Home Energy Physics Nuclear Power Electricity Climate Change Lighting Control Contacts Links



By C. Rhodes

Liquid metal electro-chemical cells are of fundamental importance for stationary multi-MWh energy storage systems. This electro-chemical technology provides a practical means for utility scale daily storage and recovery of electrical energy at either urban or rural locations.


Presently the practical electro-chemical technologies for storing energy for later recovery are lithium ion batteries, liquid metal batteries and vanadium flow batteries. Lithium ion batteries are believed to be too expensive for large scale stationary energy storage. Vanadium flow batteries are thought to be too bulky and too environmentally dangerous for widespread use in residential and small commercial applications. Hence, this web page focuses on liquid metal electro-chemical energy storage technology.

The modern sodium-sulfur-nickel chloride energy storage cell is the result of about 50 years of patient research and development in the areas of sodium-sulfur batteries, sodium-chloride (zebra) batteries and density stratified liquid metal batteries. However it is only very recently that these technologies have merged into a commercially viable technology. The purpose of this web page is to briefly describe this merged technology. A leading technical authority on the merged technology is Prof. David Sadoway of MIT.

The initial cost of implementing a liquid metal energy storage system is a strong function of the price of its components. Frequently the dominant cost is the cost of nickel. The other materials are comparatively inexpensive. The nickel and other cell materials are not consumed.

A liquid metal battery consists of a top layer of a low density low melting point metal from the left hand side of the chemical periodic table such as sodium or magnesium, a middle layer of a higher density electrolyte material that will effectively selectively transport metal ions such as a liquid salt and/or liquid sulphur and a bottom layer of higher density metals such as nickel or antimony. Older versions of sodium sulfur cells relied on a thin alumina electrolyte layer. Newer cell versions avoid the requirement for this separator by use of spontaneous material density stratification. However, use of density stratification requires that cells be stationary and have near perfect horizontal mounting.

It is important to operate an energy storage system in a manner that prevents formation of metallic whiskers at the cathlode that can reach through the electrolyte and cause electrical shorts connecting the top and bottom metal layers.

These energy storage systems can easily be fabricated in modules which lend themselves to truck or rail transport. Battery, inverter, transformer and switchgear components are easily interchanged for service. Presently the energy density is about 1 MWh / 10 tonnes. One 18 wheel boom truck load (40 tonnes) can consist of 4 MWh of Na-S-NiCl2 energy storage. There is ongoing development aimed at increasing the energy density.

Liquid metal energy storage cells operate at temperatures above 300 degrees C. Some metal combinations operate at up to 475 degrees C. The thermal insulation required to efficiently thermally isolate liquid metal batteries is bulky. Large batteries are more thermally efficient due to larger mass to surface area ratios. Generally liquid metal batteries are sized to permit truck and container transport.

If the object is to make an electro-chemical cell with the highest possible energy content per unit weight, the periodic table of elements indicates that the cell's active elements must come from the top left and top right corners of the periodic table. On the top left are hydrogen, lithium and sodium. On the top right are fluorine and chlorine. However, use of some of these elements is impractical for various reasons.

Over the temperature range of interest hydrogen is a gas that is difficult to store in quantity, except in a large, heavy and expensive pressure tank.

Lithium is not a common element and it is in high demand for use in automotive propulsion batteries, which makes it too expensive for most stationary energy storage applications.

Fluorine forms strong chemical bonds that are difficult to break for energy storage following energy discharge.

The next best choices for energy storage are sodium (Na) and chlorine (Cl). However, chlorine is a dangerous gas. The practical way to use chlorine is to weakly chemically bind it to another element. This element must be chosen so that the resulting compound does not spontaneously decompose at the highest anticipated operating temperature. Hence the chlorine is usually stored as nickel chloride (NiCl2).

In theory a simple Na-NiCl2 molten salt cell could be made which when it discharges forms NaCl and Ni. However, to enable ion mobiity such a simple cell must be operated above the melting point of NaCl (801 C). That temperature is too high to be practical.

One solution is to use sulfur to form a Na-S-NiCl2 cell. In this cell there are two ongoing reverseable chemical reactions of the form:
2 Na+ + 5 S = Na2S5++
Na2S5++ + NiCl2 = 2 NaCl + S5 + Ni++

Thus during energy discharge sodium in the anode gives up an electron to the external circuit forms sodium ions which move across the electrolyte and form NaCl near the cathlode. Nickel from the NiCl2 accepts electrons from the external circuit and plates out onto the cathlode. The sulfur acts as a pseudoelectrolyte which allows selective transport of sodium ions by forming Na2S5++.

