Dr Mike McDonagh provides insight into the factors involved in choosing and processing suitable alternative antimony alloys for lead-acid battery manufacturers and component suppliers.
Since January 2024 the price of antimony has skyrocketed (Fig 1a). The LAB industry accounts for around 15% of the global antimony consumption as shown in Fig 1b . In fact, the price of antimony has peaked at over US$55,000, which is an increase of over 400% in a matter of months.
This unprecedented price hike has serious cost implications. Many industrial lead-acid battery manufacturers, and component suppliers to this industry, have a dependence on antimony to ensure sufficient hardness, mechanical strength and in- service resilience for their lead alloys. And now they are faced with either higher product costs or a reduction of antimony content in their products.
The pressure is now on for manufacturers to reduce or eliminate the use of antimony in battery and process components.
In order to assess the problems and outcomes associated with alloy changes, a study of a component manufacturer, UK Powertech (for formation connectors) and a lead-acid industrial battery manufacturer, Microtex Energy (for grids) has begun. Four companies are involved in this:
- UK Powertech – supplier of battery formation connectors
- Microtex – manufacturer of industrial batteries
- Ecotech Energy Solutions – consultant
- BEST magazine – to report the results
The initial task in finding alternative alloys to those currently used requires an understanding of the function of antimony and the mechanism by which it provides the properties needed in these components. The items chosen are: the positive spine and negative flat plate grids used in tubular plate traction batteries manufactured by Microtex (Fig 2) and the formation connector heads used in the fabrication of UK Powertech’s formation connectors (Fig 3). To do this we need to examine the manufacturing processes and the in-service demands for both items to determine their required properties. The first is to determine the required properties for manufacture of these parts.
For both applications, the manufacturing procedures need to be addressed and the first process is the casting of the component. This is gravity casting for the lead alloy flat plate grids and pressure die casting for the spine grid and the connector heads.
In all cases, the main requirements are fluidity to fill the mould and solidification characteristics that do not create defects within the castings. The second manufacturing need is that of sufficient mechanical strength to withstand the subsequent processing stages.
For the battery grids, rigidity is needed prevent distortion and damage through the pasting and assembly stages for the flat plate grid. For the tubular spine, the handling through cropping, sleeving, tube filling and assembly also requires some mechanical resilience. The connector heads are a more robust shape and the next process of joining to the cable and over-moulding of the plastic sheath are less demanding than those experienced by the battery grids. However, hardness is important in these components for storage, transport and movement.
The second requirement is the in-service properties.
For batteries:
- Recovery from deep discharges, promoting a higher cycle life
- Creep strength to prevent grid growth causing battery failure by short circuits
- Corrosion resistance to improve battery life when high overcharge is required
For formation connectors:
- Hardness to withstand frequent knocks from fitting and removing connectors from battery terminals
- Resistance to deformation under stress to ensure a good long term fit between the connector and the pillar
- Corrosion resistance to withstand the acid environment of lead-acid battery high charging rates
- Low resistance contact surfaces even when corroded to minimise both energy losses and temperature rise during formation charging schedules
In both battery and formation applications there are mechanical and electrochemical requirements to satisfy.
To appreciate the mechanical properties needed, and to understand how they can be improved in lower antimony alloys, we need to understand how metals deform. This requires knowledge of the role of dislocations in the deformation of metals.
A dislocation is a structural imperfection in a crystal lattice and it is by movement of these imperfections, under load, that metals deform. Fig 4 shows that there are two mechanisms for dislocation movement within a metal matrix. These are grain boundary slip, and the shifting of structural imperfections in an ordered crystal lattice (Fig 4a). In the grain boundary slip mechanism, grain boundaries can slide over one another when under stress. In the second, a matrix imperfection will move one atom at a time as the stress is applied. In both cases the movement can be impeded by secondary alloying atoms.
Fig 4b shows how secondary element atoms will pin dislocations in place, making the metal more resistant to deformation. This is important to improve component strength, hardness and creep resistance. All three of these properties are interdependent. Antimony (Sb) is the primary solute atom used for pinning dislocations. Removing or reducing the Sb percentage will lower the resistance to deformation of the lead alloy. The electrochemical requirements are important, mostly for in-service use in both the connector heads and the battery grids.
To obtain similar or better mechanical properties with lower or zero Sb% alloys we can refer to the phase diagrams of both Pb-Sb and Pb-Ca alloys. These show the phase compositions of alloys as they cool from the liquid to the solid state. The phases refer to the physical state of the alloys at different temperatures and ally concentrations. Specifically, Fig 5a shows the lead rich end of the Pb-Sb diagram, the liquid and solid states and their compositions at different temperatures and solute element concentrations. In the case of lead antimony, the dislocation pinning atoms described earlier are antimony. These pinning atoms are the result of precipitation due to the difference between the solubility of Sb at the solidification temperature at its lower value at ambient temperature.
