Phinergy’s battery: energy storage, problem solved? June 6, 2014Posted by Maury Markowitz in electric cars.
Tags: batteries, electric cars, energy storage
You hear stories of new battery tech all the time, and I’ve generally learned to tune them out.
A company makes grandiose claims long before they’re bending metal, and then run into some sort of impossible problem and that’s it, they’re done. You know, like EEStor.
So why am I writing about one now? Because this is Alcoa, and they’re on the road.
If this pans out, and that’s always a big if in battery tech, then basically the storage problem is done, for cars. There are several serious limitations, but it does neatly solve the major problem of providing long-distance driving.
But is this for real? Read on!
A little background
So I’ve heard a couple of variations on the story so far, all of them with various degrees of “entirely lacking in detail”, but I have gleaned a few things and put a few more together with a little help from the Wikipedia.
So lets start off with a little high-school review. Chemistry is the exchange of electrons between atoms. We want those electrons to do something for us first. So batteries basically take two or more chemicals and physically arrange them so the only way the reaction can complete is when the electron gets from A to B, through whatever it is we want to power.
In the case of a typical dry cell, like a Duracell, the chemistry takes place between the zinc case of the battery and the atoms in a thick paste inside (which is why they’re “dry”). You get power out of them because the electrons have to get all the way through the paste to the zinc, and the battery provides a shortcut through a carbon rod in the center. That’s why there’s a little nubbin on the top, that’s the end of the rod. When you connect the rod to the case, with a wire, those electrons want to get through that wire so bad that they’ll light up your flashlight along the way.
Every battery works this way, although the specifics can vary greatly. For instance, in the sodium-sulphur battery the chemicals are molten and separated by a clever semi-porous solid material. In the case of a modern li-ion battery, the lithium metal is separated from the carbon anode with a plastic sheet. Fuel cells are basically the same, but keep their reactants, hydrogen and oxygen, in separate tanks.
The problem, or the main one, with batteries is that the amount of energy they store is small compared to their size and weight. The holy grail of the battery industry is one that improves these numbers, basically at any cost.
One way you can dramatically improve the power-to-weight ratio of your technology is to leave half of it out of the battery. And you can do that if one of the chemicals involved is oxygen, because we can get that from the air. So there’s a lot of research on various “air batteries”, like lithium-air, zinc-air, or in this case, the aluminum-air battery.
Phinergy is an aluminium-air design. They’re not the first; this tech has seen limited use in a few roles already. The problem has always been that carbon dioxide or other nasties in the air get into the battery and gum up the works. They start off working great and then slowly die, long before all the energy is gone. This is what Phinergy claims to have solved, by introducing a membrane that filters out the CO2 and protects the battery.
Now if there’s anything I’m sceptical of, it’s right here. The same membrane would be just as useful to everyone else’s battery projects too. And if you did build such a membrane, the first place you’d take it is to a lithium-air design, which has at least double the energy density. There’s been a lot of effort put into these sorts of filters. So maybe Phinergy just beat everyone. Or maybe it doesn’t really work over the long term. This is the part I’m going to be watching most closely.
Other than that the basic concept is very simple. Alcoa uses electricity from Projet Grande-Baleine to smelt aluminum into sheets. These are then placed in a watertight container along with a caustic electrolyte, the exact nature of which I don’t know. When oxygen is let into the container through the membrane, some of the electrolyte and aluminum mix, causing the aluminum to oxidize (basically rusting). To keep the reaction going, the rust is dissolved into the electrolyte, turning it into a gel, and exposing fresh aluminium. This process releases electrons, but can only continue if the electrons complete the circuit, like in the dry cell. So when you disconnect it, it stops.
All that’s left is keeping the electrolyte wet, which you do by periodically dumping a little water in the tank. Yes, water. From a hose. You gotta like that.
And the rub…
I have to point out that this battery is not rechargeable, at least not by you. The reaction turns the solid aluminum into a gel, and putting electricity back into the cell doesn’t cleanly deposit the metal back on the plates. The same basic problem is true in most batteries, to one degree or another, which is why they don’t last forever.
The system is, however, recyclable. When the battery dies, you send it back to Alcoa’s smelter in Baie Comeau, and they melt it back down into plates and re-assemble it. In terms of materials, the cycle is almost perfect, you can do this as many times as you like with essentially zero loss.
The problem is that in energy terms it’s terribly inefficient. The total round-trip is perhaps 15% efficient – 85% of the energy is lost just heating up the metal to form it back into plates. You’ll lose another 10% of that in the electric drivetrain, and there’s a tiny shipping cost, so you might be looking at 13% well-to-wheel.
Now that might sound terrible, but you have to remember that a normal car is about 13% efficient, and that’s 13% of a CO2-spewing fuel, whereas this is 13% of super-green hydro.
In the grand scheme of things, this is all win.
And now the numbers
Did you think I was going to skip the math?! Ha!
From the CBC report I see that the battery weighs about 100 kg, and provides enough power for 1,600 km of driving – or 3,200, I’ve seen both numbers. Now, of course, 1,600 km of driving could mean anything, but lets take it at face value for now.
1,600 km is 1,000 miles. The Tesla S burns 35 kWh per 100 miles, so if that’s a typical number (and it is) then we might expect this battery to be packing about (1000/100)*35 = ~350 kWh.
