Replacing All Fossil Fuels With Nuclear-Enabled Hydrogen, Cellulosic Hydrocarbon Biofuels, and Dispatchable Electricity

Fossil fuels are remarkable [1]: (1) low cost, (2) easy to store and (3) easy to transport at low costs. As a consequence, 84% of the world's energy demand in 2020 was met by natural gas, oil, and coal. They enabled billions of people to move from poverty to the middle class. Figure 1 shows the U.S. fossil-fuel system and why it is difficult to transition to a low-carbon energy system. Wells and mines produce fossil fuels. Low-cost long-distance transport of fossil fuels is done via pipelines, trains, barges, and ships. The United States stores about 6 weeks of energy to meet variable hourly to seasonal energy demands in coal piles, oil tanks, and natural gas underground storage that are relatively close to the final customer [1]. That is more than 3 million gigawatt hours of energy storage. Fossil fuels are converted into electricity, heat, and transportation services for the final customer using technologies such as gas turbines, furnaces, and internal combustion engines.

In the electricity sector, almost all energy storage is in coal piles, oil tanks, and underground natural gas storage facilities near the power plants. It is converted into electricity at a variable rate to match electricity demand. Very little electricity is stored as electricity because of cost. For example, a million gigawatt hours of battery storage would cost several hundred trillion dollars—many times the size of the U.S. economy. Very little electricity is transmitted long distances. The primary purpose of the transmission system is to assure reliability so if plant A is shut down, the load is taken up by plants B, C, and D. A transmission line typically transports only a few gigawatts of electricity whereas a large long-distance pipeline has energy transport rates of tens of gigawatts. Furthermore, pipelines operate at high capacity factors because of the availability of low-cost natural gas and oil storage near the customer to meet variable demand. Electric transmission lines have much lower capacity factors and much higher costs per unit of energy transported.

None of the proposed replacements for fossil fuels come close to duplicating their remarkable capabilities and the remarkable capabilities of the total system. Each alternative to fossil fuels has its own limitations, some of which are severe. Table 1 shows these low-carbon energy options and their technical characteristics.

The requirement in a low-carbon world is to replace fossil fuels at minimum costs to enable everyone to have a reasonable standard of living. Because fossil fuels are inexpensive, easy to store and transport costs are low, the world's energy system is relatively homogeneous. There are a few exceptions such as locations with large quantities of hydroelectricity. In a low-carbon world, nuclear reactors are the only option that can be built almost anywhere to deliver heat and electricity. Nuclear reactors have been built with air-cooling systems and thus are not dependent upon water. The energy content of a kilogram of uranium is about a million times larger than fossil fuels; thus, there is no economic barrier to transporting nuclear fuel anywhere. That same factor of a million implies that small amounts of uranium provide massive amounts of energy and sufficient uranium resources to meet energy needs. Uranium resources increase as the ore grade goes down. Because uranium costs are a small fraction of nuclear power costs, increased uranium costs substantially increase available uranium resources with small impacts on the total cost of nuclear energy [2]. Other non-fossil energy sources (wind, solar, hydro, etc.) are local energy sources with costs per unit of energy output that vary widely by location.

Energy choices will be primarily based on economics. In the last decade, energy has varied between 7 and 13% of the global economy. Only very wealthy nations can make energy choices not based on economics without major impacts on standards of living. Faced with the challenge of replacing fossil fuels, countries such as France, the United Kingdom, and Sweden have announced major nuclear programs. The same is beginning to occur in the United States. What will change is we are going from a relatively uniform global fossil-fuel energy system to a heterogeneous system where different parts of the world have very different fractions of their energy provided by nuclear, wind, solar, and hydro based on local resources and economics.

A series of studies and workshops were undertaken that examined how to use nuclear energy to replace all fossil fuels. The goal is similar to the Weinberg studies [3] in the 1960s that asked the question if nuclear energy could support a world of 10 billion people with a middle-class standard of living. There are three constraints. First, nuclear reactors are a capital-intensive technology with low operating costs; thus, reactors should operate near base-load to minimize costs. Second, the system design should allow wide variations in different energy inputs with location to minimize total costs to society. The costs of other primary energy sources such as wind, solar, and hydro are strongly dependent upon location. Last, the technologies can be implemented in a reasonable period of time. We describe three systems that use base-load nuclear reactors to replace the three primary forms of energy delivered to the customer: electricity, gaseous fuels, and liquid hydrocarbons while integrating the different low-carbon energy sources (Table 1) into an efficient low-cost, low-carbon energy system.

• Dispatchable electricity to the grid. The gas turbine using stored natural gas is replaced with base-load nuclear reactors with large-scale heat storage to provide dispatchable electricity to the grid.

• Hydrogen replacing natural gas. The primary role of natural gas today is to provide heat. Natural gas is replaced by hydrogen produced in nuclear hydrogen giga-factories enabled by the transport capacities of large-scale pipelines. Separate from replacing natural gas as an energy source, massive quantities of hydrogen are required for the production of fertilizer, steel, and hydrocarbon biofuels. That demand could be 20% of global energy production.

• Cellulosic liquid hydrocarbon fuels replacing all crude oil. There is sufficient cellulosic biomass feed stock if massive nuclear heat and hydrogen inputs at the biorefinery that (1) more than doubles the liquid fuel yield per ton of biomass feed stock and (2) enables the use of lower-grade feed stocks. The cellulosic biomass feed stock becomes primarily a carbon source to produce hydrocarbon liquids, not an energy source. The conversion processes enable recycle of plant nutrients in the biomass back to the soil with some refractory carbon to improve long-term soil productivity. This approach removes and sequesters carbon dioxide from the atmosphere resulting in negative carbon emissions.

In the United States, electricity provides slightly less than 20% of the energy to the final customer (residential, commercial, industrial, and transportation) [4]. Its production uses a larger fraction of the primary energy because of the conversion losses from primary energy sources to electricity. Nuclear reactors are used for base-load electricity production. That is a consequence of the existence of fossil fuels. Nuclear plants have high capital costs and low operating costs while fossil plants have low capital costs and high operating costs. The different economics of nuclear and fossil resulted in base-load nuclear plants with dispatchable electricity from fossil-fuel plants, primarily gas turbines burning natural gas.

