Envia claims 400 wh/kg and $125/kwh

[Jeff N]

protomech sez:
"Envia 46Ah cell: 3.2V 46Ah, 147 Wh, 0.365 kg = 403 Wh/kg. Cell dimensions 97 x 190 x 10mm."

Where did you get 3.2v? I assume this is meant to be the average discharge voltage. From Envia's website it looks like the average discharge voltage is more like the 3.65v of the EIG cells with the peak voltage being about 4.6v.

From here:
http://enviasystems.com/announcement/

When I look at the specific capacity graph at the above Envia URL in your reply I see a fairly linear voltage drop during discharge from 4.6v to 3.0v and it drops like a rock with very little remaining energy output down to 2.0v. So, roughly 4.6 / 3.0 is closer to 3.75v average than 3.2v.

[protomech]

The top left chart plots voltage vs specific cathode capacity for Li/Li+ and Envia chemstries, not cell voltage vs % DOD which is more typically seen. At least that's how I'm reading it.

Cell capacity is more than just cathode capacity anyhow, though I'm fuzzy on details.

[flar]

The top left chart plots voltage vs specific cathode capacity for Li/Li+ and Envia chemstries, not cell voltage vs % DOD which is more typically seen. At least that's how I'm reading it.

Cell capacity is more than just cathode capacity anyhow, though I'm fuzzy on details.

I'm not sure how the specific cathode capacity differs from DoD, but I'm assuming that the two graphs are a charge and a discharge - at least the insert talks about a charge to 290 mAh/g which matches the endpoint of the light green line and a discharge for 275 mAh/g which matches where the knee in the dark green line happens.  So the light green line is charging and the dark green line is discharging.

The vertical axis lists (V vs. Li/Li+) and I'm not sure what that "vs." is referring to - would that be "voltage potential vs. the positive ion state"?

Also, the insert mentions a C/20 rate which implies a plot of charging and discharging.

[flar]

From here:
http://enviasystems.com/announcement/

The Envia Systems cells are prototype lithium pouch rechargeable cells. The cells have a capacity of 46 Ah and an energy density of 400Wh/Kg. The cell's dimensions are approximately 97 mm wide, 190 mm long and 10 mm thick. The cell's approximate weight is 365 grams. Cell serial numbers are 400WhK-07-005-111205 (designated as 005) and 400WhK-07-006-111205 (designated as 006).

400 Wh/kg * 365 g / 46 Ah = 3.17 V averaged out. Close enough to 3.2 V nominal.

What's odd to me is that the cell very clearly drops in capacity over the first 3 charge cycles.  They throw a loop in to the equation by switching from 100% DOD to 80% DOD after the first 3 cycles, but if you divide it out then really the cell drops under 40 in terms of Ah after just a few cycles.  Do manufacturers really quote an unrealistically high Ah rating that is only achievable on the first couple of cycles off of the manufacturing table?  I know when I read measurements of my cell phone batteries with a battery app the mAh measured is very close to the rated mAh listed on the battery, even after several cycles.

Also, what is the "approximate" in the findings at the NSWC?  Didn't they actually weigh it?  Combine that with the fact that Envia is claiming 45Ah cells, but the NSWC claims they are 46Ah cells and something is odd about the data...the cells weren't sent in to be estimated...

[protomech]

You're right about the discharge / charge plot in the chart. At the very gentle discharge rate C/20 it looks like 3.5 - 3.6V nominal. So perhaps 3.2V is off - though it still looks a little lower than a conventional lithium voltage curve.

The first three discharge cycles for the anode (top right plot) are done at C/20, C/10, C/5. All discharge cycles after are done at C/3.

The real plot that I can grasp is the lower left plot. It shows 3 discharged @ 100% DOD, C/20 (~48 Ah), C/10 (~46 Ah), C/3 (~45 Ah). The next cycles appear to be 80% DOD, and appear to linearly degrade from 33 Ah down to 26 Ah at 460 cycles.

Humor me for a moment and let's say the nominal voltage at C/3 is 3.5 volts. 100% DOD then would represent a gradual capacity loss from 144 Wh down to 114 Wh over 460 cycles. Given an approximate cell weight of 365 g - perhaps the approximate is included because of non-production fittings or test apparatus? - that drops from 394 Wh/kg down to 312 Wh/kg over 460 cycles (79% capacity retention).

That's not great - especially in a lab environment - but perhaps it will improve before it reaches market. 400 Wh/kg and $125/kWh should allow a future Leaf to fit a 60 kWh battery pack (~150 mile range) and a future Brammo or Zero bike to fit 15-20 kWh battery packs (150-200 mile range). 460 cycles with those ranges are 70-90k miles. That's still not great for 80% in-lab capacity retention .. but replacement costs at $2k (bike) and $7k (car) averaged out are still cheap-ish (2-3c/mile bike, 10c/mile car.

[Jeff N]

What's odd to me is that the cell very clearly drops in capacity over the first 3 charge cycles.  They throw a loop in to the equation by switching from 100% DOD to 80% DOD after the first 3 cycles, but if you divide it out then really the cell drops under 40 in terms of Ah after just a few cycles.  Do manufacturers really quote an unrealistically high Ah rating that is only achievable on the first couple of cycles off of the manufacturing table?

No, not in promotional materials, but it's normal for Lithium-ion batteries to lose substantial capacity after the first few cycles because some if the Lithium gets sucked into parts of the cell structure permanently. Normally, these cycles are performed at the factory and the quoted specifications are based on the "real" capacity after the first few cycles and their losses.

[protomech]

Envia. What a mess.

