Former CEO, LG Chem Power
Prabhakar has been at the forefront of the evolution in vehicle electrification for 40 years. Prabhakar is a trained aeronautical engineer and started his career at Ford where he worked his way to Chief Engineer of Hybrid Technologies which was the lead role on the Escape Hybrid, the first electric production programme by a North American OEM. After realising lithium-ion batteries were the future, Prabhakar left Ford in 2005 and joined as CEO of Compact Power which later became LG Chem, one of the largest battery manufacturers globally. Prabhakar currently advises various electric battery, technology, and vehicle start-ups. Read moreView Profile Page
Could you start by giving us a bit of a history into different types of chemistries and formats in EV batteries?
I think there’s really the three main ones that have been there since day one, day one being somewhere around 2005. The oldest is the cylindrical format, the 18650 as it’s commonly known, the 18mm diameter, 65mm height, used in consumer electronics. It used to be used in laptops, until laptops got too thin to accommodate it. But then Tesla and Panasonic have adopted it for vehicle use. That format is better suited for NCA chemistry, which is also something that Tesla and Panasonic have selected, so that’s a combination that works together, whereas the rest of us in the industry concluded that the better format was a flat format because the size of the cylindrical cells gets limited. You cannot make it too fat, otherwise you get too much temperature difference between the center and the surface of the cell. Plus, it involves more parts, so for various reasons, most of the vendors agree that it’s better to go for cells, make them as large as possible to cut down on the number of cells, which then reduces the overhead when you make them into a pack, in terms of the battery management, thermal management, et cetera.
So, the two predominant flat-formats, which is a rectangular footprint, are either pouch or prismatic: the pouch is basically where the active material in some sense is shrink-wrapped in the enclosure; whereas, in the case of prismatic, it is enclosed in a hard metal shell. But both of these formats cannot use NCA chemistry because of the gassing that it undergoes and the pressure changes, which is what the cylindrical cell can do, and that in some sense also means we use the NMC nickel-manganese-cobalt cathode composition chemistry. So that’s the brief history of these two major chemistries and these three formats.
There’s also the LFP chemistry, lithium iron phosphate, which is something that was pushed originally by the Chinese government because it does have a little bit of a safety advantage, but it suffers from poor energy density — as much as maybe 20% compared to NMC — and as a result, for most passenger vehicle applications, NMC is the preferred chemistry. However, LFP will have its role, both in buses and stationary energy because of its advantage in safety because you can evacuate a personal car in a few seconds, a bus takes a minute, so the safety edge becomes important, and similarly, in stationary applications where the energy density doesn’t matter, so it will have its place. Plus, it does not use any cobalt, so for that reason, going forward, these three chemistries are going to be the dominant ones. NCA used by Tesla and Panasonic, NMC used by the rest of us, and then LFP for stationary and buses.
Why did Tesla choose the NCA chemistry?
Because it allowed them to get the lower cobalt faster than we could with NMC. Again, as you may recall, consumer electronics cells used to use pure cobalt, LCO, but that was bad for two reasons: toxicity of cobalt and cost of cobalt. So, we started, with vehicle application, using what’s called 333 or 111 equal-proportion nickel-manganese-cobalt and went to 532. Today, we’re at 622, which is 20% cobalt. The NCA is already below 10% and maybe closer to 6% to7% cobalt. However, Tesla had started with cylindrical cells, and there’s a bit of history behind that also. When Tesla started, none of us major battery manufacturers wanted to invest because Tesla was a question mark and was too risky an investment. But for cylindrical cells, there was a glut on the market because laptops had gotten thinner, and Panasonic actually had 40% to 50% overcapacity, so Tesla got a great deal on cylindrical cells. It was appropriate for NCA chemistry, so that combination has gone forward and now, because they put so much into making those cells work and doing a great job of system engineering on packs, they are going to stay with it for some time. However, if you fast forward as the market grows – and it may grow by a factor of ten by 2025 – and the majority of that is going to be NMC, which will then tip the scales in favor of NMC for economies of scale that Tesla had for some time. So, that’s going to change the cost equation in favor of NMC. It remains to be seen whether Tesla pivots over at some point to the flat format and NMC.
Where are we today on unique cost of the cell and the pack?
In major manufacturers for Western markets where you’re offering ten years of warranty, it’s about $120 a kilowatt/hour at the cell level and between $160 and $170 dollars a kilowatt/hour at the pack level.
That’s a total, you mean. So, $40 to $50 per kilowatt/hour for the pack? For the battery, for the systems around that, plus the cell cost, which is $120?
Right, so 75% of the pack cost is made up of cells and the rest is battery management system, thermal management system, and the manufacturing cost.
Where do you think Tesla and Panasonic are versus that industry-wide number?
It’s always been hard to pin down Tesla, and what also makes the comparison a little bit hard is the level of integration they do; the BMS, for example. When we build packs, all the battery management system is in the pack, whereas I think Tesla take some of that content out and integrates it with some of the other ECUs. So, given that, my sense is that Tesla is about the same, maybe a little bit lower on the cell cost, and correspondingly, a little bit lower on the pack. Not much more than maybe 5%. The numbers I quote are prices, so they do include a little bit of margin, both at the cell and the pack level. Whereas, I think, when Tesla quotes their numbers, they’re always cost, which will make it 8-10% lower.
So, let’s say they’re 10% lower today. As the industry scales and, like you said, the capacity in the industry is going to be, let’s say, ten times bigger in ten years; it’s pretty hard for Tesla to have an advantage at scale versus NMC.
Today, NCA has a little bit of a scale advantage. Out of the 70 gigawatt/hours capacity that was used last year for passenger vehicles, a significant chunk was NCA, maybe 30% to 40%, whereas the rest was split between NMC and LFP. But, if you fast-forward to 2022-2023, a significant chunk of the market will be in NMC for passenger vehicles, which is going from 70 gigawatt/hours to 700 by 2025. That’s actually tipped the scales in favor of NMC compared to NCA because there is still quite a bit associated with processing costs. If you look at just the cost of the metals, that makes up about 30-40% of the cost of the cathode material which, in turn, makes up around 35% of the cost of the cell. With that kind of a gap, as you get economies of scale, NMC is going to have the advantage over NCA because Tesla and Panasonic are the only players who will continue with NCA, and at some point, Tesla may consider switching over.
So, that could be a considerable disadvantage for Tesla then, if they stick to NCA?
Correct. So, the question is, when will that come through? Because the other thing to notice is that Panasonic is now tied up with Toyota, and it remains to be seen whether Toyota tries to go with NCA. Some of the smaller OEMs go with cylindrical cells and potentially, with NCA chemistry, because they don’t get the attention from the major players and, therefore, it’s easier to make packs using cylindrical cells, but that is a relatively small amount of volume. However, if Toyota decide to go that way, it will be something significant, but I don’t think they will. Many years ago, Toyota actually had a fire in their battery factory when they were using NCA, so I think they’re going to think twice before going NCA. Therefore, Panasonic, at least for Toyota and others, may switch over to NMC, so that just leaves Tesla as the main proponent of NCA chemistry. They may see the advantage of making the switch from a cost perspective because of the economy of scale.
Is there any advantage in using NCA in terms of range for Tesla?
No. I think when you say range, it really comes down to energy density, the amount of battery you need, whether it’s mass or volume, and today, it’s pretty comparable in terms of about 260-270 watt/hours per kilogram at the cell level.
So, there’s no material advantage? From the consumer’s point of view, when they’re buying a car, the key thing is range, affordability, “How far can I drive this?” So, there’s no advantage for Tesla in terms of using that battery chemistry to improve range?
Correct. The thing that Tesla has done is led the way for the industry towards this flat, rectangular battery pack, low-profile, about 100mm high or four inches, that sits underneath the load floor of the vehicle, which doesn’t screw up the ground clearance too much, nor does it raise the floor height too much and yet, it gives the vehicle very good ride and handling characteristics because of the low center of gravity. And if you see the new platforms that we use, MEP and others, that’s where the industry is headed. But given that it comes down to the amount of volume and mass that you have available to allocate to the battery, there is really no significant difference between NCA and NMC for achieving the desired range and, therefore, energy content of the pack.