This precipitation occurs over time and is explained by examination of the phase diagrams of a binary lead antimony alloy. By the time the component reaches room temperature, the solubility of Sb is reduced from 2.0wt% Sb to 0.004wt% Sb, a difference of 1.996% Sb. Much of this solute is trapped in solid solution due to the rapid cooling of the molten alloy. This precipitates out over time as a uniform dispersal of pinning atoms, causing the alloy to age harden. Comparing this to the Pb-ca phase diagram we see that in the case of Pb-Ca there is only 0.07%wt of solute Ca atoms available to precipitate. This difference contributes to the higher strength and hardness of Pb-Sb alloys compared to Pb-Ca alloys.
Looking to find substitute alloys for the applications mentioned we can split them as follows:
- Case 1 – Connector heads for UK Powertech currently 8% Sb reduce to 1.6% Sb
- Case 2 – Spine casting for Microtex currently 5% Sb reduce to 1.6% Sb; Flat plate grid casting for Microtex currently 1.6% Sb change to Pb/Ca/Sn alloy
Case one: Formation connector head
There will be three factors to consider:
- Processability
- Hardness and wear resistance
- Corrosion resistance
Connector head processability
Test samples were produced by gravity casting, however, the actual process will be pressure die casting, which is similar to the spine grid casting method. In both of these cases there is a very rapid injection by hydraulic pressure into an empty mould. This is followed by a slightly slower compaction phase, again using hydraulic pressure.
The total time from start to finish is a maximum of a couple of seconds, often less. The critical factor here, other than the casting machine settings of temperature and pressure, is the solidification profile of the alloy.
Fig 5a is a lead-antimony phase diagram. This shows that addition of Sb up to 11% will reduce the melting point of lead. This is a principal reason for antimony additions to cast lead alloy products. Fig 5a also shows the solidification of a low antimony (1.6% Sb) alloy from the liquid state. This solidification profile applies to both gravity and pressure die casting.
In the short space of time that the alloy is solidifying, the alpha phase (a solid solution of Sb in lead) is nucleating as dendrite (tree) shaped particles. This spiky tangled mess of growing alpha phase is surrounded by a low melting point eutectic liquid that enables the continued flow of alloy around the mould interior. The larger and spikier the growing alpha phase, the lower the fluidity and filling ability of the alloy. The lower the Sb content, the more likely it is to produce the larger dendritic grain structure with less eutectic liquid to enable the free flow of the alloy.
To mitigate this problem, a secondary element – selenium – is used as a grain refiner. This metal has limited solubility in lead, and rapidly precipitates out while the alloy is still in the molten state. The tiny precipitate particles act as nuclei to seed the alpha phase precipitation in the pasty region of solidification, which results in smaller finer grains.
In addition, Se also affects the free energy of the solid- liquid interface, resulting in a more rounded shape to the dendritic arms of the nucleating particles. Both of these factors combine to enable better, tangle-free passage of the alloy through the mould. It also enables liquid eutectic to flow freely between the solid metallic grains to fill any voids remaining in the casting. In terms of pure fluidity related to melting point, this can also be enhanced by tin (Sn) additions at low levels.
Connector head hardness and wear resistance
Connectors have daily operational requirement of removal and replacement on battery terminals. During the formation process, acid spray from the vent plugs contaminates the battery lids and connectors. This acid will corrode the connector heads’ contact surfaces.
They are also subjected to heat, particularly if incorrectly fitted or if there is a defect on the pillar casting that prevents proper surface contact. Unfortunately, the wear and tear does not end there. When collected from the circuit, connectors are usually thrown into bins and left in an acidic state. This can result in mechanical damage and substantial corrosion of the internal connection surface. This corrosion layer does increase the contact resistance between connector and battery terminal.
The current alloy used by UK Powertech is an 8% Sb binary alloy. This level of Sb gives sufficient hardness and corrosion resistance to provide an adequate service life. Due to the high level of Sb, it also is relatively easy to cast by both gravity and pressure die casting methods.