During energy charging sodium from the NaCl moves across the electrolyte to become elemantal sodium at the anode. The Chlorine atoms react with the nickel at the anode to form NiCl2.

The sodium, sulfur and sodium polysulfide all melt at less than 275 degrees C allowing these electro-chemical reactions operate efficiently at about 300 degrees C. However, historically the problem with 300 degree C operation was long term formation of metallic whiskers that would short out the cell after about 2000 charge/discharge cyces. This problem appears to have been conquered by Professor David Sadoway of MIT and his company AMBRI by adoption of a suitable cell geometry, by raising the cell operating temperature and by using certain proprietary materials.

A liquid metal cell typically consists of a outer electrically insulating enclosure, a low density top metal such as sodium which forms the anode, a medium density molten salt or like electrolyte such as sulfur and a high density bottom metal such as nickel or antimony which forms the cathlode.

Air is excluded. Small amounts of other substances and an excess of nickel is used at the cathlode to improve its electrical conductivity.

The sulfur electrolyte has the property that it conducts Na+ ions but does not conduct either electrons or neutral Na atoms. During cell discharge Na+ ions formed in the liquid Na combine with the sulfur to become Na2S5++ ions that pass through the sulfur layer. These ions then react with the chlorine in the denser NiCl2 to form 2 NaCl, liberating Ni which plates out onto the cathlode. After all the NiCl2 is exhausted the remaining sulfur acquires more Na and fully converts to Na2S5.

During cell charging the Na2S5 converts back to Na2S5++ which can then transport Na+ ions from the NaCl. The Na+ ions again become Na2S5++, pass back through the sulfur, acquire electrons and accumulate as elemental liquid sodium at the anode. The Cl- released by the NaCl combines with the cathlode's nickel coating to form NiCl2, releasing electrons to the external circuit. After all the NaCl is exhausted the remaining sodium polysulfide continues giving up Na until the Na2S5 is exhausted. The sulfur so released in addition to forming S8 can form NiS or Ni3S2 which must be eliminated via complete cell discharge. The primary role of the sulfur/sodium polysulfide is to act as an electrolyte, which transports Na+ ions between the liquid sodium anode and the reservoirs of solid NiCl2 and solid NaCl near the cathlode. The sulfur/sodium polysulfide also provides some energy storage.

The Na-S-NiCl2 energy storage cell operates in two stages during both charging and discharging. There is a distinct change in cell voltage and cell internal resistance at the interstage transition. The interstage transition, in combination with a current integrator, can be used to estimate the remaining useful cell charge. In general the cell is most efficient when it is converting NiCl2 to 2 NaCl during cell discharge and when it is converting 2 NaCl to NiCl2 during cell charging. At the top and bottom ends of the charge-discharge range the internal resistance is higher and nickel sulfides form or are eliminated. The nickel sulfide formation/elimination process reduces the cell's efficiency. In applications involving high charge-discharge rates the cell should be operated in the middle 80% of its operating range where the internal resistance is relatively low and where formation/elimination of nickel sulfides is not an issue.

Initially, when the cell is fully charged, all the Na is in the anode and all the chlorine is contained in NiCl2. During the cell discharge each participating Na atom in the anode gives up an electron to the external circuit and forms an Na+ ion according to the equation:
Na = Na+ + e-
The Na+ ions move into the electrolyte and react with the liquid sulfur to form Na2S++. The Na2S++ is unstable and spontaneously takes on more sulfur to form Na2S4++ or Na2S5++. The net equation is:
2Na++ + 5S = Na2S5++
The Na2S5++ moves through the sulfur, finds a NiCl2 molecule and forms 2NaCl according to the equation:
Na2S5++ + NiCl2 = 2 NaCl + 5 S + Ni++.
Each released Ni++ ion drifts to the cathlode where it neutralizes itself by acquiring two electrons from the external circuit and plates Ni metal onto the cathlode. This process continues until all the available NiCl2 is exhausted. The cathlode must be designed so that the Ni plating onto its surface does not significantly reduce its surface area.

