Monthly Archives: September 2016

Supercharging ala Carte

News leaked this week that some code for a pay as you go supercharging infrastructure is going to go live at some point. It looks like Tesla is going to lower the price of at least some of their Model S and X models by making supercharging an optional upgrade.


Some people who have done some math on the cost of supercharging now use $2500 as the cost of the upgrade. Supercharging was an option on the original Model S 60. To upgrade an old Model S 60 now costs $2500, but it cost $2000 originally. Assuming the cost for permanent supercharging will be the same, the calculations should use $2000 instead of $2500. The costs are different in different countries, but the calculations are probably similar for other countries.


Tesla is in a spot of difficulty with some state laws. Some states allow anyone with a charging station to charge by the KWh unless they are a utility. This may be the core reason Tesla is buying Solar City. Solar City is a utility, so after the merger Tesla will be able to charge by the KWh in every US state.


The US average rate for electricity is 12.73 cents/KWh, but they can vary dramatically ranging from a low average of 8.96 cents/KWh in Louisiana to a high of 27.50 cents/KWh. Because of the wide range of costs, Tesla may charge a different rate in different states, but for simplicity of calculation, let’s assume they will charge 16 cents/KWh for electricity at the superchargers. This will allow Tesla to make a little on charging which will cover the expense of maintenance and at least partially pay for expansion.


So if supercharging for the lifetime of the car is $2000 and Tesla charges 16 cents/KWh, where is the break-even point?


$2000 / 0.16 = 12,500 KWh (12.5 MWh)


That’s a lot of energy, but how much driving is this?


If you get about 3 miles/KWh that’s 37,500 miles on superchargers. If you only get 2 miles/KWH that’s 25,000 miles. If you draw on average 50 KWh per supercharger session, that’s 250 supercharger visits.


Tesla could end up charging rates on the order of other services who sometimes have to charge by the minute in states that don’t allow non-utilites to charge by the KWh. Various companies charge different rates, but I’ve found rates ranging from 39 cents/KWh to 79 cents/KWh.


For the sake of calculations, assume Tesla charges 40 cents/KWh, the numbers change a fair bit:


$2000 / 0.40 = 5000 KWh (5 MWh)

At 3 miles/KWh that’s 15,000 miles

At 2 miles/KWh that’s 10,000 miles

At 50 KWh/charge, that’s 100 supercharger visits


Another issue that comes up is if the charge per KWh is too high, supercharging will begin to not be cost competitive with gasoline on a long trip, which is a competitive disadvantage for a long range EV. I made the charts below comparing charging costs for different EVs compared to the cost of gasoline cars. The bottom line is the cost/mile.


At just 30 cents/KWh, the cost per mile for a Model S 90D essentially costs the same for a 25 mpg car at $2.50/gallon for gasoline. For this reason, I believe Tesla will keep the cost/KWh for supercharging low.


The ICE vs EVs


How Internal Combustion Engine (ICE) cars work is embedded in the DNA of our culture so deeply, we don’t even think about it. Driving electric is a different technology that works differently than ICEs. In most ways this technology is superior, but it has to be managed a bit differently than ICEs.


100 years ago there were three competing technologies for cars, ICE, battery electrics, and steam power. Each had their advantages and disadvantages.


Steam power took the longest to get going and steam powered cars were very heavy with the largest engines of any type of car. Once they got up a head of steam they were very fast though. Steam engines using the Stirling Cycle can also be more efficient than an ICE ever could be. High temperature steam from a leaking boiler is often invisible to the human eye and can be deadly if someone gets too close. The disadvantages killed off steam power for cars and that’s probably a good thing.


Electric cars had a niche, but the only batteries available were lead acid batteries, the same tech as used for the 12V battery in ICE cars today. Lead acid batteries don’t have great energy density compared to other rechargeable batteries available today, but they do well in ICE vehicles for starting the car. Early electric vehicles had very poor range and needed even longer charge times than today. 100 years ago electricity wasn’t nearly as common at it is today, so there was even more risk of getting stranded than today.


