Magnesium Hydride Slurry: A Better Answer to Hydrogen Storage

As the world moves toward a sustainable future with increasing use of intermittent renewable energy sources, energy storage will be required to replace the convenient storage capability of fossil fuels. One synthetic storage medium that appears attractive is hydrogen [1]. Renewable energy can be stored in hydrogen by breaking water into hydrogen and oxygen. There is much work going on today to produce hydrogen from renewable energy sources [2–4]. When this energy is needed, the hydrogen can be burned to produce heat or it can be used in a fuel cell to produce electricity. Hydrogen can be stored as a compressed gas at high pressure, as a cryogenic liquid, or chemically bound to another element or molecule. Other synthetic energy storage mediums are batteries, hydroelectric reservoirs, compressed air, or the synthetic production of hydrocarbons.

When hydrogen is bound to certain metals, it is called a metal hydride. Metal hydrides can be stored as powders or as mixtures of powder and liquids. Magnesium hydride (MgH₂) is a metal hydride that allows hydrogen to be cycled in and out of storage. By mixing powdered magnesium hydride with mineral oil, a slurry can be formed. Hydrogen stored and transported in magnesium hydride slurry offers a cost effective and safe alternative to other energy storage alternatives. This hydrogen storage medium is pumpable, inexpensive, and it can be stored at ambient conditions for months.

To provide reliable energy to our society, the renewable energy sources will need to be part of a larger system. Much work is underway to evaluate potential systems to make renewable energy part of a dispatchable system [5,6]. To show the value of MgH₂ slurry, Safe Hydrogen has performed an analysis of a hydrogen storage system that can be used with an intermittent energy wind farm to produce baseload or dispatchable power. The study calculated the performance of a 150 MW baseload wind power system and found that it can produce an IRR of 10% with an electricity price of $0.088/kWh. This compares well with an equivalent cost of a new coal fired power plant of $0.11/kWh and a cost for a new nuclear power plant of $0.115/kWh. The capital cost of an equivalent system using batteries would be 50–100 times more.

Hydrogen storage can be achieved by storing as a compressed gas, by storing as a cryogenic liquid, or by binding the hydrogen chemically with another element or compound. Chemical storage techniques include metal hydrides, liquid organic hydrogen carriers, and ammonia among others. The most common hydrogen storage techniques employed in industry today are compressed hydrogen and liquid hydrogen. Chemical binding processes for the storage of hydrogen have been under study for many decades but have not yet achieved substantial commercial success.

Compressed hydrogen is delivered today in bottles and tube trucks. It requires high-pressure vessels to reduce its volume. Pressure vessels made from steel are relatively heavy and store hydrogen at pressures up to about 200 bar. Pressure vessels made from composite materials are being used in hydrogen-fueled vehicles with pressures of 350 bar and 700 bar.

Compression energy makes up about 5% of the energy stored when compressing to 350 bar and about 8% of the energy stored when compressing to 700 bar [7]. One of the disadvantages of compressed hydrogen storage is that when hydrogen is moved from tank to tank, the supply tank must be sufficiently high in pressure to allow the tanks to move toward equilibrium or more compression energy must be expended to move the hydrogen from the lower pressure tank to the higher pressure tank.

Hydrogen delivered as compressed hydrogen by tube trucks has a cost of $8–$10/kg of hydrogen. Hydrogen delivered in small quantities in bottles can cost over $100/kg.

Liquefied hydrogen is delivered today in cryogenic trucks at 20 K (−253 °C) and 2–4 bar.

Liquefaction uses 35% of the energy stored to liquefy the hydrogen for 4400 kg/day plants and 31% of the energy stored for 50,000 kg/day plants [7]. New technology and larger plants are anticipated to bring this value down to 18% of the energy stored. In addition, when the liquid hydrogen must be stored for a long time, additional energy will be needed to recondense the hydrogen evaporating within the Dewar.

Liquefied hydrogen offers the least expensive commercial method to transport hydrogen today when pipelines are not available. Hydrogen delivered by Dewar trucks has a cost of $6–$8/kg of hydrogen.