The proposal is to substitute this with a 1.6% Sb alloy containing selenium (Se), arsenic (As) and tin (Sn) as additional elements. As mentioned, the Se aids in casting. However, the finer grain structure also helps to prevent cavities in the surface due to corrosion. The As acts as an age-hardening agent to increase the effectiveness of the lower %Sb in the alloy. Sn not only, as mentioned, aids in the alloy fluidity in casting, it also provides additional corrosion resistance by ensuring that the passive PbSO4 corrosion layer formed on the inner surface will still be conductive. Table 1a shows the new 1.6% Sb alloy composition with the tolerance bands for the additional secondary elements.
In addition to the secondary element additions there are also the measures of heat treatment and quenching that will raise the hardness and mechanical strength of the low %Sb connector head. Age-hardening by heat treatment for low Sb alloys is very effective.
Fig 6 is a graph showing the effect of Sb content and quenching on age-hardening. It also illustrates the benefit of post-casting heat treatment and quenching on the alloys’ hardness. This graph is based on binary Pb-Sb alloys. It shows the effect of Sb content on both as cast and quenched alloys in line one.
The effect of heat treatment at 240°C then quenching the same alloys is given in line two. In both of these cases there is no age-hardening, and the graph clearly shows that the Sb content is the dominant factor related to hardness. However, when age hardened, we get a maximum value at around 2–3% Sb for both treatments. This is of major importance in the %Sb reduction strategy.
To obtain optimum hardness values, samples of the 8% Sb binary alloy and the new 1.6% (actually 1.7wt% Sb) Sb alloys were prepared by hand-casting the connector head component under the following conditions:
- Pouring temperature: 340°C
- Mould temperature: 130°C
- Quenching water temperature: 30°C
Four sets of samples were produced:
- 8% Sb alloy cast and quenched
- 1.7% Sb alloy cast and water quenched
- 1.7% Sb alloy cast and air cooled
- 1.7% Sb alloy cast air cooled then heat treated at 240°C and quenched
The samples were then hardness tested using a local laboratory with a Vickers hardness machine The test results from the trials are given in Table 2 . The 1.7% Sb control heat treatment and quench show that the highest hardness value. The 1.7% Sb quenching direct from the mould gives a substantially higher value than the 8% Sb air cooled. The 1.7% Sb cast and air-cooled sample gives the lowest value. It was substantially less than the current 8% Sb alloy.
From these results it is evident that with quenching, and/or heat treatment, the wt% Sb in the connector heads can be drastically reduced without compromising the connector’s mechanical properties.
Connector head corrosion resistance
Another concern for a new alloy is the surface corrosion layer formed on the inside of the connector head. Fig 7 is a cross section of the head and is an SEM picture and EDX analysis of the corrosion layer of a connector in service for many months. The SEM shows an internal surface reacted with sulphuric acid to form PbSO4.
Within this layer there are shiny particles of Sb and eutectic. These particles provide good electron conduction to prevent complete passivation of the inner connector surface where it contacts the battery terminal. Reducing the Sb content in a binary alloy would effectively increase the resistance of the corroded interface connection. For this reason, tin is added which aids in both casting and providing metallic conducting particles within the corrosion layer.
Summary
The reduction of antimony from 8–1.7 wt% Sb requires further treatment and other secondary elements to provide the required processing and in-service requirements. The results show that a new alloy, with 1.7% Sb, will provide a more than adequate substitute for the binary 8wt% Sb alloy, provided that it undergoes quenching, and/or further heat treatment after casting.
Case two: Microtex
For case two, the tests were conducted on both a spine grid and a flat plate grid used in the manufacture of batteries. Fig 2 shows two typical grid designs used in industrial batteries. In this we have:
- A pressure die-cast spine with an original alloy of 5% Sb plus tin. This has been reduced to 1.6% Sb as in Table 1a.
- A flat plate grid, originally made of a 1.6% Sb alloy. This is now cast in a lead calcium tin (Pb-Ca-Sn). The alloy composition is shown in Table1a.
Following the same procedure as before, we follow the three requirements of casting, mechanical and in-service properties.
Grid casting processability
- A – pressure die-cast spine: This will be an alloy change of 5wt% Sb down to 1.6wt% Sb. The UKP connector head component is also pressure die cast, the processing implications of an alloy change to 1.6wt% Sb for that component will be the same as those for the spine casting. For this reason, the casting description will not be repeated.
- B – gravity cast grid: This is a shift from a low Sb alloy to a lead-calcium-tin (Pb-Ca-Sn) ternary alloy. This has different solidification characteristics which are described by the Pb/Ca peritectic binary diagram Fig 5b – in this case the cooling profile differs from 5a specifically in having almost no liquid solid phases in equilibrium during cooling from liquid to solid. Critically, there is virtually no dendritic growth during the solidification stage, and the difference between the solid solution concentration at the melting point and room temperature is only 0.08wt%. This gives a far smaller proclivity to age-harden than that of PbSb alloys. The alloy is therefore softer and relies on age-hardening and heat treatment. However, there are additions of tin (Sn) and aluminium (Al) that improve alloy flow and reduce secondary element loss respectively. For gravity casting, the temperatures are generally higher at all stages due to the higher melting point of the alloy (Table 1b).