This quartz article is worth a read for anyone interested in the sordid tale. Short version: Envia is likely toast.

http://qz.com/158373/envia-the-mysterious-story-of-the-battery-startup-that-promised-gm-a-200-mile-electric-car/

Bullet list:
- company CEO (Kapadia) and CTO (Kumar) are at odds, with lawsuits in flight claiming that Kumar stole technical work from a previous employer
- Envia can't replicate 400 Wh/kg performance from the 2012 Crane tests, possibly due to an improperly licensed anode
- cell durability is still very suspect .. 50% loss of capacity in 400 cycles

It sounds like there are still some technical merits to the Envia technology, but it's still far from delivering on a production-quality cell with the claimed cost and weight benchmarks. Hopefully those technical merits can be salvaged in future production cell technology.

Unfortunate as we go into 2014 .. nearly as much for the display of deception and greed as for the failure to deliver on the benchmark targets.

[Shinysideup]

Interesting and sad article. This passage caught my attention and led to my naive question:

"An additional plus for electric car batteries is the ability to draw out the lithium fast—that is the power that allows a driver to speed up immediately on depressing the accelerator. But these traits tend to work against each other—you can pack in a lot of lithium, but only draw it out of the two electrodes slowly, which means that you can drive non-stop between New York and Washington, but may be in trouble if you need to quickly maneuver out of someone’s way. Or you can choose the alternative—you can generally accelerate fast, but go only a relatively short distance on a single charge."

Why isn't it possible to have two sorts of batteries in one vehicle with a chip deciding which bank to use. Sensing the demand for quick acceleration, the chip would switch the draw to the "fast lithium" type of battery, whereas cruising at a constant speed would employ the "slow lithium" bank, giving the driver the best of both worlds?

[00049 (AKA SopFu)]

Shinyside - I always wondered about have a capacitor for the power demand, and the batteries for longevity. Most hydrogen fuel cells have to have a battery for that very reason, but why not go a step further with caps?

If you read deep enough into the story, you'll see that Envia did in fact get 400 wh/kg, just like Crane tested. But for only one cycle. That one cycle capacity is what they marketed, not the useful capacity. It's a matter of semantics, but I am very surprised that GM did not look deeper into the Crane results before investing. Reality eventually catches up.

Envia licensed the cathode from Argonne, alongside GM and others. The only thing GM lost is the performance that was promised...they still own essentially all of the valuable technology/IP Envia was using.

[flar]

Interesting and sad article. This passage caught my attention and led to my naive question:

"An additional plus for electric car batteries is the ability to draw out the lithium fast—that is the power that allows a driver to speed up immediately on depressing the accelerator. But these traits tend to work against each other—you can pack in a lot of lithium, but only draw it out of the two electrodes slowly, which means that you can drive non-stop between New York and Washington, but may be in trouble if you need to quickly maneuver out of someone’s way. Or you can choose the alternative—you can generally accelerate fast, but go only a relatively short distance on a single charge."

Why isn't it possible to have two sorts of batteries in one vehicle with a chip deciding which bank to use. Sensing the demand for quick acceleration, the chip would switch the draw to the "fast lithium" type of battery, whereas cruising at a constant speed would employ the "slow lithium" bank, giving the driver the best of both worlds?

I don't think you need 2 different kinds of batteries.  The speed at which you can draw current is only limited for a single cell, but the batteries in EVs are made up of multiple, sometimes thousands of cells (more in a car than in our bikes).  If one cell is sluggish, you combine the output of a hundred of them and you can draw as much power as you want at a time.  EV batteries already do this to customize the voltage and current they can achieve.  You just have to factor this maximum power draw into the electrical strategies in how you wire them up and how the motor controller accesses and combines the output from the battery banks. (assuming protomech will correct me if I've missed something?)

This is why the 0-60 times of the Tesla get faster as the battery pack increases in capacity.  More batteries means higher current draw...

[protomech]

Yes, it is possible to have heterogenous inputs into a motor/controller.

- energy cells and power cells in a hybrid battery
- fuel cell and battery
- engine/generator (Volt) and battery
- super-caps and battery

liveforphysics (engineer @ Zero) wrote a pretty good post detailing why super-capacitors aren't a good match with a battery pack. Short version: super-caps weigh too much and a well-designed battery pack delivers plenty of power for vehicle applications. Noone ever accused an Empulse RR (small power pack) or a Tesla Model S (large energy pack) of being short on power..

https://www.elmoto.net/showthread.php?2732-How-about-adding-a-capacitor-in-parallel-to-my-current-battery-pack-to-boost-power&p=34574&viewfull=1#post34574

As for hybrid cell pack, this is possible as well.

Consider the following two cells:

EIG C020 energy cell
- 3.7 V 20 Ah 0.43 kg
- 174 Wh/kg 1463 W/kg (@ 15% voltage sag, 10C)

EIG F007 power cell
- 3.2 V 7 Ah 0.25 kg
- 95 Wh/kg 2300 W/kg (@ 15% voltage sag, 30C)

When you build a hybrid pack, you have to worry about matching the battery voltage, etc. Ignore that for a minute.

Suppose you want to build a pack to supply 100 kW peak.

- 100% of power from F007. 100 kW, 4.1 kWh, 43.5 kg (95 Wh/kg)
- 50% of power from C020, 50% from F007. 100 kW, 8.0 kWh, 56.0 kg (143 Wh/kg)
- 100% of power from C020. 100 kW, 11.9 kWh, 68.4 kg (174 Wh/kg)

Now, the power pack is definitely lighter. And sometimes, light weight is of paramount importance.

But consider this:
- going from the pure power to the hybrid pack adds 3.9 kWh, 12.5 kg (300+ Wh per kg added)
- going from the hybrid to the pure energy pack adds 3.9 kWh, 12.5 kg (300+ Wh per kg added)

Ignoring the complexity of the hybrid pack, in most vehicle applications where you care about both energy and power, you're generally better served by a homogenous pack.