Prototype Sodium Ion Batteries in 18650 Cells

French researchers have announced a prototype of an 18650 sodium-ion battery. If you’ve bought a powerful LED flashlight, a rechargeable battery pack, or a–ahem–stronger than usual LASER pointer, you’ve probably run into 18650 batteries. You often find these inside laptop batteries and –famously– the Tesla electric vehicle runs on a few thousand of these cells. The number might seem like a strange choice, but it maps to the cell size (18 mm in diameter and 65 mm long).

The batteries usually use lithium-ion technology. However, lithium isn’t the only possible choice for rechargeable cells. Lithium has a lot of advantages. It has a high working voltage, and it is lightweight. It does, however, have one major disadvantage: it is a relatively rare element. It is possible to make sodium-ion batteries, although there are some design tradeoffs. But sodium is much more abundant than lithium, which makes up about 0.06% of the Earth’s crust compared to sodium’s 2.6%). Better still, sea water is full of sodium chloride (which we call salt) that you can use to create sodium.

The researchers are keeping some of the construction details a trade secret, so far. There’s a lot at stake as electric vehicle, and other battery uses are expected to grow significantly in the coming years. However, they are quoting a 90 Wh/kg energy density and over 2,000 recharge cycle life span. By comparison, a typical lithium-ion battery has around 110 Wh/kg and tops out at about 1200 recharge cycles. In all fairness, though, some lithium-ion batteries (like cobalt lithium-ion) can reach over 160 Wh/kg.

What does this mean to hackers? Only a few of us will be building our own batteries (although we’ve seen it more than once). However, many of us do build with portable power, and sodium-ion may be–if not the next big thing–another choice in your battery arsenal. You can see a short video about the new technology below.

69 thoughts on “Prototype Sodium Ion Batteries in 18650 Cells

    1. Metallic sodium would blow up on contact with water but that’s not necessarily what’s used in the battery. Metallic lithium blows up on contact with air but that’s generally not what’s used in lithium batteries, except for some prototypes.

      Anyway, it’s in a hermetically sealed can.

      1. Maybe you meant Lithium-Ion, but even then there are Lithium batteries, which are not rechargeable and not a prototype, that contains Lithium that will actually blowup/ignite in contact with water. Search for “Lithium Battery Bomb”.

      2. You might want to search youtube before you commit to that. :) There’s a video where you’re shown quite specifically how to take apart one of the energizer lithium batteries (non-rechargeable, I grant you) and get a nice little sheet of lithium out of it.

    1. The fact that they leave that out is disturbing to say the least. However, I don’t know that it’s necessarily malicious. The only thing I’ve found so far is a paper linked in Wikipedia that found the voltage for NaFePO4F was about 3.60 V and for LiFePO4F was about 3.55 V. I don’t know if the paper is behind a paywall, but if you can find it, there’s a great plot that shows the switch from Li to Na doesn’t really affect the voltage-discharge curve much at all.

      https://en.wikipedia.org/wiki/Sodium-ion_battery
      http://www.nature.com/nmat/journal/v6/n10/full/nmat2007.html

      1. That makes sense. The electronegativity of Sodium and Lithium are very close. The difference is about 0.05 V, with sodium producing a higher voltage, all other things being equal. Source: a peek at the periodic table.

    2. I have only seen it mentioned at one spot in a wiki for a specific chemistry of 3.6V. Stick a decent LDO made in the last few years and you get good efficiency (~91%) for 3.3V parts.

      https://en.wikipedia.org/wiki/Sodium-ion_battery
      >Sodium ion cells have been reported with a voltage of 3.6 volts

      There is usually a trade off between the capacity and # of cycles. There is only so much you can do to stuff more active chemicals into a fixed size package. I’ll believe it when they actually achieve both the claimed density and # of charge/discharge cycles in a full production parts.

      1. Stick any linear regulator in and you’ve tanked your efficiency. Sure, newer ones will have lower quiescent current, but that’s the least of your worries if you turn all of your voltage overhead into heat.

          1. The buck boost topology has slightly lower efficiency than a buck mode converter, so if it is not necessarily an advantage at the system level that you can output higher voltage. There is very little energy at the end when the battery drops below 3.3V…

            http://www.eetimes.com/document.asp?doc_id=1273123 very detailed analysis.
            >Portable Lithium-ion applications that require 3.3V for a micro hard disk drive will not always show extended battery life when substituting a buck converter with a buck-boost converter. In this test, the buck battery life exceeded the buck-boost battery life by nine minutes when powering a 3.3V output loaded to a typical micro hard disk drive level of 300mA.

            >Most Lithium-ion batteries plateau at a voltage level sufficient for a buck converter to regulate at 3.3V throughout most of the battery life. Past the plateau region, very little battery charge remains, minimizing the advantage of a buck-boost converter’s wide input voltage range. A battery with a plateau below 3.3V is better suited for a buck-boost converter and more likely to provide increased battery life for 3.3V applications.

