Vikström, H., Davidsson, S., Höök, M. 2013. Lithium availability and future production outlooks. Applied Energy, 110(10): 252-266. 28 pages
[ This is by far the best paper explaining lithium reserves, lithium chemistry, recycling, political implications, and more. I’ve left out the charts, graphs, references, and much of the text, to see them go to the original paper in the link above.
I personally don’t think that electric cars will ever be viable because battery development is too slow, and given that oil can be hundreds of times more energy dense than a battery of the same weight, the laws of physics will prevent them from ever achieving enough energy density — see my post at Who Killed the Electric Car. (and also my more-up-to-date version and utility-scale energy storage batteries in my book When Trains Stop Running: Energy and the Future of Transportation. Some excerpts from my book about lithium and energy storage:
Li-ion energy storage batteries are more expensive than PbA or NaS, can be charged and discharged only a discrete number of times, can fail or lose capacity if overheated, and the cost of preventing overheating is expensive. Lithium does not grow on trees. The amount of lithium needed for utility-scale storage is likely to deplete known resources (Vazquez, S., et al. 2010. Energy storage systems for transport and grid applications. IEEE Transactions on Industrial Electronics 57(12): 3884).
To provide enough energy for 1 day of storage for the United states, li-ion batteries would cost $11.9 trillion dollars, take up 345 square miles and weigh 74 million tons (DOE/EPRI. 2013. Electricity storage handbook in collaboration with NRECA. USA: Sandia National Laboratories and Electric Power Research Institute)
Barnhart et al. (2013) looked at how much materials and energy it would take to make batteries that could store up to 12 hours of average daily world power demand, 25.3 TWh. Eighteen months of world-wide primary energy production would be needed to mine and manufacture these batteries, and material production limits were reached for many minerals even when energy storage devices got all of the world’s production (with zinc, sodium, and sulfur being the exceptions). Annual production by mass would have to double for lead, triple for lithium, and go up by a factor of 10 or more for cobalt and vanadium, driving up prices. The best to worst in terms of material availability are: CAES, NaS, ZnBr, PbA, PHS, Li-ion, and VRB (Barnhart, C., et al. 2013. On the importance of reducing the energetic and material demands of electrical energy storage. Energy Environment Science 2013(6): 1083–1092). ]
Lithium is a highly interesting metal, in part due to the increasing interest in lithium-ion batteries. Several recent studies have used different methods to estimate whether the lithium production can meet an increasing demand, especially from the transport sector, where lithium-ion batteries are the most likely technology for electric cars. The reserve and resource estimates of lithium vary greatly between different studies and the question whether the annual production rates of lithium can meet a growing demand is seldom adequately explained. This study presents a review and compilation of recent estimates of quantities of lithium available for exploitation and discusses the uncertainty and differences between these estimates. Also, mathematical curve fitting models are used to estimate possible future annual production rates. This estimation of possible production rates are compared to a potential increased demand of lithium if the International Energy Agency’s Blue Map Scenarios are fulfilled regarding electrification of the car fleet. We find that the availability of lithium could in fact be a problem for fulfilling this scenario if lithium-ion batteries are to be used. This indicates that other battery technologies might have to be implemented for enabling an electrification of road transports.
- Review of reserves, resources and key properties of 112 lithium deposits
- Discussions of widely diverging results from recent lithium supply estimates
- Forecasting future lithium production by resource-constrained models
- Exploring implications for future deployment of electric cars
Global transportation mainly relies on one single fossil resource, namely petroleum, which supplies 95% of the total energy . In fact, about 62% of all world oil consumption takes place in the transport sector . Oil prices have oscillated dramatically over the last few years, and the price of oil reached $100 per barrel in January 2008, before skyrocketing to nearly $150/barrel in July 2008. A dramatic price collapse followed in late 2008, but oil prices have at present time returned to over $100/barrel. Also, peak oil concerns, resulting in imminent oil production limitations, have been voiced by various studies [3–6].
It has been found that continued oil dependence is environmentally, economically and socially unsustainable .
The price uncertainty and decreasing supply might result in severe challenges for different transporters. Nygren et al.  showed that even the most optimistic oil production forecasts implied pessimistic futures for the aviation industry. Curtis  found that globalization may be undermined by peak oil’s effect on transportation costs and reliability of freight.