During the discharge cycle second stage there is no NiCl2. The cell behaves as a pure Na-S cell. Na ions from the anode react with the sulfur. The sulfur, which during the discharge cycle first stage only partially converted to Na2S5 now continues to react with Na+ causing all the remaining sulfur to convert to Na2S5 according to the equation:
2Na+ + 5S = Na2S5++
Each Na2S5++ ion drifts to the cathlode where it acquires two electrons from the external circuit. This discharge process continues until all of the available sulfur has converted to Na2S5. At that point there is no energy source to drive the external circuit and the cell is fully discharged. The cell internal resistance during the discharge cycle second stage is higher than during the discharge cycle first stage. There is a surplus of Na to ensure that the process is not Na constrained.

Initially, when the cell is fully discharged, all of the sulfur is contained in Na2S5, NiS, Ni3S2 and all of the chlorine is contained in NaCl. In the charge cycle first stage the Na2S5 at the electrolyte surface breaks down according to:
Na2S5 = 2 Na+ + 5S + 2e-
and the Na+ ions move back to the Na anode where they acquire neutralizing electrons from the external circuit. The two negative charges in the sulfur attract another two Na+ ions from the NaCl reservoir, which restore the Na2S5 concentartion, leaving Cl- ions. These Cl- ions drift to the Ni cathlode where they give up their electrons to the external circuit and combine with the nickel cathlode material to form NiCl2. This process continues until all of the available NaCl is exhausted. There is a surplus of Ni to ensure that the process is not constrained by the available supply of Ni.

The free Cl- ions may also combine with any NiS or Ni3S2 according to:
2Cl- + NiS = NiCl2 + S--
6Cl- + Ni3S2 = 3NiCl2 + 2S---
thus converting any NiS or Ni3S2 to NiCl2 plus free sulfur ions.
Since the sulfur is liquid the sulfur ions can migrate to the cathlode. There the sulfur ions do not chemically combine with the nickel during the charge cycle first stage because of the preference of nickel to combine with chlorine.

During the charge cycle second stage there is no NaCl. The cell behaves as a pure Na-S cell. The Na2S5 continues giving up Na+ ions which pass through the electrolyte to the Na anode. The negatively charged sulfur ions left behind circulate and give up their negative charge at the Ni coated cathlode and form the nickel sulfides NiS or Ni3S2. The negative charge flows in the external circuit. This process continues until all the Na2S5 is exhausted, at which point the cell is fully charged. The cell internal resistance during the charge cycle second stage is higher than during the charge cycle first stage.

However, there is an issue with nickel sulfide formation during the charge cycle second stage. The sulfur and sodium polysulfide are liquids at 300 degrees C but the NiS and Ni3S2 are not liquids. Formation of nickel sulfides reduces the amounts of liquid sulfur and liquid sodium polysulfide that are available to provide cathlode structure conductivity. The nickel sulfides can be removed by completely discharging and then recharging the cell. However, while that process is taking place the cell is not fully available for random energy storage. Hence, a practical energy storage system involves multiple series connected strings of Na-S-NiCl2 cells. Then, with the aid of a suitable control system, each cell string can be completely discharged, partially recharged to clear any nickel sulfides, operated in an alternating charge-discharge mode and then charged only to the point where nickel sulfides start to form. A practical Na-S-NiCl2 energy storage system should have a microprocessor based charge controller to monitor and control the charging and discharging of each series connected string of Na-S-NiCl2 cells. This controller is essential for minimizing the nickel sulfide concentration in the cells and hence maximizing cell string performance.

The sodium in the anode must be above the Na melting point of 97.81 C in order to emit Na+ ions. The sulfur must be above its melting point of 112.8 C in order to allow Na2S5++ ions to circulate. The sodium polysulfide must be above its melting point of 275 C in order to circulate and form Na2S5++ ions. Hence in order to ensure proper operation the cell must be kept above 300 degrees C.