ICE cars were not very popular early on. They required a crank to start and broken arms from engines starting and then spinning the crank were common. They cost just as much as other types of cars and were dirty and high maintenance compared to electric cars.


Two things solidified ICE cars as the standard, first Henry Ford started making inexpensive ICE cars that more people could afford and soon after the electric starter came along. This sent the world down the ICE car route for most of the next century. Until the 1970s, cheap oil prices kept anyone from considering anything but petroleum fueled cars.


When oil prices went up in the 1970s, interest in electric cars started a new renaissance, but battery technology had advanced very little in the previous 60 years. The only other rechargeable batteries available were nickel cadmium (NiCd), which are not suitable for electric cars because they have a memory. Unless they are fully discharged before recharging, they will tend to establish the level of charge when charging started as the new “empty”.


The nature of the propulsion source is different between ICEs and electric motors. Electric motors convert a lot more of the energy in into motion than ICEs can. All thermal engines, whether steam powered or an internal combustion engine have a theoretical maximum efficiency. A steam engine can get as high as 60% efficiency, a gasoline ICE can theoretically get up to 56-61% efficiency, and diesel engines can achieve a bit higher efficiency than an ICE. However, in the real world, there are many reasons those efficiencies are never seen. All moving parts have friction, combustion is rarely 100% complete, real world fluids and gases don’t behave like theoretical ideal liquids and gases, a narrow range of RPM where the engine is most efficient, and energy is also lost due to aerodynamic forces.


In cars, gasoline engines are rarely better than 20% efficient and diesels can get a high as 40% efficient. In 2014 Toyota claimed to be working on an engine to be used in hybrids that will be capable of 37% efficiency, the best ever seen in a gasoline engine.


A standard gasoline ICE engine runs what is known as the Otto Cycle (named after the inventor), and a different type of ICE is used in hybrids called the Atkinson Cycle. The Atkinson Cycle is a bit more efficient, but it produces a lower power density. This is mostly only used in hybrids because most of the time the engine is charging the battery and the electric motor and gasoline motor together can propel the car, so the lower power density is not as critical. A car powered only with an Atkinson Cycle engine would likely have very poor acceleration. An Atkinson Cycle hybrid can get up to 30% efficiency in the real world, which is better than a pure gasoline ICE alone. Part of this efficiency gain comes from running the ICE at the most efficient RPM and load settings most of the time (when charging the battery).


Where does the lost energy go? Mostly heat, though some goes into making noise too. Most of the heat generated is released through the exhaust gases, but most ICE car engines have a liquid cooling system to absorb the heat that doesn’t get released in the gases. Without liquid cooling of some kind, heat will build up in some areas of the engine and cause it to fail. Some older ICE car engines were air cooled, but they don’t meet modern air quality standards, and some cars have the ability to run without any coolant in the engine (though it’s not recommended), this is mostly done by using extra engine oil to replace the job done by the water based coolant.


ICE engines do have one added benefit, once the engine is warm, channeling some of that heat into the car’s cabin can heat the car as a “freebie”. Running the heater actually helps the engine cooling a little bit.


ICE engines produce no torque at zero RPM (revolutions per minute), and will rip themselves apart if the RPM gets too high. If an ICE gets to too low an RPM, it will stall. There is also a sweet spot RPM where the engine is producing power most efficiently. Because of these limitations, any ICE capable of more than a few miles per hour top speed requires a transmission.


If an ICE car didn’t have a transmission, it would either stall when you tried to start moving, or it would be ripping itself apart by the time you got to about 10 mph. A transmission is a necessary evil with ICE cars. They do reduce efficiency some from extra sources of friction, as well as add weight and complexity to the car design.


In recent years car makers have tried to squeeze as much efficiency as they can out of cars by making transmissions with more gears, or CVT transmission which have an infinite number of gears. All this is to try and keep the engine in the efficiency sweet spot as much as possible.


Electric motors are a completely different technology with completely different characteristics. Real world efficiency for an electric vehicle from the wall socket to the wheels is around 75%. That is 75% of the power that came out of the wall socket is used propelling the car. The ideal efficiency of the motor alone is closer to 95%.


