Abstract

Over the last couple of years, researchers have solved several important problems associated with lithium-air batteries, including cathode clogging, anode corrosion and low voltage efficiency. Porous carbon cathodes have been found to reduce clogging and thus significantly increase the charge capacity of the battery significantly. Researchers have also developed additional electrolytes and oxygen-selective membranes that prevent corrosion by stopping water from reaching the anode, while still being permeable to oxygen. Testing of various catalysts, including bifunctional ones, has allowed both charge and discharge voltages to approach theoretical values. However, Li-air batteries still function poorly in real-life conditions, and further research is therefore required before they can be used in electric vehicles. This review describes the progress that has been made so far as well as the remaining challenges.

 

1. Introduction

Pollution from personal motor vehicles accounts for about 10% of the total CO₂ emissions from fossil fuels in the world (EDF 2006), despite the efficiency advances that have been made in internal combustion engines in recent decades. Therefore, there is a significant environmental value associated with driving electric vehicles (EVs) and non-plug-in hybrids, and policy-makers intend to increase the proportions of these vehicles in the US, which are currently only 0.1% and 3% for EVs and non-plug-in hybrids, respectively (Valdes-Dapena 2012). Today, almost every major car company has at least one electric model, and the Department of Energy (2011) estimates that there will be one million EVs on American roads by 2015. Furthermore, Pike Research predicts that EVs and hybrids will exceed 5% of the total number of cars sold in the US by 2017 (Martin 2011). However, in order for this to happen, car producers must make EVs attractive to customers so that the demand matches the increase in supply.

At this point, many drivers are reluctant to change to the greener alternative due to significant downsides associated with substituting EVs for fuelled cars, particularly low driving range and poor performance after many charges. Although the introduction of lithium-ion batteries in cars has improved both of these aspects greatly, EVs are still far behind cars powered solely by fossil fuels. The 2013 Honda Fit EV, which has received the highest efficiency rating from the Environmental Protection Agency, has an estimated driving range of 82 miles per charge from its 20 kilowatt-hour (kWh) Li-ion battery (VerticalNews 2012). Furthermore, the Nissan Leaf, which the Department of Energy (2011) predicts to be the most-sold EV in the US throughout the next five years, has a 73-mile range (VerticalNews 2012). Compared to one of the most efficient non-electric cars, such as the Volkswagen Passat TDI, which has a fuel range of almost 800 miles per tank (Ganz 2012), this is undoubtedly far from optimal. The difference is further amplified by the fact that the Honda and the Nissan must be charged for about three hours, while a fuel tank can be filled in a couple of minutes.

Since the Honda Fit EV is considered to be state-of-the-art today, it is quite extraordinary that IBM is designing an EV with a range of 500 miles per charge, which is more than 6 times the capacity of the Honda. Ultimately, this whole project comes down to the power source that IBM uses—the lithium-air battery—and particularly its energy density. In contrast to Li-ion batteries, which have heavy cathodes made of transition metals such as cobalt, manganese and nickel, IBM’s battery is based on the highly energetic reaction between lithium and oxygen and therefore does not need such a cathode. Much lighter than transition metals, porous carbon is used to trap the oxygen and facilitate the reaction at the cathode. Consequently, the Li-air battery has a theoretical energy density of 5200 watt hours per kilogram (Wh/kg), or 11,140 Wh/kg, ignoring oxygen mass (Abraham and Jiang 1996), in contrast to the typical value of 210 Wh/kg for a Li-ion battery (Zu et al. 2011). The idea of creating such a battery was introduced in 1978, but it was not put into practice for another 18 years, when Abraham and Jiang (1996) created a functional Li-air battery with a non-aqueous electrolyte (1996).

Although this accomplishment was considered a major breakthrough for not only Li-air battery development but also green technology as a whole, several problems remained to be solved at this point. First of all, reaction products clog the pores of the carbon cathode and thus limited the discharge reactions and the extent to which electrons can move from one half-cell to the other (Kuboki et al. 2005). Furthermore, the lifetime of the battery is in many cases limited by corrosion of the anode (Crowther et al. 2011). Lastly, experiments have shown that the voltage of a Li-air battery is significantly lower than the theoretical value due to high overpotentials (Debart et al. 2007). Throughout the last couple of years, however, major advances in materials science and nanotechnology have facilitated the development of solutions to these problems. In this review, the solutions to cathode clogging, anode corrosion and low voltage efficiency will be examined through extensive analysis of the most recent research on the topic, in order to identify trends and the remaining challenges.

