Manufacturers

This year has been one of heightened awareness regarding battery safety. Smartphone manufacturers now realize that battery fires are real and recalls from the field are enormously expensive. Along with this education comes a deeper realization that battery quality and the presence of hidden defects are very serious matters.

Today’s post sheds light onto the challenges of battery manufacturing and ensuing defects. These discussions are not useless academic conversations. We see manufacturing defects in batteries especially those sourced from Chinese manufacturers. We observe them more frequently in batteries with high energy density where manufacturing tolerances are challenging.

We analyzed recently a family of batteries from a China-based manufacturer that shall remain unnamed. The battery in question is in serious consideration for a possible release in a smartphone in 2017. Our analysis revealed that the battery was potentially unsafe exhibiting an elevated risk of lithium metal plating; this can lead to electrical shorts. A post-mortem dissection showed regular horizontal bands of lithium metal deposited on the surface of the graphite anode. The scary part was that this battery was quite new….it had only been used for a few days with less than 10 charge-discharge cycles. So what’s happening?

Let’s go back to basics. This previous post recaps the basic structure of a lithium-ion battery. There are two electrodes that face each other with a porous insulating layer in between meant to keep these two electrodes apart. The anode has a coating of carbon (in the form of graphite) on top of a copper layer. The cathode has a coating of a specialized metal oxide, often lithium cobalt oxide abbreviated as LCO.

The design of the battery dictates that the amount of graphite compared to the amount of LCO must be balanced. This balance, which in technical jargon is called the A to C ratio, requires that there is a small amount of excess graphite relative to LCO. Usually, in a good design, there is about 5% more graphite material than LCO.

If, for some reason, there is less graphite than there is LCO, then this creates a condition where excess lithium ions cannot be absorbed by the anode, leading to lithium metal forming on the surface of the graphite….this is called lithium metal plating. As I said before and I will say again, lithium metal is a NO NO. It is a risk for electrical shorts and it is highly flammable in the presence of water vapor or oxygen; both are precursors for poor safety.

In a good manufacturing environment, the graphite anode layer is uniform in thickness and density. In other words, anywhere along the large anode surface, the graphite maintains the proper balance with the cathode layer that sits on the opposite side.

But now imagine a scenario where the manufacturing is not well controlled such that regions of the anode (or the cathode) are not uniform — for example the thickness or the density of graphite particles vary. Now remember that this need not be a huge variation: only a mere 5% change in particle density is sufficient to create an imbalance between the two electrodes. That is precisely what happened in this Chinese battery. The figure below illustrates how these bands are related to the defects in the electrode layer.

When the manufacturer coated the anode and cathode layers, their machines created small ripples in the density and/or thickness. So when the battery was first powered, lithium metal began to form…and that was what we observed in our laboratory.

So now you will say, “Well, that was a cheap Chinese factory. Surely the more reputable incumbents do not have this problem.” True, but that is not good enough. The economics are not in their favor. I will explain.

The tolerances of manufacturing high-energy-density batteries are becoming very tight. This drives the need for newer state-of-the-art manufacturing equipment, such as machines that can coat the electrodes more uniformly. The cost of equipping new manufacturing facilities therefore increases, driving up the cost of manufacturing these batteries.

Compare this to manufacturing silicon integrated circuits. Moore’s Law makes it possible to spend billions of dollars on manufacturing facilities and yet amortize this cost over a rapidly increasing number of electronic chips. The result is the amortized cost per chip actually goes down.  Sadly, there is no equivalent of Moore’s Law in batteries. More expensive manufacturing means more expensive batteries, and consequently, a loss of market share to the low-cost Chinese battery manufacturers. It is no surprise therefore that the reputable incumbents in lithium-ion batteries like LG Chem and Samsung SDI have their sights set on the electric vehicle market where they get to build the power train, not just a battery. This is a big invitation to low-cost  and low-quality battery makers in China to continue expanding.

The combination of increasing energy density along with the enormous pricing pressures from second-tier battery manufacturers in China are invariably leading to increased incidence of manufacturing defects — and that is something smartphone and device OEMs ought to be thinking about very seriously.

A recent article published by The Verge attempted to explain the science behind the exploding Samsung Note 7 batteries. The article touches on several important aspects of battery safety but the handwaving did not really talk about much science. So this post will address a failure mode of lithium-ion batteries and how defects can form during manufacturing with catastrophic results.