For of the casting processes mentioned there is a generally neglected part of alloy metal processing, i.e. the treatment of dross. Fig 8 shows a typical method of dross removal from a lead alloy pot, with a very shiny, part-metallic mash in the dross pot.
Many of the added elements have limited solubility, Typical dross removal, takes out the secondary elements. This causes further depletion within the bulk molten metal. Dross should be stirred into the pot before removal particularly As and Se. When the surface of the melt is scraped clean to remove dross, these elements, along with tin compounds, are removed. As a result, more of the elements diffuse to the surface down the concentration gradient and are consequently depleted in the bulk metal. The correct procedure is to stir the dross back into the pot before its removal. The machine settings that will minimise loss of secondary elements are given in Table 1b.
Grids’ mechanical properties
In both spine and flat grid designs, the mechanical requirements for processing are the same- stiffness and hardness to enable the handling and stresses encountered in the filling, pasting and assembly processes. A test was devised in order to assess whether or not a PbCa flat grid is strong enough for the pasting process. Fig 9 shows how the angle of bend of a grid is a measure of its stiffness and therefore its processability. For the positive grid, however, we need to introduce an additional mechanical strength property, that of creep strength. This requirement is to minimise grid growth and corrosion resistance in service, for maximum life expectancy.
For this, we needed a test of mechanical properties within the existing factory quality checks that would predict the creep strength. We chose a tensile test to measure the mechanical strength of spine castings. This can be justified from Fig 10 This shows the relationship between the mechanical properties of yield strength (YS), ultimate tensile strength (UTS), and Youngs modulus (E). This graph describes the stress/strain relationship, or how much a metal will elastically stretch (strain) under a given load (stress). Creep resistance is related to the YS and the YS is related to the UTS of an alloy. This is an oversimplification but broadly valid.
Spine corrosion resistance
The action of charging will corrode the positive plate to PbO2. This has two detrimental actions, particularly for the positive grid:
- It will eventually corrode the entire grid to a metal oxide. At some point there will be insufficient metallic content to provide adequate conduction. At this point the battery will fail.
- The corrosion product has a higher volume than the metal alloy to which it is attached. This places a stress on the grid frame, or the spines in a tubular design, which results in an amount of strain, i.e. elongation of the grids (Fig 11). Failure can result from a short circuit when the plates in the positive group grow to touch the strap of the negative group. This is characterised by the commonly known ‘positive pillar growth’ observed in 2V traction cells.
Fig 12 shows that there is minimal corrosion of the 1.6wt% Sb alloy compared to the other alloys shown in the graph. This is a positive characteristic, particularly for a positive grid spine. The second point of creep strength, already discussed, raised the question of how to improve the mechanical properties.
Spine casting trials were conducted by Microtex to ascertain the effect of water quenching spines, produced from a pressure die casting machine. The UTS of spines from the grids were compared with those that were cast at the same time and under the same conditions, but were allowed to air-cool and were not water quenched. These were tested each day for one week on a tensile tensometer.
The results, which are at an early stage, are for the UTS measurements only. These are shown in Fig 13. After one week of ageing we begin to see a slight increase in the mechanical strength of the quenched spine compared to the air-cooled sample. The tests have just started so results are a little thin on the ground.
Financial summary
Table 4 gives the evaluations for case one of changing the alloy composition from 8wt.% to 1.6wt.% Sb alloys, and in case two of changing from a 5wt.% Sb to a 1.6wt.% Sb alloy for the positive spine casting. The flat plate grid will be evaluated once a suitable alloy has been determined.
Both the component and spine grid show a very substantial lowering of the products’ manufacturing cost. In the case of the formation connector, the cost reduction is around US$0.8 per connector.
Whilst it does not sound massive per unit, we have to bear in mind that these are a consumable and are replaced at regular intervals by battery manufacturers. Purchasing orders usually specify tens of thousands in a single order. Over the course of a year, a manufacturer can save a battery formation department upwards of US$20,000.
Regarding the spine grid alloy, this is far greater due to the weight – in fact it is US$1.07 per grid. In an average 2V traction cell with five positive spines this will amount to a very substantial US$5 per cell. In both cases these alloys will enable the manufacturers to maintain their competitiveness, and hopefully, improve their bottom line.
