Liquid Fluoride Reactors: A New Beginning for an Old Idea

Feb 19th 2009

Google Tech Talk

Dr. David LeBlanc

Physics Dept, Carleton University, Ottawa& Ottawa Valley Research Associates Ltd.

d_leblanc@rogers.com

Can Nuclear Help Solve our Energy and Climate Change Problems?

The nuclear industry would be more than happy with building 400 reactors over the next 30 years But that merely keeps the status quo

Thousands of large reactors are needed to truly make a difference.

Is there any way that existing or “next generation” designs can realistically accomplish this?

Seven Reactor Design Priorities Overall Power Costs

Capital Costs Fuel plus O&M

Safety Resource Utilization

Quantity and Quality of Starting Load Annual Requirements

Proliferation Resistance Subversion to a State Program Theft and/or Terrorist Take Over

Long Term Waste Radiotoxicity R&D Requirements Capability for Rapid Deployment

PWRs, BWRs and CANDUs Overall Costs

High capital but arguably competitive with coal Safety

Good through costly “Defence in Depth” Resource Utilization

Moderate for startup, Very poor annual Proliferation Resistance

Good if international safeguards maintained Long Term Wastes

Once Through cycle extremely poor, Pu recycle not much better

R&D - Very Little Required Capability for Rapid Deployment

Labour requirements alone limit growth

What About Gas Cooled Pebble Beds?

Costs savings possible but no great reduction in overall complexity

Massive pressure vessel and needs of decay heat removal limit core size and power

Rad Waste production very similar (with radioactive graphite as an added bonus)

Reprocessing of fuel extremely difficult, so an even worse outlook for uranium resources

From safety outlook, excellent for decay heat management but criticality events still a possible issue*

*The reverse situation is true for Accelerator Driven Systems whichguard against criticality accidents but do not address decay heat

What about other Generation IV Reactors?

Supercritical Water (LWR or CANDU) Merely modest improvements possible and extremely

challenging corrosion issues to solve Sodium Cooled Fast Breeders

More costly and higher safety concerns ~12 tonne fissile requirement Doubling time upwards of 40 years

Lead Cooled Fast Breeders Possible lower capital costs ~20 tonne fissile requirement makes rapid fleet

expansion even less likely* Gas Cooled Fast Breeders

Similar cost advantages and fissile needs of Lead Cooled Fast Breeder

*Entire U.S. supply of Plutonium from spent fuel and weapons program could only start 30 to 40 such reactors

What about Thorium?We keep hearing about that

Yes, fertile thorium producing fissile 233U has numerous advantages

Works well in any spectrum, not just a fast one like U-Pu

However any use in solid fuels, involves very difficult fuel reprocessing

What about fluid fuels? What about the Gen IV reactor that was made for thorium?

The Liquid Fluoride Reactor* Over 50 years old and never looked better

* Historically known as Molten Salt Reactors

Liquid Fluoride / Molten SaltOverview

What did the original Molten Salt Reactor Program at Oak Ridge National Labs (1950s to 1970s) end up with?

What factors guide design, how did the design evolve and why was it cutoff?

What has happened since?

Do modified priorities and new ideas change the outlook?

The Single Fluid, Graphite Moderated Molten Salt Breeder Reactor (MSBR)

General Attributes of the “Traditional” 1000 MWe Single Fluid MSBR

Design 27LiF-BeF2 - 12% ThF4 - 0.3% 233UF4

700 oC Outlet Temperature 44.4% on Steam Rankine cycle Up to 48% on Gas Brayton cycle Fission Products removed on 20 day

cycle 233Pa* removed on 10 day cycle Startup Fissile Inventory of 1500 kg Breeding Ratio of 1.06** 20 year doubling time

* 233Pa is the 27 day half-life intermediate between thorium and 233U** Meaning it produces 6% more fissile fuel than it consumes

General Benefits of Any Molten Salt Design Volatile fission products including Xenon

simply bubble out*, are removed and stored elsewhere. A spill will simply solidify