Liquid organic hydrogen storage uses an organic liquid that will absorb and desorb hydrogen depending on the temperature of the liquid. Air Products explored this option for the Department of Energy several years ago. TIAX and Argonne National Laboratory (ANL) evaluated the Air Products recommended product [8] and concluded that the liquid could supply hydrogen to a vehicle at a cost similar to that of gasoline.

TIAX and ANL concluded that the system would require about 33% of the energy contained in the fuel to release the hydrogen from the liquid.

Metal hydrides have been explored for decades. Some metal hydrides will absorb and release hydrogen at relatively low temperatures. Others require moderate temperatures. Some require small fractions of the stored energy when releasing the hydrogen. Others require large fractions.

In many metal hydride systems, the metal hydride powder is contained within a pressure vessel. The powder in the pressure vessel is supported to prevent packing of the metal hydride during dehydriding and to spread the metal hydride through the vessel so that it can be heated or cooled by heat transfer piping also contained within the vessel. These systems must be designed to act as both storage and reaction vessels.

Magnesium hydride holds about 7.7% of its mass in hydrogen but requires about 30% of its stored energy when releasing hydrogen. One form of magnesium hydride use is as magnesium hydride slurry. Metal hydride slurries have been under development by researchers at Safe Hydrogen since 1998 [9–12].

Magnesium hydride slurry is a suspension of magnesium hydride particles in light mineral oil. We have been working with slurries ranging from 40 wt.% solids to 75 wt.% solids. Slurries with 40–50 wt. % solids cycle well in the reactors that were tested. In these slurries, the particles stay in suspension for months. When settling eventually occurs, the slurry forms a soft pack, which can be easily stirred back into suspension.

The slurry looks like thick paint. The viscosity at a temperature of 30 °C is about 500 cP. The viscosity at operating temperatures is considerably lower. The slurry is described as a non-Newtonian shear thinning material.

Reaction rates are similar to those reported for powder systems. Full hydrides take 1–2 hr. Full dehydrides take 3–4 hr. Figure 1 displays the hydrogen content of our magnesium hydride slurry observed during a recent experiment. In this figure, hydrogen leaves the slurry (dehydrides), enters the slurry (hydrides), leaves slurry (dehydrides), and then enters back into the slurry (hydrides). During this test, the slurry was cycled at about 80% of its theoretical capacity during which high reaction rates were achieved. It is important that good reaction rates occur over a high fraction of the capacity which demonstrates that the slurry system can be economical when used. Reaction rates slow as the slurry approaches the lower (when dehydriding) and higher (when hydriding) limits of its capacity.

The apparatus, for this test, consists of a Parr autoclave and a high-pressure tank that supplies hydrogen to the slurry through a regulator at about 1724 kPa (250 psia). The hydriding reaction temperature is typically in the range of 280–340 °C. When dehydriding, the hydrogen produced is collected in a low pressure tank that gains pressure from about 448 to 552 kPa (65–80 psia). The dehydriding reaction temperature is in the range of 340–370 °C.

Figure 1 is calculated from the amount of gaseous hydrogen in the system (the high-pressure supply tank, the low pressure receiver tank, the reactor, and the piping). As the slurry dehydrides, gaseous hydrogen is supplied to the system. As the slurry hydrides, gaseous hydrogen is removed from the system. From the volume, temperature, and pressure of the bottles, we calculate the amount of hydrogen desorbed during dehydriding and the amount adsorbed during hydriding.

where d[H₂]/dt is the rate of change of moles of H₂, [Mg] is the number of moles of Mg, [H₂] is the number of moles of H₂, [MgH₂] is the number of moles of MgH₂, and K is the reaction rate constant.

The dehydriding follows a similar reaction rate profile for the decomposition of MgH₂ to Mg and H₂ except that a straight section in the middle of the dehydride occurs. This is caused by the heat transfer limited nature of the reaction vessel that we have been using.

