If you need to know about batteries; you’ve come to the right place Chinese flag 点击这里访问我们的中文网站 Chinese flag

The highs and lows of bipolar batteries

Dr Mike McDonaghTechnical editor, Dr Mike McDonagh takes us on a tour of the reduced-cost and improved-performance highs of bipolar batteries, and the inevitable lows of manufacturing frustrations that come with any new technology.

The first design of the modern battery was accredited to Alessandro Volta and his Voltaic pile. Not a medical condition, but a series of layered cells consisting of a copper electrode, a cloth separator soaked in salt solution and a zinc anode. The cathode is actually hydrogen, which is reduced on discharge from its ionic state to the gas H2. The zinc is oxidised from metal to whatever compound has the lowest energy of formation in the electrolyte being used. So, going old school we have:

Ζn | Ζn2+ || 2Η+ | Η2 | Cυ 

The Cu electrode is a non-participating noble metal for electron conduction only.

This was essentially a primary battery, later replaced by Gaston Plante’s rechargeable lead-acid cell with which we are all familiar. The modern design of lead-acid battery has some drawbacks, notably the added cost and weight of inter-cell connectors, plus the difficulty of obtaining high-voltage batteries due to limitations of design and production methods. It was in the 1970s that the question inevitably arose: “Why not use a voltaic series pile construction using Plante’s rechargeable lead-acid chemistry?” This would have the advantage of removing the intercell connections and the 2V cells could be stacked together to make batteries of high voltage. Theoretically, this is a simpler and cheaper way to make the batteries, as the intercell connector is removed and the simple stacking process means that large multiples of cells can be easily fabricated. On top of that, making the newly popular 48V construction would be a much simpler manufacturing affair, as would any multiple of a 2-volt cell.

This bipolar approach offers many advantages, most of which have been extolled for several decades. The lack of lead alloy intercell connections, higher volumetric and specific energy densities, lower manufacturing cost, better and more uniform active material utilisation, and lower stock inventories— as external containers for 6V and 12V monoblocs would be unnecessary. The bipolar building block would be common to all designs, albeit each bipolar frame would have a fixed capacity. 

However, methods of parallel connections between plates or batteries, to increase their capacity, are known or already practised by some commercial bipolar companies. Advanced Battery Concepts (ABC), for example, has a unique method of parallel joining of their batteries to double the capacity, without making the coupling obvious. It does sound like a technology with a lot of advantages. 

If only life in the battery universe were that simple. In Volta’s world there were no commercial considerations or performance standards to meet. The impetus behind the first patent for a lead-acid bipolar battery (BLAB) filed in 1979 was to improve the performance and reduce the lead content of the lead-acid battery. Because this was a rechargeable battery with performance requirements, the active material balance, including that of the acid, had to be optimised. This meant that it was impossible to contain the electrolyte in a fabric without leakage between isolated cells. The drawing for the bipolar construction for this patent is given in Fig 1. Immediately obvious from this is the new idea of a bipolar element. It is very different from a conventional pack element in a traditional PbA battery. 

The basic building block is shown in Fig 2. It has a positive and negative active material pad on either side of a conducting substrate. In this particular patent, the substrate is a polymer filled with carbon fibres that provide the required conducting attributes. There are also other materials, including lead stripes, which bond to the substrate and provide a conducting interface bond with the active material. On the positive side there is a Teflon coating applied between the stripes to prevent oxidation of the polymer/carbon surface of the substrate. The bipolar battery is constructed by connecting multiples of these 2-volt elements in series, up to the required battery output voltage.

