Towards Affordable, Ubiquitous Ultra-Fast Electric Vehicle (EV) Charging. Part I: Need & Battery Issues

[Disclaimer: The views expressed here are personal and meant for information/education purposes only and not representative of his employer directly or indirectly. Any mention of company names or images are for illustrative examples only. ]

This is part of a series of informational articles on e-mobility: Electric Vehicles & Ride Sharing Economics, Commercial Fleet EVs - Stealth Revolution, Towards Ubiquitous, Affordable, Ultrafast Charging, Electric & Autonomous: Synergies for a TaaS future, Towards Scalable EV Charging Infrastructure (Think Outside the Gas Station Box), Towards Scalable EV Charging: Hidden Costs of Level 2 Charging at Scale.

The pictures above show an "exception" from an "exceptional" company: two sites in California with large number of Tesla superchargers. The opportunity now is to drive technology, business models and economics to make ultra-fast EV charging "affordable" and "ubiquitous": imagine every medium-large parking lot (malls, hotels, restaurants), logistics centres, workplaces sporting not just slow charging, but ultra-fast charging, ... not just in one or two isolated stalls per location, but in many many parking spots everywhere at affordable costs, ... and charged from zero-emission sources.

This article makes a case that we can get there relatively quickly (in the next 5-10 years in developed markets) with a few iterations of technology, solution offering & business model innovation. And we will: creating a huge complementary driver of the clean-energy transition of transportation. In this first part, we articulate the need (to support the medium/long term EV growth), and technical aspects of battery pack size, charging rate, temperature, voltage/current profiles etc, and its implications on E-Mobility infrastructure business models. [Note the case made in this article makes economic sense in markets where electric vehicles (EVs) with reasonably sized batteries (> 30kWh) are taking off.]

Electric Vehicle Growth

Electric vehicles are growing rapidly, adding more than 1 million every 6 months globally from a small base, and towards a future of a fleet of tens of millions of EVs. Tesla has recently achieved 1000 cars / day of Model 3 production alone which would ship over 360K Model 3s in a year at that rate, ramping towards the 0.5-1 million/year mark and beyond for one model alone of global production.

A number of automotive OEMs are launching electric vehicles (see graphic, credit to Bloomberg New Energy Finance, BNEF). Observe that it is covering a variety of vehicle types and progressively increasing in range (100-300 miles+ per charge, which corresponds to larger battery pack sizes of 50-100 kWh in cars). We should expect beyond 2020, the trend towards larger packs to continue till a threshold of around 600 miles/charge is hit, equivalent to today's gasoline tank capacities (i.e. 200 kWh of pack, at 3 miles/kWh efficiency).

Taking just a couple of examples: Volkwagen is making a $50 billion bet by 2023 to make a large-scale transition. launches by its group companies, Audi (e-Tron), Porsche (Taycan). Porsche Taycan has a 800 V battery pack that can charge at rates up to 350 kW. Hyundai's Kona Electric has won some nice reviews (eg: this Fully Charged show) for its affordable pricing, energy efficiency (close to 5 miles/kWh or 8.2 km/kWh in practice, more than Model 3's impressive 4.1-4.2 miles/kWh), and fast charging capability. A recent bakeoff organized by Bjorn Nyland of efficiencies shows the Model 3 doing as well or slightly better than Hyundai Ioniq depending upon the configuration (the Hyundai beat most of the cars on efficiency though). Rivian, a new auto-maker revealed impressive specs for their electric pickup truck and SUV prototypes: eg: 180 kWh/400 mile range for R1T and power specs of 522-562 kW (or 700-750 hp).

Simple economics will drive a transition of ride sharing (eg: Uber) & taxi fleets electric vehicles/EVs (see prior article). I personally believe that once transportation network companies (TNCs) start adopting EVs, and capex per EV drops and the capex differential is paid off (probably 2-3 years away), it will drive down the average price-per-mile charged to users which will set off a virtuous cycle of more EV rides in Uber/Lyft/e-Taxis etc; the limit would be set by the the capex of the vehicle and number of electric vehicles on the road vs ICE engines. Commercial electric vehicle fleets are also seeing a stealth growth, as I have covered in my last article. With the growth in electric vehicles, while it is great that they can be charged at home, it is important to grow the network of affordable, ultra fast charging at scale similar to the large Tesla supercharger parking lots depicted at the beginning of this article.

