The Supply and Demand Giga-Risk for Cobalt, Lithium and Graphite in Lithium-Ion Batteries
Posted on March 4, 2016 by John Petersen
For the last 10 years we’ve been deluged with news stories and investment analyses that extoll the virtues of lithium-ion batteries and speculate on the technology’s potential to change the world’s energy landscape forever. While the occasional curmudgeon like my colleague Jack Lifton questions the availability of enough lithium or flake graphite to satisfy soaring demand from the battery industry, everybody has overlooked or ignored the most critical mineral constraint – Cobalt.
It’s a truly gargantuan challenge. A Gigarisk!
This graph from Battery University, a wonderful source for understandable unbiased information on the ins and outs of the battery industry, summarizes the typical specific energies for nine different battery chemistries.
The six lithium-ion chemistries (orange columns) use varying amounts of lithium metal in their cathode and electrolyte formulations. While there is no single “right” value, a lithium-ion battery with perfect electrochemical efficiency would need 80 grams of lithium per kWh of capacity. Since perfection doesn’t exist in the real world, the industry average is closer to 160 g/kWh.
While lithium-titanate cells (LTO) don’t use carbon-based anodes, the five remaining chemistries use about 1 kg of processed carbon per kWh of battery capacity. Avicenne Energy, a global leader in lithium-ion battery manufacturing, supply chain and market analysis, reports that the battery industry used 55,000 tons of processed anode materials to manufacture 50 GWh of lithium-ion batteries in 2014. Avicenne estimates that roughly 35,000 tons of processed anode materials were derived from natural graphite while the remaining 20,000 tons were derived from artificial graphite, amorphous carbon and other sources. Since 70% of the mine-mouth weight of natural graphite ends up as waste when the graphite is processed into battery grade spherical granules, estimated flake graphite demand in 2015 was 133,000 tons.
According to Avicenne, the lithium-ion battery industry also used 118,000 metric tons of active cathode material in 2014, which works out to an average of 2.4 kg/kWh including scrap losses. According to Battery University, LCO, NMC and NCA, the lithium-ion chemistries with the highest energy densities, all use cobalt in their cathode formulations. For LCO, cobalt represents 60% of cathode mass, or roughly 1.44 kg/kWh. For NMC, cobalt represents from 10% to 20% of cathode mass, or roughly 0.36 kg/kWh. For NCA, cobalt represents 9% of cathode mass, or roughly 0.22 kg/kWh.
Avicenne expects global battery production to climb to about 135 GWh per year by 2025. Since Tesla, LG Chem, Foxconn, BYD and Boston Power are building new battery factories that will boost manufacturing capacity to about 140 GWh by 2020, one could argue that Avicenne’s forecast is overly conservative. For purposes of this analysis, I think overly conservative is best. Supply and Demand for Flake Graphite, Lithium and Cobalt. The following table is based on Avicenne’s detailed analysis of historical trends in the lithium-ion battery space and its conservative growth forecasts. It summarizes (1) current production rates (in metric tons) for flake graphite, lithium and cobalt, (2) estimated tonnage of each material the world’s miners sold to the battery industry in 2015 and (3) estimated tonnage of each material the world’s miners will need to supply to the battery industry in 2020 and 2025.
Column 1
Column 2
Column 3
Column 4
Column 5
1
Global
Battery Industry Demand
2
Supply
2015
2020
2025
3
Flake graphite
380,000
133,000
164,000
290,000
4
Lithium metal
32,500
9,760
12,160
21,520
5
Cobalt
92,000
35,223
40,110
64,842
When I reflect on these numbers and the production dynamics described below, I can’t help but believe the lithium-ion battery industry is overdue for a close encounter of the worst kind with the mining industry.
The problem is that flake graphite is the only critical battery mineral that’s mined as a primary product. Lithium and cobalt are lesser co-products or minor by-products; and increased demand for co-products and by-products can’t move the needle for mining companies without additional sustained demand for their primary products.
Flake Graphite Production Dynamics. Graphite is typically mined from surface deposits that have graphite concentrations of 20 to 30 percent and particle sizes ranging from very fine to very large. The typical mine produces 1/3 fines (-150 mesh), 1/3 small and medium flakes (-150 to +80 mesh and 1/3 large flakes (>+80 mesh). According to the USGS, natural graphite is used extensively in brake linings, foundry operations, lubricants, batteries, refractory applications, and steelmaking.
