Lithium-Air Battery: Promise and Challenges

Lithium-Air Battery: Promise and Challenges

ABSTRACT The lithium-air system captured worldwide attention in 2009 as a possible battery for electric vehicle propulsion applications. If successfully developed, this battery could provide an energy source for electric vehicles rivaling that of gasoline in termsof usable energydensity. However, there arenumerous scientific and technical challenges that must be overcome if this alluring promise is to turn into reality. The fundamental battery chemistry during discharge is thought to be the electrochemical oxidationof lithiummetal at theanodeand reductionofoxygen fromairat thecathode. With aprotic electrolytes, as used in Li-ion batteries, there is some evidence that the process canbereversedbyapplyinganexternalpotential, i.e., that suchabatterycanbe electrically recharged. This paper summarizes the authors’ view of the promise and challenges facing development of practical Li-air batteries and the current under- standing of its chemistry. However, it must be appreciated that this perspective represents only a snapshot in a very rapidly evolving picture.

M ankind’s total power consumption is currently 14 TWand is projected to roughly triple by 2050.1At present,oil represents 34%of theworld’s total primaryenergy source. It accounts for 40% of the total CO2 emission and is a major cause of geopolitical instability.1 Since themajority of oil is used for automobile and light truck applications, a transition to an electrified road transportation system should be a societal goal of utmost importance.2 This is already beginning with the advent of hybrid electric vehicles (EVs), and will accelerate as plug-in hybrid vehicles and ultimately pure EVs are developed.

Themajor technical hurdle for the complete electrification of road transportation is the insufficient storage capacity of current batteries, severely limiting the range of practical EVs. The Li-air battery was initially proposed in the 1970s for automotive appli- cations.3 In the last year, that interest in Li-air has grown sharply, resulting in over 14 research articles published in the first quarter of 2010 alone. The Li-air battery potentially has much higher gravimetric energy storage density compared to all other battery chemistries, and this has led to strong interest in whether such batteries could be developed for powering EVs, enabling driving ranges comparable to gasoline powered automobiles.4

The Battery 500 Project by IBM and its partners, as well as many other research groups have initiated research programs on Li-air batteries to evaluate their potential as batteries for automotive propulsion applications. The “500” stands for a target range of 500 miles/800 km per charge, which translates into a battery capacity of about 125 kWh at an average use of 250Wh/mile for a standard family car. Li-air batteriesmayalso play important roles in other areas, such as powering laptops and remote sensors or for robotic uses. However, the require- ments of automotive applications shape the thrust of most researchprograms. At this stage, research focuses principally on determining the basic scientific principles underlying the opera- tion of Li-O2 batteries (rather than Li-air). Section II discusses

system level requirements for a successful EV application battery. Section III describes various architectural (chemical) approaches for Li-air batteries and compares key trade-offs. Section IV describes current experimental and theoretical studies of the aprotic Li-O2 battery, with emphasis on the air cathode. Section V discusses the anode, while section VI dis- cusses the status and challenges of using air rather than O2 as the reactant. Section VII summarizes the research we believe needs to be performed to develop practical Li-air batteries.

System Level Requirements for Automotive Propulsion and the Li-Air Battery. The key criteria for practical EV propulsion batteries are energy density, cost, lifetime (measured in years and miles), and safety.

Gravimetric and Volumetric Energy Density: The energy density of gasoline is 13000 Wh/kg, which is shown as its theoretical energy density in Figure 1. The average tank-to- wheel efficiency of the U.S. fleet is 12.6%; therefore the usable energy density of gasoline for automotive applications is approximately 1700 Wh/kg.1 This is shown as the “practical”

The key criteria for practical EV propulsion batteries are energy

density, cost, lifetime (measured in years and miles), and safety.

Received Date: April 27, 2010 Accepted Date: June 24, 2010

D ow

nl oa

de d

vi a

U N

IV O

F D

A Y

T O

N o

n Ju

ne 2

5, 2

01 8

at 1

8: 01

:3 2

(U T

C ).

Se

e ht

tp s:

//p ub

s. ac

s. or

g/ sh

ar in

gg ui

de lin

es f

or o

pt io

ns o

n ho

w to

le gi

tim at

el y

sh ar

e pu

bl is

he d

ar tic

le s.

r 2010 American Chemical Society 2194 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

energy density for gasoline in Figure 1. Since the efficiency of electric propulsion systems (battery-to-wheels) are about 90%, a 10-fold improvement of the current energy densities of Li-ion batteries, which are typically between 100 and 200Wh/kg (cell level), would bring electric propulsion systems on-par with gasoline, at least as measured by gravimetric energy density. However, there is no expectation that current batteries such as Li-ion will ever come close to the target of 1700 Wh/kg. New chemistries are required to achieve this goal.

The oxidation of 1 kg of lithiummetal releases 11680Wh/kg, not much lower than that of gasoline. This is shown as the theo- retical energy density of a Li-air battery in Figure 1. However, practical energy densities for Li-air batteries will be far less. Existingmetal-air batteries, such as Zn/air, typically have a practi- cal energy density of about 40-50%of their theoretical density. However, onecan safelyassume that even fullydevelopedLi-air cellswill never achieve such an excellent ratio, because lithium is very light, and therefore the overhead of the battery structure, electrolytes, and so forth will have a much larger impact.

Fortunately, an energy density of 1700 Wh/kg for the fully charged battery corresponds only to 14.5% of the theoretical energy content of lithium metal. It is not inconceivable that such an energy density, at the cell level, may be achievable, given intensive and long-termdevelopmentwork. The energy densityof the complete battery systemmaybeonlyhalf of the density realized at the cell level.

The volumetric energy density (measured in Wh/L) of the battery is also an important design consideration. This re- quirement is best captured by allocating a maximum volume of300L (familycar) for thebatteryand its auxiliary systems.A driving range of 500 miles (800 km) requires 125 kWh capacity (at 250 Wh/mile), thus a 300 L volume limit dictates that the specific gravity of the battery, including the space allocated for air channels inside the battery and the air- handling system, must not be less than 0.5 kg/L. Note that

the specific air flow (kg air per kWof power generated) into a Li-air battery, assuming the exhaust air from the battery had an oxygen content of 17%, is comparable to the specific airflow of internal combustion engines.

Power Density and Cost: While Li-air systems offer the promise of very high energy densities, their power density (measured in W per kg of battery mass) is currently very low. During discharge, oxygen is thought to be reduced in the presence of lithium ions to form lithium oxides, and during charge the chemical reaction reverses to evolve oxygen gas. Both reactions occur at the surfaces of the cathode. Thus very large internal surface areas, both in a microscopic and a macroscopic sense, are required. Prototype aprotic Li-air cells deliver current densities in the order of 1 mA/cm2. It will be critical to increase this current density by at least 1 order of magnitude. Even then, the macroscopic surface area to supply the total power for a propulsion battery is very large. For example, a battery with 100 kW power output at a cell voltage of 2.5 Vand a current density of 25 mA/cm2 will require a total internal surface area of 160 m2, equal to the internal surface of the human lung.