Barely 2% of the world electricity is used by transportation , where most of this is made up by trains, trams, and trolley buses.
A high future demand of Li for battery applications may arise if society choses to employ Li-ion technologies for a decarbonization of the road transport sector.
Batteries are at present time the second most common use, but are increasing rapidly as the use of li-ion batteries for portable electronics , as well as electric and hybrid cars, are becoming more frequent. For example, the lithium consumption for batteries in the U.S increased with 194 % from 2005 to 2010 . Relatively few academic studies have focused on the very abundance of raw materials needed to supply a potential increase in Li demand from transport sector . Lithium demand is growing and it is important to investigate whether this could lead to a shortfall in the future.
[My comment: utility scale energy storage batteries in commercial production are lithium, and if this continues, this sector alone would quickly consume all available lithium supplies: see Barnhart, C., et al. 2013. On the importance of reducing the energetic and material demands of electrical energy storage. Energy Environment Science 2013(6): 1083–1092.]
Aim of this study
Recently, a number of studies have investigated future supply prospects for lithium [13–16]. However, these studies reach widely different results in terms of available quantities, possible production trajectories, as well as expected future demand. The most striking difference is perhaps the widely different estimates for available resources and reserves, where different numbers of deposits are included and different types of resources are assessed. It has been suggested that mineral resources will be a future constraint for society , but a great deal of this debate is often spent on the concept of geological availability, which can be presented as the size of the tank. What is frequently not reflected upon is that society can only use the quantities that can be extracted at a certain pace and be delivered to consumers by mining operations, which can be described as the tap. The key concept here is that the size of the tank and the size of the tap are two fundamentally different things.
This study attempts to present a comprehensive review of known lithium deposits and their estimated quantities of lithium available for exploitation and discuss the uncertainty and differences among published studies, in order to bring clarity to the subject. The estimated reserves are then used as a constraint in a model of possible future production of lithium and the results of the model are compared to possible future demand from an electrification of the car fleet. The forecasts are based on open, public data and should be used for estimating long term growth and trends. This is not a substitute for economical short-term prognoses, but rather a complementary vision.
The United States Geological Survey (USGS) has been particularly useful for obtaining production data series, but also the Swedish Geological Survey (SGU) and the British Geological Survey (BGS) deserves honourable mention for providing useful material. Kushnir and Sandén , Tahil [19, 20] along with many other recent lithium works have also been useful. Kesler et al.  helped to provide a broad overview of general lithium geology.
Information on individual lithium deposits has been compiled from numerous sources, primarily building on the tables found in [13–16]. In addition, several specialized articles about individual deposits have been used, for instance [22–26]. Public industry reports and annual yearbooks from mining operators and lithium producers, such as SQM , Roskill  or Talison Lithium , also helped to create a holistic data base.
In this study, we collected information on global lithium deposits. Country of occurrence, deposit type, main mineral, and lithium content were gathered as well as published estimates for reserves and resources. Some deposits had detailed data available for all parameters, while others had very little information available. Widely diverging estimates for reserves and resources could sometimes be found for the same deposit, and in such cases the full interval between the minimum and maximum estimates is presented. Deposits without reserve or resource estimates are included in the data set, but do not contribute to the total. Only available data and information that could be found in the public and academic spheres were compiled in this study. It is likely that undisclosed and/or proprietary data could contribute to the world’s lithium volume but due to data availability no conclusions on to which extent could be made.
In order to properly estimate global lithium availability, and a feasible reserve estimate for modelling future production, this section presents an overview of lithium geology. Lithium is named after the Greek word “lithos” meaning “stone”, represented by the symbol Li and has the atomic number 3. Under standard conditions, lithium is the lightest metal and the least dense solid element. Lithium is a soft, silver-white metal that belongs to the alkali group of elements.
As all alkali elements, Li is highly reactive and flammable. For this reason, it never occurs freely in nature and only appears in compounds, usually ionic compounds. The nuclear properties of Li are peculiar since its nuclei verge on instability and two stable isotopes have among the lowest binding energies per nucleon of all stable nuclides. Due to this nuclear instability, lithium is less abundant in the solar system than 25 of the first 32 chemical elements .