In the dehydriding, the oil performs another significant function beyond keeping the particles separated; it keeps the partial pressure of hydrogen low in the reaction vessel around the MgH₂ particles. As hydrogen is produced from the reaction, it forms bubbles in the slurry. Part of the bubble is oil vapor. The bubble rises to the surface and the hydrogen and vapor leave the reaction zone. Oil is condensed in a separate condenser and returned to the slurry. The pressure in the reaction vessel is made up of the partial pressure of oil vapor and the partial pressure of hydrogen. The pressure downstream of the condenser is made up primarily of hydrogen. A low partial pressure of hydrogen in the dehydrider promotes a higher reaction rate at elevated hydrogen pressures than would be observed in powdered magnesium hydride at the same temperature and pressure.

We have observed the condenser to reduce the oil concentration in the delivered hydrogen to parts per million levels.

The slurry characteristics are shown in Table 1 for a storage system with 200 kg of hydrogen. Since 1 kg of hydrogen has about the same energy as a gallon of fuel oil, this system is similar to a 200 gal oil tank. A 200 gal oil tank is about 0.75 m³. In Table 1, HHV refers to the Higher Heating Value of hydrogen.

In addition to the characteristics shown in Table 1, the slurry has been shown to meet the Department of Transportation criterion for a label as a nonhazardous material when transported. This criterion defines hydrogen-producing materials when wet as hazardous if 1 l of hydrogen is produced within an hour per kg of material tested [13]. The slurry does not produce hydrogen at normal temperatures and pressures. If mixed with water, our tests showed a production of hydrogen of less than 10 ml over a week. Both the hydrogen rich slurry and the hydrogen-depleted slurries were tested and shown to fall into the nonhazardous category.

The material is relatively benign. If spilled, the slurry could come into contact with water. In that event, it would slowly react producing hydrogen at a very slow rate because of the low temperature (ambient conditions) and producing a byproduct of magnesium hydroxide, which is otherwise known as “milk of magnesia,” a mild basic chemical sometimes used to reduce the acidity of our stomachs. A spill can be handled in much the same way as an oil spill.

The oil medium allows the slurry to be pumped from container to container. The slurry can use the conventional liquid fuels infrastructure of tanks, pumps, and transportation. Since the liquid fuels infrastructure is highly developed, the slurry can be stored and transported inexpensively.

The mineral oil part of the slurry not only provides a medium for making the powder perform like a liquid; but it also provides advantages as a heat transfer medium. During the dehydriding and hydriding reactions, the mineral oil aids in the transfer of heat. When desorbing hydrogen from the slurry during the dehydride process, heat must be supplied to the magnesium hydride particles to break the bonds holding the hydrogen to the magnesium atoms. As a liquid, the slurry can be pumped past heat transfer surfaces significantly increasing the heat transfer rate from the surface to the slurry. When adsorbing hydrogen, during the hydriding process, the oil is able to absorb the heat of reaction. If the reaction is fast, the oil can boil as a means of carrying heat away from the particles.

In a dry powder system, when hydriding, the heat of reaction must be carried away from the particle by the hydrogen that is fueling the reaction. This does not work well and the particle temperatures frequently rise high enough to cause the powder to sinter. When sintered, the magnesium hydride particles stick together and form a porous mass. In a slurry system, the oil carries heat away from the particles and helps to keep the particles from touching and sintering.

One of the greatest advantages of the slurry approach to storing hydrogen in magnesium hydride is that the function of reacting the magnesium hydride can be separated from the function of storing it. In the powder approach, the magnesium hydride is typically contained within a pressure vessel, which also contains heat transfer tubing and a means of supporting the particles so that they do not pack at the bottom of the container. In the slurry approach, the reaction and storage functions can be separated. A reactor can be built to provide for the power rates required of the application and a storage system can be built to provide for the energy requirements of the system. The storage system can use conventional liquid fuels tanks and can be as large as needed. It does not need to be designed for the pressures and temperatures required of the reactions.

Many safety issues are also addressed by the slurry approach to using magnesium hydride. We have already mentioned the benign nature of the slurry. In addition, because we use a mineral oil that has a low vapor pressure, the slurry is difficult to ignite. It has a vapor pressure lower than that of fuel oil. We have tested this by holding a torch on the slurry. The torch must be held on the slurry for tens of seconds to ignite the slurry. After igniting, we were able to blow the flame out. It was only after the flame had burned/boiled enough of the oil away that the magnesium began to burn.

