Posted by on in Energy

The Future in Storage is Long

The Future in Storage is Long

Can you imagine producing a product of value and being forced to pay someone to take it off your hands? Well, that is the situation that some renewable energy producers have faced over the last year. Now, this typically occurs during lower demand periods, perhaps when a particularly gusty weather front rolls through. But as renewable penetration increases, the phenomenon will proliferate unless an adequate storage buffer is put in place. Fortunately this is starting to occur. According to GTM and the Energy Storage Association, US energy storage installations increased in 2015 by 243% over the previous year with the installation of 221 MW of capacity, over half of that occurring in the last quarter.

Recent installations have been primarily short duration applications such as back-up power, frequency regulation and demand charge management. In 2015, the average duration of storage installations in the US was less than 45 minutes. Proven lithium ion batteries love many shallow, high-rate efforts and these services typically provide a value far in excess of typical electricity prices. Systems are in the money today, especially when they are geared to serve multiple applications.

But as solar and wind costs continue to drop, longer duration storage is needed to perform the simple task of shifting the variable supply to the periods of demand. In order for the energy mix to continue its shift to an increasingly renewable supply, long duration storage support needs to grow in parallel. Levelized cost of storage (LCOS) is the metric that describes the lifetime cost of a storage system divided by the amount of energy delivered over that life. For a simple renewable time of day shifting application, an LCOS in the single digit cents per kWh is when the tipping point occurs.

Lithium – The Silicon of Storage?

An energy storage system is analogous to a PV solar system where the solar module is like the battery pack and then you have a variety of other items that make up the system level costs. Over the last decade, silicon solar module costs dropped much faster than even the most aggressive forecasters imagined and the same is true for lithium ion batteries over the last several years.

Historically, cost analysis has been a difficult spectator sport, as the industry likes to compare apples with oranges when talking about costs. I always like to remember the famous line, “There are liars, damned liars and battery suppliers” (who said this?). Thankfully, Tesla recently introduced some transparency into the market. You can now price PowerPack systems with a neat little online tool. Their DC battery system runs at about $500/kWh and when you tack on an inverter you are looking at a cost of about $550/kWh, uninstalled. Unfortunately this is not advertised, but when you account for inverter losses, depth of discharge limits and capacity fade (Tesla’s batteries are warrantied to 60% of capacity at year 10) the “effective” capital costs creep closer to $1,000/kWh and LCOS is still in the double digits of cents.

But there is a key question in determining when the long duration storage tipping point starts to occur. Will lithium ion continue to follow those relatively steep cost reductions of the last few years, or are we approaching the period when marginal gains guide the cost trajectory on a much shallower decline?

An analysis of where costs can still be squeezed out is helpful to answer this - and sticking to Tesla as a proxy for the industry also makes sense. We have all heard about the $100/kWh cells projected to roll out of Tesla’s 50GWh Gigafactory. Tesla currently pays about $150/kWh for cells and so from a cell cost perspective there is about $50/kWh in efficiencies to be had. The analysis below will give some clues as to when.

  • Can the Gigafactory really achieve high yield manufacturing at 50GW scale by 2020? Considering that LG Chem, (no slouch in this industry by any means) is just reaching 1GWh at their Holland, Michigan plant after about five years and the Tesla plant is not scheduled to produce its first cell until the end of this year, the timeline seems like a stretch.
  • Fortunately, when you dig into the analyst reports, much of assumed cost reduction actually comes from materials and manufacturing improvements rather than pure economies of scale. Some are simple and obvious such as thinner current collectors. Other supposedly “drop-in” changes such as the conductivity improving carbon nanotube based additives, sold by our portfolio company Cnano Technology still can take several years to fully adopt. Adoption is particularly sluggish on new electrode technologies such as lithium- and manganese-rich cathodes. Short-circuiting the lengthy task of multi-variable cell engineering has proven difficult by even the best. Increasing silicon content and therefore capacity of the anode is already reducing the cost of Tesla’s automotive packs. But the cycle life required by the grid may make this approach elusive. So while a $100/kWh cell requires many knobs fully turned, you can’t just turn them all at once and automatically assume the longevity and reliability required by the industry is maintained.
  • The cost trajectory over the last few years has already benefited from a steep decline in the cost of commodities such as copper, manganese, plastics, and aluminum - in many cases below the cost of production. Further declines are unlikely and a return to stronger global growth could push costs up and negate the technology-driven cost improvements currently underway.
  • Let’s not even consider the cost of industry specific materials such as cobalt or lithium carbonate in the face of rapid growth in demand. Lithium carbonate saw a 300% spot spike in the last few months so this is really just a wildcard in the cost reduction game.
  • Margins can’t go any lower. Similar to solar modules, cell manufacturing has traditionally been a razor thin margin game and reports are that some companies are currently selling at a loss. Compared to solar, much more of the capital in the battery space is private and presumably at some point investors will expect to see an ROI. Up the value chain, some components such as ceramic-coated separators and NMC cathodes have decent margins today. There is a wave of patent expiry around 2020 and so that may be the next area affected by of margin compression.
  • Some people look to potentially game changing approaches, such as the new semisolid thick electrode paradigm proposed by MIT spinout 24M with a target high volume cell cost of $100/KWh. Step change? Possibly. Exactly when? Wait and see.

In short, companies will continue to march along with incremental gains but the tailwinds that they have been riding for the last few years appear to be dying down. As it turns out, the more important leverage point with cell technology is in improving calendar life, cycle life and capacity fade so that the energy delivered part of the LCOS equation can start to creep up. But once again these materials driven incremental improvements will only occur over time.

Beyond The Cell

In terms of reducing upfront costs, systems level items such as enclosures and structures, electrical systems, fire-protection, thermal management, inverters etc. make up two thirds to three quarters of the total capital cost. If we look at the solar industry these system costs have been the source of major reductions over the last few years. In the storage industry, inverters certainly have headroom to come down in cost by at least half. And systems level assembly work certainly has room to be driven down the cost curve. But for items like containers, major cost reductions seem like much more of a stretch.

High temperature stability and flammability has traditionally been lithium’s crux. As a result, the cooling and fire protection requirements contribute significantly to total system costs and can represent operational headaches as well. Without changes to the core chemistry there is no getting around these costs. Materials innovations such as solid-state batteries or high temperature electrolytes and additives may have a role to play in the future. But in the near term, these large format systems will probably be the last market where these premium technologies find a home.

So does the cost parallel hold true?

At a high level, the silicon solar and lithium ion battery cost roadmap analogies seem to make sense. But digging in, there are numerous different factors at play. The cell industry is already at significant scale and with material cost reductions and margin compression behind us, the cost reductions will be much harder to find. Similar to the solar cost evolution, system level costs are more likely to have a near term impact and we could very well see a $550/kWh system drop by $100/kWh or more as a result. At this level, there is definitely a future for lithium in long duration markets but we’ll still rely on juicy short duration markets to put the projects in the black.

Of course we can’t ignore the “beyond lithium” approaches so in Part 2 of this post, I’ll discuss the throng of players trying to displace lithium from its throne.

Partner, Pangaea Ventures Ltd. Andrew has over 12 years of energy and industrial experience, recently leading several of Pangaea’s investments in the energy generation, energy storage and energy efficiency domains. Andrew holds a Bachelor of Applied Science (Mechanical Engineering) and a Masters in Business Administration degree.View Andrew Haughian's profile on LinkedIn


  • No comments made yet. Be the first to submit a comment

Leave your comment

Guest Tuesday, 21 February 2017