The addition of non-dispatchable wind and solar provides electricity to the grid based on weather patterns independent of the demand for electricity. The effects of wind and solar have been seen in places such as California where wholesale electricity prices collapse at times of high solar and wind output and increase at other times [5]. Figure 2 shows California electricity prices on a spring day in 2012 and 2017. The 2012 prices were set by fossil-fuel power plants. The large variations in electricity prices in 2017 were a consequence of the large-scale addition of solar. Simultaneously, there is increasing curtailment of wind and solar to the grid at times when excess production exceeds electricity demand.

The high cost of electricity storage makes it uneconomic to store excess electricity from wind and solar at the scale it is produced. Revenue collapse limits the scale of wind and solar deployment. Today gas turbines burning natural gas provide dispatchable electricity to assure meeting the variable demand for electricity. Natural gas turbines have low capital costs and high fuel costs. Wind and solar act as fuel savings technologies that reduce natural gas consumption [6]. Total system costs go down and then up with increased wind and solar if attempt to create a decarbonized system with just wind, solar, and storage. That is because of the high electricity storage costs for more than a few hours. System studies [6,7] show with deep decarbonization one needs a dispatchable electricity source. The studies also show the potential for large cost reductions if low-cost long-duration storage can be developed. The question is what replaces the gas turbine with stored natural gas?

Multiple developers of advanced nuclear reactors are planning to add thermal energy storage to enable base-load reactors to provide variable electricity to the grid (Fig. 3). Recent workshops examined this option [8,9]. For higher-temperature reactors, the near-term thermal energy storage material is sodium–potassium nitrate salt—the solar salt used in existing Concentrated Solar Power (CSP) plants for heat storage. The reactor is not directly coupled to the power block. Instead, the reactor takes cold salt, heats the salt, and sends it to the hot-salt tank. The power cycle takes hot salt and produces steam that produces electricity. For lower-temperature light water reactors (LWRs), there are parallel technologies that use oil to transfer heat [9] to storage. In some of these LWR systems [10], heat storage is incorporated into the steam cycle rather than separating the reactor from the power cycle. Recent studies [11] have evaluated the value of thermal energy storage for nuclear reactor systems in different parts of the United States.

The reactor is sized to match the average electricity demand. The power block with steam boilers and turbines is sized to match peak electricity demand and may be several times the power output of the reactor. It is built to non-nuclear standards because it is not directly coupled to the reactor. The power cycle is designed to minimize capital costs because its capacity factor may be 30% while the reactor capacity factor is 90%. These steam cycles are designed to operate at variable loads with lower capacity factors that are similar to those used in CSP plants that have similar functional requirements.

The power block can change power levels much more rapidly than a commercial nuclear reactor or a conventional fossil-fuel steam plant. Heat input into the power cycle is controlled by the hot-salt pump speed. The rate of power change is not controlled by the reactor. The hot salt-to-water/steam heat exchanger is much smaller than a fossil-fuel boiler with variable air flow through the boiler. Heat transfer from salt to the steam cycle is better than flue gas in a boiler to the steam cycle. These features enable faster power changes than fossil-fuel steam plants.

If there is very-low-price electricity, the power plant buys electricity to heat more nitrate salt. This is further discussed in Sec. 2.3 as a mechanism that provides a floor on electricity prices for times with excess wind and solar production. If the peak demand occurs for an extended period of time and heat storage becomes depleted, a low-cost furnace burning natural gas or, in the future, a furnace burning hydrogen or biofuels can provide the additional necessary heat. Nuclear energy with thermal energy storage enables the larger-scale use of wind and solar because of (1) assured production of electricity at times of low wind and solar conditions and (2) low-cost heat storage that raises the minimum price of electricity at times of high wind or solar output increasing wind and solar revenue.

The near-term thermal energy storage material for higher-temperature reactors is nitrate salt stored in large hot and cold storage tanks. This heat-storage system is used in CSP plants at the gigawatt-hour heat-storage scale for two reasons. First, on partly cloudy days, the power output may go up and down a dozen times as clouds pass over the solar farm. Storage provides constant heat to the power block. Second, more recently, salt storage enable solar plants to produce electricity after the sun sets. The heat-storage capital costs are $20–30/kWh of heat—an order of magnitude less than the equivalent storage costs of battery or pumped hydro storage. General Electric and TerraPower [12] have announced that their Natrium reactor to be built in Wyoming within this decade will use this commercial two-tank nitrate storage system to enable variable electricity to the grid with base-load reactor operation. The base-load electricity output is 345 MW(e) with a peak power of 500 MW(e). Several other reactor developers are proposing the same two-tank nitrate heat-storage system.

The traditional commercial nitrate heat-storage systems are economic for daily storage but not for multiday storage including weekend/weekday storage. Electricity demand decreases on weekends implying excess electricity production on weekends. Wind output has a multiday cycle. Advanced heat-storage systems are being developed that may lower costs to a few dollars per kWh of heat [8,9,13–18] to enable long-term heat storage for a week or more.

The second-generation thermal energy storage systems [13,17] are single-tank thermocline systems with (1) a hot fluid on top of a cold fluid and (2) a low-cost fill material. The fluid may be a liquid salt, a heat-transfer oil, or other fluid. Candidate fill materials include crushed rock, special concretes, and iron. Costs may be reduced by a factor of two relative to two-tank systems. In a two-tank system, half the tank space is always empty whereas, with a thermocline system, all tank space is full. Multiple variants of this technology are at the pilot plant stage of development.

Figure 4 shows an advanced thermal energy storage system for sodium reactors where the fill material is cast iron hexagons in stainless steel wrappers and the fluid is sodium from the secondary reactor loop. It is a multi-tank system that is expandable to any capacity where the hot–cold liquid-sodium interface moves between tanks in series. The hot–cold temperature difference in the storage system can be much larger than the inlet and exit temperatures from the reactor by mixing hot exit sodium from the reactor with cold sodium from heat storage to obtain the desired sodium inlet temperature for the reactor. The closely packed hexagons minimize the sodium inventory in the storage tanks to minimize inventory of expensive sodium and minimize fire risks.