If powdered metal hydrides are exposed to oxygen and moisture in air, they generally become deactivated. Magnesium hydride is not as easy to deactivate as many other metal hydrides because its reaction rate is low at ambient temperatures relative to most other metal hydrides of interest. The oil in the slurry further protects the powder from the oxygen and moisture in the air. So the slurry can be transferred from tank to tank without concern that it will be deactivated. This practice is probably not recommended though because each time the slurry is exposed to the atmosphere, there will be a small amount of oxygen and moisture caught on the surface of the slurry. This small amount will react with the hydride particles when the slurry moves through the reactors and a small amount of magnesium will be converted to magnesium oxide. Over time, the hydrogen storage capacity of the slurry will be decreased.

One of the greatest safety features is that there is very little hydrogen gas in a system when using magnesium hydride slurry. When hydrogen is needed, it is removed from the slurry and consumed by the application. When hydrogen must be stored, the slurry readily adsorbs it.

The slurry approach also has many economic benefits. The reduction in concern of deactivation means that magnesium hydride can be handled with less expensive procedures than most metal hydride powders.

Magnesium hydride slurry can use the existing liquid fuel infrastructure, including tanks, pumps, tank trucks, tank railroad cars, and barges. Since our infrastructure for liquid fuels is mature and inexpensive, the storage and transportation of magnesium hydride slurry can be inexpensive. We have estimated that magnesium hydride slurry can be shipped 500 mi for a cost of $0.68/kg H₂ by tank truck. This calculation includes the cost of the vehicles and the labor to drive, load, and unload the slurry.

Since magnesium hydride slurry can be pumped, it can be moved from one tank to another with a very low energy cost. This results in inexpensive handling.

Magnesium hydride is currently expensive to purchase. This is because it has a very small market and is produced as a powder when needed. We have recently demonstrated that we can make magnesium hydride slurry from magnesium slurry. We used a mixture of 92% magnesium slurry and 8% magnesium hydride slurry. After cycling four times, we had a magnesium hydride slurry that could store and release over 80% of the hydrogen expected based on the magnesium content. This is an important result because it shows that magnesium hydride slurry can be produced readily and with significantly improved safety. Since reduction of hazards significantly reduces the costs of production, we are confident that magnesium hydride slurry can be produced at costs similar to the costs of the magnesium and oil used in the slurry.

Additional cost savings can be achieved by using new technologies for the direct reduction of magnesium oxide to magnesium such as the solid oxide membrane (SOM) process [14,15]. This process offers significant reductions in the energy used to produce magnesium metal. The process offers a direct reduction of magnesium oxide to magnesium without the secondary cycles required of competing technologies. Since magnesium metal requires less energy to reform than aluminum when reformed in a direct process, the cost of magnesium should be less than the cost of aluminum when processed at similar scales.

Magnesium hydride slurry is not as energy dense as the automotive industry currently would like for fuel cell cars. However, it is inexpensive to store and handle so it may be better suited for other applications. We have identified five applications that look particularly attractive.

Grid scale electricity storage will be of particular interest as we expand our use of intermittent renewable energy sources such as wind and solar. We will need increasing amounts of storage to be able to provide reliable power as intermittent renewable energy sources provide greater proportions of the system energy. One way to achieve this end is to use excess electrical energy to split hydrogen from water by electrolysis. This hydrogen can be stored in magnesium hydride slurry, which in turn can be stored in a tank farm. When electricity is needed, it can be retrieved from the magnesium hydride slurry and burned in a gas turbine. We have explored the costs of this opportunity and they look attractive. A further description of this option is provided in the Grid Scale Electricity Storage Section.

To recover the costs associated with providing power during peak power periods of the day, utilities apply demand charges as part of the rate structure to certain classes of customers. Demand charges are based on the peak power used by a customer during the high demand period of the day. For sharp peaks, this charge can be much greater than the energy charge.