SodiumNa97.81 C.97
Sodium ChlorideNaCl801 C2.165-98.168 kcal/mole
Nickel ChlorideNiCl21001 C3.55-73.077 kcal/mole
Nickel mono-SulfideNiS797 C5.5
Nickel sub-SulfideNi3S2790 C5.82
SulfurS8112.8 C2.07
Sodium IodideNaI660 C3.67
Sodium HydroxideNaOH318.4 C2.13
Sodium MonosulfideNa2S1180 C1.856
Sodium TetrasulfideNa2S4275 C1.268
Sodium PentasulfideNa2S5251.8 C
NickelNi1455 C8.90
Nickel IodideNiI2797 C5.834


1) Due to their different densities the cell components tend to stratify. Over time the Ni cathlode tends to reform at the bottom of the cell. Stratification of components by density R in gm / cm^3 at 350 degrees C causes:
Na top R = 0.97
S8 next R = 1.66
Na2Sx next R = 1.85 to 1.91
NaCl next R = 2.16
NiCl2 next R = 3.55
NiS next R = 5.5
Ni3S2 next R= 5.82
Ni bottom R = 8.9

2) Note that the NaCl and the NiCl2 tend to sink in the liquid sulfur/sodium polysulfide. An important part of cell design is ensuring adequate fluid circulation in the cathlode structure. Note that Na2S5 and S8 are immiscible. Additives are required to make sulfur electrically conduct. Note that at low temperatures the density of S8 is 2.07 and is denser than the Na2Sx. The temperature at which the density of S8 becomes less than the density of Na2Sx may represent an optimum operating point.

3) The net chemical reaction is:
2Na + NiCl2 = 2NaCl + Ni
Energy is mainly stored as sodium metal and as nickel chloride. Nickel chloride is used because it stores chlorine at the working temperature with minimum binding energy. Sulfur provides some additional energy storage at the expense of nickel sulfide formation which increases cell internal resistance.

4) The energy liberated per mole of NaCl is:
98.168 kcal/mole (73.077 / 2) kcal/mole = 61.63 kcal/mole

5) 61.63 kcal/ mole NaCl X 1 mole NaCl / (22.99 + 35.45) gm
=61.63 / 58.44 kcal/gm
=1.0545 kcal/gm X 1000 gm / kg = 1054.5 kcal/ kg
= 1054.5 kcal/ kg X 4.18 kJ/kcal
= 4407.8 kJ/kg X 1kW s/kJ X 1h/3600s
= 1.22 kWh / kg NaCl

6) Active chemical weight is further increased by nickel weight.
Total active chemical weight / NaCl weight
= (2(58.44) + 58.71) /2(58.44)
= 175.59 / 116.88
= 1.5023

7) Hence theoretical maximum energy / kg is:
1.22 kWh / 1.5023 kg
= .812 kWh / kg active chemical
By comparison the presently achieved total energy storage system weight is about 0.1 kWh / kg. Thus there is an opportunity for about a four fold increase in energy storage per unit weight via improved energy storage system engineering.

8) The theoretical minimum weight of nickel required to fabricate the energy storage cell is given by:
(1.22 kWh / kg NaCl) X ((2 X 58.44 kg NaCl)/(58.71 kg Ni)) X (.454 kg Ni) / (lb Ni)
= 1.102 kWh / lb Ni

Experimental experience indicates that the actual amount of Ni used is 3 times the theoretical minimum, indicating that the Ni requirement is:
3 lb Ni / 1.102 kWh = 2.72 lb Ni/ kWh
In early September 2009 the cost of Ni was about $8.12 US / lb, giving a cost of Ni for energy storage of:
$8.12 US / lb Ni X 2.72 lb Ni / kWh = $22.09 US / kWh
At $8.12 US / lb the cost of Ni is tolerable.

9) In industrial quantities the cost of sodium is about $4.00 / kg
In industrial quantities the cost of sulfur is about $.10 / kg

10) The cost of the chemical inventory required to store 1 kWh is:
= $25.00 / kWh of electrical energy storage capacity, which suggests that the large scale fabricated battery price will likely be in the range of $100 / kWh to $200 / kWh.

11) The basic energy storage cell components are:
a) ceramic cell enclosure
b) liquid Na anode
c) proprietary liquid salt electrolyte
d) liquid S8/Na2S5
e) NaCl/NiCl2 reservoir
f) Ni cathlode

12) The classic sodium=sulfur cell failure mode used to be formation of a metallic whisker that causes an internal short circuit. AMBRI claims that its proprietary technology has solved this problem.