Resources and reserves
An important frequent shortcoming in the discussion on availability of lithium is the lack of proper terminology and standardized concepts for assessing the available amounts of lithium. Published studies talk about “reserves”, “resources”, “recoverable resources”, “broad-based reserves”, “in-situ resources”, and “reserve base”.
A wide range of reporting systems minerals exist, such as NI 43-101, USGS, Crirsco, SAMREC and the JORC code, and further discussion and references concerning this can be found in Vikström . Definitions and classifications used are often similar, but not always consistent, adding to the confusion when aggregating data. Consistent definitions may be used in individual studies, but frequently figures from different methodologies are combined as there is no universal and standardized framework. In essence, published literature is a jumble of inconsistent figures. If one does not know what the numbers really mean, they are not simply useless – they are worse, since they tend to mislead.
Broadly speaking, resources are generally defined as the geologically assured quantity that is available for exploitation, while reserves are the quantity that is exploitable with current technical and socioeconomic conditions. The reserves are what are important for production, while resources are largely an academic figure with little relevance for real supply. For example, usually less than one tenth of the coal resources are considered economically recoverable [32, 33]. Kesler et al.  stress that available resources needs to be converted into reserves before they can be produced and used by society. Still, some analysts seemingly use the terms ‘resources’ and ‘reserves’ synonymously.
It should be noted that the actual reserves are dynamic and vary depending on many factors such as the available technology, economic demand, political issues and social factors. Technological improvements may increase reserves by opening new deposit types for exploitation or by lowering production costs. Deposits that have been mined for some time can increase or decrease their reserves due to difficulties with determining the ore grade and tonnage in advance . Depletion and decreasing concentrations may increase recovery costs, thus lowering reserves. Declining demand and prices may also reduce reserves, while rising prices or demand may increase them. Political decisions, legal issues or environmental policies may prohibit exploitation of certain deposits, despite the fact significant resources may be available.
For lithium, resource/reserve classifications were typically developed for solid ore deposits. However, brine – presently the main lithium source – is a fluid and commonly used definitions can be difficult to apply due to pumping complications and varying concentrations.
Houston et al.  describes the problem in detail and suggest a change in NI 43-101 to account for these problems. If better standards were available for brines then estimations could be more reliable and accurate, as discussed in Kushnir and Sandén .
Environmental aspects and policy changes can also significantly influence recoverability. Introduction of clean air requirements and public resistance to surface mining in the USA played a major role in the decreasing coal reserves .
It is entirely possible that public outcries against surface mining or concerns for the environment in lithium producing will lead to restrictions that affect the reserves. As an example, the water consumption of brine production is very high and Tahil  estimates that brine operations consume 65% of the fresh water in the Salar de Atacama region. [ The Atacama only gets 0.6 inches of rain a year ]
Regarding future developments of recoverability, Fasel and Tran  monotonously assumes that increasing lithium demand will result in more reserves being found as prices rise. So called cumulative availability curves are sometimes used to estimate how reserves will change with changing prices, displaying the estimated amount of resource against the average unit cost ranked from lowest to highest cost. This method is used by Yaksic and Tilton  to address lithium availability. This concept has its merits for describing theoretical availability, but the fact that the concept is based on average cost, not marginal cost, has been described as a major weakness, making cumulative availability curves disregard the real cost structure and has little – if any – relevance for future price and production rate .
Production and occurrence of lithium
The high reactivity of lithium makes it geochemistry complex and interesting. Lithium-minerals are generally formed in magmatic processes. The small ionic size makes it difficult for lithium to be included in early stages of mineral crystallization, and resultantly lithium remains in the molten parts where it gets enriched until it can be solidified in the final stages .
At present, over 120 lithium-containing minerals are known, but few of them contain high concentrations or are frequently occurring. Lithium can also be found in naturally occurring salt solutions as brines in dry salt lake environments. Compared to the fairly large number of lithium mineral and brine deposits, few of them are of actual or potential commercial value. Many are very small, while others are too low in grade . This chapter will briefly review the properties of those deposits and present a compilation of the known deposits.