Graphite is not a particularly scarce mineral, but making battery grade spherical graphite granules usually requires high purities (>94%) and large flake sizes (over +150 mesh). While global graphite production is roughly 1.19 million tons per year, global production of large flake graphite that’s suitable for use in battery grade materials is roughly 380,000 tons. The process of making spherical graphite granules for lithium-ion battery anodes from large flake graphite is wasteful and it generally takes 3.3 kg of flake graphite to produce one kg of battery grade material.
In 2014 and 2015, China accounted for 65% of global graphite production. China is currently pursuing aggressive policies to clean up and consolidate its mining industry. It has also imposed a 17% non-refundable value added tax and a 20% export duty on graphite. These tax policies are designed to increase value-added manufacturing in China by giving Chinese industries a major competitive edge in the sale of products like batteries that use large amounts of graphite.
At present, the battery industry uses about 35% of global flake graphite production. Over the next 10 years, that percentage will increase to over 75% unless additional resources are brought into production. Given the usefulness of graphite in a wide variety of essential industries, I have a hard time accepting the idea that a single sector like the battery industry can lock down 75% of available supply and do so at attractive prices, much less lower prices.
While the Chinese taxes and tariffs on graphite will give their domestic battery manufacturers an advantage in the battery market, they will also give non-Chinese graphite producers an advantage in the raw materials market; particularly the market segment that supplies battery grade materials to non-Chinese battery manufacturers.
In my view, current flake graphite production dynamics provide a meaningful window of opportunity for junior graphite miners that can bring new projects online quickly; but my experience is that new mining projects always take longer and cost more than anyone expects.
Lithium Production Dynamics. Lithium is an odd “minor metal” and current global production is 32,500 tons of metal content, or 172,500 tons of lithium carbonate, per year. According to the USGS, lithium is used in batteries, 35%; ceramics and glass, 32%; lubricating greases, 9%; air treatment and continuous casting mold flux powders, 5% each; polymer production, 4%; primary aluminum production, 1%; and other uses, 9%.
According to SQM, the world’s largest lithium producer: “Salar brines are located in the nucleus of the Salar de Atacama. They contain the greatest lithium and potassium concentrations ever known, in addition to considerable sulphate and boron concentrations. From this natural resource lithium carbonate, potassium chloride, potassium sulphate, boric acid and magnesium chloride are produced.”
Operations on the Atacama Saltpeter Deposit produced potassium chloride, lithium chloride, boric acid and potassium sulfate; (Note 26.8)
Operations in the Salar del Carmen produced lithium carbonate and lithium hydroxide; (Note 26.8) and
Operations in the Salar de Atacama generated 37.5% of its total comprehensive income and the overwhelming bulk of combined income from its potassium and lithium segments (Note 19.3).
SQM also provided the following more granular disclosure of its comprehensive income classified by operating segments. (Note 26.3)
Column 1
Column 2
Column 3
Column 4
1
Revenue
Income
Segment
2
(000s)
(000s)
Margin
3
Specialty Plant nutrients
$502,300
$148,700
30%
4
Potassium
$333,800
$110,200
33%
5
Iodine and derivatives
$199,300
$60,000
30%
6
Lithium and derivatives
$160,100
$80,200
50%
7
Industrial chemicals
$79,300
$21,100
27%
8
Other
$42,200
$4,600
11%
9
Total
$1,317,000
$424,800
32%
Lithium production is SQM’s most profitable operating segment, but SQM’s operations in its Atacama Saltpeter Deposit would not make economic sense without the potassium products that generate about two-thirds of its revenue from brine operations. Therefore any decision to increase lithium production from the Salar de Atacama will be contingent on the strength of the potassium market because SQM can’t produce more lithium without also producing a lot more potassium. If increasing potassium supplies will depress prices for those products, the game won’t be worth the candle.
If SQM will have to grapple with co-product market strength issues in any decision to expand lithium production from the Salar de Atacama, I have to believe the other big companies that produce the overwhelming bulk of the world’s lithium from brines will face similar issues.
At present, the battery industry uses 30 to 35 percent of global lithium production. Over the next 10 years, that percentage will increase to around 70 percent unless additional resources are brought into production. Given the usefulness of lithium in other industries, I have a hard time accepting the idea that a single sector like the battery industry can lock down 70 percent of the available lithium supply and do so at attractive prices, much less lower prices.
In my view, current lithium production dynamics provide a real and substantial window of opportunity for junior lithium miners that can bring new projects online quickly; but my experience is that new mining projects always take longer and cost more than anyone expects.
Cobalt Production Dynamics. According to the Cobalt Development Institute, an incredible 94% of global cobalt supplies come from nickel and copper miners that produce cobalt as a minor by-product. Only 6% of global cobalt supplies come from primary cobalt mines that might be able to increase production in response to growing demand from the battery industry.
Last year, the world’s nickel miners sold $14.58 billion of nickel and $1.05 billion of cobalt, which means cobalt revenue represents 7.2% of their total revenue. It’s even worse with copper miners who sold $68.4 billion of copper and $0.92 billion of cobalt, which means cobalt revenue represents 1.3% of their total revenue.
While global cobalt production surged from 52,400 tons in 2005 to 92,000 tons in 2015, the bulk of the increase was attributable to new capacity from African copper mines. In addition to thorny regulatory issues for companies that buy minerals from African miners, production from African mines is far from guaranteed. In 2015 Glencore produced 421,900 tons of copper and 19,400 tons of cobalt from its Katanga, Mutanda and Mopani mines in Africa. In September of last year, Glencore announced the closure of Katanga and Mopani for 18 months as part of a debt reduction and modernization plan. Those closures will temporarily reduce Glencore’s cobalt production by several thousand tons.
Nickel and copper prices are both wallowing at multi-year lows due to decreased demand from China. Under those conditions, increased demand for minor by-product can’t possibly drive facilities expansion decisions.
In my view, expecting the supply side of the cobalt market to respond to the needs of the lithium-ion battery industry is like expecting a cropped tail to wag a Rottweiler.
According to the Cobalt Development Institute, the battery industry currently uses about 40% of global cobalt supplies. Over the next 10 years, that percentage will increase to about 70% unless additional resources are brought into production. While there is limited competition in the global markets for flake graphite and lithium, cobalt is one of the most useful metals going.
Essential non-battery applications include:
Superalloys for combustion turbine engines (16%)
Hardened high-speed steel for machine tooling (7%);
Hardened carbides and diamond tooling (10%);
Catalysts (7%);
Pigments (6%);
Magnets (5%); and
Others (8%).
The bottom line is that the cobalt supply chain is entirely dependent on global demand for nickel and copper. To make matters worse, cobalt demand is tremendously inflexible because of (1) the nature of the competitive uses, and (2) the lower price sensitivity of competitive users who need small quantities of cobalt for high-value products. While new primary cobalt mines may come on-line, exploration, permitting and development for a new metal mine typically involves decades of work and billions of dollars.
In my view the battery industry is careening toward a natural resource cliff at 120 mph while fiddling with the touch-screen and tweeting about its unlimited potential. The supply side of the cobalt equation is entirely dependent on global demand for nickel and copper. The demand side of the cobalt equation includes a rich variety of manufacturers that must have cobalt for products the world considers essential.
Whenever increasing demand crosses swords with inflexible supply, the inevitable outcomes are chronic shortages and substantially higher mineral prices.
The battery industry itself does have the theoretical ability to free up additional cobalt supplies by phasing out LCO chemistry, which requires 1.44 kg/kWh of cobalt, and repurposing their existing LCO factories to make cells with NMC or NCA chemistries, which require 0.36 and 0.22 kg/kWh of cobalt, respectively. But I’d love to be a fly on the wall of a meeting where one battery manufacturer tells another, “I want you to stop making LCO batteries so that I can use your supply chain for my business.”
BYD and LG Chem currently manufacture LCO batteries and they may be able to repurpose their supply chains to accommodate higher production of low-cobalt cells. Unless new market entrants like Tesla, Foxconn, and Boston Power are stealthily reconfiguring their gigafactories to make low energy LTO, LFP and LMO cells, I see no reasonable possibility that any of them will be able to lock down a secure cobalt supply chain in the foreseeable future.
Conclusion. In the lithium-ion battery industry, materials costs are 50 to 70 percent of total manufacturing costs; one of the highest ratios on the planet. Prevailing green mythology holds that lithium-ion battery prices will fall dramatically as anticipated production rates soar due to gigafactories and undefined “economies of scale.”
The last time I checked, the difference between a known mineral resource and a reliable factory feedstock was a couple decades of exploration, permitting, development and construction work, coupled with a couple billion dollars in capital spending.
Notwithstanding the prevailing green mythology, I’ve never met a miner, or for that matter a mining investor, who was incentivized to embark on a new project by expectations of lower future product prices. They’re all in it for the money. Given the current production dynamics for both lithium and cobalt, I’m firmly convinced that increased demand can only lead to higher raw material prices and excruciating shortages. Since most competitive users of lithium and cobalt are far less sensitive to raw material prices than battery manufacturers, it’s certain that they’ll protect their critical supply chains and the battery industry will either have to pay up or do without.
It doesn’t matter how big your battery factory is if you don’t have a rock solid supply chain for your critical minerals.