The first question is "What is fast / ultra-fast" charging? The answer is that "it depends" upon the size of the battery pack (assuming a well designed LiIon chemistry), the aggregate voltage of the battery pack, and the nature/quality of active cooling of the pack during the sustained charging period. C-rate is a relative measure of charge / discharge compared to battery size in rated Amp-hours. Specifically for a 120 Ah battery pack, a 120 A current is a 1C rate (charge or discharge). Now battery packs are made out of modules (24V, 48V etc) stacked up to a pack level voltage of 300Vdc-800Vdc. Keeping voltage constant, 1C charge rate for a 30 kWh battery pack is 30 kW; 2C is 60 kW. For 100 kWh pack, it is 100 kW (1C) and 200 kW (2C). Different series/combinations of cells can yield higher voltage packs with roughly same kWh capacity. For example, the BMW i3 packs by Lion BV are shown below with 96s (series) 84p (parallel) for 400Vdc configuration; or 192s44p 800V configuration. Higher voltage can take in higher power (in kW for the same C-rate), but needs to be handled more carefully for safety; and any malfunction of a cell will affect a longer series string.

Typically LiIon batteries don't like being charged at sustained rates 2C and above (explained later in this article due to a combination of thermal issues, chemical issues like concentration polarization, metal plating, electrolyte decomposition etc). The key is to keep the current (I) relative to Amp-hour rating at an acceptable C ratio (<2). The way to deliver higher power is to reconfigure the pack and wiring to make it a high voltage system so that the current is within the same C-rate limit, but at a higher voltage (as done in Porsche Taycan, Audi e-tron GT etc). Higher voltage has risks of over-voltage at a cell level which also need to be carefully managed (described later).

And this phase of fast charging is only between 20-75% of state of charge (SOC) of the battery, roughly corresponding to the constant current (CC) charging phase; the power delivered drops between 75-100% SOC. The specifics of what is allowed depends upon the OEM, vehicle, temperature, age etc and is determined dynamically by the battery management system (BMS) - the charger respects the BMS-defined constraint and delivers the necessary power dynamically.

Ultra-fast charging with a heterogeneous fleet of cars, LCVs, trucks/buses is therefore at least 50kW+ ("fast"), but typically 100 kW-350kW (cars, LCVs, MCVs, buses), and beyond for special cases (semi-trucks, long distance buses etc). Note that the numbers mentioned above are per-charge point & peak rates. If you have 10 chargepoints (with at least 100 kW), the aggregate rate can be more than 1 MW. A parking lot with 200 parking spots, offering 150 kW each spot is 30 MW aggregate power requirement at peak, but could be much lower on "average".

Ultra-fast charging is most important for the commercial & ride-sharing sector for the productivity of their EV fleets; but the larger scale and visible, convenient availability everywhere will drive up the convenience factor and spur further growth of EVs (a win-win dynamic between demand for EVs and charging infrastructure). Note that the larger battery packs (50 kWh+, and above 150 kWh) mean that charging at home with 11 kW L2 charging for larger pack sizes will be quite long indeed (eg: 185 kW for Rivian / 11 kW will imply 16+ hour charge times, not OK for commercial productivity)!

Why affordable, ubiquitous, ultra-fast charging?

If we can make the economics and deployment to work out, the consumer / end-user answer is simple: convenience & top-up ability anywhere. Initially the demands for ultra-fast charging will come from commercial vehicle fleets (logistics, airports, bus depots etc), ride sharing/taxis, and the rapid 50%+ annual growth of consumer vehicles (both in terms of number and as an insurance to reduce/erase the notion of range anxiety from consumer minds). As Transport Evolved observes what happened in the holiday 2018 weekend and indicated, the sheer number of EVs, and the relatively higher latency to charge (vs gas fueling) will also lead to more parking spots having charging ability to avoid queues building up at charging points. Retail (malls, hotels, restaurants, fast food etc) will also use fast EV charging "top up" and bundle it with their loyalty programs. In several European cities where people park cars overnight on streets (vs garages), it is important to have public top up fast charging or workplace / destination charging in lieu of home charging. Home / workplace / destination charging will continue to be AC/low speed (L1/L2) charging ranging from 7 kW-22kW.