This concept looks fairly straightforward so far, until that is, we look into the material and manufacturing requirements of these elements and those of the battery’s constituent cells. Fig 3 shows the basic elements of construction in a BLAB. The key points of concern are highlighted to show the importance of the materials used for the substrate and the cell frames making up the element compartments as well as the outer casing. As basic pre-requisites based on Fig 3, we can list the following fairly obvious requirements:

  • Keeping the acid separated between cells
  • Finding a substrate that is acid proof and conducting with a low resistance
  • Ensuring good adhesion of the AM onto the substrate
  • Keeping pressure on the substrate to prevent AM shedding and keep good AM/substrate contact
  • Having an external case that can adhere to the substrate without inter-cell electrolyte leakage
  • Joining the individual cell frames without electrolyte leaking outside of the battery case

We can add some less obvious requirements that become apparent once the above basic problems are addressed: 

  • Increasing the capacity of a cell without building up the AM block on the substrate and increasing the distance from the surface of the AM block to the substrate conductor
  • Providing sufficient pressure onto the AM to maintain a good adhesion without creating a bulging effect in the centre of the battery

Looking at each of these requirements, the first question is that of the substrate, which must be acid-proof and conducting to ensure electron transfer between the positive and negative active masses. Metals other than lead are out for obvious electrochemical reasons, and since the objective is to remove the lead mass, it makes no sense to put in a lot more lead to act as a substrate. Lead would also be ruled out, purely on the mechanical strength and low weight required to give improved energy density. 

This now opens up the realm of materials science, that is, the production of a rigid, lightweight conductor that is acid-proof and can be sealed to prevent acid leakage between cells. There has been a lot written on this subject. The 80s and 90s saw high activity levels in designing and testing different materials. 

In terms of material requirements, there are three fundamental properties that need to be incorporated into the material: it must be lightweight and rigid (mechanical properties); electrically conductive with minimum resistance (electrical properties) and resistant to corrosion and acid degradation (chemical properties). 

There has been a lot of activity in the search to find the ideal material for the substrate. These include: conducting polymers, composites of metals and conducting refractories mixed with resin; even lightweight metals clad in lead have been proposed. The most notable of the materials trialled and which still seem to be of interest are the refractories and conductive polymers. 

In the case of refractories, the conductivity can be achieved using the principle of ‘electron holes’. This is the removal of an oxygen atom from a stoichiometrically balanced crystal lattice as in a metal refractory. The non-stoichiometric crystal structure then has a surfeit of electrons left when the O2- oxygen atom leaves behind a surfeit of two electrons in the valence band to form oxygen gas, or react with a cationic atom such as hydrogen. Defects such as: missing elements, added elements (doping) or structural defects that occur in a covalent structure (such as a ceramic), can all give rise to electronic conduction either as positive holes or a surfeit of electrons. 

A strategy of partially reducing a metal oxide (ceramic) can lead to a structural defect known as an oxygen vacancy, Fig 4. This can be illustrated using the Kroger-Vink notation: 

 

With this situation the oxygen atom has to release two electrons to enter the gas phase as O2 or react with hydrogen to form H2O. Either way, it leaves behind a couple of electrons in the valence band in the atomic lattice. When these electrons are excited into the conduction band, the material conducts electricity. This was the strategy employed by Atraverda (later Ebonex), using their proprietary titanium sub-oxide material as the basis of a bipolar substrate (Fig 5). Other materials currently in commercial operations are silica used by Gridtential, plus various conducting polymers being promoted by materials manufacturers, but not yet exploited for battery use. 

There is also the deployment of a plastic frame that is perforated in the area of AM application as employed by ABC in Clare, Michigan. The frame, instead of being the conductor itself, has the perforations filled with lead, Fig 6. This enables electrons to pass through the substrate to create the conduction required for the AM reactions. They are stencil printed with lead and soldered. The process allows sheets to be different thicknesses, depending on the type of battery and the warranty required. 

The next material problem is to make a sealed cell from the substrate. This must be strong enough to maintain the forces generated by the compressed AGM and the expansion/contraction phases of cyclic duty. This seal is a vital part of the bipolar principle; any minute leak between cells creates an instant short-circuit and effectively removes the contribution of one cell. I know that in the case of Ebonex, a polymer is heat-pressed into the titanium sub-oxide/resin substrate as the substrate is made. This forms the bonding agent when attached to a plastic frame after the pasting process. In the case of ABC, the frame and substrate are one and the same, so the intercell bond question does not even arise. 

