(This article originally appeared in Charged Issue 7 – MAR/APR 2013)
Q&A with Erik Spek: The Chief Engineer at TÜV SÜD Canada on battery abuse testing, improving safety, and developing standards.
When analyzing the current state of lithium-ion technology, it helps to look back at the development of the nickel metal hydride battery (NiMH). Some argued that hybrid automotive applications would require too much power from the small packs, but they’ve performed very well. It turned out that the chemistry is extremely robust. It has a great warranty record with very few exceptions, and all told, it was a lucky choice to start off the whole industry.
In contrast, lithium-ion has seen much less time in development – about 25 years versus 100 years for the nickel electrode (albeit with primitive tools). Lithium-ion technology is far more sensitive than nickel ever was, even at early stages of its development. So, it’s not surprising that while the general public has seen the proliferation of hybrids as a non-event, lithium-ion technology in plug-in vehicles seems more problematic.
Charged caught up with Erik Spek, TÜV SÜD Canada’s Chief Engineer, to discuss the safety challenges of lithium-ion battery packs. Spek is an energy storage technology specialist with broad applications experience, primarily with batteries in vehicles and stationary systems.
Charged: How do the inherent challenges of designing lithium-ion battery packs compare to NiMH?
Erik Spek: With lithium-ion, there is a built-in fire triangle that we’re trying to overcome, and NiMH doesn’t have all three legs – fuel, heat, and oxygen. Lithium-ion has the fuel from the electrolyte (typically ethylene or propylene carbonate fluids that can be quite flammable), heat can be generated by a short circuit or other thermal event, and oxygen can be generated inside the cell, and is obviously present outside as well. You don’t have the fuel component in NiMH until much higher temperatures are achieved (the electrolyte is water based and does not act as a fuel). That’s the basic difference.
Charged: Does the increased energy density of lithium-ion add to its volatility?
ES: The biggest difference in energy density between lithium-ion and NiMH is probably 2-to-1 or 3-to-1 at best at the pack level. So, that doesn’t explain why we have so many more reactions with lithium as opposed to NiMH.
Charged: Does the size of lithium-ion packs in vehicles make design more challenging?
ES: Yes, a large lithium-ion pack is like a kindergarten class where the teacher is trying to keep the children under control. When one cell goes out of bounds the whole pack eventually gets out of control. That’s the problem with the big pack that you don’t have in consumer electronics. It’s not only the chemistry of the cells, and the cooling of the pack, but also the management of all the cells working in conjunction with each other is a major challenge.
Charged: What sort of tests do you perform to evaluate the safety of a battery system?
ES: The operator of the car typically wants four things from the battery: power for acceleration, energy for range, low price, and safety. We do a lot of testing to abuse batteries in the way our customers think they’ll be abused in real life, and the way that several standards describe. The tests are generally classed as mechanical, electrical, and thermal, and lithium-ion batteries are typically more volatile than other chemistries like NiMH in all three areas.
Within each class there are different kinds of abuse that you might expect to happen to a battery or individual cell, such as crushing, penetration, vibration while operating at the highest temperature, shock (like driving over a pothole), short circuits, over-charging, over-discharging, subjecting cells to extreme high temperature then extreme low temperature, and operating at low pressure such as in the cargo hold of a jetliner.
North America is full of pickup trucks, sooner or later electric vehicles are going to be backing up into salt water to launch boats, and the battery has to withstand that immersion. We’re exposing them to salt fog that occurs on coastal areas, and even burning them to simulate what happens when driving over a gasoline fire. All of these tests are to simulate the worst known conditions that can be found in real-life operations.
At the most basic level, we look for reactions that come from either the individual cell becoming too hot due to this abuse, or a cell undergoing an internal short circuit. The short circuit can be caused by latent manufacturing defects within a small cell or breakdown of the insulating material between the electrodes. These conditions can happen when it’s brand new, or much later after it slowly and steadily degrades, and we typically do tests on new product, or slightly used.
There is still a missing link for some of these tests: How real is it? That’s a big challenge for everyone – the OEM, the test house, the cell manufacturers. Writing test standards is a challenge since we already have standards for gasoline vehicles, and the natural inclination is to use these for battery operated vehicles. But are they ‘application relevant’?
Charged: What is the current state of battery abuse testing standards?
ES: Around 1995, some of the people who had been involved in exercises like the Ford Ecostar, with sodium sulfur batteries, and the GM EV1 started to think “What’s the next step? EVs may not be the logical next step because of the weakness in the batteries. So, lets consider hybrids as well.” They got together as the US Advanced Battery Consortium and started developing standards. The standards, at that point, talked about how to measure performance, how to measure life, and how to set up test programs. They also talked about what kind of abuse tests you’d have to do. But they considered it from the level they knew at the time, which wasn’t nearly as much as we know now and we still don’t know enough. Not much had gone wrong yet, simply because there were not a whole lot of vehicles on the road – a couple of fleets totaling a few hundred vehicles.