Electric motors produce some heat, but far, far less than an ICE. The drive train can be much simpler which reduces places where friction eats up energy. Tesla’s design places the motor between the wheels where it can drive the wheels directly.


Electric cars don’t need transmissions. A transmission may help at extremes, like over 80 mph, but it’s generally not needed. Electric motors produce lots of torque even at 0 RPM, so when the car is sitting at a traffic light or stop sign, the motors can be turned completely off. Electric motors can also work just fine over a wide range of RPM and once the motor has overcome the static friction when stationary or moving at low speed, it’s extremely efficient.


There are two types of friction, static and dynamic. If you’ve ever pushed a heavy object across a floor without wheels, you probably found it was more difficult to get the object moving than it was to keep it moving than it was to get it moving in the first place. That’s because the coefficient of static friction is always higher than the coefficient of dynamic friction, so once something is moving, it’s easier to keep it moving.


Cars have the same issues with friction. When starting or just moving at low speed, static friction is coming into play. Once everything is moving, dynamic friction is in play and the losses from friction alone go down. Aerodynamic drag is a factor that comes into play as cars get going faster. The faster the car goes, the worse the aerodynamic drag gets and it increases exponentially with speed not linearly.


Friction and aerodynamic drag affect all vehicles. ICE cars usually have a bit more friction at low speed because there are more moving parts, but in general, both of these factors affect all cars equally.


Most ICEs are geared to get the best motor efficiency around 45-60 mph. Aerodynamic drag is kicking in at those speeds, but because the engine is geared for best efficiency there, the two cancel out to some degree allowing the car to be efficient at that speed. Because of all the other inefficiencies of ICEs, aerodynamic drag is not noticed as much as with an EV.


Think of it like tracking your grade in a class in school. If you’re getting an average of 30% in the class and you get a 20% on a midterm, it’s not going to lower your grade all that much. But if you’re averaging a 90% and get a 50% on a midterm, it’s going to drop you down a grade or two. With an EV, there are so few sources of loss that the few sources of loss are really felt whereas with an ICE, another source of loss is just another one among many and the contribution from that loss source isn’t noticed as much. It’s there, it just gets lost in the noise to some degree.


The charts below show the efficiency curves for some ICE vehicles and another chart for the Model S and the Tesla Roadster. The curves for the ICEs vary a lot more, but they tend to plateau around 45 mph, then aerodynamic drag becomes a bigger and bigger factor over about 55-60 mph and the curves roll off. The Tesla curves peak around 25 mph and roll off from there.



Despite the dropping efficiency at speed, electric vehicles are significantly more efficient than ICE cars. In the graph below you can see the different EPA energy expended per mile for ICE cars, hybrids, and EVs. ICE get the biggest increase in efficiency going from city driving to highway. Hybrids see a little increase, and EVs see only a slight improvement For an EV the slight improvement comes from driving at a constant speed instead of the stop and go driving in city driving.



Additionally you can see on the graph that EVs overall get much better energy economy. ICE engines aren’t normally measured in Wh/Mi, they are normally measured in Miles per Gallon (mpg) or Kilometer per Gallon depending on where in the world you are. Gasoline has around 33 KWh/gallon, so you can convert back and forth between the two. The EPA uses 33.7 KWh/Gal for their conversion from Miles/Wh to MPGe. If you look at EVs at the EPA’s website: you will see all EVs have pretty high MPGe values. All Teslas are near or over 100 MPGe.


So why are EVs so short ranged? It’s because the energy density of batteries is much lower than gasoline. The volume of the battery pack on the Tesla Model S and X is about 96 gallons, but the energy density of the highest capacity Li-ion batteries is only about 1 KWh/gallon of volume, 1/33 the energy density of gasoline.