One way to reduce the power density requirements is to create an electric hybrid propulsion systemwhere a small capa- city but high power battery, for example, based on Li-ion technology, provides power for short periods of high demand, such as during acceleration. This is possible since the ratio of average-to-peak power demand in a car is only about 1:10.

Electrical Energy Efficiency: Current Li-air cells exhibit large overvoltages, i.e., the charging voltage is considerably higher than the discharge voltage. This corresponds to a low cycle electrical energy efficiency, currently on the order of 60-70%. Practical propulsion batteries should exhibit “round-trip” energy efficiencies of 90%. The detailed me- chanisms underlying these high over voltages are currently not understood, but hopefully can be substantially reduced by the choice of catalysts.

Liftetime and Cyclability: Current Li-air cells have been demonstrated with up to about 50 cycles with only moderate loss in capacity.5 Therefore, future research efforts need to focus on improving the capacity retention during cycling. Nevertheless, the total number of charge cycles of large propulsion batteries would not necessarily need to be high given the large energy capacity of Li-air cells. For example, a battery designed for a lifetime of 150000 miles and support- ing a 500mile rangewill need to be recharged only 300 times (full cycle equivalent). But many tons of air will have to pass through a battery during its lifetime, and even minute accu- mulations of moisture (for aprotic batteries) or products of side reactions will be detrimental.

Safety: The batteries of EVs will be held initially to extre- mely high safety standards, much higher than gasoline cars. Typical thermal runaway of a Li-ion battery due to overcharg- ing or internal shorts is not a possibility in Li-air batteries because of the rate-limited surface nature of the reaction, i.e., the reactant O2 is not stored in the battery. However, there are two other safety concerns to be considered. First, the desired, though not mandatory, use of lithium metal anodes is a well-known safety problem, since lithiummetal tends to form dendrites, which can short-circuit the battery and react

Figure 1. The gravimetric energy densities (Wh/kg) for various types of rechargeable batteries compared to gasoline. The theore- tical density is based strictly on thermodynamics and is shown as the blue bars while the practical achievable density is indicated by the orange bars and numerical values. For Li-air, the practical value is just an estimate. For gasoline, the practical value includes the average tank-to-wheel efficiency of cars.

r 2010 American Chemical Society 2195 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

aggressivelywithmany contaminants. Second, the presumed dominant reaction product of aprotic cells is Li2O2, which is a strong oxidizer. Combined with an organic electrolyte, this could lead to safety issues in an accident. However, preliminary experiments at IBM indicate that no thermal exothermic reactions between Li2O2 and common electro- lytes occur at temperatures below themelting point of lithium metal (180 �C). This safety concern does not exist in aqueous cells.

In conclusion, the requirements for large capacity auto- motive propulsion batteries are extensive, but quite well- defined. They will serve as guidelines for the research to be carried out on Li-air systems. At present, automotive propul- sion batteries are just beginning the transition from nickel metal hydride to Li-ion batteries, after nearly 35 years of research and development on the latter. The transition to Li-air batteries (if successful) should be viewed in terms of a similar development cycle.

Architectures of Li-Air Batteries. Currently, four chemical architectures are being pursued worldwide, which are outlined in Figure 2. These include threeversionswith liquid electrolytes: a fully aprotic liquid electrolyte, an aqueous electrolyte, and a mixed system with an aqueous electrolyte immersing the cathode and an aprotic electrolyte immersing the anode. The fourth approach is an all-solid-state battery with a solid electro- lyte. The fundamental electrochemistry depends upon the electrolyte around the cathode.

In an aprotic electrolyte, the fundamental cathode dis- charge reactions are thought to be

2LiþO2 f Li2O2 and possibly

2Liþð1=2ÞO2 f Li2O In an aqueous cathode electrolyte, the fundamental reac-

tions are

2Liþð1=2ÞO2 þ 2Hþ f 2Liþ þH2O ðacidic mediaÞ 2Liþð1=2ÞO2 þH2O f 2LiOH ðalkaline mediaÞ

The battery chemistry for the all-solid-state battery is not clear at this point in time, but is presumably the same as that for the aprotic electrolyte. The aprotic version was first investigated by Abraham and Jiang6 and further developed principally by Read et al.7 and Bruce et al.5,8,9 The fully aqueous version is being developed by Polyplus,10 while the mixed electrolyte configuration is also being developed by Polyplus10 and Wang and Zhou.11 A fully solid-state Li-air battery was recently reported by Kumar et al.12 Each config- uration has specific advantages and presents definite scien- tific and engineering challenges, so that the ultimate choice for the best configuration is still an open question. A brief preview of some of these issues is given below.While Figure 2 and much of the discussion here assumes lithium metal as the anode, alloy anodes are also an option for Li-air batteries and are discussed briefly in section V.

In the aprotic liquid version (see Figure 2), the lithium anode is in contact with the electrolyte and forms a stable solid electrolyte interface (SEI). At the air-breathing cathode (currently porous carbon cathodes similar to polymer electro- lytemembrane (PEM) fuel cell cathodes), insoluble Li2O2 (and perhaps Li2O) is thought to be formed via the oxygen reduc- tion reaction (ORR). There is some evidence that, with catalysts present, Li2O2 will undergo the oxygen evolution reaction (OER) at sufficiently high applied recharge voltages5

so that the aprotic configuration could be the basis for an electrically rechargeable Li-air battery. This will be discussed in detail later in this article. Both for the aqueous electrolyte configuration and the mixed electrolyte configuration, the cathode chemistry is identical, and there is no evidence at this time that the electrochemical reaction is reversible, except by mechanically removing the reaction products and replacing themwith fresh reactants. Thegreat advantageof the aqueous or mixed electrolyte configuration is that the discharge reac- tion product is soluble in H2O, eliminating the cathode clogg- ing, volume expansion, and electrical conductivity issues of the aprotic architecture.13 A difficult challenge for aque- ous andmixed systems is thedevelopmentof goodLi-ion con- ducting membranes, which protect the anode from reacting vigorously with H2O (see section V). For the aqueous system, this requires the development of a lithium metal and water- stable artificial SEIs (see Figures 2 and 3). In the mixed electrolyte system, an aprotic electrolyte is in direct contact with the lithium metal anode so that a natural SEI is formed on the lithium metal. This minimizes the difficult require- ment that the membrane protecting the lithium metal from H2O is also stable against reduction by lithium metal. Visco et al. have demonstrated an aprotic Li-air battery with an artificial ceramic SEI to protect the anode.10 This cell could be cycled over 60 times at 0.4 mA/cm2 in air with 50% relative humidity. The protection of lithium metal from contamination is discussed in section V. All four configurations will ultimately have to solve the difficult problem of developing a high throughput air-breathing system that passes O2 and keeps out environmental contaminants (e.g., H2O, CO2, N2). Because only the aprotic configuration of a Li-air battery has shown any promise of electrical rechargeability, this configuration has attracted the most effort worldwide to date, and we focus

Automotive propulsion batteries are just beginning the transition

from nickel metal hydride to Li-ion batteries, after nearly 35 years of research and development on the latter. The transition to Li-air bat- teries (if successful) should be

viewed in terms of a similar devel- opment cycle.

r 2010 American Chemical Society 2196 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

principally on this configuration for the remainder of this perspective.