Lithium mineral deposits
Lithium extraction from minerals is primarily done with minerals occurring in pegmatite formations. However, pegmatite is rather challenging to exploit due to its hardness in conjunction with generally problematic access to the belt-like deposits they usually occur in. Table 1 describes some typical lithium-bearing minerals and their characteristics. Australia is currently the world’s largest producer of lithium from minerals, mainly from spodumene . Petalite is commonly used for glass manufacture due to its high iron content, while lepidolite was earlier used as a lithium source but presently has lost its importance due to high fluorine content. Exploitation must generally be tailor-made for a certain mineral as they differ quite significantly in chemical composition, hardness and other properties . Table 2 presents some mineral deposits and their properties.
Recovery rates for mining typically range from 60 to 70%, although significant treatment is required for transforming the produced Li into a marketable form. For example, [40, 41] describe how lithium are produced from spodumene. The costs of acid, soda ash, and energy are a very significant part of the total production cost but may be partially alleviated by the market demand for the sodium sulphate by-products.
Lithium brine deposits
Lithium can also be found in salt lake brines that have high concentrations of mineral salts. Such brines can be reachable directly from the surface or deep underground in saline expanses located in very dry regions that allow salts to persist. High concentration lithium brine is mainly found in high altitude locations such as the Andes and south-western China. Chile, the world largest lithium producer, derives most of the production from brines located at the large salt flat of Salar de Atacama.
Lithium has similar ionic properties as magnesium since their ionic size is nearly identical; making is difficult to separate lithium from magnesium. A low Mg/Li ratio in brine means that it is easier, and therefore more economical to extract lithium.
The ratio differs significant at currently producing brine deposits and range from less than 1 to over 30 . The lithium concentration in known brine deposits is usually quite low and range from 0.017–0.15% with significant variability among the known deposits in the world (Table 3).
Exploitation of lithium brines starts with the brine being pumped from the ground into evaporation ponds. The actual evaporation is enabled by incoming solar radiation, so it is desirable for the operation to be located in sunny areas with low annual precipitation rate. The net evaporation rate determines the area of the required ponds .
It can easily take between one and two years before the final product is ready to be used, and even longer in cold and rainy areas.
The long timescales required for production can make brine deposits ill fit for sudden changes in demand. Table 3. Properties of known brine deposits in the world.
Lithium from sea water
The world’s oceans contain a wide number of metals, such as gold, lithium or uranium, dispersed at low concentrations. The mass of the world’s oceans is approximately 1.35*1012 Mt , making vast amounts of theoretical resources seemingly available. Eckhardt  and Fasel and Tran  announce that more than 2,000,000 Mt lithium is available from the seas, essentially making it an “unlimited” source given its geological abundance. Tahil  also notes that oceans have been proclaimed as an unlimited Li-source since the 1970s.
The world’s oceans and some highly saline lakes do in fact contain very large quantities of lithium, but if it will become practical and economical to produce lithium from this source is highly questionable.
For example, consider gold in sea water – in total nearly 7,000,000 Mt. This is an enormous amount compared to the cumulative world production of 0.17 Mt accumulated since the dawn of civilization . There are also several technical options available for gold extraction. However, the average gold concentration range from <0.001 to 0.005 ppb . This means that one km3 of sea water would give only 5.5 kg of gold. The gold is simply too dilute to be viable for commercial extraction and it is not surprising that all attempts to achieve success – including those of the Nobel laureate Fritz Haber – has failed to date.
Average lithium concentration in the oceans has been estimated to 0.17 ppm [14, 36]. Kushnir and Sandén  argue that it is theoretically possible to use a wide range of advanced technologies to extract lithium from seawater – just like the case for gold. However, no convincing methods have been demonstrated this far. A small scale Japanese experiment managed to produce 750 g of lithium metal from processing 4,200 m3 water with a recovery efficiency of 19.7% . This approach has been described in more detail by others [51–53].