Lets unpack the non-marketing aspects from first principles (see graphic below).

The power drawn by vehicles is a function of mass (M, big/small loads), vehicle speed (V), acceleration (α), "hilliness" or inclines/slopes driven on (θ) and rolling resistance (Crr), vehicle frontal area (A) & drag coefficient (Cd).

Initially the energy and power required for acceleration to a target velocity dominates. Accelerating slowly consumes less power; and faster acceleration consumes more power. Average power during acceleration can range from 60kW to the max HP/kW rating of the car (eg: 290kW for Tesla Model 3 for 394 hp); but this lasts only for a few seconds (< 10 seconds). Note that the battery charges up using "regenerative braking" when you take off your foot off the accelerator or brake. This leads to surge battery charging for short periods of a few seconds each time. In fact for heavy braking, the numbers can be large: the peak braking power for bringing a mid-sized car moving at 96.6 km/h to stop in 5 seconds can be as high as 180-200 kW (3-5C depending upon battery size).

A quick take-away is that EVs safely go through truly a lot of instances of high power, and very short duration cycles of charge/discharge. It is the "sustained" charge or discharge at high power rates (relative to battery pack) that should be handled more carefully.

Once the vehicle achieves a given speed, the power required is a function of the cube of the velocity (i.e .V^3!), adjusted for head/tail winds. Cleantechnica & ABRP show some interesting real world data (reproduced in graphics on left) on Tesla Model S and X at different speeds and temperatures.

Note 10 m/s = 22.37 miles/hour; and 30 m/s is 67 .1 mi/hour (freeway speeds). The steady speed power use is of the order of 10-25 kW (which is 1/6th-1/3rd of peak power use during rapid acceleration). It is also a comfortable C discharge rate. For a median 60 kWh battery, 25 kW is 0.42 C discharge rate.

The converse is true at very high speeds. Driving on the German autobahn at 200km/h+ is 6-8 times more energy expensive as going at 100 km/h on a regular freeway, which in turn is 6-8 times as expensive as a 50 km/h on an inner suburban road. For example Bjorn Nyland observed 680 Wh/km in autobahn in a Jaguar I-PACE. In Bjorn Nyland's tests of Tesla Model 3 (where he drove a model 3 for 500 km/310 miles) during cruise speeds of 90 km/h on a flat road in California (I-5) shows a consumption of 135 Wh/km (or 7.4 km/kWh or > 590 km for the pack size of 80 kWh; or 12.15 kW average rate of use (=135*90/1000 in 1 hour)); head winds etc can raise the consumption (Note EPA test cycle is 160 Wh/km). Anecdotal numbers for the Hyundai Kona and Ioniq indicate even more efficient numbers.

Other aspects such as ambient temperature vs battery and cabin temperature (especially winter in northern latitudes where the battery and the cabin needs to be kept significantly warmer than ambient, or summer where the battery needs to be cooled to stay away from the 45C+ and danger zones of 55C+; and cabin consumes more air conditioning) also determine residual energy left for moving the vehicle/range etc.

As the range, power, temperature graphs illustrate, use cases (eg: urban commercial package delivery) with lower speeds (V is low), reasonable weights such as a van (M low or medium) or shorter total distance will require lower amounts of energy, and smaller battery packs. Conversely, larger weights (eg: 80K pounds full load of semi-trucks or class 7 buses, or SUVs or pickup trucks with trailers), longer ranges (eg: 300-500 miles per charge) will require larger amounts of energy and larger battery packs. The graphs also illustrate the value of aggressive active thermal management of the batteries, and selective/opportunistic insulation from exterior to achieve greater temperature-related efficiencies.

Consider the 80 kWh Tesla Model 3 long range with 394 hp (= 293 kW) which is applied for ~5-10 seconds for short bursts of acceleration, implying a discharge C rate (= 293 / 80) of 3.67C (quite high, but for shorter periods of time). Peak charge C rates of 120 kW is 1.5 C (sustained for tens of minutes).