Gasoline is an amazingly dense energy source, which does mean it will be around for some years to come in some applications. However for personal cars, EVs are superior to ICE passenger cars in many ways:


  • EVs have much better acceleration. This will become even more apparent as the CAFÉ standards get stricter and stricter. ICE cars will have to get wimpier to meet new MPG standards.
  • Interior volume – By putting the batteries under the floor, Tesla was able to give much more interior space than other cars of similar size and shape.
  • Ability to refuel at home – This is still a sticking point for people who don’t have a way to charge an electric car at home such as people who don’t have a garage or live in an apartment, but with infrastructure improvements to charge on the street or in the lot of apartment complexes, this can be solved. It’s a big advantage to be able to always leave home with a “full tank” of gas.
  • Cheaper to fuel – Electricity in the US costs on average 12 cents a KWh. Coupled with the better efficiency of EVs, the cost per mile to drive an EV is significantly cheaper than an ICE.
  • Quieter – EVs only produce much less noise than ICEs.
  • Maintenance – EVs have fewer moving parts and much simpler drive trains than ICEs which can lead to low maintenance costs over the long run and much longer lasting cars.


The only downsides have to do with the speed of charging, driving very fast, and the current cost. Batteries take longer to charge than filling a tank with gas. And though the price of batteries is always coming down, battery powered cars still cost more than gas powered cars. As the cost of batteries comes down, this difference will be erased.


The time it takes to charge batteries is not normally a factor. As long as a car has enough battery capacity to allow someone to do whatever driving they need to do in a day, the down time to charge the car overnight is not usually an issue. The only time it becomes a factor is when you need to charge during a road trip. Tesla currently is the only car company with an effective long distance travel network. The superchargers go a long ways towards mitigating the hassle of charging on the road.


The last issue with driving fast can be mitigated by driving slower, though it is embarrassing by many to drive along in the slow lane with everyone screaming past. As EV battery packs get larger capacity this will become less of a factor over time.


It takes rethinking the way you drive a bit when transitioning to an EV, but in the long run it’s worth it. I’ve only driven an electric for three months at this point, but I definitely think it’s vastly superior to driving my old ICE. I’ve had a hard time parting with my old ICE for sentimental reasons, but I don’t want to drive it anymore.


Reference links:

Battery Tech Part 2

This is a continuation of my rambling about battery tech, this time focused more on Tesla’s batteries.


Tesla currently (as of 2016) uses a Li-ion cell called the 18650. The number just refers to the dimensions of the cell. It’s 18mm in diameter (a little less than 3/4 inch), and 65 mm long (about 2.5 inches). This is a standard cell size used in laptop batteries and was available with no retooling by a battery maker when Tesla set out to make electric cars. They look like a AA alkaline battery, but they are a bit bigger.


The Model S and X have had two pack types from the beginning, a small pack and a large pack. Tesla arranges the packs in modules. The small pack has 14 modules and the large pack has 16. Each pack is arranged to have 22.2 V per module by stringing 6 cells together in series (voltages add up), then arrange the pack into sets of these 6 cell packets in parallel (same voltage for each packet, but the current capacities add up). In the 1st generation, the small pack had 64 X 6 cell pieces in parallel and the large pack had 74. From the math, it appears the second generation pack has the same arrangement, but the 3rd generation pack has more cells per module (Elon Musk confirmed this for the 100 KWh pack at the announcement). In the smaller pack the unused cell positions are filled with dummy cells.


So far the battery cells have been through two generations, and the packs have been through three capacity versions (with many sub-versions of each). Both versions of the battery cells have been the 18650 format. The first version had all graphite anode and the second generation added a little silicon which boosted the current capacity of each cell.


We know a lot about the first generation cells and packs because Tesla both talked about them quite a bit when the Model S was launched and a user named Jason Hughes took apart some 1st gen packs for a home solar storage project and documented everything he found. What’s in the 2nd and 3rd generation packs is still open to some debate. Most people assume the 2nd generation packs use the same number of cells as the 1st generation pack, but I can’t find any data where anyone has opened up a 70 or 90 pack to see.