 

2. Clogging of the cathode

As the discharge products of a lithium-air battery, LiO₂ and Li₂O₂, are not soluble in a non-aqueous electrolyte, they clog the pores of the carbon cathode (Kuboki et al. 2005). This prevents new oxygen molecules from being reduced, and thus substantially lowers the charge capacity of the battery (Sandhu et al. 2007). However, Read (2002) showed that this phenomenon could be prevented by optimizing the wetting of the cathode. Strong adhesive forces between a Super-P carbon black cathode and a diethyl carbonate electrolyte caused a high contact area that allowed for a charge capacity of 2120 milliamp hours per gram (mAh/g) (Read 2002). The total surface area of the pores, however, was not found to have a significant impact on charge capacity (Read 2002). This discovery was later confirmed by Kuboki et al. (2005), who observed that pore volume was a more important factor and improved the charge density to 2433 mAh/g.

Though impressive at the time they were measured, these charge densities were not even near the theoretical potential. Supported by a more thorough understanding of the cathode on the nanometer scale, Xiao et al. (2011) discovered the advantages of using a hierarchically porous graphene cathode (2011). This material combines the discoveries made by Kuboki et al. (2005) with channels between the pores, and thus facilitates the diffusion of oxygen (Xiao et al. 2011). In addition, Xiao et al. (2011) observed that defect sites on the graphene lattice made it energetically favorable for discharge molecules to deposit individually rather than in clusters. In this way, the problems of clogging were further reduced, allowing Xio et al. (2011) to reach a charge capacity of 15,000 mAh/g (2011), in contrast to the theoretical capacity of the common LiFePO₄-cathode Li-ion battery of only 170 mAh/g (Takahashi et al. 2002).

For a long time, researchers believed that clogging reduced charge capacity due to inhibition of oxygen diffusion alone. However, Viswanathan et al. (2011) suggested that reduced charge transportation was just as significant, based on theoretical calculations and redox reaction experiments involving electron tunneling through a Li₂O₂-film. With this realization and an improved understanding of the discharge product distribution on the cathode, which was mapped by Nanda et al. (2012) through the use of neutron imaging, Li et al. (2012) proposed the use of a sulphur-doped graphene cathode. This proved to be highly effective, because such a cathode causes the discharge products to form nanorods through which charges can tunnel (Li et al. 2012).

 

3. Corrosion of the anode 

A more fundamental way of preventing clogging is to simply use an aqueous electrolyte, but this causes hydrolysis of the lithium anode. As lithium is a highly reactive metal, it reacts with the water in the electrolyte, which significantly shortens the lifetime of the battery (Kuboki et al. 2005). Ding et al. (2012) showed that this effect could be prevented by separating the aqueous electrolyte and the anode using an additional (solid state) electrolyte, such as LATP glass (Li_1+x+yAl_xTi_2−xSi_yP_3−yO₁₂), that facilitates Li-ion diffusion while blocking out water. However, lithium is unstable when in contact with LATP glass (Ding et al. 2012, Zhang J et al. 2010). Visco et al. (2007) solved this problem by introducing another boundary layer, this time between the glass and the anode.

In order to evaluate this setup, He et al. (2012) measured its stability in different conditions and reported that a relatively high temperature and a 0.5M to 1.0M LiOH electrolyte were suitable choices for such a battery. However, it appears that He et al. (2012) failed to test the long-term effects of such conditions (the time range is not mentioned under “Experimental”). Shimonishia et al. (2011) simultaneously published results showing that a much less alkaline electrolyte of 0.057M LiOH caused significant LATP degradation in the form of a conductivity decrease after only three weeks. These results are supported by Thokchom and Kumar’s observations (2007) of how P-O-P bonds in LATP are readily hydrolyzed. Thokchom and Kumar (2007) also found that LATP hydrolysis could be prevented by adding metal groups and thus replacing P-O-P with P-O-M. Furthermore, Shimonishia et al. (2011) suggested the addition of LiCl, since an increase in the concentration of Li⁺ reduces the dissociation of LiOH and thus makes the electrolyte less alkaline. This method proved to reduce corrosion of the LATP glass significantly, which is consistent with observations made by both Hasegawa et al. (2009) and Ding et al. (2012).