Mid-sized cars average about 140-170 Wh/km, or 5.9-7 km/kWh; for a median 60 kWh pack, this implies a range of 420 km. Bo (ABRP) has an interesting analysis of Tesla Model 3 variants (vs Model S) in terms of efficiency (Wh / km, range (km per charge), and temperature, speed etc), a few graphics reproduced here. More specific numbers for a bunch of cars are summarized below.

EV Specs of Popular Mid-Sized EVs: 30-80 kWh battery packs, Power rating (100-400 hp), Peak transient discharge C rate (2.5-3.5C), Peak sustained charge rates 1.25-1.75C

The range of peak discharge C rates and peak charging C rates are more tightly distributed than the nameplate kWh of battery (30-80 kWh). Battery pack sizes have been going up recently to the 60 kWh+ mark. We should expect most battery pack sizes to keep rising for the next 5-10 years, as a function of improvements in price (18-20% learning rate), specific energy or energy density (5-10% per year improvement), and chemistry or system level improvements (power vs energy and lifetime tradeoffs and hybridization with ultra-capacitors). Older cars had a peak fast charge rate of 50 kW (300-350V, 125-150A); but newer ones have higher voltage 350-450 V, higher amperage (upto 200A), which allows them to go to higher charge rates of 70-90 kW. Tesla cars and superchargers use a higher voltage and current: 480 V, 250 A for a rate of 120 kW peak.

Battery Sizes, Charging Profiles

For a couple of good video overviews of Li-Ion technology, see here (Prof. Cui, Stanford) and here (Prof. Whitacre, CMU). Here is a quick summary of the relevant aspects that matter from a fast/ultra-fast charging perspective.

Batteries involve electro-chemical reactions where the A + B -> AB is broken up into two half "redox" reactions (one oxidation, losing electrons, and another reduction, gaining electrons) at each of the electrodes; and the electrons go via external circuit. The ions (eg: Li-Ion) travels between the electrodes via the electrolyte.

A Zn-Cu battery with a salt bridge (1.1V) is shown as an illustration (it is a galvanic cell, spontaneous, non-rechargeable).

Reversible reactions implies the possibility of a rechargeable battery. In Lithium-Ion batteries, the Lithium ion (smallest possible after H+ or a proton) moves from the metal-oxide cathode to a graphite/silicon anode during charging; and during discharging moves back. To charge, the external voltage is kept higher (but limited carefully); and during discharge the terminals are connected via and external load. The battery cell voltage (E) depends upon the gibbs free energy, a function of the elements. For a good battery, A and B should have different electron attraction ability (electron negativity vs affinity); A and B should be electronically conducting. The electrolyte should be electrically insulating (i.e. blocks electron transfer), but ionically conducting (allows Li-ions to move).

For rechargeable batteries, the reaction should be reversible. Batteries also store charge in the bulk locations of the crystal structure of the anode and cathode (i.e. in the volume). Thinner film batteries reduce this volume size; but can increase power (i.e. it is called a "power cell" with faster power draw, but lower capacity). At the limit, the charge is electrostatic and only at the plates (i.e. in the area of plates and not volume of the electrodes), called an ultra-capacitor. An ultra capacitor has some similarities to a battery, since it has an electrolyte solution between plates instead of a dielectric, and much lower thickness. Since the capacitance (ability to store charge) is inversely proportional to thickness, this is better than regular capacitors. Similar to a capacitor it stores charge in an electrostatic form in the area; whereas batteries store charge in the bulk volume. Without side reactions, the ultra capacitor could have many more lifetime cycles; and deliver / receive power faster; but energy capacity is significantly lower (since a volume can store more charge than a surface).

