PART 1: Liquid Fuel Nuclear Reactors introduced the history and technology of liquid fuel nuclear reactors – the path not taken as the world followed Rickover’s forceful choice of solid fuel reactors.

Cost innovation at SpaceX

SpaceX's key innovation is cost reduction. Harvard Business Review wrote about SpaceX and founder Elon Musk. “Musk quickly zeroed in on the one area ripe for innovation: cost reduction. … The complexity of the Shuttle and its reliance on 1970's technology drove costs up. … To build a business for the long haul, SpaceX wisely recognized it must embark on the complicated and risky task of developing an entirely new rocket engine. … With small teams and far lower overhead, SpaceX was able to go from incorporation to first space flight in six years.”

SpaceX developed its Falcon 1 rocket at an order-of-magnitude lower cost – only $443 million, compared to NASA estimates of $1.4 to 4 billion. Can energy cost innovation reap similar benefits for nuclear power, still using technologies birthed in the 1970s?

Cost innovation with molten salt reactors

The molten salt reactor promises low power costs. Weinberg’s Oak Ridge associates termed it Alvin’s 3P reactor – a Pot, a Pipe, and a Pump -- because the liquid fuel reactor was so conceptually and physically simple. It is certainly difficult and venturesome to predict the future cost of such an emerging technology. The development costs will not be clear until the pilot plant is running, and production costs will depend on industry experiences with factory manufacturing. However many factors support low costs.




2012 $/watt

Sargent & Lundy




Sargent & .. ORNL TM-1060




Kasten, MOSEL reactor








McNeese et al, ORNL-5018




Engel et al, ORNL TM7207













Historical estimates. This table presents seven cost estimates to build MSRs. The $/watt is the cost of research and development and construction. The last column is inflation-adjusted to 2012 dollars. This suggests that $2/watt is a reasonable goal for commercially produced, standardized power reactors in commercial production, after initial research, development, and testing has been completed.

Atmospheric pressure operation. Radioactive materials in a MSR are near atmospheric pressure. There is no requirement for a steam-explosion-proof reinforced concrete containment dome, nor need for high-pressure piping, valves, and pressure vessels such as an LWR requires. This reduces costs for nickel-alloy Hastelloy piping, fittings, valves, pumps and other materials. It also simplifies safety engineering, for there are no pressurized radioactive materials that could be propelled from the reactor into the environment in a severe accident.

Thermal stability. As the molten salt heats and expands, the density of fissile material is reduced and the chain reaction slows. Rising temperatures increase neutron absorption and lower fission probabilities, slowing the reaction. Operational neutron-absorbing control rods are not necessary. The simple backup safety freeze plug melts at high temperatures when its active cooling stops, dumping the fuel salt into specially configured tanks where the chain reaction stops.

High thermal efficiency.MSR’s high 700°C outlet temperature enables 45% efficient Rankine or Brayton cycle power conversion, compared to LWR efficiency of 33% operating at 315°C [or to LMFBR 40% efficiency operating at 500°C]. This means that the MSR can deliver 45/33 the electricity for a thermally comparable LWR. The rejected heat dissipated by the reactor cooling system is reduced by 39% for the same electrical power, reducing costs for cooling towers, or alternatively permitting dry air cooling.

Small size. Salts are excellent coolants so the reactor is compact, reducing mass and costs. The genesis of the MSR was a nuclear reactor small and light enough to sit in an airplane fuselage, as shown in Figure 3. The airplane jet engine was the low-mass, compact, Brayton-cycle power-conversion system.

Heat capacity. The high molten salt heat capacity exceeds that of the water in LWRs or liquid sodium in LMFBRs, allowing compact geometries and heat transfer loops that make the reactor more compact, requiring less material such as Hastelloy-N or SAE 316 stainless steel, lowering materials cost.

Neutron efficiency. Neutrons are sometimes lost from the nuclear reactor chain reactions through parasitic absorption by internal structures, burnable poisons, and leakage. These amount to just 5% in DMSR, compared to 15% in LMFBRs or 22% in LWRs. Fission product Xe-135, a potent slow neutron absorber, is continually removed from the MSR salt, but remains within LWR solid fuel.