Aprotic Li-Air Battery. A typical design for aprotic Li-air batteries is shown in Figures 2 and 4. It is composed of a metallic lithium anode, an electrolyte comprising a dissolved lithium salt in an aprotic solvent, and a porous O2-breathing cathode composed of large surface area carbon particles and catalyst particles, both bound to a metal mesh using a binder.

The chemistry proposed for the aprotic Li-air battery is as follows.6 During the discharge of the cell, an oxidation reaction occurs at the anode (Lif Liþþ e-). The electrons flow through an external circuit and the lithium ions generated from this reaction reduceoxygen to formLi2O2 (andpossibly Li2O) in the cathode. The standard potential for the discharge reactionU0 is given by the thermodynamics of the reaction as U0 = 2.96 V, using the Nernst equation. At externally applied potentials

(U > U0), the reaction above is thought to be reversed, i.e., lithiummetal is plated out on the anode, andO2 is evolved (i.e., generated) at the cathode (Figure 4).The simplest net reaction

Figure 3. Schematic drawingof the lithiummetal-electrolyte inter- face choices. Both the complicated natural SEI formed by reduction of the electrolyte and anartificial SEI, e.g., Li-ion-conducting ceramic, are shown as examples. Adapted from refs 14 and 15.

Figure 2. Fourdifferent architectures of Li-air batteries,which all assume the use of lithiummetal as the anode. The three liquid electrolyte architectures are aprotic, aqueous, and amixed aprotic-aqueous system. In addition, a fully solid state architecture is also given. Principal components are as labeled in the figure. Spontaneously occurring SEIs on the lithium anode are given as dashed lines, while artificial SEIs are given as solid lines.

All four configurations will ulti- mately have to solve the difficult problem of developing a high

throughput air-breathing system that passes O2 and keeps out environmental contaminants.

r 2010 American Chemical Society 2197 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

envisioned for the aprotic battery is a two-electron process: 2LiþO2 f Li2O2 ðdischargeÞ ð1Þ

Li2O2 f 2LiþO2 ðchargeÞ ð2Þ The lithiumanode is in contactwith theelectrolyte and forms

a stable SEI,which protects themetal from further reactionwith the electrolyte. This is quite similar to the formation and proper- ties of an SEI at carbon-lithium anodes in conventional Li-ion batteries.16Examplesofaprotic electrolytesused todate include organic carbonates (ethylene carbonate, propylene carbonate, dimethyl carbonate), ethers (tetrahydrofuran (THF), dioxolane), andesters (γ-butyrolactone),which solvate lithiumsalts, suchas LiPF6, LiAsF6, LiN(SO2CF3)2, andLiSO3CF3andhavehighoxida- tive stabilities.17

A typical discharge-charge cycle of a Li-O2 cell is shown in Figure 5, and is quite similar to others in the literature.5 The cell consisted of a lithium metal foil as the anode, a 250 μm thick glass mat fiber separator, and a porous cathode constructed from high surface area Super P (Super P is a conductive carbon black and is a product of TIMCAL Graphite & Carbon) carbon particles mixed with R-MnO2 nanorods serving as a catalyst, both uniformly distributed and bound to a 1.6 mm-thick metallic nickel foam current collector with PVDF binder.

The composition of the active cathode materials by weight percentwere carbon:R-MnO2:PVDF=54:10:36. The separator and cathode were flooded with 1 M LiN(SO2CF3)2 [LiTFSI] in propylene carbonate, which served as the electrolyte. The cell construction was a spring loaded Swagelok design with active electrode areas of 1.2 cm2. The initial opencircuit voltage (OCV) of this cell is ∼3.3 V, which is significantly higher than U0. This high OCV likely reflects a mixed potential effect due to inter- calation of Li-ions into the R-MnO2 particles. This process terminates quickly as the cell is discharged because there is notmuchR-MnO2 present.Without the catalyst in the cathode, the OCV is ∼3.1 V, still slightly higher than U0 by 0.15 V. The working voltage of this cell during discharge is approximately 2.6-2.7 V, which is significantly less than U0. This difference is called the discharge overpotential ηdis. The discharge capacity of this Li-O2 cell is also much lower than its theoretical capacity given by the void volume of the cathode, ∼1%. This is most

likely due to the insolubility of discharge products in the nonaqueous organic electrolyte, which causes a buildup of discharge products on the cathode surface, inhibiting the flow of reactants (O2, Li

þ, and electrons) to the active surface. During constant-current charging of the cell, the voltage

increases to approximately 4.0 V (Figure 5). Hence the charge overpotential (ηchg) is significantly greater than the discharge overpotential (ηdis). Thus, at present, the electrical energy effi- ciency for a discharge-charge cycle is only 2.6 V/4 V=65%. It is hoped that better electrocatalysts can reduce overpotentials and increase the electrical and thermodynamic efficiency. Debart et al.8 showed at very low current densities that the addition of a relatively small mole percent of various transition metal oxide particles enhances the discharge capacity, reduces ηchg, and yet has a minimal effect on ηdis. Lu, et al.

18 showed even more dramatic positive effects of Au nanoparticles on ηdis and of Pt nanoparticles onηchg. It seems highly unlikely that catalysis, i.e., the lowering of activation barriers in the electro- chemical reaction steps, is the root cause of the observed increase in discharge capacity. However, it is possible that the lowering of ηchg and ηdis are true electrocatalytic effects (although this is difficult to rationalize if insoluble reaction products rapidly coat the catalyst particles). Hopefully, in the future, polarization studies, i.e.,measuringη(i) or i(η) can probe the nature of the electrocatalysis in quantitative detail. Prelimin- ary polarization studies at IBM confirm the existence of a moderate catalytic effect for R-MnO2 in the charge current. However, work is still in progress to determinewhetherR-MnO2 is in fact catalyzing the charging reaction shown in eq 2, or catalyzing some deleterious side reactions (see below).

There is limited direct experimental evidence that the discharge current in aprotic cells is due predominately to eq 1. The only spectroscopic evidence is the identification of the Li2O2 product by its Raman spectrum,

6,9 although this did

Figure 4. Schematic operation proposed for the rechargeable aprotic Li-air battery. During discharge, the spontaneous electro- chemical reaction 2LiþO2f Li2O2 generates a voltage of 2.96 Vat equilibrium (but practically somewhat less due to overpotentials). During charge, an applied voltage larger than 2.96 V (∼4 V is required due to overpotentials) drives the reverse electrochemical reaction Li2O2 f 2Li þ O2.