Unfortunately, the water in humid, ambient air can also cause hydrolysis of the anode. During the last couple of years, however, researchers have developed a solution to this problem that involves entirely covering the primary Li-air cells with oxygen-selective membranes, which are impermeable to water due to high hydrophobicity but allow O₂ to flow into the cell (Crowther et al. 2012, Zhang J et al. 2010). Zhang J et al. (2010) created such a membrane by laminating a 50μm-thick porous nickel sheet with a Teflon film, and found that the battery was operative for only 21 days in a relative humidity (RH) of 20%. The authors suggested that the relatively weak O₂-H₂O-selectivity was due to Knudsen-diffusion, meaning that mechanisms on the nanometer-scale—which are known to favor H₂O-diffusion—dominate over those that apply to the continuum-model. Nevertheless, Crowther et al. (2012) reported good results for fiberglass cloth coated in the same material, which suggests that structural factors that are not material-specific, such as pore size and thickness, can play a key role. They reported that their membrane allowed the battery to be operative for 40 days in 20% RH and also slowed down evaporation of the relatively volatile LiBF₄ electrolyte (2012). The latter effect was also observed by Zhang JG et al. (2010), who constructed their membrane from a different organic polymer (2010).

Zhang JG et al. (2010) found that the low gas permeability required to inhibit evaporation of the electrolyte unfortunately also prevented the battery from operating at levels of high oxygen demand (2010). Crowther et al. (2011) suggested avoiding this problem by constructing the membrane out of silicone rubber, a polymer used in contact lenses that is known for being extremely oxygen permeable.  Not only does this material completely prevent corrosion of the anode, but it also lets in almost the same amount of oxygen as when no membrane is used (Crowther et al. 2011).

 

4. Voltage efficiency

Another problem that has prevented Li-air batteries from revolutionizing green transportation so far is relatively low voltage efficiency. While the theoretical discharge voltages are 3.10V and 4.27V for non-aqueous and aqueous electrolytes, respectively (Abraham and Jiang 1996), internal resistance and unwanted reactions in the battery make the actual values lower. A significant problem is that the water in aqueous electrolytes is discharged as well, causing the battery to have a voltage value closer to that of H₂O (2.22 V) than that of the targeted reaction (He et al. 2011). For instance, Debart et al. (2007) measured a discharge voltage of 2.6V, which is considerably lower than the charge voltage of 4.8V. However, when using a manganese dioxide catalyst, they measured a charge voltage of 4.3V.

He et al. (2011) lowered the charge voltage even further, to 4.09 V, by adding LiOH to the electrolyte, making it more ionic and thereby lowering its internal resistance. Nevertheless, oxygen solubility decreases with alkalinity (Zhang C et al. 2009), so the use of LiOH is most successful when accompanied by a pH-regulating compound, such as LiClO₄ (He et al. 2011). When applying both this idea and a platinum (Pt) catalyst, He et al. (2011) achieved discharge and charge voltages of 3.32V and 3.90V, respectively. While He et al. (2011) used a 50 wt. % metal cathode, Yoo et al. (2012) measured a discharge voltage value of 3.75V when using 20 wt. %. Platinum catalysts have also been used in combination with gold (Au) in order to achieve bi-functional catalytic activity, meaning that the two elements each affect one reaction (Lu et al. 2010). The gold catalyzes the discharge reaction, whereas the platinum enhances the kinetics of the charging reaction, which gives a voltage efficiency of 77% (Lu et al. 2010). Yin et al. (2012) later developed the idea further by testing several types of this catalyst, and found that Au₂₂Pt₇₈ gave the lowest difference between charge and discharge voltage.

However, as Maass et al. (2008) point out, these catalysts make the cathode unstable, and researchers have therefore suggested using other compounds (Yang et al. 2012, Yoo et al. 2012). Since perovskites can provide a high amount of electrons (Suntivich et al. 2011), Yang et al. (2012) proposed the use of a copper perovskite as a bifunctional catalyst and measured a discharge voltage of 3.2V. Furthermore, Yoo et al. (2012) achieved a value of 3.50V when using nitrogen-doped graphene nanosheets under acidic conditions.

 

5. Conclusion 

With an energy density much higher than that of Li-ion batteries, the Li-air battery is a promising technological invention that can potentially revolutionize the EV industry. Although there were initially many problems associated with Li-air batteries, extensive research has brought the theoretical idea much closer to reality over the last couple of years. Improved knowledge of nanotechnology has facilitated the development of porous cathodes with properties that significantly reduce clogging, and researchers have almost completely solved the problems with anode hydrolysis through the use of effective diffusion membranes. Furthermore, testing of various catalysts has brought the voltage efficiency of Li-air batteries significantly closer to the theoretical potential. Nevertheless, most experiments have been conducted under optimal conditions and over relatively short periods of time. Therefore, the actual lifetime of the Li-air battery under more realistic conditions is currently on the order of only a month, which is certainly insufficient for use in a vehicle, especially since this is measured in only 20% relative humidity. In the future, more effort must be made to test and develop the Li-air battery in real-life conditions, so that this promising energy source can be brought from the laboratory to the car engine.

 

References