Power conversion. Two new power conversion systems are candidates to use MSR high temperatures. The triple-reheat closed Brayton cycle turbine mass is smaller than a comparable steam turbine by a factor of about four. Open cycle Brayton turbine engineering has been developed to a high art for the aircraft industry. A $24 million GE90 turbine delivers 83 MW – a power production cost of only $0.29/W. After perfecting, the Brayton closed cycle helium turbine costs should similarly drop relative to massive steam turbines used in LWRs. The newer supercritical CO2 turbine is even smaller, and it requires more engineering to perfect. At lower technology risk, Siemens today sells 46% efficient 620°C steam turbines and will be testing a 56% efficient 700°C steam turbine after 2015.

Waste disposal. All reactors produce the same heat from decaying fission products that cool in a few hundred years. Longer-term heat production from decaying transuranics is a cost driver for geological storage sites such as Yucca Mountain or the Waste Isolation Pilot Plant. MSR and LMFBR can produce less than 1% of the long-lived radioactive transuranic isotopes produced by LWRs.

Factory production. Commercialization of technology leads to lower costs as the number of units increase. Experience benefits arise from work specialization, new processes, product standardization, new technologies, and product redesign. Business economists observe that doubling the number of units produced reduces cost by a percentage termed the learning ratio, seen in the early aircraft industry to be 20%. Today Moore’s law in the computer industry illustrates a learning ratio of 50%. In The Economic Future of Nuclear Power University of Chicago economists more conservatively estimate the learning ratio is 10% for nuclear power reactors.

Fig 6. Illustrating a 10% learning ratio, the cost of the 1024th MSR would be about 35% the cost of the first commercial MSR.

Fig 6. Illustrating a 10% learning ratio, the cost of the 1024th MSR would be about 35% the cost of the first commercial MSR.

Aircraft example. Boeing made 477 airplanes in 2011 costing up to $330 million each. Airplane manufacturing has many of the same critical issues as manufacturing nuclear reactors: life safety, reliability, strength of materials, corrosion, regulatory compliance, documentation, design control, supply chain management, and cost.

Fig 7. Boeing, capable of manufacturing $200 million units daily, is a model for MSR production.

Fig 7. Boeing, capable of manufacturing $200 million units daily, is a model for MSR production.

Reactors of 100 MWe size costing $200 million can similarly be factory produced. Manufacturing more, smaller reactors traverses the learning curve more rapidly.

Manufacturing technology. Electronic documentation control, integrated with manufacturing, saves costs and increases accuracy. New manufacturing techniques are enabled with CAM (computer aided manufacturing), automatically converting designs to manufacturing instructions for machine tools and industrial robots.

Modular expansion. A MSR with slow neutrons is practical for 100 MW(e) and smaller power plant modules, lowering absolute capital requirements, allowing incremental capital investments for modular additions. Fast neutron leakage requires larger size LMFBRs.

Fissile inventory. Starting up a traditional 1000 MWe LMFBR requires about 10 tonnes of U-235, costing $0.50/W. Startup fissile U-235 for TerraPower’s LMFBR is about 5 tonnes/500 MWe, or $0.50/W. LWR initial fissile fuel load is about a quarter of this, costing about 13 cents/W. B&W’s new mPower small modular reactor (SMR) needs 1 tonne/180 MWe costing $0.28/W to power it for 4 years. NuScale’s natural circulation, passively cooled SMR may require 3-4 times this for a 2-year fuel cycle. David LeBlanc of Terrestrial Energy estimates fissile material such as U-235 to start up a 100 MWe DMSR will be in the range of 100-350 kg, depending on design. At $50/g this costs roughly 5-17 cents/W.

Low fuel costs.The DMSR is fueled 86% by thorium, 11% U-238 and 3% by U-235. A DMSR requires addition of fissile U-235 because its thorium-uranium conversion ratio is only about 0.8. DMSRs will likely use 19.75% low enriched uranium (LEU) such as is used in research reactors. Consuming $50/g U-235 at the rate of 200 kg per GWe-year will entail costs of only 0.11 cents/kWh. At $300,000 per ton, thorium costs of 0.0004 cents/kWh are negligible. Including fuel fabrication, the cost for LWR fuel is typically estimated to be 0.7 cents/kWh. B&W mPower and NuScale SMR fuel costs will be higher. The U-238 fuel expense for a LMFBR is negligible, but the one-time expense for fissile U-235 fuel loading costs $0.50/watt, which amortizes to about 0.5 cent/kWh cost over the fuel life. The GE Prism LMFBR, designed to burn excess weapons-grade plutonium and operating with a conversion ration of 0.8 (same as DMSR), would require fissile supplemental fuel adding about 0.11 cents/kWh.