Salts have high boiling point and operate at low pressure

A temp increase instantly lowers reactivity Fuel salt at the lowest pressure of the

circuit, the opposite of a LWR Control rods or burnable poisons not

required. Almost no excess reactivity A simple freeze plug melts upon fuel

overheating to drain the salt to critically safe, passively cooled dump tanks

*Assisted by bubbling helium through the flowing salt

General Benefits of Any Molten Salt Design

Inherently Passive Safety Low Fissile Inventory Very High Thermal Efficiency Ideal for LWR TRU waste destruction Ability to use closed thorium cycle

Only consume 800 kg thorium per GWe year TRU (transuranic = Np, Pu, Am, Cm) waste

production extremely low Much lower long term radiotoxicity

Radiotoxicity PWR vs FBR* vs LFR* *Assuming 0.1% Loss During Processing

Data and graph from Sylvain David, Institut de Physique Nucléaire d'Orsay

PWR OreLevels

Turns waste management into 500 year job, not million year

FPs Fission

Products

Proliferation Resistance The Pure Thorium – 233U Cycle

232U present in significant quantities 69 year half-life with strong 2.6 MeV gamma ray

from daughter product 208Tl Makes illicit use difficult and highly detectable No national program ever based on 233U

233U can be instantly denatured by dumping 238UF4 into the molten fuel salt

233Pa removal can lead to “clean” 233U and thus should be avoided

Only small amounts Plutonium are present, it is of poor quality and very hard to extract

*HEU=Highly Enriched Uranium (>12% 233U), LEU=Low Enriched Uranium

Proliferation ResistanceDenatured Cycles

The pure Th-233U cycle does represent the use of Highly Enriched Uranium

Running denatured by including 238U makes any uranium useless for weapons

It does mean more plutonium present but still of poor quality and hard to remove

In many designs the conversion ratio will suffer somewhat

Quality of Produced PlutoniumIsotope Proliferation

propertiesPWRReactorGrade

DMSR30 Year Once Through

MSBRPure Th – 233U cycle

238Pu Generates heat from alpha emission

1.3% 12.6% 73%

239Pu Main fissileComponent

60.3% 31.1% 9.5%

240Pu Spontaneous fissions high

24.3% 18.1% 4.4%

241Pu Fissile and adds hard gamma rays

5.6% 13.6% 4.8%

242Pu fertile 5.2% 24.3% 7.4%

Problems Specific to the Single Fluid, Graphite Moderated MSBR

Limited graphite lifetime (4 years) Fuel processing hindered by chemical

similarity of thorium and the rare earth fission products

Problem with temperature reactivity coefficient recently discovered

233Pa separation adds a unique proliferation concern

Design Basics:2 Fluid Versus Single Fluid

A 2 Fluid design has separate fluids A Fuel Salt containing 233UF4

A Blanket Salt containing ThF4

Advantages; Absence of thorium in fuel salt makes

fission product removal far easier 233Pa removal unnecessary* Strongly negative temperature and void

coefficients for the fuel salt

*Pa is effectively diluted in the blanket salt so it experiences a lowerneutron flux and has a much greater probability to decay to 233U

Converter or “Once Through” Designs

Use a mix of Low Enriched Uranium and thorium in a Single Fluid

No need for any processing of fuel salt for the life of the plant

Not able to “Break Even” but require only a small fraction the annual uranium compared to LWR, CANDU or PBMR

Choice of Neutron Spectrum Harder Spectrum

Losses to fission products drop so may leave in longer

Prompt neutron lifetime is shorter Typically requires higher fissile load

Softer Spectrum Fissile inventory can be very low Losses to fission products highest Any extra structural material must be

very limited or breeding not possible Prompt neutron lifetime is increased

Fuel Processing Basics

Fluoride Volatility Process Removes uranium from carrier salt

Fission Product Removal Essential if breeding desired Several year cycle possible in some

designs Protactinium Removal and Hold Up

Essential only for Single Fluid designs wishing to breed or break even

Fluoride Volatility

Simple bubbling of fluorine gas through carrier salt

Converts UF4 to gaseous UF6

Conversion back to UF4 done by adding H2

Well established process, known since 1950s

Fission Product Removal

1950 to 1964 Several potential methods but none

proven to be practical 1964 Vacuum Distillation

Simple system to evaporate off carrier salt and leave behind fission products

Thorium would remain with fission products, so only for a 2 fluid design with no thorium in the fuel salt

