Get Complete Project Material File(s) Now! »

The Na-MCl2 battery technology

In this first chapter, the sodium metal halide technology is described in details. Following a historical presentation, the main elements required for basic functioning are reviewed. The actual commercial solutions manufactured by the F.I.A.M.M. group are then introduced, from the cell unit level to large battery energy storage systems (BESS). In the light of this contextual overview, the objectives of my thesis will be exposed.

Invention and developments

Up to more recent developments, this section mostly highlights Johan Coetzer and Jim Sudworth’s recollection of the development of the sodium metal chloride battery, in their 2000 review Out of Africa: the story of the zebra battery (1).

The origin

The discovery in the early sixties that beta-alumina ceramics were good ionic conductor at high temperature (2), opened up the possibility of designing electrochemical cells with liquid sodium at the negative electrode, and beta alumina as solid electrolyte membranes3. Weber and Kummer (3), first described in 1967 a new secondary battery that featured a beta-alumina ceramic as the electrolyte, and liquid sulfur with sodium polysulfide in the positive compartment4 (Figure 1).
This type of accumulator provides an open circuit voltage (OCV) of around 2 V. It must be operated at a sufficiently high temperature (250-350 °C) to keep all of the active material in a molten state, and to ensure adequate ionic conductivity of the electrolyte. Because sulfur and polysulfides are insulant materials, an additional conductive media is needed in the positive compartment5 that extends from the current collecting plate to the ceramic surface. The sulfur and polysulfides are strong oxidants, whereas sodium is a highly reductive metal. The beta-alumina membrane is impervious to chemical attack by these species, but it shows limited fracture toughness, and this is a significant safety issue. Indeed, the direct reaction between positive and negative molten active materials is highly exothermic and therefore must be controlled. This highly corrosive environment also limits the choice of material for other parts of the cell (casing, leads, seal), and their durability (4). Early works that will lead to the Na-MCl2 technology (the so called ZEBRA battery), were driven by investigating alternative active materials that do not present these drawbacks.
4 It will be further referred as a NaS battery 5 Weber and Kummer have proposed the use of a porous graphite felt

The ZEBRA project: from proof of concept to prototyping

ZEBRA research dated back to the mid-seventies. It was first conducted in Pretoria (indeed ZEBRA originally stands for Zeolite Battery Research Africa), where Johan Coetzer and his group started to investigate partly chlorinated transition metal carbides matrix, as an alternative to sulfur compounds. The main benefit of using these metal chloride based materials is that they are non-corrosive, but like the sulfur based species, they are not conductive. The aim of the ZEBRA project was to find a secondary molten salt electrolyte to mediate the transfer of sodium ions between the main solid electrolyte, and the positive current collector. This intermediate molten salt had to be compatible with the latter, and must prevented dissolution of the metal chloride during cell cycling. Indeed, if the transition metal happened to be in ionic form in the melt, it could irreversibly block the ceramic conductive channels. In the early eighties, the binary sodium tetrachloroaluminate molten salt (AlCl3 + NaCl → NaAlCl4) has been identified as a strong candidate by Roger Bones and his team at the Hartwell Laboratories in the UK, after they have joined the ZEBRA project. Chloroaluminate melts are also referred as the first generation of ionic liquids. In addition to its ionic conductivity and its ability to avoid dissolution of selected metal chlorides6, sodium tetrachloroaluminate proved to be an effective asset regarding safety issues (that will be described later in this chapter).

1982-1989: bringing the technology to the next level (Beta R&D and the AAC)

