In Hypercritical Podcast #74, the topic of batteries came up, and it was asked what will batteries be like in five years (2017). Being a fan of the podcast and a battery geek I thought I’d write something up.
On the timeline, five years ago, the 18650 cell (the standard Li-Ion battery cell size, billions of cells manufactured annually) had a capacity of 2Ah. In 2012, the best 18650 cells have 3.1Ah of capacity. By 2017, I expect the standard 18650 cell to have a capacity in the range of 4.5-5.0Ah.
For lithium ion batteries the annual improvement rate in battery capacity is about 8%. This was pointed out publicly by Elon Musk, since he is running Telsa Motors I take his word for it (previously I’ve referred to this as Musk’s law, a la Moore’s law for integrated transistors). In 5 years you’re looking at 8% compounded, or about 47%. A laptop with the same sized battery will get almost 50% more battery life given the same power demands. This 8% annual rate might increase due to the tons of money being dumped into battery research and development, both public and private. Some forecasts have this annual rate increasing up to 18% starting in 2013 (page 16 of this PDF, which also has a lot of good information on the future of batteries).
But it is important to balance the energy supply from the batteries to the energy demand from the computer or handheld device to really understand what the battery of the future looks like.
The capacity increase of the cells is driven by a few factors, the main two are manufacturing improvements and chemistry improvements. Chemistry improvements are the dominant factor and what everyone focuses in on. Large changes in chemistry (NiCad to Li-Ion) only come along once every 25-30 years. Meanwhile, smaller improvements in the chemistry occur throughout the interim until the next big shift happens.
In the next five years, a few technologies will come online that will improve energy storage capacity. Two specific improvements are the silicon anode (replacing today’s carbon anodes) and electrolyte improvements to allow for higher voltage batteries (4.5-5V, up from 3.7-4.2 today). These will contribute to the almost 50% increase in capacity mentioned above. In the long term, Li-Air batteries are looked upon favorably because their initial energy storage capacity will be about ten times today’s batteries (and possibly higher as time goes on), but they wont arrive until sometime in the 2020s.
Cycle life is dependent on the chemistry used to make the battery. Around 1,000 cycles that is probably the best we can expect for a leading edge battery. There are other formulations out there (Li-Titanate batteries) that can withstand 10,000 cycles, used for 8-10 years, and be recharged at super speeds, but they have less than half the capacity (per unit weight and per unit volume) of current Li-Ion batteries. If a vendor wanted to provide for more than 1,000 cycles they would need to reduce the depth of discharge of the battery to extend cycle life. This would mean oversizing a battery (100Wh instead of 75Wh) and artificially limiting the battery to operate between 15% and 90% full. Cycle life improves logarithmically with depth of discharge, so a battery that has a 75% depth of discharge would likely see its cycle life improved by 2.5x. But this battery oversizing takes additional space and weight…
And therein lies the issue. I don’t expect anyone, let alone Apple, to keep the batteries the same size as they are now if the capacity is increased dramatically. Instead, I expect the battery size to be slimmed down to make that next generation iPhone, iPad or Macbook to be even thinner than the previous generation.
Likewise, I don’t expect the constituent parts of laptops, tablets and smartphones to consume more energy, rather less. One example is the pending switch to IGZO screens from LED LCD. IGZO screens let more backlight through the display, and reduce the power consumption of the device by reducing the power of the backlight. Figures estimate between 50-90% reduction in backlight power usage, which is one of the largest parts of the battery usage when it comes to mobile devices. The upgrade to the retina display for the iPad shows off how much power the backlight uses – the iPad 2’s non-retina display used about 2.7W of power, while the third generation iPad’s retina display used 7W of power for the backlight. By switching to IGZO screens, Apple could return to roughly the same power consumption level as they had before the retina displays. The resulting lower power draw would mean the fourth generation iPad for 2013 could return to a 25Wh battery found in the first two generations, instead of the larger and heavier 42.5Wh pack found on the third generation devices.
Other components in these devices will also become better at using less power, leading to a net drop in total energy use at the same time as batteries continue to increase in capacity. As discussed in the podcast, this means when companies make thinner and lighter devices that deliver the same usage time as the previous version, people will adopt. Those of us clamoring for 15 hour battery life on our laptops and iPads will be left wanting. The closest we’ll get is lighter and higher capacity secondary batteries.