Fission Product Removal

1968 Liquid Bismuth Reductive Extraction Ability to process for fission products

with thorium present Thorium behaves much like rare earth

fission products which makes the process very “delicate”

Complex and costly but works for Single Fluid designs

Fuel Processing Summary

For 2 Fluid design, choice of: Vacuum Distillation Salt Replacement Simplified Liquid Bismuth Extraction Can skip Pa removal since it resides in

the low neutron flux blanket For Single Fluid Design

Multi-staged Liquid Bismuth Extraction Pa removal very difficult to avoid

Skip altogether for converter design

Design Evolution at Oak Ridge National Laboratories

(ORNL)

Crucial to remember that early design priorities very different

Long Term Wastes and Proliferation Resistance low priority

Uranium resources thought scarce, thus must “breed” fuel to expand reactor fleet

Short Doubling Time* top priority in competition with Sodium Cooled Fast Breeder

*Time to breed enough excess fuel to start another reactor

A Strange BeginningAn Aircraft Reactor?

The Aircraft Reactor Experiment

Test reactor of early to mid 1950s Very high temperature 860 oC Canned BeO moderator NaF-ZrF4 carrier salt Points the way to possible power

reactors (even if the idea of an airborne reactor far fetched)

The Rise and Fall of the 2 Fluid Reactor The Late 1950s

Main focus on homogeneous reactors (no graphite or other solid moderator)

Looked at both converter reactors and thorium-233U breeders

Carrier Salt itself provides significant neutron moderation so a variety of neutron spectrums possible

All studies were spherical geometry with 1/3 inch Hastelloy N core walls

Homogenous Molten Salt Reactor Late 50s ORNL

The Rise and Fall of the 2 Fluid Reactor Early 1960s

Graphite now proven to be compatible with fluoride salts

Lower possible fissile inventory leads ORNL to change to a graphite moderated 2 Fluid design

Simplist Sphere-Within-Sphere 2 Fluid design would only allow ~ 1 meter core* Obviously too low a total power output

Complex fluid intermixing deemed necessary

*Without having the 233UF4 fraction so low as to preclude breeding

1960`s, 2 Fluid, Graphite Moderated MSBR

Intermixing of Fuel and Blanket salt in core done by graphite “plumbing”

Allows large core diameter for adequate power production

Graphite first shrinks and then swells under neutron irradiation

Extremely difficult “Plumbing Problem” to solve

Meanwhile, also in the mid 60s…Molten Salt Reactor Experiment MSRE

MSRE 8 MW(th) Reactor Chosen to be Single Fluid for simplicity Graphite moderated, 650 oC operation Designed from 1960 to 1964 Start up in 1965 Ran very successfully for 5 years Operated separately on all 3 fissile

fuels, 233U , 235U and Pu Some issues with Hastelloy N found and

mostly resolved in later years

1968: The Start of the Single Fluid Era

New Liquid Bismuth Reductive Extraction, while more difficult, can work with thorium in the fuel salt

2 Fluid concept abandoned “Plumbing Problem” left unsolved Major funding for MSBR cancelled in

early 70s. R&D extremely limited until quite recently

2 Fluid concept largely forgotten

Why was the program cancelled?First, what did ORNL do right?

Early decision to focus on Th-U cycle instead of U-Pu No need for a fast spectrum, less radiotoxicity

much better resource potential of thorium over uranium

Brought the top minds together in one place Engineers, chemists, physicists constantly

update and reinforce each others efforts Reactor safety kept at the forefront

Inherently passive over engineered solutions Early recognition that events of low probability

must be planned for

What did ORNL do wrong? Early decision to focus on Th-U cycle

instead of U-Pu In 50s to 70s, production of Pu still a top military,

government priority and by extension the AEC Brought the top minds together in one

place MSR almost unknown to experts outside Oak Ridge.