A battery pack is a series/parallel arrangement of cells (with associated active thermal cooling and protection) into modules and modules into a pack, controlled by a battery management system (BMS). An example of the Tesla battery pack used in some of their older vehicles is shown in the graphic. The 85 kWh packs had 7104 cells (18650 type) organized into 16 modules, each with 444 cells. Within each module, there are six groups of 74 cells wired in parallel; the six groups are wired in series to create a 24 V module (21.6V nominal). The 100 kWh packs have 8256 cells, organized into 16 modules, each with 516 cells. The Wh rating per cell is around 12 Wh; and Ah rating is around 3.3 Ah. The maximum power that a Tesla battery pack can can use for charging is 4.2 X N X I where N is the number of cells in the pack and I is the maximum current allowed per cell. Tesla allows I = 4A (<2C rate given the Ah rating of 3.3 Ah/cell). For 85/90 kWh packs this works out to 7,104 X 16.8=119.3 kW. For the 100 kWh packs it is 8,256 X 16.8=138.7 kW. For more info on Tesla battery packs: see here (Elektrek), here (Telsa Motors Club), here (Skie.net), here (TwoBitDaVinci), here (Xtar), and here (Cleantechnica).

The Tesla Model 3 uses newer 2170 cells with higher capacity. It has 46.7% higher volume and at least 69% higher capacity (in mAh, Wh), and possibly more depending upon chemistry improvements. Cost of making these cells are expected to be lower by 20-30%. The larger Wh per cell capacity will also allow faster super-charging rates, since it can potentially allow 9-10A per cell (vs 4A limit earlier) - which could indicate supercharging V3 at 240-250 kW (~double of prior 120kW limit), assuming 480Vdc systems: but at higher currents will need 500A cables (thicker or perhaps liquid cooled). It is possible that the voltage is increased to reduce current in V3 (have to verify. please note - this are not official numbers.).

According to Cleantechnica: "The (Model 3) pack consists of four modules — two 25s, two 23s — totaling 96s. The pack energy of 80.5 kWh consists of a total 96s (96 series) cell pack nominally the same voltage as S, 400V. With 96s, that gives a fully charged voltage per cell of about 4.17V, and at about 5A each. The reason for not making them all 24s sections has to do with battery terminals and the need to terminate negative and positive leads at the same end of the pack. The modules are nominally 90V and 86V 233 Ahr. There are 4,416 cells in a 46p, 96s arrangement (p for parallel, s for series). There are then 46 parallel 5.065 Ahr cells, for a total of 233 Ahr, and a nominal (mid-discharge voltage) of 345.5V."

Battery Pack Characteristics and Charging Rates

Interestingly there is a case for "fast" charging (50kW or more) for both EVs with small battery packs (~35 kWh or lower, at 50kW) AND large battery packs (50-100 kWh or higher, at 60-100kW+ depending upon vehicle limits). The reasons are different.

In a smaller pack, if the vehicle usage pattern involves (even occasionally) longer trips, then having a greater density of charging points is important in the range of usage of the vehicles - initially this has been in / around cities, but the networks will grow larger. Note that I am also keeping in mind both consumer and commercial vehicles with ranges of 150 miles or less. Since the user will have to stop more often, there is a time tradeoff, and therefore a preference for faster charging. Note, the C-rate refers to the charge rate in kW divided by the battery pack size (i.e. what rate will charge the pack in 1 hour). Charge rates beyond 1.5C or 2C are not recommended for short/long term pack performance/degradation reasons discussed below); which implies 30 x 1.5 = 45kW (or 50 kW fast charging limit) for most small packs, but this is consistent with a "top up" of 15 kWh (or 50% of the 30 kWh pack) in 20 min in the "safe zone" of between 20-75% of state of charge (SOC).

In a larger battery pack, the required frequency of charging will reduce (modulo home/destination slow charging), and fast charging "top up" behaviors for convenience will increase. An average user could forget to charge their EV a couple of nights, but still make up for it by either an overnight charge or a fast charge on the third day (or once a week in the limit). The depth of discharge (DoD) for the daily commute usage etc will be lower, promoting battery life. The total number of km for a given number of cycles of battery life is also higher: eg: 300 miles x 1000 deep cycles = 300,000 miles till battery drops down to 80% original rating. Ben Sullins at Teslanomics has an excellent ongoing data-driven study on practical battery lifetimes, and actual degradation.