The incentive for this design is that it enables a sodium-cooled reactor to efficiently couple to advanced combined-cycle gas turbines [18] with the capability to rapidly vary power output with base-load reactor operation [see next section]. In the 1950s the U.S. initiated the Aircraft Nuclear Propulsion program with heat transferred from the reactor to the aircraft jet engines using sodium. Tradeoff studies indicated sodium to be the preferred heat-transfer fluid if transferring heat from a reactor to compressed air inside a combined-cycle gas turbine producing electricity. In this example, as in many other cases, the power cycle choice partly drives the choice of heat-storage system.

The third-generation heat-storage systems use hot oil [14,19] or nitrate salts [13,15] for heat transfer and crushed rock for heat storage. Fluid choices are limited by the need to be chemically compatible with the rock. The capital cost goal for these Crushed Rock Ultra-large Stored Heat (CRUSH) systems is $5/kWh of heat to enable economic heat storage for a week or more—including the option of seasonal storage [16]. Crushed rock is the lowest-cost heat-storage material. Nitrate salt is used to transfer high-temperature heat from higher-temperature nuclear reactors or CSP systems. Hot oil is used if transferring heat at temperatures up to 400 °C from light water reactors (LWRs) or lower-temperature CSP systems. In the existing commercial two-tank nitrate-salt heat-storage system, 40–50% of the cost is in the nitrate salt and 40–50% of the cost is in the insulated tanks [13,15]. To reduce costs, the CRUSH system (Fig. 5) replaces (1) the tanks with a low-cost insulated building and (2) crushed rock as the heat-storage media. Nitrate salt or heat-transfer oil is only used for heat transfer. The technology is coupled efficiently to all heat-generating technologies (fission, fusion, and CSP) and is the lowest-cost heat-storage technology thus far identified.

Heat is added to the crushed rock by spraying the hot heat-transfer fluid (oil or nitrate salt) over the crushed rock sweeping from left to right (Fig. 5, upper right and lower left). The fluid trickles through the crushed rock while heating the rock to the catch pans below the crushed rock. The hot fluid spreads out as it trickles downward. The cold heat-transfer fluid is collected by the bottom collection pan to be reheated. If the heat-transfer fluid is not fully cooled by the time it reaches the collection pan (Fig. 5, lower left), the warm fluid is pumped onto the top of the next section of crushed rock to preheat cold crushed rock. To recover the heat, cold fluid is poured on top of hot rock and is heated as it trickles down to the drain pans (Fig. 5). Sections of rock are heated with hot oil or nitrate salt. There is a rock-heating wave followed by a second wave to recover heat. When either wave reaches the end of the structure, it starts over at the other end. A wave of hot fluid heats the crushed rock from left to right.

Development of CRUSH is underway in the United States. [15], Germany [13], and China [19]. The largest-scale nitrate salt tests were recently conducted in Germany raising the Technological Readiness Level to between 4 and 5. The work in China is on an oil-based system.

Multiple advanced power cycles [18,20–25] have been proposed to couple with reactors or heat storage for variable electricity production using high-efficiency combined-cycle gas turbines—the primary technology used today to produce variable electricity in the United States. The goals are (1) higher efficiency than steam cycles, (2) significantly lower capital costs than traditional steam cycles for systems that may operate only 30% of the time, and (3) faster response. The economic incentives for such power cycles only exist when there are many days of the year with large variations in electricity prices (Fig. 2). Such markets are only now beginning to appear in the United States and Europe.

In each of these systems, heat is inputted at two temperatures. Reactor and stored heat provide lower-temperature heat (in some cases to 600 °C) with a second source of high-temperature heat—a thermodynamic topping cycle that improves efficiency and power output. The second source is either a combustible fuel (hydrogen, biofuels, etc.) or high-temperature stored heat in a firebrick that can match existing peak gas turbine temperatures. Most of these systems can operate on only a reactor or stored heat at lower power outputs. In some cases, the topping cycle can increase the power output by a factor of five or more.

The near-term option is the Liquid Salt Combined Cycle^™ (LSCC) being developed by Pintail Inc. as shown in Fig. 6. It is a patented [21–25] hybrid power plant combining conventional off-the-shelf generation technologies with thermal energy storage. The more-advanced options require significant equipment development. LSCC integrates two-tank commercial molten salt thermal energy storage with combustion and steam turbines in a system that enables faster startup and lower capital costs. The near-term deployment strategy is a lower-cost longer-duration replacement for utility-scale batteries where the input is low-cost electricity. The longer-term option is that a nuclear reactor provides the heat to the storage system. This system provides an example of the types of engineering tradeoffs to significantly reduce the power cycle costs in markets where the power cycle has a low capacity factor versus the base-load reactor.

The LSCC is a modified combined-cycle power plant. In the steam cycle, the feed water is heated to the boiling point by exhaust heat from the gas turbine. The evaporative heating is provided by stored heat. The peak pressure in the Rankine cycle is lower than in traditional steam plants. This reduces the steam saturation temperature, which in turn allows more heat to be extracted from the molten salt. LSCC uses a eutectic salt (53% potassium nitrate, 7% sodium nitrate, and 40% sodium nitrite), commonly known the by tradename Hitec^® Costal Chemical Heat Transfer Salt with a low-freezing point (142 °C). The hot salt is heated to about 425 °C using (1) a nuclear reactor or (2) low-price electricity when available. The peak salt temperature is lower than 565 °C which is the typical limit for solar salt. Under these conditions, the salt is stable, non-corrosive, non-toxic, non-flammable, and does not degrade with use. The evaporative heater is small because of the very good heat transfer between a liquid molten salt and boiling water. The steam is then heated by the gas turbine exhaust and sent to a simple steam turbine (no reheat or other complex steam cycle features).