Low uranium enrichment demand. Today’s expanding worldwide fleet of LWRs increases demand for uranium and for the enrichment services to convert it from 0.7% to 4% U-235. Using DMSRs, existing global uranium production and enrichment services could supply the entire current world’s 2500 GWe electric power demand. DMSRs reduce demand for uranium enrichment services, removing an excuse for nations such as Iran to build centrifuge enrichment plants that can also make weapons material.

Inexhaustible fuel supply. Uranium is plentiful; even extracting uranium from seawater at $1000/kg (10 times today’s price) adds less than a half cent/kWh to DMSR power costs. [The LMFBR also has low uranium demand, because it breeds U-238 to fissile plutonium.] Uranium is essentially inexhaustible for both DMSR and LMFBR. World thorium reserves in costal monazite sands alone are 12 million tonnes – a thousand year supply.

Fuel fabrication. Unlike LWRs, there are no costs for producing high quality zirconium tube fuel rods to contain UO2 pellets and their fission products for centuries. The IFR requires uranium metal alloy fuel pins and molten sodium for thermal bonding within steel cladding tubes. Pebble-bed reactors using TRISO particle fuel incur costs for triple-coating millions of UO2 particles designed to retain fission products within the three redundant layers. The DMSR fuel supply form is gaseous UF6 or solid UF4 crystals, which are already intermediate products in the production of solid UO2 used in LWRs.

Fuel reprocessing. The IFR requires electro-processing and recasting of fuel pins about every three years. TWR fuel remains in the reactor 15-20 years before recladding because the fuel pins are internally reshuffled. After 15-30 years of operation, DMSR salt may be discarded or, if economic, reprocessed with the possibilities of recovering uranium, recovering fissile transuranics, or removing fission product fluorides. These processes have not been demonstrated at scale. Separation of thorium from fission product fluorides dissolved in the salt may be impractical because thorium is chemically similar.

Control systems. The number of people required to operate today’s LWRs is higher than for other forms of power production -- more than 1000 employees per GW of power output, adding about 1 cent/kWh to electricity costs. This is another cost innovation opportunity. Information systems and control systems technologies have improved immensely since LWRs were designed in the 1970s. Safety critical software techniques enable low-labor-cost operation of aircraft, helicopters, and rapid transit. The scope of MSR safety critical software is limited, because physical properties over-ride a control system failure -- a fuse-plug melts and empties the fuel salt to a noncritical-configuration dump tank. Reducing direct operator control of reactors can also avoid mistakes, such as the series of operator errors that led to the Chernobyl disaster. Security labor costs should be proportional to the possible threats, which are much lower with a non-pressurized DMSR able to withstand total loss of power. Even US ICBMs in missile silos were protected with remote electronic surveillance rather than guards.

Current research. Cost reductions are presaged by current engineering research. Compact, thin-plate heat exchangers may reduce fluid inventories, size, and cost. Possible new materials include silicon-impregnated carbon fiber with chemical vapor infiltrated carbon surfaces, and higher temperature nickel alloys. Operating at 950°C can increase thermal/electrical conversion efficiency above 50%. Such high temperatures can improve efficiency for water dissociation to create hydrogen, to lower manufacturing costs of synthetic fuels such as methanol or dimethyl ether that can substitute for gasoline or diesel oil.

Transmission costs. One cost associated with multi-GW power plants is for transmission lines to transport power hundreds of miles on low-loss high-voltage direct-current (HVDC) lines. Fewer  transmission lines are required if 100 MWe power sources such as SMRs or DMSRs are near cities and manufacturing centers. Costs for HVDC lines are roughly $1 million per mile, so the cost for energy transmission over 1,000 miles is roughly 1 cent/kWh.

LWR safety. Achieving safety with multiple defense-in-depth systems adds costs. LWRs require multiple control-rods, boron-injection shutdown systems, a large containment building to contain depressurized radioactive steam, emergency electric power systems, spent fuel cooling systems, and multiple sets of emergency cooling pumps, valves, pipes, and water sources.

LMFBR safety. In a LMFBR the fuel pins must always be immersed in liquid sodium kept in a pool vessel, contained within a redundant secondary vessel, because loss of the sodium would increase the reaction rate and lead to run-away overheating. A LMFBR also requires special safety systems because its liquid sodium metal coolant burns on contact with air, water, or concrete. About 20 LMFBRs built worldwide experienced fires. Russia’s BN-600 reactor is the only remaining commercial power LMFBR. Since inception in 1980 it has experienced 14 sodium fires.