To minimize the cost of electricity, many commercial and industrial customers are employing measures to smooth out their load over the day, or to shift large loads to reduce the peak demand. One way to do this is to make hydrogen by electrolysis during off-peak periods and to use this hydrogen to make electricity during peak periods. Such a system can also be used to supply backup power when the utility service is unavailable.

There are many applications that require backup power. The cell phone companies employ backup power to keep the cell tower system operational during grid wide blackouts. Internet server farms use backup power to keep their servers in operation when the grid power is unavailable. All levels of government and most private companies are protected by backup power to maintain critical services when the utility system fails.

Hydrogen stored in magnesium hydride slurry can compete in this application. Backup systems can either be designed to be rechargeable on site or slurry can be restocked as the hydrogen is consumed.

Since hydrogen can be used to produce electricity with minimal pollution, hydrogen fuelled power plants can be built in pollution sensitive regions. Hydrogen produced by renewable energy and transported into cities as magnesium hydride slurry provides an attractive means to produce additional power in sensitive regions when needed. Fuel cells can be used to produce this electricity if air or noise pollution is of particular concern.

Because magnesium hydride slurry can be transported inexpensively using the existing liquid fuels infrastructure and because it can be stored easily and inexpensively, it offers a cost effective means to transport hydrogen from where it is generated to where it is needed.

Currently, the U.S. packaged hydrogen market is small. The liquid hydrogen infrastructure that was built to support the National Aeronautics and Space Administration continues to supply the packaged hydrogen market admirably. But as the packaged hydrogen market expands, with the expansion of hydrogen use in fuel cell cars or with the expansion of hydrogen use in stationary fuel cells, this liquid hydrogen infrastructure will need to be expanded. Magnesium hydride slurry can perform this function at a fraction of the cost of liquid hydrogen when compared with new liquid hydrogen installations.

As noted above, grid scale electricity is a potential application for magnesium hydride slurry. We selected this application for a detailed study because it is an opportunity where economies of scale can aid in the cost competitiveness of hydrogen stored in MgH₂ slurry. The slurry can be stored in large storage vessels that are cheaper per unit storage as they get larger. In addition, the reactors will benefit from economies of scale because they will be cheaper, per unit of hydrogen reacted, as they get larger.

Another important factor is that utilities are signing agreements with some new wind farms limiting the amount of electricity that they will buy under contract. All the rest of the electricity will be bought by the utility at spot market prices. Since more wind energy occurs during the night when spot prices are low than during the day when spot prices are high, wind farms are not able to get as high a price as they might if they could supply dispatchable power or baseload power.

Hydrogen storage can enable electricity storage from intermittent sources. For this process, hydrogen will be produced by electrolysis of water when energy is available and electricity will be produced from the hydrogen when electrical energy is needed and not available from the intermittent systems.

Electrical storage can also be achieved by pumped hydroelectric systems, compressed air systems, and a wide array of batteries. Many of the alternative electrical energy storage systems offer higher round trip efficiencies than hydrogen storage systems with magnesium hydride slurry but most cost more than the magnesium hydride slurry approach or are limited by where they can be applied. Batteries cost considerably more than the magnesium hydride slurry approach. Pumped hydro and compressed air storage using caverns are limited in where they can be applied.

The process we are considering uses efficient industrial scale electrolysis machines to produce hydrogen from energy captured by a wind farm. The hydrogen is stored in magnesium hydride slurry using large-scale hydride reactors. The slurry is stored in a tank farm.

When additional power is needed, hydrogen is removed from the magnesium hydride slurry by dehydride reactors and supplied to hydrogen-fueled gas turbines. The heat required to perform the dehydriding can be provided primarily from the exhaust of the gas turbines. The gas turbine exhaust does not contain quite enough energy at the temperatures needed to supply all the energy required by the dehydride reaction. Burning some additional hydrogen in the gas turbine combustor can supply this additional energy. This heat can be transported by a liquid heat transport system, such as liquid tin, to the dehydrider.

In operation, when wind power is greater than is needed to meet contract power requirements, the contracted wind power is sent to the grid. Power above the contracted value is consumed by the electrolysis machines to make hydrogen. The electrolysis machines can ramp from 10% power to full power in milliseconds so the power output to the grid can be very constant.