The high steam temperature in a combined-cycle plant is primarily needed to manage the moisture content at the steam turbine exhaust, rather than to improve efficiency. In traditional high-pressure steam cycles, the exhaust moisture content is controlled by reheating the steam after partial expansion. In the LSCC system, high-temperature gas turbine exhaust is reserved for superheating steam while using stored energy for water evaporation at medium pressure. A simple low-cost steam cycle can be used and there is no need for high salt temperatures. The lower salt operating temperature is compatible with carbon steel tanks and piping to reduce capital costs and makes it feasible to use low-freezing point salts, instead of solar salt, to reduce operating costs. The power output of an LSCC system is about twice the power output of the combustion turbine, compared to about 1.5× for a conventional triple-pressure plus reheat combined-cycle plant. Half the heat input is from the salt heat-storage system. LSCC also uses lower-pressure steam, which reduces the thermal stresses on exhaust heat recovery components during startup.

The gas turbine can use conventional natural gas, renewable natural gas, biofuels (discussed later), hydrogen (discussed later), or electricity-to-stored heat—but half as much per unit of electricity as a conventional combined-cycle plant. Renewable natural gas [26] is made from biomass with no net addition of carbon dioxide to the atmosphere. Plants remove carbon dioxide to produce biomass, the biomass is converted into renewable natural gas, and burning the natural gas returns the carbon dioxide to the atmosphere. It is a commercial product. The alternative to burning fuel in the gas turbine is to convert low-price electricity, if available, into high-temperature stored heat in firebrick recuperators that can match the combustion temperatures of natural gas [27]². This option uses electrically-conductive firebrick as resistance heaters capable of storing heat at 1800 °C which is above fuel combustion temperatures. This variant uses a new electric-heater technology that is in the process of being commercialized³ whereas the LSCC described earlier uses only existing well-proven commercial technologies.

With all these options, the well-defined long-term storage competition includes (1) traditional combined-cycle gas turbines burning low-carbon hydrogen or biofuels that can be stored on an hourly to seasonal basis (below) and (2) pumped hydro. The pumped hydro options include traditional pumped hydro, seawater pumped hydro where the lower reservoir is the ocean (one plant exists in Japan), and pumped hydro with underground reservoirs.

Heat storage is more efficient than battery or pumped hydro storage for any energy source that generates heat—fission, fusion, or CSP. The U.S. Energy Information Administration [28] reported that the average round-trip electricity-to-electricity efficiency of utility battery systems is 82% and for hydro-pumped storage is 79%. The low round-trip efficiency is because of the multiple conversion steps in the storage process—such as in a battery from alternating current to direct current to chemical energy and back again. With a high-temperature nuclear reactor, the salt is the intermediate loop that would exist in any case. Heat normally goes from the reactor to the power cycle. Adding heat storage in the intermediate loop does not involve energy conversion steps with associated inefficiencies. There are small heat losses from the reactor through storage to the power block where heat losses depend upon storage times and system size. For equivalent duty cycles with battery and pumped hydro storage, thermal energy storage losses will be a few percent. However, because of their much lower capital costs, thermal storage systems will be preferred when need longer storage times with added energy losses. The one compensating factor is that larger systems with smaller surface-to-volume ratios have much smaller relative heat losses.

The efficiency of buying electricity, converting it to stored heat and converting the heat back to electricity is much lower. Converting electricity to heat is nearly 100% efficient but converting heat to electricity using these salt systems is nearly 40%. If one builds a nuclear reactor with heat storage, the incremental capital cost of the electric resistance heaters is very low—everything else in this thermal battery (storage tanks, power conversion block, connection to the transmission grid, etc.) already exists. The system has “two” storage systems that use most of the same equipment: (1) the efficient reactor heat-to-heat storage to power block and (2) the less efficient but very-low-incremental-cost electricity to heat storage to power block system that buys low-price electricity from the grid when available. This electricity would be from wind and solar at times of excess electricity production. The economics favor dumping very-low-price electricity into heat storage that has a low-incremental capital cost—and setting a minimum price for electricity.

The U.S. Energy Information Agency [28,29] has estimated the levelized cost of electricity produced by solar ($31.30/MWh), on-shore wind ($31.45/MWh), and offshore wind ($115.04/MWh) in good locations. These costs do not include the costs of energy storage or backup gas turbines. However, wind and solar can provide electricity less than half the time because the sun sets, and there are days with no wind; thus, most electricity in such systems is provided by the gas turbines that had to be built to compensate for these times of low production. The cost and performance limits of existing electricity storage systems are large. The levelized cost of storage batteries [28,29] is $121.86/MWh—far higher than the cost of making electricity. Furthermore, batteries are only good for energy storage on the time scale of two to 6 hours and thus are unable to provide electricity for multiple days of cloudy weather or a week of low wind conditions. Batteries can reduce the number of hours per year the gas turbines operate but do not eliminate the need for gas turbines or equivalent generating capacity.

Large-scale wind and solar impose large system costs [30–32] onto the grid in terms of resource adequacy (assured generating capacity), energy adequacy, and reliability. The economic viability of large-scale wind and solar is tightly coupled to finding an economic replacement for the gas turbine that provides assured electricity for times of low solar and wind input. Nuclear reactors with large-scale heat storage and assured peaking capacity may be the enabling technology for a low-carbon system that uses large-scale wind and solar resources.

Natural gas provides 24% of the global energy demand and 32% of the U.S. energy demand. In the United States, this demand includes about 4000 industrial users in the U.S. with heat demands above one megawatt [33]. The expectation is that a significant fraction of this natural gas will be replaced with hydrogen—another gaseous fuel. There are historical reasons for this. Starting in the early 1800s, major cities were lit with town gas, a mixture of hydrogen and carbon monoxide produced from coal. In the 1950s, natural gas replaced town gas in the United States. This conversion occurred in the 1970s in the United Kingdom. Converting to hydrogen or hydrogen with renewable natural gas would be a second hydrogen fuel transition—from town gas (hydrogen and carbon monoxide) to natural gas (hydrogen coupled to carbon as methane) to hydrogen (with or without methane).

There are two hydrogen futures: one that is assured and one that is more uncertain. The first future is hydrogen as a chemical where there are three mega markets. Today, hydrogen from natural gas is used to make ammonia and other fertilizers that feed about half the world. Ammonia consists of one nitrogen atom obtained from the air and three hydrogen atoms. Second, inexpensive hydrogen is the key technology to rapidly replace all crude oil with cellulosic drop-in liquid hydrocarbon fuels (gasoline, diesel, and jet fuel) and chemical feedstocks (Sec. 4). Third, affordable hydrogen enables rapid decarbonization of the steel industry by direct reduction of iron ore [34]. Iron and steel production is about a quarter of the global industrial energy demand and is responsible for 7% of global carbon dioxide production. The world converts 2.5 billion tons of iron ore per year into steel using carbon from coal to grab the oxygen in iron oxide to produce iron and carbon dioxide. Hydrogen replaces the use of coal. Iron ore (iron oxides) reacts with hydrogen to produce metallic iron and water (two hydrogens and one oxygen). These three chemical hydrogen markets could exceed 20% of global energy demand.

The second future is using hydrogen as a fuel. Because these applications do not depend upon the specific chemical properties of hydrogen, there are alternative heat sources. Furthermore, it is more expensive and complicated to safely deliver hydrogen to millions of smaller customers in small quantities compared to a few thousand industrial customers with trained staff.

In the United States, most hydrogen is made from natural gas where there is the option to sequester the carbon dioxide byproduct. The process chemistry results in most of the carbon dioxide leaving the process as relatively pure carbon dioxide and thus enables relatively low-cost carbon capture and sequestration. In locations with low-cost natural gas and good sequestration sites for carbon dioxide, the estimated cost of hydrogen is between $1.50 and $2.00 per kilogram. The U.S. government recently provided incentives for producing hydrogen from natural gas by sequestering the carbon dioxide. This has resulted in multiple announcements to build such plants including the announcement by Exxon Mobil to build the first billion cubic feet per day (28 million cubic meters per day) hydrogen plant with carbon dioxide sequestration in Texas.

Multiple advanced processes [35] are being developed to convert natural gas into hydrogen and carbon where the solid carbon is buried—potentially a lower-cost conversion option that would not require carbon dioxide sequestration. That would enable long-distance shipment of natural gas to regions without good sequestration sites with local conversion to hydrogen and local burial of carbon. The energy inputs to operate these processes may be from burning some of the hydrogen that is produced or external energy sources such as nuclear energy.

Hydrogen is different from electricity as an energy source. First, hydrogen is inexpensive to store on a large scale. Underground hydrogen storage facilities have operated for decades [36] in depleted oil and gas fields, saline aquifers, and salt caverns. These are the same geologies used in the 400 underground natural gas storage facilities in the United States. The large experience base with underground natural gas storage [37] provides the basis for new underground hydrogen storage facilities or conversion of natural gas to hydrogen storage facilities. In the United States, we store up to 15% of a year's supply of natural gas to meet peak winter demand. There is no need to match hydrogen production on a second-to-second or even month-to-month basis with demand since storage provides assured supply.

A single hydrogen pipeline can ship tens of gigawatts versus electricity transmission lines that are limited to a few gigawatts. This creates the option to build very large hydrogen production plants. However, transcontinental shipment of hydrogen is more expensive than natural gas because the volumetric energy density of hydrogen is several times smaller than natural gas. That fact drives toward a system with regional hydrogen production. Today, in Texas, we have such hydrogen storage facilities and pipelines that connect refineries, chemical plants, and hydrogen production facilities.

Nuclear energy is potentially competitive in this market when low-cost natural gas is not available. Hydrogen can be made by electrolysis of water or steam. Water electrolysis has been a commercial technology for over a century. High-temperature electrolysis (HTE) is the most efficient technology [38–40] where nuclear plants can provide electricity and steam—an intrinsic advantage of nuclear energy to produce hydrogen versus electrolysis of liquid water using electricity-generating technologies such as wind and solar photovoltaic. HTE is in the early stages of commercial deployment. However, hydrogen plants using either technology, from the power supplies to electrolysis cells to compressors and associated pipelines, are capital intensive. The hydrogen plant capacity factor must be high as shown in Fig. 7 to produce cheap hydrogen [41]. At high capacity factors, electricity is the primary cost. At low capacity factors, the equipment capital cost is the primary cost because of not producing much hydrogen in an expensive plant. The higher efficiency of HTE and the requirement for high capacity factors provide an economic competitive advantage to coupling nuclear reactors to hydrogen production plants compared to wind or solar with their lower capacity factors. Nuclear plants have capacity factors of about 90% versus wind nearly 41% and solar nearly 25% in the continental United States.

There are several nuclear hydrogen production options. The first option is to have the nuclear reactor produce hydrogen and peak electricity (Fig. 8). The hydrogen cost versus capacity curve enables peak electricity production during the 5–10% of the year with the highest electricity prices. This has a limited impact on hydrogen production costs because the hydrogen cost versus capacity curve is relatively flat between 80 and 90% capacity. The economic penalty incurred by lower hydrogen plant capacity factors is relatively small if electricity is diverted to the grid for a limited number of hours per year. This feature can help meet the occasional peak summer or winter electricity loads and significantly increase plant revenue—replacing those gas turbines that run for a limited number of hours per year [42,43].

The U.S. Department of Energy [44] with utilities has launched a demonstration program at four commercial nuclear power sites for hydrogen production. Three sites will use conventional electrolysis. The Prairie Island Nuclear Power plant in Minnesota will demonstrate hydrogen production using high-temperature electrolysis where the reactor provides steam and electricity.

Because of the large energy transport capacity of pipelines, we have the option to build very large nuclear hydrogen production complexes [41] on the same energy scale as oil refineries. This possibility creates a new nuclear plant production model (Fig. 9). First, build a modular nuclear reactor fabrication plant that produces or assembles reactors to be sited next to the factory with the hydrogen plant. Second, with shipyard cranes that can lift several thousand tons, move reactors from the factory to nuclear plant site by crane. Third, if the reactor needs refurbishing, transport it back into the factory.