Shallow discharges will raise the total lifetime miles to be higher. This desire for more lifetime miles, re-sale value/lower vehicle depreciation (especially in leasing or expected resale situations) will prompt "upsizing" of batteries in cars and a preference for larger packs by consumers (more broadly Bloomberg called this "Tesla stretch" behavior to spend beyond their usual wants) similar to how they buy larger flash memory sizes in iPhones (prefer 256 GB to a default of 64 GB). Ride-sharing or lending out of cars on sites like Turo for $150/day as reported by InsideEVs (or the upcoming Tesla network) will also drive up the preference for larger battery packs, and "fast charge top up" behaviors where 20-50% of battery i.e. 15-45kWh is filled up during a break for lunch/snack/coffee or a visit to a mall (i.e. in 15-30 min at charge rates of 50-90 kW).

Bjorn Nyland in his videos also suggests a strategy of charging at peak power (i.e. constant current phase) till the charge rates drops off, and then move on to the next super-charger. With more ubiquity in fast charging stations, this will allow drivers to optimize time by switching to "top-up" behaviors of ultra-fast charging between 20-70% only for periods of 5-15 mins. A 5 min "top up" ultrafast charge at 250 kW (eg: while picking up your Starbucks or McDonald's coffee) will add 20+ kWh, and at 350 kW will add ~30 kWh, good for 120-150 km (at 4-5 km/kWh efficiency).

Another reason for larger battery pack sizes is de-rating of pack capacity (and vehicle range) in winter or cold weather (by 10-20%), and lower fast charging rates due to poorer kinetics (Bjorn Nyland): a larger pack gives more comfort and "insurance" / peace of mind for the user -- both a commercial user and for a consumer. Some videos mention that if you do home charging at L1/120V (as in USA) in cold weather, there may still be a net negative drain (source: Rich Rebuilds) since the EV is trying to keep battery warm; and you may need to go to a nearby public charger to get more juice. Finally for greater loads or speeds if common (eg: towing, pick up trucks or power requirements - sports car, big pickup, or autobahn rides), users invest in larger battery packs. These same useage patterns will also drive up demand for "top up" fast charging behaviors. Further more, in large commercial vehicles (eg: bus depots, logistics, school bus fleets), the common charging pattern overnight for larger battery packs (due to heavier vehicles) is likely to be 50kW+ for "slow" charging.

Now lets look at charging profiles. Typically charging involves an initial slow start, a constant current (CC) phase between 20-75% state of charge (SoC) and a constant voltage (CV) phase where the current/power rates taper off after the battery is charged up to 80%+.

Note that in terms of productivity of charging, i.e. kWh added in unit time (i.e. kW rate of charging, or C-rate), the 20-80% is the most productive (i.e. constant current), unless you absolutely need more charge. If you look at popular EVs (eg: Tesla, Jaguar I-PACE, Kia Niro, Hyundai Kona: graph by Bjorn Nyland), it is roughly constant rate (power of charging, with slight increase in rate, this is indicative of battery voltage creeping up, with roughly constant current). Different vehicles have different rates of step down after a threshold based upon battery management policies. Note that Jaguar I-PACE has a 450 Vdc battery pack allowing it to access 80-90 kW charging rate at 200A max; Tesla has a 120 kWdc max with 480 Vdc, 250A maximum current.

Another characterization of LiIon batteries is the Voltage profile, i.e. Voltage vs Ah-capacity (charge capacity) plots. The profiles shown in the graphic indicate that that with faster charge rates, there is a reduction in Ah capacity, and resistive voltage loss (i.e. voltage gap seen in the graph betwen different charging rates).

The Ragone plot puts together the Power (density) vs Energy (density) characteristics of the battery pack under different conditions, shown in the graph.

Specifically the Ragone plot implies that you should desist from charging or discharging the battery pack at high rates (2C+) at high (>45C) or low (<10C) temperatures. Also the energy density (or charge capacity in Ah) drops by <5% for 1C charging and <10% for 2C charging. The difference is lost as heat.