When wind power is insufficient to meet the contract power requirements, hydrogen is removed from storage and used to produce the power needed to meet the contract.

Figure 2 displays the concept in graphical form. This diagram follows the discussion presented by Succar [16]. The curve that starts at 500 MW and declines gradually to zero is typical of the annual energy provided by a wind farm. In this example, for about 20% of the year, the wind farm will produce at its rated power level. For 10% of the year, there will be insufficient wind to produce any output power from the wind farm. During the rest of the year, the wind farm will produce energy between its rated power level and zero power. The area above the baseload line, at 150 MW, is the energy that is to be stored. When the wind farm is producing more than the baseload requirement, the baseload energy goes directly to the grid from the wind turbines and the remainder of the wind-generated electricity goes to electrolysis machines to produce hydrogen, which is stored. When the wind farm is producing less energy than the baseload requirement, energy is returned from storage to keep the output at the required power level, in this case 150 MW. Since there are losses in the storage system, more energy is stored than is returned.

Using hourly load data from ISO New England [17] and 10-min wind data from the NREL database for a location near Lubbock, TX [18], we built a computer model to operate as described above. The wind data were modeled rather than measured but it is representative of actual wind data. The data provided power output every 10 min. We modeled both baseload operation and dispatchable operation. In these analyses, the system was allowed to sell additional electricity if the spot price exceeded the contract price and if the system had sufficient additional capacity. This capacity could be defined by the wind supply, the gas turbine capacity, or the grid connection capacity. We limited the grid connection to about half the capacity of the wind farm though the analyses did not benefit from the cost savings of this limitation because we did not include costs for the grid connection. (Limiting the grid connection capacity to something larger than the contract power is an opportunity to reduce the cost of the installation).

In the dispatchable analysis, we assumed that the power plant would be called on to follow the load. We arbitrarily assumed that the system would be selling no power to the grid when the grid was at its annual low. We also arbitrarily assumed that the system would be selling 100% of its capacity at the annual high load peak.

A business partner of ours provided representative costs for the wind farm, the electrolysis machines, and the gas turbines. We have compared these prices to prices described by others and have found them to be competitive for the large number of units planned for purchase. We have estimated the costs of the hydrider and dehydrider using a 2/3 power law scaling approach and the cost of our laboratory unit. The cost of the slurry was based on the cost of the magnesium and oil with a factor for the additional capital and labor costs required to make the slurry.

The IRR has been determined using the calculated capital expense for the various major components. Installation costs and supporting equipment are assumed to be included in the category of other project costs, which is calculated as 15% of the capital cost. Working capital is calculated as 10% of the total capital and other project costs. Operating costs are assumed to be 1% of the capital cost per year for maintenance. In addition, there is a cost of water assumed to come from a water purification plant at a total cost of $0.77/m³. The system has been optimized by varying the number of wind turbines, the number of electrolysis machines, and the electric price required to achieve an IRR of 10%.

Income is from the contract power provided at the contract price for the electrical energy, the additional electrical energy sold at spot market prices, the renewable energy credit, and the sale of oxygen. Oxygen sales are at the current price of bulk oxygen. The total initial investment is the capital expense, the other project costs, and the working capital. An investment tax credit of 30% of the total capital cost has been assumed for the cases displayed. Cases performed using the producer tax credit required a slightly higher price of electricity to achieve 10% IRR. The IRR is calculated from this initial total capital cost minus the investment tax credit and the difference between the income and expenses over a 30-yr lifetime.

Table 2 displays cost and performance characteristics of the two cases studied. The dispatchable system uses fewer wind turbines and less hydrogen storage than the baseload system because less electrical energy is sold in the dispatchable case than in the baseload case.

Table 3 displays the amount of electrical energy sold directly from the wind, the amount sold from the gas turbines, and the amount of hydrogen produced by the electrolyzers. Both systems spill some wind but the amount spilled is small relative to the total amount produced. The baseload system spills less than 0.2%. The dispatchable system spills less than 3%. When additional electrolysis machines were added to the system to use this energy while holding the IRR at 10%, the cost of electricity increased.