Ok, so let’s sanity check this number. According to the Wiki, an aluminum-air battery should get around 1300 Wh/kg (which is one of several ways to measure energy density) and in theory might go as high as 6000 to 8000. This one weighs 100 kg and has 350 kWh, so that’s 350,000/100 = 3500 Wh/kg.
That’s high. Suspiciously high. But not ridiculously high. So I’m going to give them a pass for now.
On the upside
Now for comparison, a typical lithium-ion battery’s energy density is around 200 Wh/kg, less than 1/10th of the Phinergy design. But that li-ion is basically completely rechargeable, whereas Phinergy is basically one-shot. So how do you use it?
The average car in Canada puts on about 25,000 km a year. That’s about 70 km, total, a day, or 45 miles. Using the same numbers from the Tesla S, that’s (45/100)*35 = 15.75 kWh. But the Tesla comes with up to 85 kWh of power. Why?
Because math. This math to be specific, the poisson distribution:
Car trips follow a distribution like this. You can’t take a trip of zero miles, which is why it is still above Y=0 at X=0. But you do take some short-ish trips, to the store and back, in this case about 18% of all your trips. And then there’s the trip you take all the time, to work, 18 km away, which is about 35% of the trips on this graph. You also take some longer trips, and even some really long ones. But mostly your trips are 100 km or less.
That’s where this battery comes in
Tesla is trying to totally remove range anxiety, so they want to have a range that’s way out on the right side of the curve, the “long tail”. That way 95% of your trips don’t need a charge. So a Tesla S can go something like 400 km on a charge, even though most trips might be 70 km or less. And to do that, they need 1,200 pounds of batteries.
So let’s say this Phinergy battery actually works, and we want to re-design the Tesla to use it. Now since the Phinergy isn’t rechargeable, we only want to use it now and then. What do call now and then? Is 80% on the li-ion and 20% on the Phinergy good enough for now?
Looking at the curve above, we see that the average point is around 20, and the 80% mark is about 40 km. So if the average is 70 km, that means we get 90% of the trips with a battery that goes 140 km. 140 km is 87 miles, and with the Tesla’s 35 kWh/100 miles that means we need about 25 kWh of battery.
The Tesla’s 85 kWh battery weighs 1200 pounds. So if we reduce that to 25 kWh, we might expect it to be about (1200/85)*25 = 350 pounds (160 kg). Now we add one of these new batteries, at about 100 kg, or 220 pounds. Now the battery pack is 160+100 = 260 kg. That’s 290 kg of weight we’ve removed from the car, over half the total battery weight.
And what do we get for that? Well for 90% of your trips, there’s no change whatsoever. You drive, you recharge, the Phinergy never gets touched. But when you want to go on that long trip, there it is. And not just a few hundred miles to the nearest SuperCharger station… now you get to go all the way to Halifax.
More realistically, we expect about 10% of your trips to hit the Phinergy. But even then, in most cases it will only be a short hit, let’s say 10 kWh here and there. So you get maybe 25 to 30 medium-long trips before it’s starting to get low. And if that’s 10% of your travel, and the other 90% is off the normal battery, then you’re looking at hundreds of trips before you have to do anything.
And when it does get low, you quick change it at the local shop and get a new one. Off you go, in less time than an oil change. And there’s no oil to change, of course.
This is really no different than how the Chevy Volt works. The Volt is also sized so most trips work off the battery, and then it fires up the gas engine when the battery gets low.
The difference is the weight and complexity. The Volt has an engine, gas tank, oil, cooling system, generator and transmission. That’s in addition to the electric motor and battery.
GM doesn’t say how much all of this weighs, but its based on the GM Family 0 EcoTec, so it’s probably about 100 kg for the engine alone. I’d wager the total is somewhere between 200 and 250 kg.
That gets replaced by 100 kg of battery. And all the complexity, and hundreds of moving parts, disappear.
What about for renewables?
Now the possibility of a lightweight battery with huge capacity sounds like it might solve all the intermittency problems with solar or wind (or nuclear for that matter). However, that 15% round-trip efficiency kinda kills the economics. But, there is one other role where I think this might come into play.
If you want to take your home or cottage off the grid, you need some form of storage. Because the sunlight is so variable, especially in the winter, you need to have a lot more battery than you’d need for any single day. Two to three days is typical, and then you put a generator on top of that.
A version of this battery might eliminate that. The average Canadian house burns about 25 kWh a day, a cottage maybe half of that. So you get enough li-ion to run the house for a day, and have one of these in place of a generator. It’s lighter, smaller, makes no sound, and you don’t have to worry about whether or not you forgot to put gas in it.
The only real difference between this and the car case is the shape of the curve. I’ve yet to see the numbers for homes vs solar input, so there’s the possibility that the 80% case needs 80% normal batteries, at which point this sort of hybrid solution doesn’t make sense.
Another possible role is a pure backup system. This battery holds about two days of power, or four in my case because I’ve done everything I can to lower my use (LED lights, gas appliances, etc). Connected to an inverter this would have kept my house going all the way through the recent blackout (78 hours). It appears it can store in charged stage for very long periods, so that’s perfect for this role. That might make an interesting product for people who want blackout protection.
So, what’s the scoop?
I remain a little skeptical of their energy density numbers, and I’ve also seen a lot of similar projects get to this point and no further. So I’m definitely playing wait and see.
But on the flip side, we have Alcoa and the Quebec government both hands-in. I’d hope they’ve done some due diligence. And this is on the track, for real. This seems to suggest this is more real than, say, EEStor.
So fingers crossed. If this does work, that’s basically that for cars.