Figure 5. A single measured discharge-charge cycle for an aprotic Li-O2 cell (based on SP carbon) operated at ∼0.1 mA/cm2 current density. This gives the cell output voltage for discharge or the nece- ssary applied potential for charge at the given current density as a function of the charge per gram of carbon in the cathode (mAh g-1). Details of this cell are given in the text. Discharge and charge directions are given by the labeled arrows in the figure. The thermo- dynamic potential is given by the dashed line, and the overpotentials for discharge ηdis and charge ηchg are indicated by the arrows.

r 2010 American Chemical Society 2198 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

not allow quantitative comparison to the current. There is, however, some indirect evidence for this, i.e., that the weight gain of the cathode is consistent with what would be expected from eq 1, based on coulometric measurements of the total charge drawn from the battery5 and that the O2 consumption from the gas phase is also consistent with coulometry derived from eq 1.19 At higher discharge currents, however, there is indirect evidence from reduced O2 consumption compared to the coulometry that some Li2O must be formed as well.

19

The identification of the charge reaction Li2O2f 2Liþ O2 as the sourceof the charge current is even lessdefinitive at this point in time. O2 is evolved during recharge of a cathode packed artificially with Li2O2 particles, and X-ray diffraction studies show the disappearance of the Li2O2 from the artifi- cially packed cathode.5 In addition, scanning electron micro- scope (SEM) studies show the appearance of a white deposit, presumably Li2O2, during discharge and its disappearance during charge.6,9 While these results are certainly suggestive that the charge current after discharge of a battery is domi- nated by eq 2, this is far from a quantitatively settled issue.

Recent differential electrochemical mass spectrometry (DEMS) experiments at IBM show that, while O2 is evolved during charging of a discharged battery (especially in the presence of some catalysts), significant oxidation products (e.g., CO2) are also evolved.

20 At the high potentials required for charging, electrolyte oxidation and possibly even carbon oxidation are likely and could contribute significantly to the charging currents observed. This result also implies that the electrochemical stability of both the electrolyte and the cathode itself are key challenges in the development of practical Li-air systems. In fact, we believe that electrolyte redox stability in contact with reaction products (and intermediates) is a key unresolved issue. Organic carbonate-based solvents such as propylene carbonate or other carbonate mixtures have limited anodic oxidation stability and are prone to oxidation at lower onset potentials by typical catalytically active materials. This is very relevant for the Li-air battery system since the oxidative stability of the electrolyte could be reduced in the presence of catalysts or even Li2O2. It is possible that ionic liquidsmay have some advantages in this regard,21 but this remains to be seen.

The experiment in Figure 5 was done at low discharge and charge current densities of j≈0.1mA/cm2. Figure 6 (andmany related experiments in the literature, e.g., ref 19) demonstrates a decrease in cell capacity with increasing discharge current. Even at 1mA/cm2, current densities are significantly lower than those used in typical Li-ion batteries. This cell was similar to that described for Figure 5 except that Ketjenblack EC-600JD

(Ketjenblack EC-600JD is an electroconductive carbon black and is a product of AkzoNobel) carbon particles were used to construct the cathode without any added catalyst particles, and the electrolyte was 1M LiPF6 in propylene carbonate. Note that the discharge capacity at i=0.1mA in this cell is∼3 times higher than that for the Super P in Figure 5. There have been manystudies in the literature of the Li-O2discharge capacityof cathodesmade from different kinds of carbon at low discharge current densities (e.g., ref 22). There are conflicting conclusions as to what is most important: surface area, porosity, pore volume, and so forth.22-25 While undoubtedly all aspects are important in some way, we believe that the more important issue is why the low current capacity is limited to only a very small fraction of the overall void volume.

Thedecrease in capacitywith ihaspreviouslybeendiscussed in terms of an O2 transport limitation where O2 diffusion through the cathode flooded with electrolyte cannot sustain the electrochemical reaction rate.26,27 Thus, the electrochemical reaction occurs in a progressively smaller region close to the air-electrolyte interface as the current increases, possibly lead- ing to pore clogging, which is schematically represented in Figure 7. In principal, techniques used in PEM fuel cell cathodes to enhance oxygen transport could also be used in Li-air batteries. In fuel cells, a thin ion transport film covers the pores of an otherwise O2 gas-filled cathode so that fast gas-phase O2 diffusion throughout thecathodestructure isachieved.However, cell-level transport modeling combined with measurements on flat electrodes and flooded porous cathode cells suggests that electron transport through the Li2O2 deposit is also an impor- tant reason for capacity loss. This will be reported in detail elsewhere.28 An electron conductivity issue is not surprising since pure Li2O2 is an insulator, and its buildup during discharge could throttle the current. Obviously, the capacity loss with i is a serious problem in trying to obtain simultaneously high power density and retain high gravimetric energy density.

Figure 6 also demonstrates that increasing i also reduces battery output voltage (i.e., overpotential ηdis increases). Since

This result also implies that the electrochemical stability of both the electrolyte and the cathode itself are key challenges in the develop- ment of practical Li-air systems.

Figure 6. Discharge curves for an aprotic Li-O2 cell (based on a Ketjenblack cathode) at three currents i: 0.1 mA., 0.5 mA, and 1 mA. The curves give the cell output potential as a function of discharge capacity per gram of carbon in the cathode (mAh g-1). The dashed horizontal line is the thermodynamic standard cell potential U0. The loss in output voltage with increasing i is highlighted by the arrow labeled as ηdis(i), and the loss in capacity is indicated by the arrow as capacity (i). Details of the cell are given in the text.

r 2010 American Chemical Society 2199 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

ηdis increases logarithmically rather than linearly with increas- ing i,29 it is evident that the loss in output voltage is not simply the result of an ohmic iR drop due to some internal cell resis- tance. Rather, this implies that the charge transfer at the elec- trodes is rate limitingandhence theprincipal originofηdis. Since the anode reaction is known to be extremely fast, ηdis must be related to some kinetic activation barrier in the cathode chem- istry. The increase in ηdis with i makes achieving reasonable power density from Li-air batteries difficult. Effective catalysts that reduce the activation barriers for the discharge reaction must be discovered to minimize ηdis for high i. We do note that the use of Au nanoparticles does reduce ηdis at low i, but the effectiveness at higher i has not yet been investigated.18

All experiments above were done at room temperature (∼25 �C). Unpublished results by Read30 indicate that both the capacity and output voltage (at low i) decrease significantly with decreasing temperature in the range þ40 �C to -30 �C. The origin of these temperature dependences is not yet under- stood. Since propulsion batteries require a reasonable tempera- ture range, this is an issue that will certainly require further study.