Next is the adherence of the active material to the substrate. Again, this is a critical requirement that is a determining factor for the battery internal resistance and its cycle life. So far as I am aware, it is the use of lead or lead-calcium-tin alloy sheet, pressed into the substrate, which is universally adopted for this purpose. Again, this is another non-standard component or material to add to a battery company’s inventory.

The materials so far described as the substrate, and in part the outer frame, or case, of the bipolar battery, have to be joined and be leak-proof, both internally and externally. This is where we encounter the manufacturing problems. When reading the claims of the companies offering this technology, that ‘It can be manufactured using existing battery manufacturing equipment and processes’, I’m afraid that I am unable to stay quiet. 

So, let us examine that claim in some detail. The starting point is the production of the substrate used in the bipolar element. This will probably be a material or component not usually found in lead-acid batteries and therefore neither stocked, nor manufactured, by a battery company. This will mean either setting up a new manufacturing line to make it, or buying it in as a component from a subcontractor. In either case it would require a new manufacturing step, using new equipment and processes. 

This substrate will then need to have the active material applied, the negative AM on one side and the positive AM on the other. Before we start on the problems of doing that, there is the minor detail of achieving a good bond between the AM and the substrate. 

With a conventional lead-alloy grid there is the chemistry of the pasting and curing processes that creates a chemical bond to ensure a good low-resistance contact. This also helps to prevent AM shedding during cycling. All of the substrate materials that I am aware of, used in LAB bipolar designs, are non-lead semi-conductors. The solution that I have seen is to use a lead or lead-alloy sheet pressed into the material to act as the conducting and bonding platform for the active material. Again: an extra process to install, and an extra component for the inventory. 

The next problem is, in my opinion, perhaps the most underrated of all the issues surrounding the process engineering of lead-acid bipolar batteries. I am referring to the pasting, or active material application to the bipolar substrate. The pasting process, which can currently produce somewhere between 20 and 40 thousand plates per shift, depending on the method used, does not lend itself to producing the double-sided, pasted bipolar element. 

The conventional battery plate is produced on a continuous belt, fed either by a continuous strip or individual cast grids that are pushed, side by side, through a trowel roller or orifice paste dispenser. Either way, the whole grid is covered from top to bottom. It is not feasible to have interrupted flows of paste, to leave clear sections of the frame. Instead, scrapers are used to wipe off the excess, which leaves small but significant amounts of paste contamination on these sections. And therein lies the main problem. Bipolar substrates need to have clear edges, in order to adhere to a frame of some sort that acts as a seal between the cells. These areas have to be very clean indeed to ensure perfect high-pressure seals between adjacent cells. 

Fig 7 is a representation of a generic bipole construction. It has a conducting substrate, a gasket to take account of a non-weldable substrate material, a lead or lead-alloy sheet for AM adherence, an AM pad or block that is pasted or pressed onto the sheet and an external frame, which attaches to the substrate and also forms the outer casing when welded together.

The Ebonex approach employed a weldable polymer gasket pressed into the titanium sub-oxide/resin substrate to facilitate leak-proof adherence to a plastic frame. These bipolar plates were hand pasted with obviously a very low throughput. 

Discussions with Doug Lambert, VP Sales & Technology, Wirtz, resulted in a proposal to supply a pasting machine that could achieve the objective at very high plate-pasting rates. So it is possible. In fact, so much so that ABC has devised a pasting process that can enable battery production rates roughly similar to those of a cast-grid pasting machine. 

I say BATTERY rather than PLATE production rate for a good reason. There are in essence only two plates per cell in a bipolar construction compared with anything from 5 to 30 plates in a conventional lead-acid battery cell. With the ABC method, they have a pressing station which takes the substrate, lead adherence sheet and a pre-pasted and shaped, glass mat, active material layer to make a sandwich that is pressed together to form one half of the bipolar element. This is then flipped over where the process is repeated, but this time with the opposite polarity AM to complete the bipolar unit. Fig 8 is a photograph of a frame having the AGM with pasted side underneath pressed against the substrate.