The beta alumina tubes used for the ZEBRA project were the same one supplied by the British Rail Technical Center in Derby (BRTC, UK) for the development of sodium sulfur battery. After this program was terminated, members of the BRTC team (Jim Sudworth, Roger Tilley, and Hamish Duncan) formed Beta R&D Ltd, to develop potential commercial solutions for sodium beta battery (SBB, i.e. NaS and ZEBRA technologies). The Beta R&D company negotiated a contract with the Anglo American Corporation (AAC) from South Africa to take the ZEBRA concept to the next level of development, and to find viable manufacturing route. The first discloser of this work in the scientific literature, was published by Johan Coetzer in 1986 (5).
He proposed a list of transition metal to use as active material (Cr, Mn, Fe, Co, Ni), and presented experimental data concerning FeCl2/Fe and NiCl2/Ni cells (with practical specific energy and power up to 130 Wh.kg-1 and 100 W.kg-1 for the first cycles). The viability of the concept was also demonstrated by using multi kWh prototype batteries to power electric cars, over an eighteen month period. This short communication was followed by two more comprehensive papers published in 1987 by the Hartwell Laboratory team on an iron-based cell7 (6), and by the Beta R&D/AAC team on nickel and iron-based cell8 (7). It is evident from the late eighties scientific literature that the strongest candidates to be used in the positive electrode were iron and nickel (8)(9)(10)(11)(12)(13). The iron cell had the advantage of low cost but the nickel cell was more stable with cycling and gave higher specific power. During this period, efforts have been pursued successfully to optimize initial chlorination of the transition metals, the wetting of the liquid sodium on the beta alumina in the negative compartment, and to identify the additives needed for control of capacity loss and resistance rise. On the other hand, cost effective routes to produce the beta alumina tubes at a large scale production have been developed.

1989-1998 The industrial Scale-up (the AAB years)

In 1989, AAC formed a joint venture with AEG (then a subsidiary of Daimler Benz), AEG Anglo Batteries (AAB), to industrialize the ZEBRA technology. It was decided to focus on a redesigned
slimmer nickel cell to reach the higher power to energy ratio needed for the targeted automotive applications (Figure 2).
Figure 2: experimental C-Class Mercedes-Benz with electric drive and ZEBRA battery, 1993 Consequently, the choice of nickel determined the design of all future commercial cell types, with the positive electrode inside the beta alumina tube, as otherwise a nickel cell casing would be needed and this would be too expensive. In 1991, AAB set up a pilot line in Berlin to produce 30 Ah slimline cells (denoted SL), using new and more reliable processes to bound the collector nickel parts to the alumina components, and ultimately get a totally sealed cell unit (Figure 3).
During the following years, and while testing the battery modules on BMW and Mercedes electric vehicles (EV), other major advances were achieved by modifying the design and content of the unit cell. In 1995, the ceramic-manufacturing plant in UK was able to produce cloverleafshaped beta alumina tube (Figure 4) with increased surface area (+50 %). The associated benefit was a significant reduction of the positive electrode mean thickness while keeping the same casing.
The peak power quality of these monolith cells (denoted ML) has also been improved by substituting some of the electroactive nickel with iron (around 20 %)10. These optimizations have led to a cell with a specific energy in excess of 140 Wh·kg-1, and potentially 100 Wh·kg-1 for a battery pack. These specifications were able to fulfill the EV batteries requirements defined by the United States Advanced Battery Consortium (USABC), such as energy and power density, no maintenance, summer and winter operation, safety, failure tolerance, and low cost potential. This was publicized in a scientific paper written by Cord-H. Dustmann in 1998 (16). Performance advances during the AAB era were also formalized by Roy Galloway and Steven Haslamn, from Beta R&D, in a comprehensive scientific paper (15) published in 1998 too. Figure 5 sums up the 9 On the pictures we distinguish metal shims between the casing and the ceramic which function will be described in influence of the changes in design and active material on both the specific energy and the specific pulse power at 80 %11 of depth of discharge (DOD) of ZEBRA cells, during the AAB period.
However the same year, and despite the fact that the ZEBRA technology was ready for commercial volume production, the two parents companies of AAB chose to redirect their strategies away from the EV application field.


1999-2015 From EV applications to large stationary storage systems

In early 1999, Carlo Bianco the owner of MES-DEA (a major Swiss supplier of components for the automotive market) took this opportunity to purchase the ZEBRA battery IP technology.
He installed a new integrated plant in Stabio for the volume production of both the beta alumina tubes and the battery packs (with a projected production capacity potential of more than 30,000 batteries per year). Production started in 2001, with a 32 Ah nickel iron unit cell (denoted ML/3G), and amongst other configurations 32 Ah/64 Ah nominal capacity packs (denoted Z5, 216 cells, 17.8 kWh nominal energy and 32 kW peak power). MES-DEA are part of the CEBI 11 The USABC goal in 1998 regarding the specific pulse power was 150 W.kg-1 at 80 % depth of discharge for a two third open circuit voltage at the end of a 30 s pulse group of companies and the group’s expertise in volume production has been applied to both the cell and battery, resulting in significant weight and cost savings. MES-DEA advances over the years were publicized at the international Electric Vehicle Symposium (EVS) (14) (17) (18) (19) (20), and towards the scientific community in (21) and (22). The latest major improvements to the electrochemical performances of the cell, have been described in 2007 by Alberto Turconi (23). They are related to the composition and granulation of the positive electrode, and their effects on the electrical performances. Indeed and from 2001, MES-DEA pursued a program12 to enhanced the specific energy density of their unit cells, resulting in the increase from 94 Wh·kg-1 (C-type cell), to 120 Wh·kg-1 (P-type cell), and ultimately 130 Wh·kg-1 for the X-type cell. Xtype ranules composition is the result of:
 adding small quantities of iron sulfide to prevent capacity loss (enhanced stability with ageing),
 using a different brand of supplied nickel powder, with higher surface area to increase the peak power availability,
 reassessing the iodide additive content and granulometry, to better prevent resistance rise with cycling.