Understanding the detailed mechanism of the electroche- mical Li-O2 reaction and the origin of the existing overpoten- tials is a key first step in trying to minimize overpotentials.31 In suggesting a possible reaction path for the proposed mechan- ismLiþO2T Li2O2, a fewexperimental facts are summarized: (a) the product of the discharge reaction, Li2O2, is insoluble in the aprotic solvents used (e.g., propylene carbonate), (b) after a very brief initial transient drop in voltage, the discharge voltage remains nearly constant throughout the discharge process until the process terminates (Figure5), (c) at the endof thedischarge process,mesoscale particles (∼100nm) of Li2O2 are formed on the carbon cathode and observed by SEM,9,19 and (d) when using small surface area flat glassy carbon electrodes,∼40 nm- thick deposits electrically passivate the battery.32 These facts all suggest that the dominant electrochemistry being observed during discharge is that of Li2O2 formation on the surface of Li2O2 (and its dissociation for charge).

On this basis, a mechanism31 has been suggested invol- ving the following steps in the electrochemical oxidation of

bulk metallic lithium:

LiT Liþ þ e- ðLi electrodeÞ

Liþ þ e- þO2 þ� f LiO2� ðO2 cathode 1Þ

Liþ þ e- þ LiO2� f Li2O2 ðO2 cathode 2Þ An “*” denotes a surface site on Li2O2 where the growth

proceeds. It is essentially a neutral Li vacancy in the Li2O2 surface. The first reaction step, the dissolution of metallic lithium into the electrolyte, is assumed to be in equilibrium. Chemistry at the cathode is described as electrochemical growth of Li2O2 on Li2O2 deposits. This describes all battery chemistry except the initial nucleation of Li2O2 in the cathode structure. Implicit in this mechanism is that there must be at least some electrical conductivity from the carbon support through the growing Li2O2 film to the surface. While Li2O2 itself is an insulator, the existence of neutral Li vacancies is predicted via density functional theory (DFT) to generate some electrical conductivity.31 Since the intermediate state in the oxidation, LiO2* is a Li vacancy, one expects at least some surface conductivity and perhaps even some low bulk electrical conductivity if someof these vacancies persist in the bulkas the filmgrows.However,we certainlydonot anticipate high electrical conductivity of the native electrochemically grown Li2O2 film.

The theory of electron transfer reactions at surfaces is complex. However, trends in electrocatalysis, where species are stronglybound to theelectrode surface, arequalitativelywell described in terms of the thermodynamics of all the surface bound species in the electrochemical reaction steps, and DFT gives reasonable estimates of the free energies of all surface species and intermediates. Crystal growth generally occurs at kinks or steps on the most stable surface. Figure 8 shows the calculated free energy diagram for the growth at a step site on the most stable Li2O2(100) surface.

31 For this surface, the processes described abovemust occur twice for it to reproduce itself in growth. The effect of an applied potential U is included by adding an energy, -neU, to the calculated free energy for every intermediatewith n electrons in the electrode. ForU=0, the overall formation of Li2O2 is strongly exothermic, giving an equilibrium potential of U0= 2.47 V from the Nernst equation. This value is somewhat smaller than that derived from known standard thermodynamic data (2.96V). At a potentialU=2.03 V, all discharge steps are downhill (ΔG e 0). Thus, the over- potential for discharge can be qualitatively identified as ηdis = 2.47V- 2.03V=0.44V. In a similarmanner, all reaction steps are downhill (ΔGe0) for the charge reaction atUg3.07V. This gives ηchg = 3.07 V – 2.47 V = 0.60 V. While ηchg > ηdis, the even larger asymmetry observed experimentally for the over- potentials does not seem to be fully explained by these charge transfer steps, and its origin is still a mystery. This is of con- siderable importance if catalysts are tobedesigned to reduce the charging overpotential. In fact, given the mechanism described above, it is unclear how traditional electrocatalysis mechanisms can be effective in lowering barriers when the reaction product and intermediates are insoluble in the electrolyte.

Figure 7. Schematic representation of the air cathode and proposed chemistry at the air cathode. Various species are as labeled in the diagram. The left side of the diagram shows the electrolyte (with Liþ

ions), the porous carboncathode floodedwith the electrolyte, catalyst particles, and the product Li2O2. The expanded version shows more details of the discharge reaction in the cathode.

r 2010 American Chemical Society 2200 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

Laorie et al.33 havediscussed analternativemechanism for the discharge reaction. Instead of the mechanism discussed above, they suggest a disproportionation reaction of the intermediate, i.e., LiO2* þ LiO2* f Li2O2 þ O2. However, the free energy change ΔG for this reaction is ∼4 eV less exorgic than the last two downhill steps, so that the dispro- portionation mechanism is highly unlikely.

The Lithium Anode. Low specific capacity anodes, such as LiC6, are quite acceptable for use in Li-ion batteries, as the weight of the cathode and electrolyte dominates the battery’s overallmass and therefore energydensity. However, with lightweight Li-air cathodes replacing heavy inter- calation cathodes, a high specific capacity of the anode becomes much more important. Although metallic lithium has an extraordinarily high specific capacity, various lithium insertion/alloying materials are also possible for use as high-capacity anodes. We discuss below some of the issues associated with metallic lithium as the anode and then present current research activities to help circumvent these issues.

Lithium dendrite/moss formation upon battery cycling, which can eventually lead to shorts between the anode and cathode, has plagued the development of a long-lifetime sec- ondary battery based on lithiummetal.34,35Dendrite formation inmetal plating is caused by uneven current distributions at the metal-electrolyte interface. When lithium is immersed in an organic solvent, it spontaneously and almost immediately reacts to form a thin Li-ion conductive film on its surface. As the reaction between lithium and the solvent proceeds, a multilayer deposition of lithium salts creates a mass diffusion barrier between the lithium and solvent, inhibiting the reaction kinetics between the two and preventing further corrosion of the lithiummetal.16 This passivation layer is thewell-knownSEI and is schematically shown in Figure 3. The chemical hetero- geneity of the SEI can result in a brittle and morphologically heterogeneous structure, which can accumulate defects upon

repeated lithium stripping and plating during cycling. These defects can produce nonuniform current distributions through- out the SEI, leading to preferential deposition of lithium metal and, therefore, dendrite formation.16,36,37

In order to improve the stability of the lithium anode, several approaches involving homogeneous, highly Li-ion- conductive artificial protective layers have been developed to reduce dendrite formation during battery cycling. For example, polymer electrolytes remain a promising direction for minimizing the problems of lithium dendrite formation. For example, Seeo, Inc. is currentlydeveloping electrolytes for lithiummetal-based batteries based on novel di- and triblock copolymers.38 Their electrolytes combine the mechanical stability of a hard polymer segment, such as polystyrene, with the high Li-ion conductivity of a soft polymer segment, such as a poly(ethylene oxide) (PEO)/Li-salt mixture. These polymers will hopefully inhibit dendrite shorts (via mechan- ical blocking from the hard polymer segments) while still maintaininghigh ion conductivity. Alternative approaches use Li-ion conducting glasses or ceramic materials as the solid state electrolyte. For example, Bates et al. produced a thin film solid state battery, based on Li/LiCoO2, and a lithium phos- phorus oxynitride (LIPON) electrolyte that could be cycled 4000 times at low currents with only 2% lifetime capacity fade.39 For Li-air batteries, the best results to date use thin films of Li-ion-conducting ceramics (lithium superionic con- ductor (LISICON)-type materials) to encase lithium metal.10

This approach is shown in Figure 3 (labeled the artificial SEI). Generally, the Li-ion-conducting ceramic is readily reduced by lithium metal, and therefore a thin film of a lithium stable conducting material, e.g., Li3N or Li3P, must be inserted between the ceramic and metal.