Grosjean et al.  points to the fact that even after decades of improvement, recovery from seawater is still more than 10–30 times more costly than production from pegmatites and brines. It is evident that huge quantities of water would have to be processed to produce any significant amounts of lithium. Bardi  presents theoretical calculations on this, stating that a production volume of lithium comparable to present world production (~25 kt annually) would require 1.5*103 TWh of electrical energy for pumping through separation membranes in addition to colossal volumes of seawater. Furthermore, Tahil  estimated that a seawater processing flow equivalent to the average discharge of the River Nile – 300,000,000 m3/day or over 22 times the global petroleum industry flow of 85 million barrels per day – would only give 62 tons of lithium per day or roughly 20 kt per year. Furthermore, a significant amount of fresh water and hydrochloric acid will be required to flush out unwanted minerals (Mg, K, etc.) and extract lithium from the adsorption columns .
In summary, extraction from seawater appears not feasible and not something that should be considered viable in practice, at least not in the near future.
Estimated lithium availability
From data compilation and analysis of 112 deposits, this study concludes that 15 Mt are reasonable as a reference case for the global reserves in the near and medium term. 30 Mt is seen as a high case estimate for available lithium reserves and this number is also found in the upper range in literature. These two estimates are used as constraints in the models of future production in this study.
Estimates on world reserves and resources vary significantly among published studies. One main reason for this is likely the fact that different deposits, as well as different number of deposits, are aggregated in different studies. Many studies, such as the ones presented by the USGS, do not give explicitly state the number of deposits included and just presents aggregated figures on a national level. Even when the number and which deposits that have been used are specified, analysts can arrive to wide different estimates (Table 5). It should be noted that a trend towards increasing reserves and resources with time can generally be found, in particularly in USGS assessments. Early reports, such as Evans  or USGS , excluded several countries from the reserve estimates due to a lack of available information. This was mitigated in USGS  when reserves estimates for Argentina, Australia, and Chile have been revised based on new information from governmental and industry sources. However, there are still relatively few assessments on reserves, in particular for Russia, and it is concluded that much future work is required to handle this shortcoming. Gruber et al.  noted that 83% of global lithium resources can be found in six brine, two pegmatite and two sedimentary deposits. From our compilation, it can also be found that the distribution of global lithium reserves and resources are very uneven.
Three quarters of everything can typically be found in the ten largest deposits (Figure 1 and 2). USGS  pinpoint that 85% of the global reserves are situated in Chile and China (Figure 3) and that Chile and Australia accounted for 70 % of the world production of 28,100 tonnes in 2011 . From Table 2 and 3, one can note a significant spread in estimated reserves and resources for the deposits. This divergence is much smaller for minerals (5.6–8.2 Mt) than for brines (6.5– 29.4 Mt), probably resulting from the difficulty associated with estimating brine accumulations consistently. Evans  also points to the problem of using these frameworks on brine deposits, which are fundamentally different from solid ores. Table 5. Comparison of published lithium assessments.
One thing that may or may not have a large implication for future production is recycling. The projections presented in the production model of this study describe production of lithium from virgin materials. The total production of lithium could potentially increase significantly if high rates of recycling were implemented of the used lithium, which is mentioned in many studies.
USGS  state that recycling of lithium has been insignificant historically, but that it is increasing as the use of lithium for batteries are growing. However, the recycling of lithium from batteries is still more or less non-existent, with a collection rate of used Li-ion batteries of only about 3% . When the Li-ion batteries are in fact recycled, it is usually not the lithium that is recycled, but other more precious metals such as cobalt .
If this will change in the future is uncertain and highly dependent on future metal prices, but it is still commonly argued for and assumed that the recycling of lithium will grow significantly, very soon. Goonan  claims that recycling rates will increase from vehicle batteries in vehicles since such recycling systems already exist for lead-acid batteries. Kushnir and Sandén  argue that large automotive batteries will be technically easier to recycle than smaller batteries and also claims that economies of scale will emerge when the use for batteries for vehicles increase. According to the IEA , full recycling systems are projected to be in place sometime between 2020 and 2030. Similar assumptions are made by more or less all studies dealing with future lithium production and use for electric vehicles and Kushnir and Sandén  state that it is commonly assumed that recycling will take place, enabling recycled lithium to make up for a big part of the demand but also conclude that the future recycling rate is highly uncertain.