Despite my reservations on the feasibility of realising reasonable production rates for pasting, ABC have demonstrated that this part of the process can be achieved with commercial viability. However, the fact remains that it will require a complete overhaul of a pasting line. The next stage of curing is achieved in conventional curing ovens found in most lead-acid battery manufacturing sites. The downside is that a new process will be required due to the lack of flash drying, which leaves higher initial moisture content. On the plus side, curing times of 48 hours and under are achievable and the required curing cycle can easily be programmed into just about any modern curing oven. I know from my work at Ebonex that this process worked well, and a good cycle life was achieved in all of the design iterations that were trialled during the development of the battery’s design and its manufacture. 

The next stage of manufacturing has been the subject of much discussion and speculation since the 1970s. That is the joining of these elemental frames to prevent external and intercell leaks. Again, speaking from experience, the patented method of Ebonex was to use a lead wire pressed into a recess in the outer case of each individual frame. These bipolar frames were assembled into a battery, placing AGM sheets between them in stacks of 3, 6 or 12, or indeed any number required to achieve the target battery voltage. They were then compressed inside an induction welder to the required size and AGM pressure. The lead wires in the frame were heated by a moving magnetic coil and melted the plastic around them, creating a bond between the connected frames. This actually worked quite well and the end results were quite robust and capable of withstanding the considerable pressures generated within an AGM battery during cycling. However, although it was not a very fast process, the equipment was relatively inexpensive and two or three of these machines could produce the same output as one lead-acid battery assembly line at a comparable or even lower cost.

ABC has adopted a totally different approach. It still requires specially designed equipment and a radical departure from normal battery assembly methods. Again, the bipolar elements are stacked with AGM sheets and compressed to the right size and pressure. However, instead of welding the plastic frames together, the whole assembly is over-moulded by a plastic skin that adheres to the external sections of the frame, Fig 9. This both seals off the individual cells and prevents external leaks. The moulding process also injects plastic rods through holes in the substrates to pin the whole assembly in its place, thereby minimising cell bulging. Why is this important? Very simply, excessive bowing of the partition plates results in variation in the compression of the glass mat separator. This seriously jeopardises the efficiency of the gas recombination within the cell and the percentage saturation of the AGM. It also means that there is uneven pressure on the active material pasted onto the substrate. This presents the dual problem of maintaining good pressure of AM against the conducting substrate and, possibly, promoting AM shedding on cycling. ABC are confident that their patented design and process minimises this problem and I can quote CEO  Ed Schaffer as claiming that end-case bulging across a 27cm frame is less than 0.5mm at each end. 

However, plastic injection moulding at these sizes is not cheap. Despite this, the process is quite rapid and not a lot of these units would be required to replace a two-batteries-per-minute assembly line. Weighed against the cost of modern cast-on strap, through-the-wall lug welding, wrap stack assemblies, having a single inventory of plastic pellets rather than hectares of plastic containers stacked outside, etc., the economics definitely look favourable. 

Even so, this is only part of the point. Extra space would be needed for any system adopted. It would be entirely separate from the normal process, no intercell welding, no short testing, no lid welding and even the pressure-testing station would need minor modifications to accommodate the new size of vent hole. 

In the case of the ABC design, they have taken the view that a common chamber above the cells is the best way forward. This ensures uniform internal cell pressures and they also claim it improves the acid filling characteristics of the battery.  

This is another problem area to be considered. An AGM battery will have multiple apertures, probably greater than 10mm in diameter, which are used to introduce acid into the battery. This is where a performance advantage becomes a processing disadvantage. Small distances between plates means a low IR for the battery with better high-rate discharge characteristics. This is a signature property of bipolar designs. However, with cell spacing in the order of mm rather than cm, apertures can get pretty small— less than 2mm. This can seriously reduce the inflow of acid and seriously slow down the filling process, despite using vacuum pulsing techniques. 