Table of contents :

1.1. Invention and developments
1.1.1. The origin
1.1.2. The ZEBRA project: from proof of concept to prototyping
1.1.3. 1982-1989: bringing the technology to the next level (Beta R&D and the AAC)
1.1.4. 1989-1998 The industrial Scale-up (the AAB years)
1.1.5. 1999-2015 From EV applications to large stationary storage systems
1.2. Operating principles
1.2.1. Key elements and cell reactions
1.2.2. The front reaction hypothesis
1.2.3. Focus on the secondary electrolyte NaAlCl4
1.2.4. Additional requirements
1.3. The FIAMM commercial technology
1.3.1. The ML/3X unit cell: design, constituents and conditioning
1.3.2. Unit cell performances
1.3.3. Environmental impact and health issues
1.3.4. Battery modules: influence of cell failures to packs general design
1.3.5. Battery modules: performances and management
1.3.6. Battery lifetime and ageing
1.3.7. Safety and recyclability
1.3.8. Large energy storage systems: general design and applications
1.3.9. Objectives of the thesis
2.1 Electrical performances
2.1.1 Constant discharge voltage curves
2.1.2 Standard charge voltage curves
2.2 Cyclic voltammetry experiments
2.2.1 Full scan of the cell at very low rate Experimental and peaks initial attribution Full scan peaks analysis Discussions about these first results
2.2.2 Additional tests: partial cyclic voltammetry scans
2.2.3 Conclusions
2.2.4 Cyclic voltammetry with increasing scan rates
2.3 OCV dependence of the temperature
3.1 Isothermal electrochemical model of the Na-MCl2 cell: state of the art
3.1.1 Porous electrode electrochemical model requirements in molten salts
3.1.2 The seminal 1D model: governing equations Ohms’s law for electronic conduction ( ϕ1) Ohms’s law for ionic conduction ( ϕ2) Material Balance Charge transfer kinetics Molar average velocity of ionic species in the secondary electrolyte Porosity and volume fractions changes Specificities of the seminal model
3.1.3 Subsequent works
3.2 Integration of the thermal model
3.2.1 Thermal model requirements for an electrochemical cell
3.2.2 Potential heat effects in Na-MCl2 cells
3.2.3 Focus on the reversible heat
3.3 2D model development
3.3.1 Geometrical design considerations and meshing
3.3.2 Parameters and variables of the model Active materials microstructural parameters and quantities Secondary electrolyte melts properties Reference exchange current densities and transfer coefficients Thermal parameters and variables Geometrical parameters
3.3.3 Governing equations: practical implementation in Comsol
3.3.4 Governing equations in domain 1 (negative steel casing)
3.3.5 Governing equations in domain 2 (negative sodium electrode)
3.3.6 Governing equations in domain 3 (BASE separator)
3.3.7 Governing equations in domain 4 (porous positive electrode)
3.3.8 Governing equations in domain 5 (the carbon felt)
3.3.9 Thermal-electrochemical coupled model
3.4 Initial simulation strategies
4.1 Electrochemical two-steps mechanism simulations in isothermal mode
4.1.1 Initial tryouts
4.1.2 Simulations with constant melt composition (non-isothermal model)
4.2 Solid state process simulations in discharge mode
4.2.1 New working hypothesis
4.2.2 Baseline simulation of a constant discharge Baseline simulation for the non-isothermal model Baseline simulation analysis Baseline simulation in isothermal mode
4.2.3 Impact of the rate of discharge
4.2.4 Impact of the initial temperature
4.2.5 Impact of design changes


Related Posts