Where Einstein Meets Edison

The Cost Dilemma: Why Are Batteries So Expensive?

The Cost Dilemma: Why Are Batteries So Expensive?

Mar 14, 2011

            As battery and fuel researchers, there is one question we frequently get asked (especially during Thanksgiving and other family gatherings):  “What will it take for electric vehicles (EVs) to become affordable for the common man?”  Currently, with battery costs alone comprising an estimated 25-50% of the electric vehicle price[1], success appears to hinge on when – and if – the battery price can be lowered enough to compete with the internal combustion engine.

             To help spur improvements, the U.S. Department of Energy has set a 2014 cost target for PHEV batteries at $200-300/kWh (in this context, ‘battery’ refers to the entire packaged unit: the cells, safety mechanisms, wiring, and casing).[2] To put this in perspective, industry experts estimate that current battery prices range from $600/kWh to as high as $1100/kWh.[3]  Since an average EV gets about 4 miles per kWh, this cost means that a PHEV with a 40 mile range requires $6,000-$11,000 worth of battery today.  For an EV with a 200 mile range, the battery cost skyrockets to an astonishing $30,000-$55,000. Even if this figure were to decrease by a factor of three, the battery cost would still amount to $10,000, which means we won’t be seeing EVs with a Honda Civic or Ford Focus price point in the near future (for a 200-mile range, the Tesla Roadster — the longest-range electric vehicle on the market — is currently selling at $101,500.)

            Despite this foreboding outlook, there’s still plenty of optimism. U.S. Department of Energy Secretary Steven Chu said in a recent speech in Cancun, Mexico,[4] “Is there any hope of being competitive with an internal combustion car engine? The answer is yes. It’s not like it’s 10 years off.” However, Chu notes, “It might be 5 years off.” 

            Five years sounds reasonable enough. But how reasonable is it to expect the battery price to decrease to a third of its existing price?  To help wrap our minds around the scope of the problem, we took a look at what comprises existing battery costs.  Then, we examined what types of technological improvements might result in sufficiently reduced battery prices. Finally, we took a look at how reasonable these expectations are, given such ambitious targets.

Battery Costs and Technological Opportunities

            State-of-the-art lithium-ion batteries consist of several internal components that contribute to the cost of the pack.  The core component of every battery pack is the cell, which, according to a 2000 study by Argonne National Laboratory,[5] carries a price tag of $400/kWh. The remaining 30-50% of the cost comes from electronics, assembly and packaging. (Current costs might be somewhat lower, reflecting engineering and manufacturing improvements to cell assembly since the study was published; however, these numbers are difficult to pin down, and the Argonne analysis remains useful as an upper limit for cost analysis of batteries).

            The cell itself consists of several components including the anode and cathode materials, the battery separator and electrolyte, and all of the metal leads, windings, and safety valves that are packaged inside. The most costly component within the cell is the cathode electrode material, which is typically a transition-metal oxide that contains cobalt or, with recent efforts to lower cost and improve stability and voltage, various ratios of manganese, aluminum, and nickel. According to the same ANL report, cathodes containing cobalt constitute roughly 50% ($200/kWh) of the cost of the cell. One reason for this is the high price of cobalt, which currently fetches about $18/lb on the global market, compared to $11.7/lb for Ni, $1.60/lb for Mn, and $1.08/lb for Al.[6] This per-weight cost translates to an estimated $40/kWh for the raw cobalt alone,[7] and that number does not even take into account the purification, synthesis, and manufacturing costs required to produce an electrode from the raw metal. Additionally, that cost assumes 100% utilization of the available energy within the cell. This assumption is not entirely realistic —battery capacity does degrade over time—and therefore excess materials are generally needed above and beyond the nameplate capacity. If this excess capacity could be eliminated and if the manufacturing and processing costs were to decrease by a factor of 2-3, the cost of a cobalt-containing cathode could theoretically be reduced by the same factor. 

             Binder, the glue that links together the cathode materials, represents another sizable contribution to the cathode price.  Reducing or replacing binder with a more effective, safe, low-cost alternative could lead to substantial cost savings, a strategy adopted by startup Electrovaya, which reported a new cathode forming-method that does not require the use of expensive, toxic binder material[8].

            Lithium iron phosphate, a relatively newer cathode chemistry, represents the next step towards affordable, safe and high-power batteries and is usually touted as being cheaper than the transition metal oxides. What does ‘cheaper’ actually mean?  While it’s difficult to find comparable data on production costs for the different cathode materials, a simple comparison of market prices gives a pretty good idea.  At Radio Shack, for instance, a 900 mAh lithium-cobalt oxide battery fetches $29.99, while a similar capacity iron phosphate battery is only $6.99.  However, the discharge voltage and the available capacity (how much charge you can actually cycle to maintain a good cycle life) are different for these two cells – iron phosphate has a lower voltage, although a higher useable percentage of charge.  Considering the useable energy of the cells, you would be paying approximately $18.61/kWh for cobalt oxide cells and only $3.04/kWh for the iron phosphate cells.