This approach changes nuclear energy into a factory operation where the site hydrogen production capacity grows over 10 years and thereafter the factory produces replacement reactors. Factory fabrication [45] can dramatically lower the cost of nuclear power plants—in addition to improved economics of operation of multiple reactors at a single site and economics of scale for the hydrogen production plant. Separately, the historical experience in building nuclear power plants in multiple countries is that the cost decreases rapidly with increasing numbers of identical reactors built on a single site [46,47]. That reflects (1) the one-time cost to build the infrastructure to support nuclear plant construction, (2) identical designs that minimize engineering, and (3) site learning.

The U.S. Department of Energy's hydrogen program goal is to reduce the cost of hydrogen to a dollar per kilogram within the next decade—partly by lowering the capital costs and improving the efficiencies of electrolyzers. That would significantly lower the costs of hydrogen from nuclear, wind, and solar—lowering the cost curve shown in Fig. 7—and reduce the curvature of that cost curve with capacity factor. The levelized cost [48] of hydrogen is estimated at $0.91/kg for a giga factory producing two million tons per year of hydrogen—an option that if successful would be significantly less expensive than hydrogen from low-cost natural gas.

Crude oil products provide almost half the energy to the final customer in the United States and a third of global energy. Recent studies [49–51] have examined the potential of a nuclear-assisted cellulosic biorefinery system to produce sufficient gasoline, diesel, jet fuel, and other hydrocarbons for chemical plant feedstocks to replace all crude oil—domestic and global.

Today, most biofuels are produced from starches (corn), vegetable oils (soybeans, etc.), and sugar (sugar cane). However, the resource base of these feedstocks is limited with potential conflicts with food production. Lignocellulosic biomass (wood, straw, core stover, grasses, kelp, etc.) is the most abundant form of biomass on earth and has long been used as an energy source. It is also a source of renewable carbon that can be converted into hydrocarbon fuels. Because plants capture carbon dioxide from the air, the burning of biomass does not increase atmospheric carbon dioxide levels. About 100 billion tons of biomass, roughly containing 50% carbon on a mass basis, are created by photosynthesis each year [52]. Given a demand for additional biomass, much more cellulosic biomass could be readily and sustainably produced [50].

Using external heat and hydrogen inputs enables replacing all crude oil with liquid hydrocarbon biofuels using available cellulosic biomass resources without major increases in food or fiber prices. First, external heat and hydrogen more than double the quantities of hydrocarbon fuels per ton of biomass feedstock [50,53]. For a given amount of biofuels produced, this reduces the land requirements for biomass production by more than a factor of two—recognizing that in many cases (corn and corn stover, etc.), one simultaneously produces food and cellulosic biomass. Second, external heat and hydrogen enable the use of biomass feedstocks that are poor energy, food, and fiber sources but excellent sources of carbon for the production of biofuels. Third, the economics allows higher prices to farmers for biomass, thereby greatly increasing the supply of available biomass and benefiting rural areas.

With traditional biofuels (Eq. (1)), the cost of biofuels is driven by the cost of biomass. With external heat and hydrogen (Eq. (2)), hydrogen is the largest cost component. One can afford to pay more for biomass without large increases in the cost of the final liquid hydrocarbon. Higher prices dramatically increase the availability of cellulosic biomass feedstocks thanks to the remarkable productivity of American agriculture. For example, corn yields have gone from 20 to 180 bushels per acre in the past century [54] with corresponding increases in cellulosic corn stover. We have never asked what the capability of agriculture would be if there was a large market for cellulosic biomass. Studies [49,50] have identified multiple routes that would each increase cellulosic biomass feedstocks by hundreds of millions of tons per year. For example, extensive double-cropping [55] is not practiced today in the Midwest United States because of the lack of a market for the double crop. Double-cropping typically involves producing a high-value grain or oilseed crop such as corn or soybeans and a lower-value grass or forage crop. Double-cropping was widely practiced prior to farm mechanization in order to provide forage for horses and oxen, that is, to provide biofuels for transportation and work.

Hydrogen is used in the production of liquid fossil hydrocarbon fuels to remove sulfur and produce gasoline, diesel, and jet fuel. The prices of different feedstocks reflect the differences in the amount of hydrogen and refining required to produce the final product. Fossil feedstock prices to produce hydrocarbon liquid fuels go in decreasing order from light crude oil to heavy crude oil, to tar sands, and then to coal. Coal has the lowest price, but requires the most hydrogen to convert to liquid fuels. The same trends are true when making hydrocarbon liquids from other carbon sources. Today's hydrocarbon biofuels are primarily made from feedstocks such as plant oils that minimize hydrogen inputs. These feedstocks are in limited supply. Hydrocarbons made from abundant cellulosic feedstocks require more hydrogen. The other carbon feedstock is carbon dioxide (CO₂) from non-fossil sources (ethanol plants, ocean, atmosphere, etc.). However, it takes six hydrogens to convert each carbon atom into a hydrocarbon fuel (four to convert oxygen to water and two to convert the carbon to a hydrocarbon) versus about two to convert cellulosic feedstocks into hydrocarbon fuels. The combination of available feedstock and lower hydrogen demand is why liquid hydrocarbons from cellulosic feedstocks will become the primary source of future hydrocarbons.

Economics requires large biorefineries (equivalent or larger than 100,000 barrel/day oil refineries). Large refineries are replacing small refineries. The United States has 126 refineries, and the combined capacity of the bottom 32 refineries matches that of the largest single refinery with a capacity that is over 600,000 barrels per day [50]. When listing refineries by capacity, the refinery in the middle of the list has a capacity of 100,000 barrels of oil per day but most of the refining capacity is in much larger refineries. This is a consequence of three factors. First, there are large economies of scale. Second, large refineries have the capability to mix different crude oils to produce a feed that matches their operational requirements resulting in lower feedstock costs. Last, large refineries have the capability to produce different products over the year to match product demand and maximize revenue. Product requirements have become more restrictive in the last several decades. In the United Sates, this has included limiting gasoline vapor pressure in summer to minimize air pollution and reducing the chemical toxicity of gasoline. Tighter requirements require more sophisticated refineries. The same technical and market factors will drive biorefineries.