Practical energy storage systems contain multiple series connected strings of Na-S-NiCl2 cells. The minimum number of cells in a string is determined by the minimum required discharge voltage. Since the required charging voltage is higher than the available discharge voltage, generally two cell strings are series connected to obtain the required discharge voltage but are parallel connected for charging. For example an inverter that provides a 120V/208V 3 phase wye output can obtain its DC power from two series connected cell strings where each cell string has a minimum output voltage of:
1.41 X 120V = 169.2 volts.
Allowing for a 3.0 volt drop across switching devices and reactors increases this voltage requirement to 172.2 volts / cell string.

According to the published MES-DEA specifications the minimum operating voltage per cell is 1.722 volts, so the minimum number of cells per string is:
(172.2 volts/ cell string) / (1.722 volts / cell) = 100 cells / cell string

Application of similar design rules to a system with a 277V/480V 3 phase wye output results in a minimum output voltage per cell string of:
(1.41 X 277V) + 3V = 393.57 volts / cell string.
The corresponding minimum number of cells / cell string is:
(393.57 volts / cell string) / (1.722 volts / cell) = 228.5 cells / cell string.

Application of similar design rules to a system with a 347V/600V 3 phase wye output results in a minimum output voltage per cell string of:
(1.41 X 347V) + 3V = 492.27 volts / cell string.
The corresponding minimum number of cells / cell string is:
(492.27 volts / cell string) / (1.722 volts / cell) = 285.9 cells / cell string.

From the point of view of both electrical and thermal isolation it is convenient to package all of the cells of a cell string in a single module. From the point of view of practical handling with readily available lifting equipment, it is desirable to keep the maximum individual module weight less than 10 tons (20,000 pounds). Hence the individual cell weight including its enclosure allowance must be less than:
20,000 pounds / 300 cells = 66.67 pounds / cell = 30.2 kg / cell. An alternative is to design for 150 cells per module.

In order to meet charging/discharging voltage requirements a practical energy storage system has a minimum of 2 cell strings. In order to permit full charge-discharge cycling without imposing input or output constraints it is necessary to have additional cell strings. Generally there are either one or two modules per cell string. Each module string needs its own voltage and current monitoring for optimum charging and discharging.

Generally with Ni-S-NiCl2 energy storage systems, to accommodate the variable cell string voltages, there is one power inverter for each pair of cell strings. Each power inverter must be designed to accommodate the wide variations in DC bus voltages and to provide controlled rates of cell charging and discharging. Frequently there is a requirement for AC bus voltage control. The inverters must be fitted with integrators that monitor the Na-S-NiCl2 cell charge/discharge conditon to prevent over charging. The inverters must provide input/output signals to allow efficient control and monitoring of the energy storage system. There must be a means to ensure that the total load is appropriately and efficiently shared by the various inverters. There must be a means of automatically isolating a defective cell string or inverter such that the failure of a single unit does not cause a failure of the entire energy storage system.

Usually the energy storage system charging rate and discharging rate are controlled by an external energy management system that is programmed to minimize the facility owner's total electricity cost and/or maximize his distributed generation revenue. The energy transfer rate into or out of the energy storage system may be adjusted by changing the tap setting of the transformer between the AC line and the energy storage system.

These energy storage modules are heavy and should be installed in a location with easy boom truck access. Such locations have load bearing constraints.

Typically about 30% of the overall electrical energy that is input into the energy storage system is not recovered as electricity but instead is converted into heat. During the charging cycle there are rectification losses, switching losses and internal resistance losses. While energy is in storage there are continuous static heat losses. During the discharging cycle there are switching lossses and internal resistance losses. Ideally the energy storage system should be located such that the heat losses are recoverable and are used for a useful function such as preheating supply air or domestic hot water. However, the energy storage system must be located such that at all times all of its heat output can be safely dissipated and such that it does not impose a fire risk. The combination of access, heat dissipation and fire protection requirements makes practical heat recovery a challenge.

When the energy storage system is fully charged it contains a substantial mass of a highly reactive liquid metal such as sodium. It is essential to prevent this liquid sodium reacting with either water or air. If water penetrates a cell seal large amounts of hydrogen will be released, that can lead to an explosion in a confined space. For this reason Na-S-NiCl2 energy storage modules should be installed in a well ventilated dry above grade location where flooding cannot occur. Similarly there should be an adequate fire separation between these energy storage modules and any combustable adjacent structure.

This web page last updated May 7, 2019.

Home Energy Physics Nuclear Power Electricity Climate Change Lighting Control Contacts Links