One of my earlier posts described the inner structure of a lithium battery. In a nutshell, there are alternating material layers that form the basic structure of the battery: a sandwich of two electrodes, called the anode and the cathode, with an insulating separator between them. During manufacturing, these layers are assembled then rolled together like a cigar before they are packaged into a protective sleeve. This is a gross simplification but highlights the basic structure and assembly of the lithium-ion battery. With some minor exceptions, the manufacturing is primarily an assembly process, and does not resemble in any form the manufacturing processes used in semiconductor devices.

The first figure below shows a rudimentary drawing of the basic structure of the lithium-ion cell. The graphite anode, shown in black, sits counter to the cathode, shown in green. The separator, shown in blue, is sandwiched between the two electrodes and acts as an insulator, in other words, its primary function is to prevent internal electric shorts between the two conductive electrodes. We all know that electric shorts are not good!

One of the basic requirements in the design of the battery is for the graphite anode to physically extend beyond the edges of the cathode. In other words, the anode is wider than the cathode at every point, especially the long edges of the sheets. This is needed to maintain safety within the cell and prevent the formation of lithium metal. Intuitively, there has to be more anode material than cathode material to absorb all the lithium ions. When the anode is not properly sized, the excess lithium ions will deposit as lithium metal, and that is called lithium plating. If you would like to dig a little deeper into lithium plating, this earlier post will shed some additional insight.

In practical terms, the anode is wider than the cathode ever so slightly, only a few percents. Any extra width of the anode does not participate in energy storage. In other words, the extra width of the anode is required for safety reasons, but does not contribute to charge storage. So battery designers go to extremes to optimize the extra width of the anode for the requisite safety.

As energy density increases, these battery designers have limited choices, one of them is to reduce the width margin of the anode. This means that the additional width of the anode relative to the cathode is now at its bare minimum. Any errors in manufacturing that jeopardize this extra overlap may have dire consequences.

So now let’s examine one particular manufacturing defect where a slight misalignment between the anode and cathode occurs during the assembly process. The figure below shows the same structure as above but now the anode layer is shifted ever so slightly to the right.

At the misaligned edge, the requisite overlap of the anode relative to the cathode is now diminished or even possibly vanished. The A/C ratio at this locale drops below the requisite limit for ensuring safety. The result, as you expected, is the onset of lithium metal at this edge. The lithium metal forms on the anode edge. As the lithium metal grows in size and thickness, it ultimately punctures the separator and causes an electrical short between the anode and cathode. Boom! we now have a catastrophic failure.

So this begs the question: why did Samsung release new software that limits the maximum charge in the faulty Galaxy Note 7 to only 60% of maximum? It is because the risk of lithium metal plating heavily depends on the voltage and the maximum charge in the battery. This is evident in the voltage chart of this earlier post: the higher the voltage, i.e., the higher maximum allowed charge, the higher the risk of lithium metal plating.

I will close by reiterating one final thought. The tolerance requirements in the manufacturing of lithium ion batteries have risen sharply with increasing energy density. Short of using new materials (that still do not exist in commercial deployment), increasing the energy density means reducing all the extra space inside the battery that is not made of anode and cathode materials. These are the only two materials that store energy. Everything else is just overhead…i.e., dead weight. They are still needed for other functions and safety, but they do not contribute to storing electrical charge. So battery designers keep reducing this overhead and in the process, make the manufacturing tolerances every so tight….and that is a recipe for many disasters to come unless we start adding a lot more intelligence to the battery to avoid and mitigate these undesired situations.

The images of melted Samsung Note 7 smartphones are all over the internet. News of Samsung’s massive recall are headline news. It is embarrassing to Samsung Mobile, its marketing and engineering teams, and most certainly its executives. Consumers are wondering how could Samsung ship units with defective batteries that can catch fire.

It is easy but not right to pick on Samsung or be critical of the company at this moment. Why? because this could happen to anyone…that’s right, anyone. If you are an OEM of smartphone devices or consumer devices with lithium-ion batteries, this is the time for you to pay attention to your products because you could be next.

While this sounds ominous, the intent here is to raise safety awareness in the entire ecosystem that depends on batteries. Samsung happened to be the first unlucky company to exhibit the strains that have been accumulating now for several years. I have covered in several past posts how the battery industry has been hitting the wall. Battery materials are reaching their limits. Battery economics are not favorable. Yet, the performance demands on batteries continue to rise. All of these factors are and continue to be precursors to the situation that Samsung finds itself in.