A Panasonic 2070 voltage profile is shown in the graphic. This again shows the battery being discharged from slow to fast speeds, and seeing a resistive voltage drop at higher speeds, and significant rise in temperatures. These batteries were 4Ah rated capacity, so 8A = 2C rate, and 15A = 3.75C (note this 3.75C is also the peak HP rating ratios we saw in the list of cars earlier]

[Recall again that 1C and 2C (charge OR discharge) are relative to the aggregate battery pack size. 1C for a 33 kWh BMW i3 pack is 33 kW, and 1C for a 80 kWh long range Tesla Model 3 battery pack is 80 kW ! In other words, the capacity and performance penalty discussed here occurs at higher absolute power levels for a larger battery pack]

Part of the Ah capacity loss is due to resistive (or IR) loss shown as a drop in voltage in the above graph (and heat generated). But what is not depicted is that fast charging/discharging at consistently high rates also reduces the effective charge (Ah) capacity as well (both on the short and long term)! This part of the Ah capacity loss is due to a battery materials phenomenon called concentration polarization, where the lattice sites (either on the graphite anode or metal-oxide cathode) for Li-Ions to migrate are locally congested. These phenomena are shown in the voltage-profile graphic plotting voltage vs Ah capacity. Note that concentration polarization loss increases super-linearly with higher C-rate of charging (of course, IR or resistive losses also increase linearly with current, I).

Another phenomenon is the resistive loss which lead to the heating up of the electrolyte - if not cooled fast enough (eg: active liquid thermal management in the Tesla, Hyundai packs etc), the electrolyte can go beyond 55C and can disintegrate / oxidize. Note that at very low temperatures (< 10 C), it is better to heat up the battery prior to charging eg: via active thermal methods. Similarly the evacuation of heat at higher temperatures can help batteries like the Hyundai Kona be charged at faster rates even at higher states of charge (SOC). But great care is required !

Over voltage and heat are particularly bad for lithium ion batteries, especially at high state of charge (SOC). Over-voltage can be viewed in two halves - high voltage at the cathode (top of the graph); or low voltage at the anode (bottom of graph, courtesy, Venkat Srinivasan). Look at the top of the graph: it is not just the heat from resistive loss, but also the voltage levels locally at the cathode that can lead to chemical electrolyte oxidation or decomposition. Electrolyte oxidization is exothermic, leading to gassing and runaway/explosive conditions. These phenomena need to be more closely managed as the state of charge (SOC) goes above 75% which is why the charging pattern switches over to constant voltage.

At the anode, during the formation stage of the battery, the electrolyte is partially reduced (solvent reduction process) to create a Solid Electrolyte Interface (SEI), a thin film that has some resistance to Li-ions, but forms a electrical barrier for electrons (which is good for batteries - and is carefully managed in battery mfg). If the applied voltage during cycling (charge/discharge) is too high - at the anode, more of this layer is formed; and in the extreme, Lithium plating happens; which is dendritic in nature and can grow and create a short between anode and cathode. A visual summary of various degradation processes, and safe charging "window" is shown in the two graphics above. The key message to remember is that batteries involve complex chemical processes; and every small detail / interaction matters, and every chemistry is different: therefore respecting the limits and being in the safe zones of charging regimes!

Battery Management Systems (BMS) systems have some form of over-voltage protection, and also monitor the voltage of the battery as a function of the life of the battery. Side reactions like solvent decomposition, oxidation etc lead to irreversible loss of Li-ions leading to capacity reduction, even with calendar time, i.e. batteries lose capacity just with time (even though this is much improved in current chemistries).

Net net: Ultrafast charging is ~2C charge rate relative to LiIon battery pack size and voltage of the battery pack. Specifically, a 100 kWh battery pack at 800Vdc (eg: Porsche Taycan) can take upto 400 kW sustained charge rate safely. A Tesla 100 kWh battery pack at 400 Vdc can safely do up to 200 kW. Higher rates (eg: 475 kW) will require higher charging voltages (eg: 930 Vdc) which has been demonstrated recently by Volkswagen et al. Truck or bus battery packs that are larger can take in a larger aggregate kW rate since their battery packs are bigger.

However, one of the implications of the effective 2C charge rate limit in LiIon technology as of today is that irrespective of the size of the battery pack, even if power electronics off board and the charging cables can deliver the power, there is going to be a limit on the time (15-30 min) for a significant fraction (50%) of battery, from say 25% SOC to 75% SOC to be charged up (modulo voltage of the battery pack - higher voltage would reduce this chargeup time). Simple queuing theory (when service times are longer, you need more servers) implies that ultrafast charging needs to be affordable and widespread to allow "top-up" behaviors in 5-10 min periods, and synergized with other economic activities (eg: retail, coffee breaks, short visits) which has an implication on business models. The key is to drive tech and business models for affordability & scale.