Government relies on opinions of experts from across the country

Reactor safety kept at the forefront Oak Ridge director and inventor of the PWR, Alvin

Weinberg drew the ire of the AEC`s infamous Milton Shaw by raising safety concerns of LWRs

Weinberg fired as director and funding for Molten Salt program cut off in the early 1970s. Topic almost forbidden since then by AEC and later DOE

Post Program Developments Late 70s ORNL work focuses on Denatured

designs termed DMSRs Little funding but many vocal advocates

worldwide until being chosen as one of six Gen IV reactors in 2002

Major re-examination by large group in France Find a reactivity problem with 70s MSBR Look to decrease processing needs by allowing

breeding ratio to drop Move towards a harder spectrum without

graphite and add a radial blanket salt U.S. based effort to examine the use of clean

salts as coolant for TRISO type fuels. Many advantages if positive void can be avoided

Much worldwide work on transuranic waste destruction using Molten Salt designs

The French TMSR Thorium Molten Salt Reactor

Design has a thorium blanket but only radial, not axially (which would be very difficult)

Russian MOlten Salt Actinide Recycler and Transmuter MOSART

Updating Priorities

Fissile material for starting a large fleet of liquid fluoride reactors not a problem so no need to “breed”

Proliferation Resistance and Long Term Waste Reduction now high priorities

Should minimise the need for fuel processing and its complexity

No two experts or two nations will rank priorities the same so multiple options best

Keeping in Mind…

KISS, Keep It Simple Stupid Idea people often drawn to complex

solutions New improvements are extremely

expensive to verify in the nuclear field OTS Off The Shelf where possible RBB Rice Bowl Breaking

Unavoidable with new concepts but be aware and minimise if possible!

New Ideas for Your Consideration

Reviving the 2 Fluid Concept

A Barrier Free, Possibly Fuel Processing Free Design

Simplified Converters, expanding on ORNL`s “30 Year Once Through” design

Restating the 2 fluid Plumbing Problem

Sphere-Within-Sphere is simplest approach (or short cylinders)

BUT, if inner fuel salt lacks thorium then the critical diameter must remain small (approx 1 m with or without graphite)

Too small for power plants “Standard” conclusion was fuel and

blanket salts must be interlaced

What is the solution? Here’s a hint…

Geometry Buckling2 Ratio to Bsphere

Sphere 1

Infinite Cylinder 0.766

Finite Cylinder + If H = 10 R

0.772

Infinite Slab 0.5

2

R

2405.2

R

2405.2

R

2

H

2

a

Modified Geometry 2 Fluid Reactor* “Tube-Within-Shell”

*Patent Pending

Expands power producing volume while maintaining thesmall inner core needed for a simple 2 Fluid design

New Concepts Advantages Can use simple 2 fluid fuel processing

without the “plumbing problem” Very strongly negative fuel salt

coefficients Blanket will also have negative

temp/void coefficient as it acts as a weak reflector

Simple, transportable cores Ease of modeling and prototyping Fissile inventory of 400 kg per GWe or

even lower is possible.

Cross Section Graphite Free

Example: Graphite Free,Carbon or SiC composite for barrier

Using ORNL modeling for a 122 cm “spherical” core, 0.16% 233UF4 should be able to reach Break Even Breeding

A 122 cm sphere equates to 94 cm diameter in elongated cylindrical geometry

Assuming; Core power density of 200 kW/L 2 m/s salt velocity in core Standard 565 C/705 C for Inlet/Outlet Temp

Gives 404 MWe (911 MWth), 6.6 m core

Other Variations

Modestly higher concentration of 233UF4 (0.2% to 1%) gives excess neutrons to allow: Metal barriers such as Hastelloy N,

Stainless Steels, Molybdenum Alternate carrier salts to reduce costs

and tritium production Even greater simplification of fission

product processing. 20 year or longer removal time for fission products

Critical Issue:Core-Blanket Barrier

Viability of barrier materials in high neutron flux Much recent work in the fusion field

using same 27LiF-BeF2 salt as coolant Molybdenum, SiC/SiC or simple carbon

composites leading candidates Hastelloy N and Stainless Steels

possible with a modest temp reduction Ease of “retubing” means even a

limited lifetime still may be attractive

Fusion Structural Materials Studied

“Operating Temperature Windows for Fusion Reactor structural Materials”

Zinkle and Ghoniem, 2000

Graphite Cores

Using graphite in the core leads to a similar inner critical diameter

Many advantages but much lower power density than pure salt cores

The following embodiment does not require an encompassing barrier and only graphite experiences a significant neutron flux

Barrier Free, Fuel Processing Free? Can we run break even without a barrier

AND without fuel processing? A harder spectrum without graphite

lowers losses to fission products but neutron leakage can be excessive without a blanket salt and barrier

As well, the outer vessel wall needs protection against a high neutron flux

A common solution is a graphite liner/reflector This leads to an increase in power near

reflector and limited graphite lifetime Can surprisingly increase neutron leakage

Boron often added at outer edge of the graphite reflector to stop final neutrons

Power Distribution With and Without Graphite Reflector

Reproduced from “Transmutation Capability of Molten Salt Reactors Feed with TRUs from LWR” M. Frantoni and E. Greenspan, AWRIF 2005

Liner = Blanket Concept Core salt is Single Fluid (potentially

denatured with 238U) In early years of operation the liner is

completely absorbing but there are excess neutrons since few fission products

As fissile 233U (and/or Pu) build up, the liner begins to send back neutrons which helps compensate for losses to fission products

Ideally the liner is permanent until the end of plant life or perhaps a major mid-life refurbishment. Contained fissile in liner and salt can then be collected for the replacement plant

Liner = Blanket Concept Concept is at very early stage and

requires modeling Graphite lifetime and ability to forego fuel

processing increase with core diameter since power density drops

Question is what diameter would give a full 30 years or more for the graphite and be able to run without fuel processing

Too large and/or too high a fissile concentration may mean unacceptably large starting fissile load

DMSR Converter Reactors Starting Premise is Oak Ridge`s 30

Year Once Through Design (1980)* 1000 MWe output Startup with LEU (20% 235U) + Th No salt processing, just add small

amounts of LEU annually Low power density core gives 30 year

lifetime for graphite (8m x 8m) Similar startup fissile load as LWR

(3450 kg/GWe) Averages a Conversion Ratio above 0.8

*ORNL TM 7207 and dozens of other relevant documents available at http://thoriumenergy.blogspot.com/

Denatured Molten Salt Reactors

1810 tonne Lifetime U Ore Requirement 6400 tonnes for LWR 4900 tonnes for CANDU

At end of 30 years Uranium easily removed and reused which

would drop the lifetime ore to under 1000 t Transuranics should also be recycled ~1000 kg TRUs in salt at shutdown Assuming typical 0.1% processing loss,

just 1 kg in 30 years! Same low long term radiotoxicity as pure Th-233U cycle

Suggested Improvements Graphite Pebbles as moderator

Removes need for flux flattening Can go to smaller, higher power core Pyrolytic coatings for increased safety

Higher power density, smaller cores Ex, double and replace every 15 years Much lower starting inventory and

better uranium utilization Elongated Cylindrical Geometry

Eases fabrication and dealing with axial plenums. Tricks to limit neutron leakage

Suggested Improvements Graphite Free “Tank of Salt” Core

Can retain softer spectrum by having very low fuel concentration and let the carrier salt act as moderator (Be, Li, F)

Alternate carrier salts NaF-BeF2 low cost, low melting point

May allow stainless steel throughout loop NaF-RbF low cost, no tritium production

Simplification of entire primary loop NaF very low cost, higher melting point

These alternatives will only roughly double the neutrons absorbed in the salt from 1.5% to 3%

Reactor

Conversion Ratio

LifetimeUraniumOre (t)

AnnualUraniumOre (t)

AnnualOre Costs 50$/kg U

AnnualOre Costs5000$/kg

LWR 0.5-0.6 6400 200 8.5 million

850 million

LWR with U-Pu Recycle

0.5-0.6 4080 125 5.3 530

Sodium Fast Breeder

1.3 2400If start up on 235U

1

DMSR Converter

0.8 1820 35 1.5 150 0.02$/kwh

DMSR single U recycle

0.8 1000 35 1.5 150

DMSR alternate salt, triple the neutron losses

0.74 2140 47 2.0 200

Based on 0.2% tails, 75% capacity factor, 30 year lifetime

LWR data from “A Guidebook to Nuclear Reactors” A. Nero 1979

3.9 million$ annual enrichment costs for DMSR at 110$/SWU, Total Fuel Costs <0.001$/kwh

At $5000/kg, uranium from sea water potentially feasible and unlimited resource

Peak Uranium?