Those standards survived and they’re being used now after a few updates, but we can see the gaps that are there. We don’t know how relevant they are to real applications until we start digging deeper into them. They all do tests like I described, but those tests may not be so relevant for all cars.
Some of the cars may have the batteries really well protected, and you can’t get at them with a forklift truck or a hammer or screwdriver to poke a hole in them. So that’s on the good side.
On the bad side, for example, we tend to test these products for vibration and pothole shocks the same way we would test a transmission or an engine, which are very robust devices that don’t twist and turn due to the twisting and turning of the car. So, we need to learn how to do that for a battery pack, which in some cases is very long, up to 1.7 meters for example. A pack that long is going to react somewhat to the movement of a car under shock conditions, vibration and body twisting.
That kind of conversation is going on right now. We’re pushing what we know a little farther through efforts like the cooperative research being done by NHTSA and SAE. They are looking to find out what the standards should be, rather than saying “let’s take them where they are, and hope that they’re right.” There is also another cooperative project being conducted by NHTSA and Ford. Both of these are working on trying to establish the science behind what we should be doing for standards.
We are currently doing some testing for SAE that’s looking at how we measure the results of a certain abuse test, and how to make it become more relevant on a scientific level rather than just a straight standard.
Charged: Do you have a sense of how lithium-ion’s resilience to abuse is progressing?
ES: The commercial reality for lithium-ion began in the early 1990s with Sony, but those small cells weren’t destined for EVs at the time. It’s only in the last decade, or maybe decade and a half, that you’ve seen serious use of the technology in automotive applications with somewhat larger cells, and in the last half a decade with significantly larger cells.
In the last three years since we’ve been seriously testing abuse, we’ve seen a pretty dramatic improvement in some particular products that we’ve tested. We don’t see all of the products out there, but we deal with a good number of customers around the world. We’re getting a strong trend in our analysis that says things are definitely getting better. We can see trends in certain tests that show this industry is grappling with the issue of abuse. Cells are getting better at tests like penetration and over-charge. In fact, one manufacturer has shown evidence that it can couple over-charge with a penetration test and show very benign response – this was unheard-of for large cells three years ago.
I suspect that from what we’ve seen, improvements can be attributed to better chemistry and better materials. Both of these combined together have brought it to the point where we can say that it’s definitely improved.
Charged: Based on your previous work with other battery chemistries, and the trends you’re currently observing, could you predict the future development time needed to bring lithium-ion technology up to the level of safety now expected from other automotive batteries like NiMH?
ES: I’ll preface the answer by saying that I don’t know what the word “safe” means. Everybody uses it, but as a testing company, we don’t use it. All we know is that there are certain standards that we test products to and that they either meet or don’t meet. I think the right people to ask what safety means are probably the automotive OEMs. They are closest to knowing what the end user thinks safety means.
At the speed that we’re seeing now, it will probably take at least half a decade before we have a new family of standards that is a workable document and has confidence to be used by people who specify batteries in new electric vehicles. A minimum of five years, considering that it takes at least two years to bring a document to fruitful use.
Is it going fast enough? Is it going too slow? It depends on which camp you stand in. And I’d say that the speed at which this develops is directly related to the number of people who are qualified to work on it. And in turn, it’s a bit limited to whether or not they’re working on the right road map, at least on the abuse side of things.
Charged: Are there a limited number of qualified engineers who can work on the problem?
ES: There is a limitation on the talent pool for two reasons, I think. Number one is the investment community’s appetite for supporting it. When you talk about a talent pool to develop abuse standards, you would like to have a revenue-generating industry to support those people. At this point, you have to look really hard to see if anybody in the electric car industry is making money. So the number of people who are working on it is limited by that constraint. It’s a transition industry – it’s coming from zero. And in that short time, the speed of developing robust standards is going to be determined by that pace of development.
The second reason is a sequence issue: Who works on what? Specifically, scientists perform the upfront work setting the stage for basic developments, and engineers take it from those discoveries and develop real-world solutions. The universities are just starting to offer scientific and engineering disciplines to feed this sequence.
Charged: When you say “robust standard,” do you mean a testing protocol that each pack will have to go through?
ES: Yes, that’s exactly it. It also means testing that replicates all of the real-world scenarios. In other words, a battery in an electric vehicle that works all of the time, anywhere, with any driver who unknowingly or knowingly abuses the vehicle, under any weather conditions, in any geographic region, in any terrain and from new to old condition. Most of it covered by warranty.