucl Energy 122 328 (2018)httpsdoiorg101016janucene201808039

51 S ZHANG et al ldquoDouble-Wall Natural Draft HeatExchanger Design for Tritium Control in FHRsrdquo Proc2017 25th Int Conf Nuclear Engineering (ICONE25)Shanghai China July 2ndash6 2017

52 C FORSBERG et al ldquoFluoride-Salt-CooledHigh-Temperature Reactor (FHR) Using British AdvancedGas-Cooled Reactor (AGR) Refueling Technology andDecay-Heat Removal Systems that Prevent Salt FreezingrdquoNucl Technol 205 9 1127 (Apr 2019) httpsdoiorg1010800029545020191586372

53 B ZOHURI et al ldquoHeat-Pipe Heat Exchangers for Salt-Cooled Fission and Fusion Reactors to Avoid Salt Freezingand Control Tritium A Reviewrdquo Nucl Technol httpsdoiorg1010800029545020191681222

54 J D STEMPIEN ldquoTritium Transport and Corrosion Modelingin the Fluoride Salt-Cooled High-Temperature Reactorrdquo PhDThesis Massachusetts Institute of Technology (2015)

55 S T LAM ldquoManaging Tritium Inventory and Release withCarbon Materials in a Fluoride Salt-CooledHigh-Temperature Reactorrdquo SM Thesis MassachusettsInstitute of Technology (2017)

56 S T LAM et al ldquoTritium Management and Control UsingCarbon in a Fluoride-Salt-Cooled High-TemperatureReactorrdquo Fusion Sci Technol 71 4 644 (2017) httpsdoiorg1010801536105520171290945

57 C FORSBERG J STEMPIEN and R BALLINGERldquoTritium Removal from Salt-Cooled Reactors UsingCarbonrdquo Proc American Nuclear Society Winter MeetingWashington District of Columbia November 8ndash12 2015

58 C W FORSBERG P J McDANIEL and B ZOHURIldquoBase-Load Nuclear and Concentrated Solar Power (CSP)Air-Brayton Power Cycles with Thermodynamic ToppingCycles and Heat Storage for Variable Electricity and HeatrdquoProc ICAPP 2020 Abu Dhabi United Arab EmiratesMarch 15ndash19 2020



Abstract

I INTRODUCTION


III FHRs AND LIQUID-FUELED MSRs

IV DIFFERENCES BETWEEN FISSION AND FUSION FLIBE USE



IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment

VI MARKET CHALLENGES AND IMPLICATIONS FOR FLIBE SYSTEMS

VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References

Fusion Blankets and Fluoride-Salt-Cooled High-TemperatureReactors with Flibe Salt Coolant Common Challenges TritiumControl and Opportunities for Synergistic DevelopmentStrategies Between Fission Fusion and Solar Salt TechnologiesCharles Forsberg Guiqiu (Tony) Zheng Ronald G Ballinger and Stephen T Lam

Massachusetts Institute of Technology Department of Nuclear Science and Engineering 77 Massachusetts AvenueCambridge Massachusetts 02139

Received September 12 2019Accepted for Publication November 7 2019

Abstract mdash Recent developments in high-magnetic-field fusion systems have created large incentives to developflibe (Li2BeF4) salt fusion blankets that have four functions (1) convert the high energy of fusion neutrons intoheat for the power system (2) convert lithium into tritiummdashthe fusion fuel (3) shield the magnets againstradiation and (4) cool the first wall that separates the plasma from the salt blanket Flibe is the same coolantproposed for fluoride-salt-cooled high-temperature reactors that use clean flibe coolant and graphite-matrixcoated-particle fuel Flibe is also the coolant proposed for some molten salt reactors (MSRs) where the fuel isdissolved in the coolant The multiple applications for flibe as a coolant create large incentives for cooperativefusion-fission programs for development of the underlying science design tools technology (pumps instrumenta-tion salt purification materials tritium removal etc) and supply chains Other high-temperature molten saltsare being developed for alternative MSR systems and for advanced Gen-III concentrated solar power (CSP)systems The overlapping characteristics of flibe salt with these other salt systems create significant incentives forcooperative fusion-fission-solar programs in multiple areas

We describe the fission and fusion flibe-cooled systems what has created this synergism what isdifferent and the same between fission and fusion in terms of using flibe and the common challenges Wereview (1) the characteristics of flibe salts (2) the status of the technology (3) the options for tritiumcapture and control in the salt heat exchangers and secondary heat transfer loops and (4) the coupling topower cycles with heat storage The technology overlap between flibe systems and other high-temperatureMSR and CSP salt systems is described This defines where there are opportunities for cooperativeprograms across fission fusion and CSP salt programs

Keywords mdash Salt-cooled reactors fusion flibe salt fluoride-salt-cooled high-temperature reactor moltensalt reactor

Note mdash Some figures may be in color only in the electronic version

I INTRODUCTION

This paper describes the fission and fusion applicationsfor flibe salt with overlapping areas of science technologychallenges and future directions Figure 1 shows the pro-posed affordable robust and compact (ARC) fusion

E-mail cforsbermiteduThis is an Open Access article distributed under the terms of theCreative Commons Attribution-NonCommercial-NoDerivatives Lice-nse (httpcreativecommonsorglicensesby-nc-nd40) which permitsnon-commercial re-use distribution and reproduction in any mediumprovided the original work is properly cited and is not altered trans-formed or built upon in any way

NUCLEAR TECHNOLOGYcopy 2019 The Author(s) Published with license by Taylor amp Francis Group LLCDOI httpsdoiorg1010800029545020191691400

1



























TABLE I




628 MW(thermal)










TABLE II


Parameter Value




















TABLE III


Coolant


(degC)

AverageCore Exit

Temperature(degC)










IVC Tritium






V FLIBE SYSTEMS




VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References



























TABLE I




628 MW(thermal)










TABLE II


Parameter Value




















TABLE III


Coolant


(degC)

AverageCore Exit

Temperature(degC)










IVC Tritium






V FLIBE SYSTEMS




VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References




TABLE II


Parameter Value




















TABLE III


Coolant


(degC)

AverageCore Exit

Temperature(degC)










IVC Tritium






V FLIBE SYSTEMS




VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References













TABLE III


Coolant


(degC)

AverageCore Exit

Temperature(degC)










IVC Tritium






V FLIBE SYSTEMS




VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References






IVC Tritium






V FLIBE SYSTEMS




VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References

VA Supply Chain


VB Properties







VC Corrosion


TABLE IV
































VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References


























VD1a Gas Sparging








































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References
































VD3c Heat Pipes














































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References








































VF Equipment





VIA Heat Storage





















VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References














VIC Power Cycles





VII CONCLUSIONS



Acknowledgments




ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References

ORCID


References


































































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References








































Abstract

I INTRODUCTION






IVC Tritium

V FLIBE SYSTEMS

VA Supply Chain

VB Properties

VC Corrosion







VF Equipment


VIA Heat Storage


VIC Power Cycles

VII CONCLUSIONS

Acknowledgments

References

technologies charles forsberg, guiqiu (tony) zheng, ronald ...xylenepower.com/nt fission fusion...

Documents