There are several reasons to question the probability of high recycling shares for Li-ion batteries. Kushnir and Sandén  state that lithium recycling economy is currently not good and claims that the economic conditions could decrease even more in the future. Sullivan and Gaines  argue that the Li-ion battery chemistry is complex and still evolving, thus making it difficult for the industry to develop profitable pathways. Georgi-Maschler  highlight that two established recycling processes exist for recycling Li-ion batteries, but one of them lose most of the lithium in the process of recovering the other valuable metals. Ziemann et al.  states that lithium recovery from rechargeable batteries is not efficient at present time, mainly due to the low lithium content of around 2% and the rather low price of lithium.
In this study we choose not to include recycling in the projected future supply for several reasons. In a short perspective, looking towards 2015-2020, it cannot be considered likely that any considerable amount of lithium will be recycled from batteries since it is currently not economical to do so and no proven methods to do it on a large scale industrial level appear to exist. If it becomes economical to recycle lithium from batteries it will take time to build the capacity for the recycling to take place. Also, the battery lifetime is often projected to be 10 years or more, and to expect any significant amounts of lithium to be recycled within this period of time is simply not realistic for that reason either.
The recycling capacity is expected to be far from reaching significant levels before 2025 according to Wanger . It is also important to separate the recycling rates of products to the recycled content in new products. Even if a percentage of the product is recycled at the end of the life cycle, this is no guarantee that the use of recycled content in new products will be as high. The use of Li-ion batteries is projected to grow fast. If the growth happens linearly, and high recycling rates are accomplished, recycling could start constituting a large part of the lithium demand, but if the growth happens exponentially, recycling can never keep up with the growth that has occurred during the 10 years lag during the battery lifetime. In a longer time perspective, the inclusion of recycling could be argued for with expected technological refinement, but certainties regarding technology development are highly uncertain. Still, most studies include recycling as a major part of future lithium production, which can have very large implications on the results and conclusions drawn. Kushnir and Sandén  suggest that an 80% lithium recovery rate is achievable over a medium time frame. The scenarios in Gruber et al. , assumes recycling participation rates of 90 %, 96% and 100%. In their scenario using the highest assumed recycling, the quantities of lithium needed to be mined are decreased to only about 37% of the demand. Wanger  looks at a shorter time perspective and estimates that a 40% or 100% recycling rate would reduce the lithium consumption with 10% or 25% respectively by 2030. Mohr et al.  assume that the recycling rate starts at 0%, approaching a limit of 80%, resulting in recycled lithium making up significant parts of production, but only several decades into the future. IEA  projects that full recycling systems will be in place around 2020–2030.
The impact of assumed recycling rates can indeed be very significant, and the use of this should be handled with care and be well motivated.
Future demand for lithium
To estimate whether the projected future production levels will be sufficient, it is interesting to compare possible production levels with potential future demand. The use of lithium is currently dominated by use for ceramics and glass closely followed by batteries. The current lithium demand for different markets can be seen in Figure 7. USGS  state that the lithium use in batteries have grown significantly in recent years as the use of lithium batteries in portable electronics have become increasingly common. Figure 7 (Ceramics and glass 29%, Batteries 27%, Other uses 16%, Lubrication greases 12%, Continuous casting 5%, Air treatment 4%, Polymers 3%, Primary aluminum production 2%, Pharmaceuticals 2%).
Global lithium demand for different end-use markets. Source: USGS  USGS  state that the total lithium consumption in 2011 was between 22,500 and 24,500 tonnes. This is often projected to grow, especially as the use of Li-ion batteries for electric cars could potentially increase demand significantly. This study presents a simple example of possible future demand of lithium, assuming a constant demand for other uses and demand for electric cars to grow according to a scenario of future sales of
electric cars. The current car fleet consists of about 600 million passenger cars. The sale of new passenger cars in 2011 was about 60 million cars . This existing vehicle park is almost entirely dependent on fossil fuels, primarily gasoline and diesel, but also natural gas to a smaller extent. Increasing oil prices, concerns about a possible peak in oil production and problems with anthropogenic global warming makes it desirable to move away from fossil energy dependence. As a mitigation and pathway to a fossil-fuel free mobility, cars running partially or totally on electrical energy are commonly proposed. This includes electric vehicles (EVs), hybrid vehicles (HEVs) and PHEVs (plug-in hybrid vehicles), all on the verge of large-scale commercialization and implementation. IEA  concluded that a total of 1.5 million hybrid and electric vehicles had been sold worldwide between the year 2000 and 2010.