          2. I agree with qwerty that it would be difficult (if not impossible) to find a linear regulator that *stably* outputs 3.3V from 3.6V in real world usage. Some VLDO (Very Low Drop Out) linear regulator’s datasheet may seem to suggest they can do this, but I find that reliability/regulation suffer greatly at the lowest possible cited drop-out on the bench, and in the real world, fuggedabouddit.

            I thank tekkieneet for the paper s/he cited (I look forward to reading it fully, but I think there’s a caveat to be noted there: it was written in 2006, and we’re mere days from 2016. My favorite voltage converter/regulator *families* didn’t exist then. I agree with the principles laid out — reaching 90% efficiency with a buck-boost is tough at such a low input voltage. The %age hyped at the top of the datasheet is at optimum input.

            If I were faced with this requirement, I’d look at energy harvesting chips. They’re more expensive than regulator chips (but only by about the price of a single lithium cell), and they are designed for quite low voltage inputs (e.g. 200mv for the now 5-year old LTC30i8) , so to them a battery is pure Easy street. They tend to be limited to a low output current, but they also have a wealth of functions (conveniently pinned out) that make running a very sophisticated power management system easy.

            The energy harvesting chips I use won’t get you to you 90%+, but that generation is getting old, and I have high hopes for newer chips I haven’t yet looked at, or which may come in the VERY near future (certainly well before Na ion cells hit the market). The market demand is strong; the IC development research has been robust.

        1. SMPS has larger quiescent currents (easily in the hundred uA and higher depending on the type and sizes) as it has to charge and discharge MOSFET gate capacitances each time they switches. Some of the SMPS chips have to changes mode from PWM for light loads to get around this. In the end of your electronics are spending a lot of its time in sleep mode, those quiescent current adds up.

          You would also need to select SMPS that can get closer than 90% duty cycles as it will never get to 3.3V output even before I*R losses! They need to be in the 99% or operate in 100% with drop out modes.

      2. Do the math: 3.3V/3.6V = 91.2% The current terms cancels out.

        While you can get around 95+% for a well designed SMPS, your aren’t gaining much. Most of the newer LDO uses single digits of uA of quiescent power and have 0.1V or less drops.

        1. I am confused – is 3.6 V the nominal voltage or the maximum? In case of lithium based batteries these are not the same (commonly maximum is 4.1 for nominal 3.6, 4.2 for nominal 3.7 and there are some chemistries allowing the maximum of 4.35 V as well). Your part might be happy with 3.6 V but probably not with 4.2 V. Or am I wrong here? Please someone clarify.

      1. I wouldn’t trust the 3800 ones as they’re usually from non reputable China dealers. The 18650B’s from Panasonic/Sanyo sit around 3400-3500mAh.

        When it comes to batteries, you should definitely go with a trusted brand and Panasonic is very trusted.

        1. I agree, I bought some Panasonic and Sanyo 2600mAh cells and when I made discharge test at 1A I got exactly that capacity. When I bought some crappy Ultrafire or something like that (some even claim 10000mAh capacity) they are usually below 2000mAh, sometimes even below 500mAh. Those are probably recycled cells from dead laptop batteries.

          1. No. Probably not. Recycled cells from laptops usually come in near 2000-2500. Much better than the cheap china shit.

            Either they are simply “badly designed” cells, say a factory that hasn’t found the knack yet. Or they have a small 500mAh battery hidden in there together with some rice or sand.

          2. Accidently clicked “report” because the reply button is missing. Many other forums have the reply button on the right side.
            I did NOT want to report rewollff’s comment.
            I just wanted to say, that they do not have to be badly designed, they are just manufactured crappy. Or, more likely, they are the ones which did not pass quality control, while the other ones are sold as well-known-brand ones.

    1. Are they the usual type of “water battery”, where they’re just activated by water, not actually powered by it? Usually they have a dried soluble electrolyte, water dissolves it and allows the real power source, the metals of the electrode, to react. Once the electrodes are used up, no more power. It’s a novely, with the occasional real use in batteries that need to be stored for a long time without self-discharging.

      I’ve had a look, and the batteries there are rechargable. Not sure what the point is, or the attraction, of any one electrolyte solvent over any other. Not something anyone would really give a shit about. The environmental loveliness of saltwater fades away once you start dissolving chemicals into it. It’s not what plants crave.

      1. > Not sure what the point is, or the attraction, of any one electrolyte solvent over any other.

        Well, for one thing saltwater batteries are nonflammable, as opposed to Lithium batteries. :)

        Also, saltwater is more environmentally friendly than lead or sulfuric acid. Just saying. :)

        1. I’m not sure it’s *really* saltwater, one of the links I looked up on it showed sodium sulfate solution as the electrolyte.

          But all the components are cheap and readily available. And relatively non-toxic. Though I wouldn’t exactly call manganese oxide non-toxic, at least I don’t treat it as such when handling it in powdered form; but it’s still environmentally friendly enough that it’s used in disposable alkaline batteries. And they claim these batteries are being made, at least partly, on equipment originally intended for pressing pills.