As is often the case in life, we tend to remain complacent until a crisis hits. The crisis is here, and now. Samsung is first to feel the pain, but each and every company in this ecosystem, from consumer devices to energy storage and electric vehicles, should acknowledge the severity of the situation and participate in its solution. Again, why?

This perfect storm has been brewing for a while, in particular, the drive to increase energy density along with faster charging while making less expensive batteries. Increased energy density and faster charging operate the battery near its physical limits. In other words, the margins for error at these elevated performance levels are really thin. For example, the newest lithium-ion cells now operate at a terminal voltage of 4.4 Volts, up from 4.2 Volts a few years back. This increase in voltage is one of the underlying physical tenants of increased energy density, yet it moves the battery every so close to the edge of the safety abyss. Another example relates to charging speed: it is widely accepted now that charge rates are approaching if not exceeding 1C. Electric vehicle makers are actively exploring very fast charging for EVs. Tesla is deploying their superchargers at a fast pace. These superchargers can charge a Tesla model at up to 1.5C, i.e., put in half a tank in about 20 minutes. Fast charging wreaks havoc inside the cell if not properly managed.

So now add the push for making less expensive batteries. Battery manufacturing, unlike semiconductors, does not scale. There is no equivalent of Moore’s law. In other words, as energy density increases, the cost per Wh (per energy unit) does not decrease…au contraire, it tends to increase because manufacturing tolerances get tighter. As a result, capital expenditures go up. Combine that with low-cost, low-quality batteries coming out of China and at a fundamental level, you can see how the financials of battery companies do not look pretty. This invariably leads to changes in manufacturing processes as companies seek more efficient ways to manufacture. But when the design margin of error is so thin, it does not take much before small variations in manufacturing lead to disastrous consequences. Remember, all it took in the case of Samsung was 35 failing devices out of a total of 2,500,000 shipped to cause a recall. This is a failure rate of 14 ppm (parts per million). It is a small number but, clearly, not small enough.

This is not to say that battery manufacturing and battery technology are doomed. There are countless examples in history where engineers built far more complex systems and structures safely and economically…but usually these include a change in paradigm. For example, pause for a second and compare the first commercial airplanes with the most recent jetliners. The newest Boeing and Airbus commercial airliners are marvels in computation and software. Fly-by-wire and automated systems with redundancy are the norm today, yet these new airplanes are scantily faster than their predecessors. In other words, the industry added so much more intelligence and shifted the burden to computation. The result is that modern planes are vastly safer than ever and far more economical to operate.

This is precisely the opportunity in front of the battery manufacturers and their customers, the OEMs, to think deep and hard on how they are going to implement a lot more intelligence to manage their batteries. Kudos to Sony for recognizing this….the batteries in their smartphones carry a great deal of intelligence, perform incredibly well and are safe. I am biased here…a lot of that intelligence is from Qnovo, but that should not diminish from the importance of the point of needing intelligence to manage the vanishing margins of error that battery designers have to cope with.

Every so often, I hear at a dinner gathering the question: “Is there enough lithium?” I have already covered this question in a previous post. The answer is YES. However, an abundance of a resource in the earth’s crust does not mean that its economics will always be favorable. These depend on several factors, in particular the balance or lack thereof between supply and demand.

We know demand is rising. Batteries in consumer devices, electric vehicles, energy storage….But what about the supply? how is lithium mined? where is it mined? which countries or corporations control this supply chain? …etc. Let’s explore.

Let’s start with some basics. Lithium is a metal. With an atomic number of 3, it sits right below hydrogen in the first column of the periodic table. That means it is a very light metal. Also, it means it is an alkali metal, and as such, it is highly reactive and flammable. Lithium is not found in nature in its purest metal form. Instead, it is found in various types of deposits; the most common lithium ores are spodumene and petalite minerals (mined as pegmatite rocks), as well as lithium brine deposits that are essentially underground saline water enriched with dissolved lithium.

Spodumene is a lithium aluminum silicate, with the chemical formula LiAlSi₂O₆. In its simplest form, it is a yellow or brownish crystal but spodumene includes two gem varieties that are more precious: the pink Kunzite and the rarer green Hiddenite. Petalite, with the chemical formula LiAlSi₄O₁₀, often occurs alongside spodumene, though the latter has a higher Li₂O content and is considered the more important ore.