The first 3rd generation packs were the new 75 KWh packs introduced in early 2016 and I originally thought they might be a 3rd generation battery cell, but at the P100D launch, Elon Musk said the 75 and 100 packs use the same type of cell as the 2nd generation pack, it just uses more of them.


The exact numbers on the cells and hence the exact capacity of the battery packs is open to a bit of debate. Panasonic makes several different versions of the 18650 cell format, each with a little different chemistry. As far as I can figure out, the cells used in the 1st generation pack were NCR18650Bs.


The cells inside the Tesla packs are wholesale cells which have some Data Matrix codes on them (2D bar codes), but don’t have any other markings. I saw a post on Reddit where someone compared the codes stamped on the batteries from the disassembled 1st generation pack and the codes on a commercially obtained NCR18650B and the Data Matrix codes are the same. Using the values for the NCR18650B does come up with pack sizes that work for the 1st generation of Model S.


The same post speculated that the batteries in the 70 and 90 KWh packs were the NCR18650G which was popularized by Panasonic for a while, but then pretty much disappeared. Using the numbers for the NCR18650G does come up with plausible real pack sizes for the 70 KWh and 90 KWh 2nd generation packs.


For the 3rd generation packs, we know there are more cells in there because Elon Musk has said so, but nobody outside of Tesla knows how many more. Because the basic building black of Tesla packs is the 6 cell packet, we can assume the 3rd generation packs have some multiple of 6 cells, but beyond that it’s all educated guesswork. In the table below I generated numbers based on two different possible sets of cells. One set of numbers comes up with actual capacities that are a little under the advertised levels and one set goes a little over. I would guess the actual cell count is probably a little under rather than a little over.


The exact voltage and current capacities to use when calculating the pack capacity is also a bit of speculation. Panasonic does not publish the data sheets for the NCR 18650B or G on their website (at least I couldn’t find it, though I did find sheets for several other 18650 versions). I did find a supplier in China that had spec sheets on a wide variety of batteries from many makers.


The NCA type chemistry these cells use produce a max voltage of 4.2V, but any calculation is going to use the nominal voltage which is rated at 3.6 or 3.7V depending on the source. I used 3.6V which is the more often used number. You want to use the nominal voltage because a battery Is only going to have it’s max voltage when it is at 100% charge and it will drop to the nominal voltage fairly quickly as it discharges.


The reported current capacities vary all over the map and they can change depending on conditions. Temperature can have an effect on all batteries. In places that have very cold winters, cars don’t want to start on cold mornings because the lead acid batteries can’t provide enough power to fire the starter. In the oil fields in the north of Alaska, nobody ever shuts off a car in the winter. If they did, there is a good chance they won’t get it started again.


Li-ion batteries are also affected by temperature and you get different ratings for the cells at different test temperatures. Using the ratings for the cells measured at 25C (about 77 F) gets values that fall into the ballpark of Tesla’s advertised battery pack capacities. My numbers are in the table below.


The cell format that will be built at the GigaFactory is a larger size than the 18650 used now. The 21700 is 21 mm wide and 70mm long and talk is that the larger cell will allow for an improvement in energy density from the size change alone, and there are rumors Tesla will also be using a chemistry that improves energy density even more.


The speculation about how much energy density gain is all over the map. I have seen estimates from 10% up to 33%. I think it’s likely going to be 20-25%. A 10% increase would make the small pack 82.5 KWh and the large pack 110 KWh. A 20% increase would mean a small pack at 90 KWh and the large pack up to 120 KWh.


We’ll probably see the new packs in Model Ss and Xs in early 2017.


Even if the initial GigaFactory built packs are only a modest increase, I expect Tesla will come out with something for the S and X that is much larger by the time of the Model 3 release. They need to do something to keep the Model S distinguished from the Model 3 and the only feature the S has over the Model 3 right now is the hatch. If the Model S and X had much greater battery capacities than you can get on the Model 3, that would leave a market niche for the more expensive cars.


Battery Tech Part 1

How Batteries work


If you are more interested in specifics of Tesla batteries, that’s mostly in part 2.


When dealing with electricity, you have some terms that need defining:

Voltage – This is a potential difference between two points. You can think of it like a height in the physical world like a hill or a mountain.

Current – This is what flows from point to point. In the early days of electricity research nobody knew what was flowing, but it was later discovered to be electrons. You can think of this like a river of water in the real world.

Resistance – This resists the flow of electrons and generates heat. You can think of this as similar to friction in the physical world.

Power – This is calculated by multiplying the Voltage and the Current and you can think of it like the power of the river caused by the amount of water combined with the height of the drop. This is measured in Watts.

Energy – This is Power times time and is measured in Watt-Hours.


When you start looking into batteries you see a lot with Voltages, Current, Power and Energy. The biggest confusion is between Power and Energy in part because both have something to do with Watts, but they are different measures. When you are looking at what is going on instantaneously, you are concerned about Power (Watts), but when you are looking at the total storage capacity, you are concerned with Energy in Watt-Hours.


Power and Energy have less relationship than you might expect. Some batteries can hold a lot of Energy (Wh), but they can’t produce much current at any point in time. Alkaline batteries are this way. They can last a very long time in something like a clock or other device that uses low current, but they do not work in a high current application.


You can also have batteries with low capacity (Energy storage density), but they can deliver very high currents for short periods.


Batteries fall into two broad categories: Primary Batteries and Rechargeable. Primary batteries are use once and discard and rechargeable can be put into some kind of recharging system and recharged. We’re all familiar with alkaline batteries that are the world’s most common Primary Batteries. Almost all cars built in the last 100 years have had a lead acid rechargeable battery to start the car and keep it’s electrical systems alive when the car is turned off.


We have other rechargeable chemistries like Nickel Metal Hydride and Nickel Cadmium, but Li-ion is the type of rechargeable battery that is getting the most attention these days.


Inside a battery, you need an anode and a cathode. Each is made of a different material and the type of material used by the anode and cathode determines the voltage. Different materials have different electrode potentials, that is they contribute a certain voltage from neutral when combined with another material. Lithium has an electrode potential of -3.04V, which is one of the most negative potential materials in existence.


When designing a battery chemistry a lot of factors beyond just the chemisty come into play. There are chemistries that may work to extent in the lab, but they are impractical to put into production. At this time there are no other combinations using calcium or stronium that nake a better battery and/or more cost effective than lithium.


To get a voltage for a battery, you need a second material to make the other terminal. For example nickel-cadmium batteries use nickel hydroxide for the cathode (positive terminal) and cadmium for the anode (negative terminal). Some kind of electrolyte is in between the two to allow atoms to easily transfer from one to the other.


When a battery is in a circuit, electrons are released from the negative terminal (anode) that makes the anode more positively charged and the cathode at the other end of the circuit gains electrons and becomes more negatively charged. Inside the battery, positively charged atoms move from the anode to the cathode. When the cathode become saturated with positive ions, the battery is discharged.



For a non-rechargable battery this process is only one way. There is no way to get the atoms separated again unless you recycle the battery and reprocess the materials making up the battery. With a rechargeable battery, applying an outside power source to the battery reverses this process. Some rechargeable batteries can be discharged and recharged many times, while others are more fragile and break down a bit each time the battery is recharged. Some rechargeable batteries have other limitations on charging. For example NiCd batteries tend to develop a “memory” if they aren’t fully discharged before recharging and will begin to think it’s “empty” when it reaches that level of charge instead of empty. Other batteries like Li-ion don’t like to be fully discharged and don’t like being charged to 100% and then allowed to sit with no current draw.


Other characteristics of the battery determine how much energy the battery stores. For many batteries, the main factor is the physical dimensions of the anode and cathode. Li-ion batteries have other factors that can affect the energy density stored in the battery.


Lithium-ion battery tech might be the most complex thing in electronics today. The public (and a lot of media) lumps all Li-ion batteries as one thing, but in reality there are many types of Li-ion batteries and they can be quite different. Li-ion is an umbrella term for a whole family of batteries, each with a different application.