As we all know, once acid is introduced, the reaction with the AM is rapid and generates heat and steam that seriously impedes the further ingress of acid into the cell. The internal cell design and the acid flow within it need to be addressed to ensure proper filling. 

There is also the heat generation to consider. With a conventional LAB, there are metal connections between the cells and a lot of lead in the plate straps, which all conduct heat away from the electrolyte to the terminals. This is not the case with a BLAB. A water bath really is essential, preferably with circulation. Without this, battery temperatures could easily climb over 70°C within several minutes, especially with high-voltage and high-capacity cells (that is, bigger batteries).

Effective acid filling, and the formation programme itself, are quite critical parts of the manufacturing process. Headspace is at a premium in BLABs particularly if looking for high specific and volumetric energy densities. For this reason, headspace for acid, either for initial filling, or for electrolyte expansion in operation, is in short supply. The small gap in the bipolar cell between the substrates and AM blocks is filled by tightly compressed glass mat, which makes efficient filling more difficult. For this reason, the internal pathways for filling acid should provide minimum flow resistance. 

Logically, the first part of the acid shot will hit the top of the AM either side of the glass mat. This creates immediate heat and expansion that will block further downward ingress into the bulk of the AGM. The tendency then is to flow around the outside of the AGM and fill the mat inwards, from the bottom and sides of the cell. It is very important to maximise the amount of incoming acid and its rate of ingress. Apart from vent and internal-cell design, maximising the velocity of incoming electrolyte is important. For this reason, vacuum filling is an absolute necessity, with multiple vacuum shots within a few minutes being mandatory. The vacuum also prevents backpressure from air pockets in the AGM and remote parts of the cell from impeding the electrolyte flow.

The headspace in most designs will, almost certainly, be insufficient to hold a reservoir of acid large enough to maintain full saturation of the AGM during the formation process. A measured amount of additional acid can be held in detachable reservoirs that are fixed to each vent hole in the battery— a strategy familiar to most AGM battery producers. Whilst on charge, the acid volume will increase as water electrolysis occurs and hydrogen and oxygen bubbles are produced; and their volume displaces the acid in the glass mat, which can only move upwards and out of the vent hole. Again, the attached reservoir should be large enough to accommodate the additional acid volume. 

It goes without saying, of course, that the initial filling and reservoir acid are at the right specific gravity (SG) and volume to ensure the fully-formed battery has the correct operating SG and level of AGM saturation without causing leaks on charge. That is not so easy to get right when standard process variations in AM and acid filling weights are taken into account.

In essence, nobody doubts that a lead-acid battery in bipolar form can provide a low-cost and high-performing energy storage device. However, there are still questions around its cycle life and its ability to provide high coulombic capacities without some serious design changes. 

Although, on saying that, simply providing a very high voltage battery, combined with a DC/DC converter would help to solve one, if not both of those problems. What’s more, the whole idea of energy provision from electrochemical sources could shift from current to voltage-biased constructions. A solution that is not without cost benefits. However, the efficiency of the DC/DC conversion would be a key factor when considering that approach. 

With regards to having strange materials encompassed in a lead-acid battery, that particular problem seems to have been solved by ABC, who use only plastic and lead to achieve the required constructional parameters, Fig 10. Likewise, their over-moulding of an external case seems to have successfully resolved the twin problems of external and internal leakage. So, with the promise of seeing ABC’s BLAB design being manufactured in 2021 by Monbat and 2022 by Exide, it seems that many of the issues preventing its commercialisation are largely resolved. You can rest assured that BEST will be there when it happens, and for some time after, to monitor, what I hope to be, a new chapter for lead-acid technology in the 21st century. It’s certainly been a long time coming.

Year of publication: 
2020
Wordcount: 
4,087