Table 4 summarizes the earnings, costs, and the IRR calculated for the two projects. The IRR for the dispatchable system, assuming a contract price of $110/MWh, a 30% investment tax credit, and a renewable energy credit of $3/MWh, is 10%. The electric price for the baseload system, making the same assumption for sales, credits, and IRR is $88/MWh. In both cases, the model assumes that the amount of energy that can be contracted is dependent on the amount of energy stored in the hydrogen storage system and the assumption that the wind might not blow. The storage for both cases is sized to ensure that there will always be enough hydrogen to fuel the gas turbines at full capacity for a 2-day period even when the storage system is largely depleted.

Table 5 displays some figures of merit for this system. The systems store energy at a capital cost of $11–$12/kWh of storage capacity. The storage capacities of the systems are about 75,000 MWh for the dispatchable case and 109,000 MWh for the baseload case. The amount of energy moved through the storage during the year is 232,000–364,000 MWh. So the storage is fully cycled slightly more than three times each year.

Figures 3–10 display several of the operating parameters of the model over a period of 2 days of the year modeled. In this period, the wind farm output, shown in Fig. 3, was producing nearly 500 MW of power then during the night, the wind dropped off and the wind power production dropped to less than 100 MW.

Figure 4 shows the spot price of electricity. The spot prices vary hourly throughout the year. Figure 5 shows the electricity sold directly from the wind. When the spot price was high, additional wind electricity was sold up to the capacity of the grid connection. Figure 6 shows the electricity sent to the electrolyzers. The total power sold, Fig. 7, shows occasional power production above the 150 MW baseload contract. This was allowed when the price of electricity exceeded the baseload power contract. This extra power is represented in Table 4 as earnings spot market. Figure 8 shows the power produced by the hydrogen-fuelled gas turbine. It should be noted that as the wind power declines, the gas turbine power ramps up to meet the deficit in the wind production and the electrolysis machines ramp down because less wind power is available to produce hydrogen. Figure 9 shows the amount of wind spilled when the electrolysis machine capacity is insufficient. Figure 10 shows the hydrogen stored. The stored hydrogen increases when the electrolysis machines are producing hydrogen and decreases when the gas turbines are consuming hydrogen.

This study made several assumptions that may offer opportunities to further reduce the cost of a baseload renewable energy system.

In this analysis, we evaluated the use of a wind farm to supply the energy needed by a baseload wind energy system. The storage system has a capacity to deliver 150 MW for 30 days. This is the storage capacity that is required to provide the baseload capacity through the entire year. Since wind energy production is greater in the winter than in the summer, the storage system had to be sized large enough to carry some of the winter energy into the summer. If the system had used both a wind farm and a solar farm, the input energy could have been more uniform throughout the year, the storage could have been significantly reduced, and the costs of the storage system could also have been reduced. Solar energy farms collect more energy in the summer than in the winter because the days are longer so the total energy input to the system would have been more balanced throughout the year.

For this analysis, we assumed that the cost of magnesium hydride slurry is based on the cost of magnesium and mineral oil with a factor for production costs. We made this assumption, because we have demonstrated that we can hydride magnesium from slurries of magnesium powder and magnesium hydride powder. As the market for magnesium ramps up, new magnesium production technologies such as the SOM process [14,15] promise to reduce the cost of magnesium which will further reduce the cost of magnesium hydride slurry.

Figure 11 displays the sensitivity of the IRR on the contract price. As the price of electricity increases, the income on the contracted electricity increases but the income on extra power sold, when the spot price is above the contract price, declines since there are fewer opportunities to sell at prices above the contract price.

The 150 MW baseload power system described has an overall efficiency about 61%. About 61% of the energy produced by the wind farm is delivered to the grid. The storage system is about 30% efficient. The efficiency can be further improved by using fuel cells.

This model does not include any heat recovery from the hydriding system or the gas turbines. Heat recovery from the hydriding system and gas turbine could provide additional power that could be used to produce hydrogen or to offset some of the hydrogen consumption. We have estimated that heat recovery from the hydrider can produce an additional 8% of electrical energy into the electrolysis system. This could result in a reduction of the number of wind turbines in the system.