Engineered safety systems. The designers of LWRs and LMFBRs are aware of such risks and develop safety systems to counter them. However, the very existence of such engineered safety systems is a vulnerability in itself, as demonstrated in the Browns Ferry cooling system cable fire.

MSR safety. In contrast, MSR’s intrinsic safety derived from physical properties keeps costs low.  A molten salt reactor can’t melt down because the fuel is already molten -- its normal operating state. The salts are solid at room temperature, so if a reactor vessel, pump, or pipe ruptured the salts would spill out and solidify. There is no explosion potential because the pressure in the reactor is atmospheric. If the temperature rises stability is intrinsic due to salt expansion. In an emergency an actively cooled solid plug of salt in a drain pipe melts and the fuel flows to a critically safe, passively cooled dump tank. The transparent fuel salt can enable remote observation of the reactor internals.

Passive decay heat removal. When any reactor stops the fission products continue to decay and the heat must be removed. The MSR salt transfers heat better than water in a LWR. Material properties permit a stopped MSR to rise to higher temperatures than a LWR or LMFBR. In a LWR the zirconium fuel rod cladding releases hydrogen from steam over 500°C; sodium in a LMFBR will boil over 881°C. The MSR can withstand higher temperatures because Hastelloy melts at 1300°C and the molten salt boils at over 1400°C. The high temperature increases thermal radiation cooling and permits natural air convection. It helps that small 100 MWe scale reactors have double the relative surface cooling area compared to large 800 MWe reactors.

Fission product safety. In an MSR the radioactive noble gas fission products Kr-85 and especially Xe-135 are continuously removed to avoid parasitic neutron capture. Xe-135 decays to mildly radioactive Cs-135 retained at the reactor. Short-lived (8 day half-life) iodine-131, whose release caused thyroid cancer in over 1000 children near Chernobyl, remains locked in the ionic salt as iodide. Cs-137 and Sr-90, the 30-year half-life fission products of most environmental concern, are bound as fluorides in the salt and can not enter the environment as gasses, liquids, or dispersed particulates. After about 15-30 years of accumulation of fission products in the molten salt and at the reactor site, the fission products must be removed and sequestered, possibly in the solidified fuel salt.

MSR energy cost innovation summary

Capital costs of $2/watt and electricity at 3 cents/kWh are supported by simple fluid fuel handling, high thermal capacity heat exchange fluids, smaller components, low pressure core, high efficiency power conversion, simple intrinsic safety, factory production, the learning curve, and technologies already under development. These cost drivers are encouraging, however achieving the cost objective requires the design, development, and testing of a fully operational pilot plant MSR, followed by industry participation to achieve factory production of such power plants.

Electric energy cost model

The cost model in the book THORIUM: energy cheaper than coal analyzes realistic, unsubsidized costs of electric power from new power plants. Capital cost examples for power plants were taken from published sources. Capital cost recovery assumes an 8% cost of capital and 40 year asset lifetime generating power at 90% of capacity, except 30% for wind and 20% for solar. Separately, LMFBR capital costs have been historically presented as some multiple of the cost per watt for a LWR. Examples are 2.1 (France’s SuperPhenix), 1.4 (SuperPhenix-II), 1.5 (Russia’s BN-600), and 1.35 (BN-800). TerraPower’s objective is to be cost-competitive with LWRs ($6.4/watt). This table uses Deloitte’s 2010 estimate of €4.286 billion for a 600 MWe LMFBR ($9.5/watt).


Electricity costs from alternative sources, cents/kWh










Capital cost recovery











































Many people wish that renewable energy sources could solve the world energy/climate crises, but the costs are 3-4 times too high and obscured by a patchwork of subsidies, taxes, emissions credits, renewable portfolio standards, and feed-in tariffs. Economic fundamentals shows that renewable wind, solar, and biomass clearly can not replace coal. [Relatively benign and affordable hydropower is not shown because suitable sites are limited.]

Natural gas is displacing some coal generation in the US, but burning it also emits substantial CO2. The US AP1000 is too costly to replace coal; the cost for the Vogtle plant in Georgia is $16 billion for 2.2 GW of electric power generation capacity ($6.4/watt). [However China claims a $2/watt AP1000 cost.] Given acceptable safety and environmental attributes, cost is the dominating factor for technology choice in most nations. Cost-innovative DMSRs are the most economical source for energy. 

PART 3: Global Impacts of Low-Cost Clean Energy, to be published Wednesday, August 28, will illustrate how low cost energy can benefit civilization: checking global warming, ending energy poverty and stabilizing population, synthesizing fuels to replace petroleum, and sparking world GDP growth.