Very long battery life with rechargeable lithium ion batteries

Question

I'm building a device with a battery life of at least one year. The options are to use alkaline batteries or lithium ion (with appropriate capacity). I suppose the self-discharge of both batteries is small enough for this use. The lithium ion battery lifetime is usually measured in cycles, however I suppose in this case the shelf/calendar life is the limiting factor. I've tried searching some shelf life numbers for lithium ion, but it seems the value is 1-10 years depending on the source, also I haven't found any datasheets stating the shelf life (searched for some Samsung, Panasonic and LG 18650 cell datasheets as a reference).

1. How/where could one estimate/find performance of the lithium ion battery in this case?

The main reason to use a lithium ion battery in this device is that it can be slim compared to for example AAA batteries. However I cannot even come up with any commercial device (for comparison) that would have such battery life with a lithium ion battery. The batteries and required charging chips are rather unexpensive in my opinion. Now that I think of it, I haven't been able to find reliable source for the self-discharge either, some sources state ~2% and some 10% and the datasheets state nothing.

2. Why would one not make a device with one year battery life using lithium ion batteries? Is the reason perhaps the combination of vague shelf life, vague self-discharge and slightly higher cost?

3. Is there any research on how much battery life a device with a non-rechargeable battery should have from the usability point of view? (slightly unrelated)

Answer 1

1. Start with the manufacturers datasheet (if it does not exist, try emailing them, sometimes they just don't have it in an obvious location), this should include things like discharge cruves at different loads, ESR vs temperature, and other data that can help you plan for your exact device (small lithium like CR2320's ESR shoots up like a cliff when the temperature gets too low), be aware these values will be similar but not exact between similar size, capacity and C rating batteries, e.g. https://data.energizer.com/pdfs/lithiumcoin_appman.pdf

Take these graphs and work with them to make a calculator, I have made on for most of my previous projects in excel, decide on a clock speed, then from there enter in the expected power consumption values and work from there, its usually along the lines of (mode1 current * time in mode1 + mode2 current * time in mode2 + sleep current * sleep time between wakeup)

This gives you the ratios of power used per mode, and how much power is used per active/sleep cycle so you can simply divide the total capacity by this cycle count consumption to work out the length

2. Replacement options 3-5 years after the devices lifetime, unless you pick a common or standardised size, it could be quite difficult to keep your device running, even if you yourself have no plan to replace the battery, leave the option, be it pads on the pcb or something similar, also could make sourcing problematic in future if you need to make more,

3. Depends heavily on the device, I've made water monitoring probes that are still active 6 years in off a single CR2320, though they are now in the part of the ESR curve where they are too weak to use there wireless daily download, and instead I force download them on nice warm day every 3 months or so, (smart software can work miracles here, I kept a tally on the cycles and the temperature, and every day it works out if its safe to transmit)

Main point would be, if non-rechargable is your choice, make replacement possible, does not need to be easy, but possible, I ironically had the reverse of the normal issue, they don't make many waterproof and tiny enclosures, So mine is a normal IP54 enclosure with glue lines heatshrink fitted over the enclosure, to replace the batteries I would need to cut and peel it off, but its possible. If you glue or ultrasonic weld things, then its not really possib.e

Answer 2

The main reason to use a lithium ion battery in this device is that it can be slim compared to for example AAA batteries.

I think this is the key part of your question, which is why you find yourself in a puzzling situation.

Everything about your project seems to be hinting at primary cells. Rechargeability is seldom really needed if battery replacements are once a year.

To get the small form factor I'd suggest using several coin cells (CR2032) spread out in a way to keep your case form factor. You can stack one on top of the other if you have space. Or you can use larger, slightly less "over-the-counter" sizes like CR2450.

The shelf life of Li-Ion is a difficult to estimate thing. As others have stated, keeping a Li-Ion 100% charged for extended periods damages it. You can charge it to slightly lower voltage, like 4.1 or even 4.0 volts before putting it in your device to mitigate that to some extent.

For really long shelf-life you can look into specialist chemistries, like Li-SoCl₂. They have a really high energy density, and also came in a number of sizes including very small ones. I've successfully used a tiny cell of this type embedded in a wall light-switch hole (behind the regular switch). But these are specialist batteries and may be hard to get because of hazmat regulations, depending on your area.

So I can't really answer your question 1 (you can look here for some battery discharge data, including self-discharge, but it is for 1 month only).

For your question 2, the reason I think is almost exclusively cost. Noone would use the more expensive, trickier-to-ship Li-Ion, if alkaline cells were good enough.

Your question 3 is too broad to be answered precisely, as we don't know details of your application. I'd say you're in the best position to make that call.

My suggestion: explore the available battery sizes more. If all your options for primary cells are just AAAs, then you haven't looked enough.