Refinery processes for the conversion of cellulosic biomass into hydrocarbon fuels include Fischer-Tropsch (FT), flash pyrolysis, and direct hydrogenation [56]. There are large economics of scale associated with these processes. The Sasol FT coal-to-liquids plant in South Africa produces 150,000 barrels per day of liquid fuels. The newer Shell FT natural gas-to-liquids plant in Qatar produces 260,000 barrels per day of liquid fuels. There are other processes to convert biomass into hydrocarbons [57]. These processes produce the hydrocarbon feedstock that is the replacement for crude oil; thus, the expectation is that existing refineries will convert incrementally over time to biorefineries by blending progressively more bio-oil with crude oil into their feedstocks.

The downstream refinery processes [58] convert bio-crude oils into the different hydrocarbon liquids (gasoline, diesel, jet fuel, and chemical feedstocks) used by society. Refineries consume about 10% of the oil going into the refinery to operate the refinery [59]. For the United States, that is about a million GWh of heat per year or slightly under half the thermal heat output of the U.S. nuclear fleet. Almost all of this energy is used in the form of heat. A biorefinery will have larger heat demands, partly because of water associated with the biomass feedstocks. Large refineries are the largest single users of heat with heat demands measured in gigawatts. Constant heat input is required because the startup times for refineries are measured in days, and for many processes, load following is not an option. Those requirements match the capabilities of nuclear reactors. Dow Chemical recently announced an agreement with X-Energy [60] to build four high-temperature reactors to provide process heat for one of the chemical complexes in Texas. The same reactor systems are required to provide heat to oil or biorefineries. The only other heat option is fossil fuels with sequestration of the byproduct carbon dioxide [61].

The conversion of bio-crude oils into liquid hydrocarbon fuels requires massive quantities of hydrogen. The low-cost option today is the conversion of natural gas into hydrogen with underground sequestration of the carbon dioxide anywhere in the world with low-cost natural gas and good sequestration sites for carbon dioxide. Recent studies [62] show that the greenhouse footprint of properly designed and operated systems can approach that of green hydrogen from wind or solar but not as low as hydrogen from nuclear systems. That is because of the embedded greenhouse gas releases in the building of wind and solar systems. In the longer term, hydrogen will likely be from nuclear sources as described earlier. It is unclear if the cost of electrolysis using wind and solar will decrease sufficiently to be economically competitive (Fig. 6).

The daily biomass feedstock requirements for a large biorefinery are much larger than the unprocessed biomass that can be economically shipped directly from farms and forests to the biorefinery. Large biorefineries require local depots to convert locally produced biomass into dense storable intermediate commodities that can be economically shipped long distances to the large biorefineries. A simplified system schematic is shown in Fig. 10. The nuclear reactors that produce heat for the biorefinery must be co-located because heat can only be transported short distances. Hydrogen can be produced locally or shipped in via pipeline.

The economic shipping distances of unprocessed biomass are 50–80 kilometers. This limits traditional biofuel plants to feed input rates of about 3000 tons per day. A biorefinery with a throughput of 250,000 barrels per day oil refinery will need cellulosic biomass feed input rates approaching 60,000 tons per day. Transportation logistics and economics require depots to convert raw biomass into economically shippable commodities to the refineries [63]. There are four primary depot options where the choice partly depends upon the type of biomass.

• Pelletization. The biomass is densified and shipped by rail or barge. This process is used today to prepare and ship some types of wood long distances to power plants as fuel. The densified solid biomass would be sent to biorefineries for conversion into liquid fuels.

• Anaerobic Digestion. The biomass is fed to an anaerobic digester that produces a methane/carbon-dioxide gas mixture that is then shipped via pipeline to the refinery—plus a digestate that is returned to the soil. This process is used commercially in some countries to produce renewable natural gas [26]. Converting natural gas to liquid hydrocarbon fuels is a large-scale commercial process. The digestate recycles nutrients to the soil and improves long-term soil fertility and water/nutrient holding capacity by adding refractory carbon. Anaerobic digestion is commercial for some types but not all types of biomass.

• Fast pyrolysis. The biomass is heated quickly to produce pyrolysis oil and biochar [64]. After stabilization, the pyrolysis oil may be shipped by pipeline or train to the refinery. The biochar can be recycled into the soil to recycle nutrients or sent to the refinery for conversion into liquid fuels. This process is at the early stage of commercialization with plants in northern Europe.

• Direct hydrogenation. Direct hydrogenation [56] of crude oil is used in refineries to remove sulfur and upgrade crude oil into final products. If pipeline hydrogen becomes widely available, it becomes a local depot option. Sulfur and oxygen are column 6 elements in the periodic table with similar chemical behavior; thus, much of the existing technology to de-sulfur crude oil is applicable to removing oxygen from biomass as water. Nutrients with some carbon can be recycled into the soil. There is limited experience in direct hydrogenation of cellulosic biomass.

The depot system is a variant of the existing farm-to-market system used for many agricultural products. For example, corn is shipped from farms to elevators that may process the corn to remove weed seeds and dry the corn to enable storing the corn for long periods of time. The depot system for cellulosic biomass may also conduct other processing options. For example, alfalfa is an important forage crop for animals where the leaves contain most of the animal protein. This creates economic incentives to separate the leaves for animal food while using the stems for biofuels.

Depots create a system that can improve long-term soil properties and productivity, thus improving rural economies by capturing more value-added locally. The growing of food extracts trace nutrients from the soil for human consumption—a form of mining. In biofuels production, only the hydrogen and carbon are of value. We do not want potassium, phosphorus, and other nutrients in the biofuels. The depot system recycles nutrients as well as refractory carbon from digesters and pyrolysis units back to the soil. Experience with anaerobic digesters shows significant improvement in soil productivity over time with the return of the digestate to the soil.

The last component of the system is the sequestration of carbon from the atmosphere. Negative carbon emissions occur in two locations where the quantities of negative emissions will partly depend upon the market price for sequestered carbon.