Both the expected number of cars as well as the amount of lithium required per vehicle is important. As can be seen from Table 9, the estimates of lithium demand for PEHV and EVs differ significantly between studies. Also, some studies do not differentiate between different technical options and only gives a single Li-consumption estimate for an “electric vehicle”, for instance the 3 kg/car found by Mohr et al. . The mean values from Table 9 are found to be 4.9 kg for an EV and 1.9 kg for a PHEV.
As the battery size determines the vehicles range, it is likely that the range will continue to increase in the future, which could increase the lithium demand. On the other hand, it is also reasonable to assume that the technology will improve, thus reducing the lithium requirements. In this study a lithium demand of 160 g Li/kWh is assumed, an assumption discussed in detail by Kushnir and Sandén . It is then assumed that typical batteries capacities will be 9 kWh in a PHEV and 25 kWh in an EV. This gives a resulting lithium requirement of 1.4 kg for a PHEV and 4 kg for an EV, which is used as an estimate in this study. Many current electrified cars have a lower capacity than 24 kWh, but to become more attractive to consumers the range of the vehicles will likely have to increase, creating a need for larger batteries . It should be added that the values used are at the lower end compared to other assessments (Table 9) and should most likely not be seen as overestimates future lithium requirements.
Figure 8 shows the span of the different production forecasts up until 2050 made in this study, together with an estimated demand based on the demand staying constant on the high estimate of 2010– 2011, adding an estimated demand created by the electric car projections done by IEA . This is a very simplistic estimation future demand, but compared to the production projections it indicates that lithium availability should not be automatically disregarded as a potential issue for future electric car production. The amount of electric cars could very well be smaller or larger that this scenario, but the scenario used does not assume a complete electrification of the car fleet by 2050 and such scenarios would mean even larger demand of lithium. It is likely that lithium demand for other uses will also grow in the coming decades, why total demand might increase more that indicated here. This study does not attempt to estimate the evolution of demand for other uses, and the demand estimate for other uses can be considered a conservative one. Figure 8. The total lithium demand of a constant current lithium demand combined with growth of electric vehicles according to IEA’s blue map scenario  assuming a demand for 1.4 kg of lithium per PHEV and 4.0 kg per EV. The span of forecasted production levels range from the base case Gompertz model
Potential future production of lithium was modeled with three different production curves. In a short perspective, until 2015–2020, the three models do not differ much, but in the longer perspective the Richards and Logistic curves show a growth at a vastly higher pace than the Gompertz curve. The Richards model gives the best fit to the historic data, and lies in between the other two and might be the most likely development. A faster growth than the logistic model cannot be ruled out, but should be considered unlikely, since it usually mimics plausible free market exploitation . Other factors, such as decreased lithium concentration in mined material, economics, political and environmental problems could also limit production.
It can be debated whether this kind of forecasting should be used for short term projections, and the actual production in coming years can very well differ from our models, but it does at least indicate that lithium availability could be a potential problem in the coming decades. In a longer time perspective up to 2050, the projected lithium demand for alternative vehicles far exceeds our most optimistic production prognoses.
If 100 million alternative vehicles, as projected in IEA  are produced annually using lithium battery technology, the lithium reserves would be exhausted in just a few years, even if the production could be cranked up faster than the models in this study. This indicates that it is important that other battery technologies should be investigated as well.
It should be added that these projections do not consider potential recycling of the lithium, which is discussed further earlier in this paper. On the other hand, it appears it is highly unlikely that recycling will become common as soon as 2020, while total demand appears to potentially rise over maximum production around that date. If, when, and to what extent recycling will take place is hard to predict, although it appears more likely that high recycling rates will take place in electric cars than other uses.
Much could change before 2050. The spread between the different production curves are much larger and it is hard to estimate what happens with technology over such a long time frame. However, the Blue Map Scenario would in fact create a demand of lithium that is higher than the peak production of the logistic curve for the standard case, and close to the peak production in the high URR case.