Attractiveness of converter designs hinges on sustainability of uranium resources

As shown however, even an enormous price increase has little effect on fuel costs for a DMSR

DMSR also can be viewed as a first step with fuel processing added later if uranium is scarce

Conclusions

Molten Salt designs have inherent features that favour overall safety, waste reduction, low cost and rapid deployment

They also have great flexibility to match varying priorities Can attain much higher levels of

proliferation resistance compared to current offerings

Can run on minute amounts of thorium, or modest amounts of uranium for the utmost in simplicity

What Way Forward?

Corporate interest will always be difficult to attract No lucrative fuel fabrication contracts Min 15 year return on investment a

tough sell to shareholders (no matter how big the return may be)

Existing nuclear players have their choices in place

What Way Forward? Other Corporate Players?

Big Oil For a small fraction of current profits, can

retain their position in the energy market after “Peak Oil”

Chemical Giants A majority of the needed R&D and engineering

work would fit their skill set

Individuals with Deep Pockets? What better way for those such as Gates,

Branson, Allen, Buffet to invest in the future

What Way Forward? International Cooperation is key way

to spread the costs and rewards ITER model as rough guide but with

greater corporate involvement Likely no one design will be best for

all nations or utilities so best to move forward on several versions 95% of R&D needed would serve entire

community Nothing like competition to yield the

best results

What is Needed Short Term

Neutronic modeling Fuel Salt chemistry and corrosion studies

of various carrier salts and materials for heat exchangers or potential 2 Fluid barriers

Non-nuclear component testing of pumps, valves, heat exchangers etc.

Minor levels of funding to support these efforts (the hardest part of all!)

EXTRA SLIDES…

Two Region Homogeneous ReactorProjected breeding ratios assume thicker blanket and

alternate barrier. From ORNL 2751, 1958

Core Diameter

3 feet 4 feet 4 feet 8 feet

ThF4 in fuel

salt mole % 0 0 0.25 7

233U in fuel salt mole %

0.592% 0.158% 0.233% 0.603%

Salt Losses 0.087 0.129 0.106 0.087Core Vessel 0.090 0.140 0.109 0.025Leakage 0.048 0.031 0.031 0.009Neutron Yield 2.193 2.185 2.175 2.20Breeding ratio (Clean Core)

0.972 0.856 0.929 1.078

Projected B.R.(thinner wall)

1.055 0.977 1.004 1.091

Projected B.R.(carbon wall)

1.105 1.054 1.066 1.112

Graduate (or 4th year project) topics all would generate “publishable” results

Duplicate/refine/check early ORNL documents/conclusions (e.g., ORNL 2751) with a modern neutronic code

Determine (calculate) neutron energy distribution/total flux at the barrier of a hypothetical 0.4 GWe, tube-in-shell, 2-salt reactor – results relevant to its BR, tube durability/lifetime, and proliferation resistance (232U generation)

Determine (calculate) that reactor’s long-term BR and fissile requirements as a function of salt clean-up frequency ( i.e., FP/234U build-up in fuel salt - do not assume a salt-cleanup technological “breakthrough”

Ditto assuming that Th is added to its fuel salt (rendering it a 1½ salt system) - vary the concentration of that Th over a wide range

Ditto assuming that the 2-salt system’s uranium (both salt streams) must be “denatured” (233U/Utotal ≤ 12%) with 238U

Calculate the consequences of denaturation (how it affects radwaste generation, the reactor’s BR, and fissile requirement)

Design a barrier material-relevant test consistent with ATR’s* capabilities/limitations

Design a two-salt breeder with a simple-to-replace reactor tube

*ATR is the Advanced Test Reactor of Idaho National Laboratories. This slide is from Dr. Darryl Siemer (retired) who is attracting student interest at Idaho State University