            Other costly components of a cell are the electrolyte, which is about 24% of the cost, and the anode, which is approximately 10%.  However, opportunities to lower the costs of these components are scarcer. Very few electrolyte chemistries work well with lithium-ion batteries, so there simply isn’t much flexibility with the recipe used to make it, and progress here has been relatively slower than with the cathode chemistries.  Furthermore, scaling is not likely to reduce these costs any further as these electrolytes are already widely produced in Japan and elsewhere for use in laptop and cell phone batteries. Anodes ($40-50/kWh) are made from an already quite cheap material, graphite, which will be very challenging to beat cost-wise (although there is still room to beat graphite weight-wise and volume-wise.)  From here on down the list, the other cell components represent a relatively minor part of the price and don’t suggest a promising strategy for large-scale cost reductions; at least not until the cathode problem is solved first.

Strategies for Reducing Cost Further

            Analyzing the existing cost of producing batteries is tough enough, but predicting future cost reductions introduces a whole new level of complexity. While batteries for laptops and cell phones are produced at a volume of roughly 1 billion/year, batteries for EVs are still being made in the thousands.  What would be the impact of scale? How would improvements in manufacturing processes affect the cell costs? Where is the biggest opportunity for cost reduction?

            The answer, not surprisingly, is: it depends. In a 2009 Merit Review[9] for the Department of Energy, TIAX, LLC modeled the future cost curves for batteries based on variables ranging from cathode chemistry, amount of material used per cell, how the cells are assembled, and even where they’re made. For control purposes, they assumed battery pack volumes were high enough to sustain annual production of 500,000 vehicles a year.

            The first finding is that with the right battery chemistry, it’s feasible for costs to go as low as $300/kWh for lithium iron phosphate or lithium manganese oxide cells.  That’s in the best of scenarios: electrode coating speeds improve, manufacturing process speeds double, and market prices for metals remain within a reasonable range.  Interestingly, though, when accounting for the uncertainty in the model, no single chemistry emerges as an obvious winner for making low-cost cells. Different chemistries require different processes, and this can partly counteract the advantages from lower materials costs alone. In a particularly interesting scenario for a nickel-cobalt-aluminum oxide cell, TIAX found that actions such as increasing the assembly line speed or even moving production from the U.S. to China would result in a more significant cost improvement than if the cathode materials costs were to decrease by a factor of eight.

            This last finding is particularly poignant for battery scientists who are working passionately in the lab to find lower-cost and better-performing materials. Indeed, the introduction section of many a battery academic publication touts a material’s supposed advantage in raw material price compared to existing materials. However, these claims are not usually quantified in practice, and when looking at the big picture as TIAX and others have done, it’s obvious that there is no clear answer.

The Long View 

            The TIAX analysis highlights the stark reality of manufacturing any high-tech product: changes in process flow and materials supply/demand happen on a much shorter timescale than do great discoveries in the lab.  While there may be great solutions underway in academia, these technologies will likely take ten to fifteen years before they are commercialized.  By then, we may well see a doubling, or even an order of magnitude increase in electrode production speeds and a more efficient world market for raw materials and labor flow.  Even if we don’t, there’s no telling how government subsidies of consumer EV purchases may influence the market, as well as how people’s attitudes toward price will evolve.  It may well be that the DOE target is not reachable – and it may be that it aims too low.  (And it may be that fuel cells become a competitive option. Who knows!)

            The best-positioned companies, then, are the ones taking the long view – focusing on manufacturing a solid-quality product, with high product yields and in relatively cheap labor markets.  Successful cost reductions will occur as a suite of improvements, and must not compromise the high standards for safety and robustness that exist for current batteries. A killer technology – like a great cathode material – will help, but it won’t guarantee ultimate success.  A well-run company with a diversified strategy, as is often the case, gets a lot closer to ensuring a bright future for batteries.

____________________________________________________________


[1] Calculated from Nissan Leaf and Chevrolet Volt published numbers for vehicle price, battery pack capacity in kWh, and an assumed range of $650-$1100/kWh for battery cost.

[2] U.S. Department of Energy FreedomCAR and Vehicle Technologies Program, “Plug-in Hybrid Electric Vehicle Battery Research and Development Activities,” 2007.

[3] Peter Fairly, “Will Electric Vehicles Finally Succeed?”, MIT Technology Review, January/February 2011.

[4] http://www.bloomberg.com/news/2010-12-06/competitive-electric-car-battery-may-be-5-years-away-secretary-chu-says.html

[5] “Costs of Lithium-Ion Batteries for Vehicles,” Argonne National Lab report, May 2000. Available online: http://www.transportation.anl.gov/pdfs/TA/149.pdf

[6] http://www.metalprices.com/

[7] Calculation assumes a cell voltage of 4 V, a capacity of 140 mAh/gcathode, yielding 0.560 kWh/kgcathode or ~ 1 kWh/kgcobalt.

[8] According to 2000 Argonne Study, binder cost constitutes only ~5% of the cell cost.  While this may not be significant, this percentage is expected to be significantly higher now that other costs have come down in the past 10 years.

[9] http://www1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/es_02_barnett.pdf

 

About