• Refinery. The refinery can produce carbon dioxide for sequestration when excess low-priced biomass is available or during times of low liquid-fuel prices. This option provides variable negative carbon emissions while stabilizing the price of liquid fuels caused by variable production of biomass or changing markets for liquid fuels over time. Carbon capture and sequestration from a conventional fossil power plant is expensive because of the cost of removing dilute carbon dioxide from the stack gas. In contrast, carbon dioxide sequestration is relatively inexpensive [61]. The process chemistry of a biorefinery provides lower-cost ways to produce a relatively pure carbon dioxide stream with low sequestration costs.

• Depots. The depots can provide variable quantities of refractory carbon for carbon sequestration in the soil [65] to improve long-term soil productivity. Recycling carbon into soils provides long-term carbon sequestration. The quantities of carbon in the soil today are several times that in all of the world's biomass. Recycle of carbon and nutrients to the soil is not a new technology. A historical example of carbon and nutrient recycling is ancient settlements in the Amazon basin where local tribes created carbon residues that were added to the soils to improve soil productivity and where the carbon remains today [66]. This may become the primary method for negative carbon emissions from the atmosphere because (1) the demand for hydrocarbon biofuels implies massive carbon flows through the system so that 5–15% carbon to soils for long-term sequestration implies massive carbon sequestration and (2) economic incentives to recycle soil nutrients and carbon to improve agricultural productivity.

The implementation strategy uses existing oil refineries with modifications to front-end processing units. Oil refineries would progressively transition from crude oil to biomass feedstocks. This is beginning to happen. The Neste refinery in Finland is the first large refinery to announce plans to fully convert to biomass feedstocks by the mid-2030s. The transition is similar to transitions that have occurred at many refineries as they moved over time from processing light crude oils to heavy crude oils. Preliminary economic assessments indicate the cost of liquid hydrocarbons from cellulosic biomass will be equivalent to crude oil at $70–$80 per barrel of crude oil assuming a hydrogen cost of $2/kg. Recent studies indicate hydrogen from natural gas with carbon capture and sequestration (CCS) is less than this number [67]. The largest cost component is the cost of hydrogen, followed by the cost of biomass and refinery operations with the massive heat demand where nuclear heat is the economic choice.

Commercialization of any alternative to fossil liquid fuels will require incentives. Oil prices are highly volatile with average yearly prices in the last decade varying from $37.22 to $102.58 per barrel. The combination of (1) existing global capabilities to produce hydrocarbon products from crude oil and (2) the high volatility of oil prices makes investments in any replacement technology risky and discourages the deployment of any alternative system [68]. For liquid cellulosic hydrocarbon fuels, one option from the electric sector is “Contracts for the Difference.” In its simplest form, the government guarantees a minimum price for cellulosic biofuels to any biofuel producer for X years. If the sales price of biofuels when produced is below the guaranteed fuel price, the government makes up the difference. If the sales price of biofuels when produced is above the guaranteed price, no payment is made. There are alternative options such as clean fuel standards that have been implemented in California, Oregon, Washington, and British Columbia. The government defines allowable emissions of fossil carbon dioxide from the transport sector. Sellers of fossil liquid hydrocarbons must buy credits from low-carbon fuel producers so the allowable net fossil-fuel carbon dioxide emission goals for the year are met. The expectation is that cellulosic biofuels will be competitive—partly because (1) cellulosic biomass is the lowest-cost biomass at the required scale and (2) the economic advantages of large commercial refineries.

The primary challenge in the transition away from fossil fuels is economically replacing the three energy services provided by fossil fuels: (1) energy, (2) energy storage, and (3) energy transportability. The challenge cannot be solved by a single technology. The different energy production technologies are not interchangeable—each has different strengths and weaknesses and produces different products (electricity, heat, and biomass). The characteristics of nuclear energy are that it can provide a constant heat output, can be built almost everywhere, and is not dependent upon the availability of local energy resources (wind, solar, and water flow). The important economic characteristic is that it has high capital cost and low operating cost that favors operating nuclear plants near maximum base-load capacity.

Modern societies primarily use energy in three forms: electricity, gaseous fuels, and liquid hydrocarbons. While these energy delivery forms compete with each other to meet customer needs, it is extremely expensive to fully replace any one of these energy forms—such as proposals to electrify everything. Under any scenario, there is a massive chemical hydrogen demand to produce fertilizer, steel, and chemicals. Hydrogen maybe 20% of the energy mix partly because the massive chemical hydrogen demand creates the hydrogen infrastructure that opens up gaseous fuel markets to partly replace natural gas. U.S. oil consumption is about 18 million barrels per day and can be reduced; but, studies [51] show the cost of alternatives increases rapidly if U.S. consumption goes much below 10 million barrels per day—about 20% of the energy to the final customer. Today, electricity is somewhat less than 20% of energy to the final U.S. customer.

This leads to the requirement for three nuclear energy systems that produce: (1) dispatchable electricity, (2) hydrogen (gaseous fuel and chemical reagent), and (3) liquid hydrocarbons (fuels and chemical feed stocks). The production of cellulosic hydrocarbon liquids enables large-scale removal of carbon dioxide from the atmosphere by sequestering carbon to the soil while recycling plant nutrients. Future developments that drive economics will determine the relative size of these three energy delivery sectors—but the broad outline of such a nuclear future is likely to remain unchanged. Local conditions determine the relative contributions of nuclear, wind, solar, and hydro to meet local energy demands.

Charles Forsberg gratefully acknowledges support from the Shanghai Institute of Applied Physics (SINAP) of the Chinese Academy of Sciences and the INL National Universities Consortium (NUC) Program under DOE Idaho Operations Office Contract DE-AC07-05ID14517. Professor Dale gratefully acknowledges support from Michigan State University AgBioResearch and the National Institute for Food and Agriculture of the US Department of Agriculture. A short early version of this paper was given at an A + B conference [69].

Charles Forsberg: conceptualization, original draft preparation, reviewing, and editing; Bruce E. Dale: conceptualization, original draft preparation, reviewing, and editing; Eric Ingersoll: conceptualization, original draft preparation, reviewing, and editing. All authors have read and agreed to the published version of the manuscript.

There are no conflicts of interest. This article does not include research in which human participants were involved. Informed consent is not applicable.

The authors attest that all data for this study are included in the paper.