The baseload wind power system would provide 100% renewable energy. This compares favorably to the current system of supporting wind farms with natural gas fired gas turbines. Because of the intermittent nature of wind, most wind farms produce less than 45% of the nameplate capacity of the farm [18]. Natural gas fired gas turbines are being called upon to provide electrical energy when the wind production is less than is required. To reach a goal of 80% renewable energy, we will either need to have an excessive amount of overcapacity of wind (resulting in a large fraction of wind energy being spilled and wasted when the load cannot use it) or we will need storage.

As the capacity for renewable energy increases to larger fractions of the total installed electric generation capacity, more conflicts will arise between the intermittent energy sources and the baseload energy providers. At low load periods during the night, when the wind is blowing most heavily and the electric power system has ramped down such that only baseload providers are operating, there will be too much electrical energy available for the load. Either the wind farms or the baseload power plants will need to reduce production. When this has happened in recent years, the wind farms have been asked to feather their turbine blades because of negative impacts to the baseload power providers. Wind capacity in ERCOT during 2009 was curtailed 17% of the time [19]. Bulk energy storage can solve this problem and deliver 100% renewable energy.

The baseload wind farm system, using magnesium hydride slurry for hydrogen storage, compares well with competing electric storage technologies. The advantage of the rechargeable slurry system is that the cost of bulk energy storage is low so that large quantities of energy storage are possible in an economical system. Table 6 displays comparison characteristics of several storage technologies [20]. The systems are compared by build time, efficiency, capital cost (on a $/kWh basis and $/kW basis), and discharge time. The typical comparison criteria for generation equipment are the capital cost comparisons of cost/kWh stored and cost/kW installed. The discharge time helps to differentiate the various technologies. The H₂/slurry storage system offers a very large storage capacity that can allow very long discharge times. This places the H₂/slurry storage system in a class of its own. In addition, it does not suffer from location restrictions. Despite the high cost per kW, the system produces a high return on investment. It should also be noted that the cost per unit power is somewhat misleading as it is based on the 150 MW of gas turbine power rather than the 250 MW of system capacity limited by the grid connection or the 500 MW of wind turbine capacity. It also uses the entire cost of the system including the cost of the wind farm.

Magnesium hydride slurry has been shown to be an attractive method for storing hydrogen. The advantages of hydriding and dehydriding magnesium hydride in an oil-based slurry are significant. The oil protects the metal hydride and supports it during the cycling. It aids in the transfer of heat to and from the particles. It gives the magnesium hydride the characteristics of a liquid and opens up the opportunity to use magnesium hydride slurry with the conventional liquid fuels infrastructure.

Using the slurry approach, we have found that magnesium hydride can be a cost competitive means of storing and transporting hydrogen.

An analysis has been performed to evaluate the potential for using magnesium hydride slurry to store hydrogen produced from a wind farm. The result of this analysis shows that, for an electricity price of $88/MWh, an IRR of 10% can be achieved for a baseload wind/storage system. Further, the study concludes that the system can be configured as a dispatchable power system (one that follows the load throughout each day) for an electricity price of $110/MWh.

The baseload or dispatchable wind power system would provide 100% renewable energy and it can alleviate the conflicts that will become more frequent as the renewable energy system is expanded.

The authors would like to acknowledge the support that the U.S. Department of Energy, under cooperative Agreement No. DE-FC36-04GO14011, provided toward the early development of magnesium hydride slurry.

The authors would also like to thank the DOE/NREL/ALLIANCE for making the NREL Wind Integration Datasets available to the public. They were invaluable in modeling baseload and dispatchable wind energy systems. Funding for this work was provided by Safe Hydrogen, LLC.

Nomenclature

cP =

centiPoise

H₂ =

hydrogen

IRR =

internal rate of return

kPa =

kilopascal

kWh =

kilowatt hour

Mg =

magnesium

MgH₂ =

magnesium hydride

MW =

megawatt

MWh =

megawatt hour

psia =

pounds per square inch absolute

wt. % =

percentage by weight