Summary:

Electric vehicles are at an early inflexion point. With the stimulus provided by well-capitalized, relentless entrants like Tesla, the automotive OEMs are investing significantly in the EV transition leading to a number of launches of vehicles over the next couple of years, and increasing battery pack sizes. As discussed in prior articles, the trends of electric vehicle (EV) ride sharing/taxis, and commercial electric vehicle adoption is also at the early stages of takeoff. Autonomous technology (such as the recently launched Waymo One robo-taxi service, or autonomous EV pods/shuttles such as Navya or Local Motors Olli etc) will add a new dimension of growth in the next 18 months. An under-appreciated lever has been the role of ultra-high-speed, affordable, and ubiquitous charging infrastructure.

This article has made a strategic case for such an economic infrastructure even if battery pack sizes are small OR large. The motivations are different - small packs need to be re-charged more often at lower rates; large pack sizes can take in larger charge/power rates, and add more discretion/user flexibility in charging, but will likely lead to "top up" behaviors (i.e. charging 20-50% of the pack in a short periods of time).

While we have made a case for "Why" or "Why not", we have not discussed "How" to achieve the design objectives. Tesla's progress in supercharger deployment & technology offers good hints. The different rates of charging at different levels of SOC (state of charge) shows that the aggregate charging capacity can be shared across vehicles, even if it leads to slower than peak charging.

This "virtualization" of the charger (via a combination of power electronics and software innovation) is an important lever for affordability in the future going forward. There are several other levers; but it will be a journey of multiple iterations of technology innovations in power electronics & software/digital, system integration, deployment, new business models involving a variety of stakeholders (grid operator, asset owner, asset operator, financiers, solution provider), adding up the contribution from multiple value streams, driving adoption by commercial vehicle / ride sharing market segments, and public sector / airport / logistics / retail market segments.

All of this together will spur the future, and achieve the economic/affordability/scale goal which we have painted. It is a journey worth going after, since it involves a virtuous cycle with the growth of electric vehicles in all sizes and use cases. And it will accelerate the de-carbonization of transportation (a key contributor to global climate change mitigation) enabled by a combination of electrification, and the use of renewables and storage. The improvement in economics in turn will allow developing markets like India adopt such technologies sooner.

LinkedIn: Shivkumar Kalyanaraman 

[Disclaimer: The views expressed here are personal and meant for information/education purposes only and not representative of his employer directly or indirectly. Any mention of company names or images are for illustrative examples only. ]

Twitter: @shivkuma_k

If you like this article, please check out other articles: Towards Scalable EV Charging: Hidden Costs of Level 2 Charging at Scale, Towards Scalable EV Charging Infrastructure (Think Outside the Gas Station Box), "Commercial EV Fleets: The Stealth Growth", "EV Taxi Fleets and Ride Sharing: Poised for Huge Growth", "Shared EV Transportation in India", "Understanding the Rs. 3/kWh bids in India in 2017", "Distributed / Rooftop Solar in India: A Gentle Introduction: Part 1","Rooftop Solar in India: Part 2 {Shadowing, Soiling, Diesel Offset}", "Rooftop Solar in India: Part 3: Policy Tools... Net Metering etc..." "Solar Economics 101: Introduction to LCOE and Grid Parity" , "Solar will get cheaper than coal power much faster than you think..", "Understanding Recent Solar Tariffs in India", "How Electric Scooters,... can spur adoption of Distributed Solar in India," "Solar + Ola! = Sola! ... The Coming Energy-Transportation Nexus in India", "UDAY: Quietly Disentangling India's Power Distribution Sector", "Understanding Solar Finance in India: Part 1", "Back to the Future: The Coming Internet of Energy Networks...", "Tesla Model 3: More than Yet-Another-Car: Ushering in the Energy-Transportation Nexus", "Understanding Solar Finance in India: Part 2 (Project Finance)", "Ola! e-Rickshaws: the dawn of electric mobility in India", "Understanding Solar Finance in India: Part 3 (Solar Business Models)" , "Meet Olli: Fusion of Autonomous Electric Transport, Watson IoT and 3D Printing".

All LinkedIn Articles/Posts.