Improved efficiency can decrease the lithium demand in the batteries, but as Kushnir and Sandén  point out, there is a minimum amount of lithium required tied to the cell voltage and chemistry of the battery.
IEA  acknowledges that technologies that are not available today must be developed to reach the Blue Map scenarios and that technology development is uncertain. This does not quite coincide with other studies claiming that lithium availability will not be a problem for production of electric cars in the future.
It is also possible that other uses will raise the demand for lithium even further. One industry that in a longer time perspective could potentially increase the demand for lithium is fusion, where lithium is used to breed tritium in the reactors. If fusion were commercialized, which currently seems highly uncertain, it would demand large volumes of lithium .
Further problems with the lithium industry are that the production and reserves are situated in a few countries (USGS  in Mt: Chile 7.5, China 3.5, Australia 0.97, Argentina 0.85, Other 0.135]. One can also note that most of the lithium is concentrated to a fairly small amount of deposits, nearly 50% of both reserves and resources can be found in Salar de Atacama alone. Kesler et al.  note that Argentina, Bolivia, Chile and China hold 70% of the brine deposits. Grosjean et al.  even points to the ABC triangle (i.e. Argentina, Bolivia and Chile) and its control of well over 40% of the world resources and raises concern for resource nationalism and monopolistic behavior. Even though Bolivia has large resources, there are many political and technical problems, such as transportation and limited amount of available fresh water, in need of solutions .
Regardless of global resource size, the high concentration of reserves and production to very few countries is not something that bode well for future supplies. The world is currently largely dependent on OPEC for oil, and that creates possibilities of political conflicts. The lithium reserves are situated in mainly two countries. It could be considered problematic for countries like the US to be dependent on Bolivia, Chile and Argentina for political reasons . Abell and Oppenheimer  discuss the absurdity in switching from dependence to dependence since resources are finite. Also, Kushnir and Sandén  discusses the problems with being dependent on a few producers, if a problem unexpectedly occurs at the production site it may not be possible to continue the production and the demand cannot be satisfied.
Although there are quite a few uncertainties with the projected production of lithium and demand for lithium for electric vehicles, this study indicates that the possible lithium production could be a limiting factor for the number of electric vehicles that can be produced, and how fast they can be produced. If large parts of the car fleet will run on electricity and rely on lithium based batteries in the coming decades, it is possible, and maybe even likely, that lithium availability will be a limiting factor.
To decrease the impact of this, as much lithium as possible must be recycled and possibly other battery technologies not relying on lithium needs to be developed. It is not certain how big the recoverable reserves of lithium are in the world and estimations in different studies differ significantly. Especially the estimations for brine need to be further investigated. Some estimates include production from seawater, making the reserves more or less infinitely large. We suggest that it is very unlikely that seawater or lakes will become a practical and economic source of lithium, mainly due to the high Mg/Li ratio and low concentrations if lithium, meaning that large quantities of water would have to be processed. Until otherwise is proved lithium reserves from seawater and lakes should not be included in the reserve estimations. Although the reserve estimates differ, this appears to have marginal impact on resulting projections of production, especially in a shorter time perspective. What are limiting are not the estimated reserves, but likely maximum annual production, which is often missed in similar studies.
If electric vehicles with li-ion batteries will be used to a very high extent, there are other problems to account for. Instead of being dependent on oil we could become dependent on lithium if li-ion batteries, with lithium reserves mainly located in two countries. It is important to plan for this to avoid bottlenecks or unnecessarily high prices. Lithium is a finite resource and the production cannot be infinitely large due to geological, technical and economical restraints. The concentration of lithium metal appears to be decreasing, which could make it more expensive and difficult to extract the lithium in the future. To enable a transition towards a car fleet based on electrical energy, other types of batteries should also be considered and a continued development of battery types using less lithium and/or other metals are encouraged. High recycling rates should also be aimed for if possible and continued investigations of recoverable resources and possible production of lithium are called for. Acknowledgements We would like to thank Steve Mohr for helpful comments and ideas. Sergey Yachenkov has our sincerest appreciation for providing assistance with translation of Russian material.