Why is Li-ion getting all this attention? For rechargeable technologies, Li-ion can have the best energy density per weight of any battery type. Lead acid batteries are about 40 Wh/kg. Their energy density per volume (cu ft) is not bad, but because they use a fair amount of lead, the energy density by weight it poor. NiCd batteries are about 60 Wh/Kg and NiMH (nickel-metal hydride) batteries are about 90 Wh/kg.


NiMH is a special case in the battery world. It was a popular chemistry for cars in the 90s, it can produce high currents and has a high safety level, plus it’s a lot cheaper than li-ion batteries. However it is limited in car use because of who owns the patent. The battery chemistry was invented in the 1980s, and GM bought the patent in 1994 and then later sold it to Texaco just before Texaco was bought by Chevron. Among the patents Chevron bought was a patent originally obtained by GM in the 1990s for a pure EV battery pack using NiMH batteries.


Chevron then banned the use of NiMH batteries in pure EVs, only allowing them for hybrids. (Bring that up the next time someone says the oil companies aren’t trying to prevent alternate fuel vehicles!) Chevron’s patent on the cell technology itself has expired, but the patent on the car battery packs is valid until late 2020.


Because of Chevron’s blocking, EV carmakers looked to Li-ion batteries, which are much more expensive, but have energy densities ranging from 80 Wh/Kg to over 250 Wh/Kg. Ultimately Chevron has forced the market to improve Li-ion batteries and Tesla to build the Gigafactory to lower the cost of cells as much as possible.


As mentioned above, Li-ion batteries are a complex and unusual technology. Lithium has some inherent advantages with a high electrode potential as well as being one of the lightest materials in the universe. Lithium is the lightest solid at room temperature of all the elements. However pure lithium has the nasty attribute of being very flammable when exposed to oxygen and water, even water vapor in the air.


Research turned to using lithium compounds usually with a carbon graphite anode and some kind of compound for the cathode. When charged, ions of lithium are loosely attached to the carbon graphite cathode and as the battery discharges, the lithium ions cross through the electrolyte to bond with the compound making up the anode.


Lots and lots of research has gone into the materials used in these batteries. The only thing that they have in common is there is a lithium ion that crosses the electrolyte when the battery releases energy. Most Li-ion batteries made up until recently used all graphite for the cathode, but some recent advancements have added small amounts of silicon to the graphite which increases the number of lithium ions the cathode can hold. This doesn’t change the voltage (which is determined by the electrode potential difference between the lithium ions and the material the cathode is made from), but it does increase the energy density.


As with many materials with batteries, silicon has drawbacks. Graphite doesn’t change it’s physical dimensions much when it absorbs a lithium atom, silicon expands when it absorbs lithium which stresses the material and can lead to the battery falling apart a little bit on every discharge cycle and shortens the life of the battery. Recent advancements have allowed small amounts of silicon to be added without risk of destroying the battery. Tesla started using batteries with a little silicon in mid-2015 when they introduced their second generation of batteries.


For the anodes, the picture is even more complex. There are many known compounds used for different types of Li-ion batteries. Each has different characteristics. Each Li-ion chemistry is a trade-off between the specific energy (amount of energy it can hold), the specific power (the amount of current it can discharge at once), safety (some chemistries are more prone to catching fire than others), how it performs at different temperatures, its life span, and the cost.


Tesla uses a chemistry that uses a nickel-cobalt-aluminum oxide which has the best specific energy, but isn’t the best for the number of charge and discharge cycles, and it is more prone to fire than many other chemistries. Tesla has compensated for these limitations with very careful software management, good battery cooling, as well as good battery protection.


There have been a few Tesla fires, but they are very rare compared to internal combustion engine car fires. A study of US car fires between 2003 and 2007 found an average of 287,000 car fires in the US per year, which comes out to around 32-33 per hour. A total of 5 Teslas have caught fire world-wide in 4 years of production (3 in 2013 and 2 in 2016).

If you want to learn more about the nitty gritty details with batteries, Cadex Electronics maintains a wonderful site with more than most people would ever want to know about them: