Running the numbers on battery specs

July 18, 2007

I’ve been doing some research lately on the different battery technologies that are being formulated for EV use – all proven lithium ion chemistries and cell formulations. If you have a modicum of understanding about math, electricity and engineering, you might find what I found out rather interesting, because there’s stuff out there that isn’t well explained – or well publicized.


1) Lithium ion with cobalt oxide cathode- This is the most proven technology – energy densities up to 200Wh/kg! It’s the technology being used in the Tesla Roadster due to its high capacity and wide availability. But there are drawbacks – the cells are inherently unstable, and can explode or burn when overcharged or physically damaged. That’s why Tesla has extensive technology in place to protect the LiCoO2’s from both the environment and themselves. Cobalt is also rare, expensive, and found in China. It’s also used in the widening Chinese steel industry, so there’s competition for it on the horizon.

Here’s where the numbers show some interesting things:

The Tesla pack has a capacity of 56000 Watt-hours. With 6,831 laptop cells in place that translates to each cell containing about 8.2 Watt-hours. Divide that by the typical LiCoO2 voltage of 3.7V, and the individual cell current capacity comes out to 2.2 Amp-hours. In other words, these. A cell with an energy density of 175Wh/kg.

So that’s the culprit cell. Now, multiply each one of these sucker’s weight (46.5 grams) by 6,831 – and you’ve got 317.6kg, or 698 pounds.

But wait! The total pack weight is well known to be 450kg, or 990 pounds. In other words, the other parts of the system – the extensive monitoring and protection components – weigh in at 292 pounds! That’s 30% of the entire pack weight!

What this means is that even though Tesla’s choice was driven by finding the best high energy density, putting these cells together into a cohesive package is still a very heavy and complicated venture.

2) Lithium iron phosphate

This chemistry is newer, but has a compromise. The advantage is incredibly good power density, as well as plentiful, environmentally-friendly materials. Also, it’s got great safety characteristics. The cells don’t explode when damaged and are much more stable under high charging/discharging stress. The folks at Killacycle have never replaced a single cell. But the cells have lower capacity than traditional lithium ion cells.

Again, the numbers present an interesting case once you get past the surface. Detractors to this chemistry – ahem, Tesla – point to the low energy density as a deterrent. However, citing energy densities as being “less than half” those of cobalt cells isn’t accurate – in two ways, at that.

But first, the caveat. Large format cells, like Valence’s U-Charge, DO have energy densities of less than 90Wh/kg. Similar pitiful numbers are observed in C-cell LiFePo4’s (down to 60Wh/kg!)

However, a scientifically better comparison is to not just compare different chemistries, but different chemistries using the same cell format. An 18650 format lithium iron phosphate battery from Valence has an energy density of 117Wh/kg, versus a cobalt oxide’s of 175. That is nowhere near as bad a difference as critics make it out to be. Additionally, the chemistry’s theoretical maximum tops out at 140Wh/kg (80% of lithium ion’s most stable offering). Automotive format cells of this capacity from A123 are anticipated to be used in the Chevy Volt. So saying that one winds up with “twice the weight” as Martin Eberhard has stated, is entirely inaccurate.

What is more, the weight of the eventual battery pack also has to factor in the heft of the safety equipment. Cooling and padding still help to prolong any pack’s life, of course, but given LiFePo4’s superior stability, a less radical amount of technology is needed to control it versus cobalt oxide. Thus, a less energy dense pack may still wind up with a great capacity rating for the whole car – not to mention a greater safety rating.

3) Lithium nanotitanate spinel

AltairNano’s new anode technology, despite being marketed as revolutionary, is actually predicated on existing, 10-year old innovations in manganese-spinel cathode technology. Both these chemistries have lower energy density, but offer extremely fast charging and good safety characteristics.

All the numbers I’ve punched come out to an energy density of around 85Wh/kg (based on a 900-pound, 35,000Wh battery pack). This isn’t that great, but one has to consider the clear difference that cell format makes in changing this number. Valence’s large format cells are fairly comparable to Altair’s in terms of capacity, and their 18650’s are considerably higher. Additionally, the large format cells that Altair and Phoenix uses have NO active temperature or safety management, if any at all. Put that in a heavy, steel truck or SUV, and the fact that you get a range of “about 130 miles” is pretty impressive. Putting the same pack in a much lighter car undoubtedly would push the range higher – well into Tesla territory.



  1. I found another company that markets a novel cell chemistry – Enerdel (http://enerdel.com/content/view/105/88/) makes Lithium titanium dioxide anode cells with a manganese spinel oxide cathode. A trend these cells seem to share with Altair’s cells is much lower cell voltage (2.5V!). Long life seems to be another defining characteristic – 95% of capacity after 1000 cycles.

    I can’t make much in the way of conclusions, but this still stresses the importance of using good cell design to maximize any particular chemistry. If the cells are easy to package and cool, and recharge quickly while still offering decent specific energy, they’ll be a viable contender.

  2. How do the numbers compare to Lithium Technology Corp. large cell format? They have a lithium iron phosphate (LiFePO4) cells, which they claim is the largest cells of their kind in the world.

  3. Thank you for a very good summary and analysis. Have you considered the effect that the availability of prismatic LiFePO4 cells would have on your analysis? A123, in conjunction with GM, is testing “… a prototype cell, a silvery slab about the size of three decks of cards placed side by side …”

    A deck of cards measures 2.5″ x 3.5″ x 5/8″, or 90 ml in volume. So, these rectangular 270 ml cells would represent quite a space savings and likely some weight (and cost) savings over the 18650 (cylindrical) cell format which has a volume of 16.5 ml. Additional space and weight savings are expected at the battery pack level just from the lessened complexity of integrating hundreds of stable prismatic cells versus thousands of the less stable 18650 cells. However, these savings are likely to be enough to allow a direct replacement of the Tesla’s LiCoO2 battery pack with a LiFEPO4 pack.

    The space savings of this prismatic cell may allow a direct replacement of most, if not all, of the 56 kWh of 18650 cells but there would be a weight penalty. At 175 Wh/kg, the Tesla’s 698 pounds of LiCoO2 cells equates to 1,044 pounds of LiFePO4, at 117 Wh/kg – a 346 pound, or 50%, battery weight penalty. Some weight savings in the Tesla’s 292 pounds of monitoring and protection components may reduce the penalty to, say, around 250 pound. This 9% weight gain from 2700 to 2950 pounds would cause further watering down of the Tesla’s range from 200 to 190 miles due to a ~ 9% increase in rolling friction but none in aerodynamic drag. Tesla Motors took quite a bit of flak when it decreased the Roadster’s range from 250 to 200 miles. So, Tesla would be very reluctant to announce yet another range reduction.

    However, if one substituted the superior chemistry of LiFePO4 over that of LiCoO2 into the Tesla Roadster, it would just about double its battery life, currently listed at 100,000 miles, especially if it is driven hard. This superiority may also allow the Tesla to maintain its 4-second 0 to 60 mph despite a 9% weight gain from 2700 to 2950 pounds. The environmental benefits, cycle life, power, safety and cooling considerations, and the simpler construction of a battery pack using 1/10 the cells should make LiFePO4 chemistry a prime candidate for all future EV designs. The cost and supply of cobalt is also an issue. This is bit OT, but do you have any comments on the supply of lithium?

  4. Pete-
    Thanks for your comment.

    1) Regarding the flat “prismatic” cell format – that reference is the first and only place I’ve seen this info. So good find!

    However, as far as I know they are still focusing mainly on the 32-series cylindrical cells based off of their M1 batteries – except larger:
    http://www.a123systems.com/newsite/index.php#/products/cells32series/. The cylindrical cells, despite being – as you say – more complex to package, are already well-suited to A123’s manufacturing infrastructure, and according to sources quoted in the March 2007 issue of Automotive Engineering International, are also easier to cool. This might be due to the excellent thermal conductivity of the metal can, or the surface area of the can.

    No reason they couldn’t be experimenting with flat designs though, so it is definitely something to keep an eye out for.

    As far as making a direct comparison of a “card deck” test cell to an 18650 cell, A123 actually uses the 26650 format for its power tool cells, which have a volume of ~34.5 mL. Even so, it’s not a great comparison, since we can’t assume that the prototype flat cell is just a rolled-out, flattened A123 cell. It could just be a slab of A123 material of any given weight. Without knowing the weight and capacity of the slab to go along with the volume, we can’t truly judge what improvement -if any – has occurred.

    2) Material constraints:

    Lithium is a rather common element on earth (#33), and is used in a wide array of applications, from batteries to medication. I haven’t seen any reports on current lithium shortages or price hikes to indicate a possible constriction for the automotive industry. The technical processes – and geopolitical economics – of lithium purification and battery manufacturing seem to be the major obstacles right now.

    As funny as it may seem, copper might be more of a problem on the horizon. Many advanced AC motors are going to be making increased use of copper rotors, and there is reportedly a big copper shortage looming. In a local newspaper today, actually, I read about thieves who broke into a college dorm complex specifically to steal copper fixtures: http://www.flathatnews.com/news/958/thieves-steal-copper-from-campus

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: