05 agosto 2019
One of the strategic themes expected to shape our long run operating environment is the electrification of transport.
This is the second episode in a series on this vital topic.
The first episode in this series updated our long run forecasts on the penetration of electric vehicles (EVs) in the light duty fleet.
This episode turns its attention to the types of batteries that are expected to power this fleet over time – and the amount of nickel that will be required to produce them.
The batteries that will drive the mass market EVs of the future will be expected to provide adequate energy to replicate or exceed the performance of an internal combustion engine; and to do so while:
- storing and discharging sufficient energy to sustain that performance for a minimum of 200 miles on a single charge;
- being light and compact enough to not unduly compromise on performance, cabin comfort and storage;
- exhibiting an adequate cycle life calibrated to the life of the vehicle itself;
- being rechargeable in a reasonably competitive period of time;
- doing all of the above safely; and
- delivering these services at the bargain cost of less than $100 per kWh.
Putting that in other terms, a battery capable of enabling an EV takeover of the light vehicle market must be powerful, energy dense, thermally stable, capable of being recharged thousands of times, and reach a high charge in a matter of minutes. It must also use raw materials that are in reliable supply and are reasonably priced.
So, how did we go about addressing this multi-faceted problem?
The first task is of course to assess the current set of commercialised technologies on this range of criteria. Within the dominant Lithium-ion “genus”, the key battery chemistry “species” are Lithium-Nickel-Manganese-Cobalt (hereafter NMC); Lithium-Nickel-Cobalt-Aluminium (NCA); Lithium-Iron-Phosphate (LFP); Lithium-Manganese Oxide (LMO); and Lithium-Manganese-Nickel Oxide (LMNO). The proportions of each material in these chemistries are denoted as shares of ten. For example, for the NMC 5-3-2 “sub-species”, that means 5 units of nickel, 3 of manganese and 2 of cobalt in the battery cathode.
Examples of the commercial use of these chemistries are the first generation Nissan Leaf (LMO); Tesla (NCA), Chinese brands like first generation BYD (LFP), while NMC chemistries are commonly used in the EV offerings of western automakers, such as the Chevy Bolt.
It is useful to visualise our assessment of the relative strengths and weaknesses of these species and sub-species. Our preferred method for doing so is to use five axes to capture in a single snapshot the absolute and relative benefits of each technological option. The origin on the chart represents the uncompetitive frontier (in our estimation), and a score of 9 at the extremity of each axis (corresponding to the US Department of Energy’s standard 1-9 scale of technological readiness) represents the best of breed.1
The simple lesson from this exercise is that no battery is perfect.
The simple lesson from this exercise is that no battery is perfect. Any choice that battery manufacturers or automakers take therefore comes with trade-offs. The battery chemistry that is ultimately adopted must be able to mitigate its weaknesses (at acceptable cost) while leveraging its competitive strengths.
In our minds, the reason the industry is converging on nickel rich lithium ion (right panel above) and away from NMC 1-1-1 and LFP chemistries (left panel above), is that nickel rich chemistries present the best opportunity to achieve the positive performance characteristics that are required, and its weaknesses (principally thermal stability) are manageable at an acceptable cost. Within the nickel rich category, at the margin, we favour NMC 8-1-1 over NCA. That is due to the former’s cycle life advantage, and the existing economies of scale, which we expect will ultimately out-weigh the latter’s likely raw material cost advantage (A versus M) and its mildly better density characteristics.2
That is why we expect NMC 8-1-1 to be the “workhorse” technology of the EV future.
The fact the diffusion of NMC 8-1-1 does not require a basic scientific breakthrough, as do competing embryonic chemistries, we consider to be a huge advantage in what is ultimately likely to be a scale game in mass market EVs.
Given relatively swift progress to date moving from NMC 1-1-1 to 5-3-2 and 6-2-2, we see the progression and ultimate commodification of 8-1-1 as an iterative engineering problem with respect to improving safety and cycle life. Displacing cobalt units with nickel units will see the cost and energy density parameters take care of themselves. The fact the diffusion of NMC 8-1-1 does not require a basic scientific breakthrough, as do competing embryonic chemistries, we consider to be a huge advantage in what is ultimately likely to be a scale game in mass market EVs.
So much for enabling the EV “take-off”. In the long run, well beyond the cost competitive tipping point versus internal combustion engines, as EV penetration gains pace across all light vehicle segments, we expect that a multiplicity of chemistries will emerge. This is a logical outcome of both a concerted research effort as well as the varied demands of customers in each of these segments.
We expect the greatest variety in the EV high case. That reflects the way we think about breakthrough and incremental innovation across all of our forecasts. Outlays on R&D respond to the profit motive and available cash flow, which are both elevated in high case worlds. In low case worlds, cash flow is constrained and firms tend to “settle” for incremental innovation in existing processes and technologies, with little financing available for “moon shots”. In such a world, there is a large degree of technological lock-in, with R&D aiming predominantly at thrift. In the high, nothing is fixed, everything is up for grabs and a major disruption is entirely possible.
Our views on some of the early stage technologies that might break through in time are summarised visually below. Note that Li-S stands for Lithium-Sulphur; LMR-NMC stands for Lithium-Manganese Rich; LNO stands for Lithium-Nickel Oxide.3
The result of all of the above is a range of views on the future battery underpinnings of the EV world to match our unit sales and market share projections.
Before that information can be turned into a projection of total demand for primary nickel units, you need two more things (well, four actually, given we don’t work in mid-points, only in ranges). They are (1) low and high case views on demand for non-battery uses of nickel, which after all, absorb around 95% of nickel units today, and (2) low and high case views on the supply of secondary nickel through changing recycling rates of batteries over time.
The result of this exercise is summarised below. In short, our expectation based on our analysis is that by 2050, either the primary nickel market will increase by a little more than 50% in the low case; or it will almost triple in the high case (and indeed, it would triple at the projected peak in the early 2040s before a combination of a saturated share of light vehicle sales and large scale battery recycling really kicks in).
The importance of the recycling assumption becomes clear in the shape of the high case demand trajectories for total primary demand (left panel) and batteries (right panel).
That is not the end of the story from a nickel point of view. The majority of first-use nickel demand today comes from the stainless steel sector. As is well known, the rise of “nickel pig iron”, which is acceptable to most end-users in stainless applications despite containing very small quantities of nickel itself, has drastically re-shaped the cost curve over the last decade or so. We expect ongoing increases in market share in stainless applications for such low quality, “class II” products.
Our customers in the plating, alloying and battery precursor fields, on the other hand, demand higher quality “class I” nickel products. As battery precursor demand grows from consuming less than 5% of total primary nickel units today to 40-50% of the market at mid-century, the expected need to induce primary supply capable of meeting “class I” demand at reasonable cost will become obvious.
Looking at the upstream segment of the nickel value chain, we see opportunities for nickel sulphide resources to produce attractive margins under the range of conditions described above.
Nickel sulphides are relatively scarce vis-à-vis laterite nickel deposits, and tend to be higher in grade. They are also advantaged with respect to the cost required to convert the ore into one of the products suitable for entering the battery supply chain.
That is why we are seeking to add to our nickel sulphide resource base, with a focus on exploration in prospective areas close to our integrated Nickel West operations in Western Australia, and then, perhaps, somewhat further afield. If we are successful, we will create new growth options for our portfolio to take advantage of this trend, as it unfolds.
 Cost: Values are scaled relative to the estimated price of the lithium-sulphur (Li-S) battery pack.
Performance: Values are scaled relative to the energy density of the solid-state NMC battery (with an 811 cathode).
Safety: Values are scaled relative to LFP cathodes.
Cycle Life: Values are scaled relative to the LFP cathode which has a cycle life of ~2000 cycles.
Technology Readiness: TRLs are BHP judgements within the context of the US Department of Energy’s framework for technological readiness.
 Other factors in the NCA versus NCM debate include public source versus private intellectual property and the number of baking cycles required to produce the cathode active material.
 This is far from an exhaustive list. We also consider other solid-state chemistries, metal-air, redox flow and other low-to-mid TRL options, as well as anode innovations. We also note that this analysis does not account for performance at very cold temperatures. We expect that EVs will become competitive with ICE later in harsh climates, due to diminished performance on range, heating and charging once temperatures reach -10 to -20 degrees. We have incorporated this into our regional penetration forecasts.