Spodumene is found in low concentrations in pegmatite rock deposits (these are rocks, like granite, that formed millions of years ago in the final stages of the crystallization of magma as it cooled down). Spodumene mines follow traditional drill-and-blast methods that expose unweathered zones of the pegmatite rock ore. The spodumene ores containing about 3 to 5% Li₂O  are then processed in neighboring plants into high-grade lithium sludge-like concentrates that are then crated for shipment. This is what a refinement or a conversion plant receives for further processing into the final product, either lithium carbonate or lithium hydroxide. The latter is typically used in the manufacture of batteries using NCA or NCM cathodes (like the ones used in the Tesla electric vehicles) whereas lithium carbonate is the preferred material for batteries with LFP cathodes (widely used in China) or LCO cathodes (the typical cathode material in consumer applications). This older post reviews the different types of common cathode materials used in lithium-ion batteries.

Lithium brine deposits are processed in different and far less expensive ways. Much like salt ponds used to make table salt, the brine, often holding a concentration of 200 to 1,400 mg per liter of lithium, is pumped to the surface and stored in a succession of ponds where evaporation results in a higher concentration of lithium salt. This drying process can last 9 to 12 months and yields the required 1 to 2% concentrate of lithium. This is then further refined at chemical plants into the final product, again lithium carbonate or lithium hydroxide. A lithium brine field might require $150-$300 million in capital expenditures, whereas a spodumene mine could easily require 5 to 10 times that amount. But as we will see below, there are advantages to spodumene mines.

The primary mines for spodumene are in Western Australia with the Greenbushes mine being one of the largest. Africa also boasts of additional mines with the Manono-Kitolo mine in the Democratic Republic of Congo being a notable one. There are also known deposits in the US, Canada, Europe (e.g., Austria, Serbia, Russia) either still unmined or in small mining operations. In contrast, brine deposits are largely concentrated in South America, with the Uyuni field in Bolivia and the Atacama field in Chile containing enormous reserves. The provinces of Catamarca, Salta and Jujuy in Argentina also hold significant reserves. The USGS reports that Australian and Chilean mines each produced in 2015 a total of 13,400 and 11,700 metric tons, respectively, accounting for 77% of the world’s production of lithium. An interesting tidbit: The United States government, through its Defense Logistics Agency Strategic Material, held in 2015 a strategic reserve of 150 kg of LCO and 540 kg of NCA battery cathode materials.

The Greenbushes mine as seen from space (Google Earth). The opening of the mine on the right is about 2000 ft (or 600 m).

Known world’s deposits of spodumene tend to be smaller than those of brine deposits, around 10 million tons of lithium for spodumene vs. over 25 million tons of lithium for brine. However, the diverse geographical distribution of spodumene deposits makes pegmatite mining less susceptible to supply chain disruption and a more reliable source of lithium.

Companies mining for lithium have seen their fortunes rise in the past 10 years. In Chile, anecdotally called the Saudi Arabia of lithium, Sociedad Quimica y Minera (SQM) is the largest lithium mining company in the world. It just formed a new joint venture with Western Lithium and Lithium Americas, two smaller US operations. In the US, FMC Corp. and Abermale  Corp. (through its acquisition of Rockwood Lithium) are two large players. In Australia, Talison Lithium is very prominent. These four companies account for about 55% of the world market of lithium, with Chinese chemical companies, such as Ganfeng Lithium and Tianqi,  accounting for an additional 45%.

Given the rising demand for lithium, it is not surprising that the four major Western suppliers all announced significant expansions in their production of lithium carbonate and hydroxide. But as supply tries to keep up with demand, spot prices of lithium ore have hit a near high price in excess of $600 per tonne, up from a long term average near $400 per tonne. Goldman Sachs estimates that these prices translate to a spot-market product price of $20 per kg for lithium hydroxide up from the average $9 per kg that Korean and Japanese battery makers were typically paying. As supply and demand balance out in the coming years, lithium hydroxide pricing will return to more normal levels, but these normal levels could easily be above $10 per kg. In other words, lithium is getting more expensive — but one thing is very likely, as lithium gets more expensive, batteries continue to get commoditized placing serious financial pressures on battery manufacturers.

If you are an investor, however, looking for a pure lithium play, you are a little out of luck. That is because many of the mining companies tend to be diversified chemical conglomerates, and lithium, as a commodity, still does not have futures contracts or swaps, leaving equities as the only play.