Another significant advantage of the ceramic based artifi- cial SEI is that it protects the lithium metal from all atmo- spheric contamination. An example of the remarkable barrier properties of this type of ceramic was demonstrated by Visco et al.,15 who successfully used this protective ceramic to enable a primary Li-seawater battery. The strong barrier properties of the ceramics are also essential for enabling the aqueous, the mixed aqueous-aprotic, and the all-solid-state electrolyte architectures of Li-air batteries discussed in sec- tion III and Figure 2. A pure-polymer artificial SEI is unlikely to achieve the same level of contamination protection for the lithium for reasons discussed in section VI.

Although the ceramic Li-ion conductors offer excellent contamination isolation of the lithium metal, they are brittle and can only be formed reproducibly in small sheets (on the order of a few square inches). These sheets are typically relatively thick (on the order of a few hundred micrometers) to instill mechanical integrity and therefore limit power density because of the iR voltage drop at higher currents. Thinner protective layerswould lead to higher Li-ion conducti- vity and better power densities. In addition, the brittle nature of these materials could potentially result in barrier defects (cracks) upon charge-discharge cycling, leading to parasitic reactions at the anode surface. Kumar et al. have begun to investigate ceramic polymer composites for the lithiummetal barrier in their all-solid-state Li-air battery.12 In principle, such a composite barrier could produce advantages of both

Figure 8. Free energy diagram calculated from DFT for the electro- chemical reactionproposed at the cathode for the Li-air battery. Two formula units are added during discharge (left to right) or removed during charging (right to left). The free energies are shownatdifferent potentials: U = 0 is the open circuit potential, U = 2.03 V is the highest potential for which discharge is energetically downhill for all steps, and U = 3.07 V is the lowest potential where charging is energetically downhill for all steps. From ref 31.

r 2010 American Chemical Society 2201 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

polymers (excellent process ability) and ceramics (excellent barrier properties).

Given the safety issues associated with cycling metallic lithium-based batteries, major research efforts have focused on producing a variety of safer anode alternatives and led to the development of the first commercially available Li-ion battery by Sony in 1990. Current state-of-the-art LiC6 intercala- tion anodes have been engineered to provide very high cyclabil- ity and safety. Similarly, insertion anodes comprised of lithium titanate (Li4Ti5O12) have been considered for automotive appli- cations as a result of their acceptable high-rate performance. Furthermore, lithium titanate’s lithiated and delithiated cell structure is almost identical,making it verysafeduringcycling.40

Although insertion anodes exhibit reliable cyclability, their spe- cific capacities are an order of magnitude lower than that of metallic lithium. However, lithium alloying provides a route to anodeswith specific capacities similar to that ofmetallic lithium. Unfortunately, these compounds are severely limited for prac- tical use as a result of the dramatic volume changes associated with the alloying process, which eventually leads to degradation of the anode.

Li-Air vs Li-O2 Batteries: Membranes. Most experi- ments to date on rechargeable aprotic batteries have em- ployed O2 rather than air to avoid unwanted parasitic reac- tions with components, such as water, carbon dioxide, carbon monoxide, andnitrogen, in ambient air. Thus, removal of these components remains a significant challenge that must ulti- mately be overcome to allow Li-air battery operation for repeated cycling and long-term use. One such method envi- sioned to deliver O2 or highly O2-enriched gas from air to a Li-O2 cell has been an “oxygen-diffusion” membrane that selectively permeates O2 while retaining other gaseous spe- cies.41 However, the incorporation of such a membrane re- mains a critical challenge for high-rate applications, as water vapor is much more permeable than O2 through any mem- brane, regardless of the membrane’s physical properties (including its relative hydrophobicity).42 Table 1 presents variousphysical parameters that serve as reasonablemeasures of a gas’s condensability (Tc) and molecular size (Vc, kinetic diameter). Of particular interest, water molecules are much smaller than oxygenmolecules, indicating that preferential O2 permeability over water through porous membranes, where gas separation is controlled by size exclusion mechanisms such as Knudsen diffusion and Pouseille flow, is not possible.

Additionally, gas separation through nonporous polymeric membranes is governed by the solution/diffusion mechanism, where permeability of any gas through a polymer is the product of its ability to dissolve into a polymer (i.e., its solubility) and its ability to diffuse through the polymer (i.e., its diffusion rate).42

In other words: P = DS, where P is a gas’s permeability coefficient through a nonporous polymer film, D is its diffusion coefficient, and S is its solubility coefficient. For light gases, such as those listed in Table 1.1, a gas’s diffusion coefficient through a polymer film is generally inversely correlated to its kinetic diameter raised to apositive power, and therefore,water usually has a higher diffusion coefficient than O2 through polymeric membranes. Furthermore, an Arrhenius-type correlation exists between a gas’s solubilility coefficient and its condensability in any polymer.42 As a result,water,whose condensability ismuch higher than that of O2,, will also bemuchmore soluble than O2 in anymembrane. Therefore, water permeability though dense polymer films tends tobesignificantlyhigher (sometimesmany orders of magnitude) than oxygen permeability. Nonetheless, by preferentially permeating water instead of oxygen, mem- branes still hold promise to deliver highly oxygen-enriched, dehydrated air for EV applications. For example, a possible configuration for air dehydration employs a hydrophilic mem- brane with a very high H2O permeability compared to O2 permeability.43 As a result, O2 is retained on the high pressure side of the membrane, where it can be sent to the battery or a further dehydration step. Of course, the size andweight of such a system is critical when considering the Li-air battery energy density; however, because of water’s high permeability rate, a compact dehumidification membrane system is potentially an attainable goal.

While preliminary experiments on Li-air batteries do suggest the promise of being able to provide rechargeable batteries with much higher gravimetric energy densities than currently available, there are challenges that must be over- comebefore this promise canbecome a reality. At present,we limit the consideration to an aprotic liquid electrolyte since this is the only version that has been suggested to be electrically rechargeable.

We list below the key research that in our opinion is necessary for the development of a practical electrically rechargeable Li-air battery and its subsequent commerciali- zation.

1 Quantitative understanding of the electrochemical re- actions and their relationship to the discharge/charge currents. This is the key to quantitatively demonstrating chemical reversibility and understanding Coulombic efficiency of the battery in cycling.

2 Development of oxidation-resistant electrolytes and cathodes that can withstand high oxidation potentials in the presence of O2. This is also essential for chemical reversibility and Coulombic efficiency in the battery cycling.

3 Understanding the nature of electrocatalysis for Li-air batteries where insoluble products are formed and the development of cost-effective catalysts to reduce overpotentials for the discharge and charge reactions. This is key to enhancing power density in discharge, electrical efficiency in a discharge-charge cycle, and ultimately in cycle life (due to possible electrolyte oxidation).