          I wonder about the cell voltage, which I couldn’t find. If the anode is carbon as described, that would tend to limit the charge voltage to less than the electrolytic threshold of water, about 1.23-1.25V, else the liberated oxygen breaks down the anode to produce CO2. Maybe there’s something else to it though, I only have limited experience with electrochemistry.

    1. the argument is not good, but I think studying the batteries is good, for human knowledge, maybe finding interesting stuff, and maybe even worse, they might find an interesting detail on sodium chemistry that would advance lithium or another chemistry.

    2. Sodium is much cheaper and more accessible than lithium. The question isn’t whether we’ll completely run out of either; with recycling and careful use, absolute exhaustion is unlikely. But the question is how many dollars per kWh are we willing to spend for temporary bulk storage of electricity, and why would we choose the more expensive (lithium) solution?

      Lithium, being higher up on the periodic table, does have a weight advantage over sodium. For things like quadcopters, where weight is at a premium, and the cost of the battery is cheap relative to the overall cost of the device, it’s always going to have the advantage. There are other portable and transportation uses where the weight advantage is worth the money. But for stationary applications, like off-grid storage of home solar electricity, weight doesn’t matter, and a significant price cut would be welcome.

      1. Yup, there’s a huge, enormous, potential market for massive amounts of battery storage, cars being the obvious one. If you can bring the price down a lot, it could be world-changing. Sodium’s all over the place.

        1. Weight also matters quite a lot in cars although less than in multicopters. In a multicopter the weight needs to be compensated by a constant use of current, in cars the current is used during accelerations, but in essence in both cases consumption is proportional to the weight.

      2. The problem is still a matter of safety. It is not that Sodium is anymore dangerous than Lithium but when every you store large amounts of energy you run the risk of that energy being released in an uncontrolled manner.
        It does not matter if it is a fuel tank in a car, a propane tank, flywheel, or battery. All can go boom, bang, and or woosh. The trick is to learn to manage the risk. For now I would rather see storage batteries at the utility level than in every home until we work out the safety issues.

    3. There is enough lithium in the world, but there isn’t enough production capacity for lithium.

      One Tesla factory sucks up 15% of world lithium production, and there’s a future demand for thousands of them. Everything from electric cars to grid storage needs batteries in a thousand times greater volume than we’re making today, but the lithium production industry can’t scale up that fast.

  1. There’s plenty of lithium, the ocean is full of it. The only problem is that there are few deposits, lithium is really spread out. China has a few operations that pull lithium out of sea water. We’d run out of iron/aluminium etc to make the chassis before we ran out of lithium to make the car’s battery. Same goes for stationary storage.

    1. Aluminum and iron are the 3rd and 4th most abundant elements in the Earth’s crust, with abundance around 80,000 and 50,000 ppm, respectively. Lithium has an abundance of about 20 ppm, or at best about 1/2000th the abundance of iron or aluminum.

      With around 50 pounds of lithium per Tesla, each car would have to use around 160,000 pounds of aluminum and 100,000 pounds of iron to cause the exhaustion of either of those elements before they exhaust the lithium supply.

      Not that I’m suggesting we’re about to run out of any of these, but we’re absolutely not going to run out of iron or aluminum before we run out of lithium.

      But the real question isn’t about running out completely. It’s about the cost of recovery.

      We mostly use copper instead of silver for wires, even though silver is more conductive, and even though we’re in no danger of running out of silver. Why settle for copper? Because copper is almost as good as silver, and it’s a lot cheaper. Likewise, in batteries, sodium holds the promise of being about as good as lithium, but a lot cheaper.

    1. “Took 33rd place in the race” tells you nothing about just how well you did relative to first place. Similary “33rd most abundant” tells you nothing about just how abundant it is.

  2. Unless these have the exact same charging characteristics as the LI 18650, it is a mistake to put them in the same package. This happened with the CR-123 with the LI having a charge voltage twice that of the common alkaline. Mistakes can be made, and when they do, it’s bad.

    1. 18650 means: 18mm diameter and 650 tenths-of-a-mm height. You understand 2032 now too. The prefix in CR2032 tells you the chemistry. There are a bunch of different 2032 battery chemistries. Similarly there are alkaline and NiNH AA and AAA batteries.

      Anyhow…. All those battery formats are convenient in different applications and no single one battery chemistry can monopolize a single battery size.

    2. The voltage is almost the same, so is the capacity. Dunno about charge / discharge rates, but since they’re being suggested as a replacement for Lithium, and since the other figures are similar, and Lithium and Sodium are in the same alkali metal group, I’d guess they’re going to be similar enough to be mostly interchangeable.

  3. The real savior/killer of this tech is what kind of “C” rating they can get out of them. If they can manage a safe and fast charge and are capable of expending a lot of amps at once this thing could rule the market, if not it isn’t going to be very useful in the crazy high drain devices we use Li Ion for now like vehicle batteries, cell phones, high output flashlights ect.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.