A young woman, Anna Crail, was flying on 19 March of this year on an Alaska Airlines flight from Seattle to Honolulu. About 90 minutes prior to landing, her iPhone 6 suddenly broke out in flames causing panic in mid flight. The fire was rapidly extinguished by the flight attendants, but not without leaving the airline and the FAA searching for answers. This is one of several safety-related battery incidents that are becoming increasingly common. There are countless reports on hoverboards that are catching fire. While safety-related incidents involving Apple iPhones appear to be sparse in the media, there are increasing reports of Android-type mobile devices posing serious safety hazards, especially in Asian geographies. This post provides first insights on the factors that impact the safety of the lithium-ion battery in mobile devices.

First, let’s start with some background material. Three key design factors heavily influence the lithium-ion battery in a smartphone: i) Higher charge capacity, i.e., more mAh; ii) Faster charging ; and lastly iii) thin profiles.

Higher capacity is driving an increase in energy density at the rate of about 5 to 7% per year. The energy density of lithium-ion batteries when the first iPhone was launched in 2007 was near 400 Wh/l. Today’s state-of-the-art mobile batteries are in the range of 600 to 700 Wh/l. These higher energy densities are associated with two important physical parameters: higher terminal voltage (near 4.4V, up from 4.2V ten years ago) and significantly higher current flux inside the batteries (a lot more ions are making the journey between the electrodes) elevating the risk of damage within the battery.

Fast charging is now an expected feature in mobile devices, at least those on the higher end of the spectrum.  A charge rate of 1C corresponds to a 50% charge in 30 minutes…1C or faster is becoming the norm. Faster charging also means a lot more ions are making the journey within the battery between the two opposite electrodes. Again, faster charging = higher risk of battery damage.

Lastly, thin profiles of smartphones are pushing the battery to ever thinner dimensions. The next generation of smartphones are employing batteries that are a mere 3 to 4 mm thick, with this figure being pushed down even further where possible. Thin batteries create a slew of headaches for engineers…their performance tends to be inconsistent; manufacturing non-uniformities are amplified; and the current flux (and corresponding ion density) within the battery is also pushed to higher levels.

So what do these really mean? they mean that the perfect safety storm is brewing if proper care, battery intelligence, and diagnostics are not implemented.  One of the first consequences of higher energy density (especially higher operating voltages near 4.4V) is the increased risk of formation of lithium metal on the carbon anode (the negative electrode in the battery during charging). This is called lithium plating. Combine high energy density with fast charging and thin batteries, and the risk of lithium plating becomes dangerously significant. But lithium metal (i.e, in its molecular form, not in its ionic form) is highly flammable especially in the presence of oxidants. Additionally, spurs of lithium metal can cause electrical shorts within the battery…both of these mechanisms have seriously hazardous consequences.

The next figure illustrates the safety risk and its relationship to charge rate and energy density. Right around 1C and 600 Wh/l, the battery may become a safety hazard, especially in the absence of proper and diligent designs. Some battery manufacturers are better than others, with batteries made by Chinese manufacturers being the most prone to increased safety risks. Some device manufacturers (OEM) choose to sacrifice some battery specifications to gain a little safety margin. For example, some OEMs reduce the operating voltage of the battery as it ages, for example, from 4.4V down to 4.35V or less. This means that you will be robbed of mAh without being told. Your smartphone may have a great battery (say for example, 3000 mAh) at the beginning of its life (and when it operated at 4.4V), but a few months in its operation, the maximum voltage is intentionally reduced to 4.35V thereby reducing the capacity by 150 – 200 mAh; in other words, your battery is now only about 2800 mAh. Ouch! That’s not good, especially when you are not aware of it.

The safety risk also depends on temperature, rising rapidly with lower temperatures…and by low temperatures, I really don’t mean sub-freezing temperatures. The vast majority of battery safety tests are conducted at room temperature, usually near 25 °C (77 °F). What is considered as “low temperature” for a battery is 10 °C (50 °F) or lower. Right around this temperature range, the probability of plating of lithium metal soars creating serious hazards.

For the time being, batteries catching fire have been mostly limited in frequency and consequences. But with rising energy density and charge rates, the safety hazard is slated to become a lot more serious in the near future. Look for smartphone OEMs that are investing in the proper solutions to give you an excellent battery experience AND a safe one too.