4 Development of new nanostructured air cathodes that optimize transport of all reactants (O2, Li

þ, and electrons) to the active catalyst surfaces and

Table 1. Physical Properties of Common Air Gases42

gas critical

temperature, Tc [K] critical volume, Vc [cm

3/mol] kinetic

diameter, dk [Å]

O2 154.6 73.5 3.46

N2 126.2 89.3 3.64

H2O 647 55.9 2.65

CO 134.5 90.1 3.76

CO2 304.2 91.9 3.3

r 2010 American Chemical Society 2202 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

provide appropriate space for solid lithium oxide products. This is required to maintain capacity at higher power densities. A new realization is that mini- mizing difficulties due to electron transport through the lithium oxide solid products in the cathode is important.

5 Development of a robust lithium metal or lithium composite electrode capable of repeated cycling at higher current densities. This will most likely require development of a protective layer that limits the dele- terious effects of environmental contamination on the lithium and inhibits dendrite growth.

6 Development of high throughput air-breathing mem- branes (or other mechanisms) that separate O2 from ambient air in order to avoid H2O, CO2, and other environmental contaminants from limiting the lifetime of Li-air batteries.

7 Understanding the origin of the temperature depend- ences in Li-air batteries and minimizing their adverse effects.

AUTHOR INFORMATION

Corresponding Author: *To whom correspondence should be addressed.

Biographies Dr. G. Girishkumar is currently an electrochemist at the IBM

Almaden Research center. He obtained his Ph.D. (2003) degree from the Indian Institute of Science. After completion, he was selected for a post doctoral position at the Notre Dame Radiation Laboratory. Subsequently, hemoved to Air Products and Chemical., Inc. as a Senior Research Scientist, (2005-2009) working in the area of the Stabilife family of electrolyte salts for rechargeable lithium-ion batteries. His present research focuses on electrochem- istry of rechargeable batteries and fuel cells.

Dr. Bryan D. McCloskey joined the IBM Almaden Research Institute as a Post-Doctoral Research Associate in 2009. He received his B.S. at the Colorado School of Mines in 2003 and his Ph.D. in Chemical Engineering at the University of Texas at Austin, where he was a National Science Foundation Graduate Research Fellow in 2009. His research interests focus mainly on molecular transport through polymer membranes.

Dr. A. C. Luntz is currently consulting with IBM Almaden on the Battery 500 project and is also Adjunct Professor of Physics at both the University of Southern Denmark and at Aarhus University. Prior to this, he spent 25 years at Almaden Research Laboratory before leaving for an academic career in Denmark in 1994. Principal current research interests are in surface chemical dynamics, espe- cially related to heterogeneous catalysis.

Sally A. Swanson, an advisory scientist in Nanoscale Science & Technology, has worked for IBM Research – Almaden since 1982. She has over 25 patents on materials for semiconductor manufac- turing, electrophoretic and OLED displays, molecular electronics, and lithography. Her current interests include lithium/air batteries and DNA origami projects.

Dr. Winfried Wilcke is a Program Director at the IBM Almaden Research Center. He holds a Ph.D. in nuclear physics, and has worked at the University of Rochester, LBL, and Los Alamos. He played a key role in IBM’s development of supercomputers, served as CTO of HaL (64-bit Sparc), led the IBM Icecube project, and initiated the Battery500 lithium/air project.

ACKNOWLEDGMENT We thankfully acknowledge all the other members of the IBM Research Almaden and Z€urich teams working on Li-air batteries. Their work and discussion have contributed greatly to the opinions expressed in this perspective. The members are D. Bethune, M. Hart, C. Scott, B. Shelby, M. Sherwood, B. A. Smith, C. Larson, K. Virwani, G.Wallraff, H.-C. Kim, Q. Song, and R. Miller in San Jos�e, California and A. Curioni and T. Laino in Z€urich, Switzerland.

REFERENCES

(1) Richter, B.; Goldston, D.; Crabtree, G.; Glicksman, L.; Goldstein, D.; Greene, D.; Kammen, D.; Levine, M.; Lubell, M.; Savitz, M.; Sperling, D. Energy Future: Think Efficiency; American Physical Society: College Park, MD, 2008.

(2) Mackay, D. J. Sustainable Energy – Without the Hot Air; UIT Cambridge, Ltd: Cambridge, U.K., 2009.

(3) Kinoshita, K. Metal/Air Batteries. In Electrochemical Oxygen Technology; Kinoshita, K., Ed.; John Wiley & Sons, Inc.: New York, 1992; pp 259-306.

(4) Symons, P.; Butler, P. Introduction to Advanced Batteries for Emerging Applications. Sandia National Laboratory Report SAND2001-2022P, April 2010. Available at http://infoserve. sandia.gov/sand_doc/2001/012022p.pdf.

(5) Ogasawara, T.; D�ebart, A.; Holzapfel, M.; Nov�ak, P.; Bruce, P. G. Rechargeable Li2O2 Electrode for Lithium Batteries. J. Am. Chem. Soc. 2006, 128 (4), 1390–1393.

(6) Abraham, K. M.; Jiang, Z. A Polymer Electrolyte-Based Re- chargeable Lithium/Oxygen Battery. J. Electrochem. Soc. 1996, 143 (1), 1–5.

(7) Read, J. Ether-Based Electrolytes for the Lithium/Oxygen Organic Electrolyte Battery. J. Electrochem. Soc. 2006, 153 (1), A96–A100.

(8) D�ebart, A.; Bao, J.; Armstrong, G.; Bruce, P. G. An O2 Cathode for Rechargeable Lithium Batteries: The Effect of a Catalyst. J. Power Sources 2007, 174 (2), 1177–1182.

(9) D�ebart, A.; Paterson, A. J.; Bao, J.; Bruce, P. G. R-MnO2 Nanowires: A Catalyst for the O2 Electrode in Rechargeable Lithium Batteries.Angew. Chem., Int. Ed. 2008, 47 (24), 4521– 4524.

(10) Visco, S. J.; Katz, B. D.; Nimon, Y. S.; DeJonghe, L. C. Protected Active Metal Electrode and Battery Cell Structures with Non- aqueous Interlayer Architecture. U.S. Patent #7282295, 2007.

(11) Wang, Y.; Zhou, H. A Lithium-Air Battery with a Potential to Continuously Reduce O2 from Air for Delivering Energy. J. Power Sources 2010, 195 (1), 358–361.

(12) Kumar, B.; Kumar, J.; Leese, R.; Fellner, J. P.; Rodrigues, S. J.; Abraham, K. M. A Solid-State, Rechargeable, Long Cycle Life Lithium-Air Battery. J. Electrochem. Soc. 2010, 157 (1), A50–A54.

(13) Kowalczk, I.; Read, J.; Salomon, M. Li-Air Batteries: AClassic Example of Limitations Owing to Solubilities. Pure Appl. Chem. 2007, 79 (5), 851–860.

(14) Peled, E.; Golodnitsky, D.; Ardel, G. AdvancedModel for Solid Electrolyte Interphase Electrodes in Liquid and Polymer Electrolytes. J. Electrochem. Soc. 1997, 144 (8), L208–L210.

(15) Visco, S. J.; Nimon, Y. S. Active Metal/Aqueous Electrochemi- cal Cells and Systems. U.S. Patent #7645543, 2010.

(16) Aurbach, D. Review of Selected Electrode-Solution Interac- tions Which Determine the Performance of Li and Li Ion Batteries. J. Power Sources 2000, 89 (2), 206–218.

(17) Xu, K. Nonaqueous Liquid Electrolytes for Lithium-Based Re- chargeable Batteries. Chem. Rev. 2004, 104 (10), 4303–4418.

r 2010 American Chemical Society 2203 DOI: 10.1021/jz1005384 |J. Phys. Chem. Lett. 2010, 1, 2193–2203

PERSPECTIVE pubs.acs.org/JPCL

(18) Lu, Y.-C.; Gasteiger, H. A.; Parent, M. C.; Chiloyan, V.; Shao- Horn, Y. The Influence of Catalysts on Discharge and Charge Voltages of Rechargeable Li-Oxygen Batteries. Electrochem. Solid-State Lett. 2010, 13 (6), A69–A72.

(19) Read, J. Characterization of the Lithium/Oxygen Organic Electrolyte Battery. J. Electrochem. Soc. 2002, 149 (9), A1190–A1195.

(20) McCloskey, B. et al. Work to be submitted for publication, 2010.

(21) Kuboki, T.; Okuyama, T.; Ohsaki, T.; Takami, N. Lithium-Air Batteries Using Hydrophobic RoomTemperature Ionic Liquid Electrolyte. J. Power Sources 2005, 146 (1-2), 766–769.

(22) Yang, X.-h.; He, P.; Xia, Y.-y. Preparation of Mesocellular Carbon Foam and Its Application for Lithium/Oxygen Bat- tery. Electrochem. Commun. 2009, 11 (6), 1127–1130.

(23) Beattie, S. D.; Manolescu, D. M.; Blair, S. L. High-Capacity Lithium-Air Cathodes. J. Electrochem. Soc. 2009, 156 (1), A44–A47.

(24) Xiao, J.; Wang, D. H.; Xu, W.; Wang, D. Y.; Williford, R. E.; Liu, J.; Zhang, J. G. Optimization of Air Electrode for Li/Air Batteries. J. Electrochem. Soc. 2010, 157 (4), A487–A492.

(25) Tran, C.; Yang, X. Q.; Qu, D. Y. Investigation of the Gas- Diffusion-Electrode Used As Lithium/Air Cathode in Non- aqueous Electrolyte and the Importance of Carbon Material Porosity. J. Power Sources 2010, 195 (7), 2057–2063.

(26) Read, J.; Mutolo, K.; Ervin, M.; Behl, W.; Wolfenstine, J.; Driedger, A.; Foster, D. Oxygen Transport Properties of Organic Electrolytes and Performance of Lithium/Oxygen Battery. J. Electrochem. Soc. 2003, 150 (10), A1351–A1356.

(27) Sandhu, S. S.; Fellner, J. P.; Brutchen, G. W. Diffusion-Limited Model for a Lithium/Air Battery with an Organic Electrolyte J. Power Sources 2007, 164 (1), 365–371.

(28) Albertus, P. Physics-BasedModeling of the Li/Air Battery, with Comparison to Experiments. ECS Meeting Abstracts, 2010 (to be published).

(29) Yang, X.-H.; Xia, Y.-Y. The Effect of Oxygen Pressures on the Electrochemical Profile of Lithium/Oxygen Battery. J. Solid State Electrochem. 2010, 14 (1), 109–114.

(30) Georgi, D. Lithium Primary Continues to Evolve. Available at http://www.batteriesdigest.com/lithium_air.htm (accessed April 15, 2010).

(31) Hummelshoj, J. S.; Blomqvist, J.; Datta, S.; Vegge, T.; Rossmeisl, J.; Thygesen, K. S.; Luntz, A. C.; Jacobsen, K. W.; Norskov, J. K. Communications: Elementary Oxygen Electrode Reactions in the Aprotic Li-Air Battery. J. Chem. Phys. 2010, 132 (7), 071101.

(32) Girishkumar, G. et al. Work to be submitted for pulication, 2010.

(33) Laoire, C. O.; Mukerjee, S.; Abraham, K. M.; Plichta, E. J.; Hendrickson, M. A. Elucidating the Mechanism of Oxygen Reduction for Lithium-Air Battery Applications. J. Phys. Chem. C 2009, 113 (46), 20127–20134.

(34) Whittingham, M. S. Electrical Energy Storage and Intercala- tion Chemistry. Science 1976, 192 (4244), 1126–1127.

(35) Brandt, K. Historical Development of Secondary Lithium Batteries. Solid State Ionics 1994, 69 (3-4), 173–183.

(36) Cohen, Y. S.; Cohen, Y.; Aurbach, D. Micromorphological Studies of Lithium Electrodes in Alkyl Carbonate Solutions Using In Situ Atomic ForceMicroscopy. J. Phys. Chem. B 2000, 104 (51), 12282–12291.

(37) Choi, N.-S.; Lee, Y. M.; Cho, K. Y.; Ko, D.-H.; Park, J.-K. Protective Layer with Oligo(ethylene glycol) Borate Anion Receptor for Lithium Metal Electrode Stabilization. Electro- chem. Commun. 2004, 6 (12), 1238–1242.

(38) Singh, M.; Gur, I.; Eitouni, H. B.; Balsara, N. P. Solid Electrolyte Material Manufacturable by Polymer Processing Methods. U.S. Patent Application #12271829, 2009.

(39) Bates, J. B.; Dudney, N. J.; Neudecker, B.; Ueda, A.; Evans, C. D. Thin-Film Lithium and Lithium-Ion Batteries. Solid State Ionics 2000, 135 (1-4), 33–45.

(40) Palacín, M. R. Recent Advances in Rechargeable Battery Materials: A Chemists Perspective. Chem. Soc. Rev. 2009, 38, 2565–2575.

(41) Zhang, D.; Li, R.; Huang, T.; Yu, A. Novel Composite Polymer Electrolyte for Lithium Air Batteries. J. Power Sources 2010, 195 (4), 1202–1206.

(42) Matteucci, S.; Yampolskii, Y.; Freeman, B. D.; Pinnau, I. Transport of Gases and Vapors in Glassy and Rubbery Polymers. In Materials Science of Membranes; Yampolskii, Yu., Pinnau, I. , Freeman, B. D., Eds.; JohnWiley & Sons: New York, 2006.

(43) Baker, R. Future Directions of Membrane Gas-Separation Technology. Membrane Technology; Elsevier: Chichester, U.K., 2001; Vol. 200.


Comments are closed.